Datasets:

disi-unibo-nlp

/

SciLay

like 0

Follow

DISI UniBo NLP 11

10.3390/ani12020152

PMC8772571

Creation of new meat-based pet food and pet treats continues to grow at a steady annual rate within the pet food industry. Poultry co-products are often overlooked due to their poor quality and low customer acceptance. However, poultry co-products pose great potential and added value to the pet food industry. Two of the most common poultry co-products (wooden breast and carcass frames) often directed towards pet food were used in a fresh pet food formulation. Due to variations in meat quality because of the wooden breast and carcass frames, a hydrocolloid was utilized to improve fresh pet food characteristics. A hydrocolloid is a type of protein that when added to meat products aids with binding and stabilization of the pet food. For the current study, the combination of sodium alginate and encapsulated calcium lactate pentahydrate (ALGIN) was used. Due to the perceived poor quality of wooden breast and carcass frames, it is plausible that the addition of hydrocolloids can combat the undesirable characteristics. Results from the current study suggest that the impact of ALGIN in poultry co-product pet food combinations does not severely alter shelf-life characteristics of a fresh pet food. However, the inclusion of varying amounts of wooden breast and ground carcass frame can impart a greater impact on shelf-life characteristics in fresh pet food by altering surface color and lipid oxidation.

Poultry co-product chicken frames (CF) and wooden breast (WB) along with ingredient technology use may bring enhanced value to the pet food industry. Therefore, the current study focused on evaluating CF and WB combinations along with sodium alginate and encapsulated calcium lactate pentahydrate (ALGIN) inclusion within a fresh pet food formulation under simulated shelf-life conditions. Fresh chicken frames (CF) and boneless-skinless wooden breast (WB) were ground and allocated randomly to one of ten treatment combinations with either 0.5 or 1.0% added ALGIN. Ground treatments were placed into a form and fill vacuum package and stored using a reach-in refrigerated case for 21 days. Packages were evaluated for instrumental surface color, lipid oxidation, water activity, and pH on days 1, 3, 7, 14 and 21 of the display. Packages of pet food were lighter, less red, and more yellow (p < 0.05) with increasing percentages of CF regardless of ALGIN inclusion, whereas pH was greater (p < 0.05) and lipid oxidation was less (p < 0.05) with increasing percentage of WB. Water activity increased (p < 0.05) when WB and ALGIN inclusion increased. The current results suggest that the use of ALGIN in a poultry co-product pet food formulation can improve shelf-life characteristics such as surface color and lipid oxidation in fresh pet food.

1. IntroductionCreative development and availability of meat-based food and treats for pets in the retail space continue to increase at a rapid pace. According to American Pet Products Association, sales within the U.S. Market in 2020 for pet food and treats alone was $42.0 billion and the estimated sales for 2021 is to be $44.1 billion [1]. Due to the variety of pet food options, pet food customers are provided with an almost limitless number of options based on packaging, price, size of the product, ingredient preference, nutritional benefits, and pet breed options. A focus of the pet food industry has traditionally relied on animal by-products from the rendering industry. All animal material that is considered inedible for human consumption is further processed for animal agriculture purposes such as fertilizer, animal feed and pet food treats and/or food. Type of animal by-products that are rendered include offal, fat, blood, bones, meat trimmings and viscera. Within the poultry meat industry, inedible by-products account for approximately 28% of the live weight of a broiler chicken [2]. Poultry co-products are often undervalued throughout the meat industry due to their poor quality and low customer acceptance. However, it is plausible that these co-products could be a potential key component in adding value to the pet food industry.The poultry industry has struggled with a muscle myopathy identified as wooden breast (WB) resulting in tough muscle texture. This muscle abnormality has and continues to have a negative impact economically on the broiler industry [3]. WB is characterized as a myopathy that presents an overtly harder chicken breast, with sections being swollen and pale resulting in decreased meat quality, yield, and consumer/customer acceptance [4]. Initially, visual characterization of WB was identified by the white striations throughout the breast meat [5] because of excessive accumulation of lipids and connective tissue [6]. Severely affected chicken breasts caused by muscle fiber degeneration have been downgraded by the processor and directed to be further processed within the meat and poultry industry that may include deli meats, sausages, emulsified products, and even pet food. Pet food buyers are influenced by similar organoleptic attributes when buying food for their pets such as aroma, texture, and color. In addition, uniformity and branding are important drivers that may suggest higher quality. Studies have reported that pet food customers are highly influenced by the appearance of the pet food or treat and a darker brown color seems to influence pet owner purchasing intent [7]. Another poultry by-product gaining consideration for use in possible pet food formulations is chicken frames (CF) because of it is nutritional properties [8]. However, when nutritionally evaluating chicken meat and bone meals, these co-products consist of a greater protein level (50 percent) and contain more saturated fatty acids when compared to other rendered meat [9].Due to the low quality of WB and CF, it is plausible that the addition of hydrocolloids to these meat co-products may aid in creating more desirable characteristics making the co-products suitable for fresh pet food and pet treat applications. Hydrocolloids are used to define a range of proteins and polysaccharides that contain an affinity for water. Hydrocolloids are used throughout the meat and food industry to aid in a variety of formulation functions that include gelling, binding, coating, thickening, stabilization of pH, enhanced heat resistance, salt tolerance, as well as reduce undesirable effects caused when reducing fat and salt content [10,11,12]. Interestingly, gelling of food products via hydrocolloids allows for the product to become more stable [13]. Out of the wide range of hydrocolloids within the meat and food industry today, the combination of sodium alginate and encapsulated calcium lactate pentahydrate (ALGIN) was selected for the current study. Sodium alginate has been commonly used within the food industry in the development of functional food products for human and companion animal consumption. Sodium alginate is an irreversible hydrocolloid that is derived from brown algae, a polysaccharide that is composed of 1–4, β-D mannuronic acid and α-L guluronic acid sugar residues [14,15]. An advantage of using the algin/calcium gel combination, has reported improvements in binding cooked or raw meat products and can prevent the potential negative effects of added salt on product quality [16,17,18]. To our knowledge, there is very little literature about the effects of ALGIN when included in poultry co-product formulations such as CF and WB with the intended purpose of creating pet food products. Therefore, the aim of this study was to evaluate the inclusion rate of ALGIN on meat co-product formulations of CF and WB and subsequent impact on fresh pet food shelf-life characteristics.2. Materials and Methods2.1. Raw MaterialsBoneless-skinless wooden breast (WB) meat and de-boned chicken frames (CF) were purchased from a commercial poultry processing facility in North Alabama. Classification of WB was conducted by trained plant personnel at the time of grading using the scoring methods described in [19]: (0 = normal; 1 = moderate, 2 = severe; and 3 = extreme). Fresh, raw materials were transported under refrigerated conditions (2 °C) to the Lambert-Powell Meats Laboratory at Auburn University and stored in the absence of light at 2 °C for 36 h. Fresh, raw poultry materials WB and CF were ground twice using a commercial meat grinder (Hollymatic 3000, Thompson Meat Machinery, Queensland, Australia). To aid in reducing the surface area of the raw materials, CF was ground once through a 9.52 mm grinder plate then a second time through a 4.76 mm grinder plate (SPECO 400, Schiller Park, IL, USA). Fresh boneless-skinless WB meat was ground twice through a 4.76 mm grinder plate (SPECO 400, Schiller Park, IL, USA). After grinding, ground WB and CF were weighed and randomly allocated to 10 treatment batches (N = 22.67 kg/treatment) with two replications (n = 11.34 kg/replication). Treatment combinations (Table 1) of WB and CF (COMB) were mixed in a commercial mixer (AFMG 48, Biro, Marblehead, OH, USA) for 5 min while slowly incorporating sodium alginate and calcium lactate (ALGIN). After mixing, each treatment was portioned into 454 ± 0.5 g bricks using a vacuum stuffer (Model VF608plus, Handtmann, Biberach, Germany).2.2. Treatment PackagingAfter portioning, fresh pet food bricks were packaged using a Reiser form and fill vacuum packaging machine (Optimus OL0924, Variovac, Zarrentin, Germany). Each package was placed into a commercial film (WINPAK, Winnipeg, MB, Canada) that was comprised of a forming film with a standard barrier consisting of 175 μm Nylon, EVOH and enhanced polyethylene coextrusion and a non-forming film layer was comprised of 60 μm Nylon, EVOH and polyethylene coextrusion. The oxygen transmission rates (OTR) for the forming and non-forming films were 0.4 cc/sq. m/24 h and 1.2 cc/sq. m/24 h, respectively.2.3. Simulated Retail Display ConditionsPackages of fresh pet food were stored in a pull-door, self-service refrigerated 3 tiered LED lighted display case (Model 178GDC49HCB, Avantco Refrigeration, Lancaster, PA, USA) for up to 21 days. Storage temperatures during the simulated display were monitored using a data-recording device (TD2F, Thermoworks, American Fork, UT, USA) with probes placed within the center of each shelf. Refrigerated case temperatures averaged 2.1 ± 1.25 °C. The continuous LED (308 lux) lighting of each shelf was measured in the display cases with a hand-held light meter (Model ILT10C, International Light Technologies, Peabody, MA, USA). Packages of pet food were distributed evenly across the shelving and rotated daily from top to bottom and front to back within the display cooler to reduce temperature variation and simulate consumer package shifting at the retail counter.2.4. Instrumental Color MeasurementInstrumental color readings were collected by scanning each sample through the packaging [20], at three separate locations on each package using a HunterLab MiniScan XE Plus Colorimeter (Model 45/0-L, Hunter Associates Laboratory Inc., Reston, VA, USA). Color readings (L*, a* and b*) were recorded using Illuminant D65, a 10° observer with a 25 mm diameter aperture using the Commission Internationale de l’ Eclairage (CIE L*a*b*) color scale [21]. Prior to capturing objective surface color readings for fresh color, the colorimeter was calibrated on each sampling day 1, 3, 7, 14, and 21 using the standard black and white tiles (L*, 0 = black, 100 = white; a*, −60 = green, +60 = red; b*, −60 = blue, and +60 = yellow).2.5. Thiobarbituric Acid Reactive Substances (TBARS)On days 1, 3, 7, and 14 of the simulated display period, packages of fresh pet were removed from the display cases and frozen at −80 °C until TBARS analysis could be completed. Prior to analysis, samples were placed into a refrigerated cooler 4 °C and thawed for 12 h. After thawing, pet food was removed from the packaging material and prepared for 2-thiobarbituric acid reactive substance using a modified version [21]. In duplicate, approximately 5 g of each package was homogenized with 8 mL of cold (1 °C) of 50 mM phosphate buffer (pH of 7.0 at 4 °C) that contained 0.1% ethylenediaminetetraacetic acid, 0.1% n-propyl gallate, and 2 mL trichloroacetic acid (Sigma-Aldrich, Saint Louis, MO, USA). After homogenizing, samples were filtered through Whatmann No.4 filter paper and duplicate 2 mL aliquots of the clear supernatant were transferred into 10 mL borosilicate tubes, mixed with 2 mL of 0.02 M 2-thiobarbituric acid reagent (BeanTown Chemical, Hudson, NH, USA), and boiled at 100 °C for 20 min. Immediately after boiling, tubes were placed into an ice bath for 15 min. Finally, absorbance was measured at 533 nm with a spectrophotometer (Turner Model—SM110245, Barnstead International, Dubugue, IA, USA) and then multiplied using a factor of 12.21 to derive the TBARS value (mg of malonaldehyde/kg of fresh meat) [22]. The value of 12.21 was obtained previously from a standard curve using a known malonaldehyde solution measured across multiple absorbencies.2.6. Fresh pH and Water Activity (aw)Packages of fresh pet food from each treatment were chosen randomly from the display cabinets on days 1, 3, 7, 14, and 21 to be analyzed for pH and water activity (aw). Prior to collecting pH readings, the pH meter was calibrated using 2-point standard buffers (pH 4.0 and 7.0). Duplicate packages from each treatment were opened and using a pH electrode attached to a pH meter (Model HI199163, Hanna Instruments, Woonsocket, RI, USA) was inserted into the ground pet food. Measurements for pH were collected in triplicate from each packaged and averaged. For water activity samples, 4 g of ground pet food was removed from each package in duplicate inserted into the plastic sample container and analyzed using a benchtop water activity meter AUQALAB 4TE (Dew Point Model, METER Group, Inc., Pullman, WA, USA) which uses the dew point principal method [23].2.7. Statistical AnalysisAn analysis of variance was computed using a generalized linear mixed model (GLIMMIX) procedure with statistical analysis software (SAS Institute, Inc. Cary, NC, USA) version 9.4. Fixed effects for ground chicken frames (CF) and wooden breast (WB) and ALGIN inclusion percentage along with their interaction were evaluated. Least-squares means were computed for all variables, and when significant (p ≤ 0.05) F-values were observed, least-squares means were separated using pair-wise t-tests (PDIFF option).3. Results and Discussion3.1. Instrumental Fresh ColorThe combination (COMB) of WB and CF and ALGIN across DAY of simulated display presented an interaction (p < 0.05) for instrumental fresh surface color lightness (Table 2). Surface color of pet food packages was lighter (p < 0.05) with greater percentages of CF regardless of ALGIN or day of display. However, as the duration of display increased, surface color lightness became darkest (p < 0.05) with increasing percentages of WB (Table 2). In addition, an interaction of COMB × ALGIN × DAY occurred for instrumental surface color redness (Table 3). Packages of pet food formulations with greater percentages of WB were redder initially (p < 0.05), whereas packages of CF remained lighter throughout the entire display period regardless of ALGIN inclusion (Table 3). Lastly, an interactive influence of COMB × ALGIN × DAY for surface color yellowness occurred (Table 4). Packages of pet food formulations were more yellow (p < 0.05) throughout the entire display period when the percentage of CF was greatest and ALGIN was only 0.5% within the formulation. Moreover, as the concentration of WB increased, surface b* became greener (p < 0.05). Fresh surface color of pet food remains an enigma within the retail market because the intended user of the food is not visually appraising the product in the same manner fresh edible meat products are assessed. However, it is plausible that purchasers of fresh pet food (pet owners) will continue to use surface color as an indicator of wholesome and freshness of pet food constructed with fresh meat ingredients that can deteriorate during a storage period. In a similar study [24] on surface color differences between cooked and raw WB samples with non-affected boneless, skinless, breast fillets, it was reported that the chicken breasts with severe WB can have greater redness (a*) values, which appears consistent with our findings in the current study. In additional studies, it has been determined that boneless, skinless WB fillets often have more connective tissue and a greater percentage of white striations throughout the filet often causing an altered surface color of the meat [5,24,25,26]. It was expected that the formulations with higher percentages of WB would result in greater lightness (L*) values due to less muscle myoglobin influence and greater hemorrhagic lesions throughout the affected WB meat. However, the current results indicate lightness (L*) values for treatment combinations containing more WB in the formulations darker, redder and greener. A similar reported that WB fillets had significantly greater (p < 0.05) lightness values (L*) and yellowness values (b*), which are inconsistent with our findings [25]. The inconsistency in surface color results may be attributed to the addition of CF, severity of WB within the COMB, but does not suggest that ALGIN imparted surface color changes throughout the simulated display period.3.2. Thiobarbituric Acid Reactive Substances (TBARS)An interaction (p < 0.05) of COMB × ALGIN × DAY for TBARS values occurred during the simulated retail display period (Table 5). As a result of fresh pet food quality declining throughout the display period, TBARS values were only measured through 14 days of the simulated display. TBARS values were greatest (p < 0.05) on day 14 when ALGIN and WB combinations were 1.0 and 100%, respectively. With limited results available from previous studies, the values from the current study provide a foundation to lipid oxidation changes that may occur during storage of fresh pet food. Previous lipid oxidation findings for WB have suggested that oxidation can be variable [27], whereas frozen storage of cooked chicken sausage formulated with WB can range from 0.14 to 2.00 mg malonaldehyde/kg [28]. Countless studies in chicken [29,30] and beef [31,32] suggest TBARS values may exceed 3 mg of malonaldehyde/kg in fresh or cooked meat samples. Current values for TBARS values align with previous studies and provide a baseline for future studies. Consumer perception of meat products for wholesomeness and freshness at the time of use may be partly influenced by lipid oxidation [33]. However, it is necessary that additional research on fresh pet food shelf-life using various ingredient technologies that may or may not improve lipid oxidation be investigated.3.3. Fresh Pet Food pH ValuesThere was an interactive (p < 0.05) effect of COMB × ALGIN × DAY on pH values of fresh pet food during a simulated retail display (Table 6). Pet food containing more than 75% WB in the formulation resulted in greater (p < 0.05) pH values throughout the 21 day simulated retail period. It has been reported that WB tends to have greater pH values because of muscle degeneration and implications on glycogen content minimizing lactic acid formation in postmortem muscle [34,35,36]. It has been noted that CF pH ranges tend to fall within 6.5 to 6.9 [37]. It is plausible that the variation noted in surface color of the current study are attributed to the influences of fresh pH values.3.4. Water Activity (aw)There was no interactive (p > 0.05) impact of COMB × ALGIN × DAY on water activity during the simulated retail display. However, an interactive influence of COMB × ALGIN (Table 7) on water activity occurred. Water activity increased with increasing usage of ALGIN and percentage of WB in the formulation. As expected, ALGIN improved water holding capacity as WB inclusion increased. It is known [38] that water holding capacity is less in WB leading to further implications on water activity. In addition, there was a COMB × DAY interaction (p < 0.05) on water activity (Table 8). Ground fresh pet food with a greater percentage of WB resulted in greater water activity throughout the simulated retail display. Lastly, the interactive (p < 0.05) effect of ALGIN × DAY on water activity provides further support that the use of hydrocolloids in a meat system can aid water retention (Table 9). At the conclusion (day 21) of simulated retail display, water activity was greater (p < 0.05) in fresh pet food containing 1.0% ALGIN. Previous results support the use of hydrocolloids in a meat system for improving water retention [39].4. ConclusionsLittle to no information regarding fresh pet food shelf-life studies currently exists within the research arena. Therefore, presentation of current findings provides a brief snapshot into the use of current ingredient technologies from the meat and food industries that may be considered as viable tools for formulating fresh pet food. The current results suggest that the inclusion of ALGIN on poultry co-product pet food combinations involving CF and/or WB can improve fresh surface color characteristics. However, the combination of CF or WB used can alter surface color lightness and redness regardless of ALGIN. Regardless, the current results suggest that ALGIN with either CF or WB can be utilized in the formulation of a fresh pet food. Additional research focused on the optimal shelf-life storage period of fresh pet food is needed.

animals : an open access journal from mdpi

10.3390/ani13091442

PMC10177147

In mammals, sperm acquire fertilization ability after capacitation in vitro or when in the female reproductive tract. Adenosine triphosphate (ATP) is required for sperm capacitation through two main metabolic processes, oxidative phosphorylation (OXPHOS) and glycolysis. This study incubated dairy goat sperm with different concentrations of ROT, FCCP, TIG, and AMPK inhibitors. Sperm motility attributes, ATP content, pyruvate and lactate levels, mitochondrial permeability transition pore fluorescence intensity, MMP, protein synthesis, and ferroptosis were analyzed. The results showed that glucose starvation inhibited ferroptosis by activating the LKB1/AMPK signaling pathway and promoted the motility and linear motility of dairy goat sperm, thereby promoting fertilization.

In mammals, sperm acquire fertilization ability after capacitation in vitro or when in the female reproductive tract. The motility patterns of sperm undergo continuous changes from the moment of ejaculation until fertilization in the female reproductive tract. In vitro, hyperactivated motility can be induced through high glucose mediums, while in vivo, it is induced by oviduct fluids. Conversely, sperm maintain linear motility in seminal plasma or uterine fluids that contain low glucose levels. In dairy goat sperm, energy metabolism associated with capacitation depends on the energy sources in vitro, seminal plasma, or the female reproductive tract, especially the glucose levels. However, there is little experimental knowledge that glucose levels affect sperm energy metabolism in dairy goats. To clarify these hypotheses, we incubated dairy goat spermatozoa with different concentrations of rotenone-glucose (ROT), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP), and tigecycline (TIG) in vitro. Sperm motility attributes, ATP content, pyruvate and lactate levels, mitochondrial permeability transition pore fluorescence intensity, mitochondrial membrane potential (MMP), and protein synthesis were analyzed. Sperm motility patterns changed from circular to linear under low glucose conditions compared with those in high glucose conditions and showed a significant improvement in progressive motility and straight line speed, whereas lactate and pyruvate levels and MMP decreased remarkably. Incubation of spermatozoa with ROT, FCCP, and TIG inhibited sperm mitochondrial activity, protein synthesis, oxidative phosphorylation, and ATP levels, thereby reducing sperm motility, including the progressive motility, straight line speed, and total motility. Simultaneously, incubation of spermatozoa with Compound C under low glucose conditions significantly decreased the ATP levels and MMP, as well as liver kinase B1 and AMPK protein expression. Under low glucose conditions, sperm mainly produce ATP through mitochondrial OXPHOS to achieve high speed linear movement, inhibit ferroptosis through the LKB1/AMPK signaling pathway, and further maintain energy metabolism homeostasis.

1. IntroductionSperm are highly specialized cells with unique compositional, morphological, and functional properties. After ejaculation, mammalian sperm acquire fertilization ability after capacitation in the female reproductive tract. Fertilization in mammals is a complex process influenced by multiple factors, especially because the site of ejaculation is at a distance from the site of fertilization [1]. Therefore, sperm migration, capacitation, hyperactivation, and acrosome reaction are all necessary for successful fertilization. In addition, the components in the female reproductive tract change during different reproductive stages. Under normal physiological conditions, the vagina mainly contains glucose, lactic acid, and glycerol, whereas the uterus and oviduct mainly contain glucose, pyruvic acid, and lactic acid [2]. When sperm enter the female reproductive tract, the components in the female reproductive tract are converted into fructose, citrate, glucose, lactic acid, and free amino acids [3]. These substrates of energy metabolism are necessary for spermatogenesis, maturation, and fertilization. As a substrate for sperm preservation in vitro, the glucose concentration should be in accordance with that in the female reproductive tract. Besides glucose, sperm also utilize other substrates such as fructose and sorbitol via glycolysis to maintain motility [4,5]. Sperm motility is driven by ATP from cytosolic glycolysis, mitochondrial OXPHOS, or both [6].Sperm motility depends on the movement of sperm flagella and is driven by ATP production. Mitochondria are the most important organelles for ATP production in sperm. In human sperm, the ATP required for sperm motility is mainly generated from glycolysis. Although OXPHOS can produce ATP, it is not enough to sustain sperm movement during high motility [7]. Both glycolysis and OXPHOS are necessary to maintain vigorous motility. Incubation of mouse sperm in glucose-free media with a mitochondrial OXPHOS inhibitor (FCCP) and mitochondrial respiratory chain inhibitor (ROT) significantly reduced ATP levels and progressive motility [5,8,9]. Sperm motility parameters and the mitochondrial membrane potential (MMP) decrease significantly in sheep after treatment with FCCP and the glycolytic inhibitor 2-DOG, whereas sensitivity to lipid peroxidation (LPO) increases remarkably [10]. However, ATP production in horse sperm relies on mitochondrial OXPHOS in contrast to that in other mammals [11,12,13,14,15]. Therefore, sperm produce ATP for successful fertilization via different metabolic pathways in different energy conditions. However, the changes in motility patterns and selection of energy metabolic pathways for ATP production in dairy goat sperm in different energy conditions are unknown.5′-AMP-activated protein kinase (AMPK) is a key regulator of energy homeostasis, which is achieved by stimulating ATP production or inhibiting ATP consumption to maintain energy homeostasis [16]. Glucose starvation is a classical activating condition of AMPK that maintains cell survival and redox homeostasis through multiple pathways [17]. In rat or mouse Sertoli cells, AMPK activation induces an increase in glucose uptake and lactate production, whereas the expressions of glucose transporter 1 (GLUT1) and monocarboxylic acid transporter 4 (MCT4) reduce significantly [18]. Platycodin D inhibits ferroptosis induced by high glucose in HK-2 cells, downregulates ACSL4 and TFR1 expressions, and upregulates FTH-1 and SLC7A11 expressions [19]. Furthermore, glycolysis and pentose phosphate pathways are also inhibited by glucose starvation, which induces oxidative stress that is characterized by an increased production of reactive oxygen species (ROS). This leads to interference with the antioxidant system, redox dysregulation, and cell death [20]. Hence, in this study, sperm motility patterns at different concentrations of glucose were measured, and sperm motility parameters were evaluated using a computer-assisted sperm analysis (CASA) system. We also determined the MMP, the opening fluorescence intensity of mitochondrial permeability transition pores (mPTP), and the protein expression after incubation with a mitochondrial respiratory chain inhibitor (rotenone), a mitochondrial OXPHOS inhibitor (FCCP), and a mitochondrial translation inhibitor (TIG) to confirm the metabolic pathways at different glucose concentrations. Lastly, we explored the relationship between ferroptosis-induced glucose starvation and oxidative damage.2. Materials and Methods2.1. Chemicals and ReagentsAll chemicals and reagents were purchased from Sigma-Aldrich unless specified otherwise. All antibodies were used for incubation following the manufacturers’ instructions and were obtained from the following vendors: COX-1 (13393-1-AP, Proteintech, Wuhan, China), COXVB (11418-2-AP, Proteintech, Wuhan, China), ND3 (ab192306, Abcam, Cambridge, UK), NRF1 (12482-1-AP, Proteintech, Wuhan, China), SLC7A11 (26864-1-AP, Proteintech, Wuhan, China), GPX4 (67763-1-AP, Proteintech, Wuhan, China), DHODH (14877-1-AP, Proteintech, Wuhan, China), β-actin (20536-1-AP, Proteintech, Wuhan, China), horseradish peroxidase (HRP)-labeled goat anti-mouse immunoglobulin (Ig)G (CW0102S, CWBIO, Beijing, China), HRP-labeled goat anti-rabbit IgG (CW0103S, CWBIO, Beijing, China), and enhanced chemiluminescence (ECL) fluid (SuperSignalTM West Femto Maximum Sensitivity Substrate, 34095, Thermo Fisher Scientific, Waltham, MA, USA).2.2. Animals and Semen CollectionSemen was collected using an artificial vagina in Guan Zhong dairy goats (n = 5) from Shaanxi Aonick Dairy Goat Breeding Co., Ltd. (Weinan, China), that were raised under the same feeding condition as in Weinan, Shaanxi Province (34°76′ N, 109°17′ E). The semen of dairy goats was collected using the artificial vagina method and the collection frequency was three times a week on Monday, Wednesday, and Friday mornings. Semen samples with a deformity rate of less than 5%, normal morphology of sperm, motility greater than 0.8, and semen density exceeding 1.5 billion/mL were suitable for this study. Subsequently, the semen samples were pooled to eliminate individual variation and used for subsequent experiments.2.3. Ethics StatementAll animal experiments were approved by the Animal Ethics Committee of Northwest A&F University. All experiments with dairy goats were implemented strictly according to the Guide for the Care and Use of Laboratory Animals.2.4. Semen ProcessingDilute fresh sperm with a modified solution consisting of 172 mM glucose, 90 mM lactose, 58 mM sodium citrate, 1000 IU/mL penicillin G potassium, and 1 mg/mL streptomycin was passed through a 0.22 μm filter. The glucose concentration (172 mM) in the modified solution was determined to be 40%. A portion of lactose was used with glucose to prepare different doses of glucose bulking agents (0, 86, 172, 258, 344, and 430 mM, i.e., 0%, 20%, 40%, 60%, 80%, and 100%), as sperm do not use lactose as an energy substrate. Sperm were incubated with different concentrations of glucose medium at 37 °C for 1 h to evaluate whether dairy goat sperm were involved in the mitochondrial oxidative phosphorylation mode of energy supply. Moreover, sperm were incubated in a low glucose group containing 86 mM glucose and different concentrations of ROT, FCCP, and TIG at 37 °C for 5 h to evaluate the role of mitochondrial oxidative phosphorylation in the regulation of mitochondrial energy metabolism function in dairy goat spermatozoa. Finally, Compound C (AMPK inhibitor) was added to low and high glucose diluents and incubated in a water bath at 37 °C for 5 h to evaluate whether the low glucose dilutions activated AMPK to maintain energy homeostasis and inhibit sperm oxidative damage.2.5. Sperm MotilitySperm motility parameters were assessed using a CASA system (HVIEW-SSAV8.0, FuZhouHongShiYeSoftware Technology Co., Ltd., Fuzhou, China). Briefly, 10 μL of the samples was dropped onto a preheated glass slide and covered with a coverslip. Then, sperm motion was captured in at least five different regions using dynamic video acquisition of CASA on a heated platform at 37 °C. Each region was required to contain at least 200 sperm counts. Lastly, sperm motility parameters were assessed using the analysis software of CASA.2.6. Biochemical AssaysAll tests were conducted per the respective manufacturers’ instructions for each kit and were obtained from the following vendors: reactive oxygen species (ROS) assay kit (CA1410, Solarbio, Beijing, China), malondialdehyde (MDA) assay kit (BC0025, Solarbio, Beijing, China), pyruvate (PA) content assay kit (BC2205, Solarbio, Beijing, China), lactic acid (LA) assay kit (A019-2-1, Nanjing Jiancheng, Nanjing, China), ATP assay kit (S0026, Beyotime, Shanghai, China), MMP assay kit (M8650, Solarbio, Beijing, China), and bicinchoninic acid (BCA) protein assay kit (PA115, TAINGEN, Beijing, China). Finally, the related indexes were detected by a multifunctional enzyme label instrument (Synergy H1, American Berten, VT, USA).2.7. Mitochondrial Permeability Transition Pore (mPTP) Fluorescence IntensitySperm samples were centrifuged at 1000× g at room temperature for 5 min and the supernatant was discarded. The sperm density was diluted to achieve a final concentration of 1 × 106 sperm/mL by adding an appropriate volume of Calcein AM (1000×) staining solution and fluorescence quenching solution or ionomycin (200×) control, and incubated in the dark for 30 min. Then, the supernatant was removed by centrifugation at 1000× g at room temperature for 5 min and slowly resuspended twice at 37 °C. Subsequently, the samples were smeared and analyzed using flow cytometry (FACS Melody, BD Biosciences, Franklin Lakes, NJ, USA) within 1 h. The excitation wavelength was 494 nm and the emission wavelength was 517 nm.2.8. Immunofluorescence AssaysSperm samples were washed with phosphate-buffered saline (PBS), spread on a glass slide with poly-D-lysine, dried naturally, and fixed in 4% paraformaldehyde for 10 min. After drying, the sperm were penetrated with PBS containing 0.5% Triton X-100 for 30 min, washed three times for 5 min each, and blocked with 5% bovine serum albumin (BSA) in PBS-T for 30 min. Next, the slides were incubated with the primary antibody (1:100) overnight at 4 °C. Subsequently, the slides were incubated with the second antibody (1:200) at 37 °C for 1 h in the dark. The sperm nuclei were stained with DAPI at room temperature for 1 min in the dark and washed three times with PBS for 5 min each. Lastly, the samples were observed using a fluorescence microscope (LECIA-DM6 B, LECIA Co., Ltd., WETZLAR, Hessian, Germany).2.9. Western BlottingProtein concentrations were determined using a BCA protein assay kit (PA115, TAINGEN, Beijing, China). Proteins were separated using FuturePAGE™ 4–12% (ET15412Gel, ACE Biotechnology, Nanjing, China) sodium dodecyl sulfate–polyacrylamide gel electrophoresis, transferred to polyvinylidene fluoride membranes, and blocked with 5% BSA for 1 h at room temperature. The primary antibody was diluted and incubated with the membrane overnight at 4 °C. After incubation, the membrane was washed three times for 10 min each using Tris-buffered saline tween (TBST) and incubated with the second antibody at 37 °C for 1 h. After washing with TBST, detection was performed using enhanced chemiluminescence (ECL) (34095, Thermo Fisher Scientific, Waltham, MA, USA) using a Gel Doc XR System (BioRad, Hercules, CA, USA) according to the manufacturer’s specifications. Lastly, the intensity of the protein bands was analyzed using ImageJ software.2.10. Statistical AnalysisData are expressed as the mean ± standard error of the mean from at least 3 independent experiments. A statistical analysis among groups was performed using PRISM (version 6, GraphPad), followed by post hoc test using the Student’s t-test. The mean difference at p < 0.05 was considered statistically significant (* p < 0.05, ** p < 0.01, *** p < 0.001, and **** p < 0.0001. ns, with no apparent difference).3. Results3.1. Sperm Energy Metabolic Pathway and Sperm Motility Patterns Are Changed in Different Concentrations of GlucoseThe CASA system can detect the motility and movement trajectory of sperm. Sperm motility tracks changed into circle-like tracks in high glucose media (>60%), and the progressive motility and straight line velocity reduced significantly. However, the total motility and straight line velocities in the low glucose groups (20% and 40%) were significantly higher than those in the other groups (Figure 1A–D). To further determine the energy metabolic pathway, the levels of pyruvate, lactic acid, ATP, and MMP in the sperm of dairy goats were assessed at different glucose concentrations. The levels of pyruvate, lactic acid, and MMP were significantly higher in the 40% glucose group after incubation for 1 h compared with those in the control group, whereas the levels decreased remarkably in the high glucose group (Figure 1E–H, p < 0.001).3.2. Addition of Rotenone to the Culture Reduces ATP Content and Straight Line MotilityTo further validate the role of OXPHOS in energy metabolism in dairy goat sperm, rotenone (an inhibitor of mitochondrial respiratory chain complex I) was incubated with sperm samples for 5 h in this study. The results from CASA showed significant reductions in forward sperm motility and straight line velocity, but no significant difference was observed in the total motility of sperm (Figure 2A–D). Moreover, there was a considerable decrease in the fluorescence intensity in a dose-dependent manner with respect to mPTP and ATP levels with the addition of rotenone (Figure 2E–G, p < 0.05). Therefore, the expression of the marker proteins of ETC and OXPHOS was determined using immunofluorescence staining and Western blotting. Immunofluorescence staining results showed that COX-1 and COXVB were mainly distributed in the middle segment of the sperm tails (Figure 2H) and that the expression of COX-1, COXVB, ND3, and NRF1 decreased significantly after treatment with rotenone (Figure 2I–M, p < 0.001). Taken together, it could be inferred that rotenone inhibits the regulation of mitochondrial OXPHOS in dairy goat sperm.3.3. The OXPHOS Uncoupler FCCP Reduces Sperm Motility and ATP LevelsTo further explore the effect of OXPHOS on energy metabolism in dairy goat sperm, the samples were treated with a specific inhibitor of mitochondrial OXPHOS (FCCP). The sperm motility, forward motility, and linear velocity significantly decreased in a dose-dependent manner after treatment with FCCP for 5 h (Figure 3A–D, p < 0.0001). The fluorescence intensity of mPTP and the level of ATP in mitochondria also decreased significantly (Figure 3E–G, p < 0.01). Furthermore, we observed that COX-1, COXVB, ND3, and NRF1 expressions in mitochondria decreased obviously (Figure 3H–L, p < 0.001).3.4. The Mitochondrial Translational Inhibitor TIG Reduces MMP and ATP Levels in a Dose-Dependent MannerTo further confirm that the straight line motility pattern in dairy goat sperm relies on mitochondrial transcription and translation, the mitochondrial translation inhibitor TIG was incubated with sperm. There were no significant differences in sperm motility at the 5 h point, but the progressive motility and straight line velocity of sperm decreased significantly (Figure 4A–C, p < 0.0001). Moreover, ATP levels and MMP decreased noticeably after incubation with TIG (Figure 4D,E, p < 0.0001), indicating that the inhibition of mitochondrial translation reduces the progressive motility and straight line velocity of sperm. Additionally, COX-1, COXVB, NRF1, and TFAM expressions in mitochondria decreased significantly (Figure 4F–J, p < 0.0001). Therefore, energy metabolism and the linear motility pattern are closely related to transcription and translation in mitochondria.3.5. Low Glucose Conditions Inhibit Ferroptosis-Induced Oxidative Damage by Regulating the Liver Kinase B1 (LKB1)/AMPK Signaling PathwayHigh glucose media cause a reduction in cell viability as well as ROS accumulation, LPO, and outer mitochondrial membrane rupture, eventually inducing ferroptosis [21]. Therefore, low and high concentrations of glucose were used to determine whether low glucose conditions could inhibit ferroptosis and maintain the redox balance by activating AMPK in dairy goat sperm. Sperm motility in high glucose media did not change obviously compared with that in low glucose media at the 1 h point, as determined using CASA, but decreased significantly after high glucose or erastin incubation for 3 h or 5 h; moreover, ROS and MDA levels increased significantly (Figure 5A–C, p < 0.01). In addition, the intensities of LKB1, AMPK, glutathione peroxidase 4 (GPX4), and family 7 member 11 (SLC7A11) proteins decreased significantly in high glucose conditions or when treated with erastin (Figure 5D,E, p < 0.001). Therefore, high glucose levels could induce a reduction in sperm motility and lead to the accumulation of ROS and LPO in dairy goat sperm, which, in turn, caused ferroptosis. Glucose is the primary energy source for cell types, and AMP-activated protein kinase (AMPK) is an energy sensor. In low energy conditions, cells can be isomerically activated by adenosine monophosphate (AMP) and adenosine diphosphate [22]. Thus, an AMPK inhibitor (Compound C) was used to determine whether low glucose levels could inhibit ferroptosis in dairy goat sperm by regulating AMPK. A significant decrease was observed in ATP levels and MMP in the low glucose group (p < 0.01) but not in the high glucose group (Figure 5F,G). Furthermore, Western blotting revealed that the expression of LKB1 and AMPK proteins decreased noticeably in high glucose conditions or after treatment with Compound C (Figure 5H–J, p < 0.001).4. DiscussionMammalian sperm are produced through spermatogenesis and serve the purpose of fertilizing an oocyte. After ejaculation, sperm require energy to meet their demands. This energy can be obtained through the importation of exogenous substrates or the use of endogenous sources. Adenosine triphosphate (ATP), an important energy source that supports sperm capacitation and migration in the female reproductive tract, is generated from two main metabolic pathways, namely glycolysis and oxidative phosphorylation (OXPHOS). They occur in different compartments in sperm flagella, both of which are located in the principal piece [23]. The metabolic pathways of energy in sperm change in the presence of different metabolic substrates. In addition to these changes, the sperm motility pattern associated with capacitation also changes and is called hyperactivation [24]. The hyperactivated status is always characterized by a high curvilinear velocity and a high lateral amplitude [25]. However, the relationship between the energy metabolic pathway and motility patterns in dairy goat sperm remains unclear. In this study, we established a model by incubation of sperm with different concentrations of glucose and found that not only the progressive motility and straight line velocity of sperm but also ATP levels and MMP increased significantly in low glucose media. This finding indicated that sperm regulate the energy metabolism pathway based on changes in the metabolic substrates and that low glucose levels activate OXPHOS in the mitochondria to maintain the normal functions of sperm. In terms of the changes in oxygen and substrate levels in the female reproductive tract, sperm adapt to the fluctuations in exogenous substrate levels via different energy metabolism pathways [26]. Qiu et al. found that among the substrates tested, glucose and pyruvate were better than lactate at maintaining goat sperm motility. Pyruvate enters goat spermatozoa through monocarboxylate transporters and is oxidized by the tricarboxylic acid cycle and electron transfer to sustain sperm motility [27]. Zhu et al. reported that GSK3α/β was expressed in the head acrosome and the mid, main end of the tail of goat sperm and that it regulated goat sperm motility and the acrosome response by mediating energy pathways in glycolysis and oxidative phosphorylation [28]. Thus, these different energy metabolic pathways are interconnected in sperm, and different concentrations of glucose can selectively activate glycolysis as well as OXPHOS in the mitochondria. Moreover, pyruvate and lactate are involved in the glycolysis pathway. In this study, the progressive motility of sperm was found to be regulated by OXPHOS in the mitochondria, which is similar to the physiological state of sperm in the female reproductive tract after ejaculation. The semen, uterus, and oviduct contain high levels of pyruvate, lactate, and amino acids, and sperm can utilize these metabolites through OXPHOS for energy metabolism, further enhancing their progressive motility and speed of access to reach the site of fertilization. Zhu et al. demonstrated that ATP generated from OXPHOS in the mitochondria in low glucose conditions induced progressive motility in boar sperm. On the other hand, incubation with CRP, a mitochondrial translation inhibitor, inhibited mitochondrial translation in sperm and decreased their progressive motility but not total motility, suggesting that motility patterns in sperm depend on the substrates of energy metabolism [29]. Moreover, high glucose concentrations led to a transient inefficiency in the late stages of both glycolysis and NADH accumulation and decreased the conversion flux from fructose-1,6-bisphosphate to pyruvate and the ATP content [30].Mitochondria are called the motive source of cells and play a key role in energy metabolism through the tricarboxylic acid (TCA) cycle, β-oxidation, and OXPHOS [31]. During glycolysis, a glucose molecule produces two ATP molecules, whereas in the TCA cycle or OXPHOS pathway, a glucose molecule produces 32 ATP molecules. OXPHOS is composed of four respiratory complexes (complexes I to IV) and ATP synthase (complex V) and plays a key role in MMP across the mitochondrial inner membrane [32]. Rotenone is a strong inhibitor of complex I of the mitochondrial respiratory chain. Heo et al. demonstrated that rotenone induces mitochondrial dysfunction by inhibiting SIRT1 during oocyte maturation in pigs in vitro, thereby reducing mitochondrial activity and ATP generation and increasing ROS production [33]. There was no significant difference in sperm motility after incubation with rotenone, but there was a significant decrease in mitochondrial activity and ATP levels [29]. Besides, FCCP is a classic uncoupler of mitochondrial OXPHOS. Davila et al. showed that sperm motility, speed, and the number of live sperm decreased significantly in horses after incubation for 3 h with FCCP, but the MMP and ATP levels were unaffected [15]. Blanco-Prieto et al. have reported that FCCP significantly increases oxygen consumption and decreases the total motility of boar sperm [34]. In this study, dairy goat sperm were incubated with rotenone and FCCP separately. After treatment with rotenone, their total motility did not change much, but the progressive motility and straight linear speed reduced significantly. The addition of FCCP decreased the overall survival and mPTP permeability in a dose-dependent manner. Both these treatments significantly reduced the MMP and ATP content. Furthermore, the protein expression related to the mitochondrial respiratory chain determined using Western blotting showed that incubation with rotenone and FCCP resulted in the disruption of the ETC in the mitochondria and inactivation of ATP synthase [35]. On the other hand, the relationship between mitochondrial function and ATP generation from OXPHOS is especially close, and mitochondrial activity is influenced by proteins from the mitochondrial respiratory chain. Zhu et al. found that in a low glucose environment, 13 genes involved in mitochondrial transcription and translation were remarkably upregulated, which, in turn, improved mitochondrial OXPHOS and progressive motility in boar sperm [29]. Consistent with the above findings, tigecycline (a mitochondrial translation inhibitor) inhibited the expression of proteins in the mitochondrial respiratory chain in our study, decreased the MMP and ATP content, and reduced the progressive motility of dairy goat sperm. Glucose and fructose are essential components as a semen preservation diluent, but excessive glucose induces cellular damage [36]. High glucose concentrations induce cell apoptosis, ferroptosis, necrosis, or other types of cell death [37,38]. Ferroptosis is a new form of oxidative cell death that is induced by small molecules and especially by the imbalance of lipid ROS in cells [8]. Ferroptosis is also induced by other factors, including the inhibition of glutathione synthetase or GPX4, leading to the accumulation of LPO products and ROS and eventually resulting in cell death [39]. In our study, incubation with either high glucose levels or erastin for 3 h significantly reduced sperm motility, increased ROS and MDA levels, and decreased the expressions of LKB1/AMPK, SLC7A11, and GPX4. However, in low glucose media, there was no significant difference in sperm motility after incubation for 3 h, indicating that the normal functions of sperm could be maintained by accelerating ATP production through the AMPK pathway [40]. Glucose starvation is a typical activator of AMPK and an important regulator in energy homeostasis. The LKB1/AMPK signaling pathway is critical in maintaining cell survival under low glucose conditions [20]. Metformin (an activator of AMPK) was found to reduce the levels of LPO-stimulated steroid, restore spermatogenesis, and increase sperm motility in the testes of diabetic or obese rats [41]. During cryopreservation of chicken sperm, the addition of an AMPK activator, metformin or AICAR, improved AMPK phosphorylation after thawing, whereas LPO and ROS production was reduced [42]. Additionally, the activation of AMPK in goat sperm improved sperm motility, plasma membrane integrity, and the acrosome reaction; maintained normal levels of MMP, lactate acid, and ATP; and enhanced the activity of AMPK, PK, and lactate dehydrogenase. However, the addition of Compound C inhibited the AMPK pathway and induced the opposite effects [43]. In the present study, the addition of Compound C to the low glucose media led to reductions in ATP levels and MMP as well as a decrease in LKB1/AMPK expression compared with that in the high glucose group. These results suggested that the LKB1/AMPK signaling pathway can be activated to maintain energy homeostasis in sperm during energy stress.5. ConclusionsEnergy metabolism in dairy goat sperm is realized through the mitochondrial OXPHOS pathway under low glucose conditions, and the high speed linear motility and straight line velocity improved significantly. Moreover, in this study, the results of incubation with rotenone and FCCP confirmed that the progressive motility of sperm mainly relies on OXPHOS in the mitochondria and that ATP generation from OXPHOS is closely related to mitochondrial functions. Low glucose conditions promote transcription and translation in the mitochondria and activate mitochondrial OXPHOS to supply energy to sperm. Moreover, the LKB1/AMPK pathway is activated to maintain energy homeostasis and inhibit ferroptosis-induced oxidative damage. Therefore, the high speed linear motion induced by low sugar is a new factor to improve insemination in artificial insemination of livestock and human fertility treatments. As a convenient and inexpensive method, sperm dilution with a low-sugar insemination solution improves the rapid linear movement of sperm within the female reproductive tract, thereby facilitating fertilization.

animals : an open access journal from mdpi

10.3390/ani11113183

PMC8614503

Complement regulation is related to fetal growth during early pregnancy in humans. The results show that C1q increased during early pregnancy, and C1r, C1s, C2, C3 and C5b enhanced at day 25 of gestation, and C4a and C9 increased on days 13 and 16 of pregnancy in the ovine maternal spleen. In summary, a complement pathway participates in maternal immune regulation during early gestation.

During early gestation in humans, complement regulation is essential for normal fetal growth. It is supposed that a complement pathway participates in maternal splenic immune regulation at the early stage of gestation in ewes. The aim of this study was to analyze the effects of early pregnancy on the expression of complement components in the maternal spleen of ewes. In this study, ovine spleens were sampled on day 16 of nonpregnancy, and days 13, 16 and 25 of gestation. RT-qPCR, Western blot and immunohistochemical analysis were used to detect the changes in expression of complement components in the ovine maternal spleens. Our results reveal that C1q was upregulated during early gestation, C1r, C1s, C2, C3 and C5b increased at day 25 of gestation and C4a and C9 peaked at days 13 and 16 of gestation. In addition, C3 protein was located in the capsule, trabeculae and splenic cords. In conclusion, our results show for the first time that there was modification in the expression of complement components in the ovine spleen at the early stage of gestation, and complement pathways may participate in modulating splenic immune responses at the early stage of gestation.

1. IntroductionA complement pathway is the replenishment of immunological processes, and is implicated in immune surveillance, cell homeostasis and tissue development [1]. The complement system participates in the complex tolerance and clearance processes, and complement system dysregulation results in insufficient clearance of the fragmentation of the placental tissue and preeclampsia in humans [2]. The activity of the complement system is enhanced systemically during pregnancy, but the activity is inhibited at the placenta for pregnancy maintenance [3]. Complement regulation is necessary for a successful pregnancy, and some complement components are favorable in the normal fetal growth at the stage of pre-implantation to placentation [4]. However, during early pregnancy, complement regulation in the spleen is unclear in ewes.There is selective immunosuppression in the uterus and peripheral tissues induced by conceptus signals, which are essential for the fetus to evade maternal immune detection and elimination in domestic farm animals [5]. The spleen has a wide range of immunologic functions and plays key roles in the initiation of adaptive immunity [6]. Interleukin-33 (IL-33) participates in innate and adaptive immunity, and the IL-33 receptor is upregulated in splenic B cells during normal gestation in mice [7]. Our previous studies demonstrate that interferon-stimulated genes (ISGs), progesterone (P4) receptor, P4-induced blocking factor, tumor necrosis factor β, interleukin-2 (IL-2), IL-5, IL-6, IL-10, cyclooxygenase 2, aldo-keto reductase family 1, member B1, melatonin receptor 1 (MT1), gonadotropin-releasing hormone (GnRH) and its receptor are upregulated. However, MT2 is decreased in the ovine maternal spleen at the early stage of gestation [8,9,10,11,12,13]. It is presumed that a complement pathway is involved in the immune tolerance regulation of the maternal spleen at the early stage of gestation in sheep. The aim of the present research was to analyze the expression of complement components in the maternal spleen at the early stage of gestation in ewes, which may be useful for understanding the establishment of the immune tolerance mechanism of the maternal spleen.2. Materials and Methods2.1. Animals and Experimental DesignThe experimental design and tissue collection have been described in detail before [9,10,11]. Briefly, 35 ewes with similar age, BCS and parity received the same diet, and controlled internal drug-releasing devices (InterAg, Hamilton, New Zealand) were used for estrus synchrony. After estrus detection (day 0), six ewes were randomly selected for nonpregnant ewes and slaughtered at day 16 after estrus (day 16 of the estrous cycle, DN16) for spleen collection. The other 29 animals were exposed to fertile rams, which ensured there were at least 18 pregnant ewes. The pregnant animals were randomly killed at days 13, 16 and 25 after mating for spleen collection (n = 6 for each group). Days 13, 16, and 25 of pregnancy (DP13, DP16, and DP25), and day 16 of the estrous cycle (DN16), were chosen because the maternal spleen was under the different effects of interferon-tau and/or P4, or not on these days [12]. There was a morphologically normal conceptus for pregnant ewes at slaughter. Cross sections (0.5 cm thick) of the spleens were collected for subsequent mRNA and proteins analysis, and immersed in 4% paraformaldehyde in PBS for subsequent immunohistochemistry.2.2. RT-qPCR AssayThe total RNA isolation and cDNA synthesis have been described in detail before [12]. The specified primers were designed for the ovine (Ovis aries) complement component genes (Table S1) and synthesized by Shanghai Sangon Biotech Co., Ltd. (Shanghai, China). The primer matrix experiments were used for determining the optimal primer concentrations. The amplification efficiency of each primer sequence was in an acceptable range. A SuperReal PreMix Plus kit (Tiangen Biotech, Beijing, China) was used for quantitative PCR in a CFX96 real-time PCR system (Bio-Rad, Hercules, CA, USA), and GAPDH was applied as a housekeeping gene. Relative transcript abundances of complement component genes were analyzed utilizing the 2−ΔΔCt analysis method [14]. The mean CT value of DN16 was applied to normalize the mRNA expression level of DP13, DP16 and DP25.2.3. Western BlotThe proteins were isolated from ovine splenic samples as described previously [13]. Samples were isolated by SDS-polyacrylamide gel electrophoresis and transferred electrophoretically on PVDF membranes (Millipore, Bedford, MA, USA). The membranes were blocked with skim milk, and incubated with the antibodies (Table S2) at a dilution of 1:1000. Then, the membranes were incubated with an HRP-conjugated anti-mouse secondary antibody (Biosharp, Hefei, China BL001A, 1:2000). An HRP chemiluminescence kit was applied for detecting blots. The blot intensity was analyzed by Quantity One V452 (Bio-Rad Laboratories)., Hercules, CA, USA). The relative intensity of the blots was calculated and normalized with a reference protein (GAPDH) using an anti-GAPDH antibody (Table S2, 1:1000).2.4. Immunohistochemical AnalysisSplenic tissue was dehydrated and embedded in paraffin wax. The paraffin-embedded sample was sectioned in 5 μm-thick sections, mounted onto slides, deparaffinized and rehydrated. Sections underwent antigen recovery in boiling 0.01 M citric buffer and were treated with 3% H2O2 for blocking endogenous peroxidase activity. After blocking nonspecific binding sites using 5% normal goat serum in PBS, the sections were incubated with the anti-C3 antibody (Table S2, 1:200). The negative control sections were incubated with an antiserum-specific isotype instead of the anti-C3 antibody. The specific binding site was detected utilizing a DAB kit (Tiangen Biotech). In the end, a light microscope (Nikon Eclipse E800, Tokyo, Japan) with a digital camera DP12 was applied for image capture, and the photo was analyzed independently by four investigators. The staining intensities of the splenic samples were analyzed through the images in a blind manner and assigned an immunoreactive intensity (from 1 to 3) [11].2.5. Statistical AnalysisLeast-squares ANOVA applying Mixed and General Linear Model procedures of the Statistical Analysis System (SAS Institute, Cary, NC, USA) was applied to analyze the relative abundance levels of mRNA and protein. The PROC UNIVARIATE procedure (SAS Institute Inc.) was used for testing the data normality, and the non-parametric test was applied in cases not achieving normal distribution. Data were expressed as least squares means. Data obtained from the different spleens of animals were analyzed for the main effects of day and pregnant status and the interaction between the main effects. The Duncan method was applied to analyze multiple comparisons. A p-value < 0.05 was considered significantly different.3. Results3.1. Expression of Complement Component mRNA in the SpleenIt is revealed in Figure 1 that the C1q mRNA increased significantly during early pregnancy, and reached the highest level at DP25 (Figure 1; p = 0.0023 between DN16 and DP25; p = 0.0087 between DP13 and DP25; p = 0.0063 between DP16 and DP25). The relative values of C1r, C1s, C2, C3 and C5b mRNA upregulated at DP25 (p = 0.0034 between DN16 and DP25, p = 0.0037 between DP13 and DP25 and p = 0.0035 between DP16 and DP25 for C1r; p = 0.0082 between DN16 and DP25, p = 0.0076 between DP13 and DP25 and p = 0.0021 between DP16 and DP25 for C1s; p = 0.0071 between DN16 and DP25, p = 0.0065 between DP13 and DP25 and p = 0.0073 between DP16 and DP25 for C2; p = 0.0075 between DN16 and DP25, p = 0.0081 between DP13 and DP25 and p = 0.0036 between DP16 and DP25 for C3; p = 0.0087 between DN16 and DP25, p = 0.0079 between DP13 and DP25 and p = 0.0032 between DP16 and DP25 for C5b), but C1s, C3 and C5b mRNAs were downregulated at DP16. In addition, the relative levels of C4a and C9 peaked at DP13 and DP16, but declined at DP25 in the maternal spleen (p = 0.0053 between DP13 and DP25, and p = 0.0061 between DP16 and DP25 for C4a; p = 0.0069 between DP13 and DP25 and p = 0.0064 between DP16 and DP25 for C9).3.2. Protein Expression of Complement Component 9 in the SpleenFigure 2 and Figure S1 show that early gestation stimulated the expression of C1q protein (p = 0.0017 between DN16 and DP13; p = 0.0092 between DN16 and DP16; p = 0.0081 between DN16 and DP25). The values of C1r, C1s, C2, C3 and C5b proteins upregulated at DP25 (p = 0.0061 between DN16 and DP25, p = 0.0059 between DP13 and DP25 and p = 0.0068 between DP16 and DP25 for C1r; p = 0.0089 between DN16 and DP25, p = 0.0097 between DP13 and DP25 and p = 0.0046 between DP16 and DP25 for C1s; p = 0.0031 between DN16 and DP25, p = 0.0076 between DP13 and DP25 and p = 0.0083 between DP16 and DP25 for C2; p = 0.0037 between DN16 and DP25, p = 0.0043 between DP13 and DP25 and p = 0.0014 between DP16 and DP25 for C3; p = 0.0049 between DN16 and DP25, p = 0.0047 between DP13 and DP25 and p = 0.0028 between DP16 and DP25 for C5b). In addition, the C4a and C9 proteins were expressed at DP13 and DP16 (p = 0.0034 between DP13 and DP25, and p = 0.0029 between DP16 and DP25 for C4a; p = 0.0051 between DP13 and DP25, and p = 0.0048 between DP16 and DP25 for C9). However, it was undetected at DN16 and DP25.3.3. Immunohistochemical Location for C3 ProteinC3 protein was limited to the capsule, trabeculae and splenic cords (Figure 3). The staining intensities in the splenic samples for C3 were zero (negative), two (strong), two (weak), one (strong), and three (stronger) for the negative control, the spleens from DN16 and spleens from DP13, DP16 and DP25, respectively (Figure 3).4. Discussion4.1. Early Pregnancy Changed the Expression of Complement Components in the Maternal SpleenOur data reveal that C1q expression was upregulated during the early stage of gestation. Paternal deficiency of C1q results in fetal growth restriction in a preeclampsia-like gestation in mice [15]. The complement protein C1q is expressed in gestational tissues during pregnancy, which is involved in regulating the mother’s immune responses [16]. Therefore, the increase in C1q during early gestation participates in an adjustment of the maternal immune responses that is essential for pregnancy maintenance in ewes.There were increases in C1r and C2 in the maternal spleen at DP25 in this study. In cattle, there is an upregulation of C1r mRNA in the luminal and glandular epithelial cells during the preattachment period [17]. C1r is undetected in the implantation site of the preeclampsia model [18], suggesting that C1r expression in the implantation site is beneficial for successful pregnancy in mice. C2-deficient patient is susceptible to developing to thrombocytopenia during early gestation [19]. Therefore, the increases in C1r and C2 in the maternal spleen may be necessary for normal early gestation in ewes.C1s and C3 declined at DP16, but increased at DP25 in this study. Amnion tissue explants can synthesize C1s, which contribute to the local host defense in humans [20]. C1s protein is expressed in chorionic tissue, and secreted extracellularly, which is related to the tolerance of the allograft fetus in humans [21]. In metazoans, C3 interacts with many complement factors and non-complement proteins to play an essential role in immune regulation [22]. C3 protein is downregulated significantly in the serum of pregnant women with neural tube defects compared with the women carrying normal fetuses during pregnancy [23]. Plasma concentration of the C3 protein is increased gradually throughout pregnancy compared with nonpregnant women, suggesting that C3 plays a key role in normal placentation [24]. Therefore, it is possible that upregulation of C1s and C3 at DP25 may be related to placentation.Our data reveal that C5b was downregulated at DP16, but upregulated at DP25. C5 is cleaved to generate C5b and C5a by the C3-cleaving enzyme, and C5b participates in immune responses through assembling with downstream complement components [25]. C5b-9 is localized in the surface of syncytiotrophoblasts, intervillous fibrin and decidual vessels, which contributes to placental formation [26]. There is a higher level of C5b-9 in the villous trophoblast of placentas from normal pregnant women compared with the patients with preeclampsia [27]. Therefore, the increase in C5 at DP25 may be essential for placentation.C4a and C9 peaked at days 13 and 16 of pregnancy in this study. Human chorionic gonadotropin stimulates the C4a gene expression in the baboon endometrium, which is implicated in regulating the decidual immune environment at the time of implantation [28]. The frequency of C4 ‘null’ alleles is increased in recurrent spontaneous abortions, suggesting that C4a is necessary for a successful pregnancy in humans [29]. However, there is a high concentration of circulating C4 in preeclamptic patients compared with that in normal nonpregnant women [30]. C9 is the major component of the membrane attack complex that forms pores in the plasma membrane of target cells to participate in the adaptive immune response [31]. After the first trimester of gestation, the serum level of C9 is downregulated in women with normal pregnancy compared with females with preeclampsia [32]. Therefore, it is suggested that the upregulation of C4a and C9 on DP13 and DP16 is related to implantation, and the downregulation of C4a and C9 at DP25 is helpful for pregnancy maintenance.The spleen participates in the blood defense against blood-borne pathogens, and is also partly implicated in the immune responses [33]. The immunohistochemistry for C3 proteins was mainly located in the capsule, trabeculae and splenic cords, and the staining intensities for C3 protein were changed during early pregnancy. There is a change in the activation state of peripheral CD4 lymphocytes in the maternal spleen during the preimplantation period of gestation in mice [34]. Early gestation changed the expression of the P4 receptor, ISGs, prostaglandin synthases, T helper cytokines, MTs, GnRH and its receptor in the ovine maternal spleen in our previous papers [8,9,10,11,12,13]. C3 plays key roles in three complement pathways and is directly related to immune regulation and immune defense activation [22]. Therefore, it is suggested that a complement pathway is related to modifying maternal splenic immune tolerance through C3-dependent signaling during early gestation in ewes.4.2. Possibility for Clinical ImplicationsIn this study, there was upregulation of C1r, C3 and C5b in the maternal spleen, but the expression of C4 and C9 was downregulated at DP25. C1r expression is decreased in the implantation site of the preeclampsia model [18], and C2 deficiency results in thrombocytopenia during early pregnancy [19]. Neural tube defects in pregnant women lead to downregulation of the C3 protein in the serum compared with normal pregnant women [23]. Normal women have a lower level of C5b-9 in the villous trophoblast of the placentas [27]. However, the concentration of circulating C4 in preeclamptic patients is higher than in normal women [30], and the serum level of C9 is decreased in women with normal pregnancy after the first trimester of gestation [32]. Therefore, it is suggested that the agonists for C1r, C3 and C5b, and the antagonists for C4 and C9 may be used for improving reproductive efficiency after early pregnancy.5. ConclusionsEarly pregnancy induces changes in the expression of complement components in the maternal spleen, and these changes can be used as signs of early pregnancy in ewes. Furthermore, the C3 protein was limited to the capsule, trabeculae and splenic cords. In summary, complement pathways may be related to accommodating maternal spleen immune responses at the early stage of gestation, and the detection of C1q, C4a and C9 may be used for early pregnancy diagnosis in ewes.

animals : an open access journal from mdpi

10.3390/ani11051382

PMC8152225

This paper examines scientific evidence on the positive effects of donkey milk consumption on human health and its possible therapeutic applications. The most investigated clinical use of donkey milk is in feeding infants with food allergies, in whom donkey milk is well tolerated in the 82.6–98.5% of cases. Donkey milk has shown several beneficial properties, including immunomodulatory activity, antioxidant and detoxifying effects, modulation of the intestinal microbiota, and lowering of blood sugar and triglycerides, which have been tested almost exclusively in experimental animals. Inhibitory actions on microorganisms have been also observed in vitro studies. This literature review highlights the need for new clinical trials to collect stronger evidence about the positive effects observed in experimental models which could lead to new therapeutic applications of donkey milk in humans.

The increase of knowledge on the composition of donkey milk has revealed marked similarities to human milk, which led to a growing number of investigations focused on testing the potential effects of donkey milk in vitro and in vivo. This paper examines the scientific evidence regarding the beneficial effects of donkey milk on human health. Most clinical studies report a tolerability of donkey milk in 82.6–98.5% of infants with cow milk protein allergies. The average protein content of donkey milk is about 18 g/L. Caseins, which are main allergenic components of milk, are less represented compared to cow milk (56% of the total protein in donkey vs. 80% in cow milk). Donkey milk is well accepted by children due to its high concentration of lactose (about 60 g/L). Immunomodulatory properties have been reported in one study in humans and in several animal models. Donkey milk also seems to modulate the intestinal microbiota, enhance antioxidant defense mechanisms and detoxifying enzymes activities, reduce hyperglycemia and normalize dyslipidemia. Donkey milk has lower calorie and fat content compared with other milks used in human nutrition (fat ranges from 0.20% to 1.7%) and a more favourable fatty acid profile, being low in saturated fatty acids (3.02 g/L) and high in alpha-linolenic acid (about 7.25 g/100 g of fat). Until now, the beneficial properties of donkey milk have been mostly related to whey proteins, among which β-lactoglobulin is the most represented (6.06 g/L), followed by α-lactalbumin (about 2 g/L) and lysozyme (1.07 g/L). So far, the health functionality of donkey milk has been tested almost exclusively on animal models. Furthermore, in vitro studies have described inhibitory action against bacteria, viruses, and fungi. From the literature review emerges the need for new randomized clinical trials on humans to provide stronger evidence of the potential beneficial health effects of donkey milk, which could lead to new applications as an adjuvant in the treatment of cardiometabolic diseases, malnutrition, and aging.

1. IntroductionDonkey milk (DM) has been historically considered a therapeutic food in both Western and Eastern cultures. Hippocrates (460–370 BC) [1] and Pliny the Elder (23–79 AD) [2] were among the first to describe DM health benefits. The beneficial properties of DM are also reported in traditional Chinese medicine manuals [3].The studies of Dr. Parrot of l’Hospice des Enfants Assistès (Paris, France) in the nineteenth century were probably the first scientific approach to the use of DM in infant feeding. Dr. Parrot fed children affected by congenital syphilis directly from the donkey’s udder. In particular, Dr Parrot carried out the first controlled trials on breastfeeding with DM, comparing DM with cow and goat milk, recording the milk intakes, the weight gains of the children, and analysing DM chemical composition. Dr Parrot’s studies led to the creation of a donkey farm for the purpose of feeding orphaned children [4].In recent decades, alongside with the rediscovery of its potential beneficial effect, DM is becoming popular in Europe, especially in Croatia, France, Hungary, Italy, the Netherlands and Serbia, and in several Asian countries. In particular, China is a large donkey meat and milk producer, and the donkey industry has become important in rural China. DM is considered the natural milk with the closest composition to human milk in terms of lactose content and protein and amino acid profile [5]. Scientific studies aimed at clarifying the composition of DM and the presence of functional compounds have increased over the past years. Until now, investigations have been carried out to highlight potentially bioactive substances, such as polyunsaturated and omega 3 fatty acids [6,7], functional proteins [8,9], vitamins [10,11], polar lipids [12], phytosterols [13], and the milk compositional variability [14,15].New knowledge has emerged leading to the development of studies focused on testing in vitro and in vivo the potential effects of DM in humans. This paper examines the scientific evidence regarding the effects of DM on human health and its possible applications as an adjuvant in the treatment of cardiometabolic diseases, malnutrition, and aging. 2. Use of DM in Allergic ChildrenDM has nutritional similarities with human milk (HM), the gold standard for infant feeding, in terms of average protein content (about 18 and 21 g/L in DM and HM, respectively) [5]. Caseins (CN), which are main allergenic components of cow milk (CM), are less represented in DM and HM (56% and 30% of the total protein in DM and HM vs. 80% of CM). A major similarity is also linked to the primary structures of αS1-, β- and κ-CN, which are closely related in the HM and DM [16]. The main CN fraction of DM is β-CN (about 6.11 g/L; 62% of total CN), whose concentration is between the minimum values found in HM (1.25–4.72 g/L) and the maximum in CM (11.85–12.87 g/L). The mean concentration of αS1-CN in DM is 2.54 g/L (26% of total CN), higher than HM (0.33–0.50 g/L) and lower than in CM (8.52–9.16 g/L); αS2-CN and κ-CN are minor component of DM [9].Simulated in vitro digestion showed that donkey CN has rapid degradability and an almost complete digestibility both when artificial [17,18] and gastrointestinal fluids human fluids [19,20] were used.This can also explain the reduced allergenicity of DM since food protein allergenicity is linked to the survival of allergens in the gastrointestinal tract.On the other hand, the main DM protein fraction is made up by whey proteins (WP), which have shown multiple beneficial metabolic and antimicrobial properties [21]. β-Lactoglobulin (β-LG) is the most represented WP (6.06 g/L; 73% of total WP), absent in HM but present in CM in a concentration of about 5 g/L. The α-lactalbumin (α-LA) content in DM (about 2 g/L) is similar to HM (3 g/L) [9].The similarities between HM and DM are at the basis of a well-known application of DM: its use in the diet of children suffering from allergies to CM proteins (CMPA). In this regard, most clinical trials on humans proved the tolerability and efficacy of DM in these patients (Table 1).Most trials in the literature are single-arm prospective longitudinal studies [22,23,24,25], while there are still few randomized studies involving a control group [26]. The papers available so far have included a limited number of subjects (<100 enrolled patients) with a rather wide age range, on average between 27 months and five years [24,25], suffering from CMPA, food protein-induced enterocolitis syndrome induced by cow milk (CM-FPIES), and multiple food allergies.The reported tolerability of DM in allergic children varies between 82.6% [23] and 98.5% of patients [22]. In addition, a clinical study in infants under six months of age, conducted on a low number of subjects (six infants) with CM-FPIES, has found that DM is well tolerated also in younger patients, with no participants showing allergic reactions [27].Even though encouraging results support the consumption of DM in children with food allergies, some cases of hypersensitivity reactions to DM have been reported in children and in two adults suggesting caution in allergic subjects [28,29,30,31].animals-11-01382-t001_Table 1Table 1Studies on the tolerability of donkey milk in children.Study DesignNumber of ChildrenMean Age Duration of Diet Tolerance OutcomeReferenceDouble-blind placebo-controlled food challenge30 with the IgE- and non-IgE-mediated CMPA 12.5 years (from 0.6 to 3.8 years)3 months96%[26]Prospective study; double-blind, placebo-controlled food challenge46 with IgE- and non-IgE-mediated CMPA36 months (from 12 to 149 months)24 months82.6% of the total patients (78.8% of the children with IgE-mediated CMPA)[23]Prospective study92 highly-problematic children with IgE- and the non-IgE-mediatedCMPA27.3 months (from 7.5-to 121.5 months)48 months87% children with non-IgE-mediated CMPA (20/23) 91.3% with IgE-mediated CMPA. (63/69)[24]Open challenge70 children including patients with prior anaphylaxis to CM 35.2 ± 5.3 months (from 6 months to 18 years)/98.5%[22]Open challenge70 children with proven IgE-CMPA;11 patients with proven IgE-FPIES5.2 ± 5.3 years (from 6 months to 18 years);4.73 ± 1.68 months (from 3 to 8 months)/98.7%[25]Open challenge6 with CM-FPIES 23.6 months (from 1.5 to 6 months)/100%[27]Open challenge30 with IgE- and non-IgE-mediated CMPA4,5 years (from 6 months to 11 years)/96%[31]1 CMPA: cow milk protein allergy; 2 CM-FPIES: food protein-induced enterocolitis syndrome induced by cow milk; 3 CM: cow milk.From the point of view of palatability [23,25], DM is well accepted by the children. The good palatability is probably related to the high concentration of lactose [5] and to the fact that allergic children often follow restrictive nutritional plans and monotonous diets due to multiple food allergies. Regarding nutritional efficacy, research shows that (Table 2), despite the majority of allergic subjects have negative weight and length/stature Z-scores due to feeding difficulties, growth parameters improve after DM supplementation [23,24,26,31]. The positive effect of DM on growth is probably related to the ability of the milk to fill some nutritional gaps in the diet of treated subjects [26]. Even in infants under one year of age, if properly integrated, DM did not show negative effects on the growth [25]. Nonetheless, the fat percentage of DM is usually three times lower than HM and CM, and consequently its energy content is also lower (about 40 vs. 62 and 65 kcal/100 g, respectively) [5]. According to nutritional recommendations of different international organizations, fat should provide 40–60% of the daily energy intake in pre-weaned children (between 0 and 6 months) and should be gradually reduced to 35% in children 2 years old [32]. Therefore, since DM has a lower fat and caloric content compared to other milks used in infant feeding, fat needs to be supplemented in children consuming an exclusively milky diet, e.g., by adding vegetable oils.The Diagnosis and Rationale for Action against Cow’s Milk Allergy (DRACMA) guidelines report that equine milks can be considered valid substitutes for CM even if they cannot be a treatment of choice for CMPA [33]. The choice of alternative milks should take into account the individual clinical profile of the child allergic to CM, particularly as concerns age, severity of symptoms, sensitivity to CM proteins and associated food allergies [34]. Recently, a line of research regarding the application of DM as a fortifier in the feeding of premature newborns has developed [35,36]. Human milk fortification is a routine clinical practice for feeding preterm infants to ensure protein and energy intakes of critical importance in preterm infants. To evaluate the effect of HM fortification with DM compared to CM, Bertino et al. [35] designed an open label, randomized controlled clinical trial on a total of 156 very preterm newborns (gestational age <32 weeks; very-low-birthweight <1500 g) giving two isocaloric and isoproteic diets. The authors found that a DM-based fortifier seems to improve feeding tolerance, with a similar auxological outcome in the first 21 days of enteral diet compared to a CM-based fortifier. A follow-up analysis on 122 of these children at 18 months of age reported that the fortifier derived from DM have similar long term auxological outcomes compared with the standard CM-derived fortifier [30]. Finally, ancillary studies of Bertino et al. [35] and Peila et al. [37] have also shown that the DM compared to CM fortifier reduced the episodes of gastroesophageal reflux (GER), which frequently occur in very-low-birthweight infants [36]. GER is worsen by food intolerance and can be associated with cardiorespiratory symptoms [38]. In the paper by Cresi et al. [37], very preterm infants taking DM also had a lower frequency of weakly acidic reflux (characteristic of GER) compared to the control group. Given the lower buffering capacity of DM compared with CM, DM did not affect the physiological acid reflux. The authors [37] state that DM minor buffering effects on gastric pH could be a protective factor, preventing infections and necrotizing enterocolitis in premature infants.animals-11-01382-t002_Table 2Table 2Studies on the effects of donkey milk on the growth of infants and children.Study DesignNumber of ChildrenAge Diet Auxological OutcomeReferenceProspective study16 with IgE-CMPA 1 and 6 CM-FPIES 2 20 months (range 9–79 months).Integrated with DM for 6 monthsNo negative influence [25]Randomized controlled trial156 preterm infants (77 assumed DM 3–fortifier)11 days (median age) DM- fortifier vs. CM 4- fortifier; isocaloric and isoproteicdiets for 21 days Similar auxological outcomes than control group [35]Randomized controlled trial122 children (77 assumed DM fortifier)18 monthsDM–derived fortifier vs.CM fortifierSimilar auxological outcomes than control group[37]1 CMPA: cow milk protein allergy; 2 CM-FPIES: food protein-induced enterocolitis syndrome induced by cow milk; 3 DM: donkey milk; 4 CM: cow milk.3. Immunomodulatory Effects DM is particularly rich in lysozyme (LZ) [8], an enzyme that breaks the peptidoglycan layer of Gram-positive bacteria. The average LZ concentration of DM is on the average 1.07 g/L (13% of total WP) [9], similar to HM (0.3–1.1 g/L; [39]) and higher than CM, in which negligible amounts of LZ have been reported. LZ activity for DM ranges from 1670 to 11,531 U/mL [8,40], while it is barely detectable in CM (0.0292 U/mL) and the highest in HM (about 39,000 U/mL) [41].According to by Mao et al. [42], the WP fractions containing LZ are responsible for DM immunomodulatory effect. Whether other components contribute to this action, such as α-lactalbumin (α-LA) and lactoferrin, remains to be determined. In fact, α-LA seems to regulate the overall immune function infants [43]. Recently, investigations on milk oligosaccharides have also shown positive effects on immune system development. However, research on oligosaccharides regards mostly HM, whereas DM is less studied [44].DM have shown immunological activities in vitro tests [42,45] and randomized controlled studies in animal models [46] and humans [47], in whom DM can induce the release of some cytokines, proteins that regulate the inflammatory and immune response to infections (Table 3 and Table 4). DM has been shown to increase cytokines involved in the regulation of innate immunity and the onset of local acute inflammatory response: interleukin 1 (IL-1) [42,45,47], interleukin 6 (IL-6) [42,47] and tumor necrosis factor α (TNF-α) [42,47] both in vitro [42] and in vivo [47]. Differently, Jiang et al. [46] have reported an inhibition of TNF-α in mice with inflammatory bowel disease (IBD).Furthermore, the WP fraction of DM with a molecular mass >10 kDa has been shown to stimulate the production of specific immunity regulatory cytokines such as interleukin-2 (IL-2) and interferon γ (IFN-γ) by murine splenocytes [42]. DM has also induced the release of interleukin-10 (IL-10), which is responsible for reducing inflammatory reactions, helping the elimination of pathogens and reducing the infection damage [47].The inhibition of the expression of inflammation mediators, in particular interleukin 13 (IL-13) and the already-mentioned TNF-α, has been observed by Jiang et al. [46] on mice with inflammatory disease.The only study concerning the immunomodulatory effects of DM in humans has involved elderly subjects (14 healthy aged subjects; from 72 to 97 years), in which DM vs. goat milk was administered [47]. These authors observed that the administration of DM (200 mL/day for one month) acts as an enhancer of the acute phase response in humans. Therefore, DM daily use may be recommended in the diet of immuno-compromised elderly patients. 4. Potential Antioxidant and Antihypertensive EffectsThe antioxidant activity of DM has been tested in double-blind randomized studies on animal models [17,48,49] (Table 4). DM-treated rats have shown an enhancement in antioxidant defense mechanisms and detoxifying enzymes [17,48,49].Specifically, Li et al. [17] found that DM intake tended to increase the superoxide dismutase (SOD) activity in the plasma of diabetic rats compared to untreated rats. SOD enzyme alternately catalyzes the dismutation (or partitioning) of the superoxide (O2−) radical into ordinary molecular oxygen (O2) and hydrogen peroxide (H2O2). Furthermore, Li et al. [17] showed that total anti-oxidation capacity is also improved in diabetic rats treated with DM compared with the untreated group, towards values seen in healthy (control) rats.In mouse models, improvements in glutathione/glutathione disulfide ratio in liver (i.e., an oxidative stress marker) were also observed, as well as increased activities of liver detoxifying enzymes (glutathione-S-transferase—NAD (P) H: Quinone Oxidoreductase) [48,49].Antioxidant activities directly measured in DM and fermented DM (kefir) by mean of ABTS (2, 2′-Azino-bis (3-ethylbenzothiazoline-6-sulphonic acid) and DPPH (2, 2 diphenyl-1- picryl hydrazyl) assays have been found higher in kefir than raw DM and increased after in vitro simulated gastrointestinal digestion [18]. A role of fermentation in DM antioxidant activity has also been ascribed to bacteria, in particular Enterococcus faecium DM33, as fermented milk containing this bacterium exhibited the strongest antioxidant activity [18].However, antioxidant activities could be related to the release of bioactive peptides through the enzymatic hydrolysis of proteins. A peptidomic study [50] identified 1330 peptides from commercial donkey milk, mainly coming from β-CN, αS1-CN and serum amyloid A protein. Moreover, β-LG I and lactoferrin can be source of milk peptides, while α-LA and LZ are resistant to gastrointestinal enzymes [16]. DM peptide fractions tested by in vitro bioassays have shown antioxidant activities [19,51].Many peptides in DM have typical characteristics of angiotensin-converting enzyme (ACE)-inhibitory peptides, potentially reducing the activity of ACE [16]. DM fractions containing different peptides confirmed angiotensin converting-enzyme inhibitory actions when tested by in vitro bioassays [50,51].None of the potentially bioactive peptide identified in DM by Zenezini Chiozzi et al. [51] exactly matches sequences of known bioactive peptides. However, potential ACE inhibitory peptides (namely MPFLKSPIVPF) had a similar sequence and the same length to a confirmed antihypertensive peptide (namely MPFPKYPVQPF) which was previously found in Gouda cheese.Therefore, milk-derived bioactive peptides may potentially decrease the formation of angiotensin II and increase bradykinin levels, which have vasoconstrictor and vasodilator properties, respectively. The actions act synergistically in lowering blood pressure. In this regard, fermented DM (with Lactobacillus casei DM214) showed ACE-inhibitory activity in vitro [18]. In addition, the release nitric oxide (NO) from peripheral blood mononuclear cells (PBMCs) treated with DM was observed in vitro. Since nitric oxide (NO) is a strong vasodilator, a role for DM in the prevention of atherosclerosis has been proposed [45]. 5. Effects on Glucose Metabolism and Potential Coadjutant Action in the Diabetes TreatmentLactose is the main carbohydrate in DM, as in HM and CM (about 70, 60 and 49 g/L of milk respectively) [5,41] and it is responsible for the osmotic equilibrium between blood and alveolar lumen in the mammary gland. Lactose can assume two anomeric forms (α-lactose, and β-lactose) on the basis of the glycosylic bond (1,4) that connects the carbon atom 1 of galactose and the carbon atom 4 of glucose [52].Lactose intolerance is common in the adult population mostly due to the loss of intestinal lactase; its prevalence has great geographical variability [53]. Many intolerant individuals can tolerate low levels of lactose in their daily diet (about 5–10 g of lactose distributed throughout the day) and in general the use of fermented products and of lactase supplements can overcome the problem [52]. To the best of our knowledge, there are no specific studies regarding the use of DM in subjects with lactose intolerance. Lactose-free DM is currently not available on the market.Lactose, WP, and bioactive peptides in DM have been shown to be involved in insulin response to glucose [54]. In particular, LZ and α-LA from DM may have a role in the prevention and treatment of diabetes [17]. α-LA is a highly represented protein in DM (on the average 1.22 g/L; 14.69% of total WP) and HM (2.6–4.2 g/L) and plays a key role in lactose synthesis in the mammary gland.Beneficial glucometabolic properties of DM have been observed in animal studies [17,48,49]. However, the active components of DM and the biological mechanisms underlying these effects are still under study.Trinchese et al. [48] tested DM in healthy rat groups taking different isoenergetic diets and observed that the diet supplemented with DM improved glucose disposal and insulin resistance, leading to reduction of glucose levels and better tolerance to glucose loads, compared with the groups not receiving milk-based supplements (control) and taking CM.Positive effects of DM on glucose metabolism were also observed in rats with streptozocin-induced type 2 diabetes. In this animal model, DM powder supplements reduced blood glucose levels and insulin resistance after four weeks. Remarkably, the anti-diabetic effect of DM was similar to metformin treatment in most biochemical parameters [17].Along with insulin resistance, pancreatic β-cell dysfunction is typically involved in diabetes development and progression [55]. In this regard, it is noteworthy that DM was able to improve the viability of damaged clonal β-cells (mouse insulinoma β-pancreatic (MIN6) cells) [17].The DM beneficial effects on glucose metabolism are, at least in part, attributable to:(1)reduction of inflammatory status and leptin/adiponectin ratio. The animals treated with DM showed a reduction in serum inflammatory mediators and in the leptin/adiponectin ratio [48]. These two hormones, derived from adipocytes, are involved in lipid metabolism, energy homeostasis and inflammation [56,57]. A high leptin to adiponectin ratio is related to insulin resistance [58] and a decrease in adiponectin was found linked to the onset of type 2 diabetes in animal models [59].(2)enhancement of antioxidant defense mechanisms [17], which protects against the development of insulin resistance.(3)modulation of mitochondrial dynamics that impacts on mitochondrial metabolism. Alteration of mitochondrial dynamics, function and efficiency has impact on several pathological conditions including metabolic diseases such as obesity and type 2 diabetes [60]. DM-treated rats showed more abundant, larger and electron-dense mitochondria in the skeletal muscle at electron microscopy analysis [49]. These characteristics have been associated to more active mitochondria with higher respiratory capacity and improved glucose metabolism [61].(4)down-regulation of two gluconeogenesis key enzymes: phosphoenolpyruvate carboxykinase 1 (Pck1) and glucose-6-phosphatase (G6PC) [17].6. Effects on Lipid MetabolismSome nutritional peculiarities of DM support its use in low-calorie diets and in the management of dyslipidemia. In fact, as previously discussed, DM has a lower calorie and fat concentration (fat ranges from 0.20% to 1.7% in DM) compared with other milks used in human nutrition. In addition, the amount of saturated fatty acids (SFA) in DM is significantly lower than CM (3.02 g/L vs. 26.27 g/L, respectively), while the UFA: SFA ratio is higher (0.75 vs. 0.41, respectively). Furthermore, DM is the richest source of alpha-linolenic acid (C18:3 n-3;ALA) among farm animal milks (about 7.25 g/100 g of fat). C18:3 n-3 ALA is a precursor for long n-3 fatty acids and has beneficial health effects [41]. In their controlled studies on murine models, Li et al. and Trinchese et al. [17,48] also found that DM has beneficial effects on lipid metabolism. In DM-fed animals, significantly lowered blood triglycerides and reduced fat accumulation have been observed, which were attributed to beneficial effects on the skeletal muscle. In fact, skeletal muscle mitochondria of DM-fed animals showed increased respiratory capacity and fatty acid oxidation [48].This effect is due to:An increase in oleylethanolamide (OEA) in the skeletal muscle and in the liver [48]. OEA increase is probably related to the high concentration of palmitic acid in the sn-2 position of the triacylglycerol backbone of DM [5,62]. This type of esterification is similar to that occurring in HM and allows a more effective C16:0 absorption since 2-monoacylglycerols of SFAs are more easily absorbed than free fatty acids (FFA). OEA has been identified as an important regulator of lipid metabolism and can enhance fatty acid oxidation in rats [63].Enhancement of carnitine palmitoyl-transferase (CPT) activity: Increased respiratory capacity in the skeletal muscle is likely related to an enhancement of CPT activity, which would further increase the entry of long-chain FFAs into the mitochondria, stimulating fatty acid oxidation [48]. CPT is a mitochondrial enzyme responsible for the formation of acyl carnitines by catalyzing the transfer of the acyl group of a long-chain fatty acyl-CoA from coenzyme A to l-carnitine. This reaction allows the increase in lipid oxidation for the movement of the acyl carnitine from the cytosol into the intermembrane space of mitochondria.Modulation of mitochondrial function, efficiency, and dynamics: Mitochondrial uncoupling is a dissociation between membrane potential generation and its use for ATP synthesis [64]. Mitochondrial uncoupling dissipates the proton gradient across the inner membrane and creates a futile cycle of glucose and fatty acid oxidation without generating ATP [65], thereby increasing lipid oxidation and reducing intracellular lipid content [66]. Mitochondrial uncoupling induces a less efficient utilization of lipid substrates. This decline in mitochondrial energy efficiency may also contribute to fat burning. Promoting this inefficient metabolism that generates heat instead of ATP, mitochondrial uncoupling can serve as a potential treatment for obesity [64].7. Antiproliferative and Antitumor EffectThe literature contains conflicting reports regarding the relation between the consumption of milk and dairy products and cancer. As regards DM, only a few studies have investigated a possible antiproliferative and antitumor capacity with heterogeneous results.Indeed, Mao et al. [42] have observed an antiproliferative and antitumor effect of DM WP on A549 human lung cancer cells in a dose-dependent and time-dependent manner. From observations on murine splenocytes, the same authors have concluded that DM WP kill tumor cells through activation of lymphocytes and macrophages.A recent study in vivo reports that DM reduces primary tumor size and inhibits breast tumor progression in 4T1 mice by inducing apoptosis [67].According to Esener et al. [68], the release of NO could mediate the DM tumoricidal activity. NO release by PBMCs after DM exposure was also observed by Tafaro et al. [45].Although the exact role of NO in cancer biology is not fully understood, it seems that high NO concentrations exert a controlling influence on immune-mediated antitumor activities, whereas low concentrations facilitate cell survival and proliferation [69]. On the other hand, NO is a potential oncogenic molecule that promotes neovascularization and reduces blood flow in tumor tissues. Moreover, high concentrations of NO can directly cause DNA damage [68,70].A randomized study in rats with Ehrlich ascites carcinoma (EAC) found that administration of DM kefir for 10 days reduced the carcinoma volume and increased the number of apoptotic cells compared with the control group [68]. The effects of DM kefir on tumor volume and apoptosis were ascribed to the down-regulation of the NO synthase enzyme (isoforms iNOS and eNOS). In particular, iNOS levels were markedly higher in the control and DM groups compared to the DM kefir group [68]. However, in contrast with the findings of Mao et al. [42] and Li et al. [67], Esener et al. [68] reported that DM was not effective on the carcinoma and the groups of rats treated with unfermented DM showed less numerous apoptotic cells. 8. Protection of the Intestinal Barrier and Modulatory Effect of the Intestinal FloraMilk oligosaccharides are a complex class of bioactive carbohydrates without direct nutritional value [71].The nutritional importance of oligosaccharides in milk is due to their prebiotic role (reported for breast milk). In HM, oligosaccharides serve as a substrate for beneficial gut microbiota by acting as prebiotics and inhibit the intestinal adhesion of pathogenic microorganisms, thus limiting the onset of enteric infections. Despite the possible protective role of oligosaccharides on the intestine and the interest in nutritional applications in specific categories of consumers, DM oligosaccharides have been poorly studied [44].Currently, the results of studies on animal models suggest that DM, and in particular DM LZ, could play a role in the treatment of IBD [46,72,73].LZ is present in higher quantity in DM compared to ruminant milks and seems quite resistant to human gastrointestinal enzymes (75% resistance) in vitro [74]. Although human studies in infants are still controversial [43], investigations in mouse models show that LZ can reach the intestinal tract intact [72].Jiang et al. [46] carried out a randomized study on rats gavage fed with fractions of DM containing WP with different percentages of LZ. After 14 days of treatment, colitis was induced in mice (via dextran sulfate sodium). The authors found a protective action of DM WP on the disease: a reduction of symptoms and the improvement of lesions in treated mice compared with controls. The active component responsible for these actions was found to be LZ, while α-LA and β-LG had no significant effects on colitis symptoms. The DM action was linked to a reduction in the mediators of inflammation, but also to the protection of the intestinal mechanical barrier function [46].Anti‑inflammatory properties of DM on the intestine have been reported in a randomized study on a model of ileitis in mice (ileitis induced by indomethacin) [72]. In this study, oral DM treatment attenuated the severity of symptoms and macroscopic and microscopic damages (drop of body weight, reduction in the length of the small intestine, increase in fecal lipocalin-2). The authors [72] ascribed these actions to the normalization of the intestinal immunity function, in particular to the expression of antimicrobial peptides by the Paneth cells which contribute directly to reduce dysbiosis.Both raw and thermized DM counteracted chronic stress-induced intestinal damage, gut hyper-permeability and inflammation in mouse models of chronic stress [73]. Yvon et al. [73] highlight the importance of DM LZ activity in the reduction of intestinal damage. DM was also found to modulate the intestinal microbiota and increase the microbial diversity in healthy rats [49] and in mouse models with IBD [46]. A positive modulation of the microbiota was also observed in the study of Yvon et al. [72], though this effect was linked to indirect actions on the intestinal immunity of the host rather than direct actions on the microflora.According to Penders et al. [75], the beneficial effects of DM on the microbiota could represent a key source of immune development and regulation in early life and could have a preventive role against the development of atopic dermatitis.animals-11-01382-t003_Table 3Table 3In vitro studies on the beneficial effects of donkey milk.Experimental ModelEffectsReferenceMouse insulinoma beta-pancreatic (MIN6) cellsAnti-diabetes action:DM in the medium (500 μg/mL) improved the viability of damaged pancreatic beta-cells[17]DM 1 and fermented DM samplesAntioxidant activity of fermented DM samplesAntihypertensive effect (ACE-inhibitory activity) in fermented DM [18]Murine splenocytesImmunological modulation: increase in IL-1, IL-6, TNF-α, IL-2 and IFN-γ[42]A549 human lung cancer cellsAnti-proliferative activity induced by DM whey protein (MW 2 > 10 kDa), Human peripheral blood mononuclear cellsImmunological activities: increase in IL-1 and IL-10[45]1 DM: donkey milk; 2 MW: molecular weight.animals-11-01382-t004_Table 4Table 4In vivo randomized controlled studies on the beneficial effects of donkey milk.Experimental ModelTreatmentEffectsReference Balb/c mice with induced colitis3 DM 1 whey fraction (5%, 20% and 50% of lysozyme) for 14 daysImmunological activities: inhibition of IL-13 and TNF-αImprovement in the intestinal barrier and modulatory effects on the gut microbiota.[46]14 elderly subjects ( from 72 to 97 years old)200 mL/day of DM for one monthImmunological activities: increase in IL-1, IL-6 and TNF-α[47]Wistar rats3 g/kg day of DM powder for 4weeks.Antioxidant effects: tendency to increase SOD 2 activity in the plasma of diabetic rats[17] Improvement of metabolism:Reduction in the blood glucose on type 2 diabetic rats and in insulin resistance Wistar rats 48 mL/day of DM, for 4 weeks Antioxidant effects: improvements in oxidative stress markers in the liver; increased activities of liver detoxifying enzymes, increase of antioxidants [48,49] Improvement of metabolism: improved glucose disposal; decrease of blood triglycerides and of fat accumulation in muscles; modulation of the intestinal microbiota Swiss albino mice. with Ehrlich ascites carcinoma tumour0.5 mL/day of DM or kefir of DM for 10 daysAnti-proliferative activity: reduction in tumor volume and increased number of apoptotic cells in the groups treated with fermented DM, not in the groups treated with unfermented DM[68]C57BL/6 miceileitis induced Orally treated withDM with the same total daily activity of lysozyme, i.e., 11800UI in a total adjusted volume of 0.4 (± 0.05) mL for 7 days Reduction of dysbiosis by mean of stimulation of the intestinal innate immunity[72]1 DM: donkey milk; 2 SOD: superoxide dis-mutase.9. Antibacterial PropertiesDM showed antibacterial properties that may be linked to a synergistic activity of LZ, lactoferrin and some FFA such us lauric, oleic and linoleic acids [76,77]. Lactoferrin exhibits antibacterial, antiviral, antifungal and antiprotozoal activities [39]. Differently from HM, lactoferrin is a minor component in DM (0.097–0.133 g/L), has poor thermal resistance [78] and is easily digested by gastric and duodenal juice. Thus, it has been suggested that lactoferrin can play its biologic role in vivo mainly through its bioactive peptides derived from digestion [79].In addition to lactoferrin, α-LA, highly concentrated in DM, may contribute to inhibit the growth of potential pathogens, as reported both in vitro and in vivo for the human protein [39].The antibacterial effect of DM could also be mediated by the microflora of the milk itself. DM contains Lactobacillus plantarum that was described to produce bactericidal bacteriocins [80]. In addition, antimicrobial activities were found to be higher in fermented than raw DM and were further increased after in vitro simulated gastrointestinal digestion [18], suggesting that the formation of bioactive peptides may play an additional role in this effect.Several authors confirm the efficacy of DM in inhibiting the growth of specific foodborne pathogens in vitro (Table 5), specifically Gram + bacteria such as:(a)Listeria monocytogenes (2230/92, ATCC 19111; ATCC: 13932), which was inhibited at concentration of 1% by DM, in vitro digested DM [19] and also in situ, in artificially contaminated milk [76,81].(b)Staphylococcus aureus (ATCC 8095; ATCC: 6538). DM (50 folds concentrated) was active at minimal lethal concentration of 64 mg/mL [82], also in situ [76,81,83]. However, some authors [82,84] found that this antimicrobial activity is reduced by digestion.(c)Enterococcus faecalis (DSM 2352), which was inhibited by hydrolyzed DM [84].animals-11-01382-t005_Table 5Table 5Studies on effects against microorganisms of donkey milk.MicroorganismExperimental ModelReferenceListeria monocytogenes (2230/92; ATCC 19111; ATCC: 13932)digested in vitro DM 1 and DM at concentration 1% on microtiter plates;[19] in situ[76,81]Staphylococcus aureus (ATCC 8095) (minimal lethal concentration of 64 mg of DM concentrated to 50 folds/mL)agar well diffusion[82]Staphylococcus aureus (ATCC 8095 ATCC 25923 ATCC: 6538) and (DSM 25923) (by hydrolyzed DM)in situ[76,81,83,84]Enterococcus faecalis (DSM 2352)hydrolyzed DM milk tested by inhibition halos test on agar plates[84]Salmonella enterica serovar choleraesuis (CGMCC 1.1859)Agar diffusion test [85]Salmonella serovar enteritidis (ATCC 13076). and serovar Typhimurium (ATCC 14028)In situ[81,83];Shigella dysenteriae(CGMCC 1.1869)agar diffusion test and in situ[85]Microsporum canis, and Microsporum gypseum (failed to grow at a concentration of 60% and 70% of donkey milk respectively)microdilution test[77]Trichophyton mentagrophytes and T. rubrum (minimal lethalconcentration 32 mg of 50 folds concentrated DM/mL)agar well diffusiontest[82]Echovirus type 51 mg of DM WP2 fractions/mL in the medium of growth of infected culture of human intestinal epithelial cell line Caco-2[84]1 DM: donkey milk; WP: whey protein.Relating to Bacillus cereus, DM showed variable efficacy depending on the strain [82,84]. For example, DM was less active against B. cereus DSM 4384 [84], while B. cereus RT INF01 appeared very resistant [19].An antibacterial activity was also reported against the Gram—Salmonella enterica serovar choleraesuis (CGMCC 1.1859), serovar enteritidis (ATCC 13076) and serovar Typhimurium (ATCC 14028). DM was active against these bacteria on the agar diffusion test [85] and in artificially contaminated milk [81,83]. Shigella dysenteriae (CGMCC 1.1869) was also sensitive to DM on the agar diffusion test and in situ. In fact, the counts of viable S. dysenteriae decreased to below detectable levels in artificially contaminated milk [85].Regarding Escherichia coli, the results in the literature are conflicting and probably related to a strain-dependent activity. In fact, some authors [19] showed that DM and in vitro digested DM caused growth reduction on E. coli (EPEC) 10208355 during its stationary phase at concentrations of 0.6 and 1.0%. Moderate antibacterial activities against E. coli strain ATCC 25922 were also observed by Koutb et al. [82] with milk concentrated 50 times (minimal lethal concentrations of 128 mg/mL). In studies in situ on artificially contaminated milk, an inhibitory activity of DM on the development of E. coli has been reported [83]. Differently, the growth of the strain C84010 was not inhibited by DM on agar diffusion assay [85], and the toxicogenic E. coli DSM 8579 was very resistant to in vitro digested DM [84].10. Antifungal and Antiviral PropertiesDM proved to be inhibitory against fungi (Table 5), particularly some dermatophytes [77,82], with the potential to prevent and control the infection of these zoonotic fungi in humans.Microsporum canis, and Trichophyton mentagrophytes failed to grow in 60% DM and Microsporum gypseum appeared to be sensitive to 70% DM [77]. Also, Koutb et al. [82] reported a minimal lethal concentration of 32 mg/mL against T. mentagrophytes and T. rubrum, using a 50-fold concentrated milk, while no growth inhibition was observed when testing Candida albicans. The activity against dermatomycotic fungi was not affected after digestion of DM with pepsin by Koutb et al. [82].The only paper that investigated the antiviral activity of the DM reports its effectiveness against Echovirus (Enteric Cytopathic Human Orphan virus) Type 5 [86]. Echovirus is a small, non-enveloped, single stranded RNA virus, belonging to the genus Enterovirus of the Picornaviridae family, acquired by fecal–oral contamination, and infecting the gastrointestinal tract as the primary organ [87]. Infections with echoviruses have been associated with a wide variety of neurological and exanthematic diseases [88].Among the different protein fractions of DM tested on human intestinal epithelial cell lines (Caco-2) infected with Echovirus Type 5, WP showed the greatest inhibition on virus replication [86]. In particular, DM antiviral activity on echovirus type 5 seems due to a synergic action of high molecular mass WP, such as lactoferrin, lactoperoxidase, serum albumin, and immunoglobulins [86].11. ConclusionsSeveral clinical studies report that DM shows high tolerability in children with food allergies, while DM immunomodulatory properties have been described in animal models and in a single study in humans. Research on murine models shows that DM modulates the intestinal microbiota, enhances antioxidant defense mechanisms and detoxifying enzymes, and is effective in controlling blood sugar and dyslipidemias. Although the first in vitro study on the antiproliferative and antitumor effect of DM yielded promising results, the few available trials in animal models show conflicting findings. Finally, in vitro studies describe inhibitory actions of DM on bacteria, viruses and fungi. So far, the observed beneficial properties of DM have been tested almost exclusively in vitro and in animal models and have been mostly related to some WP. From this literature review, there emerges a need for new randomized clinical trials on DM consumption in humans to provide stronger evidence of its potential beneficial health effects, which could lead to new applications of DM as an adjuvant in human medicine.

animals : an open access journal from mdpi

10.3390/ani11082204

PMC8388470

The red imported fire ant (RIFA, Solenopsis invicta) is an exotic pest that can harm humans and animals, cause economic loss to agriculture, and damage ecosystems. In the present study, we devised a practical method to train detection dogs without introducing live RIFAs and an alternative way to correctly identify RIFA-invaded zones. Both live RIFA-experienced and inexperienced detection dogs successfully indicated RIFA-scented filter papers and live RIFAs with a high positive indication rate (>93%) and low false response rate (2%), and also performed successfully in field studies. In addition, the RIFA-scented filter papers can be stored at −20 °C and 4 °C at least 13 weeks for dog identification. Using filter paper as a RIFA odor bearer for detection dog training and RIFA identification is an effective and economical method in order to decrease the risk of RIFA introduction into uninvaded or eradicated areas.

The red imported fire ant (RIFA, Solenopsis invicta) is an exotic aggressive pest that is notorious for its ability to seriously harm humans and animals, cause economic loss to agriculture, and damage ecosystems. This is the first study to validate the capability of filter paper adsorption as a feasible odor bearer of RIFAs and evaluate its use in detection dog training. Two live RIFA-experienced detection dogs achieved a mean 92% positive indication rate (PIR) on RIFA-scented papers with a relatively low false response rate (0.8%). The similar accuracies in recognizing live RIFAs (96%) and scented papers (92%) suggest that a filter paper is an effective odor reservoir. After training with live RIFA and scented filter papers, both RIFA-experienced and inexperienced detection dogs successfully indicated filter papers that were scented with at least 10 RIFAs for 4 h with a high PIR (>93%) and low false response rate (2%). Detection dogs correctly recognized the filter papers scented by 10 RIFAs for 24 h with a 97.6% PIR. Even for scented samples stored at −20 °C and 4 °C for 13 weeks, the positive indication rates (PIRs) were as high as 90%. These results suggest that filter paper is an effective RIFA odor bearer, and the scent can be maintained at least 13 weeks for dog identification. After RIFA-scented paper training, detection dogs showed high (>95%) PIRs for both RIFA-scented paper and live RIFAs and also successfully performed field studies. Using filter paper as a RIFA odor bearer is an effective and economical method for detection dog training and RIFA identification.

1. IntroductionThe red imported fire ant (RIFA) Solenopsis invicta Buren is an exotic aggressive pest that is notorious for its ability to seriously harm humans and animals, cause economic loss to agriculture, and damage ecosystems [1]. Originally from South America, S. invicta invaded the United States in the 1930s and then spread through the southern United States and Caribbean islands [2,3]. In the early 2000s, RIFA territory expanded across the West Indies to the Pacific region; they were first discovered in New Zealand and Australia and later in Taiwan, Hong Kong, Macao, and China [4,5,6,7,8,9].The ecological and economic impact of RIFA invasion and infestation is enormous [1]. Without strong competitors, fire ants promptly become the dominant ant species in infested regions. The populations of native ants and arthropods are reduced, and native vertebrates such as mammals, birds, and herpetofauna also experienced adverse effects [10]. Immediate burning and itching at sting sites are the characteristic experiences, and hypersensitivity reactions and possibly secondary infections can be very harmful to humans [11,12].Fire ants, Solenopsis xyloni, were first noted to damage electrical equipment in 1940. Telephone failures in Texas were attribute to the ants’ removal of wire insulation and their tendency to nest inside equipment, resulting in excessive internal current flow and short circuits [13]. Over time, problems in highway department signal cabinets and failures in air-conditioning units, telephone services, and traffic equipment were reported with increasing frequency [14,15,16]. RIFAs were later demonstrated to be attracted to electric fields [17].Besides visual inspection of RIFA mounds, bait traps are the primary method to detect fire ants. Oil baits were tested as attractive chemicals in the 1970s [18,19]. Ali and Reagan (1986) demonstrated that molasses and peanut oil attract RIFAs over short and long exposure periods, respectively [20]. In a monitoring study of RIFAs on a tropical fish farm, cotton cosmetic pads moistened served as a sucrose-based ant attractant [21], while toothpicks dipped in peanut butter were utilized as an oil-based ant attractant [22]. Food lures including heated meat products and corn/potato chips were also used to attract RIFAs and estimate their habitat range and abundance [23,24,25]. While previous access can confirm RIFA appearance, it is less efficient in locating ant nests, especially emerging and small mounds. Moreover, ant detection efficacy using the bait trap method decreases to 30% when RIFA net density is ≤15 nests per ha [22].Scent-detection dogs are used to detect non-biological scents (e.g., explosives and land mines) and biological scents (e.g., human, animal and plant scents) [26]. They have also been trained to detect destructive insects including the gypsy moth (Porthetria dispar L.), red palm weevil (Rhynchophorus ferrugineus), screwworm (Cochliomyia hominivorax Coquerel), bed bug (Cimex lectularius L.), sylvatic triatoma infestans (Hemiptera: Reduviidae), and multiple termite species [27,28,29,30,31,32].Trained dogs can identify the odor of RIFAs and freely search for nests in the field [33,34]. Our group demonstrated that trained detection dogs still achieved a 93% correct RIFA indication rate even when other ant species were introduced [35]. Dogs show great competence in detecting emerging and smaller RIFA nests in low-density locations [35]. Because of the invasive risks and because live RIFAs may not be available for training, we devised a practical method to train detection dogs without introducing live RIFAs and an alternative way to correctly identify RIFA-invaded zones.We describe a method using filter paper to adsorb RIFA odors for detection dog training. Filter paper was first tested for its feasibility as an odor bearer of RIFA. The odor concentration and time required for scent adsorption of filter paper were tested, as well as the appropriate storage temperature and time. Dogs trained with odor-adsorbed filter papers were then tested for live RIFA detection and in a field trial.2. Materials and Methods2.1. CaninesFive beagle dogs (one beagle mix) aged 3–7 years old, three females (two spayed) and two neutered males, were tested in this study. They are referred to as Dog A, B, C, D, and E (Table 1). These dogs were family companions and screened for certain characteristics such as self-confidence, physical soundness, food drive, sociability, intelligence, and ability to be trained. After passing the screen for detection dogs, they were trained for odor recognition ability. Dog handlers walked the dogs on leashes and indicated the objects with hiking sticks. The dogs were taught to respond to the RIFA scent by sitting in front of the detection target (passive response). Food and verbal praise were given as positive reinforcements to correct responses to encourage performance. Dogs A and B previously received RIFA training and had acquired the capability to correctly differentiate RIFA, Solenopsis invicta, from the other local species of ants, Crematogaster rogenhoferi, Paratrechina longicornis, and Pheidole megacephala [35]. Dogs A and B had served as live RIFA detection dogs for more than 3 years.2.2. Live RIFAsRIFAs were collected from mounds in Taiwan as previously described and then cultivated in plastic buckets containing soil [34]. Double-distilled water absorbed in cotton balls was provided, and egg puddings (ingredients mainly eggs, sugar, and milk) and an occasional fried chicken thigh with skin (~13% fat) were supplied as food. Each 50 mL centrifuge tube capped with a 1.7 × 1.7 cm2 square copper sieved opening was prepared in a separate sealed bag to avoid non-target odor contamination (Figure 1A). In subsequent experiments, pieces of tissue paper were placed in the buckets to attract ants and then put into reclosable plastic bags that were filled with CO2 for anesthesia. The paralyzed ants were then removed into tubes for further testing.2.3. Odor-Adsorbed Filter PapersFilter paper (No. 2, 90 mm, ADVANTEC Toyo, Ltd., Tokyo, Japan) was used as a scent reservoir. Live RIFAs were placed in a 50 mL centrifuge tube with a 1.7 × 1.7 cm2 square copper sieved opening. Then one filter paper and the centrifuge tube with live RIFAs inside were put together into a plastic zip bag (28 × 42 cm2) and sealed (Figure 1B). These procedures avoided live RIFAs escaping from the centrifuge tube, but allowed the filter paper to be exposed to the ant scent. The number of ants and adsorption time are described in each section of the different trials. After adsorption, the paper was moved into a new, fully capped and sealed centrifuge tube with forceps and used as a target sample. Unscented filter paper placed in a fully capped and sealed centrifuge tube was used as a non-target sample. All target and non-target samples were placed in a 25 °C incubator for 1 h before testing. Right before the test, the caps of the tubes with treated or untreated papers were opened one at a time and randomly placed into metal cans for dog detection (Figure 1C).2.4. Scent Lineup MethodThe scent lineup method was applied in a work room (Figure 1D). Seven to 10 metal cans were lined up at 90 cm intervals, with 210 cm between lines to avoid interference between samples and to supply sufficient space for the handlers and dogs. In a line, only 0–2 metal cans were randomly and discontinuously set as target samples, and the other samples were non-target samples. A blank run (centrifuge tubes with unscented filter paper placed in all metal cans) was performed before each trial start. A double-blind procedure was used; neither the dogs nor their handlers were aware of the positions of the target samples. The handlers were informed if the dog responded correctly by another person who was holding a picture card with Yes/No symbols at the corner of the work room. The dogs were reward by the handler immediately when they responded correctly to the target samples.2.5. Statistical AnalysesThe data were analyzed by utilizing SAS Enterprise Guide 6.1 software (SAS Institute, Cary, NC, USA). P values less than 0.05 were considered significantly different.2.6. Trial 1: Feasibility of Using Filter Paper to Carry and Represent RIFA OdorCentrifuge tubes containing 10 live RIFAs inside and tubes with a filter paper scented by 100 live RIFAs for 4 h were prepared as targets; a tube containing untreated filter paper inside was used as a non-target. For each test, we prepared 60 samples including five with 10 live RIFAs, five with 100 live RIFA-scented paper, and 50 samples with untreated paper. Five replicates of the tests were performed by Dogs A and B, which were live RIFA-experienced detection dogs that had not previously been exposed to scented filter paper. Positive indication was defined as a correct response to tubes with live RIFAs/treated filter papers, and a false response was when the dogs responded to untreated filter papers. The average positive indication rate (PIR) was then calculated to determine if filter paper can be used as an odor reservoir and recognized by detection dogs. The PIRs for live RIFAs and scented filter paper were compared with Mann–Whitney U tests.2.7. Trial 2: Odor-Adsorbed Filter Paper—Odor Intensity and Adsorption TimeTwo live RIFA-experienced detection dogs (A and B) and three inexperienced detection dogs (C, D, and E) were enrolled in this trial. Filter papers scented by 100 live RIFAs for 4 h were used to train these newly recruited dogs for RIFA odor recognition. The training followed the procedures in our previous publication, “Indoor Training for Red Imported Fire Ant Odor Recognition and Identification” [35]. After 20 consecutive correct responses, dogs were considered to have passed the precursor exercise and were qualified to participate in the following tests.2.7.1. Dog Detection and Odor IntensityThe target samples were tubes each containing a filter paper scented with 10, 50, or 100 RIFAs for 4 h. A tube with untreated filter paper inside was used as a non-target. For each test, a total of 50 samples were prepared, including 5 samples with scented papers and 45 samples with untreated papers. Five replicates of the tests were performed by each detection dog. The average PIRs in discerning filter paper scented by 10, 50, or 100 RIFAs for 4 h were calculated and compared with Kruskal–Wallis tests.2.7.2. Dog Detection and Odor Adsorption TimeTarget samples were tubes containing filter papers scented by 10 RIFAs for 24 h, 6 h, 1.5 h, 22.5 min, 5.6 min, and 1.4 min (decreased by 4 times). A tube with untreated filter paper inside was used as a non-target. A total of 30 samples were prepared for each test: 6 samples with scented papers and 24 with untreated papers. Five replicates of the tests were performed by each dog. The mean PIRs in discerning filter paper scented by 10 RIFAs for 24 h, 6 h, 1.5 h, 22.5 min, 5.6 min, and 1.4 min were calculated. The association between the odor adsorption time of RIFA-scented filter papers and positive indication rates (PIRs) were tested by calculating Spearman rank correlation coefficients.2.8. Trial 3: Odor-Adsorbed Filter Paper—Storage Time and TemperatureDogs A, B, and C participated in this trial. The target samples were filter papers scented by 10 RIFAs for 24 h. These samples were prepared and stored at three different temperatures (−20 °C, 4 °C, and 25 °C) for eight storage durations (0 [samples tested on the collection day], 1, 3, 5, 7, 9, 11, and 13 weeks). In total, 24 different conditions were tested. A tube containing untreated filter paper was used as a non-target. For each test, a total of 30 samples were prepared (5 target and 25 non-target samples). Before each test, each sample was placed in an incubator at 25 °C for 1 h. The PIR of each dog and the mean PIR and standard deviation in each condition were calculated. The PIRs were compared using the nonparametric Kruskal–Wallis test. Spearman correlation coefficients were calculated to test the association between storage time and dog PIR, and the data for the three different storage temperatures were analyzed separately.2.9. Trial 4: Odor-Adsorbed Filter Paper Trained Dogs Detecting Live RIFAsDogs C, D, and E were inexperienced in live RIFA detection and trained using odor-adsorbed filter paper. After achieving 90% PIR in training, they were qualified to be tested on live RIFA detection. A total of 30 samples were prepared for each test, including one sample of paper scented with 100 live RIFAs, one scented with 10 live RIFAs, one scented with 50 live RIFAs, one scented with 100 live RIFAs, and 26 samples with untreated paper. Each dog completed five rounds with five tests each, and PIRs were calculated. Kruskal–Wallis tests were performed to compare the PIRs of newly trained dogs for the different samples.2.10. Trial 5: Field Trial—RIFA Detection in CulvertsThe navigation lights at Taoyuan International Airport, Taiwan are above underground wire culverts covered with five steel plates (Figure 2). Nine rows of culverts were selected, and the methods of corn/potato chips attraction and visual mound inspection were used to identify RIFA activity. In each Culvert, 3 uncapped tubes containing potato chips were placed for 2 h. In one row, approximately 200 live RIFAs were identified inside this culvert, but the other eight rows were clear. Then, filter papers were hung by using cotton thread from hand holes under each steel plate to adsorb odor in the nine culvert rows. There were 5 filter papers placed for RIFA scent adsorption in each culvert. After 24 h, the odor-adsorbed filter papers were collected, sealed in separate centrifuge tubes, and sent to the lab at 4 °C in less than 24 h. After incubation at 25 °C for 1 h, the samples were tested by four trained dogs (A, B, C, and E). This trial was practiced twice, and the PIR and false response rates were evaluated.3. Results3.1. Trial 1: Feasibility of Using Filter Paper to Carry RIFA OdorThe mean PIRs of two live RIFA-experienced detection dogs (A and B) for detecting 10 live RIFAs in tubes were both 96.0%. After detecting a tube with a filter paper scented by 100 live RIFAs for 4 h, Dogs A and B had PIRs of 88% and 96%, respectively. Their false response rates were 1.2% and 0.4% (Figure 3). The discrimination results were similar for samples of live RIFAs and RIFA-scented papers (p = 0.37). The results support the hypothesis that filter paper can be used as an odor reservoir that is recognized by detection dogs.3.2. Trial 2: Odor-Adsorbed Filter Paper—Odor Intensity and Adsorption Time3.2.1. Dog Detection and Odor IntensityThe five dogs successfully detected filter papers scented by 10, 50, and 100 RIFAs for 4 h. The mean PIRs were 95.2%, 94.4%, and 93.6%, respectively, and the false response rate was 2% (Table 2). There was no significant difference among PIRs for papers scented by 10, 50, and 100 RIFAs (p = 0.84).3.2.2. Dog Detection and Adsorption TimeThe mean PIR for all five dogs detecting tubes with a filter paper scented by 10 RIFAs for 24 h was 97.6% and tended to decrease with shorter scented time. The PIRs for detecting tubes with a filter paper scented by 10 RIFAs for 6 h, 1.5 h, 22.5 min, 5.6 min, and 1.4 min were between 64.5% and 75.3%, and the false response rate was 3.4% (Figure 4). The non-significant Spearman correlation coefficient value of 0.29 suggests a weak positive linear relationship between the PIR and filter paper scented time (r = 0.29, n = 30, p = 0.12).3.3. Trial 3: Odor-Adsorbed Filter Paper—Storage Time and TemperatureThree dogs sniffed scented filter papers stored at 24 different temperatures and durations. The samples stored at −20 °C and 4 °C were detectable throughout the 13-week storage period with a >90% PIR. Conversely, the 90% positive detection rate for samples stored at 25 °C was only achieved in the first week; the detection rate decreased with longer storage (Figure 4). The mean positive detected rates of the samples stored at three different temperatures were not significantly different between samples tested on the date of collection and those stored for 1 week (p = 0.92). After the samples were stored for 3, 5, 7, 9, 11, for 13 weeks, the mean positive detection rate of the 25 °C samples was lower than the rates of those stored at −20 °C and 4 °C (p < 0.05). Moreover, the detection rate of the 25 °C samples significantly decreased with longer storage (p < 0.0001), but no differences were found between those stored at −20 °C (p = 0.21) and 4 °C (p = 0.40) (Figure 5).3.4. Trial 4: Odor-Adsorbed Filter Paper Trained Dogs Detecting Live RIFADogs C, D, and E were trained in the present study and all achieved 100% PIR. They were then tested with 10, 50, and 100 live RIFAs. Their respective mean PIRs were 100%, 99%, 100%, and 96% in detecting paper scented with 100 live RIFAs and 10, 50, and 100 live RIFAs, with a mean false response rate of 3.0% (Figure 6). There were no differences among the positive detection rates in discriminating paper scented with 100 live RIFAs and 10, 50, and 100 live RIFAs (p = 0.28).3.5. Trial 5: Field Trial—RIFA Detection in CulvertsScented samples from RIFA-invaded underground wire culverts at Taoyuan International Airport were collected and sent for double-blind odor recognition. The mean PIR of dogs A, B, C, and E asked to identify RIFA-scented paper collected from the target culvert row was 87.5%, and the average false response was 7.8%. There were no significant differences among the dogs’ capabilities in discriminating the target (p = 0.68) and non-target samples (p = 0.76).4. DiscussionGiven their great capability for scent recognition, detection dogs play significant roles in locating both harmful insects and threatened and rare species. They have been utilized to detect invasive plants and animals such as the spotted knapweed (Centaurea stoebe), small Indian mongoose (Herpestes javanicus), and brown tree snake (Boiga irregularis) [26,36,37,38].In the present study, two live RIFA-experienced dogs did not show different capabilities for detecting 10 live RIFAs in tubes, tubes with a filter paper scented by 100 live RIFAs for 4 h, or tubes with untreated paper. In addition, there were no significant differences in scent-detection capability (both PIR and false response rate) among the five dogs based on the data in trials 2, 3, 4, or 5. These results indicate a high ability of dogs trained for RIFA scent detection; all animals met the minimum requirement with a PIR of at least 90% and a false negative positive rate ≤10% [27].Several tools have been used as odor adsorbents and transmitters, and as training aids for canine detection training [39]. Recently, more canine training aids were studied, such as odor capture-and-release materials for explosive odorants and ovarian cancer cell lines for ovarian cancer detection [40,41]. The gypsy moth (Porthetria dispar L.) sex attractant pheromone disparlure has been applied to different objects for training, and dogs’ capacity to retrieve disparlure-treated sticks was higher than for treated rubber balls or tin cans [31]. Exudates from screwworm-infested wounds placed on balls and larvae-scented balls were used as target samples for dog training [32]. For termite detection, dogs were taught to associate terry cloth towels with termite scent [27]. A filter paper impregnated with bed bug scent using pentane extraction was 100% indicated by trained dogs and therefore useful for detector training [29]. Despite these reports, there was no scientific evidence validating the utility of filter paper odor adsorption. The present study is the first to show that it is a competent odor transporting medium.Carpet squares have been used as a scent carrier of residual human odor in forensic studies. Three well-trained cadaver dogs displayed excellent validity, predictive value, and accuracy for identifying positive samples [42]. More recently, grave soil was reported to hold residual human odor and can be recognized by properly trained dogs [43]. Commercially available pseudoscents from Sigma-Aldrich (St. Louis, MO, USA) have also been applied as canine training aids to facilitate the detection of corpses and narcotics such as cocaine, heroin, and marijuana. In the field of human forensic science, cotton napkin (ARATEX™, Chlumtex) have been used in routine casework [44]. However, studies on scent carriers for RIFAs are still rare, and the stability of scent samples used in animal chemosignaling was possibly violated during analytical procedures by storage protocols, storage containers, light exposure, storage time and different temperature treatments [45,46].RIFAs produce several semiochemicals such as venom alkaloids, cuticular hydrocarbons, and trail pheromones [47]. In the present study, dogs were trained to recognize the odor of RIFAs on filter paper. The carried scent may contain various organic compounds, and further research is necessary to determine exactly what the trained dogs smelled.Detection dogs have been utilized to detect the invasive brown tree snake (Boiga irregularis). This reptile has substantially reduced or even eradicated native animal populations, but on Guam it threatened electrical utilities and human safety [36]. Given the risk of snakes on Guam, the government of Hawaii imported one live, sterile male brown tree snake to train its detection dogs. To secure this invasive damaging species, a radio transmitter was implanted into the snake, and a double-secured container and snake-catching equipment were used for restraint. Live fire ants should not be imported into an RIFA-free zone, and are a biosecurity risk. For example, RIFA colonization in Japan has not been confirmed; however, RIFAs have been reported in 37 locations in 14 prefectures in Japan and mainly entered by the shipping containers imported from southern China [48]. Our study suggests a practical method to train detection dogs without introducing live RIFAs and a quick and safe way to correctly identify RIFA-invaded zones. The results demonstrate that filter paper is a useful aid for detection dog training in uninvaded areas. Moreover, filter paper can be used to collect and transmit other odors for dog training. Future studies on the capacity of scented filter papers transported by plane will be necessary since air pressure changes with altitude.According to the legislation of the Taiwanese Ministry of Transportation and Communication, Civil Aeronautics Administration, dogs are not allowed to enter restricted areas in airports for RIFA searching. Our results suggest that detection dogs trained with filter paper have a good positive detection rate and low false response rate. Therefore, this method can be applied to train dogs to detect RIFAs in quarantined or segregated areas. Detection dogs might not be able to enter mechanical facilities and go near electrical equipment; in these scenarios, filter paper could be used to adsorb the scent and then sent for dog interpretation. This approach would minimize risk to detecting dogs and save manpower, time, and money.Using canine detectors for RIFAs is more effective than other methods of detection when the ants are at low density [35]. The results of using canine detectors and RIFA-scented filter paper in the present investigation reveal a practical method on RIFA detection for regions or countries where fire ants have not invaded. The use of RIFA-scented papers as an odor reservoir for dog training could increase detection efficiency, achieve economic savings, and decrease the risk of RIFA introduction into uninvaded or eradicated areas.

animals : an open access journal from mdpi

10.3390/ani11071995

PMC8300099

Notwithstanding extensive research into fertility problems in mares, pregnancy rates have remained low mainly because of endometrial inflammation (endometritis). In the field of equine research, endometrial explants have been used to carry out in vitro studies of the mare’s endometrium. However, there has been no wide-ranging assessment of relative stability of this model over time. The aim of this study was to perform an in-depth transcriptomic assessment of endometrial explants over a culture period of 72 h and assess if they are representative of the whole mare. Explants at 24 h demonstrated significant changes when compared to biopsies at 0 h as expected. Even though gene expression changes were seen between 24 and 48 h of culture, prior to this window changes were dominated by the effects of explanting and culture and subsequently, transcription was generally compromised. Our results, therefore have defined the optimal period when explants can be used to study equine endometritis and how the endometrium is modulated during inflammation. It highlights the use of abattoir derived samples to understand the physiology and pathophysiology of the equine endometrium, negating the need to collect repeated uterine biopsies from living mares.

Persistent mating-induced endometritis is a major cause of poor fertility rates in the mare. Endometritis can be investigated using an ex vivo equine endometrial explant system which measures uterine inflammation using prostaglandin F2α as a biomarker. However, this model has yet to undergo a wide-ranging assessment through transcriptomics. In this study, we assessed the transcriptomes of cultured endometrial explants and the optimal temporal window for their use. Endometrium harvested immediately post-mortem from native pony mares (n = 8) were sampled (0 h) and tissue explants were cultured for 24, 48 and 72 h. Tissues were stored in RNALater, total RNA was extracted and sequenced. Differentially expressed genes (DEGs) were defined using DESeq2 (R/Bioconductor). Principal component analysis indicated that the greatest changes in expression occurred in the first 24 h of culture when compared to autologous biopsies at 0 h. Fewer DEGs were seen between 24 and 48 h of culture suggesting the system was more stable than during the first 24 h. No genes were differentially expressed between 48 and 72 h but the low number of background gene expression suggested that explant viability was compromised after 48 h. ESR1, MMP9, PTGS2, PMAIP1, TNF, GADD45B and SELE genes were used as biomarkers of endometrial function, cell death and inflammation across tissue culture timepoints. STRING assessments of gene ontology suggested that DEGs between 24 and 48 h were linked to inflammation, immune system, cellular processes, environmental information processing and signal transduction, with an upregulation of most biomarker genes at 24 h. Taken together our observations indicated that 24–48 h is the optimal temporal window when the explant model can be used, as explants restore microcirculation, perform wound healing and tackle inflammation during this period. This key observation will facilitate the appropriate use of this as a model for further research into the equine endometrium and potentially the progression of mating-induced endometritis to persistent inflammation between 24 and 48 h.

1. IntroductionPersistent uterine inflammation (endometritis) has been extensively investigated in horses as it is the main cause of subfertility in mares and the third most important clinical problem affecting adult horses according to equine practitioners in the USA [1,2]. However, for many years pregnancy rates in broodmares have remained low, between 50% and 65% per oestrous cycle, despite extensive research into fertility problems [3,4]. Nearly 23,000 Thoroughbred (TB) mares were sent to the stud in Great Britain and Northern Ireland in 2019, yet only 61% produced a live foal [5]. It has been reported that 15–40% of a normal population of TB mares suffer from persistent endometritis, reducing pregnancy rates by 13% [6,7].In the field of equine research, endometrial explants have been used to carry out many in vitro studies of the mare’s endometrium [8,9,10]. Tissue culture of explants has distinct advantages over isolated cell culture models. The technique saves time as cell culture requires several days or weeks to reach confluence [11]. Utilizing intact explants is particularly important. Most tissue explant models mechanically chop the tissue aiming for better oxygenation and perfusion of nutrients, but the chopping process itself leads to damage and disruption of tissue architecture [11]. By using intact explants, the extracellular matrix (ECM) remains intact which helps to maintain normal cell differentiation and functioning. Chopping will release damage-associated molecular patterns (DAMPS) which are intracellular and extracellular molecules typically released by the ECM after cell death and/or injury. DAMPS modulate the innate immune system, triggering inflammation even under sterile conditions [12,13], distorting the observations obtained because chopping heavily wounds the tissue. As a result, some ex vivo bovine studies have adopted an intact biopsy endometrium model to better mimic the whole cow [11,14]. Therefore, the present study used the equine endometrial explant culture proposed by Nash et al. [15], but substituted the mechanical tissue chopping by harvesting intact explants using sterile punch biopsies, as advocated by Borges et al. [11].Harvesting endometrium from uteri collected at an abattoir negates the need to perform repeated uterine biopsies from living mares. Further, the large surface area of endometrium provided by each mare can be entirely harvested so that several different inflammatory stimuli and/or treatments of various concentrations can be concurrently studied, maximizing technical replication. An equine endometrial explant culture system composed of epithelial cells, stromal cells and resident leukocytes has been previously optimised. As with the whole animal, cultured explants respond to inflammatory stimuli by secreting markers of inflammation such as prostaglandin F2α (PGF2α) [10,15,16]. Schwinghamer and colleagues [17] studied equine endometrial explants in culture for periods of 12, 24 and 48 h to investigate whether explants undergo significant changes during culture. They suggested that as early as 12 h of culture explants undergo cellular damage based on the increase of lactate dehydrogenase (LDH) activity, and at 48 h of culture light microscopy showed that some degenerative changes take place in the luminal epithelium and in the epithelium lining deeper endometrial glands. Nonetheless, there has been no assessment of global gene expression changes to intact equine endometrial explants during the culture period.The aim was to undertake a transcriptomic assessment of ex vivo endometrial explants collected from native pony mares cultured over a period of 72 h. It assessed whether explants were representative of the whole mare in the pre-breeding, non-inflammatory state and established whether the transcriptome stabilizes once in culture. A temporal window relevant to persistent mating-induced endometritis was established in the current study, evidencing potential for using the endometrial explant tissue culture model for wider endometrium studies in horses. Therefore, the present study facilitates and informs the use of this model for a more in-depth study of equine endometritis to understand how the endometrium is modulated during inflammation.2. Materials and Methods2.1. Animals and Sample CollectionUteri were collected from a random selection of native pony mares presented for euthanasia at a commercial abattoir for reasons unrelated to this study. Mares were slaughtered by free bullet followed by exsanguination of the jugular vein. Management of the animals, age or reproductive history were unknown. Corresponding blood samples were collected immediately after death from the jugular vein and stored in a plain vacutainer (367895, BD Vacutainer, Plymouth, UK). To estimate the stage of the oestrous cycle for each mare at the time of the death, uteri and cervices were physically and visually assessed as previously described by various authors [15,18,19]. Briefly, uteri and cervix were assessed for colour, appearance and tone whilst ovaries were immersed in a bucket with water and analysed using an ultrasound scanner for assessment of ovarian structures and follicle measurements (Table S1). The stage of oestrous cycle was retrospectively confirmed by measuring serum progesterone (P4) concentrations from the blood sample collected from each mare. Mares at the follicular phase of the oestrous cycle (n = 8) were utilized in this study to represent the stage of cycle that clinical endometritis is likely to occur. At the abattoir, endometrial biopsies were collected from each uteri within two hours of death, with the aid of a sterilized biopsy instrument (Equivet uterine biopsy forceps 62 cm, 141965, Kruuse, Guildford, UK), placed into RNALater (10437114, Fisher, Leicestershire, UK) and kept at 4 °C for 24 h during which samples were transported to the laboratory. After 24 h, RNALater was removed and tissues were stored at −80 °C until RNA extraction. These biopsy tissues were considered to be “0 h” endometrial samples in RNA-sequencing (RNA-seq) analysis. Uteri and blood samples were stored on ice and transported back to the laboratory within 6 h of collection. At the laboratory, serum P4 concentration was measured by an enzyme-linked immunosorbent assay (ELISA) kit (EIA 1561, DRG Diagnostics, Marburg, Germany) following the manufacturer’s instructions to confirm the phase of the oestrous cycle for each animal at the time of death.2.2. Endometrial Cytology, Histology and Tissue CultureTo verify that uteri used were not inflamed, samples intended for cytological analysis were collected at the abattoir from each tract, using a sterile cytobrush (Cytobrush plus GT, C0112, Coopersurgical, Trumbull, CT, USA). Cytobrushes were rolled onto plain microscope slides and these were immediately airdried and subsequently fixed and stained with eosin and methylene blue (Shandon Kwik-Diff Kit, 9990700, ThermoFisher Scientific, Loughborough, UK). Specimens were evaluated by recording the number of neutrophils per high power field (hpf; 400×) across 10 fields [20,21]. Findings were classified as: normal (0 to 2 neutrophils/hpf), moderate inflammation (2 to 5 neutrophils/hpf) or severe inflammation (>5 neutrophils/hpf) [22,23,24]. Uteri showing moderate to severe inflammation were discarded from the study.After transportation back to the laboratory, a biopsy was sampled from the uterine body and immediately placed in Bouin’s Fixative Fluid (10821910, Fisher, Leicestershire, UK). After fixation, tissues were subjected to several changes of 70% ethanol until their yellow colour disappeared. Tissues were then sectioned and placed into histological cassettes (U4635-1CS, Sigma Aldrich, Exeter, UK) and subjected to dehydration, clearing and infiltration. Sections of 8 μm were cut using a microtome (Minot 1212 rotary microtome, Ernst Leitz GmbH, Wetzlar, Germany) and transferred onto a microscope slide (10149870, Fisher, Leicestershire, UK) and placed on a hot plate at 40 °C to dry. Sections were de-waxed, followed by rehydration in descending concentrations of ethanol and stained with Harris haematoxylin and alcoholic 1% eosin for 5 min. Slides were mounted with Histomount solution (008030, Invitrogen, Paisley, UK) and a coverslip. Slides were left to dry and analysed under a light microscope. Histology slides were assessed by a Boarded Certified Veterinary Pathologist following the Kenney classification system [25] to indicate pathological or degenerative endometrial changes. As a result of the histological assessment, mares showing pathological or degenerative endometrial changes correspondent to category III were discarded from tissue culture. The Kenney classifications for the mares used in this study can be found in Table S2.Working under aseptic conditions, punch biopsies were collected from the uterine horns (8 mm Kai Biopsy punch, BP-80F, Northumbrian Medical Supplies, Newcastle Upon Tyne, UK), immediately placed in warm supplemented Hank’s Buffered Saline Solution (HBSS) (14175053, Fisher, Leicestershire, UK) and left in the incubator (5% CO2, 38 °C) while biopsies from other uteri were harvested. Once biopsies were collected from all uteri, the tissue culture was assembled in a laminar flow hood (Microflow Biological, Bioquell, Andover, UK). Biopsy tissues were washed twice in un-supplemented HBSS and cultured by placing individual biopsies into one well of a six-well culture plate (10578911, Fisher, Leicestershire, UK). By following previously described protocols [11,15], biopsies were placed at the bottom of the well or on top of a lens-tissue lined wire platform and cultured in 3 mL and 4.25 mL of supplemented William’s Liquid E Medium (William’s phenol red free, 500 mL, A12176-01, Fisher, Leicestershire, UK), respectively. Explant cultures were performed in triplicate for each timepoint per animal. For all experiments, tissue explants were incubated at 38 °C in 5% CO2 in air. After periods of 24, 48 and 72 h explants were removed from the well and stored in 1.5 mL of RNALater for a 24 h period at 4 °C. After 24 h in RNALater explants were removed and stored at −80 °C until RNA extraction. William’s Liquid E Medium (500 mL) was supplemented with 0.01 µg/mL epidermal growth factor (EGF) (10605-HNAE-250, Fisher, Leicestershire, UK), 0.1 mg/mL neomycin (N6386, Sigma, Exeter, UK) and streptomycin (S6501, Sigma, Exeter, UK) solution, 2 mM of L-Glutamine (25030-032, ThermoFisher Scientific, Loughborough, UK), 2.5 µg/mL amphotericin B (A2942-20ML, Sigma, Exeter, UK), 5 mL of insulin-transferrin-selenium (ITS) (10524233, Fisher, Leicestershire, UK) and 50 mL of batch tested foetal bovine serum (FBS) (Invitrogen FBS, 10695023, Fisher, Leicestershire, UK). The supplemented medium was stored at 4 °C for a maximum period of 7 days. HBSS (100 mL), was supplemented with 0.1 µg/mL EGF, 1 mL of 10 mg/mL neomycin and streptomycin solution, 2 mM of L-Glutamine, 2.5 µg/mL Amphotericin B, 1 mL of ITS and 10 mL of FBS. Supplemented HBSS was stored at 4 °C for a maximum of 2 days.2.3. RNA Extraction, Quantification and SequencingTotal RNA was extracted from endometrial samples using an RNA purification kit (GeneJET RNA Purification Kit, Thermo Scientific, Bishops Stortford, UK) following the manufacturer’s instructions. The NanoDrop 1000 Spectrophotometer (Thermo Scientific, UK) was used to check RNA concentration and purity. In addition, RNA samples were fractionated on a 1% agarose gel to assess RNA integrity which was indicated by visualization of distinct 28S and 18S ribosomal RNA bands.Complementary DNA (cDNA) synthesis, fragmentation and library preparation were performed using the Illumina TruSeq Stranded mRNA kit (20020594, Illumina, San Diego, CA, USA) according to the manufacturer’s instructions to prepare the dual-indexed next-generation sequencing libraries. Briefly, from each sample poly-A messenger RNA (mRNA) was purified, fragmented, and reversed transcribed into cDNA. The cDNA was end-repaired and connected to adaptors that contained unique indexes for each sample. The connected products were amplified by a polymerase chain reaction (PCR) (98 °C for 30 s, 15 cycles of 98 °C for 10 s, 60 °C for 30 s, 72 °C for 30 s, 72 °C for 5 min and then cooling to 4 °C), followed by purification with AMPure (A63882, Beckman Coulter, High Wycombe, UK) to remove PCR reagents and adaptor dimers. Product quantity in ng/µL was assessed using a Qubit fluorescence spectrophotometer (Thermo Fisher Ltd.). The final library dilution was 8 pM. The diluted library was loaded, bound, and amplified onto a flowcell using an Illumina cBOT platform and then transferred to the HiSeq2500 and paired-end sequencing runs in the 2 × 216 bp format.2.4. Data Processing and Gene Expression AnalysisThe pair-end sequencing produced forward and reverse reads for each sample, totalling 64 raw reads. Raw reads were subjected to quality analysis using FastQC software (version 0.11.2) [26] and filtering using Trimmomatic software (version 0.33) [27] for sequencing adapters removal (i.e., ILLUMINACLIP), bases were removed from the beginning and end of the reads if the quality was below 30 (i.e., LEADING or TRAILING = 30). A sliding window trimming approach was performed once the average quality within the window fell below 30 (i.e., SLIDINGWINDOW 4:30), reads were dropped if they were below a specific length of 100 bp (i.e., MINLEN = 100), and a specified number of 10 bases from the start of the read were removed (i.e., HEADCROP = 10). The quality of the trimmed reads was assessed again by FastQC. Reads were mapped to the Equus caballus reference genome (Ensembl, EquCab3.0; GCA_002863925.1) using Bowtie software (version 2.2.3) [28] and TopHat software (version 2.0.14) [29]. The number of reads/fragments assigned to genomic features for each sample and the Equus caballus annotation file available from Ensembl website (version EquCab3.0) using the FeatureCounts software (version 1.5.2) [30]. Sequencing data are provided in Table S3. The concordant pair alignment rate to the annotated equine genome from all samples was between 86.2% and 93.6%.Differentially expressed genes (DEGs) were identified using the DESeq2 R/Bioconductor version 1.28.1 [31,32]. Benjamini-Hochberg (BH) false discovery rate (FDR) method [33] was performed in R (version 4.0.3) to correct for multiple testing. Genes were considered differentially expressed at an FDR of 0.05 with a Log2 fold change (Log2FC) greater or equal to ± 2. DEGs were submitted to the Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) database [34] (http://string-db.org, accessed on 5 February 2021) for network analysis and gene associations. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways [35] (http://www.genome.jp/kegg, accessed on 9 February 2021) were also used to identify significantly enriched pathways. Principal component analyses (PCA) and heatmaps were generated using R (version 4.0.3).3. ResultsThe global gene expression of equine endometrial explants (n = 8 horses) cultured in vitro for periods of 24, 48 and 72 h were compared to autologous ex vivo 0 h biopsies, representing the whole mare. Patters of gene expression were indicated by PCA based on all gene expression, not solely DEGs (Figure 1). Samples obtained at 0 h were separated from all others across the major source of variation along PC1. There was lesser variation along PC2 with the samples with the 24 h samples being discrete from the 48 and 72 h group which could not be distinguished.To identify the sources of variation, DEG analyses were undertaken, comparing between time points. When comparing the transcriptomes of explants cultured for 24 h to those of whole mares’ controls (0 h), 707 genes were significantly differentially upregulated and 859 downregulated. In contrast, when comparing explant tissues cultured for 48 h relative to 24 h, only 53 genes were significantly upregulated while 27 genes were downregulated. Finally, the transcriptome of explants cultured for 72 h revealed no differentially expressed genes compared to 48 h samples. Table 1 compares the number of expressed genes (background genes) and the number of DEGs between timepoints, whilst Figure 2 and Figure 3 graphically represent data featured in Table 1. It should be noted that unlike the 0–48 h period relatively few genes were expressed within the 48–72 h window.Following the STRING functional enrichment analysis of DEGs between 0 and 24 h, a total of 41 KEGG pathways were found to be significantly enriched (Table S4). Table 2 features KEGG pathways that are relevant to the current study. These indicated prominent changes linked to inflammation, immune system, cellular processes, environmental information processing and signal transduction. STRING enrichment analysis of DEG between 24 and 48 h retrieved a total of five significantly enriched KEGG pathways (Table 3). These suggested cell cycle events, cellular processes, and ECM remodelling.Genes previously used as biomarkers of endometrial function [17] were also studied to better understand explants’ functionality once in culture. The oestrogen receptor, ESR1, was upregulated at 24 h vs. 0 h (FDR = 1.48 × 10−15; Log2FC = −2.5), however it was not differentially expressed at 48 vs. 24 h (FDR = 0.35; Log2FC = 0.5). Whilst PTGS2, prostaglandin-endoperoxide synthase 2, was upregulated at 24 h vs. 0 h (FDR = 1.97 × 10−27; Log2FC = 6), it was not differentially expressed at 48 h vs. 24 h of culture (FDR = 0.99; Log2FC = −0.01). The possible initiation of cell death was considered by examining the expression of GADD45B (growth arrest and DNA-damage-inducible beta I) and PMAIP1 (phorbol-12-myristate-13-acetate-induced protein 1) [17]. GADD45B was not differentially expressed at 24 h (FDR = 3.2 × 10−4; Log2FC = 1.6) nor at 48 h (FDR = 0.35; Log2FC = −0.6) using the usual comparisons. PMAIP1 was upregulated at 24 h (FDR = 1.64 × 10−13; Log2FC = 3.1) but not differentially regulated at 48 h of culture (FDR = 0.79; Log2FC = 0.2). Figure 4 displays a heatmap showing the expression pattern (count matrix) of the biomarker genes: ESR1, MMP9, PTGS2, PMAIP1, TNF, GADD45B and SELE. This indicted a transient increase in expression by most biomarkers at 24 h whilst ESR1 was suppressed. At 48 and 72 h consistent elevated expression of MMP9, a matrix metallopeptidase, was seen in each horse explant. Expression of PTGS2 (prostaglandin processing gene) and PMAIP1 was induced at 24 h and maintained in each horse explant up to 72 h. Correlation analyses indicated that ESR1 was inversely correlated with all biomarker genes and most clearly with PTGS2 (Figure S1). In turn, PTGS2 was positively correlated with all biomarker genes but ESR1.4. DiscussionEndometritis is a major cause of infertility in mares. Whilst endometritis has been the subject of many studies that investigate pathophysiology, the development of a viable explant model that maintains multicellular integrity would represent an important advance to facilitate further study of the equine endometrium and better understand endometritis. Preliminary approaches to develop such a model system have been proposed but these lacked the extensive validation that is possible using transcriptomic approaches. In this study, we describe distinctive phases of transcriptomic responses during explant culture.In the present study, the upregulation of ESR1 at 24 h of culture implied that explants cultured for the first 24 h cannot be used to study oestrogen-responsive genes [17]. However, ESR1 is not differentially regulated at 48 h of culture, implying that explants can indeed be used in the study of oestrogen responsiveness in the window of 24–48 h of culture. Along with ESR1, PTGS2 was also used as a biomarker of endometrial function as it encodes an enzyme that plays a role in the conversion of arachidonic acid to prostaglandin endoperoxide, which is an endometrial biomarker of inflammation [17,36]. PTGS2 upregulation at 24 h of culture indicates arachidonic acid processing, possibly linked to inflammatory events during the first hours of culture due to explant removal from uterus. Nonetheless, PTGS2 was not differentially regulated at 48 h of culture, possibly indicating that at this time inflammation was stable. These findings might suggest that this tissue model offers a short window of 24–48 h when further experiments which use challenged explants to investigate endometritis-like events can be performed as suggested by Schwinghamer et al. [17].Growth arrest and DNA-damage-inducible beta I gene, GADD45B, that regulates growth and apoptosis [37] was not regulated at 24 h nor at 48 h of culture. PMAIP1 (Phorbol-12-Myristate-13-Acetate-Induced Protein 1) which encodes a pro-apoptotic protein [38] was upregulated at 24 h of culture but it was not differentially regulated at 48 h. As previously described by Schwinghamer et al. [17], both GADD45B and PMAIP1 can be used as biomarkers of cellular death in equine endometrial explant systems. This implied that cell death was triggered during the first 24 h of culture, but that apoptosis was not likely to happen at 48 h. These observations also support relative stability in the explants in the 24–48 h time frame. However, by 72 h only a very low number of genes were expressed (804 at 72 h vs. 15290 at 48 h; Table 1) which aligns with Schwinghamer et al. [17] observations that degenerative changes occur in cultured explants after 48 h of culture. Thus, our transcriptomic experiments suggest a relatively short window of 24–48 h to analyse endometrial tissue and to study endometritis using the explant system in the future.During the initial 0–24 h culture phase there were a series of transcriptional changes that were consistent with stress. The TNF signalling pathway, that plays a role in inflammatory cell proliferation, differentiation, survival and cell death [39], was upregulated at 24 h of culture (Figure S2). TNF induces the expression of E-selectin, SELE, that is involved in the localization and development of inflammation [40,41]. The upregulation of SELE is consistent with the upregulation of TNF-mediated inflammation during the first 24 h of culture seen in the TNF signalling pathway. Increases in pro-inflammatory interleukin-1β (IL–1β) and chemokine C-C motif ligand 20 (CCL20) also suggests that inflammation was a feature of the explant system. Therefore, such observations are consistent with inflammatory events being prominent in the explant culture during the first 24 h. The upregulation of MMP9 should also be seen as wider change in the ECM that features in our transcriptome during the first 24 h of culture. It is involved in the breakdown of ECM and other normal physiological and pathological processes [42,43,44]. MMP9 has been linked to a reduction in connective tissue deposition, in restoring microcirculation and it also plays a role in wound repair [42,45,46]. The upregulation of MMP9 seen at 24 h suggests that the explants have started healing wounds caused by mechanical removal from the uterus and that they are trying to restore microcirculation to support their survival in the tissue culture environment.The ECM is a network of collagen, fibronectin, enzymes, proteoglycans and glycoproteins [47,48,49] that is continually remodelled to control tissue homeostasis, cell growth, migration, differentiation and morphogenesis [50,51]. ECM molecules are then able to control the magnitude of inflammation based on their ability to bind to different pattern recognition receptors (PRRs). Equally, immune cells release enzymes and inflammatory mediators which can further degrade ECM or can alter its composition [52,53,54]. In our explant system, most collagen coding genes were downregulated (COL17A1, COL21A1, COL2A1, COL9A1, COL26A1, COL1A2, COL4A4 and COL4A6) and only two were upregulated (COL12A1 and COL20A1) at 24 h. Type IX collagen plays an important role in the formation of a collagen network and it maintains cartilage integrity and organization [55], which might suggest that at 24 h of culture explants are building a collagen network to recover from mechanical trauma. Downregulation of type IV collagen leads to an upregulation of MMP9 and thus related to cell migration and invasion [56]. COL12A1 has been associated with ligament ruptures in women [57], therefore its upregulation at 24 h of culture might be related with tissue rupture after excision from the uterus. On the other hand, type XX collagen is still not well understood and its function is relatively unknown [58].An interplay between pro- and anti-inflammatory events is also suggested by examination of interleukin (IL) expression patterns in the cytokine–cytokine receptor interaction pathway (Figure S3). Most ILs were found to be upregulated at 24 h of culture (IL1A, IL1B, IL1R2, IL1RN, IL21R, IL23A, IL24, IL2RA, IL31RA, IL36A, IL36G, IL4R and IL6) whilst IL25 (also known as IL17E) was the only IL shown to be downregulated. It has been shown that during gut inflammation the production of IL25 protein is diminished [59]. This can suggest that during the first 24 h of culture the explants are indeed undergoing inflammation, thus the downregulation of IL25. The IL1 family group induces a network of proinflammatory cytokines, which regulate and initiate inflammatory responses [60]. IL24 belongs to the IL10 family, and is released by immune cells such as monocytes, macrophages and T cells [61]. IL24 is suggested to control cell proliferation and survival and it participates in wound healing [62]. IL23A is a protein encoded by IL23 that plays an important role in inflammatory responses. It also increases angiogenesis and upregulates MMP9 [63] as seen in this explant tissue model. Most ILs found to be upregulated at 24 h of culture were from the IL1 family, demonstrating that inflammation was a feature of explants during the first hours of culture. CCR7, C-C motif chemokine receptor 7, was upregulated at 24 h of culture and the C-C chemokine ligand family (CCLs) were differentially expressed. CCL2, CCL11, CCL20, CCL26 and CCL27 were all upregulated whilst CCL24 was downregulated. During physiological and pathological conditions, specially during inflammation, CCLs play an important role in the recruitment of immune cells [64]. CCL24 is also involved in pathological processes such as inflammation and fibrosis, but it has been found that CCL24 blockade can be used therapeutically to inhibit cell activation and pulmonary inflammation [65]. Therefore at 24 h there could be a recruitment of immune cells due to the local inflammation produced after explant removal from uteri by CCLs. These events should be considered as part of a modulatory event influencing explant cell growth in culture.When comparing the transcriptome at 48 to the 24 h, IL23A and IL24 were downregulated whilst CSF3R was upregulated in the cytokine–cytokine receptor interaction pathway. CSF3R, colony-stimulating factor 3 receptor, is involved in the production, differentiation and function of granulocytes and it also plays a role in cellular processes such as surface adhesion and recognition [66,67]. A downregulation of interleukin-coding genes leads to the hypothesis that the inflammatory process due to tissue injury was mounted, tackled and it was returning to normal by 48 h of culture. At this stage the only TNF signalling pathway output that was modulated was MMP9. As outlined, MMP9 is a pro-angiogenic factor that is involved in ECM remodelling and that coordinates epithelial wound repair by removing fibrinogen matrix and stimulating collagen contraction [42,46]. Therefore, wound repair is still likely to be a feature of explants at 48 h. At 48 h of culture collagen (COL1A2 and COL24A1) and fibronectin (FN1) genes were also upregulated when compared to the transcriptome of explants cultured for 24 h. Fibronectin binds cells with collagen fibres and cell-surface integrins and is involved in cell adhesion, growth, migration, differentiation, and other processes such as wound healing [68]. As fibronectin expression is directly linked to epithelial cell growth [69], it would implicate explant cell growth at 48 h of culture.A limitation of this work was undoubtedly the lack of medical history for the animals from which uteri were collected as these were sourced from a commercial abattoir. It was important that uteri used did not have any inflammation at the time of collection that may adversely affect the data. As such the uteri collected were from native pony mares to represent a population that had not been intensively managed. These ponies were collected for commercial purposes unrelated to our study, from semi-feral populations that were rounded up from the mountains of South Wales, UK, in the days before being transported to the abattoir. Unlike domestic horses being sent to the abattoir, these were not being slaughtered due to old age, injury or illness, but were otherwise healthy individuals, less likely to have underling inflammatory processes present at the time of death. However, as these semi-feral populations receive minimal management or intervention, no medical history was documented or available to this project. In terms of uterine health specifically, we insured that uterine inflammation was not apparent at the time of tissue collection using cytological and histological assessment of each uterus used. Any animal that had evidence of inflammation using these techniques was retrospectively disregarded from the study. The cytological and histological analyses conferred confidence that, without a detailed medical history, the uteri of these mares were not undergoing inflammation at the time of tissue collection.5. ConclusionsThis study provides the first comprehensive in vitro assessment of the pre-breeding, non-inflammatory-challenged, global gene expression analysis of cultured equine endometrial explants and autologous ex vivo biopsies. It highlights the utility of abattoir-derived samples to understand the physiology of the equine uterus. The upregulation of interleukins, ESR1, PTGS2, PMAIP1, SELE and TNF is indicative that explants were undergoing stress and inflammatory events during the first 24 h of culture. Nonetheless, at 24 h explants also showed an upregulation of MMP9, which suggests that explants were undergoing microcirculation restoration and wound healing after removal from the uterus. The downregulation of interleukin-coding genes and an upregulation of MMP9 seen at 48 h of culture suggests that inflammatory processes were mounted and tackled and that the tissue was restoring homeostasis between 24 to 48 h of culture. GADD45B and PMAIP1 were not differentially expressed at 48 h of culture, indicating that cell death was not occurring at a significant level at this time. At 72 h of culture only a limited number of genes were expressed by the explants when compared to the other timepoints, suggesting that degenerative changes take place in the tissue after 48 h of culture.We suggest that cultured explants may be a suitable exemplar to further study the equine endometrium with particular application for modelling equine mating-induced endometritis progression to persistent inflammation between 24 and 48 h once in culture. Following future validation at the protein level, the explant tissue culture model may be used for wider studies applied to endometritis in the mare and potentially also in dogs and pigs as well as for investigating conditions underlying the endometrium.

animals : an open access journal from mdpi

10.3390/ani11082361

PMC8388726

Two-dimensional speckle-tracking echocardiography represents an advanced imaging technique that allows the analysis of global and regional myocardial function, cardiac rotation and synchronicity using deformation imaging. It has gained growing importance over the last decade, especially in human medicine as a method of evaluating myocardial function. This review aims to give an overview of the current understanding of this technique and its clinical applicability in the field of veterinary medicine with a focus on early detection of left ventricular dysfunction in dogs.

Two-dimensional speckle-tracking echocardiography (2D–STE) is an advanced echocardiographic technique based on deformation imaging that allows comprehensive evaluation of the myocardial function. Clinical application of 2D–STE holds great potential for its ability to provide valuable information on both global and regional myocardial function and to quantify cardiac rotation and synchronicity, which are not readily possible with the conventional echocardiography. It has gained growing importance over the past decade, especially in human medicine, and its application includes assessment of myocardial function, detection of subclinical myocardial dysfunction and serving as a prognostic indicator. This review illustrates the fundamental concepts of deformation analysis and gives an overview of the current understanding and its clinical application of this technique in veterinary medicine, with a focus on early detection of left ventricular (LV) dysfunction in dogs.

1. IntroductionStandard echocardiography is a widely used non-invasive method for evaluating myocardial function. In veterinary medicine, left ventricular (LV) fractional shortening (FS) is the most commonly used echocardiographic measurement of systolic function, readily obtained from LV chamber dimensions [1,2]. However, FS does not reflect the true systolic function since it only assesses myocardial shortening in a radial direction at a pair of specific myocardial segments even though myocardial contraction occurs in multiple directions [2]. Additionally, FS is known to be influenced by loading conditions [2]. To overcome these shortcomings, deformation imaging was developed to provide a comprehensive analysis of myocardial function, and the most widely used technique in both human and veterinary medicine is the 2D–STE.2. What Is Two-Dimensional Speckle-Tracking Echocardiography?The 2D–STE technique is advanced echocardiography that assesses myocardial function by quantifying myocardial deformation. It is based on the formation of “speckles” (natural acoustic markers) generated by interactions between myocardial tissue and ultrasound beams in standard gray scale two-dimensional (2D) echocardiographic images [3,4]. Tracking these speckles from one frame to another enables the analysis of myocardial movement throughout the entire cardiac cycle and provides information on deformation, which are measured as strain and strain rate (SR) [3,4].Strain (Lagrangian strain, ε) is a unitless measurement of deformation of the myocardium over time, expressed as percent change from its original dimension. It is calculated as ε = L − L0/L0, where L is the instantaneous length, and L0 is the initial length [4,5]. SR (ε’) is the temporal derivative of strain expressed as s−1, representing the rate of deformation [4,5]. The lengthening, thickening and clockwise rotation of the myocardium are expressed as positive deformation, whereas shortening, thinning and counterclockwise rotation are negative [3].3. Advantages and Disadvantages of 2D–STE3.1. AdvantagesThe basis of 2D–STE is the B-mode image and strain parameters are calculated from the 2D displacement of the myocardium along the myocardial wall, whereas tissue Doppler-derived strain parameters are dependent on the angle between the beam and the direction of myocardial movement [3,4,6]. Therefore, 2D–STE is considered angle independent, which allows evaluation along different spatial orientations [3,4]. Additionally, since the strain and SR are measured using 2D intra-tissue velocities, these deformation parameters are also independent of cardiac translational movement and tethering, and this permits the differentiation between an active myocardial deformation and a passive displacement of the myocardium [3,4]. Furthermore, 2D–STE allows the simultaneous evaluation of global and regional myocardial function [4]. A six-segment model is the most frequently used, where the LV myocardium is divided into six segments, and regional strain and SR are the average values within each segment [5]. Not only does segmental analysis allow the identification of regional myocardial abnormalities, but it also allows the quantification of LV synchronicity, which is identified by the variability among each segment [7]. Global strain and SR are the average values of all segments, which represent the overall myocardial function [4,5].The analysis of 2D–STE is performed using offline software, and its measurement involves the manual or semi-automatic tracing of the myocardial border by the operator. Then the software algorithm automatically tracks the speckles throughout the cardiac cycle. This semi-automatic method results in lower inter- and intra-operator variability of 2D–STE analysis [4].3.2. DisadvantagesThere are several factors to be wary of when using 2D–STE to analyze myocardial function. One is the potential measurement variability produced by software differences in the analytical algorithms used by different vendors [5,8]. Farsalinos et al. compared the strain measurements of 62 human volunteers having variable LV function obtained from nine different vendors, and they found moderate, but statistically significant, vendor variability with a maximum absolute differences up to 3.7% (p < 0.001) [8]. It was concluded that these should not have a major impact on clinical interpretation; however, ideally, the same echocardiographic machine and software should be used for subsequent examinations [8]. Additionally, offline analysis requires high-quality image data obtained at a high frame rate (optimally, 50–70 frames per second), and it is time consuming compared to conventional echocardiography, making 2D–STE disadvantageous for routine clinical use [4,6].Lastly, although to a lesser degree than FS, studies of both humans and dogs have shown that strain and SR are also partially dependent on loading conditions; therefore, care should be taken when interpreting the measurements of deformation in disorders involving altered loading conditions [6,9,10]. It should be noted that parameters of deformation analysis are not measurements of contraction but rather an estimation of systolic function [6]. 4. LV Deformation and 2D–STE During cardiac contraction, LV deformation occurs in three spatial orientations: systolic shortening and diastolic stretching in the longitudinal and circumferential planes, and systolic thickening and diastolic thinning in the radial plane (Figure 1) [4]. Additionally, in systole, there is a counterclockwise movement of the apex and a clockwise movement of the base; in diastole, the direction is the opposite. This is known as rotation (Figure 2) [4,11].4.1. Longitudinal DeformationLongitudinal strain is evaluated in apical 2-, 3-, 4-chamber views, and it represents myocardial deformation along the longitudinal axis (Figure 3a). In humans, it has been shown that global longitudinal strain measurements are superior, if not comparable, to the LV ejection fraction (EF) for inter- and intra-observer variability [8]. Additionally, polar projection of the longitudinal strains, the so-called bull’s-eye map, obtained from superimposing apical 2-, 3-, 4-chamber views, allows evaluation of regional myocardial changes and is useful in visualizing regional homogeneity [12].Longitudinal fibers of the myocardium consist mainly of sub-endocardial fibers and most significantly contribute to the longitudinal myocardial function [13,14,15]. This has been demonstrated in both humans and dogs as these sub-endocardial fibers are shown to be most susceptible to ischemic changes since they are located the furthest from the epicardial coronary blood supply [14,15]. In humans, sub-endocardial changes to the myocardium are seen in ischemic injuries and in the first stages of various diseases such as arterial hypertension [4]. Therefore, longitudinal strain is considered to be a sensitive indicator of LV dysfunction and suitable for the early detection of sub-endocardial change. 4.2. Radial & Circumferential Deformation Radial and circumferential strains are evaluated in parasternal short axis views at the level of basal, papillary muscles and apex (Figure 3b,c). Like the longitudinal deformation, a bull’s-eye map can be created by superimposing the parasternal short axis views of the base, papillary muscle and apical levels (Figure 4).Myocardial contractility occurs in radial and circumferential directions, and radial and circumferential strains have been shown to be sensitive indicators of myocardial contractility in dogs [16]. It must be noted that radial and circumferential strains also increase to compensate for longitudinal dysfunction. This has been documented both in humans and dogs, indicating that increase in contractility does not always suggest improvement of myocardial function [4,17].4.3. Rotational DeformationCombining the two opposing movements at the apex and the base results in the twisting of the heart during systole and untwisting during early diastole along the long axis [4,11]. The rotational deformation is measurable using two parameters, twist and torsion [5]. Twist, expressed in degrees, is the difference in the systolic rotation in apical and basal levels of the short axis view, whereas torsion, expressed in degrees/cm, is the value of twist normalized to the length of LV cavity, which is twist divided by the distance between the apex and base [4,8]. LV rotation is thought to be a sensitive indicator of altered LV function in humans [4]. However, the task force recommendations by the European Association of Cardiovascular Imaging (EACVI) and the American Society of Echocardiography (ASE) consider both parameters to be poorly defined and stated they should be used with caution [8]. The terms “twist” and “torsion” are often used interchangeably in the veterinary literature, in most cases referring to the LV twist [11,18,19,20].4.4. SynchronicityLV dyssynchrony, defined as an uncoordinated contraction of LV, has detrimental effects on regional myocardial perfusion, metabolism and electrophysiology, and leads to compromised global LV systolic function and energy insufficiency [21]. In humans, it is a known major contributor of heart failure (HF) and a powerful predictor of mortality in patients with HF [21]. Although evaluation of LV dyssynchrony is feasible in all three directions of strain analysis, radial strain allows the most accurate detection of LV dyssynchrony and is considered most useful for humans [4,7]. It must be noted that there is no standardized method for the analysis of LV dyssynchrony. There is a variety of parameters in both the human and veterinary literature. The two most common are the maximal time delay between peak systolic strain of the earliest and latest segments and the standard deviation of time-to-peak systolic strain across the segments [7,10,22,23,24,25]. 5. 2D-STE in Veterinary Medicine5.1. Validation of 2D–STE in Canine ModelsTable 1 shows a list of studies that investigated 2D–STE using canine experimental models. Sonomicrometry, which allows the measurements of the distance between ultrasonic crystals implanted in the myocardium, is often used as a reference method to validate 2D–STE. In the canine myocardial ischemic model, created by the transient occlusion of the left anterior descending coronary artery, 2D–STE showed adequate sensitivity in detecting changes in myocardial function produced by the alteration of regional myocardial blood flow, which had good correlation with the sonomicrometry [11,26,27,28]. A study by Adachi et al. demonstrated that during acute coronary blood-flow occlusion, a reduction in radial strain was observed to a lesser degree compared to the longitudinal and circumferential strains; however, the differences were statistically insignificant [13]. This may be due to a greater standard deviation in the radial strain, which may have resulted in lower accuracy [13]. Moreover, Stendahl et al. showed a similar correlation between 2D–STE and sonomicrometry in both radial and circumferential strains (r = 0.56 and r = 0.55, p < 0.001, respectively); nevertheless, the circumferential strain had better correspondence with a smaller bias [27]. These results suggest that the strain values from longitudinal and circumferential directions may be more favorable. The validation of twist between 2D–STE and sonomicrometry also showed good agreement (r = 0.94, p < 0.001) [11]. The usability of 2D–STE in dogs with HF was evaluated using tachycardia-induced myocardial dysfunction model [22,25,29,30,31]. These studies showed significant changes in systolic parameters of both conventional echocardiography and strain analysis in dogs with HF, which reflected the myocardial changes caused by high-electrical pacing [22,25,31]. With the analysis of rotation, the twist decreased significantly with the development of HF and returned to normal upon recovery, which suggests that the twist is a good indicator of myocardial function [30,31]. Arita et al. revealed that EF and tissue Doppler-derived dyssynchrony parameters were able to detect significant changes only in dogs with HF (with and without dyssynchrony) but not in dogs with dyssynchrony without HF [22]. Moreover, the M-mode derived dyssynchrony parameter was only able to detect significant changes in dogs with HF with dyssynchrony [22]. In comparison, both the radial and circumferential strains were able to detect significant differences in all three groups [22]. These strains not only detected dyssynchrony in HF better than the tissue Doppler- and M-mode-derived dyssynchrony parameters, but they effectively detected dyssynchrony without HF. It should be noted that Hamabe et al. observed a significant dyssynchrony only in the radial direction, and Arita et al. detected dyssynchrony in both directions with similar accuracy but with less variability in the radial direction [22,25]. These results may suggest the better sensitivity of radial strain to detect dyssynchrony. animals-11-02361-t001_Table 1Table 1Studies of canine experimental models with two-dimensional speckle tracking echocardiography (2D–STE).StudyModelObjectiveParameterOutcomeMyocardial ischemic modelAdachi et al. [13]LAD coronary artery occlusion (n = 13)Investigate changes in parameters of 2D–STELS, RS, CSReduction in RS occurred later than LS and CSNo significant differences in diagnostic accuracyAmundsen et al. [26]LAD coronary artery occlusion (n = 9)Validate 2D–STE against SMlong-, short-axis strainsGood correlation between STE and SMLong-axis strain: r = 0.90, p < 0.001Short-axis strain: r = 0.79, p < 0.001Helle-Valle et al. [11]LAD coronary artery occlusion (n = 13)Validate 2D–STE against SMTwistGood correlation between STE and SMTwist: r = 0.94, p < 0.001Pirat et al. [28]LAD coronary artery occlusion (n = 7)Validate 2D–STE against SMLS, CSGood correlation between STE and SMLS: r = 0.83, p < 0.001CS: r = 0.88, p < 0.001Stendahl et al. [27]LAD coronary artery occlusion (n = 7)Validate 2D–STE against SMRS, CSModerate correlation between STE and SMRS: r = 0.56, p < 0.001CS: r = 0.55, p < 0.001CS showed better correspondence with smaller biasTachycardia-induced heart failure modelArita et al. [22]Dyssynchrony without HF (D) (n = 12)Dyssynchrony with HF (DHF) (n = 9)HF with narrow QRS (HF) (n = 8)Evaluate synchronicity using CE, TDI and STESynchronicityEF: Significant difference in HF, DHF (p < 0.05)TDI: Significant difference in DHF (p < 0.05)M-mode: Significant difference in DHF (p < 0.05)RS, CS: Significant difference in all D, DHF, HF (p < 0.05)Hamabe et al. [25]RV pacing at 250 bpm for 3 weeks (n = 5)Compare CE with STERS, CS, synchronicitySignificant reduction of RS and CS (p < 0.01) Dyssynchrony only observed in radial direction (p < 0.05)Kusunose et al. [30]RV pacing at 220 bpm for 4 weeks (n = 7)Establish normal values of STEQuantify impact of tachycardia-induced cardiomyopathyLS, RS, CS, twistDecrease in LS, RS, CS (p < 0.001) and twist (p < 0.05)Most profound effect in the LV apex (p < 0.001)RS had the largest relative decrease (p < 0.05)Wong et al. [31]RV pacing at 230–250 bmp for 2–4 weeks (n = 6)Evaluate the effect of pacing using STELS, CS, twist, torsionLS, CS, twist and torsion significantly decreased with HF and improved with recovery (p < 0.05)Dyssynchrony modelMochizuki et al. [24]LBB ablation (n = 10)Determine diagnostic value of dyssynchrony parameters of STESynchronicityDyssynchrony parameters showed significant increase after ablationRS allowed detection of dyssynchrony with high sensitivity and specificitybpm, beats per minute; CE, conventional echocardiography; CS, circumferential strain; EF, ejection fraction; HF, heart failure; LAD, left anterior descending; LBB, left bundle branch; LS, longitudinal strain; LV, left ventricle; RS, radial strain; RV, right ventricle; SM, sonomicrometry; TDI, tissue-Doppler imaging.5.2. 2D-STE in Clinically Healthy DogsThe normal values of strain analysis in dogs are comparable those of healthy humans (Table 2) [32,33]. Normal radial strain values show greater variability compared to longitudinal and circumferential strains, and the source of this variability is thought to be technical rather than biological [32]. The 2D–STE values in dogs tend to increase from base to apex with the highest value at the apex, which is also seen in healthy humans where alteration in this base-to-apex gradient of LV deformation is associated with various cardiac pathologies [20,30,34,35,36]. Chetboul et al. was the first to provide data on strain analysis in awake dogs [10]. Radial strain and SR of 37 clinically healthy dogs revealed adequate intra-observer repeatability (within-day variability) and reproducibility (between-day variability) with a coefficient of variation (CV) of less than 10% [10]. Likewise, a longitudinal strain had a CV of intra-observer repeatability and reproducibility and of inter-observer repeatability that was less than 10%, with the exception of inter-observer repeatability of LV free wall, which was 15.1% [37]. No correlation was found between age and systolic parameters of 2D–STE; however, between young and old dogs there were significant differences in diastolic deformations (early and late diastolic RS) that are consistent with increased ventricular stiffness and delayed relaxation resulting in reduced diastolic function, which is commonly observed with age [10,19]. Conflicting results in heart rate (HR) were reported, Chetboul et al. described a positive correlation with radial strain and SR (r = 0.41, p = 0.01; r = 0.56, p < 0.001, respectively), whereas, a study with controlled HR using right-atrial pacing showed no significant changes with increased HR [10,38]. Additionally, no correlation was found between body weight (BW) and strain parameters [10]. Twist had a reasonable intra-observer variability (within-day CV of 16.37% and between-day CV of 6.84%) and appeared not to be affected by HR or BW [18]. In healthy humans, twist was shown to increase with age, but studies in normal dogs revealed no correlation with age, which could be attributed to small sample sizes [18,19,20,39]. The 2D–STE analysis in dogs showed LV synchrony, and the synchrony parameters appeared to be independent of age, HR and BW [10,23].animals-11-02361-t002_Table 2Table 2Normal values of two-dimensional speckle tracking echocardiography (2D–STE) in clinically healthy dogs.StudySubjectLS (%)RS (%)CS (%)Twist (°)Synchronicity (ms)FS (%)Chetboul et al. [10]MBn = 37-46.7 ± 12.2(26.1−69.2)--R STI: 15 ± 15(0−49)38.4 ± 6.0(30.0−49.0)Chetbou et al. [18]MBn = 35---8.4 ± 3.8(2.5−18)-8.4 ± 3.8(2.5−18)Griffiths et al. [23]MBn = 10----R STI apex: 41.8 ± 17.9base: 36.2 ± 19.0R TsSD apex: 17.7. ± 7.2base: 13.8 ± 7.3-Hamabe et al. [25]Beaglesn = 5-31.96 ± 7.12−15.44 ± 1.5-R STI: 44.25 ± 17.6C STI: 41.63 ± 12.6834.7 ± 4.6Kusunose et al. [30]Mongrelsn = 25−18 ± 439 ± 20−17 ± 48.1 ± 4.4-EF: 61 ± 8Pedro et al. [34]Great Danesn = 39-47.18 ± 12.00−16.73 ± 2.58--27.8 ± 5.5Smith et al. [40]MBn = 20-43.9 ± 8.54−20.9 ± 3.15--40 ± 6.16Suzuki et al. [19]BeaglesYoung (n = 17)Old (n = 15)−14.8 ± 3.1(−8.9 to −21.9)−14.9 ± 4.7(−4.3 to −23.2)52.4 ± 11.1(27.7−70.0)50.1 ± 12.3(21.4−64.1)−19.4 ± 4.4(−10.1 to −25.8)−17.6 ± 2.5(−13.3 to −21.2)14.7 ± 4.6(7.4−28.0)13.6 ± 5.8(3.9−25.2)-37.4 ± 7.3(27.6−58.8)37.6 ± 7.0(25.4−48.7)Suzuki et al. [17]MBn = 20−19(−23 to −14)55(47−60)−23(−27 to −20)--39(35−41)Wess et al. [37]MBn = 100IVS: −16.89 ± 4.26LVFW: −15.18 ± 5.86-----Westrup et al. [20]Irish wolfhoundsn = 46−16.2 ± 3.0(−22.2 to −10.2)Apical: 45.1 ± 10.4 (24.3−65.9)Basal: 36.9 ± 14.7 (7.5−66.3)Apical: −24.8 ± 6.2(−12.8 to −38.5)Basal: −15.9 ± 3.2(−22.3 to −9.5)11.5 ± 5.1(1.3−21.7)--Data are expressed as mean ± standard deviations (range). C, circumferential; CS, circumferential strain; EF, ejection fraction; FS, fractional shortening; IVS, interventricular septum; LS, longitudinal strain; LVFW, left ventricular free wall; MB, mixed breed; R, radial; RS, radial strain; STI, synchrony time index; TsSD, standard deviation of time to peak systolic segmental motion.6. Clinical Application of 2D–STE 6.1. Cardiac Disorders6.1.1. Myxomatous Mitral Valve Disease (MMVD)MMVD is the most common acquired cardiac disease in dogs, characterized by chronic myxomatous degeneration of the mitral valve resulting in valvular dysfunction with secondary mitral valve regurgitation (MR) [41,42]. Although, LV systolic dysfunction is an important prognostic indicator, altered hemodynamic loading conditions in MMVD make using conventional echocardiography to assess LV function a challenge [42,43].Smith et al. evaluated asymptomatic dogs with Stage B2 MMVD using 2D–STE (Table 3) [40,44]. When compared to the control group, the Stage B2 group had a significantly higher HR, greater LV size and LV systolic function, including both radial and circumferential strains, but LV dysfunction could not be identified [40]. Zois et al. observed similar results in increased strains, SRs and twist for dogs with congestive heart failure (CHF) compared to dogs with minimal or no MR (Table 3) [45,46]. These parameters increased with the severity of MMVD, which suggest augmented LV function [45,46]. However, longitudinal and radial strains and longitudinal SR showed curvilinear relationships with the left atrial-to-aortic ratio (LA/Ao), illustrating a decrease in LV function in dogs with CHF and severe left atrial enlargement [45]. These studies were unable to demonstrate LV dysfunction prior to the onset of clinical signs of CHF, and the hyperdynamic values of strain analysis may reflect a compensatory mechanism important for the preservation of LV function [46].6.1.2. Dilated Cardiomyopathy (DCM) DCM is the most common myocardial disease in dogs, characterized by progressive chamber dilation and impaired myocardial contractility [52,53,54,55]. It is well known that the asymptomatic “preclinical” DCM phase extends up to several years before any symptoms appear, during which the diagnosis can be challenging [52,53,54]. A study with 50 Great Danes diagnosed as preclinical DCM revealed an overall decrease in radial and circumferential strains and SRs at the base, papillary muscle and apical levels, with the greatest difference observed at the papillary muscles (Table 3) [34]. Additionally, a similar base-to-apex gradient of 2D–STE values was observed, but it was reduced in comparison with the clinically normal Great Danes, suggesting reduced systolic function in dogs with preclinical DCM [34]. Ro et al. reported serial changes of N-terminal pro-brain natriuretic peptide (NT–proBNP) and 2D–STE with disease progression observed in a Golden Retriever diagnosed with DCM and sub-aortic stenosis (SAS) (Table 3) [16]. Concomitant SAS makes this an atypical DCM; however, the dog presented with mainly DCM rather than SAS characteristics, and it fulfilled all proposed criteria for diagnosis of canine DCM [16]. It was speculated that the concomitant SAS resulted in worsened myocardial dysfunction and cardiac remodeling than DCM alone and resulted in a much shorter survival time [16]. Improvement in clinical signs, HR, NT–proBNP level and echocardiographic parameters of LV contractility were observed with treatment; however, segmental dyskinesia in the apical segment was detected with a regional analysis of 2D–STE. This regional deterioration of myocardial function, despite the increased contractility, suggested that the overall improvement of myocardial function did not necessary reflect the improvement of all myocardial segments. It also demonstrated that the regional analysis of 2D–STE was able to detect segmental myocardial dysfunction which was undetectable with NT–proBNP or conventional echocardiography [16]. They also revealed that the longitudinal strain and SR were the most sensitive and accurate indicators of the myocardial damage detected by NT–proBNP [16]. On the other hand, radial and circumferential strains and SRs were the most sensitive indicators of myocardial contractility, and these parameters also increased as compensation for longitudinal dysfunction, which was consistent with the findings of others [14,16,17]. Increased contractility does not always equate to improved myocardial function; therefore, it is necessary to evaluate all three directions of 2D–STE for a precise assessment of myocardial function [16]. 6.1.3. Patent Ductus Arteriosus (PDA)PDA is the most common congenital heart defect in dogs, resulting from the failure of the ductus arteriosus to close, a normal fetal structure that shunts blood from the pulmonary artery to the aorta by bypassing the nonfunctional lung and normally closes soon after birth [56]. After birth, the rise in systemic pressure and the drop in pulmonary artery pressure cause blood to flow through the PDA from the aorta to the pulmonary artery (left-to-right shunt), resulting in pulmonary over circulation and volume overload of the left atrium and LV [56]. A study by Spalla et al. compared dogs with PDA with healthy controls and found a significant increase in LV dimensions, indicating LV overload in the PDA groups (Table 3) [47]. An increase in preload increases contractility by the Frank–Starling low, but EF and FS were not different between the two groups, which suggests the possible underestimation of contractility in dogs with PDA [47]. On the other hand, the 2D–STE parameters of longitudinal, radial and circumferential directions showed significant differences between the two groups, possibly suggesting that 2D–STE is a more sensitive indicator of systolic function [47]. Subsequently, Spalla et al. evaluated changes associated with the surgical closure of the PDA (Table 3) [48]. Ductal closure resulted in significant decreases in conventional echocardiographic parameters of LV dimensions and contractility including EF and FS as a result of the decreased preload and increased afterload [48]. A similar result was observed by Hamabe et al., where diastolic dimensions and FS decreased significantly (Table 3) [9]. Additionally, the PDA closure resulted in a significant reduction in radial and circumferential strains and SRs, whereas longitudinal strain and SR did not show any change [9,48]. The decrease in preload resulted in reduced tension on the myocardial wall and the radius of the LV, leading to reduced contraction of the radial and circumferential fibers, which may explain the reduced radial and circumferential strains and SRs [48]. On the other hand, longitudinal strain and SR are sensitive to early changes in systolic function, so the lack of changes may suggest that the PDA closure was not associated with systolic dysfunction [15,48]. Decreased contractility associated with the PDA closure, observed as decreased FS and radial and circumferential strains was most likely due to sudden changes in loading conditions. Such results suggest that the radial and circumferential strain parameters may be indicators of myocardial contractility, but they are also at least partially influenced by the loading condition. 6.2. Non-Cardiac Disorders6.2.1. Systemic Inflammatory Response Syndrome (SIRS) SIRS is a clinical syndrome of infectious or non-infectious origin causing secondary multiple-organ dysfunction or death due to the excess release of cytokines. Cardiac dysfunction in SIRS has been demonstrated by an increase in cardiac biomarkers, such as NT–proBNP, cardiac muscle troponin T (cTnT) and lactate in dogs [57,58]. Although FS and LA/Ao have been shown to correlate significantly with survival-to-discharge, the detection of systolic dysfunction in dogs with SIRS has not been possible with conventional echocardiographic parameters, such as FS and EF [49,57]. A study by Corda et al. demonstrated that the endocardial longitudinal strain was able to identify systolic impairment in mild to moderate stages of SIRS, but it was not detected by the conventional echocardiography (Table 3) [49]. The endocardial longitudinal strain was significantly reduced (p = 0.001), without affecting the epicardial longitudinal or the radial strains [49]. The sub-endocardial myocytes are located the furthest from the epicardial coronary artery, and therefore it is considered most vulnerable to ischemia [14,15]. Consequently, SIRS-associated microvascular alterations resulting in myocardial ischemia may contribute to reduced endocardial longitudinal strain [59]. 6.2.2. Hyperadrenocorticism (HAC) HAC is an endocrinological disorder characterized by the chronic elevation of glucocorticoid in the blood. Systemic hypertension, LV hypertrophy and myocardial fibrosis are commonly recognized cardiovascular changes in humans and dogs with HAC [60,61]. In veterinary medicine, LV hypertrophy is reported anywhere from 47.3 to 68% of dogs with HAC [50,60]. Myocardial fibrosis results in increased LV stiffness and impaired LV relaxation causing LV diastolic dysfunction, which can be observed with the conventional echocardiography [50,60,61]. A study by Chen et al. showed a significant decrease in global strains and global peak systolic and early diastolic SRs in longitudinal and circumferential directions, suggesting impaired systolic and diastolic function in dogs with HAC (Table 3) [50]. Similar findings were reported in humans, where increased myocardial fibrosis in HAC patients resulted in both LV systolic and diastolic dysfunction [61]. In the study by Chen et al., conventional echocardiographic parameters of LV systolic function failed to detect any significant differences between dogs with HAC and controls, whereas 2D–STE was able to reveal a decrease in systolic function [50]. Such findings suggest that dogs with HAC may have subclinical systolic dysfunction that may be undetectable with conventional echocardiography [50].6.2.3. Parvoviral Enteritis (PVE)PVE, caused by the infection of canine parvovirus (CPV-2), is presented as acute hemorrhagic gastroenteritis with high fatality in young dogs [62,63]. CPV-2 is transmitted via the fecal-oral route, and targets cells with high mitotic activity for viral replication, such as the intestinal epithelium and lymphoid tissue [62,63]. Myocarditis may also occur if the infection occurs within the first 2–3 weeks of life, when rapid myocardial cell proliferation takes place [62,63]. Moreover, myocarditis may also occur secondary to PVE, resulting from sepsis caused by the disruption of gastrointestinal barrier [64]. The hypovolemic state of the PVE caused by dehydration results in reduced preload, making detection of impaired myocardial function with conventional echocardiography difficult [51]. A study by de Abreu et al. demonstrated the presence of systolic dysfunction in dogs with PVE, indicated by impaired strains and SRs, while the conventional echocardiographic parameters failed to detect any changes (Table 3) [51]. Impairment of the longitudinal strain and SR at both endocardial and epicardial levels and circumferential strain and SR at the endocardium were observed in all dogs with PVE [51]. On the other hand, impairment of the circumferential strain in the epicardium was observed only in dogs that died from PVE, while circumferential SR in the epicardium remained normal [51]. Furthermore, regional analysis revealed the lowest circumferential strain and SR in the mid-septal epicardial segment in non-surviving dogs [51]. In fact, circumferential SR in that particular segment of less than 0.95 s−1 made possible the distinction between dogs with severe PVE and non-surviving dogs with 100% sensitivity and specificity [51]. Such results suggest that longitudinal strain and SR allow early detection of myocarditis in PVE, and a circumferential SR below 0.95 s−1 at the mid-septal epicardial segment may indicate a grave prognosis [51]. 7. The Future of 2D–STEIn human medicine, the importance of 2D–STE analysis in evaluating myocardial function has been recognized, and a consensus document has been published by the EACVI and ASE to standardize deformation analysis [5]. Moreover, improvement in technology has allowed the development of 3D analysis. Similarly in veterinary medicine, the 2D–STE analysis has attracted great interest, expanding outside the simple analysis of LV and assessment of myocardial function. For example, several studies described the utility of 2D–STE for studying right-ventricular, right-atrial, and left-atrial function in dogs [65,66,67,68,69,70,71]. Additionally, studies demonstrated how 2D–STE analysis can be used to predict survival in MMVD and the onset of atrial fibrillation [72,73]. 8. Conclusions2D–STE analysis has gained increasing significance in human medicine over the past decade, and it is also becoming a promising tool for evaluating myocardial function in veterinary medicine. There is growing evidence that superior deformation parameters can serve as an indicator of myocardial function compared to conventional parameters for early diagnosis of myocardial dysfunction and as a possible prognostic indicator in both human and veterinary medicine. Additional information provided by the analysis of rotation, synchronicity and bull’s-eye maps by the 2D–STE analysis allow more comprehensive and accurate evaluation of myocardial function. There is no doubt that further development will allow deformation analysis to mature and play an important role in routine clinical use, with better diagnostic accuracy and reliability for dogs.

animals : an open access journal from mdpi

10.3390/ani11051240

PMC8146606

Disease of the heart muscle (cardiomyopathy) is very common in the domestic cat and may result in several severe outcomes. These include formation of a thrombus in the left atrium which migrates to the hindlimb cutting off the blood supply, a condition called aortic thromboembolism. Affected cats present with hindlimb paralysis and extreme pain, often requiring euthanasia on humane grounds. Several factors are known to predispose to thrombus formation, including damage to the inner cellular lining of the atrium which exposes proteins that initiates thrombosis. We studied the expression of one such protein called von Willebrand Factor in the left atrium of cats with and without cardiomyopathies and at different stages of disease severity. We found that expression increased in cats with advance disease. Obtaining a greater understanding of the role this protein has in thrombus formation may allow development of novel antithrombotic agents to help prevent this devastating consequence of feline cardiomyopathy.

Aortic thromboembolism (ATE) occurs in cats with cardiomyopathy and often results in euthanasia due to poor prognosis. However, the underlying predisposing mechanisms leading to left atrial (LA) thrombus formation are not fully characterised. von Willebrand Factor (vWF) is a marker of endothelium and shows increased expression following endothelial injury. In people with poor LA function and LA remodelling, vWF has been implicated in the development of LA thrombosis. In this study we have shown (1) the expression of endocardial vWF protein detected using immunohistofluorescence was elevated in cats with cardiomyopathy, LA enlargement (LAE) and clinical signs compared to cats with subclinical cardiomyopathy and control cats; (2) vWF was present at the periphery of microthrombi and macrothrombi within the LA where they come into contact with the LA endocardium and (3) vWF was integral to the structure of the macrothrombi retrieved from the atria. These results provide evidence for damage of the endocardial endothelium in the remodelled LA and support a role for endocardial vWF as a pro-thrombotic substrate potentially contributing to the development of ATE in cats with underlying cardiomyopathy and LAE. Results from this naturally occurring feline model may inform research into human thrombogenesis.

1. IntroductionFeline aortic thromboembolism (ATE) is a severe complication that can occur in 11.6–21% of cats with myocardial disease [1,2,3,4,5,6]. It happens when a thrombus usually originating in the left atrium (LA) or left atrial appendage (LAA) dislodges and obstructs a branching artery of the aorta. Common clinical signs frequently relate to the pelvic limbs and include severe pain, paraparesis, paraplegia, absence or reduced strength of femoral pulses, and cyanotic paws. ATE in cats carries a poor prognosis, with over 60% of affected cats being euthanised on presentation at first opinion practices [7].Virchow’s triad describes three factors that contribute to venous thrombosis: blood stasis, endothelial injury, and hypercoagulability [8]. Abnormal findings associated with Virchow’s triad have been reported in cats with a variety of cardiomyopathies, which provides insights into the potential pathogenesis of feline ATE. For instance, development of spontaneous echo contrast (SEC) in the LA of cats with cardiomyopathy is associated with decreased LAA blood velocity and blood stasis [9]. A hypercoagulable state has been suggested in cats with HCM [10] and histopathologic evidence of LA endothelial damage has been observed in cats with CHF [11]. A further indication of endothelial dysfunction in cats with ATE is that plasma arginine, the precursor to nitric oxide critical for endothelial health, is lower in these cats than in than cats with cardiomyopathy alone [12]. However, the contribution of endothelial injury in the left atrial endocardium to the development of ATE has not been studied at the sub-cellular level in cats with cardiomyopathy at different stages of their disease process and only to a limited extent in humans with heart disease.vWF is synthesised in endothelial cells and megakaryocytes and is an adhesive protein that plays an important role in thrombosis through the formation of multimers [13,14]. Endothelial vWF is either secreted into the plasma constitutively, stored in rod-shaped specialised compartments called Weibel-Palade bodies within the endothelial cells, or deposited in the subendothelium [15]. vWF of platelet origin is stored in alpha-granules within the platelet and is released upon activation. The main functions of vWF are to aid binding of platelets to exposed subendothelium and to assist in platelet aggregation via the glycoprotein Ib receptor on platelets [16,17]. The majority of plasma vWF originates from the endothelium, and in human patients the plasma concentration of vWF increases in various thrombogenic diseases reflecting endothelial injury [18]. Similarly, in cats with ATE and cardiomyopathy, the elevation in circulatory vWF has been suggested to be associated with endothelial damage [10].In human patients with atrial distension due to a variety of cardiac conditions, expression of circulating and LA endocardial vWF protein is increased and associated with the severity of LA blood stasis and atrial remodelling and has been suggested as a predisposing factor for thrombogenesis [19,20,21,22,23].Previous studies have also confirmed the pathogenic role of vWF in both arterial and venous thrombosis. High haemodynamic forces present in the systemic arterial system result in platelets binding to vWF through integrin αIIbβ3 for the initiation of aggregation [13,24], while patients with vWF deficiency as a result of von Willebrand disease are partly protected against arterial thrombosis [25]. vWF also contributes to thrombosis in veins, through formation of an extracellular scaffold with platelets, leukocytes, and fibrin to trap erythrocytes [26]. Mice deficient in vWF were protected from induced deep vein thrombosis [26]. Although fibrin is the best studied factor involved in trapping red blood cells (RBCs) in slow blood velocity environments [27,28,29,30], a recent in vitro study showed that vWF can bind to RBCs directly and the degree of binding is increased when blood stasis is present [31].Currently, there is limited information regarding the contribution of endocardial vWF to thrombosis in cats with myocardial disease. However, given the pro-thrombotic action of vWF in both rodent models and human patients described above, it is likely that vWF also plays an important role in the development of LA thrombosis in cats. The aim of this study was to investigate the expression of vWF in the LA of cats at different clinical stages of myocardial disease.2. Materials and Methods2.1. Study PopulationCats with clinical signs related to cardiomyopathy were enrolled based on their clinical presentation of ATE, CHF or both. The control cats and those with preclinical cardiomyopathy (without clinical signs of disease) were recruited from referral and first opinion cases euthanised for a variety of reasons. Cardiomyopathy was confirmed based on characteristic cardiac structural changes on gross and histopathology using criteria previously described [32,33] (see supplementary material Table S1 for histopathological diagnostic criteria for feline cardiomyopathies) in addition to complete or point of care (POC) echocardiographic examination [34,35,36] (see Table S2 for echocardiographic diagnostic criteria for feline cardiomyopathies). Briefly for full echocardiographic examinations, all measurements were taken over three different cardiac cycles and averaged. Measurements taken included left atrium to aortic ratio (LA/Ao), maximal left ventricular freewall thickness in diastole (LVFWd), maximal interventricular septum thickness in diastole (IVSd), presence of systolic anterior motion of the mitral valve (SAM), presence of spontaneous echo contrast (SEC), and presence of a formed thrombus in the LA. SAM was defined as anterior motion of either septal or both mitral valve leaflets during systole toward the LVOT using the right parasternal long axis five chamber view on review of 2D cineloops [37]. All echocardiographic examinations were performed by a veterinary cardiology diplomate or resident under direct supervision (see supplementary material Table S5 for echocardiographic views used [38,39]). POC examinations were performed in the emergency setting by a veterinary ECC diplomate or resident in training and facilitated measurement of LA/Ao as described above and a subjective assessment of left ventricular wall thickness in diastole.Cats were divided into four groups: (1) Control group: cats without structural and histopathological cardiac changes, (2) Subclinical group: cats with subclinical cardiomyopathy (some cats had mild LAE), (3) CHF group: cats with CHF attributable to cardiomyopathy with LAE, (4) ATE group: cats with ATE attributable to cardiomyopathy with LAE, irrespective of whether presenting with concurrent CHF.Inclusion and exclusion criteria are listed in Table 1.Diagnosis of ATE was based on clinical signs including acute fore/hind limb(s) paresis or plegia, loss of palpable femoral pulses, cold limbs, and cyanotic nail beds in association with myocardial disease identified by cardiac imaging [40]. CHF was diagnosed based on radiographic evidence of cardiogenic pulmonary oedema, ultrasonographical evidence of pleural effusion in association with cardiomyopathy, and left or bi-atrial enlargement. Cats with cardiomyopathy identified by cardiac imaging and gross and histopathology but showing no clinical signs of heart disease were determined to have subclinical cardiomyopathy.2.2. Sample CollectionThe heart was harvested within 30 min of euthanasia and flushed with slowly running tap water and each chamber was further gently flushed using a 20 mL syringe to ensure all blood was removed to facilitate optimal fixation in 10% buffered formalin. The LA samples were collected from the LA free wall. Where a formed thrombus in the LA was identified, it was carefully removed and placed into 10% buffered formalin, while an in situ thrombus in the LAA was harvested without being removed from the LAA. After gross pathological examination, sections of the left atrial free wall and cross-sections of the ventricles at the heart base, midwall, and apex were embedded in paraffin wax in a routine manner for preparation of slides for haematoxylin and eosin and Masson’s trichrome staining for histopathological examination. The LA sample blocks were used to prepare slides for immunostaining.2.3. Fluorescent ImmunostainingAfter dewaxing, rehydration, antigen retrieval with citric acid (pH6) at 95 °C for 10 min, and blocking with 10% goat serum, slides were incubated with primary antibodies (1:500 Rabbit polyclonal IgG against vWF, Sigma, Gillingham UK; 1:25 Mouse monoclonal IgG1 against CD41 platelet marker integrin αIIb, clone B-9, Santa Cruz, CA, USA) at 4 °C overnight. An hour-long incubation of slides with 1% Bovine serum albumin/Tris-buffered saline (TBS) suspended 4′,6-diamidino-2-phenylindole (DAPI) (1:100 nuclei stain, Sigma) and secondary antibodies (conjugated with Cyanide Dyes, Cy2 1:100, Cy3 1:500, Jackson ImmunoResearch, Ely, UK) was completed at room temperature. IgG isotype control (1:400 Rabbit polyclonal IgG, Abcam, Cambridge, UK) was used to replace primary antibody against vWF to ensure specific binding. For assessment of nonspecific binding of secondary antibodies, reagent control was carried out by omitting all primary antibodies.2.4. Image AnalysisAll the slides were examined under a Leica DMRA2 microscope (Leica, Wetzlar, Germany) connected to a monochrome camera (AxioCam, Oberkochen, Germany) and 3 images covering the entire length of the LA endocardium were taken. The green fluorescence shown in grey scale from endocardial endothelium was selected free-hand and measured using ImageJ (https://imagej.net/Welcome, accessed on 28 February 2018). The three measurements were then averaged to give a final number representing the detected fluorescence of the endocardial sample. Exposure time was fixed for channels that detected fluorescence from DAPI and vWF for all slides. All the slides were coded so the observer (WCC) was fully blinded when analysing the images and the order in which they were viewed was randomised by a second person (DJC). The fluorescent signals from both IgG isotype control and reagent control were graphed for reference.A Leica DM4000B with DFC550 colour microscopy camera (Leica, Wetzlar, Germany) was used for light microscopy. The microscopes and cameras were controlled using the Leica Application Suite Version 4.12 (Leica, Wetzlar, Germany).2.5. Statistical AnalysisAll statistical analyses were performed on GraphPad (version 8) (GraphPad, San Diego, CA, USA). Histogram and Shapiro-Wilk test were used for inspection of data distribution and normality. ANOVA test with Tukey post-hoc test and Chi-squared test with Yate correction were used to assess the difference in age and sex. Kruskal-Wallis test with Dunn’s multiple comparison test was used to compare the fluorescent intensity of labelled vWF between groups. Difference with a p-Value < 0.05 was considered significant.3. Results3.1. AnimalsLA free wall samples from 39 cats were used for quantification of endocardial vWF. Thrombi were also retrieved from the LA of 3 (out of the total 39 cats) for immunohistological investigation. The LAA thrombus in situ was acquired from an extra cat that was collected later in the timeline thus not included in the quantification of endocardial vWF (Cat ATE 12 in supplementary material Tables S3 and S4). The inclusion and exclusion criteria for the 4 separate groups of cats are shown in Table 1. The cats in the control group were generally younger but this was not statistically significant (p = 0.137). Male cats were overrepresented in cats with cardiomyopathy (p = 0.009). Further demographic information about the cats used in the study are given in Table 2.See supplementary material Tables S3 and S4 for details of clinical presentation, cardiac imaging and histopathological diagnosis for each cat used in this study. One cat in the CHF group (Cat CHF 7 in Tables S3 and S4) had been given clopidogrel prior to presentation.3.2. Localisation of vWF Protein in the Left Atrial SamplesIn all cats, vWF could be observed to a variable degree in the vascular endothelium (Figure 1A,C). In the majority of cats with cardiomyopathy and clinical signs (groups ATE and CHF), microthrombi, defined as thrombi only visible on microscopic examination, were identified on the vascular endothelium and occasionally on the endocardium. Conversely, microthrombi were rarely seen in the control and subclinical groups. Identifiable components of these microthrombi using immunohistofluorescence were vWF (green), platelets (reddish orange), RBC (mild autofluorescence), and leucocytes (blue) (Figure 1).3.3. Quantification of Endocardial vWF Expression in Left Atrial SamplesThe intensity of vWF fluorescence at the endocardium (Figure 2) was quantified using ImageJ for comparison between groups. To avoid quantifying the fluorescence signals from platelets where vWF can also be detected, the slides were double immunostained for vWF and the platelet marker integrin αIIb (Figure 3).Representative images of vWF immunolabelling in different groups of cats are shown in Figure 4. Medians of the detected fluorescence intensity of endocardial vWF in the left atrial endocardium from the four groups from left to right as shown in Figure 5 were Control group: 30.8 (IQR 28.1–34.2), Subclinical 39.6 (IQR 31.8–59.6), CHF group: 46.0 (IQR 36.6–56.8), and ATE group: 44.7 (IQR 34.9–54.6). The fluorescence intensity was significantly higher in the ATE and CHF groups compared to the control group.3.4. Characterisation of vWF in Thrombi Obtained from LA and LAAThrombi were found in the LA or LAA in three cats (CHF 8, ATE 9, and ATE 12 in Tables S3 and S4) at postmortem. The necropsy images of the thrombi and the hearts can be found in Figure S1. The microscopic images of the thrombi, two retrieved from the LA and one remaining in situ in the LAA, are shown in Figure 6. These thrombi were immunostained for vWF (green) and platelets (reddish orange) and were counterstained with the nucleus stain (DAPI). WBC can be identified by the blue round-shaped nucleus stained by DAPI. RBC can be easily identified by their autofluorescence in the reagent controls (Figure 6, autofluorescence column) [41]. The autofluorescence of RBC was markedly weaker compared to the fluorophores bound to the antibodies used for labelling vWF, platelets and nuclei and thus appeared relatively dark in Figure 1, Figure 6, and Figure 7. The composition of each thrombi was very different in terms of relative proportions of the main components and how different components were organised. An enlarged image of the upper right quadrant of the LA thrombus from CHF 8 (Figure 6) is used as an example to show the various patterns of organisation (Figure 7A). Different patterns of vWF expression were also observed and enlarged images of these different patterns of expression are described in Figure 7B.4. DiscussionCats with underlying myocardial disease are known to be at a greater risk of thromboembolic events which frequently results in euthanasia on humane grounds [6,7]. Given the high prevalence of cardiomyopathy in cats, which for HCM is estimated at up to 17% in outbred animals and even greater in certain pedigree breeds [33], it is important to obtain a greater understanding of the left atrial remodelling process that predisposes to thrombus formation.This is the first study to quantify and characterise vWF in the LA endocardium and LA thrombi from cats with cardiomyopathy at different stages of their disease process. LA samples from cardiomyopathic cats with LAE and clinical signs (CHF and ATE), but not those from subclinical cardiomyopathic cats had increased endocardial vWF protein compared to control cats (Figure 5), which suggests an association of endocardial vWF elevation with an advanced stage of cardiomyopathy. The immunohistological evaluation of the thrombi revealed that vWF was extensively involved in thrombus formation and organisation of both microthrombi and the three macrothrombi, which were visible on gross pathology within the LA(A) in cats CHF 8, ATE 9, and ATE 12 shown in Figure S1.The finding that vWF localised to some vessels in the LA samples is consistent with previous report of differential expression of endothelial vWF in mice where vWF expression varied not just in different organs but even in the vessels in the same vascular tree [42]. The minimal detection of vWF in the endocardium in the control cats is similar to that in a human report where no or only minimal focal immunostaining for vWF was present in the LAA endocardium from patients without cardiac abnormalities [19,43]. It is also striking that there were no female cats in our ATE group, which is consistent with previous reports showing a male predisposition to ATE and to an increase hazard of an ATE associated death. However, the possibility of this finding being due to chance cannot be ruled out [7,44].Importantly, our finding in cats with advanced myocardial disease are similar to two human studies where patients with hemodynamical disturbance in the LA secondary to various cardiac causes showed elevated endocardial vWF compared to the non-cardiac patients using immunohistological microscopy [19,20]. In the first of these studies, a significant increase in the expression of vWF in the endocardium of atrial appendages was identified in human patients with a variety of congenital and acquired heart diseases irrespective of the presence of atrial fibrillation. Furthermore, increased vWF expression correlated with the degree of platelet adhesion and thrombus formation [19]. In the second study involving patients with valvular and non-valvular atrial fibrillation, expression of vWF protein in the endocardium correlated with the degree of structural remodelling in the atrial wall. In addition, endocardial vWF appeared important for platelet adhesion/aggregation on the endocardium and intra-atrial thrombi formation suggesting that increased endocardial vWF may contribute to LA thrombogenesis [20]. It is recognised that inflammation per se can cause thrombosis via a VWF-mediated mechanism by elevating the level of VWF, enhancing the reactivity of VWF, and modulating the levels and activities of regulatory molecules such as the metalloprotease ADAMTS13 [45]. Heart failure is an established cause of systemic inflammation, and the increased expression of vWF in our cohort of cats with clinical signs may at least in part be related to this pro-inflammatory milieu, however the relationship between circulating concentrations of vWF and heart failure remains unclear in human studies [46,47,48] and that between endocardial vWF and inflammation is not examined.The immunostaining of the micro- and macrothrombi supports a pivotal role for vWF in thrombus formation. Thrombosis is a multifactorial process which usually involves an abnormality in one or more of the following factors: endothelial damage and dysfunction, disturbed blood flow, and hypercoagulability. Thrombus formation begins when platelets adhere to the exposed subendothelial matrix by means of vWF that was previously deposited on the sub-endothelium by endothelial cells. Once adhered, the stimulated platelets then secrete more mediators such as ADP and thromboxane A2 that further activate additional platelets leading to propagation of the thrombus [49,50].Traditionally, two categories of thrombus have been described based on their gross appearance and composition. A red thrombus that forms under low shear stress and is primarily composed of erythrocytes and fibrin, and a white thrombus that forms under high shear stress and contains a large number of platelets and few erythrocytes. The former mostly comprise venous thrombi where stasis of blood is evident and endothelial injury is not an absolute requirement for thrombus formation. Conversely, white thrombi are usually confined to the arterial system and form after the severe endothelial disruption such as plaque erosion or rupture as part of the atherosclerosis process [51]. In the present study, platelets were found to be an important component in the thrombi despite being formed in the environment of slow blood flow and low shear stress found in the enlarged LA. Different patterns of vWF expression in the thrombi were seen in or around particular cellular components forming different structural patterns around the same cell type, such us bead-like or web-like patterns around erythrocytes. vWF is known to fold or unfold under different conditions of shear stress and can therefore display various properties and functions as a result of exposing certain binding sites for adhesion or uncovering its cleavage site [52,53]. These various structural patterns formed by the thrombi might imply that the local environment within the LA where the thrombus developed altered over time, affecting vWF function and thereby the composition of the thrombus over its development. More research is needed to investigate how and why the feline cardiogenic thrombi showed components of both a traditional red and white thrombus.There are a number of limitations to the study. First, the origin of the detected endocardial vWF protein in the LA samples remains undetermined. We cannot rule out the possibility that the LA endocardial vWF we detected was originally secreted by platelets or endothelia elsewhere in the body and delivered via the circulation. Second, although all the control cats had structurally normal hearts, a number did have other clinical potentially inflammatory conditions that may have affected endothelial vWF expression. However, most systemic inflammatory diseases are associated with increased level of vWF, which is contrary to the finding of low endocardial vWF in the control cats in our study. Third, in the subclinical group, four out of the nine cats had mild LAE. However, when we further sub-divided these cats into (1) subclinical with LAE, and (2) subclinical without LAE, there was no significant difference in endocardial fluorescence intensity between these subgroups and the controls. Fourth, to better evaluate the structure of the thrombi, sliced-though slides with co-immunostaining for fibrin would be helpful to gain greater understanding of thrombus organisation. However, this was beyond the scope of the study and not performed.5. ConclusionsThese results provide evidence for increased endocardial vWF in cats with advanced cardiomyopathy and support a potential role for endocardial vWF as a pro-thrombotic substrate.

animals : an open access journal from mdpi

10.3390/ani12050593

PMC8909620

Cytauxzoon felis was first described more than 40 years ago in the US, later (in the 21st century) similar pathogens were detected in Eurasian and Iberian lynxes, European wildcats and domestic cats in Southern and Central Europe. Our findings have shown the previously unrecorded presence of this parasite in Russia (50 km from Moscow). We described the crucial decrease in the number of leukocytes and erythrocytes, as well as in hemoglobin concentration, throughout the captive serval’s disease, and their increase during the animal’s recovery over six months. Molecular genetic methods allowed us to detect and describe this parasite in four cat species in captivity. The analysis showed high genetic variability and high occurrence of the parasites, which suggests their presence in free-ranging domestic cats and wild felids in Russia.

Over the last two decades, Cytauxzoon spp. has been conquering Eurasia, although this fact has only been brought to light through recent more intensive research after the discovery of C. manul in Pallas’ cat. In Europe, Cytauxzoon was detected mainly in southern countries and later in central Europe. This pathogen has now been found in Russia for the first time (50 km from Moscow), this being the most northern sighting in Eurasia. A captive serval (Leptailurus serval) was found to be infected. Hematological analysis showed a crucial decrease in the number of leukocytes and erythrocytes, as well as in hemoglobin concentration. Genetic analysis confirmed the presence of Cytauxzoon spp. in serval blood at the beginning of the disease period. The identical pathogen was found in one bobcat at the same breeding center. Two other haplotypes of Cytauxzoon spp. were obtained from domestic cats at the same location, identical to the samples from Italy. One new haplotype, which was sequenced for the first time, was found in 7/7 investigated Amur wildcats (100%). The high occurrence and diversity of these pathogens suggest that they are present in free-ranging domestic cats and wild felids in Russia, and may be considered a potential threat to the endangered species. Current research shows that the genetic diversity of this pathogen may be even higher than it was suggested previously. Further genetic research is necessary to describe the diversity and phylogeny of this pathogen in Eurasia.

1. IntroductionThe distribution of pathogens can have a significant effect on host populations, in some cases leading to epizootics [1,2,3] and the death of some individuals or a significant part of the population [4,5]. However, this is just the tip of the iceberg. Most often, pathogens, without leading directly to the death of the host organism, can provoke a decrease/change in the activity of the immune system [6] and physical condition in general [7], as well as leading to problems with reproduction [8]. Among endangered wild felines, the populations of which are relatively small (i.e., Amur tiger (Panthera tigris altaica) or Far-Eastern leopard (P. pardus orientalis)), the death or exclusion of even a few individuals from reproduction may lead to significant negative consequences for the entire population. Although the routine monitoring of the pathogens in endangered felines [9,10,11,12] is well established, the emergence of a new pathogen in a population can lead to significant changes in existing approaches.Cytauxzoon felis is a protozoon of the Piroplasmida order (Theileriidae family), an intracellular blood parasite that affects primarily leukocytes, then erythrocytes. For the first time, Cytauxzoon felis was identified in 1976 in Missouri, USA [13], in the blood of a bobcat (Lynx rufus), which was supposed to be a reservoir of this pathogen. Seropositivity to C. felis in populations of the bobcat is estimated to be 79% [14]. Originally, it was theorized that this parasite occurs only in North America, mainly in the United States. In addition to the bobcat, it was found in the wild in the cougar (Puma concolor) in North America [14]. The bobcats usually remain asymptomatic and serve as reservoirs of the parasite. In contrast, infected domestic cats (Felis catus) usually succumb to the infection within 9 to 15 days. Cytauxzoon felis exists in two distinct forms: an erythrocyte phase (merozoites) and a myeloid cell phase (schizont). It is the schizogonous phase that causes the severe clinical disease in domestic cats. Infected domestic cats show clinical signs, such as pyrexia, anorexia, dehydration, depression, icterus and hepatosplenomegaly. In its normal course, the infection by this parasite leads to severe anemia, a decrease in the number of leukocytes and erythrocytes, a sharp decrease in the activity of the animal, loss of appetite, emaciation and death. Domestic cats usually die within 24 to 48 h upon presentation to veterinarians, in the presence or absence of supportive therapy [15,16,17,18,19,20]. However, recent research has indicated that this scenario can also present the other way around: bobcats may suffer and die from severe acute cytauxzoonosis [21], and in contrast, the domestic cat can recover from acute cytauxzoonosis and carry the parasite for a long period of time [22,23,24]. In this case, the domestic/stray cats may stay as the reservoir of the parasite and can be considered as a potential threat for the other species of felines in the wild and in captivity. The deaths of tigers [25,26] and lions (Panthera leo) [27] from acute cytauxzoonosis in captivity can serve as an example.Cytauxzoonosis (as other piroplasmosis) is a tick-borne disease and some species of ticks play an important role in the circulation of the infection. The natural route of C. felis transmission to intermediate felid hosts occurs via the blood-feeding of infected ticks harboring parasite sporozoites [28,29]. For example, successful Cytauxzoon felis transmission studies have occurred using Amblyomma americanum adults, acquisition-fed as nymphs on an experimentally infected domestic cat, or Dermacentor variabilis adults fed as nymphs on a bobcat [30,31]. In the experimental study, Cytauxzoon felis was successfully transmitted to domestic cats by A. americanum nymphs acquisition-fed as larvae on the donor cat [32].In Eurasia, a similar pathogen was identified in Pallas’ cat (Otocolobus manul) in Mongolia in 2003 [33]. Although it was named Cytauxzoon manul by the authors, there is still no generally accepted view of it as a separate species or as a strain of Cytauxzoon felis. Later this parasite was described in many European countries and this process still continues. The aim of this study is to describe the distribution of Cytauxzoon spp. in Eurasia with the precise analysis of the newly discovered northernmost case of cytauxzoonosis, and its similarities and differences to other cases.2. Materials and Methods2.1. Study Area and ObjectsThe description of the most northern case of cytauxzoonosis was carried out at the Severtsov Institute of Ecology and Evolution biological station in Tchernogolovka (50 km north-east from Moscow), where studies have been carried out for 25 years on the behavior and physiology of Eurasian lynxes (Lynx lynx) [34,35,36,37], Amur wildcats (Prionailurus bengalensis euptilura) [38,39,40] and domestic cats (Felis catus) [41,42]. In addition to these species, the Tchernogolovka station keeps bobcats, and since 2019, they have also kept caracals (Caracal caracal), servals (Leptailurus serval) and one ocelot (Leopardus pardalis). The husbandry conditions of these animals have been described several times before [34,38,41].In May 2019, along with other animals, a 4.5-year-old female serval was transferred to the station. The female was obtained from a private collection and was contained for the entire summer period in the outdoor treeless enclosure of 74 m2, with an adjacent wire-mesh cage of 6 m2. The enclosure and the cage each contained a wooden house measuring 70 × 70 × 110 cm. On 6 October 2019 the female was immobilized and transferred to an indoor complex of winter enclosures. The size of the cage was 2 × 2 × 2.3 (length × width × height) m with a 70 × 100 cm shelf located 40 cm above the floor. A male serval was kept in the neighboring cage, and the winter housing also contained caracals and the ocelot. Each animal had access to the outdoor enclosure (12 m2), which was open when the air temperature was higher than −10 °C. Access to water was ad libitum, and the food ration included 700–900 g of chicken meat daily.Other animals kept at Tchernogolovka were selectively tested for the presence of this parasite in the blood by molecular genetic methods, including 10 Eurasian lynxes (7 adults and 3 cubs at the age of 10 months), 23 adult domestic cats, 7 Amur wildcats and 1 bobcat. Blood was collected during the routine veterinary procedures (e.g., health check, vaccination) and analyzed as described below. From two of these animals, the samples were collected after their death (a 17-year-old bobcat and one lynx cub). In total, we tested 56% of the felids kept at the Tchernogolovka station.2.2. Hematological AnalysisWhen transferring the serval, a blood sample was taken for hematological analysis. Later, this sample (collected on 6 October 2021) was considered as a basal level on the graphs. We collected blood using K3EDTA tubes. The methods of hematological analysis of the entire blood and smears process have been described earlier [40]. In the course of the disease the smears were micropsied with the Leica 5000D (Leica, Wetzlar, Germany) using 1000× magnification (100 × 10). Atypical insertions were detected and counted in blood cells and later the presence of parasites was confirmed by DNA analysis. The number of parasites per 100 red blood cells (RBC) was counted on the smears and recalculated on 1 mL of blood (based on the total number of RBC obtained from the hemoanalyzer).2.3. Molecular Genetic AnalysisWe stored samples (whole blood) for the genetics at −20 °C, and then performed DNA extraction with the InviMag Blood DNA Mini Kit/KF96 (Invitek Molecular, Berlin, Germany) on KingFisher Flex workstations (Thermo Fisher Scientific, Waltham, MA, USA). To detect Cytauxzoon spp. infection, we performed a PCR with the 5× master mix (Dialat, Moscow, Russia) and the primers 5′-TGAACGTATTAGACACACCACCT-3′ and 5′-TCCTCCCGCTTCACTCGCCG-3′ on the second internal transcribed spacer (ITS-2) of ribosomal DNA, according to Brown et al. [43]. The PCR fragments of required length were purified by excision from the 3% agarose gel using the Cleanup mini kit (Evrogen, Moscow, Russia). DNA sequencing was performed with forward and reverse primers using Big Dye 1.1 kits on an ABI 3130 genetic analyzer (Applied Biosystems, Waltham, MA, USA) in a POP7 polymer. Nucleotide sequences with a length of 232 bp were obtained from 14 animals and analyzed. Sequences are registered in NCBI (accession numbers: MZ242218–MZ242221). We compared these sequences with those of other similar Cytauxzoon sp. isolates available in NCBI. The percentage of identities of the ITS-2 sequences obtained in this study was compared with published sequences calculated in the Mega X program using the p-distance method. The phylogenetic trees were constructed in MEGA X by the neighbor-joining method (NJ) using the Tamura-Nei model [44].3. Results3.1. First Detection of Cytauxzoon spp. in RussiaBy mid-November (16th, zero point), the serval’s behavior in captivity changed; she became inactive, moved slowly and began to regularly refuse to eat. Anesthesia of the animal and clinical examination was carried out 11 days after the zero point. Hematological analysis showed a sharp decrease in the number of white blood cells (WBC). The total number of leukocytes was almost five times less than in October at 2.2 × 109 vs. 10.6 × 109 cells/L (Figure 1). The total number of WBC decreased mainly because of the decrease in the number of neutrophils (almost 9 times less) (Figure 1) and monocytes (more than 9 times less) (Figure 2). The number of lymphocytes remained practically unchanged (a 20% decrease) (Figure 2). In the same period, the indicators characterizing the ability of blood to carry out oxygen transfer (the total number of red blood cells (RBC), hemoglobin content and hematocrit) almost halved (Figure 3). In the blood of the animal, the concentration of detected parasites reached 1.70/100 RBC or 65.5 × 109 cells /L (Figure 4).The first treatment was started in early December (4 December 2019), seven days after the first detection of the parasite. The animal received 70 mg of azithromycin dihydrate (Sumamed®, Pliva Hrvatska, Zagreb, Croatia) with food once per day for ten days, and Atovaquone 125 mg + Proguanil 50 mg (Malarone, GlaxoSmithKline, Brentford, UK) three times per day (every 8 h) for ten days. The course of treatment was completed on 13 December 2019. Blood sampling a week later (35 days after zero point) showed that the total concentration of WBC, as well as the concentration of all their forms (neutrophils, lymphocytes and monocytes), increased during the treatment period (leukocytes by 2.5 times, individual forms by 1.6–3.5 times). At the same time, the indicators for RBC, hemoglobin and hematocrit decreased by 15–28% compared to the previous sampling period. The concentration of parasites in the blood increased significantly and amounted to 7.54/100 erythrocytes or 207.3 × 109 cells/L. The next course of treatment included Imidocarb dipropionate 40 mg (Pyro-Stop, Apicenna, Balashiha, Russia) and erythropoietin 500 mg (Erythropoietin, MTH, Moscow, Russia). The course of treatment was started on 30 December 2019 (44 days after the beginning of disease). Erythropoietin injections were conducted every three days (5 times). Imidocarb dipropionate was used twice with a one-week interval. One week after treatment (66th day) a blood sample was taken and showed a significant decrease (more than 5 fold) in the concentration of the parasite in the blood to 1.37/100 RBC. The serval began to eat actively.Blood samples were collected one week after this treatment, on the 64th day. Due to the problem with the hemoanalyzer, we did not estimate the number of white and red blood cells; however, we were able to show a sharp decrease in the parasites’ number (Figure 4). The next blood sampling was performed on the 105th day. Body mass increased to 8.2 kg (13%). An increase in the number of WBC, RBC and hemoglobin by 20–50% was noted. The concentration of parasites was 1.28/100 RBC or 60.3 × 109 cells/L. It turned out that the number of erythrocytes, as well as the concentration of hematocrit/hemoglobin, reached maximum levels in the animal and did not change significantly over the next five months. Although the number of WBC had increased, their concentration was significantly lower than that in autumn in a healthy animal, accounting for 60% of the total number of WBC (from 45% for monocytes to 97% for lymphocytes). However, the animal was considered rather healthy and was paired with a male two days later (on 107th day), the mating took place on 110th day.The following samples were collected from a pregnant female on 148th day, 38 days after mating. By that point, the concentration of parasites decreased threefold in comparison to the day of the first detection. The number of RBC and hemoglobin and hematocrit levels did not change in the animal when compared to the previous sampling; however, the number of WBC increased significantly (9.34 × 109 cells/L—almost reaching the level before the disease), primarily due to neutrophils (increase of 1.5 times) and monocytes (2 times). The female gave birth to a stillborn kitten 80 days after mating. The data were obtained from a video camera and the female ate the kitten several hours after birth.The next routine examination of the animal was carried out 245 days after zero point (18 July 2020) and no parasites were detected in the animal’s blood. During this period, there was a slight decrease in the total number of WBC and neutrophils, but the number of lymphocytes and monocytes increased slightly instead. The number of red blood cells and the level of hemoglobin remained stable, while the level of hematocrit in the animal increased by about 27%.3.2. An Occurrence of the Parasite in Potential Hosts and Its Genetic DifferencesGenetic analysis conducted with the described primers confirmed the presence of Cytauxzoon spp. in the blood of the serval (the obtained sequence was loaded to NCBI (accession number MZ242219)) (Table 1). The presence of the Cytauxzoon spp. identical to that in the serval was detected in one animal—the bobcat, which died at the age of 17 years. Thus, the presence of the pathogen of the same haplotype was confirmed in two animals (serval female and bobcat female) kept at the Tchenogolovka station. Three other haplotypes of Cytauxzoon spp. were obtained through sequencing of other samples. One of them, NCBI accession number MZ242218, was discovered by us for the first time (Figure 5). It was typical for Amur wildcats and was found in 100% of animals (7 of 7 tested individuals). One of them arrived from the Russian Far East two years before sampling, six others were born at the Tchernogolovka station. Two other haplotypes were obtained from domestic cats—NCBI accession number: MZ242220 from three cats and NCBI accession number: MZ242221 from two cats (respectively, 13.0% and 8.7% of the 23 tested cats). The parasite was not detected in all tested blood samples of Eurasian lynxes. All sequences analyzed in our work are unique or rare and were described only in a small number of animals with insertions of 198 bp long in the ITS-2 region, for example, NCBI accession numbers: HQ 383877 and JF 330260 from bobcats from Pennsylvania and Kentucky, USA [45]. The percentage of identities of the ITS-2 sequences obtained in this study compared with published sequences ranged from 0.00% to 94.67%. The strongest differences were found between the MZ242218 and HQ383908 sequences (Figure 5).4. Discussion4.1. The Current Data on Cytauxzoon spp. Distribution and HostsFor the first time, Cytauxzoon spp. in Eurasia were described by US veterinarians [33]. Four Pallas’ cats were trapped in Mongolia and directly transported to Oklahoma (USA). All were examined within 10 days of their arrival. At the time of arrival, parasitemias were below 1.0%. The significance of this parasite to the health of free-ranging Pallas’ cats in Mongolia was not established. Later in Europe, this pathogen (Cytauxzoon spp.) was detected in domestic cats in Spain [39]. Molecular data analysis indicated that Spanish Cytauxzoon felis (cat isolate) were 98% identical to Cytauxzoon spp. from Mongolia [33] and 95% identical to African Cytauxzoon felis. Five years later Cytauxzoon spp. were detected in France in one domestic cat out of 116 tested [46,47]. The first clinical case report of Cytauxzoon spp. infection in a domestic cat in France was described 8 years later [48]. The animal received treatment, showed remission and recovered but one month later was brought to the vet clinic with the Cytauxzoon spp. infection, confirmed by PCR and smear tests. By this time the similar strain (or species) was found to be quite common in Italian domestic cats [49]. Cytauxzoon spp. infection was detected by 18S rRNA gene PCR in 23% and by blood smear examination in 15% of 118 tested domestic cats. The 18S rRNA gene sequences obtained were 99% identical to the Cytauxzoon spp. from Spanish, French and Mongolian wild and domestic cats [49], and 93% to Cytauxzoon felis [50].In the last five years, the findings of this parasite in domestic cats have been reported in Portugal [51] and Switzerland [52]. In Portugal, there was an occurrence of a clinical case with a lethal outcome, despite the intensive care and treatment [51]. The parasite was genetically identical to the samples described in Spain and France [44,45]. In Switzerland this parasite was obtained from three kittens in the same litter and the transmission of the parasite with blood transfusion was confirmed for the first time [52].Moreover, during the last 20 years Cytauxzoon spp. has been found in wild felines as well—in Eurasian lynx in Romania [53], in the Iberian lynxes (Lynx pardina) in Spain [54,55,56], and in the European forest cat (Felis silvestris) in Italy [57] and in Romania [53]. In the Iberian lynx, seropositivity to Cytauxzoon spp. was estimated at 15% [55], and the parasite was found in all four tested Eurasian lynxes [46]. In European wildcats, the serum prevalence to this parasite was 50% [53]. The distribution of the pathogen is described primarily in the eastern and southern countries of Europe, as well as in Mongolia and China [58]. Based on the data on Cytauxzoon felis and all other piroplasms, this parasite penetrates the cat organism (including domestic cats) only through a tick bite [24]. This suggests that infection can occur if animals are in close proximity to each other. The second option is that animals (carrier and infected) should inhabit approximately the same place, albeit at different periods of time, determined by the lifespan of the ticks that received the parasite when biting the infected animal. The distances that ticks may cross are extremely small compared to the distances of the host animals. The home range size of domestic cats is up to 2.94 km2 [59], of bobcats, 20.2 km2 [60] and of the Eurasian lynx (Lynx lynx), 1500 km2 [61]. Apparently, adult and dispersing felines are the main source of the spread of this pathogen in the wild. However, all these data and suggestions are obtained based on North American parasites. The way of transmission of European Cytauxzoon spp. has not been described so far. Obtaining these data is a crucial goal for the understanding of the life cycle of these parasites in Europe.In Russia, this pathogen has never been reported before, neither in the south of the European part (Caucasus, inhabited by the Central Asian leopard (Panthera pardus saxicolor) as well as Eurasian lynx, European wildcat, domestic cat), nor in the border areas with Mongolia (snow leopard (Panthera uncia), Pallas’ cat, domestic cat).4.2. The First Detected Case in RussiaWe assume that the infection of the female serval occurred at the Tchernogolovka station. The first blood sampling, carried out in October 2019, did not reveal the presence of the parasite in this female (neither by microscopy of the blood smears nor by PCR). However, already in November we noted a sharp deterioration in blood parameters. It is difficult to assess what was the source of the pathogen. The design of the open-air cage complex practically excludes the contact of animals with stray cats. At the same time, the North American species (Cytauxzoon felis) is carried mainly by ticks, which can travel distances of tens of meters. The bobcat, a confirmed host of the identical parasite, was kept 70 m from the serval enclosure, which was not an insurmountable obstacle for the tick. In addition, the transfer of ticks on the keepers’ clothes cannot be ruled out. However, in this case the disease manifested itself approximately a month after the bite, which is not typical for most cases of the Cytauxzoon spp. infection, which usually develops much faster [15,20]. An alternative is the possibility of infection with the pathogen from animals transferred with the serval to the winter enclosure (an ocelot, a serval male and six caracals). Although genetic analysis of the blood of these animals for the presence of the pathogen was not carried out, the analysis of blood smears and the observation of the physical condition of the animals did not reveal signs of infection. Thus, there are two unproven hypotheses of serval infection—either a long (about a month) incubation period of the disease, or the latent course of the infection in a number of felines. From another point of view, as we know, the transmission routes, including the vector tick, as well as the clinical signs, are not well resolved for the European Cytauxzoon spp. Thus far, it is difficult to suggest another way of pathogen transmission because all infections by piroplasms are related to ticks. However, further studies are necessary in order for us to understand the transmission routes for the European species of Cytauxzoon.This is the first known case of the discovery of Cytauxzoon spp. in a serval, which does not allow an effective assessment of the typical course of the disease. Its course and mortality differ significantly in domestic cats and lynxes/bobcats. The latter, apparently, may be asymptomatic, the animals acting as reservoirs of the pathogen [14], but see [24]. Despite the described isolated cases of disease and death in other wild feline species [14,25,26], it is not possible to assess the danger of the pathogen for these species and the typical course of the disease. The two forms/phases were described for Cytauxzoon felis: the erythrocyte phase (merozoites) in blood and the myeloid cell (schizogonus) phase. The latter causes the most severe consequences for its hosts. Thus far, the schizogenous phase has not yet been identified in European Cytauxzoon spp. The serval is a species that is evolutionarily closer to the lynx than to the domestic cat [62], which theoretically may affect the course of the disease. However, in this clinical case, the animal suffered from this disease severely. At the same time, despite the two types of treatment, parasites in the serval’s blood were recorded for at least five months. Nothing is known about the effect of this pathogen on pregnancy, however, the female serval gave birth to at least one stillborn kitten. By the middle of pregnancy, the concentration of WBC in the serval reached normal, which can partly be explained by the stimulation of the female’s immunity by embryos during this period [42].In addition, this is the first confirmed case of detection of this pathogen in Russia. Based on the confirmed distribution of this pathogen outside Russia, it can be expected to occur in domestic and wild cats in the Caucasus region and in Transbaikalia (along the border with Mongolia), however, the parasite was identified in the Moscow region. One study [63] mentions that Cytauxzoon was discovered by PCR in one Amur wildcat and in one Asiatic badger (Meles leucurus) in the Russian Far East, but it did not describe the year, precise region and methods of the study.4.3. The Diversity of Cytauxzoon spp. in the Animals at the Biological Station TchernogolovkaThe Tchernogolovka biological station has, for many years, supported a large colony of Eurasian lynxes that can act as carriers and reservoirs of the pathogen, however, the presence of the pathogen was not detected in any of these animals. The only animal of those tested that was positive for the pathogen, was the old bobcat obtained in 2004 from Moscow Zoo. Surprisingly, the haplotype of the parasite (NCBI accession number: MZ242219) was similar to that which was described for the domestic cats in China [58].We also found three more haplotypes (or possibly species) of Cytauxzoon spp. in cats at the Tchernogolovka station. Two sequences of Cytauxzoon spp. haplotypes, obtained from domestic cats (NCBI accession numbers: MZ242220 and MZ242221), were identical to ITS-2 sequences available in NCBI (accession numbers: MN513351, MN513350 and MN513349), which were obtained in Italy. Although the authors did not present the link with the publication, we assume that they were obtained from domestic cats [54,55] and European wildcats [62]. Both of these haplotypes were found at the Tchernogolovka station in domestic cats. Another new haplotype was found in Amur wildcats. The Amur wildcats were kept far away from the domestic cats (about 200 m) for several years, which likely prevented the transfer of parasites between these two colonies. Alternatively, there is supposed species-specificity of host–parasite relations in Cytauxzoon. The percentage of infected animals was higher in the Amur wildcat than in the other two species (domestic cat and Eurasian lynx).During the last two decades, Cytauxzoon spp. has been found in Eurasia and a number of studies have described its presence in many European countries, Mongolia and China [28,51,52,53,54,55,56,57,58]. It was shown that this parasite is different from the North American Cytauxzoon felis and more similar to the Cytauxzoon manul. These findings belong mainly to the southern latitudes (Mongolia, China, Southern Europe). Only in the last few years has this parasite been found in central Europe (Switzerland and Germany [52,64]). However, the recording of Cytauxzoon spp. in the surroundings of Moscow is the most northern finding of the pathogen in Eurasia. This record was obtained in captivity, however, the infection of the serval suggests tick transfer (like in other piroplasms) of the pathogen at this latitude as well. Moreover, the wide distribution of ticks in northern areas of Russia and the wide distribution of Eurasian lynxes and stray cats, suggest that Cytauxzoon spp. should occur at higher latitudes as well. It is probable that the lack of information here can be explained by the absence of studies in these areas.The high diversity of this pathogen was detected at the Tchernogolovka station, as well as its occurrence in different cat species (for example, one haplotype was obtained only from an Amur wildcat, but 100% of animals had this parasite). This partial exclusiveness (host–species specificity, but the same haplotype was detected in bobcats and servals) led us to the theory regarding the high diversity of these haplotypes/species and the difficulties in interspecies pathogen transfer. Gallusova et al. [53] suggested this idea several years ago. A more detailed study of the health status of wild and stray cats over large areas in Russia will definitely allow us to describe the genetic diversity of this pathogen. A recent study [64] suggested that Cytauxzoon in Europe includes at least three different species. One of them (C. europaeus) was described both in the European wildcat and Eurasian lynxes, but two others were collected only from the European wildcats (C. otrantorum and C. banethi). One of them was found only in Romania [64]. Although, we used other genetic primers in our project, which makes our data incompatible with this study. However, the high occurrence and high haplotype diversity of Cytauxzoon spp. at the Tchernogolovka station suggest that this pathogen may occur in wild and domestic felids in Russia. Further studies of the pathogen and its genetic diversity are necessary to assess its distribution in Russia and the degree of threat to cats, both in various breeding centers and in the wild, primarily in the southern regions of the country, in order to identify the species of these pathogens and their pathogenicity to domestic cats and wild felids.The latest findings of the Cytauxzoon spp. in Europe may give an impression that the parasite has dispersed intensively in the European region in the last two decades. Following the order of publication, the pathogen seems to have spread to north-eastern Europe (from Spain [54] and Italy [49] to Germany [52] and Romania [64]), and has now been discovered in Russia, near Moscow. The parasite appears to be conquering the continent. However, increasing the intensity of the studies of this pathogen after the first finding in Europe may lend this explanation more credence. The last findings of the pathogen in the samples of European wildcats collected at the end of the 20th century confirm the presence of this parasite in France during that period [65]. Cytauxzoon spp. are widely distributed in Europe, but our knowledge about it is relatively scarce in contrast to C. felis. The phylogeny of European Cytauxzoon spp. [64], its distribution, the vectors of the pathogen transmission and the details of its life cycle (the existence of the schizont stage) create a necessity for further investigation of this parasite over the whole range.5. ConclusionsThe latest studies indicated the presence of Cytauxzoon spp. in Eastern and Southern Europe. This study described the first case of infection by this parasite at the biological station in Russia in a captive serval (Leptailurus serval), the most northern case of pathogen detection. It also showed a clinical picture of the animal recovering from the disease. Hematological analysis showed a crucial decrease in the number of leukocytes and erythrocytes, as well as in hemoglobin concentration, which recovered only six months later. The peak of parasite concentration was observed two and a half months after infection. Genetic analysis confirmed the presence of the identical strain of Cytauxzoon spp. in one bobcat at the same breeding center. Three more strains of Cytauxzoon spp. were obtained there from domestic cats and Amur wildcats. Further genetic analysis is necessary to validate these strains with the new classification of Cytauxzoon spp. proposed by Panait with co-authors [64]. The high occurrence and diversity of these pathogens in captivity suggest that they are present in free-ranging domestic cats and wild felids in Russia, and may be considered a potential threat to the endangered felid species.

animals : an open access journal from mdpi

10.3390/ani11072028

PMC8300141

Poultry products are popular meat products in the United States for both retail and food service sectors. Food service operators typically obtain food products in bulk as they utilize products quickly and at a high volume. Typically, chicken arrives to food service operators frozen in bulk packaging and is thawed or slacked by storing it in refrigerated temperatures (2 °C to 4 °C) to be used over several days while maintaining its acceptability for cooking and serving to consumers. Ensuring a product is safe to consume is the most important factor in the food industry. This study measured the microbial growth on marinated chicken tenderloins that were aged after slaughter, bulk-packaged, frozen, then slacked for 132 h. At no time during the slacking period did any samples reach the limit (6 log) of unsafe microbial growth. Psychotropic bacteria grew at each sampling time and the tenderloins aged for 4 and 5 days post-slaughter surpassed all other treatments. As no samples surpassed the spoilage threshold, it is suggested that slacking is a safe method of thawing chicken tenderloins for up to 8 days post-slaughter.

The objective of this study was to validate the shelf-life of marinated and frozen chicken tenderloins. Treatments were randomly assigned to the age of the tenderloins post-harvest, days aged (DA): DA4, DA5, DA6, DA7, and DA8. Microbial analyses were used to analyze the growth of aerobic, psychotropic, and lactobacilli bacteria to assess the shelf-life of bulk-packaged chicken tenderloins. Tenderloins were sampled fresh, then vacuum tumbled in a marinade. After marination, the tenderloins were sampled with the remaining tenderloins packaged and frozen (−25 °C). After freezing the chicken tenderloins were slacked in a refrigerated cooler (2.2 °C) for up to 132 h (h) and sampled at 36 h, then every 24 h following. After marination, each treatment significantly (p < 0.05) decreased in aerobic and psychotropic counts except DA4. During slacking, no treatment crossed the threshold of 106 CFU/mL (Log 6) set for this study. Though none crossed the threshold, treatments DA4, DA5, and DA6 had significant (p < 0.05) increases in aerobic bacteria after 7 days of age. The psychotropic bacteria continuously grew at each sampling period, with DA4 and DA5 surpassing the other treatments (p < 0.05) at 108 h and 132 h reaching 105 CFU/mL. Every treatment remained below the spoilage threshold, suggesting that this method of storage is suitable for chicken tenderloin shelf-life.

1. IntroductionThe consumption of chicken products in the United States has dramatically increased to approximately 240 kg per person since 1959. With such stark increases in consumption, it has been reported that over 42% of chicken consumed is sold through a food service outlet [1], where many steps occur to create this product for consumers. Moreover, it has been reported [2,3] and is widely known that excessive handling of meat products by consumers and food service providers, specifically poultry, through the processing and manufacturing stages results in a greater likelihood that microbial loading can occur [4,5]. The testing of raw and finished poultry products for initial microbial load of a portioned product from start to finish can aid in developing prediction models for meat manufacturing companies. Research has concluded that microbial growth (106 to 107 CFU/g) can create signs of spoilage organism development, including but not limited to off odors and slime production [6,7]. Aerobic, anaerobic, and psychotropic microflora are common reasons for a meat product to spoil [2]. A common spoilage microorganism in the meat industry, Pseudomonas, is a gram-negative psychotropic bacteria which has been reported to grow in cold storage environments, including in coolers or refrigerators [8]. Additionally, gram-negative bacteria require higher water activity for growth survival, which can be found in poultry products [4]. Lactic acid bacteria (LAB) is another common category of bacteria which can be identified on food products including meat and poultry. By using microbial sampling methods, the deterioration of poultry meat products can be conducted to assess raw materials and the packaging methods used in distribution channels. Flat packaging has been utilized in the food service industry due to its speed, ease, and ability to control package volume. In the use of flat packaging within the poultry industry, the product is typically frozen then thawed in refrigerated temperatures (2 to 4 °C) for several days, often referred to as slacking. Additionally, the majority of the poultry processed in the United States is marinated in a solution containing water, salt, and phosphate to increase water holding capacity to promote a juicier product [6]. This is achieved by injecting the solution and possibly tumbling a product under vacuum, which aids in the distribution of the solution through the product and allows for protein extraction [6]. Thus, the objective of this study was to validate the shelf-life of vacuum-tumbled marinated chicken tenderloins after frozen storage and slacking through the analysis of microbial growth.2. Materials and Methods2.1. Raw MaterialsFresh chicken tenderloins were obtained 48-h post-mortem, hand trimmed, weighed, and sorted from a commercial processing facility (Foundation Food Group; Gainesville, GA, USA). After portioning, a total of 680.4 kg of chicken tenderloins (226.8 kg per replication × 3 replications) were placed inside a plastic liner and then divided into 150 qt insulated coolers (Igloo, 105.75 cm × 47.48 cm × 51.44 cm, Katy, TX, USA). Ice was placed in the cooler prior to the addition of the tenderloins, with a plastic liner used to ensure no moisture migrated from the ice to tenderloins. A ThermaData series II Temp Logger T2C (2 Ext. Removable Probes, American Fork, UT, USA) temperature data logger was inserted into each cooler in two different adjacent locations to monitor product temperature during transportation. Chicken tenderloins were transported to the Lambert-Powell Meat Laboratory at Auburn University. Upon arrival, tenderloins were allocated randomly to 3 replications (226.8 kg per replication) and kept in dark storage for 48 h at 2 °C.2.2. Treatment Allocation, Marination and PackagingChicken tenderloins (226.8 kg) were weighed and randomly assigned to 1 of 5 treatments (45.36 kg) based on days aged (DA) post-harvest: DA4, DA5, DA6, DA7, or DA8. Each treatment had 2.27 kg removed for microbial analysis after an initial assignment. On each of the corresponding days post-harvest (4, 5, 6, 7, or 8), each treatment group (43.09 kg) was subjected to marination by vacuum tumbling in a proprietary blend (1.64 kg) including: water, salt, modified corn starch, and monosodium glutamate (MSG) for 6 min at 4 rpm. After marination, 2.27 kg of tenderloins were again removed for microbial analysis. The remainder of the chicken tenderloins in each treatment group (40.82 kg) were packaged into blue plastic bags (C and E Supply LLC, 13 × 20 + 1.5″ LIP Blue Bag) (2.27 kg/bag) and pressed flat by hand. Flat packing was achieved by pushing any remaining air out of the bag and folding the top flap of the bag over with no actual seal. This entire process was performed with 3 replications.2.3. Product Storage and SlackingEach blue bag was placed into a blast freezer (−25 °C) where they remained for 8 d until their treatment slacking processes began. After 8 d of storage in the blast freezer, each treatment was removed and placed into a walk-in cooler (4 °C ± 2 °C) for 36 h to begin the slacking process. After 36 h, three (n = 3) blue bags were removed and used for analysis, then every 24 h following (up to 132 h) 3 blue bags were removed for analysis, to simulate the food service operator thawing specifications. When blue bags were removed from the cooler they were subjected to microbial analysis.2.4. Microbial AnalysisChicken tenderloins (n = 2) were aseptically removed from each blue bag immediately after removal from the refrigerated cooler using a modified procedure of the American Public Health Association [9]. Tenderloins were placed into a Nasco Whirl-Pak filter bag (55 Oz. Filter Bag 7.5″ × 12″, Nasco, Fort Atkinson, WI, USA) with 50 mL of phosphate buffered salt (PBS) and hand massaged for 1 min. Following massaging, 1 mL of the PBS solution was extracted from the filter bag with a serological pipette tip and placed into a dilution tube containing 9 mL of PBS to create serial dilutions. Each dilution was mixed using an analog vortex mixer (VWR International, Radnor, PA, USA). After serial dilution, 100 μL was extracted from each dilution tube and the filter bag and placed onto 3 media types: aerobic plate (Difco™ Plate Count Agar; Becton, Dickinson, and Company; Tempe, AZ, USA); psychotropic plate (Difco™ Plate Count Agar, Becton, Dickinson, and Company; Tempe, AZ, USA); and de Man Rogosa and Sharpe agar (Difco™ Lactobacilli MRS Agar; Becton, Dickinson, and Company, Tempe, AZ, USA). Aerobic and psychotropic plates were incubated in a Jeio Tech, Inc. incubator (Model IB-05G, Jeio Tech Inc., Woburn, MA, USA), with aerobic plates incubated at 37 °C for one day and psychotropic plates incubated at 8 °C for 7 days. MRS plates were incubated at 37 °C for two days in anaerobic chambers (MGC AnaeroPack® System, Rectangular Jar 7.0 L; Mitsubishi Gas Chemical Co., Inc., Chiyoda, Tokyo, Japan) which contained two oxygen scavenger packs (GasPak™ EZ; Becton, Dickinson and Company, Sparks, MD, USA) to reduce free oxygen. Each sample had 3 serial dilutions plated per plate type for a total of 9 plates per sample. After the incubation period, plate colonies were counted on a Reichert Quebec Darkfield Colony Counter (Depew, NY, USA) and recorded. The best plate was taken from each sample and then converted to CFU per mL of rinsate.2.5. pH AnalysisOn each day of sampling, pH was measured using a probe style-pH meter (H170 Hach pH meter, Hach, Loveland, CO, USA), calibrated each day using a standard 4.0 and 7.0 buffer solution prior to collecting microbial and instrumental color samples. Chicken tenders (n = 2) were removed from their respective package and a stainless-steel probe was inserted into the geometric center of each tender. The average reading of three readings from each tender within each treatment across all sampling days was recorded.2.6. Instrumental Surface Color AnalysisFresh surface color was measured using the Commission International de’Eclairage (CIE) spectrum for lightness (L*) using a HunterMiniscan XE Plus (MSXP-4500C; Hunter Laboratories, Reston, VA, USA). Color measurements were measured in duplicate on the surface of the chicken tenders and the mean value was recorded for each tenderloin.2.7. Statistical AnalysisFor this study, 3 replications were conducted with each microbial plate type being plated 3 times per replication. Microbial data was converted to log10 CFU/mL rinsate prior to statistical analysis. Media type was considered a repeated measure and days aged and slack time were fixed effects. The data were analyzed using a PROC GLIMMIX of SAS 9.4 (SAS Institute, Inc., Cary, NC, USA) and LS Means were separated using the Tukey–Kramer adjustment with α = 0.05.3. Results3.1. Aerobic Microbial Analysis ResultsResults from the microbial analysis utilizing aerobic plates are presented in Table 1. Fresh tenderloins from DA4 were the only treatment to significantly (p < 0.05) decrease in aerobic plate count (APC) values by 0 h, however those values increased by 36 h and were similar to fresh tenderloin values. Chicken tenderloins from DA4, DA5, and DA6 when sampled fresh had significantly (p < 0.05) lower APC values than DA7 and DA8. Although DA7 and DA8 both started with values significantly greater than all other treatments, they were the only treatments with significantly (p < 0.05) lower APC values at 132 h than their fresh APC values. After marination, all treatments had numerically lower APC values, however only DA4 had significantly (p < 0.05) lower values at 0 h. By 36 h slack time, DA4 and DA5 had significantly (p < 0.05) lower APC values than all other treatments, however both treatments significantly (p < 0.05) increased from 36 h to 60 h. By 132 h slack time, DA4 had the highest APC values significantly (p < 0.05) greater than all treatments other than DA7, however all other treatment were not statistically (p > 0.05) different from each other at that time.3.2. Lactic Acid Bacteria Microbial AnalysisResults from lactic acid bacteria (LAB) microbial analysis are presented in Table 2. Tenderloins from DA4, DA5, and DA6 did not significantly (p > 0.05) differ for LAB values however, LAB values for these days were all significantly (p < 0.05) less than LAB values from fresh DA7 and DA8 tenderloins. From 60 h to 84 h slack time, LAB values from DA4, DA6 and DA7 all significantly (p < 0.05) increased, where DA5 did not change and DA8 values were significantly (p < 0.05) less. Similar to APC, DA4 tenderloins significantly (p < 0.05) increased in LAB values from fresh analysis to 132 h of slack time and was significantly (p < 0.05) greater than all other treatments at the end of the slacking period other than DA7 tenderloins, which started with the greatest APC values. Interestingly, DA5 and DA6 LAB values at 132 h slack time were significantly (p < 0.05) less than any other treatments at that time. Treatments DA4 and DA6 did not reach 101 CFU/mL until 36 h slack time, where DA8 values were over 3 times higher. Overall, slack time nor days aged did not affect the development of lactic acid bacteria as no trend was found.3.3. Pyschotropic Microbial AnalysisResults from the psychotropic plate count (PPC) microbial analysis are presented in Table 3. For this analysis no treatment reached 106 CFU/mL, which was determined at the start of the study as the threshold for non-consumable products. At 36 h, DA4 tenderloins had significantly (p < 0.05) lower values than all other treatments at that time, however, no other treatments were different from each other. By 60 h slack time, no matter the treatment, there were no significant (p > 0.05) differences in PPC values throughout all treatments. At 84 h, DA5 was not significantly (p > 0.05) different from the values of DA6 tenderloins, however DA4 and DA5 values were both significantly (p < 0.05) greater than the tenderloins from DA7 and DA8, but not from each other. This trend continued into sampling at 108 h and 132 h, as DA4 and DA5 PPC values were significantly (p < 0.05) greater than any other treatments at those times. At 132 h both DA4 and DA5 treatments approached 106 CFU/mL.3.4. Storage Effect on pHThe storage time and age of chicken tenderloins following harvest greatly (p < 0.05) influenced the meat pH (Table 4). The meat pH was greatest initially (p < 0.05) on DA4, and the lowest (p < 0.05) as the duration of storage in refrigerated conditions extended through DA7 and DA8. 3.5. Instrumental Surface ColorSurface color (L*) values for chicken tenderloins during the simulated shelf-life period are presented in Table 5. There was an interactive effect (p < 0.05) of age post-harvest and slack time on the surface color. Tenderloins were darkest (p < 0.05) on DA5 and lightest (p < 0.05) on DA8 of the simulated shelf-life period. As the duration of the storage period increased, the surface color of the chicken tenderloins became lighter (p < 0.05). 4. DiscussionAerobic and psychotrophic (except for DA4) microbial loads for the chicken tenderloins declined following marination with a water, salt, modified food starch, and monosodium glutamate. It is plausible that the addition of the marinade resulted in a lowering of the surface pH of the chicken tenderloins causing microorganism growth to be limited. Previous research comparing the microbial loads of raw chicken breasts to salted chicken breasts reported that the salted chicken breasts had reduced quantities of bacteria [6]. Additionally, the addition of the wet marinade could have had a diluting effect on the microbial load. Immediately after marinating, the chicken tenderloins were packaged and placed into a blast freezer (−24 °C). It has also been discussed that storage temperatures can alter bacterium growth [10]. The extremely cold storage temperatures can cause injury to bacterial cells which would then require favorable storage conditions (temperature) for repair [11].Psychotrophs had the greatest microbial level throughout this study, likely due to the colder storage temperatures maintained throughout the storage periods. It appears that as the chicken tenderloins thawed, the psychotrophs were first to reach a favorable temperature for repair and grow much more rapidly than lactic acid or aerobic microbial organisms. The psychotropic bacteria tend to grow faster in colder temperatures (0 to 20 °C), while mesophilic bacteria grow faster in warmer temperatures (20 to 45 °C). A shelf-life experiment conducted compared how different temperatures (2 to 20 °C) affected the growth of psychotropic and mesophilic aerobic bacteria on portions of chicken breast [12]. The psychotropic microbial load was greater than the aerobic bacteria at the initial and last sampling periods, suggesting that colder temperatures during storage slow but do not eliminate microbial growth throughout shelf-life periods [12].It has been described that at low refrigeration temperatures, psychotropic bacteria can dominate the competition and that mesophiles may survive the cold conditions but not grow [6]. Several studies have explained that the greatest influencer of microorganism growth is microbial competition [12,13]. Pseudomonas spp. is a very common specific spoilage organism that is an aerobic psychotroph. This bacterium is a common culprit in the poultry spoilage realm and the environment of this study favors bacteria like Pseudomonads and other aerobic pyschotrophs [8]. Rancid odors and slime production are signs of spoilage that are attributed to bacteria like Pseudomonads and can often be detected through sensory analytical methods [14]. According to a previous study, lactic acid bacteria did not create off odors equal to or greater than psychotropic bacteria, indicating that that sensory qualities are an important aspect of spoilage [4]. Moreover, when microbial counts exceed 108 CFU/g, the production of slime associated with off odors is likely a contributing factor to the decomposition of muscle tissue [15].The changes in postmortem muscle pH during the simulated storage period tend to agree with previous studies suggesting that changes in microflora, packaging atmosphere, and temperatures during storage period can impart changes to muscle pH [16,17,18]. Moreover, pH has been linked to influencing the many changes that occur in poultry meat quality which can include color, water-holding capacity, tenderness, and juiciness. In addition, surface color has been linked to the muscle pH of breast meat, with darker meat recording lower pH values [19]. The results within the current study agree with previous studies that higher pH values can produce a lighter surface color and lower pH values are associated with darker surface colors of chicken meat [20].Surface color of meat is an important factor used to evaluate freshness or wholesomeness at the time of cooking. The surface color changes in lightness (L*) values that occurred during the storage period agree with previous studies [19,20]. Several factors have been reported to influence the surface color of chicken meat, including gender, age, and freezing conditions. In addition, broilers are harvested at younger ages resulting in less total myoglobin in the muscle tissue and a subsequent lighter surface color [21]. The surface color changes reported within the current study agree with previous studies where pH and storage temperature can impart light scattering as a function of protein denaturation [22].5. ConclusionsIn conclusion, spoilage organisms never achieved a spoilage threshold for this study of 6 log CFU/mL. The limited microbial growth and changes to surface color that occurred suggest that these findings could be utilized by the food service industry for chicken products that have been aged for longer periods post-slaughter prior to freezing. These results could bolster research used within the poultry industry in an effort to reduce waste/losses within the food service sectors, as products could be held under frozen, refrigerated, or even slacked conditions for a greater period of time without causing detrimental impacts to the freshness and wholesomeness of the chicken tenderloins. However, additional efforts are needed to identify methods of packaging and storage times that could further improve the quality attributes of taste and surface color of poultry products stored in refrigerated temperatures for extended periods.

animals : an open access journal from mdpi

10.3390/ani13111812

PMC10251952

Animal models continue to be necessary in many research fields, accompanied by ongoing ethical discussions regarding animal welfare. Therefore, we describe, in detail, our daily practice focused on the improvement of animal welfare (such as handling, enriched environment, study design, and experimental procedures), which results in a weight gain over time that has been shown to be an indicator of well-being. We also describe the reduction in the number of animals needed for our projects, thanks to the establishment of longitudinal studies.

Longitudinal studies on mouse models related to Alzheimer disease (AD) pathology play an important role in the investigation of therapeutic targets to help pharmaceutical research in the development of new drugs and in the attempt of an early diagnosis that can contribute to improving people’s quality of life. There are several advantages to enriching longitudinal studies in AD models with Positron Emission Tomography (PET); among these advantages, the possibility of following the principle of the 3Rs of animal welfare is fundamental. In this manuscript, good daily experimental practice focusing on animal welfare is described and commented upon, based on the experience attained from studies conducted in our Nuclear Medicine department.

1. IntroductionAlzheimer disease (AD) is a debilitating disease that causes progressive decline in cognitive and motor function, significantly reducing quality of life [1]. It is characterized by an early decrease in brain glucose metabolism [2], as well as the presence of amyloid plaques and neurofibrillary tangles. In addition, histological studies demonstrate that neuroinflammation is also a key feature of the AD brain [3]. Amyloid plaques are surrounded by activated astrocytes that produce reactive oxygen and nitrogen species, which may contribute to AD [4]. Animal models of AD are useful for studying these changes and the progression and development of the disease, with the goal of finding new diagnostic and treatment strategies. As the factors involved in the development and progression of the disease are different depending on the time during which the disease starts to manifest, different animal models for longitudinal studies are available.Each model reflects biological features related to AD. Involvement and the effects of all these factors during the time of disease development are of high interest. Positron emission tomography (PET) is a molecular imaging method offering a large variety of radioactively labelled substances targeting different biological structures or processes. It offers non-invasive measurement of the radiotracer concentration in tissue and is clearly translational, since the exact same tracers and imaging technology can be applied in animal models, as well as in research and clinical use in humans. For example, PET remains one of the few methods to allow direct assessment of the human central nervous system (CNS) pharmacology, providing information on target engagement and supporting dose selection [5]. Thus, it has an increasing role in studying the biochemical and physiological dynamics of the CNS. The goal of the 3Rs Principle [6] is to avoid animal experiments (Replacement), to limit the number of animals (Reduction), and to limit their suffering in tests to an absolute minimum (Refinement). From the perspective of the design and development of a longitudinal study, the principle of the 3Rs plays a fundamental role in the research of all the strategies aimed at maintaining animal welfare for the entire time of the experiment. Due to their shorter lifespan, life-course results are obtained much more quickly in animal models, which is especially important when studying aging and transgenerational disease transmission [7]. Unfortunately, the complex interactions between organs and cells within their regular environment still necessitate the use of animal models in some cases. For example, in order to establish cause–effect relationships, connections between systems and the onset of age-related pathologies need to be studied. Since animal experiments cannot be completely replaced in our research on the development of AD, we implemented different strategies, following the principles of Reduction and Refinement. Improving the quality of life for experimental animals and reducing their stress and pain are among the goals that must be considered when it is not possible to replace animals, according to the principle of the 3Rs. EU Directive 2010/63 states that the animals should have “space of sufficient complexity to allow expression of a wide range of normal behavior” [8] (Annex III, Section A, paragraph 3.1.). For this reason, we aim for the enrichment environment we use to reflect the normal living conditions of the animals as much as possible. Currently, the term enrichment environment (EE) refers to various objects, or a combination of them, that can be added to the bedding and nest material (which are now considered basic components). These components play the roles of cognitive, sensory, social, and motor stimulators, which promote the interaction of animals with objects and with each other [9]. The scientific literature is full of examples of benefits due to EE. For example, it can prevent barbering [10], it reduces the likelihood of alopecia [11], and it decreases the expression of abnormal repetitive behaviors and anxiety [12], as well as the development of depressive-like phenotypes [13]. We can also assume that EE prevents or mitigates the onset of boredom-like symptoms in mice, since, in humans, this can be triggered by predictability, monotony, and confinement, and similar phenomena in rodents indicate that boredom in laboratory animals is real [14]. Considering the gender difference, groups of males in an environment with excessive EE demonstrated a higher occurrence of aggression [15]. On the other hand, without adequate EE, the conditions of their natural habitat would not be reproduced (spread out, keep away, escape, hide, and predict the occurrence of aggressive encounters). In order to prevent aggressive events in male mice, in our facility, they are always located in a cage with littermates with the same genotype, or with around one week of age difference with the same genotype, as it is known that this can prevent aggression and, consequentially, the necessity of single housing [16]. In this report, we summarize methods of ensuring animal welfare in longitudinal studies characterizing age-related processes in mouse models of AD. In addition to the non-invasiveness of PET imaging, a number of established techniques have been adapted, which contribute to the 3Rs in repeated measurements.2. Materials and Methods2.1. Animal ModelsIn our longitudinal studies over recent years, we used the following mouse models:APPSL70 mice, transgene of the Amyloid Precursor Protein (APP) in Swedish and London mutations [17] (Figures 5–8 Section 3.1, Section 3.2 and Section 3.3) (n = 92, half group was treated);APP-NL-GF mice knock-in model for the APP gene in Swedish, London, Arctic, and Iberian mutations [18] (Section 3.2) (n = 79);APPPS1 and PS2APP mice, mutated, respectively, for the APP gene and for the genes of presenilin 1 (PS1) and 2 (PS2) [19] (Section 3.2) (n = 46 and n = 36);APPPS1 X Trem 2 mice, transgene of APP, as well as of PS1 for triggering receptor expressed on myeloid cells type 2 (TREM2) [20] (Section 3.2) (n = 33);P301S mice, mutated for the microtubule-associated protein (TAU), which is abnormally aggregated in neuronal and glial cells in AD [21] (Section 3.2) (n = 46);C57BL/6J wild–type mice (WT) as control group (Section 3.2 and Section 3.3) (n = 46).In the present study concerning the principle of the 3Rs applied to longitudinal studies in AD research, the APPSL70 is the mouse line used the most for examples.2.2. Study DesignWe performed longitudinal studies in AD mouse models, in which each animal acted as its own control. The models consisted of PET/CT scans from 3 to 5 time points, with 2 to 4 different radiotracers characterizing changes in biological targets up to the age of 12 to 18 months. Furthermore, a Morris water maze (MWM) test was integrated before or after the last PET/CT time point in order to study spatial learning. At the end of each study, an intracardial perfusion with PBS was performed after intraperitoneal injection of a solution of 300–500 µL (weight-dependent) ketamine/xylazin (4 mL Ketamine 10%/1 mL Xylazine 20 mg/mL from Serumwerk Bernburg Germany, up to 24 mL with NaCl 0.9%), for the preparation of samples for histological and biochemical analyses (Figure 1).2.3. Positron Emission TomographyIn each animal, we used 2 to 4 of the following 18F labelled tracers: [18F] D2-Deprenyl for reactive astrocytes which surround the Beta Amyloid plaques [3]; [18F] Florbetaben for beta amyloid accumulation; [18F] Ge-180 for the 18-kDa translocator protein (TSPO), as its local upregulation is a sensitive marker for the microglial activation in AD brains [22]; and [18F] UCB-H for synaptic density (synaptic loss or synaptic sprouting) [23]. PET imaging was performed under constant anesthesia with isoflurane (1.5% at 1.5L oxygen flow per minute) with a Nanoscan PET/CT (Mediso Ltd., Budapest, Hungary). For anatomical information, the system was equipped with an X-ray Computed Tomography System (CT) in line with the PET scanner. After induction of anesthesia with isoflurane, the eyes were protected from drying out by topical application of an eye ointment (Bepanthen, Bayer AG, Leverkusen, Germany). In order to judge the depth of anesthesia before the radiotracer´s injection, surgical tweezers were used to check whether the inter-toe reflex could still be triggered. If it could no longer be triggered, a micro-catheter was inserted into the lateral tail vein (30 G needle, 7 cm plastic tube, 30 G attachment, flushing with 0.9% isotonic saline solution); the correct position of the indwelling venous cannula for the application of the radiotracer was checked by administering a small amount of isotonic saline solution (20–30 µL) into the catheter. The injected radioligand solution consisted of a total of 150 µL, and contained approximately 20 MBq of radioactive tracer. PET/CT measurement was carried out up to 60 min post injection: for [18F] Ge-180 and [18F] Florbetaben, the measurement was carried out for 30 min., 60 and 30 min. after radiotracer uptake, respectively (static scan); those for [18F] D2-Deprenyl and [18F] UCB-H were carried out immediately post injection, for 60 min (dynamic scan). Up to four animals were measured in parallel, and image data were generated for each animal from head to tail. After the experiment, the mice were placed in a fresh temporary cage with food and water, warmed by heating mats. The animals were returned to their home cages only when they were fully awake.The radioactive waste coming from the animals (feces, urine) and all the equipment used on the workspace (e.g., syringes, paper tissue, paper mats, and gloves) were collected in black boxes, upon which were written the isotope used and the date and time of the experiment. These boxes were then collected by the radiation safety personnel of the department.2.4. Water MazeThe Morris Water Maze (MWM) is useful to test hippocampal-dependent learning, including acquisition of short- and long-term spatial memory [24]. Typically, it consists of a six-day trial, and it has to be conducted by the same operator in the same room in order to reduce odor trail interferences [25]. In brief, the first test day served for acclimatization to the visible platform (5 min per mouse). Thereafter, the mice underwent five training days, during which each mouse had to perform four trials per day, with the platform visible on the first training day, and the platform hidden under water for all other training days. After the trial, the mice were placed in a heated box to dry. The test day entailed a single trial with complete removal of the platform. The trial length on all training and test days was set to a maximum of 70 s. The video tracking software EthoVision® XT (Noldus) was used for analyses of escape latency, platform frequency, and attendance on the platform quadrant during the trial. In our longitudinal studies, the test was conducted in a room adjacent to both the cage facility and the room where the preparation and PET scans took place, and the mice resided in a cabinet in the same behavioral room for the duration of the trial. This minimized the stress caused by this additional procedure and by moving animals to other, more distant, locations. The personnel who carried out the MWM test received appropriate training from a veterinarian before the experiment and was also part of the team that regularly deals with animal welfare, as well as with the preparation and execution of PET scans.2.5. Enrichment EnvironmentIn our facility, the male mice are housed (in groups of 3–4 siblings, with one week age difference and same genotype) with the same EE as the females. They are more intensively monitored (for too much grooming, self-isolation attributable to the attempt to escape from the attacks of the other mice, and the presence of wounds) so as to prevent the onset of high levels of aggression, but still keeping them in a condition to be able to maintain a good level of play and sociopositive behavior. The female mice are housed in groups of 3 to 5. All of the mice are housed in large IVC-Techniplast cages (425 mm × 266 mm × 185 mm), with a 12:12 h light: dark cycle, humidity of 45–65%, and temperature of 23–26 °C. All toys are changed once a week. The food (standard diet from Sniff Spezialdiäten GmbH) is placed ad libitum, as is water. Mice were allowed to acclimatize to their environment for 7 days before any experimental procedure. The EE can be divided into categories (Figure 2) [26]. In our studies, we used the basic nest (2A), structural EE (2B), foraging EE (2C) and housing EE (2D):2.6. HandlingWe use a handling method that reduces stress, and it can be used both during experiments and in daily practices (such as weighing, animal checks, and changing cages) and does not need any additional material (for example, tunnels). Once the animal is picked up by the tail from the cage, with a quick but gentle twist of the hand, it is immediately placed in the cupped hand (Figure 3).2.7. Monitoring Signs of StressAnimals are monitored for any sign of stress using the Mouse Grimace Scale. It consists of a standardized behavioral coding system with high accuracy and reliability in which no-pain pictures and descriptions are compared with pictures of moderate and severe distress via mouse facial expression (e.g., orbital tightening, nose bulge, cheek bulge, ear position, and whisker change) [27]. In our longitudinal studies, the mice are scanned two or, maximum, three times a week for each time point, and they are always observed in the postanesthesia phase, until they have completely recovered. They are also checked on the day after the procedures according to the Mice Grimace Scale, in order to identify the onset of visible postprocedural stress symptoms. It is known that this scoring system allows for the identification of the degree of pain in mice, but also shows that they manifest their pain using facial expressions [26]. Furthermore, the MGS measures not only pain, but also distress, fear, and discomfort, which can be detected by the ear and eye score [28]. Mice hide signs of pain and suffering to avoid becoming prey, and, generally, they show only subtle signs of suffering and pain, such as weight loss [29]. In addition to the parameters shown in Figure 4, the weight of each mouse is monitored up to once per week during all the experimental phases, as it has been demonstrated that mice gain weight when the stress level is lower, and this can be achieved with appropriate EE and enough ventilation (in IVC cages) [30].2.8. AnesthesiaIn contrast to human studies, the imaging of small animals generally requires anesthesia [31]. In our longitudinal studies, we use isoflurane because it is well suited for animal PET scans for up to 6h of measurement time [32]. Repeated isoflurane anesthesia causes only mild short-term distress and impairment of well-being, mainly in the immediate postanesthetic period [33]. This can be kept under control by monitoring the respiratory rate and heart rate, and keeping the body temperature of the animals constant throughout the process in order to prevent the onset of hypothermia. Maintaining the temperature and anesthesia at a constant level during the experiment minimizes any potential pain, suffering, distress, or lasting harm to the animal [34]. In a study conducted by Baier J et al. in 2020 [35] on repeated MRI scans, using isoflurane for general anesthesia and its maintenance three times a week for four weeks, the mice showed no alterations in animal welfare due to the repeated procedures. Another study [36] demonstrated that nest-building activity is not altered by a second exposure to isoflurane.2.9. Mouse Hotel BedUsing a four-mouse bed (Mediso Ltd., Budapest, Hungary) it was possible to scan multiple mice (up to four) simultaneously (Figure 4). The use of this chamber is advantageous in many aspects:The chance to perform more scans in the same day using radioactive isotopes with a short half-life (18F: 109, 7 min). This is especially important in longitudinal studies with a large number of animals because all of the scans have to be performed respective of the age-related time point.There is a lower injected volume. This is important, as the 18F-based tracers decay quickly. Since the activity has to be the same for each animal, the volume of the injection needs to be increased according to the radioactive decay in sequential scans. Thus, the volume to be injected would be above the maximum limit allowed for these animal experiments. Using the 4-mouse bed, this problem is circumvented, since four mice are injected and scanned at the same time (instead of four individual scans one after the other).This particular bed for four animals is equipped with a heating system that keeps the temperature of the bed constant at about 37 °C, which is very important during anesthesia, as hypothermia can arise, and also because the temperature can have an effect on the biodistribution of the injected radiotracer. In addition, there is a monitoring system for cardiac and respiratory function that allows continuous monitoring in each animal.3. Results3.1. Longitudinal Studies Enable the Analysis of the Same Animal with Different Radioligands during Progressive AgingFigure 5 shows an example of PET, CT, and Fusion PET/CT of four mice imaged simultaneously.Figure 6 shows an example of beta amyloid in a transgenic mouse brain over time with the same radiotracer ([18F] Florbetaben): the same mouse was scanned at 6, 9, and 12 months of age. Coronal slices of the brain from the same animal after the injection of four different radioligands at 12 months of age with a reference MRI (Magnetic Resonance Imaging) used as anatomical reference are also included.3.2. The Longitudinal PET Studies follow the Principle of Reduction by Russell and BurchThe graph in Figure 7 shows the number of mice used by our research group in longitudinal studies in recent years, and the number of mice that would have been necessary, in theory, for the same projects using separate groups for each time point.3.3. Gentle Handling and EE Reduce Stress in Longitudinal StudiesFor all the experiments, the MGS resulted in a score of zero.We observed that, during the entire stay of the animals in the facility (9 months or more), the nest material was regularly used, as were the other components of the EE. The combination of gentle handling and EE makes the animals more docile and less stressed, even during the time before/during/after the experimental procedures. The graph in Figure 8 shows the development of the weight in a female AD mouse model (APPSL70, n = 28) and in wild-type controls (n = 17), each receiving PET scans with four different radiotracers at three different time points, with a MWM test before the last time point. We observed an average of 13% mortality during this experiment of up to six months, independent of the mouse line.While the body weight of wild-type mice increased continuously, the body weight in the APPSL70 AD model did not increase further after ~10 months of age. A very similar weight gain (9.6% mean percentage of gain weight from 3 to 6 months of age and 11.8% from 6 to 9 months of age) was observed in the APPPS1 group, which consisted of 11 animals. No mouse, either wild-type or transgenic, had to be excluded from the trials due to stress/fear factors, and no animal developed signs of anxiety/stress over six months of study (up to one year of age for mice).4. DiscussionWe report on aspects of implementation of the 3Rs Principle in longitudinal mouse studies for AD research. Of note, the reported methods and findings concerning handling and housing are not restricted to AD research. Repeated PET/CT measurements, combined with a behavioral test, proved feasible in AD mouse models and wild-type mice without losing animals due to increased stress, up to 12–18 months of age. Reduction in the number of animals used for experimentation is achieved by longitudinal, multi-tracer PET scans in the same animal; thus, each mouse is its own control, and the results of behavior tests can also be correlated for each individual animal. No episodes of severe weight loss were detected during the study (according to our approved scoring system). The weight loss observed in the APPSL70 AD model is compatible with the findings in previous studies on different AD mouse models, showing that weight loss is connected to the reduction in body adiposity, an increase in energy expenditure, and a decrease in food efficiency connected to the progression of AD [37,38]. It is known that stress has a negative effect on the results of the MWM [39,40], and in some studies, it has been hypothesized that the stress from less frequent handling before the test can be responsible for the failure of some mice in finding the platform [41]. It is also known that the induction of stress produces learning deficits in the MWM [42,43]. In this context, considering that routine handling influences animal welfare and the relationship that is established between the handler and the animal, we observed that the non-stressful handling method assures low variability in the results obtained from the test. In all procedures, we adhered to the same animal handling procedure. The standard method, picking up the mice by the tail, induces more anxiety and stress compared to more gentle methods, such as cupping and tunnel handling [44,45]; it was also previously demonstrated that gentle handling can foster a better relationship between the handlers and the rodents, and, if implemented as the standard of care, handling can reduce depressive symptoms in mice, producing data that are more reliable and, in general, improve the animals’ well-being [46]. Ueno et al. [47], in 2020, demonstrated that repeated exposure of the mice to the experimenter´s hand before conducting behavioral tests allows them to become accustomed to it and to reduce anxiety about high altitudes. In addition, it has been previously shown that reducing mouse anxiety due to handling contributes to a reduction in the number of animals required for experiments [47]. The same handling method (described in Methods 2.6) is used to put the mice under general anesthesia before the radiotracer injection. Thus, no immobilization method is necessary. How much our long-term gentle handling (up to six months stay in the facility) effectively reduced stress in both male and female mice will need to be shown in a further study including non-gentle handling control mice. While there are several advantages to using the four-mouse hotel chamber for PET imaging, it has the limitation of keeping one single flow rate for all animals. Thus, individual adaption is not possible, but could be improved in a new design. Providing enriched environments in sufficiently large cages also contributes to decreasing mortality, as has been demonstrated by comparison to conventional housing [48]. We used established, qualitative measures of stress, as well as animal weight, as a surrogate of well-being. Quantitative measures, such as corticosterone level, could not be integrated into the study protocol, since the mice are sacrificed only at the end of the project, in order to follow the Reduction Principle.5. ConclusionsIn the context of longitudinal studies, animal welfare assumes particular relevance, since the animals reside for long periods in the facility during their progressive aging and, at the same time, are involved in all of the experimental phases. Non-invasive imaging using PET allows repeated measurements of multiple biologically relevant tracer distributions over time in the same animal, thus significantly reducing the number of animals required. Reduction is achieved with the establishment of longitudinal studies that use radioactively labelled molecules in very small-volume solutions without pharmacological effect. Refinement is ensured by the possibility of conducting PET/CT scans with four animals at a time, a varied and stimulating enriched environment, and a standardized handling method during all phases of the experiment with a minimal number of experimenters.

animals : an open access journal from mdpi

10.3390/ani11041074

PMC8070664

Physiological stress increases the activity of the hypothalamic-pituitary-adrenal (HPA) axis and the secretion of cortisol, which might cross the placenta and affect foetal development. Stress in sows can be affected by management factors such as enrichment, different feed systems of the housing accommodation, and is reflected in the salivary cortisol concentration. It is unclear how stressed the sow must be before there is an impact on foetal growth, but higher levels of cortisol might affect the maturity of piglets at birth as well as their birth weight. Therefore, it could be beneficial to accommodate gestating sows in the least stressful manner, not only for piglet performance but also for sow welfare. Cortisol concentration in sows seems to be influenced by a combination of parity and feed systems, but its connection to those factors, as well as to foetal development, warrants further investigation.

The aim of this study was to investigate herd cortisol levels as an indicator of stress during gestation in three different feeding systems. Twelve commercial Danish herds with 800 to 3050 sows were included, with either free-access feeding stall (Stall), floor feeding (Floor), or electronic sow feeding (ESF; n = 4 herds per system). Saliva samples were collected from 30 sows/herd in the gestation unit for cortisol analysis with an average of 67.2 gestation days for ESF, 72.4 days for Floor, and 68.6 days for Stall. Data on piglet birth weight (PBW) and the percentage of intrauterine growth restricted (IUGR) piglets from 452 litters (9652 piglets, 8677 liveborn) from all 12 herds were obtained on the saliva collection days. The cortisol levels in saliva increased throughout gestation (p < 0.01), and lower concentrations were observed among sows belonging to Stall (4.80 nmol/L), compared to Floor (7.03 nmol/L) and ESF (7.87 nmol/L), and that difference was significant as an independent effect in the case of ESF (p < 0.01). There was no difference between Floor and ESF or Stall and Floor (p > 0.05). An interaction was observed between parity and feeding system, with parities 4–5 in ESF herds having lower levels than other parities within the ESF system (p = 0.02).

1. IntroductionPiglet mortality is an ongoing concern of the Danish pig production, and piglets that have a low piglet birth weight (PBW), suffer from intrauterine growth restriction (IUGR) or are born of litters with a high within-litter variation of PBW (PBWCV) have an increased risk of dying before weaning [1,2,3]. It is therefore of interest to investigate factors that could potentially influence and result in a low PBW, a high PBWCV, or a high occurrence of IUGR piglets so that these parameters can be improved. Piglets suffering from IUGR have not received enough nutrients during development, and their brain is prioritised for the survival of the organism [4], giving them their characteristic head shape [2,5,6]. It is not known when during gestation this takes place although recent studies suggest differences in the development of porcine foetuses already at day 28 of gestation (Strathe et al., unpublished). Physiological stress increases the activity of the hypothalamic-pituitary-adrenal (HPA) axis and the secretion of glucocorticoid hormones, namely cortisol [7], which might cross the placenta and affect foetal development [8]. The exact mechanisms behind this are unknown, but it has been suggested that a higher transfer of cortisol across the placenta will limit foetal insulin-like growth-factor-1 (IGF-1) since cortisol is an inhibitor for IGF-1 and, thereby, affects foetal development [9]. Further, preterm growth-restricted babies display alterations in the growth hormone (GH)–IGF-1 axis (with increased GH and low IGF-1 concentrations) [10]. It is, therefore, possible that increased levels of cortisol could inhibit growth (measured by birth weight) and development (measured by IUGR). Several biological factors could potentially affect cortisol levels, and thereby growth and development of the foetus and ultimately sow performance. For example, Roelofs et al. [11] reported higher salivary cortisol concentrations for primiparous sows when compared with multiparous sows, and Strawford et al. [12] reported that concentrations are lowest in the intermediate sows (2nd to 3rd parity), compared to younger (1st parity) and older sows (≥4th parity). In addition, time of sampling during gestation as well as feeding system might influence salivary cortisol concentrations. Anil et al. [13] found that cortisol levels were higher at day 108 than at days 28, 56, and 84 of gestation in individually fed sows, whereas Holt et al. [14] found a decrease between days 40 and 80 of gestation. For sows housed in an Electronic sow feeding (ESF) system, saliva cortisol concentrations were higher at day 108 than at days 28 and 56, but not higher than at day 84 of gestation [13]. Moreover, it was reported that cortisol concentration measured by hair samples of sows in late gestation increases with litter size [11]. Thus, saliva cortisol concentrations during gestation might vary, depending on biological circumstances.There was also evidence that feeding systems and housing could influence salivary cortisol levels and that this, in turn, might influence the piglet performance. Merlot et al. [15] found higher levels of cortisol in sows housed under barren systems (a conventional French system on slatted floors) compared to an enriched system with larger pens and deep straw. This difference could be explained by higher social stress and frustration due to an inability to perform rooting behaviour and to satisfy hunger [15]. Additionally, sows housed in an ESF system scored higher skin lesions than sows housed in gestating stalls due to persistent fighting around the ESF stations [16]. Repeated competition around feeding might, therefore, affect the level of stress in the sow and result in high levels of maternal cortisol, which could decrease PBW [17]. Further, the amount of feed or lack off could potentially increase stress, as Amdi et al. [9] found that restricted fed gilts had higher levels of salivary cortisol than ad libitum fed gilts. The differences in concentration levels of cortisol might therefore be explained by different feeding systems and result in differences in PBW, PBWCV, and the occurrence of IUGR piglets between sows of different feed systems.It, therefore, seems imperative to investigate if different feeding systems can cause different levels of stress to the sow and if the stress is a contributing reason for impaired foetal growth. The aim of this study was, therefore, to investigate the correlation between the overall herd stress level and PBW, IUGR, and PBWCV on selected sows farrowing on the same sampling day. For that, two hypotheses were developed: the level of stress in the sow during gestation, measured by cortisol, can (1) be affected by the feeding system, and (2) affect foetal development, measured by the percentage of IUGR piglets in a litter, PBW, and PBWCV at herd level.2. Materials and Methods2.1. Ethics Statement All animals originated from commercial production facilities. No measurements were made that were outside of the standard industry animal husbandry techniques, and the animals were cared for in compliance with local legal standards. The health and welfare of all animals were monitored throughout the sampling days by farm staff, according to the farms’ standard operating protocols and veterinary recommendations. 2.2. Animals and DesignThe study was conducted on 12 commercial Danish pig herds with herd populations ranging from 800 to 3050 sows. Three different feeding systems could be observed in the gestation units: Free-access feeding Stall (Stall), Floor feeding (Floor), and Electronic sow feeding (ESF; n = 4 herds per system). The herds were selected based on the feeding system in the gestation unit and, secondly, a large herd size in order to record as many litters in each parity over three days as possible. Seven of the herds were selected due to their involvement in other SEGES Danish Pig Research Centre research projects (herd A, C, D, F, I, K, and L), and the remaining herds (B, E, G, H, and J) by contact to a local pig production advisor. Additionally, the herds were selected due to their geographical position, so that farm visits in early mornings and evenings were possible. The herd size and disease status of the herds can be found in Table 1. The study included production data from 8677 liveborn piglets, from a total of 9652 piglets from 452 litters. All sows were crossbred Danish Landrace × Danish Yorkshire (parity 1 to 10 (mean ± sd.; 3.84 ± 1.97)), artificially inseminated with semen from DanBred Duroc boars. Data were collected between mid-September and mid-December 2019. Recordings were made successively over the three days, with the most farrowing in one given week on each herd resulting in the average recording of 40–76% of all farrowing in the weekly farrowing batch of the herds. 2.3. Management RoutinesThe daily management routines were performed as usual in the individual herds, and recordings were collected early in the morning to prevent disturbing the work routines of the employees. In all herds, sows were housed in a crated system through the nursing period and confined in locked or free-access stalls for four weeks after insemination. The size of the groups in the gestation unit, as well as the number of daily feedings, are listed in Table 1. The herds followed the Danish legislation for stocking density, percentage of slatted floors, and enrichment. Sows in herd G were housed in a deep litter section in the gestation unit. Gilts in herd A were not housed in a Stall system as the sows but in pens with nine gilts in each and fed by liquid feed in a trough. The feed used for gestation diets was formulated to meet or exceed the feed recommendations for gestating sows in all herds [18], and a reduced version of diet formulations can be found in Supplementary Table S1.2.4. Saliva CollectionAt least 30 saliva samples were collected from sows in the gestation unit of each herd on the days the production data was collected. Therefore, the saliva samples did not correspond to the sows that farrowed the piglets included in the study. Saliva was collected with the use of a Salivette (Salivette plain, Sarstedt, Leicester, UK), which consisted of a cotton bud that fits inside a centrifuge tube [19]. The sow chewed voluntarily on a cotton bud until it was moisturised, as this method allowed cortisol sampling in an easy and stress-free manner [19,20]. The saliva samples were collected in the morning between 08:00 a.m. and 11:30 a.m., which was either before or at least half an hour after feeding. The sows were selected randomly; however, in herds with ESF, sows seen eating were not selected to avoid contamination of saliva samples. The stocking density in the pen was noted (Table 1), as well as sow ID, parity of the sows, and days from insemination. The saliva samples were centrifuged (CM-6MT; ELMI Ltd., Riga, Latvia) at 1000× g for 10 min at room temperature within four hours of collection. Afterwards, the centrifuged saliva samples were transferred to Eppendorf tubes and frozen at −18 °C until later analysis. 2.5. Data on Farrowing SowsFor each sow with a newborn litter, sow ID, date of farrowing, parity, backfat thickness, number of total born, liveborn, stillborn, and mummified piglets were recorded. Backfat thickness from the P2 site, number of days from last weaning to first insemination, length of the previous lactation, and length of gestation were also recorded but not used for this specific study (data not shown). Feed curves and diet formulations were collected at the herds. 2.6. Recordings of PigletsRecordings of piglets were carried out as soon as possible after farrowing had ended and before litter equalisation, so they were no older than 24 h at the time of weighing and IUGR scoring (Figure 1). When possible, both live- and stillborn piglets were individually scored as either normal, mild IUGR (mIUGR), or severe IUGR (sIUGR), and the sex of the piglets was recorded. Some dead piglets were removed before registrations were carried out, and in this case, the number of stillborn piglets was only counted by employees, and sex and IUGR score were not noted. Dead piglets were classified as either stillborn or liveborn but dead and, if possible, the cause of death was noted for liveborn but dead piglets. Dead and wet fully formed piglets with the periople still present on the hooves were noted as stillborn. Test of inflation of the lung tissue was not performed. Piglets were noted as liveborn but dead if the above-mentioned criteria of stillborn piglets were not fulfilled. The reason of death was noted as either (1) crushed if visible trauma or subcutaneous edema appeared on any part of the body; (2) euthanised if clear signs of head trauma due to euthanisation was visible; or (3) others if no signs of either (1) or (2) could be detected. All liveborn (including liveborn but dead) piglets were individually weighed by placing the piglet in a bucket hanging on a digital weight (5 g weight interval; Ryom Digital Hanging Scale, Hatting, Denmark). To minimise the risk of disease spreading, a new digital weight was used at each herd (accuracy of each weight; ± 25 g deviation). Piglets were scored as either normal, mild IUGR (mIUGR), or severe IUGR (sIUGR). The parameters for sIUGR and mIUGR were based on modified characteristics from Chevaux et al. [21], Hales et al. [5], and Engelsmann et al. [22]. The primary parameters characterizing IUGR piglets were defined as steep/dolphin-like forehead, narrow hind part, and low birth weight (below 1100 g). Secondary parameters were defined as bulging eyes, wrinkles perpendicular to the mouth, spiky hair, and unstable mobility. The sIUGR piglets showed all primary parameters distinctively, had at least one of the secondary parameters, and a weight of no more than 1050 g. Piglets characterised as mIUGR had the primary parameters with a weight of a maximum of 1100 g and no more than one of the secondary parameters. A normal piglet had none of the parameters and weighed more than 650 g. Figure 1 illustrates the distinction between sIUGR, mIUGR, and normal piglets according to the shape of the head and the hind part.2.7. Salivary Cortisol AnalysisSamples with insufficient amounts of saliva (<100 µL) were excluded from the analysis resulting in between 19 and 31 samples analysed for each herd. Saliva cortisol was measured by ELISA (Saliva Lab Trier, daacro GmbH & Co. KG, Trier, Germany), and the inter-assay was 3.35 CV %. Two samples were above the upper limit of quantification of 82.77 nmol/L, which might be due to blood contamination, and were therefore excluded from further analysis.2.8. Data Management and Statistical AnalysisAll data management and analysis were performed in RStudio Version 1.2.503 © 2009–2019. The dependent variable in the analysis was salivary cortisol concentration (nmol/L) for individual sows. The study unit was, therefore, the sow. Outliers in the cortisol level data were removed based on two standard deviation criteria, which resulted in the exclusion of 17 observations. Salivary cortisol concentrations were log-transformed to improve the distribution of residuals. The independent variables initially tested for model inclusion were gestation days, feeding system, herd, parity, litter size, and average piglet weight. Parity data as a numeric variable did not follow a linear relation to the outcome and, therefore, was categorised as four groups to improve the fitting: 0–1, 2–3, 4–5, and ≥6. All variables were also checked for confounding and interactions. Model building was based on univariable tests, for which statistical significance was accepted at p < 0.05, and 0.05 < p < 0.10 was considered a tendency.All variables which showed a significant effect or a tendency were included in the multivariable models. The variables were then tested for significance as part of mixed-effects models, and ANOVA tests were run between different combinations of variables to assess any differences in fit and explanatory power. The following linear mixed-effects model has the final structure used to estimate the effect of parity and gestation days on the saliva cortisol levels:Yi= µ + βj + H(l)+ F(k) × αi + εijkl(1) where Yi is the response variable (saliva cortisol level); µ is the overall mean (intercept); βj is the fixed effect of gestation days (j = (1, 2, …, 113)) explained as a covariate; Hl is the random effect of the herd (l = (A, B, …, L)); Fk × αi is the interaction term between feed system (k = (Stall, Floor, ESF)) and parity (i = (0–1, 2–3, 4–5, ≥6)); and εijkl is the residual error component, which is assumed to be independent and normally distributed.The model was run with the help of package lme4 in R [23] (lmer), and the random effect structure was assessed using the Restricted Maximum Likelihood (REML) estimation (RMLE=TRUE). Results were presented as least square means (lsmeans) and their standard errors (SE), as obtained with the help of package lsmeans [24]. Although the difference in degrees of freedom at the different levels of a mixed-effects model did not allow for a clear evaluation of significance through the computation of p-values, a reasonable approximation could be obtained using the package lmerTest, which used Satterthwaite’s degrees of freedom method [25]. These approximated p-values will be presented here, along with the regression estimates.3. Results3.1. Salivary Cortisol ConcentrationsThe average salivary cortisol concentrations across the feed systems at different parities are presented in Table 2, and the results of the linear mixed-effects regression can be found in Supplementary Table S2. To facilitate comprehension, cortisol concentrations are presented as natural values (nmol/L), as the exponential of the logged coefficients and means produced by the model. Statistically significant independent effects on salivary cortisol concentrations were observed for gestation days and feeding system. It was estimated that, for each one-day increase in gestation, the sow’s salivary cortisol concentration should increase by 1.00 nmol/L on average (p < 0.01). Figure 2 illustrates salivary cortisol concentrations (untransformed observed data) between 30 and 100 days of gestation for individual feeding systems. The plotted values were obtained using a loess smoothening function, with the grey bands corresponding to an approximation of a 95% confidence interval calculated as each smoothed value ± 1.96 × standard errors.Lower concentrations of salivary cortisol were observed, on average, among sows belonging to the feeding system Stall (4.80 nmol/L) when compared to Floor (7.03 nmol/L) and ESF (7.87 nmol/L), and that difference was statistically significant as an independent effect in the case of ESF (p < 0.01). There was no significant independent difference between Floor and ESF or Stall and Floor. A significant interactive effect was also observed between parity and feeding system, with parities 4–5 in ESF herds having a lower increasing effect than other parities in this same system (p = 0.02). This could be interpreted as—although sows in herds with an ESF feeding system had, on average, significantly higher concentrations of salivary cortisol than sows in Stall herds—the difference in concentrationwas not as large when Stall sows of any parity were compared specifically with ESF sows in the fourth or fifth parity (Table 2). A similar trend was also observed for ESF sows in parity 0–1 (p = 0.08). No interactive effects involving the Floor feeding system or other parity groups were observed.Variation between herds (random effect) accounted for an additional standard deviation of 1.19 nmol/L to sow-level predictions (Table S3). On that note, the average cortisol concentration in herd G (5.88 nmol/L) was reasonably lower than in F (8.99 nmol/L), J (8.32 nmol/L), and L (8.44 nmol/L), and in the other herds within the ESF system. For a matter of diagnostics, the models were also run without that herd. Omitting herd G increased the mean cortisol level of the ESF system and, consequently, increased the coefficients and significance level for the independent and interactive effects related to the feeding system. It did not, however, alter which predictors were found important in the previous model, nor uncovered new relationships, and the authors considered that the observed values for herd G were valid as it differed from the other ESF herds by being the only one that uses deep straw bedding. Therefore, the final analyses included that herd. Other tendencies or differences between parities within feed systems were not detected.3.2. Production DataThe average days in gestation when the saliva samples were collected were 67.2 days for ESF, 72.4 days for Floor, and 68.6 days for Stall. The average herd salivary cortisol level and the recorded average herd PBW, sIUGR, mIUGR, and PBWCV from the sows that were in the farrowing unit the day the saliva samples were collected in the gestation unit can be seen in Figure 3. The sows from which the production data was collected were not the same as the ones providing cortisol samples; therefore, no analyses at the sow level were possible. The data were summarised at the herd level to obtain an ecological overview, but the resulting sample size of 12 herds was too small for any kind of inference. 4. DiscussionThe aim of this study was to measure stress levels on herd level measured by cortisol in sows housed with different feed systems. Salivary sampling provided a non-invasive, low-impact sampling of cortisol, and was considered just as representative as blood cortisol for the short-term levels of stress in the animal [19,20]. In the present study, no connections between herd cortisol levels and the results of PBW and the occurrence of IUGR piglets (both severe and mild) were evident. However, it must be stressed that this was an ecological study as the cortisol levels were not from the same sows that the piglet information came from, but it was the overall herd performance compared to the overall herd cortisol level during gestation. Therefore, several factors were not taken into account, such as genetics, individual variation, exercise, and stocking density that could all influence cortisol levels. The observed increase of cortisol throughout gestation was most likely a biological response as cortisol peaked around parturition [26] in sows and dropped again throughout the first few weeks of lactation [27].It is not known what level of severity of cortisol exposure is required to have an effect on PBW. In previous studies, the effect of prenatal stress was investigated by manipulating maternal cortisol concentrations with oral administration of hydrocortisone acetate (HCA), injection of adrenocorticotropic hormones (ACTH) [28], or by stressful events such as rough handling or frequent social mixing [29,30]. Neither rough handling, social mixing, or ACTH injection affected the PBW, whereas sows subjected to HCA had lighter piglets [17,28,30,31]. Hydrocortisone acetate was metabolised directly to cortisol in the body [32], whereas ACTH stimulated the adrenal cortex to release cortisol [7]. Rough handling and social-mixing-imposed stress on the sows, which could be detected by increased cortisol concentrations in saliva or blood, though not to the same level as administered ACTH did [30,31]. The sows in the study by Kranendonk et al. [17] were given HCA for 30 days in early, mid, or late gestation. The offspring of the sows, which received the treatment in early and late gestation, had the lowest PBW whilst the offspring of the sows treated in mid-gestation were intermediate in relation to the control group, which had the highest PBW [17]. The contradictory results might be explained by the timeframe in gestation that the treatment was given and the severity of the prenatal stress imposed. The levels of cortisol of the stressors were much higher in the HCA study than what other studies reported ([9,20]. The HCA treatment elevated the salivary cortisol levels to 23.3–29.9 ng/mL [17], equal to 64.3–82.5 nmol/L, which were very high compared to the level reported by Cook et al. [20] of sows stressed by handling and transport (11.2 nmol/L) and the levels found in the current study. The HCA was also administered every day for 30 days, whereas the sows were exposed to the other stressors less frequently. At last, HCA had the lowest impact in mid-gestation [17], which was the period that one of the studies with ACTH treatment and rough handling was conducted [30]. When ACTH was administered to sows in early pregnancy, it tended (p = 0.09) to affect the PBW [28]. This indicated the importance of the time of stressors. Therefore, it was likely that HCA created high levels of cortisol continuously for a period long enough to damage the foetal growth, which was not the case for the other stressor studied. This could indicate that maternal stress must be quite severe or perhaps long-term to affect foetal growth, even though it was unclear how stressed the sow must be in a commercial setting before that happens.Feeding systems did affect saliva cortisol concentrations, but the effect on productive performance seemed more complex. Maternal saliva cortisol concentrations were higher in the morning (9:30 a.m.) prior to feeding than later in the day (12:30 and 15:30 p.m.) and were also affected by diet restrictions and body condition [9]. Amdi et al. [9] found that gilts fed a restricted diet (1.8 kg/day) from day 25 of gestation had higher saliva cortisol levels (8.50 nmol/L) than gilts fed a high feeding level (3.5 kg/day, 5.00 nmol/L) in the same period. In addition, thin gilts (14.8 mm backfat) also tended to have higher saliva cortisol levels, 7.34 nmol/L, than fat gilts (20.2 mm) at 6.5 nmol/L [9]. Additionally, feeding frequency (one or two feedings per day) was found to not affect saliva cortisol concentrations of individually stalled gilts in a study by Holt et al. [14]. The average level of saliva cortisol in this study was in line with earlier studies on gestating sows [9,20]. Even though the cortisol levels in sows under the Floor and ESF systems (7.03 and 6.36 nmol/L, respectively) were higher than those in Stall (4.01 nmol/L), these levels were still close to earlier reports of individually stalled gestating sows subject to no stressors (5.8 nmol/L [20]), and also below the levels reported for gestating sows subjected to common stressors such as handling and transport (11.2 nmol/L [20]). This was in agreement with Anil et al. [13], who found that at any given point in gestation, salivary cortisol concentrations were higher in sows housed in the ESF system than for sows housed in individual stalls, but the differences in cortisol concentrations between systems did not seem to affect PBW [13]. This could indicate that, in general, the level of stress in the sows was not high enough to affect the development of the piglets even though it differed between systems. One of the farms (G) had straw in the ESF system and, on average, lower levels of salivary cortisol than the other farms with the ESF system. This indicated that several factors, for example, housing and management in the different feed systems, should be considered together. In addition, cortisol production showed large inter-individual variation, which had a considerable genetic basis [33], and this should, therefore, also be considered. Additionally, stocking density could influence welfare indicators, but with no difference in cortisol between groups in growing pigs [34], suggesting multiple factors should be included in order to assess the full physiological response of the feeding systems. In addition, it could be discussed if the absolute values were the most accurate way of assessing cortisol levels or if differences from, for example, a herd baseline level would be more accurate due to individual variation. We had previously reported salivary cortisol values in differences from a baseline value [27], and perhaps herd cortisol levels need to be assessed continuously throughout a longer period to establish a herd baseline.It was shown that dynamic group management in ESF did not affect maternal saliva cortisol concentrations when compared to static groups [12], but increasing floor space allowance (from 1.4 m2/sow to 3.0 m2/sow) decreased aggression and plasma cortisol levels around mixing [35,36]. An enriched environment (deep straw, 3.4 m2/sow) lowered the maternal saliva cortisol concentrations in both early and late gestation compared to sows in conventional housing (slatted floor, 2.4 m2/sow) [15,37,38]. In agreement, contradicting effects on PBW were reported in three different studies, which tested the effect of enriched or conventional housing [15,37,38]. One study reported lower PBW from conventionally housed sows [37], contradicting another study by the same authors, where they found no effect of housing [15]. Even though enriched or conventional gestational housing had no effect on mean PBW (~1.5 kg), a higher percentage of lower birth weight piglets (<1.2 kg) was reported from sows of conventional housing [15,38] and a higher percentage of heavy birth weight piglets (>2.0 kg) was seen from the enriched facility [38]. The sows in enriched housing had salivary cortisol concentrations of approximately 3.45–4.14 nmol/L, whereas the concentrations were approximately 6.90–9.66 nmol/L for sows in conventional housing (converted values from ng/mL from Quesnel et al. [38]). This suggested that the maturity of the piglets at birth was positively affected by lower salivary cortisol during gestation. As the cortisol levels in the current study were not measured from the same sows the piglets were born from, the levels of cortisol were only indicative of the general level of stress in the different systems, with no direct connection to the observed litters. 5. ConclusionsFeeding systems during gestation seemed to have an effect on the stress level of sows, as reflected by their level of salivary cortisol. However, it was not possible to confirm that the differences in stress between systems had an impact on PBW or on the occurrence of sIUGR piglets at herd or litter level due to limitations in the study design. Further studies are warranted to investigate this relationship as this could potentially improve welfare and productivity in pig production systems in the future.

animals : an open access journal from mdpi

10.3390/ani12070916

PMC8997071

With the progress of society, the demand for meat is increasing; therefore, how to improve the efficiency of breeding, reduce environmental pollution, and reduce the cost of breeding have become urgent priorities. Abdominal fat is a redundant part of the production of chickens. Thus, reducing abdominal fat deposition is increasingly vital in chicken breeding. Studies have shown that G0S2 is a crucial gene in regulating fat metabolism, and its single nucleotide polymorphism (SNP) was significantly associated with chicken production traits in previous studies. The present study aimed to identify SNPs of the G0S2 gene and analyze whether they were associated with chicken carcass traits, including abdominal fat weight. The present study results can provide practical information for molecular marker-assisted breeding of chicken carcass traits.

Gene single nucleotide polymorphisms can be used as auxiliary markers in molecular breeding and are an effective method to improve production performance. G0S2 is a key gene involved in regulating fat metabolism, but little research has been conducted on this gene regarding its role in poultry. In this study, the specialized commercial partridge chicken strain G0S2 gene was cloned and sequenced, and the relationship between the SNP sites on G0S2 and the carcass traits of chickens was investigated. The results showed that a total of seven SNPs were detected on G0S2 (g.102G > A, g.255G > A, g.349C > T, g.384A > G, g.386G > A, g.444G > A, g.556G > A). Two sites are located in the coding region and five sites are located in the 3′-UTR. SNPs located in the coding region are synonymous mutations. g.444G > A has a significant correlation with abdominal fat weight. The chickens with AG and GG genotypes have the highest abdominal fat weight, while the AA genotype is lower. The g.102G > A genotype has a significant correlation with live and abdominal fat weight. The live weight and abdominal fat weight of the chickens with AA and AG genotypes are at a higher level and have a larger gap than the GG genotype. Chickens with the AA genotype in g.556G > A had the lowest fat weight. The results of present study can provide practical information for molecular marker-assisted breeding of chicken carcass traits.

1. IntroductionMeat is a good source of protein and energy for human beings. As the global population increases, the demand for meat is also rising [1]. However, at the same time, discussions on the environmental impact of the poultry industry and the efficiency of resource utilization have never stopped. Chicken meat is one of the most consumed meats. At the same time, compared with mammals, chicken production emits less waste. Developed countries such as the United States, the United Kingdom, Canada, and Japan closed live poultry trading markets, and poultry meat circulates in chilled products. In China, consumers prefer to buy live poultry for slaughter. However, the circulation of live poultry may lead to outbreaks such as influenza in birds. It is a general trend for poultry meat to be marketed chilled. Carcass traits are essential traits for chilled whole chicken products that can visually present the product to the consumer. However, carcass traits can only be measured after chicken slaughter, which is challenging to use for breeding production.Adipose triglyceride lipase (ATGL) is the rate-limiting enzyme that hydrolyzes triacylglycerol in adipocytes. The G0S2 gene is a regulator of the ATGL gene, which can directly bind to ATGL to inhibit the activity of ATGL and reduce the rate of lipolysis [2,3]. In cattle, researchers have verified that G0S2 is related to the fat content of cattle by detecting the expression level of G0S2 in the muscles and different intramuscular fat content in bulls [4]. G0S2 is highly expressed in pig liver, and to a lesser degree, in omental adipose tissue and suet fatty tissue in pigs. Moreover, two SNPs affect back fat thickness (BFT) [5]. In poultry, the function of G0S2 is the same as that in mammals. Adipose tissue in the body can act as an inhibitor of ATGL to regulate changes in lipolytic activity [6]. In the meantime, it has been reported that G0S2 knockout chickens can reduce abdominal fat deposition without affecting other traits, and the fatty acid composition in their blood and abdominal fat is changed [7].Skeletal muscle is an energy-consuming machine, and fat tissue acts as a buffer pool between energy intake and consumption, and when intake exceeds consumption fat is deposited [8]. G0S2 has been reported to inhibit ATGL activity in mouse and human skeletal muscle and plays an essential role in regulating lipid metabolism and substrate oxidation [9]. In mice, deletion of G0S2 resulted in a significant reduction in relative body weight gain and a significant increase in serum glycerol levels [10].Partridge chicken is one of the essential endemic varieties in China, famous for its good flavor [11]. The partridge chicken used in this experiment is a specialized commercial strain of broiler bred by KwangFeng Industrial Co., Ltd. (Guangzhou, China) and South China Agricultural University. It aims to ensure the meat quality and flavor while optimizing the skin color and other properties suitable for the chilled market. The results of this study may provide helpful information for GOS2 gene marker-assisted selection in chicken production.2. Materials and Methods2.1. Experimental Animals and Determination of Their Carcass TraitsThis study took local Chinese flocks of partridge chickens as the research object. All samples and carcass trait data were collected in Guangzhou KwangFeng Industrial Co., Ltd. Guangzhou KwangFeng Industrial Co., Ltd. is the animal experimental unit operated under South China Agricultural University. All animal experiments in this study were conducted in strict accordance with the guidelines of the Guide to Animal Welfare in China. These experiments (approval number: SCAU#2021F074) were carried out under the approval of the College Animal Science, South China Agricultural University (Guangzhou, China). We made every effort to reduce the suffering of animals.The G genealogy was constructed with 24 male spotted chickens and 86 female yellow chickens as the ancestors; the M3 genealogy was constructed with 76 male spotted chickens and 16 female yellow chickens. This experiment selected 389 F2 generation partridge chickens from G genealogy, and 327 F2 generation partridge chickens were selected from M3 genealogy. Seven hundred sixteen healthy genealogy chickens were used to measure their carcass traits. All chickens were raised in stepped cages to 90 days of age, and they were randomly given foot numbers and slaughtered according to standard procedures. Meanwhile, blood samples were collected and stored at −20 °C. The measured indicators included live weight, dressed weight, half-bore weight, full-bore weight, breast muscle weight, leg weight, wing weight, foot weight, head weight, heart weight, liver weight, stomach weight, abdominal fat weight, shin length, shin circumference (the circumference of the tape measured around the shin), and body oblique length. The measurement of chicken carcass traits and segmentation method referred to Introduction to Animal Husbandry and Chicken carcass segmentation [12,13]. Among them, the breast muscles, legs, wings, feet, shin length, and shin circumference were all measured on the right part of the chicken.2.2. DNA ExtractionWe used an E.Z.N.A.® SQ Blood DNA Kit (Omega, Georgia, GA, USA) to extract blood genomic DNA. The concentration of the extracted DNA was measured with a spectrophotometer (Thermo Fisher, Waltham, MA, USA), and the concentration was adjusted to 100–200 ng/μL with double distilled water.2.3. G0S2 Fragment AmplificationWe downloaded the G0S2 full length fragment on NCBI (https://www.ncbi.nlm.nih.gov, accessed on 20 March 2021), and then used NCBI Primer-BLAST (Primer designing tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/)) to design PCR amplification primers. The total volume of PCR was 30 μL, and the amplification system was 3 uL partridge chicken gDNA at a concentration of 100–200 ng/uL; 1.2 μL upstream primer (Tsingke Biotech, Beijing, China) at a concentration of 10nM; 1.2 μL downstream primer at a concentration of 10 nM; 9.6 μL double distilled water; 15 μL Green Taq Mix (Vazyme, Nanjing, China). The primer sequences of PCR were upstream primer 5′-GCTACACTAACGTGCCCCTC-3′, downstream primer 3′-TTACTGCCCACAGGCGTTC-5′. The PCR program was pre-denatured for the first 5 min, denatured at 95 °C for 15 s, followed by annealing at 58 °C, extension at 72 °C for 50 s (32 cycles), and final extension at 72 °C for 5 min at the end of the cycle. It was stored temporarily at 4 °C until off the machine. Then, the amplified samples were sent to Tsingke Biotech Co., Ltd. for sequencing.2.4. Statistics and AnalysisThe sequence peak map file returned by the sequencing was put into Seqman in DNASTAR Lasergene software (DNASTAR, Madison, WI, USA) to open, compared with the downloaded sequence, and the genotype of each SNP was recorded. We used Microsoft Excel (Microsoft Corp., Redmond, WA, USA) to calculate the allele frequency, genotype frequency, Chi-square value, Hardy–Weinberg p-value, theoretical heterozygosity, actual heterozygosity, PIC value, and purity of each SNP Degree. Haploview 4.1 was used to analyze the linkage disequilibrium of the collated data [14].Association analysis of SNPs (or haplotypes) and carcass traits were performed with mixed linear models and were established for each SNP and carcass trait using SPSS 26 software. The mixed linear model used is as follows:Y = μ + G + H + eY: phenotypic values of carcass traits, μ: the overall population mean, G: the fixed effect of genotype or haplotype, H: the fixed effect of hatch, e: random residual error. Correction for multiple comparisons was performed using the Bonferroni method.3. Results3.1. SNP and Genotyping of G0S2 GenePCR was used to amplify fragments of G0S2, which contained all coding regions and all fragments of the 3′-UTR. The amplified DNA fragments of 716 individuals were sequenced by second-generation Sanger sequencing, and the base sequence and base peak map of the G0S2 fragment of 716 individuals were obtained. After comparison and statistics, the sites with genotype frequency greater than 85% and allele frequency less than 1% were removed, and a total of seven SNP sites were screened out on this G0S2 segment for further research. Among them, two SNP sites are located on exons, and five SNP sites are located on the 3′-UTR. Two SNP sites located on exons did not change the coding nucleotides and belonged to synonymous mutations. We list the position information of the obtained SNP sites on the chromosome in Table 1. The peak maps of all three genotypes of SNPs are shown in Figure 1.Among the seven SNPs screened, three genotypes were detected by observing the statistical sequencing peak map. We analyzed the statistical results and calculated the genotype frequency and allele frequency of these seven SNPs, as shown in Table 2. We define the allele downloaded from the database as wild type and the mutant allele obtained by sequencing and screening as mutant type. In the five SNPs of g.102G > A, g.255G > A, g.349C > T, g.384A > G, and g.556G > A, the frequency of wild-type alleles is greater than that of mutant alleles. In addition, in the four SNPs of g.102G > A, g.255G > A, g.349C > T, and g.556G > A, the GG type (CC type) is dominant. In the SNPs of g.384A > G, the distribution of the three genotypes of GG, AG, and AA is relatively close. However, the frequency of wild-type alleles of g.386G > A and g.444G > A is less than that of mutant allele frequency; wild-type allele frequency is only 0.175 and 0.117, respectively. In g.386G > A and g.444G > A, the AA genotype is dominant. We also calculated and studied a series of genotype frequency and allele frequency parameters. We performed the Hardy–Weinberg Chi-square test on genotype frequency and allele frequency. The Hardy–Weinberg Chi-square test p-value shows that only the SNP of g.386G > A is in Hardy–Weinberg equilibrium. Except for g.386G > A, the genotype frequency and allele frequency of the remaining SNPs have a large gap between the actual observed value and the theoretically inferred value, and the p-value is less than 0.05, which is not in the Hardy–Weinberg equilibrium state. We surmise this may be the result of artificial breeding. Among the seven SNPs, g.349C > T, g.386G > A, and g.444G > A are low-level polymorphisms, and g.102G > A, g.255G > A, g.384A > G, and g. 556G > A are moderately polymorphic.3.2. Correlation Analysis of Chicken Carcass TraitsTo understand the relationship between various carcass traits in chickens, we calculated the Pearson product-moment correlation coefficient between the traits, as shown in Table 3. The results showed significant correlations among the live weight, dressed weight, half-bore weight, and full-bore weight, and the correlation coefficient was high. In addition to wing and stomach weight, abdominal fat weight was significantly correlated with other carcass traits. However, their correlation coefficients are all at a low level. This phenomenon suggests that abdominal fat needs to be considered an independent aspect in poultry breeding. Although breast muscle weight was significantly correlated with other carcass traits, its correlation coefficient was low. In addition, live weight was significantly correlated with full-bore weight, wing weight, foot weight, shin length, and body oblique length, and had a high correlation coefficient. We speculate that in this strain of chickens, the relationship between chicken body size and meat production rate is lower. The nutrients are used to increase body size, not muscle hypertrophy. In addition, interestingly, there was a significant negative correlation between abdominal fat weight and foot weight, head weight, and shin circumference.3.3. Association Analysis of G0S2 Gene SNPs and Carcass TraitsTo study the relationship between the screened SNPs and carcass traits, we performed an association analysis of GOS2 genotype information with 16 traits. The results of the association analysis are shown in Table 4. The results show that, except for head weight, breast muscle weight, foot weight, and body oblique length, g.102G > A, g.255G > A, g.384A > G, g.444G > A, and g.556G > A, with the remaining 12 traits such as live weight, carcass weight, and abdominal fat weight, are all at significant levels. It is worth noting that the AA genotype in g.102G > A has the highest level in these 12 traits. g.349C > T significantly correlates with live weight, dressed weight, leg weight, heart weight, and liver weight. G.386G > A is only significantly related to live weight, heart weight, liver weight, abdominal fat weight, and tibia length. Among these traits, the AG genotype is the dominant genotype. It should be noted that, except for g.349C > T, all other SNPs are significant with abdominal fat weight. Another point of importance is that the AG genotype in g.444G > A has the highest abdominal fat weight, and the AA genotype in g.556G > A has the lowest abdominal fat weight. Breast muscle weight is only associated with two SNPs, g.444G > A and g.556G > A. Among them, g.444G > A is significantly associated with breast muscle weight, and g.556G > A is significantly associated with breast muscle weight. The AG genotype in g.444G > A has the highest breast muscle weight. In addition, the head weight trait does not correlate with all SNPs.3.4. Linkage Disequilibrium Analysis and Haplotype Analysis of SNPs of G0S2 GeneThe results of linkage disequilibrium analysis for seven SNPs are shown in Figure 2. The LD map shows seven SNPs of the G0S2 gene, and three vital LD regions were detected.The analysis results of these three blocks show that block1 has five haplotypes, block2 has four haplotypes, and block3 has three haplotypes. The composition and frequency of the haplotypes are shown in Table 5. In order to study the relationship between these haplotypes and slaughter traits, we conducted a joint analysis. The joint analysis results are shown in Table 6 (we excluded haplotypes with a frequency less than 5%). Our results show that the data of the three haplotype combinations of H1H1, H6H6, and H10H10 are lower than those of other groups, and there are significant differences. We can pay more attention to the production and weed out this type during breeding.4. DiscussionThe world’s demand for meat continues to increase, and how to use fewer resources to produce more meat products has become a top priority. Carcass traits are important economic traits of the poultry breeding industry. They can be presented directly to consumers. At the same time, it is also a simple and direct manifestation of the economic benefits of breeding companies. In a pedigree named TT, live weight at 42 days has a high genetic correlation with carcass and carcass part weight but a low genetic correlation with organ weight [15]. Meanwhile, there are also related studies showing that higher body weights and prime cut traits have a higher genetic correlation [16]. Although the pedigrees in the studies mentioned above are different from our study, the conclusions are partially similar and support each other. According to the correlation results between our carcass traits, it is more consistent with the conclusions of the studies mentioned above, that is, the live body weight has higher correlation with half-bore weight, full-bore weight, breast muscle weight, thigh weight, wing weight, and foot weight, but lower correlation with each organ.Thus far, many studies have found that genome-wide SNP screening, genome-wide haplotype screening, quantitative trait locus (QTL), and SNPs can have a certain impact on the improvement of poultry traits or the improvement of production performance [17,18,19,20,21,22]. However, compared with livestock and other mammals, the improvement of poultry carcass traits started late. Therefore, it is very important to have more research results to enrich the breeding strategy to improve the characteristics of poultry meat carcasses.In the previous study of our team, we found four SNPs in G0S2: g.102G > A, g.111C > T, g.197G > A, and g.255g > A. Association analysis with production traits showed that g.102G > A and g.255G > A were significantly associated with head width and leg muscle color. g.102G > A was significantly associated with shank diameter at 63 days. g.197G > A was significantly associated with shank diameter at 49 days, the crude protein content of leg muscle, and other traits [23]. The two sites, g.102G > A and g.255G > A, are the same as the coding region in this study. However, the traits associated with the two studies were not the same. The previous studies mainly focused on production traits, while this study focused on carcass traits. The selection of carcass traits can be more directly reflected in the slaughtered chickens, which fitted with the trend of unified slaughtering and chilled for sale. Compared with previous studies that only detected SNPs in coding regions, this study detected five SNPs in non-coding regions. SNPs in non-coding regions also significantly affect transcriptional efficiency, gene expression, and splicing dysregulation [24,25].Through previous studies, G0S2 is considered one of the key regulatory genes in fat metabolism [24,26]. Studies have shown that G0S2 directly binds to ATGL to reduce the hydrolysis of fat [25]. In the previous study of our team, there were similar results. Knockdown of G0S2 can activate ATGL, thereby promoting triacylglycerol hydrolase activity. At the same time, it affects the FABP4 gene, thereby inhibiting PPARγ and further reducing adipogenesis [27]. Poultry’s abdominal fat is currently considered a by-product of no economic value. Excessive deposition of abdominal fat will affect the economic benefits of production. Abdominal fat deposition and muscle weight gain are related [28,29,30]. Therefore, we believe that the use of G0S2, a key gene in fat metabolism, as a molecular marker for carcass traits, assists in improving poultry carcass traits. It has a certain meaning in production. Although some laboratories have been able to breed G0S2 knockout chickens, the breeding process usually takes two years or more. It takes a long approach to knock out genes from the laboratory to production and use, and the use of molecular markers for selection is a better choice.In summary, in this study, partial fragments of poultry G0S2 gene were amplified by PCR, and seven SNPs were obtained (g.102G > A, g.255G > A, g.349C > T, g.384G > A, g.386G> A, g.444G > A, g.556G > A). In addition, the relationship between the G0S2 gene polymorphism and 16 carcass traits was analyzed. Association analysis found that g.444G > A has a significant correlation with abdominal fat weight. The AG and GG genotype have the highest abdominal fat weight, while the AA genotype is lower. The g.102G > A genotype has a significant correlation with live and abdominal fat weight. The live weight and abdominal fat weight of the AA and AG genotypes are at a higher level and have a larger gap than the GG genotype. The AA genotypes of g.102G > A, g.255G > A, and g.384G > A are significantly different from the GG genotype, and the AA genotype had significantly higher values of the traits than the GG genotype. However, g.386G > A, g.444G > A, and g.556G > A have opposite trends. In addition, the three haplotypes of H1H1, H6H6, and H10H10 are also worthy of our attention in production. Therefore, this study finds that the seven SNPs, g.102G > A, g.255G > A, g.349C > T, g.384G > A, g.386G > A, g.444G > A, and g.556G > A, can be used as additional molecular markers for carcass improvement, providing a theoretical basis for poultry breeding.5. ConclusionsIn summary, this study detected seven SNPs of the G0S2 gene and analyzed the relationship between their polymorphisms and chicken carcass traits. At the same time, the correlation between each carcass trait was investigated. The results indicated that abdominal fat weight and breast muscle weight were significantly correlated with other carcass traits, but the correlation coefficients were all at low levels. Chickens with AA genotype in g.102G > A had the highest level among 12 traits, except for head weight, breast muscle weight, foot weight, and body oblique length. Chickens with the AG genotype in g.444G > A had the highest abdominal fat weight, while chickens with the AA genotype in g.556G > A had the lowest. These SNPs can be used for marker-assisted selection to improve carcass traits in partridge chickens. This will be helpful in breeding programs aimed at improving chicken carcass traits.

animals : an open access journal from mdpi

10.3390/ani11082310

PMC8388390

The commercial canine ehrlichiosis vaccine is still not available. A limitation of developing vaccines is the knowledge of the humoral immune response against E. canis infection is still unknown. In this study, we designed the peptide vaccine candidate, namely GP194-43, for inducing hyperimmune serum in rabbits. The rabbit sera were used to examine for E. canis infective inhibition in macrophage-like cells (DH82). A decrease in E. canis infection (<50% of infection) was observed on DH82 cells in the treated group with GP194-43 antiserum on the third day of the post-infection period. Cytokine genes in DH82 with/without rabbit sera were investigated and showed marked up-regulation of the IFNG expression level in DH82 cells in the treated group with GP194-43 antiserum. This study provides the preliminary information of immune response against E. canis of immunized animals and directions for genomic or/and proteomic studies involved in phagocyte-cell mediated immune response.

In terms of its veterinary importance, vaccine development against Ehrlichia canis is needed. However, the effect of developing vaccines on humoral immune response against E. canis infection is still unknown. Novel GP194-43 was synthesized according to E. canis GP19 epitope prediction. To restrict any loss and/or illness in the host animal, rabbits were used in this study to produce GP194-43 hyperimmune sera. The effect of GP194-43 hyperimmune sera on neutralization was examined in vitro by determining the inhibition of E. canis infection of the macrophage-like cell line (DH82) in the presence of the sera. Four groups of DH82 cells received differing treatments. These included E. canis experimentally infected DH82 cells, E. canis-infected DH82 cells with control rabbit serum (untreated group), E. canis-infected DH82 cells with GP194-43 rabbit antiserum (treated group) and uninfected cells (negative control group), respectively. The treated group developed a decrease (p < 0.01) in the percentage of E. canis infected cells after 3 days post-infection at 48.57 ± 1.28. In addition, real-time PCR analyses of cytokine mRNA expression involved with the macrophage, humoral, and cellular immune responses were conducted. The findings revealed an upregulated expression of IFNG in the treated group during the infection. This study demonstrated neutralization in the GP194-43 peptide hyperimmune sera of immunized rabbits. Notably, IFN-γ production could be effectively promoted in canine macrophages in relation to the activation of macrophages and adaptive immune responses. The results of this study indicate the potential for the use of this immunogen in further investigations involving immunized and infected dogs as E. canis host species.

1. IntroductionCanine monocytotropic ehrlichiosis (CME) is an important infectious disease occurring in dogs. It is commonly transmitted by the brown dog tick, Rhipicephalus sanguineus [1,2]. The etiological agent of this disease is Ehrlichia canis, a small, gram-negative, obligatory intracellular bacterium in the family Anaplasmataceae [3]. The agent invades and propagates in monocytes and macrophages, which can then lead to a variety of hematological and clinical signs [4,5]. The acute phases of CME are characterized by high fever, depression, anorexia, lymphadenomegaly, and splenomegaly, as well as clinical signs that include dermal petechiae, ecchymosis, and epistaxis [6]. Thrombocytopenia is the most prominent hematological change that occurs in the acute phase [7]. The chronic phase is characterized by pancytopenia due to suppression or destruction of bone marrow [8]. The clinical signs in the chronic phase are similar to those of the acute phase but are associated with a greater degree of severity [6]. E. canis infections in dogs have been reported in North and South America, Europe, Africa, and Asia [9,10,11,12,13]. Additionally, there has been evidence of E. canis infections in humans in certain specific regions [14].Vaccines for E. canis are needed; however, many obstacles have impeded their development including identification of ehrlichial antigens, an understanding of the relevant genetic and antigenic variabilities, and a lack of animal models that reflect the immune responses of the hosts [15]. Previous studies have shown that inactivated and attenuated vaccine candidates for CME were capable of provoking a humoral response, but only partial clinical protection was achieved in dogs challenged with the virulent strain [16,17]. The peptide vaccine is among a variety of modern vaccines that represent a potential strategy for the prevention and treatment of pathogenic diseases [18,19,20]. There are a number of noteworthy advantages associated with this vaccine relevant to its low cost, ease of synthesis, and inherent level of safety. All of these attributes are considered extremely attractive features [21]. In the case of Ehrlichia spp., recombinant P29 protein and P28-19 peptides have been developed and demonstrated an ability to protect against E. muris in mouse models [18,22]. Since there is no available vaccine against E. canis, evidence of the role of the humoral immune response in immunized dogs, compared with those that have been infected, is limited.Ehrlichial antigenic proteins have been selected based on the degree of reactivity of the proteins with antibodies obtained from infected animals [23]. Major immunodominant E. canis antigens can be recognized early in the infection period as 19-, 37-, 75-, and 140-kDa proteins [24,25]. The 19-kDa protein has been identified as glycoprotein (GP19) and has revealed orthologs with variable-length PCR target proteins (VLPT) of E. chaffeensis [26]. E. canis GP19 consists of a serine/threonine/glutamate (STE)-rich patch at the amino-terminal that contains species-specific antibody epitopes that were strongly recognized by the serum collected from an E. canis-infected dog [26]. Additionally, GP19 was found to be one of the most conserved proteins (98.8–100% identities) among the geographically dispersed E. canis strains [27,28,29]. Due to the fact that a high degree of conservation and immunoreactive ability is attributed to E. canis GP19, this antigen has been recognized as an interesting target for diagnostic and vaccine development.In this study, the GP19 epitope of E. canis was predicted. Subsequently, the GP194-43 peptide was synthesized to produce hyperimmune serum to examine the capability of neutralization against E. canis infection in vitro. The canine macrophage-like cell line (DH82) that was isolated from a dog with histiocytic sarcoma was representative of permanent macrophage morphology and excellent phagocytic cells. Furthermore, this cell line has been reported for use in the successful culture and infection of E. canis [30,31]. In addition, cytokine mRNA expression levels involved with macrophage, humoral and cellular immune responses were evaluated. Different types of cytokines have been discovered including interferons (IFN), interleukins (IL), and the tumor necrosis factor (TNF) [32,33]. Expression of the cytokine subunit genes, particularly IFN-γ (IFNG), IL4 (IL4), IL-10 (IL10), IL-12 subunit beta (IL12B), IL-13 (IL13), and TNF (TNF), were investigated in this study to gain a greater understanding of the immunomodulatory action of the GP194-43 antiserum against E. canis infection.2. Materials and Methods2.1. Construction of GP194-43 PeptideE. canis GP19 epitope sequences were determined based on protein conformation flexibility, hydrophilicity, and protein conformation. The prediction of the GP19 epitope containing the region was performed using Antibody Epitope Prediction-IEDB Analysis Resource (http://tools.immuneepitope.org/bcell/, accessed on 12 April 2020). A 40-amino-acid epitope sequence at the N-terminal region containing the STE-rich patch of E. canis GP19 was selected and then labeled as the GP194-43 peptide. The 3D structure of the GP194-43 peptide was predicted using Phyre2 protein modeling, prediction, and analysis [34] (Figure S1). The GP194-43 peptide sequence was synthesized (Bio-Synthesis, Inc., Lewisville, TX, USA), resuspended in molecular biology grade water (0.8 mg/mL), and then kept at −20 °C until being used in the process going forward.2.2. Production of Rabbit Antiserum against GP194-43Rabbit GP194-43 antiserum was generated against the GP194-43 peptide. New Zealand white rabbits were divided into 2 groups based on relevant immunogen formulations (n = 2 rabbits each group). The first group was comprised of two rabbits that were immunized with 100 µg of the GP194-43 peptide/500 µL with an equal volume of Montanide (SEPPIC, Paris, France). The rabbits in the second group received only 500 µL of Montanide (control serum group). The immunized rabbits were intramuscularly administered 4 times at 2-week intervals. All rabbits were observed for clinical signs and behavioral changes that may have occurred at the pre- and post-immunization stages. These animal experiments were approved by the Animal Care and Use Committee (ACUC), approval No. S27/2560.2.3. Evaluation of Rabbit Antibody against GP194-43 Using ELISAAn indirect enzyme-linked immunosorbent assay (ELISA) was performed to monitor the specific rabbit sera IgG against the GP194-43 peptide of E. canis. Briefly, 96-well immunoplates (Nunc-Immuno Plate MaxiSorp, Intermed, Roskilde, Denmark) were coated with the synthetic GP194-43 peptide (1 ng/well) and then incubated overnight at 4 °C. The plates were washed three times with phosphate buffer saline (PBS) containing 0.05% Tween-20 (PBST) and then blocked with blocking buffer (5% skim milk in PBS) for 1 h. A volume of 100 μL of GP194-43 immunized and control rabbit sera were added to each well; they were then incubated at 37 °C for 1 h. The plates were developed with secondary goat anti-rabbit IgG horseradish peroxidase (HRP) conjugated antibody (Dako Cytomation, Glostrup, Denmark) and incubated at 37 °C for 1 h. Subsequently, the 3,3′,5,5′-tetramethylbenzidine (TMB) substrate (SeraCare Life Sciences, Gaithersburg, MD, USA) was added to each well, and they were then incubated at room temperature for 15 min away from light. The reaction was stopped by adding 50 µL of 2 N H2SO4. Optical density (OD) was measured spectrophotometrically at 450 nm using an Accu Reader Microplate reader M965 (Metertech, Taipei, Taiwan R.O.C.). The cut-off value was determined following the formula employed in the previous study; cut-off value = mean negative control +3 standard deviations (SD) [35].Rabbits were euthanized to collect blood samples one week after the last round of immunizations according to the AVMA Guidelines for the Euthanasia of Animals: 2020 Edition [36]. Blood samples were centrifuged to separate the sera and the pooled rabbit sera from each group were then stored at −20 °C until being used.2.4. Canine Macrophage-Like Cell Lines and Culture ConditionsCanine macrophage-like cells (DH82, ATCC® CRL-10389TM) were maintained in Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) that had been supplemented with 10% heat-inactivated fetal bovine serum (FBS; Gibco, Thermo Fisher Scientific, Waltham, MA, USA) in 25-cm2 tissue culture flasks at 37 °C and 5% CO2 in a humidified incubator. Five milliliters) in 25-cm2 tissue culture flasks at 37 °C and 5% CO2 in a humidified incubator. Five milliliters of cell-culture media were changed every 3–4 days. The cells were subcultured when about 90% confluence was reached; they were then observed under an inverted microscope (CK40; OLYMPUS, Olympus Optical, Tokyo, Japan). Cell detachment was performed using 1 milliliter of 0.25% Trypsin-EDTA, phenol red, 1X (Gibco, Thermo Fisher Scientific).2.5. E. canis Culture and Stock ConditionsE. canis BF W053712X + 5 (PTA-5811™, ATCC®, Manassas, VA, USA) was used in this study and cultured in the 4th passage of DH82 according to the method employed in previous studies [4,37]. The monitoring of E. canis growth was performed at 1-week intervals by Giemsa staining via microscopic examination. Cells were propagated continuously until a level of 80% of infected cells was reached. Bacterial collection was performed manually by applying the previously described method [4]. Briefly, the approximate number of 5 × 106 DH82 cells that were infected with E. canis were detached using 1 mL of 0.25% Trypsin-EDTA; they were then incubated for 15 min at 37 °C. An equal volume of DMEM was added to the flask and the cell suspension was passed 10 times through a bent 26 G needle to rupture the DH82 cells and release the intracellular bacteria mechanically. The suspension was centrifuged for 5 min at 1500× g. The supernatant containing cell-free bacteria was collected and the total volume was then adjusted to approximately 1 mL per 1 × 106 cell-free Ehrlichia. The cell-free Ehrlichia was used in invasive detection and cytokine gene investigation in the next process. A total of 1 mL of the 1 × 106 E. canis-infected DH82 cells were collected in 2-mL cryopreserved tubes with 5% dimethyl sulfoxide (DMSO) in DMEM and were then frozen at −80 °C until being used.2.6. Infective Inhibition of Ehrlichia canis in DH82 Cell Line2.6.1. Infective Detection Using Light MicroscopyDetection of infective inhibition using light microscopy was modified from the method that had been previously described [38]. Briefly, DH82 cell viability and number were checked with 0.4% Trypan Blue. The approximate proportion of viable cells based on cytospin was found to be over 95%. The number of DH82 cells was adjusted to 1 × 105 cells/mL and the cells were then deposited into 33 mm corning® culture dishes (Corning Inc., Corning, NY, USA) and fixed with 22 × 22 mm coverslips. The cells were placed at the bottom of the wells and they were then incubated at 37 °C and 5% CO2 for 24 h to achieve cell adhesion. The dishes were then washed 3 times with 2 mL sterile 1 × PBS (pH 7.4) and used for E. canis invasive detection.The treatment group was divided into 4 groups according to the different treatments. These different treatment groups included E. canis-infected DH82 cells (positive control group), E. canis-infected DH82 cells with control serum (untreated group), E. canis-infected DH82 cells with GP194-43 antiserum (treated group), and uninfected DH82 cells (negative control group), respectively. A volume of 100 µL of 1 × 105 cell-free Ehrlichia was added to 400 µL of the cell media (positive control group), 400 µL of control serum (untreated group), and GP194-43 antiserum (treated group). The specimens were then incubated at 37 °C and 5% CO2 for 1 h. Then, the suspension was added to the monolayer of DH82 cells that were adjusted to 1 × 105 cells in the dishes, and they were incubated at 37 °C and 5% CO2 for 24 h. Additionally, 1 × PBS solution was used as a negative control (uninfected DH82 cells). After 24 h, non-adherent bacteria were removed by washing the specimens twice with 2 mL of sterile 1 × PBS solution. After being washed, the coverslips were fixed in 4% formaldehyde solution and stained with Diff-Quik staining (Sigma-Aldrich, St. Louis, MO, USA). The experiments were performed in duplicate for each group. The coverslips were then observed under a light microscope.2.6.2. Infective Detection Using an Immunocytochemistry (ICC) TechniqueThe detection of E. canis infected cells was achieved following the ICC method applied in the previous study [39]. The number of DH82 cells were adjusted to 1 × 105 cells per ml. Furthermore, 1 × 104 DH82 cells in 100 µL cell culture media were cultured in duplicated 96-well plates for 24 h in a humidified incubator at 37 °C and 5% CO2 to allow for cell adhesion. The wells were then washed 3 times with 100 µL PBS/0.1% Tween-20 (PBST; pH 7.4) and used for invasive detection.The treatment group was divided into 4 groups as has been mentioned above. For the positive control, and the untreated and treated groups, 100 µL of 5 × 104 cell-free Ehrlichia was added to 200 µL of cell media, 200 µL of control serum, and GP194-43 antiserum, respectively. The specimens were then incubated at 37 °C for 1 h. In the negative control group, PBS solution was used instead of Ehrlichia. Additionally, 100 µL (1 × 104 cell-free Ehrlichia) of the suspension was added to the monolayer of DH82 cells in the media and incubated at 37 °C with 5% CO2 for 1 or 3 days. The media were then discarded, and non-adherent bacteria were removed by washing the bacteria three times with 100 µL of PBST. The cells were fixed with cold 4% paraformaldehyde for 15 min. Endogenous peroxidase was blocked using 3% H2O2 in 1 × PBS for 10–30 min. The cells were then blocked with 3% bovine serum albumin (BSA; AMRESCO, Solon, OH, USA) and washed three times with PBST. A solution comprised of 1:200 dilution of the primary antibody (rabbit GP194-43 antiserum) was added to the plates and they were incubated at 4 °C overnight. The specimens were then washed once with PBST and incubated in a solution of 1:400 dilution of horseradish peroxidase (HRP)-conjugated goat anti-rabbit IgG (Sigma-Aldrich) for 1 h at RT. The plates were washed three times with PBST and then 75 µL of stable DAB was added (Invitrogen™, Carlsbad, CA, USA) for 2–5 min before being rinsed with tap water. The results were observed under an inverted microscope (400X, CK40; OLYMPUS, Shinjuku City, Tokyo, Japan) and infected cells were stained reddish-brown. A semi-quantitative scoring system for the stained cells was adopted from the system that had been previously described [40]. The average percentage of E. canis infected cells that were stained reddish-brown using the ICC technique was determined manually in 3 random fields and were then scored as follows; 0 = negative, 1 = 10% of the cells with positive staining, 2 = between 10–50% of the cells with positive staining, 3 = more than 50% of the cells with positive staining.2.7. Cytokine Gene Expression ProfilesTo investigate the effects of the GP194-43 peptide on cytokine expression, a volume of 100 µL of 1 × 105 cell-free Ehrlichia was added to 400 µL of the cell media (positive control group), control serum (untreated group), or GP194-43 antiserum (treated group). The specimens were then incubated at 37 °C and 5% CO2 for 1 h. The suspension was added to the monolayer of 1 × 106 DH82 cells in duplicate 24-well plates. In the negative control group, PBS solution was used instead of Ehrlichia. Plates were incubated at 37 °C, 5% CO2 for 1 or 3 days. After incubation, samples from each well were combined to achieve a cell density of 1 × 106 cells. RNAlater® solution was then added to the cells to preserve total RNA. RNA extraction was performed using a Total RNA Extraction Kit (RBCBioscience, Taipei, Taiwan) following the manufacturer’s instructions. RNA yields and concentrations were measured using a DU 730 nanoVette UV/Vis Spectrophotometer (Beckman Coulter, Brea, CA, USA). For cDNA synthesis, 2 μg of total RNA were used as a starter in 20 μL of a reaction mixture containing 2 μM Random Hexamer, 2.5 mM deoxynucleotide triphosphates (dNTPs), 5 × RT buffer, 40 U RNase inhibitor (Invitrogen™, Carlsbad, CA, USA), RNase-free water and 200 U BioscriptTM reverse transcriptase (Bioline, Memphis, TN, USA). The cDNA synthetic reaction was carried out in a T100™ thermal cycler (Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s instructions.Approximately 50 ng of the cDNA samples were quantitatively analyzed for the mRNA transcription of gene coding that involved interferon-gamma (IFNG), interleukin 4 (IL4), interleukin 10 (IL10), interleukin 12B (IL12B), interleukin 13 (IL13), and tumor necrosis factor-alpha (TNF), while the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a reference gene. Real-time RT-PCR was performed using a SensiFAST SYBR Hi-ROX Kit (Bioline, London, UK) following the protocol described by the manufacturer and by using CFX96 Touch™ Real-Time PCR (Bio-Rad, Hercules, CA, USA). Specific information regarding primer pairs used in this study is presented in Table S1. Sequences of the primers were designed by Primer3plus; primer synthesis was performed by Integrated DNA Technologies (IDT; Coralville, IA, USA). After gradient-PCR, the real-time RT-PCR cycling conditions were applied according to the initial denaturation at 94 °C for 5 min followed by 40 cycles at 94 °C for 30 s, 59 °C for 10 s and 72 °C for 20 s, ending with a final extension at 72 °C for 5 min for all genes. After completion of the assessment of the real-time RT-PCR cycles, dissociation curves were analyzed to confirm that the PCR product obtained by real-time RT-PCR included the correct components. The specificity of real-time RT-PCR on the specified genes was confirmed by 2% agarose gel electrophoresis. Analysis of relative gene expression was calculated from the Ct of the target gene and the reference gene (GAPDH). The expression levels (fold-difference) were reported using the 2−ΔΔCt method [41].2.8. Data AnalysisThe data from all three experiments were combined and analyzed to identify any outliers using the robust regression and outlier removal (ROUT) method. The normality test was completed using the Shapiro-Wilk test prior to performing statistical analysis. Statistical analyses were performed using one-way ANOVA or multiple t-tests that employed GraphPad Prism 8.2.0 (GraphPad Software, Inc., San Diego, CA, USA). Results of the statistical analyses were considered significant in all experiments when p < 0.05. Multiple comparisons of each pair were made using the Holm–Sidak method. Results were reported as mean plus standard error (SE). The cytokine genes obtained from differential expression data were used to create a network of genes using CytoScape with a GeneMANIA plugin [42].3. Results3.1. Rabbit Antibody Response to GP194-43 PeptideAll immunized and non-immunized rabbits displayed no clinical signs and behavioral changes at the pre- and post-immunization stages. The OD values of the immunized and control rabbit sera are shown in Figure 1. Following immunization of the rabbits, two peaks of antibody levels against the GP194-43 peptide were detected 14 days after the first immunization and 7 days after the third immunization. The average OD values and SD of immunized rabbit sera weekly recorded at 0.240 ± 0.025, 1.661 ± 0.065, 2.037 ± 0.012, 1.960 ± 0.013, 1.885 ± 0.005, 2.202 ± 0.010, 1.982 ± 0.043 and 2.191 ± 0.022, respectively. The average OD values and SD of non-immunized control rabbit sera were 0.271 and 0.038, respectively. Thus, the cut-off point of the indirect ELISA detecting rabbit antibody against GP194-43 of E. canis was 0.385.3.2. Infective Inhibition of E. canis to DH82 Cell Line3.2.1. Positive E. canis Infection Detected by Light MicroscopyWith the exception of the negative control group, E. canis cells infected with microorganisms within the membrane-bound compartments (morulae) were observed in the DH82 cytoplasm under light microscopy. This was attributed to a heavy degree of infection (1 × 105 cell-free Ehrlichia) on day 1 of the post-infection period.3.2.2. E. canis Infective Inhibition Detected by ICC TechniqueWith the exception of the negative control, infection of E. canis was graded with a score number of 3 (>50% positive cells) in all groups on day 1 of the post-infection period (Figure 2A). The percentages of E. canis infected cells in the positive control, and the treated and untreated groups, were 90.925 ± 1.178, 90.065 ± 2.772, and 92.918 ± 1.459, respectively (Figure 2B).Notably, stimulated DH82 indicated different percentages of E. canis infected cells on the third day of the post-infection period (p < 0.01, Figure 2B). Furthermore, the treated group revealed lower percentages of positive staining cells with DAB staining (10–50% positive infected cells; score number 2. The percentages of E. canis infected cells in the positive control, treated and untreated groups were 92.291 ± 2.269, 48.573 ± 1.282, and 90.967 ± 3.132.3.3. Relevance of GP194-43 Antiserum on Cytokine Gene ExpressionGene expression levels in the E. canis-infected group revealed overall down-regulation among cytokine genes IFNG, IL10, IL12B, and IL13 (Figure 3). In contrast, IL4 displayed a degree of upregulation (p < 0.01) when in the E. canis-infected group compared with all groups throughout the post-infection phase of the experiment (Figure 3B).IFNG gene expression was maintained following E. canis infection in the presence of GP194-43 antiserum with a degree of up-regulation of IFNG on day 3 (p < 0.001) when compared with uninfected and E. canis-infected group, likewise compared with E. canis infection in the control serum (p < 0.01) (Figure 3A). IL13 gene expression in the treated group revealed significantly downregulated levels of expression when compared to the uninfected group throughout the experiment (p < 0.01) (Figure 3E). The levels of gene expression of the other cytokine genes involved with the helper Th response (IL4, IL10, and IL12B) and the pro-inflammatory cytokine gene (TNF gene) in the treated group revealed no significant differences in the expression levels when compared to the uninfected group (Figure 3).4. DiscussionDespite the importance of veterinary medicine and the possibility of zoonotic pathogens affecting humans, very few studies have been conducted on vaccine development against E. canis infection. GP194-43 peptide serves as an immunogen in that it mimics the antigen structure, GP19 of E. canis, which can then elicit the antibody of infected dogs within 14 days. This model was constructed and used to produce rabbit GP194-43 antiserum in the present study [21,43]. Additional immunizations (two-week intervals) were given to raise the concentration of the antibodies. After the 3rd round of immunizations, immunized rabbit antibody titers had risen to a peak within a week of the post-immunization period. Rabbit GP194-43 antiserum containing a specific antibody was used to determine the effects of the GP194-43 antiserum to macrophages against E. canis infection in the canine macrophage. The 4th passages of DH82 cells were representative of canine macrophage-like cells that could be used to control relevant cell characteristics, including cell infection and gene expression [30,44].As was the case in previous studies, the passive transfer of serum containing Ehrlichia-specific immunoglobulin G (IgG) to E. chaffeensis-infected SCID mice pre- and post-infection displayed an effective degree of elimination of Ehrlichia [45,46]. Our present study reinforced the role of GP194-43 antiserum against E. canis and showed the effects of elimination of Ehrlichia in canine macrophages on the third day of the post-infection period. Results of the current experiment suggest the possible importance of the following mechanism. Macrophages could be triggered by GP194-43 antiserum with regard to the neutralizing activity against E. canis infection by a decrease in the number of positive infected cells by the third day of the post-infection period. Since this study focused on the early infection of E. canis, the effects of inhibition of E. canis infection of the macrophage-like cell line (DH82) in the presence of the antisera are still unknown in the context of long-term infection. Further studies on the neutralization role of the antibody against GP19 immunogens and E. canis infection, and between early and long-term infection, are needed to understand the effect of vaccine candidates against E. canis infection.The expression of cytokine genes involved in macrophage-adaptive immune responses in DH82 cells was unraveled in this study. With the absence of GP194-43 antiserum, the E. canis infected DH82 cells revealed significantly downregulated levels of expression of IFNG and IL12B genes that involved with macrophage-Th1 cells. These results were consistent with the findings of a previous study that had reported a low level of detection of IL-12p70 and IL-12p40 in mice spleen during virulent IOE lethal infection [47]. Contrastingly, the IFNG expression of E. canis infected cells in the presence of the GP194-43 antiserum (treated group) displayed significantly up-regulated levels on day 3 when compared to all groups (Figure 3A). A decrease in E. canis infection was also observed on the third day of the post-treatment period; accordingly, the IFNG expression in the treated cell group could be related to the elimination of the intracellular microorganism, E. canis. Moreover, IFNG is the only cytokine gene investigated in this study that shows significantly different expression between the treated and untreated groups. The findings of the present study indicate the strong possibility that GP194-43 antiserum was able to promote IFN-γ production in DH82 cells against E. canis infection. In addition, the outcomes of the present study support the use of IFN-γ as an indicator of protective immunity in animals immunized with Ehrlichia immunogens, as was evident from the findings of the previous studies [48,49].The cytokine gene expression levels that were involved with macrophage-Th2 and B cells, including IL4, IL10, and IL13 genes in DH82 cells, were also investigated. The results displayed no difference of these genes in uninfected cells, E. canis infected DH82 with and without GP194-43 antiserum. Notable, the E. canis infected DH82 cells revealed upregulated levels of expression of IL4 which reported a negative relationship with IFNG expression. With regard to vaccine development against Ehrlichia spp., the previous study revealed that the killed vaccine of E. ruminantium elicited both CD4+ and CD8+ subsets to produce IFN-γ in the absence of IL4, indicating a type 1 response [50]. However, the relation of these two genes has not been reported in canine macrophages. Investigation of the macrophage-adaptive immune response in animals immunized with the GP194-43 peptide is further needed for a clearer understanding of the protective ability.The overproduction of IL-10 and TNF-α was reported to be associated with toxic shock-like syndrome and mortality in animal models of fetal monocytic ehrlichiosis, according to the findings of previous studies [47,51,52,53]. In the present study, there were no significant differences in IL10 and TNF gene expression levels among E. canis infected cells (positive, treated, and the untreated group). Further studies are recommended on the expression of the anti-inflammatory and pro-inflammatory cytokine genes in various cells or animal models over a longer period of time to reach a definitive conclusion on the effects of the antiserum on these genes.Data on the gene expression levels for both up- and down-regulation trends were used to analyze the network of relationships using GeneMANIA plug-ins [54]. The results revealed evidence of a relationship among query genes in this study (IFNG, IL4, IL10, IL12B, IL13, and TNF), as well as among the genes in the public dataset of functional association networks. Predictions of known gene co-expression and physical interactions in the cytokine activity pathway with query genes (brown nodes), leukocyte differentiation (blue nodes), leukocyte proliferation (lilac nodes), lymphocyte proliferation (violet nodes), and inflammatory responses (yellow nodes) were defined by the presence of additional genes that are related to the individual query genes (Figure 4). In addition, some genes were indirectly present in the network, namely CSF1, IL10RA, DGKA, IFNGR1, IFNGR2, IL21R, IL24, IL4R, and FLT3LG. Our analyses of the cytokine networks suggest the possible linkages of the cytokine genes identified in this study with additional cytokines involved in the E. canis infection responses with the collaboration of other macrophages and other immune cells. Therefore, cytokine detection in different immune cells, including macrophages, CD4+ and CD8+ T cells, observed in animal models would be needed to better understand the role of the GP194-43 peptide as an immunogen in the development of a vaccine against E. canis.5. ConclusionsSynthetic GP194-43 peptide was used as an immunogen to produce rabbit GP194-43 hyperimmune serum in the present study. The GP194-43 antiserum was examined for E. canis infective inhibition in macrophage-like cells (DH82). A decrease in ehrlichial infection was observed in the GP194-43 antiserum treated group when compared with the positive control and the untreated group on the third day of the post-infection period, as was observed through the use of the ICC method. Cytokine genes involved in macrophage-adaptive immune responses were also investigated. The present findings highlight the expression level of IFNG displaying marked up-regulation in the treated group. These findings can fill the existing gaps in terms of the recorded knowledge of immune response against E. canis infection in both treated and untreated macrophages with GP194-43 antiserum containing the specific antibody. This study also indicates potential directions for future studies on the genes involved in phagocyte-cell mediated immune response. Notably, the relative macrophage, humoral and cellular immune response in both in vitro and in vivo studies will be needed for further verification.

animals : an open access journal from mdpi

10.3390/ani11092477

PMC8470321

Current estimates have determined that one in five reptile species are already threatened with extinction. The aim of this study was to determine if pregnant mare serum gonadotropin could be used in leopard geckos (Eublepharis macularius) to stimulate the production of sperm and increase their testosterone concentrations. This information may aid in the development of breeding programs for endangered gecko species in the future that will be useful for conservation efforts. Our results demonstrated that this hormone did stimulate sperm production and increased testicular size in leopard geckos. However, it did not increase testosterone concentrations under the current conditions between October–December in the Northern hemisphere. Ultimately, future studies are needed to further characterize the annual reproductive cycle of leopard geckos.

Reptiles are highly susceptible to anthropogenic activities as a result of their narrow geographical ranges and habitat specialization, making them a conservation concern. Geckos represent one of the mega-diverse reptile lineages under pressure; however, limited assisted reproductive technologies currently exist for these animals. Exogenous pregnant mare serum gonadotropin (PMSG) has been found to exhibit follicle stimulating hormone-like action and has been routinely used to alter reproductive hormones of vertebrates in assisted reproductive protocols. The purpose of this study was to determine the effects of serial injections of 20 IU and 50 IU PMSG on circulating testosterone concentrations, testicular dynamics, and semen production in a model species of gecko. Twenty-four captive-bred, adult, male leopard geckos (Eublepharis macularius) were divided into three treatment groups and administered a once-weekly injection of either PMSG or saline for a total of nine weeks. Ultrasonographic testicular measurements, electrostimulation for semen collection, and venipuncture were performed on days 0, 21, 42, and 63. Right unilateral orchidectomies and epididymectomies were performed in all animals on day 63; tissues were submitted for histopathology. PMSG treated geckos had significantly higher testicular volumes and weights, spermatozoa motility, and spermatozoa concentrations compared with controls. However, there were no significant differences in testosterone concentrations by treatment or time. Under the conditions outlined, PMSG is effective at stimulating spermatogenesis and increasing testicular size, but not effective at increasing testosterone concentrations in the leopard gecko between October–December in the Northern hemisphere.

1. IntroductionGeckos are one of three extant mega-diverse lineages of squamate reptiles originating from the major radiations that began diversifying around 200 million years ago [1]. While most gecko species have only been discovered over the last 40 years, as new molecular technologies are developed to aid in our understanding of gecko diversity, many species may already have gone extinct [1]. Modern extinction rates have increased sharply over the past 200 years and correspond to the rise of the industrial society [2]. Reptiles are especially susceptible to anthropogenic threats, including habitat loss and degradation, unsustainable trade, introduced invasive species, environmental pollution, and climate change [3,4,5]. Extinction model estimates that include expected climate changes and habitat losses suggest that as many as 76% of reptiles will be committed to future extinctions by 2050 [6], while current estimates have determined that one in five reptile species are already threatened with extinction [3]. However, due to the rigorous and time intensive process of determining a species’ conservation status, counts of “officially” recognized endangered and threatened species are likely to grossly underestimate the actual number of imperiled species [4].Reproduction is core to survival, so understanding how an animal breeds is fundamental to conserving species, populations, and, indirectly, the vitality of entire ecosystems [7]. Successful application of assisted reproductive techniques for enhancing propagation, such as artificial insemination, in vitro fertilization, embryo transfer, and germplasm cryobiology, are directly related to the amount of basic reproductive information available for each species [8]. However, despite the bewildering array of reproductive modes exhibited by reptiles, there is little information on the physiology and hormonal control of reproduction for most species [9]. For example, the ability to collect spermatozoa in reptiles is a key initial step in developing assisted reproductive technologies (ART) for these animals; however, in lizards, successful semen collection has only been achieved in nine species through either electrostimulation or manual massage, including: green iguanas (Iguana iguana), Grand Cayman blue iguana hybrids (Cyclura lewisi × nubila), veiled chameleons (Chamaeleo calyptratus), panther chameleons (Furcifer pardalis), leopard geckos (Eublepharis macularis), Chaco spiny lizards (Tropodurus spinulosis), Texas rock lizards (Sceloporus torquatus), McCann’s skinks (Oligosoma maccanni), and common house geckos (Hemidactylus frenatus) [10,11,12,13,14,15,16,17]. Furthermore, the administration of exogenous mammalian gonadotropins to male lizards in order to stimulate steroidogenesis and spermatogenesis has only been attempted in <0.14% (10/6905) of all lizard species [18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]. Prior studies attempting to evaluate the effects of exogenous hormone administration on squamate testes have been terminal, and most of these studies have not examined the effects of the exogenous hormones on circulating testosterone concentrations or semen parameters, which will be paramount in establishing ART in threatened and endangered species in the future.In mammals, follicle stimulating hormone (FSH) acts at the Sertoli cell to stimulate spermatogenesis, while luteinizing hormone (LH) stimulates steroidogenesis at the level of the Leydig cell. In contrast, a “one gonadotroph, two-cell” theory is generally accepted in reptiles, whereby one gonadotropin, or a gonadotropin complex, is responsible for stimulating both spermatogenesis and steroidogenesis, functions carried out by FSH and LH independently in mammals [19,23,24,33]. In squamates, FSH has been found to stimulate both spermatogenesis and steroidogenesis, while LH has been found to have a similar qualitative effect, although to a lesser degree [21,25,26,28,29,31,34,35,36,37,38]. Exogenous hormones, including pregnant mare serum gonadotropin (PMSG) and human chorionic gonadotropin (hCG), have been previously used with success to develop assisted reproductive programs for mammals, birds, and amphibians. In reptiles, PMSG has been found to have FSH-like activity, whereas hCG has been noted to demonstrate more LH-like activity, although conflicting findings exist on gonadal and endocrine responses to hCG [21,25,26,28,29,31,34,35,36,37,38].PMSG has been found to be capable of stimulating testicle growth, spermatogenesis, and steroidogenesis in some squamates, including Indian spiny-tailed lizards (Uromastix hardwicki) [19,23,39], oriental garden lizards (Calotes versicolor) [30], the common agama (Agama agama) [21], and the little brown skink (Leiolopisma laterale) [24]. In Indian spiny-tailed lizards, PMSG induced marked elevations of both testicular and plasma androgen concentrations [23], and it was several times more potent than hCG at stimulating testosterone synthesis [19]. PMSG was also found to stimulate a four-fold increase in testicular weights and luminal diameter of the testis tubules in common agamas compared with controls, while hCG had a much smaller qualitative effect on testicle size and tubule diameter [21]. Unfortunately, testosterone concentrations were not measured in the common agamas. Male oriental garden lizards administered either hCG, PMSG, or a combination of the two hormones during the quiescent phase were found to have increased testicular weights and diameters, in addition to increased seminiferous tubule diameters and spermatids as the abundant germ cell element [30]. Additionally, cholesterol levels in the testicles were found to be lower in animals receiving PMSG and hCG, suggesting that both exogenous hormones possessed the ability to stimulate spermatogenesis and steroid hormone production in this species of lizard. Exogenous LH, hCG, FSH, FSH + LH, or PMSG administered to the male little brown skinks during the quiescent phase found that all of the mammalian gonadotropins, with the exception of LH, increased the interstitial cell number, stimulated interstitial cell hypertrophy and cytoplasmic granulation, as well as increased epididymal and sexual segment epithelial heights [24]. Based on these findings, PMSG may have FSH-like activity in reptiles, and in some species, FSH stimulation is secondary to preparation of the male reproductive tract with hCG.The purpose of this study was to determine the effects of serial injections of an exogenous hormone, PMSG, at two different concentrations, 20 IU and 50 IU, on circulating testosterone concentrations, testicular dynamics, and semen production in a model species of gecko, the leopard gecko. The leopard gecko was chosen as a model gecko species due to their abundance, ease of maintaining in captivity, and previously established reproductive seasonality [40]. Refining these techniques in a common species will be important before applying them to threatened or endangered species. Our primary objective was to determine an effective dose of PMSG to increase plasma testosterone concentrations, testicular volumes and weights, the likelihood for semen collection, and spermatozoa motility and concentration with weekly dosing over nine weeks. The hypotheses tested in this study were: (1) PMSG administration would increase testicular volumes and weights; (2) testicular volumes measured on ultrasound would positively correlate with actual testicular volumes; (3) testicular volumes and testosterone production could be effectively determined through non-lethal methods; (4) semen samples could be consistently collected from the leopard geckos by means of electrostimulation; (5) administering exogenous PMSG would increase semen collection success, spermatozoa concentration, and motility; (6) PMSG administration would be associated with a higher prevalence of morphologically normal spermatozoa; (7) PMSG treated geckos would have significantly more histologic changes in the testis and epididymis consistent with spermatogenesis; and (8) PMSG would significantly increase circulating plasma testosterone concentrations.2. Materials and Methods2.1. Ethics StatementThis longitudinal experimental study was performed in accordance with the rules and regulations established by Louisiana State University’s (LSU) institutional animal care and use committee (protocol # 20-043).2.2. Study SpeciesTwenty-four captive-bred, adult, male leopard geckos were used for this study. The ages and previous sexual histories of the male geckos used in this study are unknown, however, no male in this study had previous contact with females for at least three years while housed at LSU. The sample size was determined using the following a priori information: (1) an alpha = 0.05, a power = 0.80, a 2:1 ratio of PMSG to control animals, an expected difference of >40 pg/mL testosterone concentrations between PMSG and control animals, and a standard deviation (SD) of 20 pg/mL for each group, and (2) an alpha = 0.05, a power = 0.80, a 2:1 ratio of PMSG to control animals, an expected difference of at least 1 × 106 spermatozoa/mL between PMSG and control animals, and a SD of 7.5 × 105 spermatozoa/mL for each group.2.3. HusbandryAnimals were individually housed at Louisiana State University in 43 × 21 × 25 cm clear, plastic containers that were separated with dividers to prevent visualization of other geckos. The environmental temperature range and humidity in the climate-controlled room were 28–29 °C (83–85 °F) and 30–40%, respectively. The geckos were housed on a paper substrate and provided a hiding area and water bowl. A 12-h photoperiod was provided with standard fluorescent lighting. The geckos were fed a diet consisting of gut-loaded and dusted house crickets (Acheta domesticus), black soldier fly larvae (Hermetia illucens), and mealworms (Tenebrio molitor) (Fluker Farms, Port Allen, LA, USA) three times weekly; the amount offered was based on 2% of the geckos’ body weight. Physical examinations were performed on each leopard gecko prior to initiating the study to confirm that they were healthy.2.4. Experimental DesignA prospective experimental study was conducted from October–December (2020); timing was based on the expected non-breeding period for leopard gecko reproduction in the Northern Hemisphere (September–December) [40]. Twenty-four adult male leopard geckos were randomly divided into three treatment groups (group 1: control [saline], n = 8; group 2: 20 IU/animal [PMSG], n = 8; group 3: 50 IU/animal [PMSG], n = 8) using a random number generator (random.org). The chosen dosages of PMSG (pregnant mare serum gonadotropin, sterile filtered white lypolized powder, 1000 IU, ProSpec-Tany TechnoloGene LTD, Rehovot, Israel) were based on previous work performed in reptiles [19,23,39]. A new bottle of PMSG was reconstituted with 2 mL sterile water (Hospira Inc., RL-4428 Lake Forest, IL, USA) to 500 IU/mL for use in the geckos each week, and each animal was administered the appropriate dose subcutaneously over the left epaxial region (shoulder) once weekly for 9 weeks. Control animals received a subcutaneous injection of sterile 0.9% saline at either 0.04 mL or 0.1 mL at the same injection site to mimic the volumes of the 20 IU and 50 IU PMSG doses, respectively. All injections were administered under manual restraint.All other procedures, including non-invasive testicular measurements, electrostimulation, and venipuncture, were performed under general anesthesia with isoflurane prior to the first PMSG and saline injections (time 0, baseline) and then again once every three weeks (days 21, 42, and 63) of the experiment. Right unilateral orchidectomies with removal of the epididymis were performed in all animals on day 63. The geckos were anesthetized using an induction chamber with 5% isoflurane (Fluriso, VetOne, Boise, ID, USA) and 3 L oxygen/minute. Once the geckos lost their righting reflex, they were removed from the chamber and maintained on 3% isoflurane and 2 L oxygen/minute administered via face mask. Heart rates and respiratory rates were monitored throughout sample collection, with geckos spontaneously ventilating throughout the procedure.2.5. Non-Invasive Testicular MeasurementTesticular measurements of the right testicle were recorded using the Sonoscape S8 (Sonoscape, Centennial, CO, USA) with the 10–15 mHz linear array hockey stick probe as previously described (Figure 1) [12,41]. The right testicle was selected for measurement because it was more easily visualized on ultrasound; visualization of the left testicle was challenging due to superimposition of the gastrointestinal tract. Testicular length and width were recorded on days 0, 21, 42, and 63 of the study (Figure 2). The distance from the cranial to caudal poles of the testicle represented the length. Width was measured in the same view and included the distance from the dorsal to ventral borders of the testicle at its midpoint. Testicular volume was estimated using the following equation: V(mm3) = 0.52 LW2 [42].2.6. Blood CollectionWhole blood was collected from the ventral tail vein or cranial vena cava of each gecko using a heparinized 25-gauge needle fastened to a 1mL syringe on days 0, 21, 42, and 63 of the study. A total of 0.2 mL whole blood was collected at each time point, ensuring total blood volume collected was <0.8% body weight. Blood samples were placed into lithium heparin microtainers (B-D Vacutainer Systems, Franklin Lakes, NJ, USA) and separated into components using centrifugation at 4000× g for 8 min. Plasma was aliquoted into 2 mL cryovials (VWR International, Radnor, PA, USA) and frozen at −80 °C (−112 °F) until it was analyzed for plasma testosterone concentrations.2.7. Electrostimulation/Semen Collection/Semen EvaluationSemen was collected using electrostimulation on days 0, 21, 42, and 63 of the study. The vent and cloaca were cleaned with a Kimwipe (Kimberly-Clark Professional, Corinth, MO, USA) to remove debris. While anesthetized, each animal was electrostimulated using a 360° circumferential metallic probe (20 mm length, 3 mm diameter) connected to a variable amperage power source [13,41,43]. An intromission was defined as the process of fully inserting the metallic portion of the probe into the vent and directing it cranially. Based on ultrasound measurements, the probe length was sufficient to reach the caudal pole of the testicles. Animals were electrostimulated by performing three series of intromissions: 15 intromissions at 0.1 mAmps, 15 cloacal intermissions at 0.15 mAmps, and 15 intromissions at 0.2 mAmps. A three-minute break was provided in between each series of intromissions. Electrostimulation was discontinued following the collection of a semen sample. Any fluid observed in the cloaca following a series of intromissions was collected with a 2–20 µL single channel pipettor. Each sample was evaluated for the presence or absence of spermatozoa by placing the fluid directly on a glass slide with a cover slip and reviewing it under light microscopy (100× and 400×) at ambient temperature. If spermatozoa were visualized, the intromission number required for successful collection of an ejaculate, in addition to semen color and volume collected, were recorded. Motility of spermatozoa was determined by estimating the percentage of progressively motile spermatozoa to the nearest 5% in 5 high powered fields (magnification, 400×). All motilities were measured by a single reviewer (MM). The sample on the microscope slide and coverslip were washed into a 2 mL microcentrofuge tube (VWR International, Radnor, PA, USA) with a 1:40 dilution of formal saline. Spermatozoa concentration was determined with a Neubauer hemocytometer with phase contrast microscopy (magnification, 400×). Spermatozoa were counted in all 25 cells on each side of the hemocytometer and the total number of spermatozoa calculated by multiplying by the dilution factor. Spermatozoa morphology were recorded for each sample when at least 50 spermatozoa could be assessed. The number of morphologically normal spermatozoa, in addition to sperm with folded tails, kinked midbodies, detached heads, retained proximal droplets, coiled tails, head defects, and distal droplets were evaluated. All spermatozoa counts and morphology were performed by a single reviewer (AM).2.8. Unilateral Orchidectomy and Epididymectomy ProceduresOn day 63 of the study, all twenty-four geckos underwent a surgical procedure to remove their right testicle and epididymis for morphometric measurements and gross and histopathological assessment. Each gecko was already anesthetized with isoflurane inhalant gas for venipuncture, ultrasound, and electrostimulation, and was administered subcutaneous injections of dexmedetomidine (Zoetis Services LLC, Parsippany, NJ, USA) 0.025 mg/kg, hydromorphone (Hospira, Inc., Lake Forest, IL, USA) 0.5 mg/kg, and meloxicam (OstiLox, VetOne, Boise, ID, USA) 0.3 mg/kg for additional sedation and analgesia just prior to surgery. Anesthesia was monitored throughout the procedure by measuring the respiratory rate, Doppler heart rate, and presence/absence of muscle tone and reflexes. The geckos were placed in dorsal recumbency, and their surgical site (ventral right abdominal region) was aseptically prepared with chlorohexidine scrub (VetOne, Boise, ID, USA) and 0.9% sterile saline. A #11 scalpel blade (Bard-Parker, Aspen Surgical Products, Inc., Caledonia, MI, USA) was used to make an initial paramedian incision on the right side of the abdomen, and Metzenbaum scissors were used to extend the body wall incision (3–4 cm). A Lone Star self-retaining retractor (Cooper Surgical Inc., Trumbull, CT, USA) was used to enhance visualization within the coelomic cavity. The ventral aspect of the intra-abdominal fat pad was immediately visualized upon entering the coelomic cavity. Gentle retraction of the fat pad revealed that the thin-walled urinary bladder was adhered to the dorsal wall of the fat pad. Medial displacement of the intestines using a cotton tipped applicator (Puritan Medical Products, Guilford, ME, USA) revealed the right testicle (Figure 3) and epididymis along the dorsal body wall (Figure 4). Once in the visual field, the thin mesorchium at the cranial pole of the testicle was gently grasped with atraumatic forceps to aid in the exteriorization and visualization of the testicle. A small hemoclip (Titanium ligating clips, Weck, Morrisville, NC, USA) was placed on the testicular artery and veins to control hemostasis, and the testicle was dissected from the remainder of the mesorchium for removal. The remaining epididymis was then grasped at its cranial end and traced caudally so an additional hemoclip could be placed and the epididymis removed. Unfortunately, it was challenging to remove the testicle and epididymis en bloc. Sterile cotton tip applicators were used to apply pressure for additional hemostasis as necessary, and the abdomen was flushed with sterile saline prior to closure. The body wall was closed with 4-0 Maxon (Coviden, Mansfield, MA, USA) in a continuous pattern, and the skin was also closed with 4-0 Maxon using a horizontal mattress pattern. Sterile skin glue (GLUture, Zoetis, Kalamazoo, MI, USA) was applied to the incision to reduce seepage. A subcutaneous injection of atipamezole (Zoetis Services LLC, Parsippany, NJ, USA) 0.5 mg/kg was administered to reverse the dexmedetomidine. The geckos were monitored post-operatively until all reflexes had returned and they were able to ambulate normally. Each animal received an additional injection of hydromorphone 0.5 mg/kg subcutaneously the following day, in addition to 0.3 mg/kg meloxicam subcutaneously once daily for three consecutive days to minimize discomfort. Animals were observed daily for 6 weeks, post-operatively, for any negative side effects associated with the surgical procedure, including anorexia, depression, discharge or swelling at the incision site, dehiscence, and lack of energy or ambulation.2.9. Gross and Microscopic Assessment of the Reproductive TractTesticular and epididymal tissues were immediately rinsed with sterile 0.9% saline upon removal from the body cavity and pat dried using a Kimwipe. The testicle and epididymis were weighed separately to the nearest milligram using an analytical balance, and testicle length and width were measured using digital calipers. A gonadosomatic index (GSI) was calculated using the following formula: (testicle weight [g]/body weight [g]) × 100. Snout-vent length (SVL) and snout-tail length (STL) were also obtained while the geckos were anesthetized.The testicles and epididymides were fixed in 10% neutral buffered formalin, routinely processed, and embedded in paraffin, and 5 μm sections were stained with hematoxylin and eosin for histological analysis. Germ cell identification was performed under light microscopy by reviewing five sections of seminiferous tubules per animal. For micrometric measurements, the slides were scanned using a digital slide scanner (Nanozoomer C9600-02, Hamamatsu Photonics, Hamamatsu City, Japan). The measurements were taken using Aperio ImageScope software (Leica Biosystems, Buffalo Grove, IL, USA). Five measurements were used to define the diameter of the epididymis; ten measurements to define epididymal epithelial height; twenty measurements to define the diameter of the seminiferous tubules; five measurements to define the numbers of interstitial cells; ten measurements to define the interstitial cell nuclear diameter from each animal. The numbers of interstitial cells were counted from triangular interstitial areas formed by three sections of seminiferous tubules. For the diameter of seminiferous tubules, diameters of round sections, or short axes of elongated sections were measured. Intraepithelial secretory granules of the epididymal epithelial cells, intraluminal spermatozoa in the epididymis, interstitial cell cytoplasm, and vacuolation in the testes were graded from 1 to 3. All samples were reviewed by the same author (JL).2.10. Testosterone AssayAn enzyme immunoassay kit (EIA) (Arbor Assay DetectX Testosterone K032-H5, Ann Arbor, MI, USA) was used to measure plasma testosterone concentrations. This assay has been previously validated by the authors in leopard geckos, and the same methods were followed in the present study [32]. All samples were processed at the conclusion of the study. The published sensitivity for this assay is 9.92 pg/mL, with a limit of detection at 30.6 pg/mL. Based on previous validation, samples were diluted at 1:20 [32]. To determine the repeatability of the assay, intra- and inter-assay coefficients of variation (CV) were measured. Intra-assay CV was measured by examining the CV of each sample run in duplicate, while inter-assay CV was measured by analyzing the same samples on different plates. Values with CV < 15% were considered data, while samples with CV > 15% were re-analyzed.2.11. Statistical AnalysisThe distributions of the data were evaluated using the Shapiro–Wilk test, skewness, kurtosis, and q-q plots. Data that did not meet the assumption of normality were log transformed for parametric testing. Data that were normally distributed are reported by the mean, SD, and minimum-maximum values (min–max), while non-normally distributed data are reported by the median, 25–75%, and min–max. Mixed linear models were used to determine if there were differences in the body weight, spermatozoa concentrations, testicular volume, ejaculate volume, testosterone concentrations, and spermatozoa motility by time and treatment. Leopard gecko was included in the model as the random variable, while time and treatment were fixed factors. Separate models were created for spermatozoa concentration to evaluate the treatment variable with three levels (saline, 20 IU, 50 IU) and two levels (saline, PMSG). One-way analysis of variance (ANOVA) testing was used to determine if post-surgical testicular volume, weight, or GSI differed between the treatment groups and the controls. Least significant difference tests were used for any post-hoc comparisons if the ANOVA was significant. Levene’s test was used to assess for homogeneity of variance. If no difference was noted when the three levels of treatment were compared, an independent samples t-test was used to determine if differences existed between the controls and PMSG treated animals (20 IU and 50 IU animals combined). One-tail testing was used for these comparisons. The same analyses were used to determine if seminiferous tubule diameter, epithelial height, interstitial cell nuclear diameter, interstitial cell number, and epididymal diameter differed by treatment groups. Kruskal–Wallis tests were used to determine if secretory granule content, intraluminal spermatozoa, interstitial cell cytoplasm, and cytoplasmic vacuoles on post-surgical testicles differed by treatment groups. If not significant when comparing all three groups, a Mann–Whitney test was used to make the same comparisons for saline versus PMSG combined treated geckos. A generalized linear model for an ordinal logistic response was used to determine if the number of intromissions required to collect a semen sample differed by treatment group or time; the same model was used to determine if the number of mature spermatozoa differed by treatment group and sections samples. Generalized linear models for linear responses were used to determine if spermatozoa morphologic characteristics differed by treatment or time. Pearson’s correlation coefficients were calculated to determine if body weight, testicular weight and volume, and SVL were correlated. SPSS 24.0 (IBM Statistics, Armonk, NY, USA) was used to analyze the data. A p ≤ 0.05 was used to determine significance.3. ResultsThere was a significant difference in body weight over time (F = 7.94, p < 0.0001), but not by treatment (F = 1.32, p = 0.284) or interaction of treatment and time (F = 0.33, p = 0.888). Body weights on day 63 were significantly lower than baseline (p < 0.0001), 21 days (p < 0.0001), and 42 days (p = 0.012); there was no difference in body weights between the other sampling periods (all p > 0.081) (Table 1). Body weight was positively correlated with snout vent length (R: 0.534, p = 0.007), but not snout tail length (R: −0.09, p = 0.673).There was a significant difference in testicular volume measured by ultrasound by treatment (F = 4.62, p = 0.006) and time (F = 3.98, p = 0.034) but not the interaction of treatment and time (F = 0.73, p = 0.625). Testicular volumes by ultrasound were significantly larger (p = 0.011) in the 20 IU (median: 63.62, 25–75%: 49.14–76.14, min–max: 20.10–141.52) treatment group compared to baseline (median: 38.47, 25–75%:25.50–53.68, min–max: 13.05–81.47). There was no significant difference in testicular volume by ultrasound between the 50 IU (median: 52.77, 25–75%: 35.15–70.07, min–max: 14.72–122.91) and control groups (p = 0.107) or the 50 IU and 20 IU groups (p = 0.272). Testicular volumes by ultrasound were significantly larger on day 63 compared to baseline (p < 0.0001) and day 21 (p = 0.04), and day 42 testicular volumes were significantly higher than baseline (p = 0.04) (Table 2).There was a significant positive correlation between post-surgical testicular volume and testicular weight (R: 0.936, p < 0.001); however, there was no correlation between post-surgical testicular weight and body weight (R: 0.014, p = 0.949) or SVL (R: 0.095, p = 0.659). There were significant differences in post-surgical testicular volume (F = 3.53, p = 0.024) and testicular weight (F = 2.76, p = 0.043) by treatment group. The saline group had significantly smaller post-surgical testicular volumes (20 IU, p = 0.009, 50 IU, p = 0.036) and weights (20 IU, p = 0.018, 50 IU, p = 0.043) compared with the PMSG treated animals (Table 3). There were no significant differences in post-surgical testicular volumes (p = 0.243) or weights (p = 0.314) between the PMSG treated geckos. The GSI was significantly higher (F = 4.1, p = 0.028) in the PMSG (mean ± SD: 0.13 ± 0.05, min–max: 0.06–0.24) treated geckos compared with the control geckos (mean ± SD: 0.08 ± 0.06, min–max: 0.01–0.18).Testicular volume by ultrasound was positively correlated (R: 0.672, p < 0.001) to post-surgical testicular volume. Post-surgical testicular volume was always higher than ultrasound measured testicular volume, except for one case. There was no significant difference in testicular volume difference between treatments (F = 0.923, p = 0.413). On average, post-surgical testicular volumes were 1.7 ± 0.62 (min–max: 0.13–2.54) times larger than ultrasound measured testicular volumes.Electrostimulation was successful in 85.4% (82/96) of the cases over 63 days. Semen was not collected in 6/32 (18.7%) events for the control and for 20 IU groups, while only 2/32 (6.2%) attempts in the 50IU group were unsuccessful. The majority (10/14, 71.4%) of the unsuccessful electrostimulation events were during the baseline sampling; 3/14 (21.4%) and 1/14 (7.1%) negative events were from the 21 and 42 day sampling periods, respectively. There was no significant difference in the number of intromissions required to collect semen by treatment (Χ2 = 1.3, p = 0.529) or time (Χ2 = 2.4, p = 0.502). There was no difference in ejaculate volume by treatment (F = 0.34, p = 0.718), time (F = 1.25, p = 0.296) or the interaction of time and treatment (F = 0.24, p = 0.961). Because there were no differences in ejaculate volume, the data were combined for a single reference (median: 2.0 µL, 25–75%: 2.0–2.0, min–max: 0–5.0).There was a significant difference in motility by treatment (F = 4.89, p = 0.018) and time (F = 4.7, p = 0.014). The interaction of treatment by time was not significant (F = 0.958, p = 0.477). Motility was significantly higher in the 20 IU (p = 0.006) and 50 IU (p = 0.049) groups compared with the controls (Table 4). There was no significant difference in motility between the two PMSG groups (p = 0.345). Spermatozoa motility was found to be significantly higher in the 42 and 63 day sampling periods compared to baseline and 21 day samples (Table 5).There was a significant difference in spermatozoa concentrations over time (F = 6.7, p = 0.002) and for the interaction of time by treatment (F = 2.5, p = 0.054). There was no significant difference in spermatozoa counts by treatment (F = 3.25, p = 0.064) when evaluating all three treatments (saline, 20 IU, 50 IU), but it approached significance. Spermatozoa counts were significantly higher at 42 (baseline, p = 0.009; 21 days, p = 0.002) and 63 days (baseline, p = 0.006; 21 days, p = 0.004) compared to baseline and 21 days (Table 6). There were no significant differences in spermatozoa counts between baseline and 21 days (p = 0.703) or 42 and 63 days (p = 0.749) (Table 6). When evaluating the model with the treatment variable at two levels (saline, PMSG [combined 20 IU, 50 IU]), treatment (F = 4.34, p = 0.042) and treatment by time (F = 4.35, p = 0.038) were found to be significantly different, while time was not (F = 2.65. p = 0.114). Spermatozoa concentrations were significantly higher in the PMSG treated animals (p = 0.042) compared to the saline treated animals (Table 6).The presence of normal spermatozoa was not impacted by time (X2 = 0.8, p = 0.845), but was significantly different by treatment (X2 = 10.0, p = 0.007) (Table 7). There were significant differences in the presence of folded tails (X2 = 22.8, p < 0.001) and kinked midbodies (X2 = 22.9, p < 0.001) by time, but not by treatment (folded tail: X2 = 0.7, p = 0.698; kinked midbody: X2 = 0.8, p = 0.661) (Table 8). There was no significant difference in the likelihood of distal droplets (time: X2 = 1.4, p = 0.569; treatment: X2 = 0.4, p = 0.829), head defects (time: X2 = 1.5, p = 0.523; treatment: X2 = 2.3, p = 0.313), detached heads (time: X2 = 2.8, p = 0.422; treatment: X2 = 0.9, p = 0.623), retained proximal droplets (time: X2 = 5.2, p = 0.157; treatment: X2 = 2.1, p = 0.345), or coiled tails (time: X2 = 2.9, p = 0.393; treatment: X2 = 1.1, p = 0.561) by time or treatment group (Table 9).There was a significant difference in the seminiferous tubule diameters (F = 4.4, p = 0.025) by treatment group, with PMSG treated gecko seminiferous tubule diameters being higher (mean: 229.9 µm, SD: 16.8, min–max: 200.5–264.9) than saline controls (mean: 205.6, SD: 40.9, 135.1–255.6). There were no significant differences in epididymal diameter (F = 0.9, p = 0.356), interstitial cell nuclear diameter (F = 0.07, p = 0.785), interstitial cell number (F = 0.629, p = 0.436), or epididymal epithelial height (F = 2.5, p = 0.067) between saline and PMSG treated geckos, although epididymal epithelial height approached significance. There was a significant difference in epididymal intraluminal spermatozoa (z = −1.6, p = 0.045) between treatment groups, with abundant spermatozoa found in 78.6% (11/14) of PMSG treated geckos and only 40% (2/5) of saline treated geckos. There were no significant differences in secretory granule content (Z = −0.394, p = 0.347) between treatment groups. There were no significant differences in cytoplasmic vacuolization (Z = −2.43, p = 0.417) or interstitial cell cytoplasm (Z = −2.57, p = 0.417) between saline and PMSG treated geckos. There was no significant difference in the presence of mature spermatozoa in the post-surgical testicles based on sample section reviewed (X2 = 3.7, p = 0.491) or treatment group (X2 = 3.9, p = 0.142). Mature spermatozoa were found to be abundant (80.8%) in the post-surgical testicles; fewer samples were found to have moderate (10.8%), few (6.7%), or absent (1.7%) mature spermatozoa. Spermatogonia, round spermatids, elongate spermatids, and primary spermatocytes were present in all three groups of geckos.There was no significant difference in testosterone concentrations over time (F = 2.1, p = 0.139), treatment group (F = 0.703, p = 0.507), or the interaction of time and treatment group (F = 1.6, p = 0.220). Because there was no difference in time or treatment groups, leopard gecko testosterone reference intervals (Table 10) were established for the months of October, November, and December according to the American Society of Veterinary Clinical Pathologists [44]. The Tukey’s test was used to screen for outliers [44]. For October, there were two outliers: gecko 19 (358.2 ng/mL) and gecko 4 (524.2 ng/mL); these data were removed for reference interval determination. There were no outliers for November or December. MedCalc 17 (MedCalc Software, Ostend, Belgium) was used to determine the central 95th percentile of the data. Because the data were not normally distributed, the central 95th percentiles of non-normally distributed data were determined using non-parametric methods established by the Clinical and Laboratory Standards Institute [45]. Confidence intervals of 90%, for the lower and upper limits, could not be determined for this data set due to the number of samples being less than 120.4. DiscussionThe results of this study confirmed the majority of the authors’ original hypotheses, except two: that animals administered PMSG would have histological changes associated with increased testosterone production and that there would be higher circulating plasma testosterone concentrations in PMSG treated animals compared to controls. Administration of PMSG was found to increase testicular volume and weight, and final testicular volumes measured on ultrasound were positively correlated with actual testicular volumes. Testicular volume and circulating testosterone concentrations were determined non-lethally in this model species of gecko, giving hope for future conservation programs with threatened and endangered species. Electrostimulation was determined to be an effective method to collect semen repeatedly in leopard geckos, and semen collection, spermatozoa concentrations, and motility all increased over time in PMSG treated geckos. While seminiferous tubule diameters were significantly increased in PMSG animals, indicating that there was a gonadal effect on spermatogenesis, there were no differences in interstitial cell number, nuclear diameter, cytoplasmic vacuolation, amount of cytoplasm, or secretory granule content in the epididymal epithelium. The results of this study confirm that PMSG can have a direct impact on the male leopard gecko reproductive tract from October–December in the Northern Hemisphere. The ability to obtain pharmacological control over the reproductive system of geckos will enable scientists to manipulate the reproductive cycle to reduce dependency on natural breeding seasons in these animals.Gecko body weights decreased over time. The diet offered to the geckos remained constant over the course of the study; thus, the weight loss noted could not be attributed to access to energy but instead some other factor(s). Physiologic stress, characterized by an increase in glucocorticoid synthesis and catabolism of stored energy, can be associated with routine restraint and handling [46,47,48,49]. The authors’ attempted to reduce gecko handling over the course of the study by limiting injections to once weekly and sampling frequency to once every three weeks. It is possible that a shorter study with more frequent dosing could reduce overall handling and should be considered in the future. Additionally, while behavior was not monitored over the course of this study, wild male reptiles often migrate short distances for breeding purposes [50], which can be energetically costly. In the authors’ experience (SMP, MAM), it is not uncommon for captive male squamates to reduce food consumption during the breeding season. Based on the high testosterone concentrations measured in all geckos, regardless of treatment, it is possible that the weight loss was attributed to the prenuptial reproductive cycle of the geckos. Increasing the availability of food during periods of increased reproductive activity may help offset weight loss; however, if these animals have reduced food consumption for physiologic reasons, we may just need to expect weight loss during this period of the reproductive cycle. Results of this study also indicate that gecko body weight and SVL are correlated, but that body weight and STL were not correlated. This discrepancy is likely due to the differences in tail length in these animals, and possible variation between normal and re-grown tails. Thus, SVL is a better indication of weight than STL in leopard geckos.PMSG increased testicular sizes in treatment animals compared to those administered saline, as measured by elevated GSI, testicular weights, ultrasound and post-operative testicular volumes, and correlating with a higher degree of sperm production in PMSG animals. Spermatozoa are produced in the testicles, and testis size in reptiles is maximal at the time of spermiogenesis, suggesting that large testes are indicative of a high spermatozoa production at the individual level [51]. In the common agama, a four-fold increase in GSI (mean GSI 0.88) was observed in animals receiving PMSG compared to control animals after 21 days [21]. While the difference in GSI was not as dramatic in the leopard geckos, a nearly two-fold increase was observed in animals administered PMSG compared to controls, suggesting that PMSG administration had a significant impact on spermatogenesis.PMSG administered to leopard geckos was also successful at increasing testicular volumes as measured by ultrasound over time, but these results were not dependent on the dosage of PMSG administered. Other studies, evaluating the use of PMSG in lizards, did not vary their doses of PMSG in order to determine their effects [19,21,23,24,30,39], however, PMSG was found to increase testicle size and promote spermatogenesis in these studies. Effective dosages of PMSG in lizards have ranged from 1 IU in Leiolopisma laterale [24] to 100 IU in the Agama agama [21] and demonstrated histologic changes at the level of the testis and epididymis. Thus, the effects of PMSG in lizards may not be dose dependent. While no current attempt has been made to standardize dosing, standardization will be necessary in order to develop functional reproductive programs in the future.Testicular volume as measured on ultrasound was positively correlated to the post-surgical testicular volume, suggesting that ultrasound is a viable option for non-invasive monitoring of testicular volume in a small lizard species. Ultrasound was initially selected by the authors as a non-invasive method for measuring testicle size to develop a clinical, ante-mortem method to assess the reproductive cycles of male reptiles. Previous studies evaluating the effects of exogenous hormone administration on testicle size and function have relied on post-mortem measurements to determine their effectiveness [19,21,22,23,24,28,29,30,31,38,39,52,53]. By using ultrasound, it is possible to conduct these types of studies on threatened and endangered species. The equation used in this study to estimate testicular volume was based on the volume of an ellipsoid [42] and only requires two measurements (length and width) to estimate volume. This was advantageous because obtaining a second image to evaluate testicle width was challenging in these small lizards. This method for obtaining non-invasive testicular measurements was also found to be effective in veiled and panther chameleons [41]. However, in contrast to the chameleon study, the right testicles of all leopard geckos were removed, weighed, and measured to calculate actual testicular volumes, allowing for the validation of this equation and comparison between gross and ultrasound measured testicles. Actual testicular volumes were found to be higher than ultrasound measured testicular volumes (except in one animal); thus, ultrasound measurements may underestimate actual volume. Reptile testes are intracoelomic, elongated, and cylindrical in shape [54]; lie dorsocaudal to the liver; are suspended by the mesorchium [55]. Due to their position in the body cavity, accurate identification of the testicle borders may be reduced by the super-imposition of the gastrointestinal tract, urinary bladder, liver, and intra-abdominal fat pads. Initially, the authors of this study had planned to measure both testicles via ultrasound in each animal, and then randomly remove either the left or right testicle for histopathologic assessment. However, the left testicle was often difficult to visualize due to interference from the gastrointestinal tract, thus, the more consistently visualized right testicle was selected for routine measurement and histologic assessment in all geckos.While a previous study successfully used electrostimulation in the leopard gecko to collect semen samples at a single time point [13], the results of the current study confirm that electrostimulation is a safe and effective means of reliably collecting repeated semen samples from leopard geckos over time. Electrostimulation was successful in producing semen samples in 85.4% of the attempts made in both the control and treatment gecko groups. Electrostimulation of Texas rock lizards (Sceloporus torquatus), Chaco spiny lizards (Tropidurus spinosus), and green iguanas produced similar successes, with semen collected in 77%, 94%, and 88% of samples collected, respectively [10,14,15]. Lower results were reported for panther (55%) and veiled (50%) chameleons [12]. Sampling was performed during the non-breeding season for the geckos, but during the breeding seasons of the other lizards. These findings affirm that electrostimulation can be used to collect semen from leopard geckos with a high degree of success during the non-breeding season regardless of additional exogenous treatments.Unlike mammals, the neuronal pathway that controls ejaculation in reptiles is unknown, but anatomic similarities between the urogenital systems of mammals and reptiles suggest they have similar innervations [10]. Additionally, a recent study in a porcine model suggested that electrostimulation directly activates pelvic musculature rather than neural mechanisms [56]. While the results obtained in the leopard geckos and other lizards support this idea, more studies are ultimately needed to confirm the underlying mechanisms.Isoflurane anesthesia was used in this study due to the perceived discomfort associated with electrostimulation [10,41]. While cattle routinely undergo electrostimulation without anesthesia, and electrostimulation was successfully utilized in the spiny lava lizard without sedation [14], the authors find that the significant amount of muscle contraction associated with the procedure must cause some discomfort. Additionally, in birds, it was determined that while there are perceived intra- and interspecific behavioral variations to electric impulses, the second intromission series of electrostimulation was found to induce much more agitation and vocalization in sampled males than did the first series of intromissions [57]. Follow-up studies attempting to objectively measure the degree of discomfort in reptiles are warranted. When considering anesthetics for this type of procedure, it is important to consider potential sequelae to treatment. For example, opioids have been associated with adverse effects on spermatozoa and should be avoided [58,59]. The authors only used an opioid, hydromorphone, after the final semen collection (day 63) and immediately prior to the orchidectomy for pre-emptive analgesia to limit any impact on spermatozoa. More studies are needed to determine the potential impacts of anesthetics and analgesics on reptile spermatozoa.Spermatozoa concentration and motility were found to increase over time and by treatment during the course of this study. However, no difference in ejaculate volume was observed based on treatment or time. PMSG has been found to possess FSH-like activity, promoting spermatogenesis in some lizard species [21,23,24,30,39]. Additionally, reptile spermatogenesis may take 5–8 weeks to complete [60]. The results of the present study coincide with these previous findings. Spermatozoa concentrations and motility increased over time in the PMSG treated geckos, with both being significantly higher six (42 days) and nine weeks (63 days) after initiating treatment. The initial median motility of spermatozoa collected at baseline was 0%, but this improved to 45% at day 63 (9 weeks) of the experiment. In other lizard species, spermatozoa motilities were found to be 78% in both green iguanas and spiny lava lizards, 70% in McCann’s skinks, 51–93% in the tegus (Tupinambis merianae), and 0–100% in veiled and panther chameleons [10,12,14,16]. While the potential mechanisms behind PMSG activity on spermatozoa motility are not currently understood, motility is a trait of mature sperm and may be impacted by differences in anatomy and physiology. In lizards, spermatozoa pass from the epididymis into the ductus deferens and gain maximum motility in the distal segment where they accumulate; testicular spermatozoa have poor motility (1%) [61]. PMSG may have a secondary impact on increasing leopard gecko spermatozoa motility as they undergo the maturation process and are expelled. The effects of PMSG on leopard gecko spermatozoa concentration and motility were not dose dependent, and future studies may aim to use the lowest effective dose of exogenous gonadotrophin to elicit an effect. Some authors believe that reptile spermatozoa motility is tied to some induction agent [62], and thus may be why higher average motilities have been observed in other studies that have taken place during normal breeding seasons.Median ejaculate volume in this study was 2.0 µL. Ejaculate volumes of the leopard gecko were similar to those obtained in veiled (2.0 µL) and panther chameleons (2.9 µL), but were lower than those observed in other studies (4.6 µL in the Texas rock lizard) [12,15]. However, even larger lizards, such as the green iguana, produce small ejaculate volumes (median 50 µL) [10]. Low semen volumes in reptiles have been associated with a lack of accessory sex glands [10]. However, despite the low ejaculate volumes, the concentration and motility of spermatozoa did not appear to be adversely affected. The small volumes can limit the number of tests that can be done to evaluate the sample; however, extending the sample to increase sample volume can help overcome this deficiency [10].There were more morphologically normal spermatozoa in PMSG treated leopard geckos compared to the control animals. Folded tails were the most common morphological defect, and became more prevalent over time, while kinked midbodies were the second most common anomaly, but improved over time. The electrostimulation technique used in this study was similar to that employed in green iguanas, leopard geckos, and chameleons [10,13,41]. However, how electrostimulation affects the male lizard reproductive tract, and where the semen is dispelled from, is unknown. Manual manipulation techniques, such as those employed in the New Zealand gecko [17], or a combination of manual manipulation and electrostimulation may be more successful at obtaining a physiologic ejaculate with less morphologic defects and contamination in the leopard gecko. Other squamate species have been found to have higher proportions of morphologically normal spermatozoa than were observed in the leopard geckos. In the Chaco spiny lizard, no morphological abnormalities were observed in semen samples [14], whereas green iguanas, corn snakes (Pantherophis guttatus), veiled chameleons, and panther chameleons had 94%, 75%, 56.5%, and 55% morphologically normal sperm, respectively [10,12,63]. However, the majority of these studies were performed during the breeding season of the species examined, when the normal spermatogenic cycle occurs, and morphologically normal spermatozoa would be expected. Despite expected elevations in testosterone concentrations amongst leopard geckos, the authors suspect they had not officially entered into the breeding season yet (January–September, Northern Hemisphere), and their reproductive tracts were still being primed prior to spermatogenesis. Thus, electrostimulation may have expelled spermatozoa from the epididymides prior to complete maturation, resulting in a larger proportion of secondary morphological defects. The higher proportion of morphologically normal spermatozoa in PMSG treated animals further supports the authors’ theory that this hormone has FSH-like activity in leopard geckos and is capable of stimulating spermatogenesis in these animals outside of their normal breeding period.The histological findings of this study confirm that PMSG had an impact on spermatogenesis, but did not significantly impact steroidogenesis. The measurements collected in this study, including the diameters of the epididymis and seminiferous tubules, presence of germ cell developmental stages, epididymal epithelial heights, numbers of epididymal intraepithelial secretory granules, numbers of interstitial cells, and abundance of interstitial cell cytoplasm and cytoplasmic vacuoles, were based on previous histologic descriptions in other lizards [24,30,60]. Seminiferous tubule diameters were larger in animals receiving PMSG compared with controls. Additionally, intraluminal spermatozoa in the epididymides were nearly twice as abundant in animals receiving PMSG (Figure 5) compared to saline (Figure 6). However, there were no significant differences in the interstitial cell numbers or nuclear diameters, nor the epidydimal epithelial heights between the saline and PMSG treatment groups, suggesting that no additional Leydig cell hyperplasia or hypertrophy took place in animals receiving PMSG. The findings of this study are in contrast to others in reptiles that directly measured testosterone concentrations and testicular histology following the administration of PMSG [19,21,23,24,30]. In these previous studies, significant increases in circulating testosterone were measured following PMSG administration, and histology of the testes and epididymides noted an increase in the interstitial cells, suggesting probable androgen production. However, in these studies, sampling was conducted during the quiescent phases of reproduction for each species. The breeding season of leopard geckos in the Northern Hemisphere begins as early as January and extends to late September [40]. However, despite this study taking place firmly within the proposed non-breeding phase for this species, the animals used in this study possessed high mean baseline testosterone concentrations averaged over the three months of the study (79.4 ng/mL). Based on these results, it is possible that the leopard gecko follows a three-phase reproductive cycle, similar to those exhibited in the Caspian bent-toed gecko (Cyrtopodion caspium) and the house gecko (Hemidactylus flaviviridis). However, it is also possible that, similar to the common gecko (Hemidactylus brooki), leopard geckos may be spermatogenically active throughout the year [64]. The seasonally breeding Caspian bent-toed gecko was described as having three phases of spermatogenesis: the active, transitional, and inactive phases [65]. Additionally, house geckos possess a three-phase reproductive cycle, characterized by quiescent, recrudescent, and active phases [66]. The quiescent phase of the reproductive cycle in the house gecko was characterized by flaccid, small testes, no spermatogenic activity, and low plasma steroid concentrations, while the recrudescent phase, occurring in September–October, demonstrated increasing testicular mass, increased primary and secondary spermatids, a rise in plasma steroid concentrations, and increased steroidogenic factors in the Sertoli and Leydig cells. The active phase, occurring in November–May, exhibited large numbers of mature spermatozoa in all sections of the epididymis, peak plasma steroids, and fully developed ultrastructural steroidogenic features [66]. Conversely, previous findings in common geckos from India determined that, although there was significant variation in testicular mass between different months of the year, the testes were spermatogenically active throughout most of the year, with the exception between June and September, during the wet season, where few animals possessed abundant spermatozoa [64]. Smaller sizes of Leydig cell nuclei were also observed from May–August in these geckos, suggesting reduced androgen output during low spermatogenic activity. Lizards, in general, exhibit prenuptial spermatogenesis [64], with spermatozoa being produced prior to mating. Thus, PMSG likely did not further increase testosterone production during this period of early spermatogenesis since it was already elevated naturally during this phase, further supporting the idea that leopard geckos likely follow a prenuptial pattern of reproduction. Based on results obtained in the current study, the authors propose that the leopard geckos were either in the late stages of the recrudescent phase of spermatogenesis, or they may exhibit more continuous spermatogenesis throughout the year, and PMSG acted to further stimulate spermatogenesis in these animals. Ultimately, extending sample collection from January through September will be needed to confirm one of these theories.The mean testosterone concentrations measured in the leopard geckos during this study were higher than baseline testosterone concentrations recorded in other lizards, including veiled (12.93 ng/mL) and panther chameleons (11.64 ng/mL) [12], house geckos (15 ng/mL) [66], and green iguanas during their reproductive season (29.7 ng mL−1) [67]. Additionally, the range of testosterone concentrations observed in the leopard geckos (11.1–465.7 ng/mL) varied widely and were not dissimilar to previously reported mean testosterone concentrations for this species (87.6–139.69 ng/mL) and Madagascar ground geckos (Paroedura picta) (3.30 to 144.22 ng/mL) [68,69]. Thus, some species of gecko may normally exhibit a high variability in testosterone concentrations between individuals, or it is possible that the breeding season of these animals is less well defined than previously thought in captive situations. For example, when in captivity, the Madagascar ground gecko has been found to breed continuously [70,71,72] and male leopard geckos that have had previous sociosexual experiences were found to express higher circulating androgen concentrations than naïve males [68]. Measuring testosterone concentrations over the course of the reproductive cycle will be necessary to determine if leopard geckos have a three-phase reproductive cycle, characterized by active, recrudescent, and quiescent phases, and to better understand comparisons with other species.Other possibilities for the lack of testosterone stimulation in the leopard geckos receiving PMSG may be that a higher dose or more frequent dosing is required in these animals to elicit an effect. The doses selected for this study, 20 IU and 50 IU, were standardized to animal rather than an IU/kg basis. This was done because hormones tend to flood all available active sites at the level of the tissue, causing a ceiling effect. The dosages selected in this study were thought to be mid-range, with the aim to use the lowest effective dosage of hormone required to elicit an effect. Previous studies evaluating the effects of PMSG in lizards used more frequent dosing (daily or every other day) with shorter durations of administration (2 days to 21 days) [19,21,23,24,30]. However, in a recent study evaluating the effects of hCG administration in veiled chameleons, it was determined that weekly injections of 100, 200, and 300 IU hCG were sufficient to maintain elevated plasma testosterone concentrations over a month-long period [12]. Additionally, in oriental garden lizards, spermatogenesis was not impacted when 5 IU of PMSG, hCG, or a combination of the two hormones were given daily for 10 days, leading the authors to consider the need for higher doses or a longer period of administration [30]. Based on these results, weekly injections of PMSG were selected for use in leopard geckos with the aim of maintaining elevated plasma testosterone concentrations over a longer period of time to be able to more fully assess the impact on spermatogenesis, since it has been determined that spermatogenesis in reptiles may take between 5–8 weeks to complete [60]. Future studies should consider administering higher dosages (e.g., 100 IU/animal) or administering PMSG on a daily or every other day basis over shorter time periods to determine dosing efficacy. Additionally, blood was sampled 7 days following the most recent PMSG injection, which may have been too long of a time period to catch a peak increase in plasma testosterone concentrations following administration of PMSG. Lastly, the failure of PMSG to stimulate additional testosterone production may be due to the possibility that leopard geckos do not follow the one gonadotroph, two-cell theory of reproduction. Based on the results of this study, PMSG has demonstrated its FSH-like effect in this species by stimulating spermatogenesis. However, testosterone concentrations did not increase concurrently as expected. Ultimately, PMSG administration at a time of quiescence and low baseline testosterone concentrations will be needed to more accurately determine the effect of exogenous PMSG on Leydig cell testosterone production to further characterize if leopard geckos are a species that follows the one gonadotroph, two-cell theory.There were several limitations associated with this study. Our lack of understanding of the leopard gecko reproductive cycle in captivity limited our ability to perform this study during a time of low testosterone production in order to evaluate the effect of PMSG on stimulating testosterone production in this species. Future studies may consider measuring study subject testosterone concentrations, prior to recruitment into the study, to ensure levels are truly associated with a quiescent phase of reproduction. Additionally, more frequent blood collection or sampling within 24 h of PMSG treatment may help identify a peak in circulating levels. Limitations associated with the measurement of testicular volumes via ultrasound were also present in this study. The superimposition of the gastrointestinal tract, urinary bladder, liver, and intra-abdominal fat in the location of the testes made visualization of the testicle borders challenging. Measurement error could have also occurred with the ultrasound unit and caliper placement when measuring the small testes; however, measurements were all performed by the same two authors each time (AM and MM) to limit bias. Another limitation was associated with the suture material used to close the surgical incisions. A quarter (6/24) of the leopard geckos experienced suture reactions to the 4-0 Maxon and required subsequent repair. In all of these cases, the suture material was intact and extruded through the skin or body wall. Based on these findings, the authors do not recommend its use in leopard geckos. Other limitations of this study were associated with the small ejaculate volumes, mechanical losses due to semen evaluation technique, sample contamination, and sperm clumping. Ejaculate volumes of the leopard gecko were similar to those reported in veiled (2.0 µL) and panther chameleons (2.9 µL), but they were lower than those observed in other studies (4.6 µL in the Texas rock lizard, 50 µL in the green iguana) [10,12,15]. The low volume may be attributed to the smaller relative size of the leopard geckos in comparison to the other lizards, in addition to the methodology used to evaluate and collect the samples. Ejaculates were obtained and placed directly on a slide with an overlying cover slip to confirm the presence or absence of spermatozoa and to characterize motility and any contamination present before being washed into a cryovial with 10% buffered formal saline. This step likely led to decreased sample volume for analysis and potentially led to a reduction in spermatozoa concentrations. Additionally, the consistency of white ejaculates was more viscous than samples that were more clear and were found to clump to a higher degree; this also could contribute to falsely lowering concentrations in some samples. Samples were often contaminated with feces and urine, which could have yielded artificial decreases in sperm concentrations. Other studies evaluating lizard spermatozoa used an extender prior to evaluation rather than evaluating raw samples [10,14,15]. A semen extender was not used in this study due to concerns for over-diluting the small semen volumes; however, future studies should consider extenders to increase sample volume and allow for additional sample testing.5. ConclusionsIn conclusion, PMSG was found to stimulate spermatogenesis in captive leopard geckos between October–December in the Northern Hemisphere; however, it had no effect on testosterone production during these same months. Sampling time may have affected testosterone concentrations, and more frequent sampling may be necessary. This exogenous gonadotropin was found to increase testicular volume and weights, in addition to sperm concentration and motility. Ultrasound can be used to safely and accurately measure testicular volumes ante-mortem in leopard geckos. Electrostimulation is also an effective tool to collect serial semen samples in the leopard gecko. Future studies are needed to further characterize the annual reproductive cycle of leopard geckos to gain insight into the specific phases of this prenuptial reptile.

animals : an open access journal from mdpi

10.3390/ani13101698

PMC10215328

In order to reliably record pollution from microplastics within animals, indicators for a monitoring program must be found. For this purpose, we collected invertebrates, fish, and sediment cores at 10 sites along the Wadden Sea coast of Lower Saxony, Germany; determined the amount of microplastics; and made recommendations for potential bioindicator species, based on the sampling conditions and results. The species studied included lugworm (Arenicola marina), common periwinkle (Littorina littorea), blue mussel (Mytilus edulis), and European flounder (Platichthys flesus). In total, microplastics were detected in 88% of the specimens and in each sediment core sample. Regarding the polymer composition, eight different types of microplastic were identified. Based on the results, the species blue mussel and European flounder are recommended for microplastic monitoring in biota.

Monitoring strategies are becoming increasingly important as microplastic contamination increases. To find potentially suitable organisms and sites for biota monitoring in the German Wadden Sea, we collected invertebrates (n = 1585), fish (n = 310), and sediment cores (n = 12) at 10 sites along the coast of Lower Saxony between 2018 and 2020. For sample processing of biota, the soft tissue was digested and the sediment samples additionally underwent a subsequent density separation step. Microplastic particles were identified using Nile red and fluorescence microscopy, followed by polymer composition analysis of a subset of particles via µRaman spectroscopy. All investigated species, sediment cores, and sites contained microplastics, predominantly in the morphology class of fragments. Microplastics were found in 92% of Arenicola marina, 94% of Littorina littorea, 85% of Mytilus edulis, and 79% of Platichthys flesus, ranging from 0 to 248.1 items/g. Sediment core samples contained MPs ranging from 0 to 8128 part/kg dry weight of sediment. In total, eight polymers were identified, predominantly consisting of polyethylene, polyvinylchloride, and polyethylene terephthalate. Considering the sampling, processing, and results, the species Mytilus edulis and Platichthys flesus are suitable species for future microplastic monitoring in biota.

1. IntroductionWith the global increase in plastic and microplastic pollution, the establishment of monitoring strategies is currently the focus of national and international strategic developments [1,2,3,4]. Implementation of monitoring is primarily inhibited by the lack of uniform standard operating procedures for all environmental matrices. Furthermore, a successful implementation of monitoring programs must ensure the derivation of valid baseline and threshold values at a spatially and temporally representative resolution.Microplastics (MPs) are commonly defined as solid synthetic polymer particles with a length of the maximum dimension below 5 mm [5,6,7] and are further subcategorized based on size, morphology, and chemical composition [8,9]. Marine environments are considered to act as a final sink for plastics [10,11], whereas terrestrial freshwaters act as transport pathways for plastics from land to the ocean, e.g., [12]. As for abiotic compartments, studies investigating the impacts of microplastics on organisms mainly focus on marine species [13], addressing the abundance and spatial distribution of MPs in species of different trophic levels and/or potential consequences associated with the uptake of microplastics and, consequently, the uptake of possibly toxic chemical compounds added or adsorbed to microplastics [14,15,16], as well as potential monitoring strategies [17,18,19].Studies conducted in the greater North Sea and Baltic Sea area have considered fish species (23 publications found on google scholar for a total of 62 species from 2013 to 2022). Pelagic species such as Clupea harengus (Atlantic herring), Trachurus trachurus (Bastard mackerel), and Sprattus sprattus (European sprat) were considered in 13 of the selected publications on 34 species [20,21,22,23,24,25,26,27,28,29,30,31,32]. Moreover, 10 of the publications considered 27 demersal species such as Melanogrammus aeglefinus (haddock), Pleuronectes platessa (European plaice), and Platichthys flesus (European flounder) [23,24,25,27,29,30,31,32,33,34].In terms of studies about MP abundance in invertebrates in the North and Baltic Sea, 17 publications on 18 species were considered. Despite the different and partly changing habitats of the investigated species, mainly filter feeders, especially Bivalvia, were included. Most of the available studies were on Mytilus edulis (blue mussel) (n = 15 publications, including [24,26,35,36,37,38,39]), followed by Magallana gigas (Pacific oyster) (n = 4, [24,37,39,40]). Gastropods were included in several studies. Here, the most studied species was Littorina littorea (common periwinkle), as the dominant grazer and suspension feeder (n = 4, [24,26,37,41]). Little studied in the North and Baltic Seas are the polychaetes, with two studies each addressing MP ingestion by Arenicola marina (lugworm) [24,39] and Hediste diversicolor (iridescent sea annelid) [42,43].Sediment studies of the North Sea mudflats are very limited, and the MP concentrations vary significantly between studies. In particular, the heterogeneity of the applied methods of sample preparation and MP identification, and the heterogeneity of the lower detection sizes between 1.2 µm and 500 µm, have to be considered. In addition, different reference values, such as concentrations per weight or volume, and different reporting of concentrations in terms of ranges, means, or medians were used. All these factors largely prevent comparison of the results.The monitoring of MP in biota is currently receiving increased attention on behalf of the Marine Framework Strategy Directive [44] and regional sea conventions, such as the Convention for the Protection of the Marine Environment of the North-East Atlantic (OSPAR) [45] and the Helsinki Commission (HELCOM) [46]. Among the potential species for MP monitoring in biota, mussels are the most common that are currently taken into consideration [19,47,48,49]. However, in terms of the identification and evaluation of MP monitoring species, only a few studies have considered multiple phyla and species. This poses a problem, since a reasonable monitoring strategy should take into account a selection of suitable indicator organisms, as representative as possible of the regions of interest. Another problem in the planning of suitable monitoring strategies is the lack of comparable data. Only if valid data sets are available is it possible to decide on the localization of monitoring sites and the required monitoring frequencies.In the present study, we analyzed microplastics in biota along the German North Sea coast, as a first approach to a possible future monitoring strategy in the area.The objectives of the current study were accordingly:Analyzing the nature and extent of the occurrence of microplastics in biota in the North Sea coastal waters of Lower Saxony.Evaluation of species–specific differences and possible correlations with species- and individual-specific parameters.Evaluation of spatial differences and influencing factors of the occurring microplastic concentrations in biota between selected study stations along the coast of Lower Saxony.Provision of recommendations for a future monitoring strategy on microplastics in biota for the Lower Saxony coastal waters, with regard to the selection of indicator organisms, station selection, monitoring frequency, and analysis methodology.2. Materials and Methods2.1. Study Area and SamplingInvertebrate and fish species were investigated within a project on microplastic abundance and distribution in biota of the Wadden Sea coastline of Lower Saxony, Germany on behalf of the Lower Saxony State Agency for Water Management, Coastal Defence, and Nature Conservation (NLWKN, Oldenburg, Germany). The sample areas are located within the Wadden Sea National Park, which stretches from the Dutch border westsides to the river Elbe eastsides, covering a total area of 3450 square kilometers [50]. The flat coastal region of the Wadden Sea consists of large areas of tidal flats that regularly dry out [51]. The climate is characterized by a warm and humid temperate climate, with a mean annual temperature of 9.7 °C and an annual precipitation of 752 mm (station Norderney, [52]).Invertebrates and sediment cores were sampled in autumn 2019 and summer 2020, while fish were sampled in the summers of 2018 and 2020. The locations were selected corresponding to the sampling stations of campaigns within the Federal and State Measurement Programme on contaminants (Figure 1). The stations differ in exposure and location in relation to water/tidal currents, as well as potential microplastic sources, tidal flat types, and anthropogenic pressure (Table 1).Invertebrates: A total number of 1585 individuals were taken at 6 sampling stations, comprising Arenicola marina (n = 308, pooled by ~5 individuals resulting in 62 pooled groups), Littorina littorea (n = 583, pooled by ~15 individuals—39 pooled groups), and Mytilus edulis (n = 694, pooled by ~5 individuals—139 pooled groups). The pooling of organisms was performed in order to receive a microplastic signal well above the detection limit. The number of individuals per pooled group was determined based on the experience of the comparative study on the Schleswig-Holstein Wadden Sea coast [24]. The individuals that were pooled together showed similar characteristics with respect to their size and were all taken at the same sampling spot.The species were selected according to discussions with the Lower State Agency for Coastal Protection, meeting the criteria of being species with predominant abundance throughout the regions and representing different feeding types. Arenicola marina (polychaeta) represents an almost ubiquitous deposit feeder in the Wadden Sea, preferring muddy to sandy substrates and mainly staying at about 30 cm depths. Littorina littorea (gastropoda) is a common grazing feeder in the region that can be found on stony surfaces or on sandy to muddy substrates closer to the shoreline of the mudflats. Mytilus edulis (bivalvia) is the predominant filter feeder, with frequent occurrence in stony shoreline sections and in mussel banks, but it also occurs on the surface of the preferred muddy to sandy substrates. Depending on the different habitats, not all species could be sampled at all stations (see Table S1a Supplementary Material).Invertebrate samples were taken manually, rinsed with filtered MilliQ-water, and immediately frozen at −18 °C.Fish: A total of 310 individuals of Platichthys flesus were collected from four fishing areas. Fish species were caught during regular fish monitoring cruises carried out by Nowak GmbH on behalf of the Lower Saxony State Agency for Water Management, Coastal Defence, and Nature Conservation. Fish individuals were placed into polyethylene bags and frozen at −8 to −18 °C and transported in freezer boxes for further processing in the laboratory.Sediment: Sediment samples were retrieved using polyvinyl chloride (PVC)-corers with a diameter of 5 cm, down to the maximum feasible depth of about 30 cm.2.2. Laboratory Analysis—Sample TreatmentInvertebrate: Every individual was photographed, and their length, width, net weight (weight of dissected soft tissue), and gross weight (weight of the total individual including shell if present) were recorded. For the extraction of microplastics from individuals, the soft tissue was removed from the shell for Mytilus edulis and Littorina littorea. The weight of the tissue per individual was determined with an analytical balance to an accuracy of 0.01 mg and transferred to previously pooled entities. Arenicola marina was transferred as whole individuals into beakers.Fish: For the analysis of Platichthys flesus, only the gastrointestinal tract was taken into account. Tissues were weighed and transferred into precleaned glass beakers, each containing the tissue of n = 1 individual.For digestion of biogenic organic matter, 10 mL of a solution consisting of potassium hydroxide (KOH) and sodium hypochlorite (NaClO) per gram of wet weighed tissue was added. The digestion solution was prepared according to Strand and Tairova, using a mixture of 150 mL NaClO (6–14%), 300 mL KOH (10 M), and 550 mL MilliQ-water [53], pre-filtered <1.2 µm to minimize contamination. Samples were incubated for 48 h at room temperature, and in case of visually detectable incomplete digestion, the samples were additionally heated to 40 °C and agitated at low speed. The digested sample suspensions were passed over a stainless steel analytical sieve with a mesh size of 20 µm, rinsed with MilliQ-water, and filtered in a stainless steel filtration unit (Sartorius Combisart) with paper filters (VWR, qualitative filter paper 413, 5–13 µm particle retention). Finally, filters were transferred for drying into acetone-rinsed glass petri dishes.Sediment: For microplastic analysis the sediment cores were cut into 5 cm sections. In total, 56 subsamples were generated (3 to 6 sediment horizons per sediment core, depending on the total depth). To avoid contamination from the PVC tube, the outermost part of the sample was not used for microplastic analysis. The samples were homogenized, and 50 mL was measured and weighed into a glass beaker [53]. Digestion of biogenic organic matter was performed via a modified protocol according to Hengstmann et al. [54]. Then, 100 mL of hydrogen peroxide (H2O2, 10%) was added and left to react for 7 days at room temperature. Following a rinsing step over a 20 µm sieve, a second digestion step with the addition of 50 mL NaClO (6–14%; volume ratio 1:3) and a reaction time of 48 h at 40 °C was included. Subsequently, a density separation with sodium iodide (NaI, 1.6 g cm−3) was undertaken in a specific glass separation column. The sample/NaI suspension was shaken overhead 12 times and left to settle for 15 min. The settled sediment was carefully extracted through the outlet at the bottom of the column. The remaining supernatant was filtered onto a paper filter (VWR, qualitative filter paper 413, 5–13 µm particle retention) with a stainless steel filtration device (Sartorius Combisart) and placed into acetone-rinsed petri dishes.Blank samples were processed throughout the analyses according to the sample series, and the resulting values were subtracted from the sample results.2.3. Identification of Particle Characteristics and Polymer CompositionFor identification, the Nile red staining method in combination with fluorescence microscopy (AxioLab A.1, Carl Zeiss Jena, TRITC HC Filterset (AHF), 2.5×) was applied [55,56]. Particle dimensions of potential MP particles were recorded, and particle morphology was determined. A random subset of particles (n = 208) was investigated for polymer composition using µRaman spectroscopy (DXR2xi Raman Imaging Microscope, Thermo Fisher Scientific, Waltham, MA, USA).2.4. Statistical AnalysisThe data resulting from the invertebrate pooled groups were recalculated to the number of individuals per group (items/individual) and the total weight of tissue of all grouped individuals (items/g). Statistical analyses were performed using R statistics (R Core Team 2021, Vienna, Austria, version 4.1.2) in an R Studio environment (RStudio Team 2021, Posit PBC, Boston, MA, USA version 2021.09.1) and IBM SPSS Statistics (IBM Corp. 2019, version 26.0, Arbank, NY, USA). A Shapiro–Wilk test was applied to test the normal distribution of microplastic concentrations. Since the data were non-parametric, median values are given throughout the manuscript (all descriptive parameters are provided within the Supporting Information). Depending on the distribution of the parameters, tests for differences in means were performed using a Kruskal–Wallis test followed by Bonferroni correction, and correlation analyses of microplastic concentrations with individual-specific parameters were calculated according to SPEARMAN. The significance level was set at α = 0.05 (α = 0.01 when indicated). Results were visualized using the R library ggplot2 [57].2.5. Background ContaminationPrecautions were taken to minimize background contamination. Therefore, all chemical solutions and MilliQ-water were filtered (691, VWR International, 1.6 μm retention), glass materials were used and pre-rinsed with acetone and MilliQ-water. Beakers were covered during standing hours, cotton laboratory coats were worn at every processing step, as well as during the integration of procedural blanks (n = 94). The humidity was increased and air filtration devices were used in the laboratory [58,59].3. Results3.1. Procedural BlanksAn average contamination of 3.9 particles (comprising 3.7 fragments and 0.2 fibers) was found in the procedural blank samples investigated alongside the invertebrate samples (n = 82). Blank samples were assigned to the analyzed composite sample and subtracted accordingly. Concerning the sediment procedural blanks (n = 12), a total number of six fragments and six fibers were found. The mean MP value of the blanks was subtracted from the results of the sediment samples; resulting in the subtraction of 2.5 fragments and 3.2 fibers per sample.3.2. Invertebrates and Fish3.2.1. Microplastic Distribution and ConcentrationsA large proportion of individuals from all species were affected by MP contamination (88%). The highest ratio was determined for Littorina littorea (94%), whereas lowest number of affected individuals was determined for Platichthys flesus (79%, Table 2).The MP concentrations detected in invertebrate species varied considerably (Table S1a–c descriptive statistics according to items/g and particle morphology). In the following, median concentrations are given in particles per gram of analyzed tissue (items/g). Please refer to Table S2a–c for particle concentrations per individual. All levels of significance were set to α = 0.05.Considering all species investigated, the concentrations ranged from 0 to 248 particles per g of analyzed tissues (given as items/g in the following). Amongst species, the concentrations followed the order from the lowest concentrations in Platichthys flesus (median over all samples 0.5 items/g), Littorina littorea (median 2.5 items/g), Mytilus edulis (median 3.0 items/g), and the highest concentrations in Arenicola marina (median 4.7 items/g).Regarding the sampling stations and considering the different measuring campaigns, the highest MP concentrations in Arenicola marina (Figure 2a) were detected at Leybucht 2019 (median 40.8 items/g), Neuharlingersiel 2019 (median 39.2 items/g), and Cappel-Neufeld 2019 (median 11.4 items/g); followed by Knockster Tief 2020 (median 10.0 items/g), Knockster Tief 2019 (median 4.3 items/g), Tettens 2020 (median 3.9 items/g), Tettens 2019 (median 3.1 items/g), and Cappel-Neufeld 2020 (median 1.1 items/g); with the lowest at Neuharlingersiel 2020 (median 0.4 items/g) and Leybucht 2020 (median 0.4 items/g). Regardless, no significant differences were detected. In Littorina littorea (Figure 2b), the highest amounts of MP particles per g analyzed tissue were similarly found in Jadebusen 2020 (median 8.1 items/g), Neuharlingersiel 2019 (median 5.4 items/g), and Leybucht 2019 (median 4.9 items/g). Lower concentrations were found in Knockster Tief 2019 (median 2.8 items/g), Leybucht 2020 (median 2.4 items/g), Jadebusen 2019 (2.2 items/g), and Knockster Tief 2020 (median 1.8 items/g). Lowest concentrations were found at Cappel-Neufeld 2019 (median 1.9 items/g), Neuharlingersiel 2020 (median 1.0 items/g), and Cappel-Neufeld 2020 (median 0.7 items/g). Considering all MP particles, regardless of their morphology, a significant difference in concentrations could be observed between individuals from Neuharlingersiel and Cappel-Neufeld (p = 0.025). Mytilus edulis (Figure 2c) specimens showed the highest concentrations at Neuharlingersiel 2019 (median 12.2 items/g), Knockster Tief 2019 (median 5.0 items/g), and Leybucht 2019 (median 4.5 items/g), followed by Cappel-Neufeld 2019 (median 4.1 items/g), Jadebusen 2019 (median 4.0 items/g), and Tettens 2019 (median 3.9 items/g). The stations with the lowest concentrations were Knockster Tief 2020 (median 2.1 items/g), Tettens 2020 (median 1.7 items/g), Cappel-Neufeld 2020 (median 0.5 items/g), Jadebusen 2020 (median 0.2 items/g), Neuharlingersiel 2020 (median 0.1 items/g), and Leybucht 2020 (median 0.0 items/g). Significant differences in fiber concentrations were detected between Neuharlingersiel and the stations Knockster Tief (p = 0.028) and Jadebusen (p < 0.001), as well as between Jadebusen and the stations Leybucht (p = 0.002) and Tettens (p = 0.015). The fish species Platichthys flesus (Figure 2d) showed the highest contamination at the station Außenjade 2018 (median 1.0 items/g), followed by Baltrum 2020 (median 0.5 items/g), Borkum 2018 (median 0.6 items/g), Baltrum 2018 (median 0.5 items/g), Außenweser 2020 (median 0.5 items/g), Außenweser 2018 (median 0.4 items/g), and Borkum 2020 (median 0.1 items/g). Differences in concentrations were significant between the station Jadebusen and both Außenweser (p = 0.002) and Borkum (p = 0.002).MP concentrations varied according to the year and season of the sampling campaigns. In the case of Arenicola marina individuals, significant differences between the sampling campaigns at Leybucht (2019 median 40.8 items/g, 2020 median 0.4 items/g, p < 0.001), Neuharlingersiel (2019 median 39.2 items/g, 2020 median 0.4 items/g, p < 0.001), and Cappel-Neufeld (2019 median 11.4 items/g, 2020 median 1.1 items/g, p = 0.007) were detected. Littorina littorea samples showed a significant difference at Neuharlingersiel (2019 median 5.4 items/g, 2020 median 1.0 items/g, p = 0.009). Across all sampling stations, the MP concentrations were significantly different for Mytilus edulis (p ≤ 0.001–0.033). Regarding Platichthys flesus, a significant difference was seen at Borkum (2018 median 2.4 items/g, 2020 median 0.66, p < 0.001).Correlations with regards to the MP concentrations per g net weight and basic data of the specimen could be detected. For Arenicola marina, a significant positive correlation between the amount of fibers and individual net weight (r = 0.518, α = 0.01), length (r = 0.484, α = 0.01), and width (r = 0.344, α = 0.01), as well as between the amount of fragments and individual net weight (r = 0.211, α = 0.01), length (r = 0.192, α = 0.01), and width (r = 0.145, α = 0.05) was found. For Mytilus edulis, significant positive correlations were found between fibers and individual net weight (r = 0.386, α = 0.01), length (r = 0.464, α = 0.01), and width (r = 0.428, α = 0.01), as well as between fragments and individual length (r = 0.133, α = 0.05). A positive correlation was also observed for Platichthys flesus between fragments and net weight (r = 0.120, α = 0.01). No significant correlations were detected for Littorina littorea.3.2.2. Particle CharacteristicsThe ingested particles predominantly consisted of fragments (92%), followed by fibers (5%) and microbeads (3%) (Table 3). The overall size distribution of fragments reveals that 44% were smaller than 50 µm, with progressively increasing frequencies with decreasing particle sizes. Fragments displayed a median length of 55 µm and ranged from 20 µm to 1907 µm. The fibers showed no particular size distribution, ranging from 45 µm to 4990 µm (median 573 µm). Microbeads were found in all species and were smaller than 1000 µm, with a median length of 37 µm.Concerning the particles found in Arenicola marina, the smallest fragment sizes at the stations were identified at Leybucht (median 47 µm), followed by Knockster Tief (median 52 µm), Neuharlingersiel (median 54 µm), Cappel-Neufeld (median 57 µm), Tettens (median 57 µm), and Jadebusen (median 64 µm). Comparing the sampling campaigns, significant differences in particle lengths only occurred at the station Cappel-Neufeld (p < 0.001, α = 0.05). No significant differences between sampling stations for fiber length could be determined; however, the size of fragments differed significantly between the stations Leybucht and Jadebusen (p = 0.013, α = 0.05), Knockster Tief and Cappel-Neufeld (p = 0.026, α = 0.05) and Jadebusen (p < 0.001, α = 0.05), Neuharlingersiel and Cappel-Neufeld (p = 0.025, α = 0.05) and Jadebusen (p < 0.001, α = 0.05), and Tettens and Jadebusen (p = 0.018, α = 0.05). The median lengths of fragments found in Littorina littorea were smallest at Leybucht (median 47 µm), followed by Jadebusen (median 48 µm), Knockster Tief (median 54 µm), Cappel-Neufeld (median 55 µm), and Neuharlingersiel (median 59 µm). However, no significant differences in fibers nor in fragments according to the different sampling stations and sampling campaigns were found. Regarding the fibers found at Mytilus edulis, no significant differences in size between sampling locations were detected. Considering the fragments, the median particle size were smallest at Cappel-Neufeld (48 µm), followed by Neuharlingersiel (53 µm), Knockster Tief (55 µm), Tettens (57 µm), Jadebusen (62 µm), and Leybucht (81 µm). The differences in size were significant between sampling locations (p ≤ 0.001–0.003, α = 0.05), except for Neuharlingersiel vs. Knockster Tief and Tettens, Knockster Tief vs. Tettens, and Jadebusen vs. Tettens and Leybucht. The smallest fragments in Platichthys flesus were recorded at Baltrum (median 51 µm), followed by Außenjade (median 56 µm), Außenweser (median 67 µm), and Borkum (median 68 µm). All particle sizes differed significantly (p < 0.001, α = 0.05) between sample locations, except for Baltrum vs. Jadebusen and Borkum vs. Außenweser.A random subsample of 140 particles were analyzed for polymer composition using µRaman spectroscopy. Of these, 134 particles were verified as synthetic polymers with four different polymer types. The predominant type found in the samples was PE (59%), followed by PET (38%). One PP and one PVC particle were also found in Littorina littorea. The distribution per species is shown in Figure 3. In consideration of the particle morphology, the analyzed fibers (n = 47) consisted exclusively of PET.3.3. Sediment Cores3.3.1. Microplastic Distribution and ConcentrationsThe sediment samples across all stations and depths contained MPs ranging from 0 to 8128 particles per kg dry sediment (median 2775 part/kg DS).Considering the findings per station and sampling year (Figure 4), Tettens 2019 showed the highest MP concentration (24,362 part/kg DS), followed by Jadebusen 2019 (21,855 part/kg DS), Leybucht 2020 (18,073 part/kg DS), Leybucht 2019 (17,643 part/kg DS), Jadebusen 2020 (17,527 part/kg DS), Neuharlingersiel 2019 (15,705 part/kg DS), Knockster Tief 2020 (11,510 part/kg DS), Knockster Tief 2019 (11,154 part/kg DS), Neuharlingersiel 2020 (8868 part/kg DS), Tettens 2020 (6500 part/kg DS), Cappel-Neufeld 2020 (6348 part/kg DS), and Cappel-Neufeld 2019 (3224 part/kg DS) (Table S3: descriptive statistics of MP concentrations in sediment according to locations and depths). Regarding the depth profiles, no MP distribution pattern could be detected.3.3.2. Particle CharacteristicsFragments represented the dominant morphology class, with 86%, followed by fibers (10%) and microbeads (4%). The size of fragments varied between 20 µm and 1383 µm, with a median length of 46 µm. The particle amount increased with decreasing particle size, resulting in 56% of all particles being smaller than 50 µm. Fibers ranged from 58 µm to 4706 µm (median 749 µm), without a particular size distribution. Regarding the microbeads, the sizes were smaller than 400 µm, with a median size of 44 µm, and were found across all stations within different depths, with the exception of the 2020 Tettens core, which did not contain microbeads in any layer.Concerning the particle sizes, the smallest fragment sizes at the stations were identified at Leybucht (median 47 µm), followed by Knockster Tief (median 52 µm), Neuharlingersiel (median 54 µm), Cappel-Neufeld (median 57 µm), Tettens (median 57 µm), and Jadebusen (median 64 µm).The polymer composition was analyzed for a random subsample of 68 particles using µRaman spectroscopy. Polyvinyl chloride (PVC 68%) was the dominant polymer type, followed by polyethylene terephthalate (PET 14%), polystyrene (PS 6%), poly(methyl methacrylate) (PMMA 3%) and polyethylene (PE 2%), polypropylene (PP 2%), polyamide (PA 2%), and polyoxymethylene (POM 2%). The fibers analyzed (n = 6) consisted exclusively of PET, whereas the microbeads analyzed (n = 6) were made of PVC (n = 3) and PS (n = 3).Regarding the distribution of depth profiles (Figure 5), it is noticeable that PVC particles were found at every depth level.4. Discussion4.1. MP Concentrations in Invertebrates and FishThe percentage of affected individuals revealed that MP contamination is ubiquitous. However, the MP particle concentrations varied according to the investigated species, locations, and seasons.In general, the highest concentrations detected for the species Arenicola marina were very pronounced for the sampling season in early autumn 2019 at the stations Leybucht and Neuharlingersiel. The station Leybucht is characterized by its location in the wider estuary of the Ems river, behind the barrier isles of Borkum, Juist, and Norderney, and therefore represents an area of elevated sediment deposition rates. Neuharlingersiel is also located behind barrier isles (Langeoog and Spiekeroog); however, it is not affected by potential MP inputs via a river or estuary. Neuharlingersiel represents a station with high tourist frequentation and is one of the central ferry ports on the German North Sea coastline. The exposure to MP sources and the specific sediment deposition conditions here are likely to play an important role concerning the ingestion of MPs through the deposit feeder Arenicola. However, the MP concentrations at both stations during the sampling season summer 2020 were within the same range as at the other sampling stations, and no significant differences were detected. Taking a closer look at the dependencies of MP concentrations and individual characteristics, a negative correlation between MP concentrations and individual weight, length, and width was detected (correlation coefficients −0.70, −0.54, and −0.49, α = 0.01); whereby, the smaller the individual, the higher the contamination. This contradicts the hypothesis that MPs in Arenicola marina only reflect a distinct sediment background signal and is supported by the finding that the Arenicola marina individuals taken at the stations Leybucht and Neuharlingersiel in autumn 2019 were significantly smaller than those sampled in summer 2020. This was most likely due to seasonal movements of Arenicola marina from the mid-tidal levels to upper shore regions, because of the migration of juveniles [60], which should be considered in future studies, to avoid or to further investigate this bias. As shown in the mesocosm experiments, Arenicola marina practices size-selective feeding, also leading to the accumulation of MPs in the feeding layer between 10 and 15 cm sediment depth [61]. Such an accumulation depth could not be verified within the sediment cores; however, it has to be considered that other bioturbation species and highly specific deposition mechanisms are present in the mudflats.In general, our findings for MP concentrations in Arenicola marina are higher than those found in the study of Van Cauberghe et al., who found 1.2 ± 2.8 items/g (lower size cutoff 5 μm) in Arenicola marina on the French–Belgian–Dutch coastline [39]. A previous study from the North Sea of Schleswig Holstein showed a concentration of 3.6 items/g (mean) in Arenicola marina, which is also higher than the results found in the present study [24]. These differences are predominately related to methodological differences, especially in terms of the lower cut-off size. Compared to the study carried out in the German Wadden Sea along the coastline of Schleswig-Holstein [24], the concentrations for the southern Wadden Sea coastline investigated in this study are almost twice as high; however, when comparing using a targeted lower size cut-off recalculated to 63 µm, the results are in line with the median concentration of 2.3 items/g in Schleswig-Holstein and 2.1 items/g in Lower Saxony.In terms of the particle characteristics and the significant differences in fragment lengths identified in Arenicola marina between the sampling stations, the conditions of different MP exposures from potential sources, and above all the sedimentation conditions and the underlying basin region expositions, morphologies, and currents during tides are most likely responsible. This is underlined by particles that were significantly larger in the eastern sampling stations of the Weser estuary compared to the more western stations, which additionally all are located behind the barrier isles along the German southern North Sea coastline. The subset of MP particles from biota samples analyzed for their polymer composition were mainly made of lightweight PE fragments, whereas the fibers were made of PET. This finding applies for all samples derived from biota samples; however, it has to be considered that the total amount of particles analyzed was low, due to a lack of resources, and thus cannot be considered fully representative.MP concentrations detected in Littorina littorea individuals were lower compared to Arenicola marina and in line with the concentrations detected for Mytilus edulis. As shown for Arenicola marina, the MP concentrations at the stations Leybucht and Neuharlingersiel and as well Jadebusen (not investigated for Arenicola marina) exhibited the largest concentrations and differences according to the sampling campaigns. This increases the evidence of the stated influence of station characteristics, such as exposure to potential MP sources and specific conditions of hydrodynamics and sedimentation. No correlation between MP concentrations and individual characteristics of size and weight was present. Comparable data for Littorina littorea have been found on the west coast of Ireland, with highest concentration of microplastics at the location of Blackhead (2.96 ± 2.92 items/g, lower size cutoff 1.2 μm) [41]. This value is comparable with the station Knockster Tief (median 2.8 items/g), one of the locations in this study exhibiting lower concentrations. A much higher concentration was found in the Dutch river delta, with a total of 20 items/g (lower size cutoff 0.7 μm) for 10 pooled individuals [37], as well as along the Schleswig-Holstein coastline in Germany [24], with a mean of 15.90 items/g (lower size cutoff 63 μm). The latter study revealed concentrations several magnitudes higher compared to this study, accounting for 5.5 items/g in Schleswig-Holstein, especially when recalculated to the respective size range (>63 µm), which results in an MP concentration of 0.9 items/g in Lower Saxony.Differences in particle characteristics regarding Littorina littorea in terms of length of morphologies were not significant. In comparison to Arenicola, no size-selective feeding behavior was demonstrated for the grazer Littorina. The polymer composition of MP particles found in Littorina littorea were similar to the findings for the other investigated species, as described for Arenicola marina.The same pattern across sampling stations and campaigns as for Arenicola marina and Littorina littorea was also shown for Mytilus edulis; however, higher concentrations were also present at the stations Knockster Tief and Cappel-Neufeld. The higher MP concentrations in samples from the sampling campaign carried out in autumn 2019 compared to summer 2020 (p < 0.001) points to the importance of defining the right time and season for Mytilus sampling. It is known that environmental factors, such as algal concentration and acidification, influence the filtration performance of Mytilus edulis, which probably also played a role here [62]. Data collected from different seasons (January, May, and July) also demonstrated that the clearance rate of Mytilus edulis was independent of seasonal temperature [62], which may also have affected the microplastic intake fluctuations between seasons. In autumn, the metabolic rate was observed as low and the energy absorbed was equal to the energy demand, while, in spring, stress results in rapid decline of oxygen consumption [63]. In this regard, it has to be considered that, up to now, no harmonized criteria in terms of seasonal- or event-based timing of sampling have been defined. Environmental factors such as temperature affect filtration rates [64], and seasonal differences can also be influenced by events such as heavy rain and land runoff or higher river discharges [65,66,67,68].Mytilus edulis is one of the most investigated species in terms of MP concentrations. Compared to findings on the French–Belgian–Dutch coastline (0.2 ± 0.3 items/g, lower size cutoff 5 μm, [39]), Norway (1.85 ± 3.74 items/g, lower size cutoff 15 mm, [15]; 0.97 items/g, lower size cutoff 70 μm, [19]), the United Kingdom (0.7 and 2.9 items/g, lower size cutoff 1 μm, [38]), and Denmark and the Netherlands Wadden Sea (0.32 items/g, lower size cutoff 5 μm, [69]), our results are slightly higher, with a median of 3.0 items/g. Compared to our study along the northwestern German coastline [24] and considering a common lower size cutoff of 63 µm, the results are in line, with 1.4 items/g (Schleswig-Holstein) to 1.2 items/g (Lower Saxony). Again, major differences can be related to different methodological approaches, such as digestion and density separation approaches and particle identification methods. Above all, the different lower size cutoffs, varying from 1 μm [39] to 150 μm [15], affected the results significantly.Mytilus edulis was the only species showing a correlation between fibers and the individuals characteristics of net weight, length, and width. This raises the question of whether this filter feeder is more prone to uptake, due to a specific physical behavior, of fibers in suspension compared to fibers that are buried in sediments or adhered to the sediment surface layer. Based on our findings, this cannot be fully explained and should be further investigated within dedicated laboratory exposure tests. The same applies for the differences of fragment sizes according to sampling stations, which most likely relate to hydrodynamic conditions during tides. The polymer compositions determined according to a subset of particles are similar to those found for the other investigated species, consisting predominantly of PE (fragments) and PET (fibers).Even considering different methodological approaches, a global comparison of MP abundance in biota showed that the results for both invertebrates and fish from the North Sea and Baltic Sea had significantly lower concentrations than, for example, in areas of the Mediterranean Sea and the Asian region [70,71,72,73,74].MP concentrations detected in the digestive tract of Platichthys flesus also varied according to the sampling area (with the lowest concentrations in the area of Außenweser) but not according to the sampling campaign. Differences between the areas can be attributed to their location directly within the estuary of the rivers Ems (Borkum) and Weser (Außenweser) and in differences related to the distance to the shoreline, with the maximum distance represented by Baltrum, which is furthermore located behind the barrier isles of Norderney and Baltrum. Compared to other studies on Platichthys flesus in the North and Baltic Seas, the concentrations determined here were considerably higher. The results from other studies ranged on average from 0.05 [33] to 2.04 [71] particles/individual, compared to the 12 particles/individual determined here along the coast of Lower Saxony. The differences here were largely due to methodological differences. For example, Rummel et al. set a lower detection limit of 500 µm [32], which leaves out the majority of particles between 20 and 500 µm considered here. This also applies to studies that did not specify a lower detection limit, but which can be estimated to be >300 µm with manual selection of potential particles using tweezers [31,32]. In general, studies on MPs in fish species of the North and Baltic Seas largely found no differences between the respective stations studied, even as a function of nearshore and offshore regions [24,28,29,31,33]. A lack of significant differences between the sampled stations was also noted by Bråte et al. [22], who however identified a “hotspot” within the port of Bergen. Lenz et al. considered MP abundances in demersal and pelagic fish species (Gadus morhua and Clupea harengus), comparing habitats, marine areas, and coastal vs. offshore sites [28]. In this study, microplastics were found in 30% of all Clupea harengus individuals at coastal sites and in 16% of individuals at offshore sites, with a significantly higher exposure of individuals in the North Sea compared to the Baltic Sea. Clupea harengus, as a pelagic species, had significantly higher concentrations than the demersal species Gadus morhua. In addition, Foekema et al. concluded that for the seven species studied in their investigation, there were spatial differences, with higher levels in the southern North Sea compared to the northern North Sea with generally low concentrations [74]. Moreover, in a methodologically comparable study along the Schleswig-Holstein coastline of Germany, no spatial differences or gradients were found with respect to the concentrations occurring in fish [24]. A lack of spatial differences was also shown by [20] for Clupea harengus and Sprattus sprattus in the Baltic Sea. In addition, they did not detect significant temporal differences over an 18-year period (1987–2005) nor between the species studied or for sampling times [20]. Comparing demersal and pelagic species, it was demonstrated that demersal species, and in particular Gadus morhua (Atlantic cod), had higher abundances of MP than pelagic species [28,32,33,75]. However, other studies found no significant difference between these habitats [30]. Positive correlations were found between MP concentrations and individual size [20]. In this context, the differences that occurred between spring and summer sampling times were attributed to increased foraging, with seasonal increases in individual size. In contrast, a study of MP ingestion by Gadus morhua and Pollachius virens (coalfish) in Icelandic waters showed no relationship of MP concentrations with individual size or weight, and no relationship with gut filling, in contrast to [23].In terms of particle characteristics, the differences in particle length were small; however, the significant differences between single stations cannot be explained here based on the location of the sampling regions, in terms of estuaries, barrier isles, or distance to the shoreline. It is assumed that both MP concentrations and particle characteristics in fish reflect the highly variable conditions in marine waters, which are less suitable for interlinking compared to the MP in invertebrates and the influencing factors, e.g., particle deposition. Similarly to the polymer composition of MP particles found in invertebrates, the MP in Platichthys flesus mainly consisted of PE (fragments) and PET (fibers).4.2. MP Concentrations in Sediment CoresMP concentrations determined within sediment cores from the invertebrate sampling stations did not show clear distribution patterns with depth, only at the station Cappel-Neufeld could a tendency of declining concentrations with increasing depth be identified. The mudflats of the Wadden Sea form a highly active layer that is, not only characterized by sediment relocations through tides [76], but is above all strongly affected by bioturbation processes. The feeding technique of species can affect the potential MP uptake, assuming that filter feeders are more prone to ingestion than deposit feeders [77]. Furthermore, the bioturbation activities of species vary, leading to particle diffusion or promoting the burial of MPs [78]. Due to this and the resulting large and varying range of values, and as only 12 sediment cores were retrieved, we did not assess a potential correlation between sediment and invertebrate MP concentrations. To assess the suitability of potential MP monitoring in mudflat sediments, further research especially addressing spatially high-resolution approaches in this highly active layer of sediment relocation is required.To compare the results from sediment analyses, only the results from the 0–5 cm layer were taken into account, since in most studies only the top layer of sediment is considered. The MP concentrations ranged from 1813 (Neuharlingersiel 2020) to 5750 part/kg dry weight (Cappel-Neufeld 2020) and were considerably higher compared to a study from the Dutch coast assessing MP particle concentrations >10 µm, which determined mean values of 770 part/kg dry weight in a sediment sample of the Danziggat at the south of the isle of Ameland [38].The polymer composition of MP particles within the sediment cores revealed a tendency of increasing MP particle abundance and higher material density with increasing depth. However, the number of analyzed particles was not representative enough to prove the significance of this statement. Strikingly, the polymer composition in sediments compared to those determined for biota differed, with increased amounts of high-density polymers such as PVC in sediments, which could not be detected for the biota samples. However, it would be expected that the composition in the deposit feeder Arenicola marina, in particular, would be similar to that of the sediment samples. Whether this bias is based on the potential selective feeding of Arenicola marina cannot be assessed here, due to the low particle numbers assessed for polymer composition, but indicates the necessity for further dedicated research in this regard.4.3. Suitability of Investigated Species for MP-BiomonitoringRegarding the suitability of the studied species for MP biomonitoring purposes, we identified several criteria as being met: (i) spatial representation for the German North Sea coast, Europe-wide, and globally; (ii) feasibility in terms of sampling and sample processing; and (ii) species-specific limitations regarding morphology and habitat mobility.The habitat of Arenicola marina is widespread throughout Europe and extends over the northern Atlantic Ocean (also occurring along the east coast of America), arctic regions, the southern Baltic Sea, and regions of the Mediterranean [79,80]. Sampling is more time consuming than for other species but nevertheless very straightforward; within one tide, about 20 to 50 individuals can be collected. Further sample preparation via digestion in the laboratory proved to be easy, as the entire individual can be dissolved without prior dissection. To ensure the detection of concentrations with signals significantly above the detection limit, pooling of samples with three to five individuals is recommended. The determined concentrations per individual Arenicola marina were almost normally distributed, but taking into account the individual weights, they were clearly distributed in a right-skewed manner. For the development of a baseline and threshold values, a normal distribution in terms of comparability of sites should be aimed for. If necessary, this could be achieved by transforming the values when considering the reference unit particles per weight. With regard to species-specific characteristics, it should be taken into account that the individual sizes of Arenicola marina vary at different stations. A minimum size of 6–10 cm length at sampling and the pooling of individuals of similar length in two size classes are recommended. However, MP concentrations in Arenicola marina are highly dependent on the current gut filling status and, therefore, more or less reflect a sediment signal. Furthermore, we could identify a bias in terms of individual length. However, Arenicola marina represents a key indicator species in the specific ecosystem of the Wadden Sea, they should not be considered as a biota monitoring species for MP according to their time consuming sampling process and inconsistent data results, which possibly fluctuate with the gut filling status.Littorina littorea is very well suited as an indicator species, due to its ubiquitous occurrence with good statistical representativeness of the value distributions and, in particular, due to the detection of potential synergies with other contaminants. A limitation is the low number of studies on this species under real environmental conditions, which indicates a low acceptance for monitoring programs. Littorina littorea is widely distributed globally and omnipresent along the German coast, with a lower occurrence along the Baltic sea coast [81]. Littorina littorea does not occur in the same abundance in every location selected or does not appear at all. Sampling is uncomplicated and quick. About 50–150 individuals can be collected in a very short time, depending on their abundance at the respective site. Laboratory processing is time-consuming and challenging, since the shell and muscle tissue need to be separated. During this process, it cannot be excluded that tissue residues adhering to the shell remain. Ideally, this preparation step should be further improved, which poses a challenge, since an initial treatment with hydrochloric acid to destroy the shell may lead to destruction of synthetic polymers and should therefore be refrained from. Mechanical disruption prior to digestion of the entire individuals, including shell fragments, and subsequent removal of the remaining shell debris could be considered. Littorina littorea shows overall concentrations per individual and per weight of an approximately normal distribution, which could be further improved by statistical transformation of the values. This suggests the suitability of Littorina littorea as an indicator species. Another potential option is using Littorina littorea as a species for evaluating near-surface accumulations from sedimentary or mineral surfaces, as they are predominantly grazers. However, it should be taken into account that the MP particles recorded here reflect the local characteristics of these surfaces. They result from the topographic position and the exposure in the terrain, although they can, for example, control the expression of algae populations. Accordingly, if Littorina littorea is considered for MP monitoring in biota, it is recommended that sampling should be conducted simultaneously to the monitoring of tributyltin (TBT) at the same locations and times. In addition to improving the comparability of stations further offshore, this would also result in synergies between the potential exposure to pollutants and MP.In the context of MP monitoring, Mytilus edulis is a very good indicator candidate. Valid data are generated through pooled samples and the differentiation of dissimilar size classes that, in association with contaminant monitoring, provide valuable clues to the potentially harmful effects of MP. The inclusion of Mytilus edulis in biota monitoring has a high scientific acceptance and is already being considered for the development of such monitoring strategies [18]. Mytilus edulis is widely distributed along the German coast of the North Sea and Baltic Sea. Globally, the occurrence of this species is also nearly universal and includes the Baltic Sea, Northeast Atlantic, Mediterranean Sea, east and west coasts of North America, southern coastal sections of South America, southern coasts of Australia, and isolated occurrences along the coasts of Southeast Asia [82]. Sampling of Mytilus edulis is site-specific and simple when abundances are sufficient. This also applies to the sample preparation in the laboratory with easy and quick dissection of the soft tissue. The distribution of MP values in Mytilus edulis related to individuals and weight is approximately normal and can be further improved through suitable statistical transformation. Mytilus edulis is the only species studied in this investigation showing a significant difference in terms of individual characteristics for dimensions and weights. As a consequence, it is recommended that two different size classes should be considered as part of potential monitoring of MP in Mytilus edulis. The potential correlation between MP concentration and individual length or weight should also be recorded, particularly with respect to specifying valid and consistent reference units of MP load per individual vs. MP load per weight of tissue analyzed. Since the duration of the filtering phases of bivalvia (i.e., during tides) is a relevant factor, the respective topographic position above sea level and the related filter duration should also be documented. As a bivalvia and filter feeder, Mytilus edulis is in focus as a representative species and can therefore also reflect MP concentrations in waters. Similarly to Littorina littorea, future measurement campaigns using Mytilus edulis should seek a synergy with pollutant monitoring. Sampling of Mytilus edulis in synchronization with contaminant monitoring of mussel beds is strongly recommended.Lastly, Platichthys flesus also turned out to be well suited for MP monitoring. Additionally, the inclusion of another pelagic fish species should be discussed. Platichthys flesus represents demersal fish and occurs universally and in high abundance along the German coasts of the North Sea and Baltic Sea. Relevant occurrences also exist along the east coast of North America [82]. Sampling and sample preparation are simple but require greater resources for vessel-based sampling campaigns. Furthermore, a very good synergy with the monitoring of contaminants in biota can be achieved. The integration of fish species in MP monitoring has a wide scientific acceptance; however, the current state of research more often includes pelagic species such as Clupea harengus.5. ConclusionsMPs were present in all investigated invertebrate, fish, and sediment samples of the German North Sea coastline of Lower Saxony. Differences in MP concentrations and characteristics according to different sampling locations were most likely attributable to deposition and accumulation patterns driven by tides, currents, and exposure to potential MP sources within river estuaries.Considering the results of the MP concentrations detected in terms of a spatial representation of the German North Sea coast, Europe wide, and globally; the feasibility in terms of sampling and sample processing; and species-specific limitations regarding morphology and habitat mobility, this study suggests Mytilus edulis as a filter feeder and Platichthys flesus as a demersal species for MP monitoring of biota. The inclusion of another pelagic species, such as Clupea harengus, should be contemplated. Littorina littorea may also be considered; however, further preliminary studies need to be conducted based on individuals collected during TBT effect monitoring. Arenicola marina is not considered suitable for MP monitoring in biota, due to the dependency on the current status of gut filling with sediment of individuals. The inclusion of sediment analyses from the upper sediment layers is recommended; however, it is not recommended to include depth profile sampling using sediment core sections, due to the highly active bioturbation processes in mudflats.

animals : an open access journal from mdpi

10.3390/ani11113281

PMC8614444

To better understand why retaliatory leopard killings caused by human-wildlife conflict happen in rural farming communities in South Africa and how to prevent them, this study interviewed conservationists, officials, and farmers living in a small village in the Western Cape Province. The respondents described four main problems that led to these killings: (1) the government’s response to the problem of human-leopard conflict is slow and unwilling; (2) this response was not effective; (3) there were inadequate resources to correctly respond to these killings; and (4) there was a lack of laws and their application as well as strong distrust between everyone involved, making it even harder to deal with the problem. Local community members had various innovative ideas that can be implemented to better handle the problem of human-leopard conflict in their region, which are highlighted in this article. Coupled with the criminological techniques proposed in this research, the problem of human-leopard conflict can be significantly reduced with local ideas and resources, in both the region and in other parts of the world that suffer from similar problems.

Retaliatory killings caused by human-wildlife conflict have a significant impact on the survival of leopards. This study explores the reasons for retaliatory killings of leopards by interviewing community members in a small village in South Africa that experienced high incidences of human–leopard conflict. The semi-structured interviews focused on the reasons why retaliatory leopard killings occurred and how to best mitigate the situational factors that triggered these killings. Respondents cited four main problems that fueled these killings: the government’s response to human–leopard conflict was slow and unwilling; this response involved inefficient methods; there were inadequate resources to respond to these killings; and there was a clear lack of laws or their application. Local stakeholders provided a range of innovative strategies to reduce human-leopard conflict and retaliatory killings. While all parties expressed different reasons why these solutions were or were not effective, their conclusions were often similar. The distrust that existed between the parties prevented them from recognizing or accepting their common ground. Based on existing human–wildlife conflict mitigation techniques and solutions identified by local stakeholders, this article explores how criminological techniques, including situational crime prevention, can help identify and frame effective interventions to reduce the number of illegal leopard killings driven by human-wildlife conflict.

1. IntroductionIllegal killing of wildlife is a global threat to species conservation [1,2,3]. From small animals to megafauna, countless species are affected by illegal killing. These animals can be killed by various easily accessible tools, such as firearms, traps, pitfalls, nets, and poisons [4,5,6]. Some species are illegally killed for consumption of their meat or for the products they can produce, such as skins, shells, tusks, and horns [7,8,9,10]. These products can be used for medicinal, ornamental, and cultural reasons [11,12,13,14,15]. Animals can also be killed as a response to human-wildlife conflict, which stems from multiple reasons, such as high human population density, competition for resources due to habitat encroachment and disturbance, destruction of crops, predator-game conflicts, and predator–livestock conflicts [2,16,17,18,19,20,21]. Predation on livestock is the most common type of human-wildlife conflict [22,23,24]. Livestock predation can cause severe threats to the livelihoods of villagers, especially in certain parts of Africa, Asia, and South America [16,25,26,27].Retaliatory killings are a reactionary response from wildlife attacks on livestock or people or destruction of crops. Human-wildlife conflict, and as a result, retaliatory killings, occur globally and affect multiple species, including carnivores and, more specifically, leopards [28,29,30]. Due to habitat encroachment by humans, some leopards are forced out of their land [31]. If a leopard is unable to move to another habitat due to limited availability of wild land and protected areas, the animal will revert to hunting alternative prey including livestock [32,33,34]. Most of the carnivorous species that fall victim to retaliatory killings prey or are suspected of preying on livestock [35,36,37]. A species that is directly impacted by human–wildlife conflict and retaliatory killings is leopards [28,38]. The illegal killing of leopards occurs throughout their range in mid- to southern–Africa, parts of the Middle East, and Asia [16,39,40]. In some locations, leopard killings are opportunistic and are clustered in areas with human-wildlife conflicts between farmers or villagers and leopards [41]. Leopards can move to other habitats to avoid human disturbance; however, this disturbance often involves habitat fragmentation, which leaves these animals vulnerable to extinction through other stressors like territoriality issues with other animals, illness, inbreeding, environmental disasters, etc. [31,42,43,44]. Human–leopard conflict is especially prevalent in South Africa, and leopard populations are already at risk of extinction in the Eastern and Western Cape Provinces because of habitat fragmentation and resulting lack of genetic diversity [25,33,45,46]. However, leopards have a positive impact on the ecosystem by performing many functions, such as population control through predation [47]. For these reasons, it is important to preserve leopard populations by mitigating human-leopard conflict to reduce the number of retaliatory killings.Environmental criminology, the framework adopted for this research, has successfully been applied to wildlife crime, for example, to understand and prevent parrot poaching, illegal, unreported, and unregulated fishing, and wildlife poaching, using the techniques of situational crime prevention [48,49,50,51,52,53]. All these studies examine the illegal killing of wildlife and propose specific solutions to address such criminal activity through understanding the circumstances that facilitate wildlife crime and the use of situational crime prevention techniques to address it. Environmental criminology refers to a family of theories (the rational choice perspective, choice-structuring properties, and situational crime prevention) that “share a common interest in criminal events and the immediate circumstances in which they occur” [54] (p. 1). These theories focus on explaining the immediate circumstances that make the commission of crime possible rather than the criminal [55]. As such, their overarching goals are to (a) analyze the patterns of crime; (b) understand the crime event; and subsequently, through this understanding of the crime patterns and events, (c) prevent and control crime through specific techniques that modify the risk-benefit analysis that offenders make when choosing to commit a crime [54]. This qualitative study adds to the body of environmental criminology literature on wildlife crime by exploring the reasons why retaliatory leopard killings occur in a small village in South Africa. It also applies the situational crime prevention framework to make recommendations on mitigation and prevention strategies that can be locally implemented to address retaliatory leopard killings more effectively. 2. Materials and MethodsThe interviews for this study took place in the Western Cape Province of South Africa, near a small farming community, with the exception of one interview which was conducted at a private nature reserve, two and a half hours east of the community. Most of the surrounding land is primarily fynbos habitat, divided into meat and dairy livestock farms, with the rest of the land designated as crop and game farms. Most landowners are white and speak English, while farm hands tend to be black and speak Afrikaans. The authors picked the area because it was experiencing significant human-leopard conflict, did not attract much tourist revenue, and lacked government oversight. Reports of retaliatory leopard killings were plentiful at the time and, as such, it was an ideal location to study tolerance of leopard killing [56].A total of 16 participants living in or near the community were interviewed during July and August 2011: seven (7) livestock farmers, three (3) conservation non-governmental organization (NGO) staff, and six (6) government officials. A local NGO provided the names of farmers and government officials to interview, and the researcher used snowball sampling to identify additional participants. They recommended a combination of farmers who applied leopard-friendly livestock farming techniques and farmers who opposed their conservation work as even those who opposed were eager to share their point of view. In one such case, a farmer took the researcher on his cattle rounds to show her his daily routine. The conservationist practitioners interviewed came from local NGOs, private wildlife reserves, and the Western Cape Province government wildlife service, Cape Nature, and two other related departments (The Department of Economic Development, Tourism, and Environmental Affair for Free State Province, and the Tourism and Parks Agency, Mpumalanga Province) in Free State and Mpumalanga Provinces. Given the small, close-knit nature of the community, being vouched for by a community member allowed farmers to share leopard problems more freely with the researcher and made snowball sampling necessary.The researcher contacted the recommended individuals by phone and asked if they would be willing to talk about “leopard poaching” in the region; all, but one person, agreed. Emphasis was placed on the fact that the research was not affiliated with any local actor, respondents’ identities would be kept confidential, and the interview was meant to be a fact-finding mission to understand why leopards were being killed in the area and to benefit from respondents’ insider knowledge of the situation. The overarching goal of this research was to understand the problem through a crime prevention lens by focusing on the (rational) choices made by those who illegally killed leopards and to design prevention interventions.The semi-structured interviews with local residents and conservation practitioners lasted about two hours, with the researcher visiting respondents at their homes after calling to set up an appointment. By this time, the researcher had been residing in the area for two months and was, therefore, tangentially known to the community. Although a list of interview questions was drawn up prior to the interviews, the flow of conversation determined the topics covered with occasional prompts based on the questions (see Appendix B for the list of pre-prepared interview questions). All respondents were told they could stop the interview and withdraw their consent to participate at any time, as well as skip any questions they did not wish to answer. Only cursory notes were taken during interviews to put respondents at ease and consisted of a few key words to jog recall. Conversations were typed up from memory, based on the key words written, immediately following the interview. Two of the government officials were unavailable in person and answered questions via email, with that correspondence taking place between April and September 2011, and were added to the in-person interview data.A content analysis of the interview data was conducted in Atlas.ti using a series of codes, generated inductively, to identify major themes respondents discussed and attitudes they held as to why retaliatory leopard killings occur in the study site and what contextual factors triggered them. An inductive method limited preconceptions about the data, allowing novel concepts to emerge [57]. Open coding generated descriptive and conceptual codes by reviewing small segments of the interviews and comparing them to each other. These codes were then refined using axial coding where “code labels and the data linked to them are rethought in terms of similarity and difference” and only the codes that best illustrated concepts and relationships found in the data were kept [57]. The remaining codes that emerged from this process fell under five (5) broad categories: (1) distrust; (2) ecological beliefs; (3) human-leopard conflicts; (4) illegal killings and; (5) solutions. The Table 1 provides an overview of the specific codes and quotations included under them within these broader conceptual categories.3. ResultsA significant majority of farmers and NGO staff (9 out of 10) in the study area suggested that local leopard killings are common, and tolerance for this crime is relatively high. Of the ten respondents, only one farmer denied that retaliatory killing took place. Contrary to the interviews with farmers and NGO staff, however, the majority of South African government officials denied that any such killings took place in the area. Leopards in the area were primarily shot or caught via gin traps which are spring traps used to capture an animal’s leg and left to die of dehydration and stress. When a farmer lost livestock to a leopard, he called a community member with hunting dogs to chase the leopard on his property into the open or up a tree where it could be shot. According to one government official, once hunted, farmers buried the evidence in the bush, instead of selling the skins. 3.1. Reasons Why Retaliatory Leopard Killings Take PlaceRetaliation for Livestock Predation. When asked what they would consider to be the primary reasons for the killing of leopards, the respondents suggested that these killings by farmers took place in retaliation for livestock predation. Conversations with farmers revealed significant anger at leopards, with several farmers calling leopards “criminal” for attacking their private property when wild prey was plentiful. The general feeling was that livestock losses to other causes, such as tick fever, were acceptable, but there was significantly lower tolerance for leopard predation on livestock. This feeling was exacerbated by the belief that human-wildlife conflict had not been a problem in the area until 20 years ago and so farmers felt that leopards had suddenly invaded what had been pristine livestock farming territory. Distrust in the Government. Respondents expressed feelings of anger from the perceived economic hardships imposed by leopards, ongoing land tenure and land-use issues, and the lack of government support. Distrust also emerged as an overarching emotion expressed by farmers and government officials.Two levels of distrust were apparent: (1) general distrust and (2) human-leopard conflict solution-specific distrust. General distrust between local stakeholders stemmed from historical events in the region. Farmers were angry that they received no support from the South African government (from 1867 to 1947, the mining industry subsidized farming but no longer do). Expanding African markets and government protections turned South Africa into one of the few countries that exported food despite fickle rainfall [58]. South Africa is no longer the farming superpower it once was, especially since its economic decline from the 1960s to the mid-1990s, and its history of extractive policies and overgrazing that destroyed farmlands and their topsoil [58]. Farmers felt that they should continue to receive subsidies to keep their crops and beef competitive on the world market. Adding to this general distrust was solution-specific distrust because farmers felt that the government had failed to adequately respond to human-leopard conflict while government officials defended their solutions. Specific complaints about solutions from farmers, government officials, and NGO staff were very similar, although none of the parties recognized this. The general distrust made it impossible for parties to see common ground and exacerbated ongoing tensions over human-leopard conflict solutions proposed in the area, fueling solution-specific distrust.All respondents’ complaints focused on four main problems: (1) the government’s response was slow and unwilling; (2) the government’s response involved inefficient methods; (3) there were inadequate resources to respond and; (4) there was a clear lack of laws or their application. Table 2 provides more detail on each of these emerging themes. Other Reasons. According to the farmers interviewed, additional reasons for illegally killing “problem leopards” were that permits to legally kill them took too long to obtain. Farmers did not want to risk more losses and feared not being able to catch the leopard if they waited. In contrast, NGO staff and government officials specified that farmers sometimes killed leopards out of spite without much proof that the leopards were responsible for the attacks on their livestock. This disconnect in perspective fueled general distrust and solution specific distrust amongst all involved.3.2. Seeking Government Help When Dealing with Human-Wildlife ConflictFarmers’ Perspectives. Farmers also provided more specific reasons as to why they did not call government officials when they had a problem leopard. For example, farmers complained that government officials used cage traps that were too small, reducing the chances of capturing problem leopards on their property. In the event that a problem leopard was caught in a cage trap, farmers argued that government officials did not relocate the animal far enough to prevent it from coming back to their farm. One farmer asked a veterinarian to tag a leopard trapped by the government on his land. Despite government assurances that they would relocate the leopard too far for it to return, when the farmer trapped “another problem leopard” with government help a few weeks later, the veterinarian confirmed that this “other leopard” was actually the original animal, which had returned to the farm after relocation. Farmers also thought that government officials were feeding leopards beef during translocations and that this habit was causing them to eat livestock upon their release. Whether or not these complaints were legitimate, the interviews made clear that farmers had strong feelings and preferences on the types of human-leopard conflict solutions that should be put in place, and they felt that their views were not being heard.Government’s Response to Farmers’ Concerns. Despite the disagreement between farmers and government officials and NGO staff about the size of cage traps, during the interviews, government officials did acknowledge broader failures in their responses to human-leopard conflict. Some recognized that the techniques often used, such as trapping or bell collars, were not always effective because they did not consider each farm’s specific terrain or livestock farming practices. Government officials were aware that predation prevention measures were time consuming and burdensome for farmers to implement effectively. They also recognized that the penalties for illegal leopard killings were not a deterrent given the rarity of prosecution and the lack of other easy solutions to human-leopard conflict.Government officials also acknowledged that they struggled to respond fast enough to human-leopard conflict complaints because of trained personnel shortages and lack of equipment. On the other hand, farmers rarely received monetary compensation from the government and its haphazard response to human-leopard conflict left them feeling like the government was powerless and uninterested in helping. Any help received was often inconsistent, breeding more distrust between the parties. The maelstrom of anger, distrust, and inefficiencies in response created a toxic environment where simple agreement on the incidence of retaliatory killings or even the most effective solutions was impossible.3.3. Solutions Proposed to Deal with Human-Wildlife Conflict Involving LeopardsRespondents were asked to elaborate on the various solutions that were either proposed or that they or others had implemented to deal with human-leopard conflict in the study area. The respondents were also asked if they thought these solutions were effective and, if so, why. The reasons provided by farmers versus those of government officials and NGO staff varied. Despite ongoing distrust between the parties, there was a significant amount of unacknowledged agreement on what solutions did or did not work. These reasons are summarized in Table 3. While all parties expressed the reasons why these solutions did not work in different ways, their arguments had a lot in common even if the distrust between them prevented them from recognizing or accepting their common ground.For example, when asked about the effectiveness of translocation, farmers argued that translocation was a temporary solution, as translocated leopards quickly returned to their original territories, a finding that has been verified through a study conducted in India [17]. Meanwhile, government officials and NGO staff believed that translocation stressed leopards and sparked inter-leopard conflict over territory, hurting leopard survival. Inter-leopard conflict can drive leopards to return to their original territory, so all respondents, in essence, agreed that translocation could result in leopards returning to their original territory, rendering the solution ineffective. Nevertheless, some government officials and NGO staff wanted leopards trapped and released without proof that fear of humans would keep them away from livestock. Farmers rejected this solution outright. This willingness to disregard evidence of ineffectiveness aggravated ongoing distrust between farmers and government officials and NGO staff. Farmers and government officials also agreed that killing a damage-causing leopard was sometimes a viable solution but differed on when to do so (NGO staff did not agree). Farmers preferred to shoot first to avoid further immediate conflict. However, Government officials and NGO staff preferred to try non-lethal methods before killing. All respondents recognized that alternative predation prevention methods, like Anatolian sheep dogs, donkeys, and wire/bell collars, were inconsistently successful and required careful implementation and evaluation. There was general agreement on why and when leopard predation and human-leopard conflict occurred. The interviews reveal that farmers and government officials and NGO staff disagreed on who/what was to blame for these conditions: farmers tended to blame poor prevention methods and environmental conditions, while government/NGO staff tended to blame farmers. Given that government officials and NGO staff often advocated for the prevention methods that farmers found ineffective and that farmers felt unsupported and targeted by government officials and NGO staff when they resorted to lethal predation prevention methods, this difference in who was to blame fueled both general and solution specific distrust between the parties. In turn, this distrust made it impossible for parties to see their common perspectives in what were the best solutions to prevent human-leopard conflict creating a vicious cycle of failed responses.Collectively, farmers conjured up the most innovative ideas to prevent leopard livestock predation including farmer-sponsored insurance schemes and certain product designations for farms that use leopard-friendly predation prevention methods. Farmers also suggested diversifying farming practices to avoid being vulnerable to livestock losses and rotating livestock around to different pastures to avoid repeat predation. Government officials’ main approach was to research other countries’ solutions to human-wildlife conflict, but it appeared more promising to harness farmers’ practical knowledge to find innovative solutions to human-leopard conflict. However, at the time of the interviews, the anger and distrust between farmers, government officials, and NGO staff was preventing such collaboration.4. Discussion and Conclusions4.1. Summary of FindingsThis exploratory qualitative research was designed to understand why retaliatory killings occur in a small village in South Africa and how to best prevent them. Through in-depth interviews, this research was able to identify facilitating and mitigating factors for these killings as well as predation prevention methods that have a strong potential to limit retaliatory killings based on empirically tested crime prevention techniques.Our results showed that farmers made rational choices about when to kill leopards. These choices, triggered by leopard predation on their livestock, were shaped by a number of choice-structuring properties, among which were government’s unwillingness to respond to human-leopard conflict; the government’s use of inefficient methods in situations when it did respond to this conflict; the lack of overall adequate resources to respond and; the clear lack of laws or their application when dealing with the conflict. All local stakeholders in this study had strong opinions about what solutions would or would not work to prevent opportunities for livestock predation and future leopard retaliatory killings. This research uses situational crime prevention as a framework through which to understand how and why these various proposed solutions would prevent and control illegal leopard killings or leopard-livestock predation.4.2. Study LimitationsOne important limitation of this research is its small sample size. There were significant challenges associated with locating individuals from the stratified groups who were willing to talk to the researcher, particularly given the remote location of farms. Given the exploratory nature of this research and the methods designed to collect in-depth data on the knowledge and experiences of the respondents, as well as the fact that the researchers did not seek generalizability, it was decided that a sample size of 16 respondents was sufficient. For studies that have small sample sizes, it is important to supplement such findings by citing supporting literature, and the findings from these interviews reflect the existing literature related to human-wildlife conflict [59]. Additionally, a large body of social science research supports the assumption that for qualitative research, a large sample size or interview pool is not always necessary [60,61,62,63]. A significant amount of social science research on “hard to reach” populations has been conducted in the past with small sample sizes despite which the researchers provided valuable insights about the studied phenomena [64,65,66,67]. The snowball sampling method for this research is another potential limitation, but snowball sampling was necessary to get such a tight-knit rural community to talk to a foreign researcher [68]. Respondents needed to hear from another community members that talking with the researcher was acceptable and encouraged to open up in interviews. The use of the word “poaching” could also have influenced the respondents’ answers given the sensitive nature of the topic and the fact that this terminology can sometimes be stigmatizing. When speaking with the respondents, the researcher made sure to explain that there was no stigmatization intended and then shifted to whatever terminology the respondent preferred. The amount of sensitive and detailed information given during in-person interviews and the range of opinions expressed, including controversial ones, suggest that most respondents were honest in their responses. The choice to only write down key words during interviews and write-up the content immediately afterwards likewise made respondents more comfortable and allowed for more open communication about their problems with leopards. The downside of this method was possible content error when writing up interviews from memory, but the general points made by the respondents would have been hard to forget and Halcomb et al. note that recording or transcribing interviews is not necessary for quality data collection [69]. Any hesitation on content was noted in the post-interview write-ups and that information was excluded from the analysis.4.3. Recommendations for Retaliatory Leopard Killing PreventionOur study site had a culture of distrust among actors which impeded conservation goals and made it difficult to acknowledge common ground or agree on solutions to human-leopard conflict and retaliatory killings. This finding is similar to findings related to the lack of trust in the relationship between local communities and wildlife enforcement in Botswana and Uganda [70,71]. Local residents, like the South African farmers interviewed, have innovative and effective ideas for solutions. These solutions are more likely to be accepted and successful because they consider the local lifestyles, locally available resources, mentalities, and physical environment, and are informed by past failures. Addressing the culture of distrust, a key choice-structuring property for illegal leopard killings is essential for protecting leopards because any response to human-leopard conflict and retaliatory killings will require a coordinated effort between local populations, national wildlife management authorities or government officials, NGOs, and species experts [72]. Conflict transformation combined with criminological approaches, such as situational crime prevention, can inform effective interventions and address this distrust [73]. While there was significant distrust between farmers, government officials and NGO staff, one of this research’s key findings is that there was a lot of common ground between parties. There were also many innovative ideas on how to best address the problem of leopard livestock predation and retaliatory leopard killings. The solutions the interviewees proposed were often based, albeit inadvertently, on effective crime prevention interventions as defined by situational crime prevention in other contexts (for a review of the effectiveness of these SCP interventions, see [74]). These solutions worked to block the opportunities for either leopards to prey on livestock or for farmers to kill leopards in retaliation for predation. Below, we describe, based on proposed solutions and crime prevention principles, some examples of what should be kept as best practices to prevent both retaliatory leopard killings and its trigger, human-leopard conflict.When designing prevention-based solutions to these problems, interventions should first focus on preventing leopard predation on livestock. Donkeys protect the herd, are already semi-accepted as a predation prevention method in South Africa, and are cheap to buy. Predator-proof corrals, ideally with wire mesh roofs and fencing at least a foot deep in the ground to prevent leopards from jumping or digging into enclosures, can also be used [75,76]. “Lion lights” with motion sensors could also deflect predators away from livestock corrals [77]. Ideally, these methods are applied in tandem for maximum protection. Reducing the opportunity for committing retaliatory leopard killings can then be done in two ways: (1) making tools less accessible, like banning over-the-counter sales of pesticides used to poison livestock carcasses, and; (2) immediately investigating any cases of missing collared leopards to increase the risk that an offender is caught before the disposal of the body.Livestock owners can also identify and manage leopard predation risk factors in their environment. A simple cellphone application that predicts likely predation locations (e.g., based on collared leopard movements or previous instances of predation) could help livestock owners decide where to graze their livestock and when to corral them for safety. This could reduce leopard attacks on livestock and people, thus limiting provocations for retaliatory killings. Game scouts can also track leopard movements and warn farmers or guard livestock when leopards are nearby (similar to the Lion Guardians’ model) [78]. Communities should implement predation response teams to immediately address complaints of human-leopard conflict and work with livestock farmers to create predation prevention plans that work for them, as implemented for human-tiger conflict near India’s Corbett Tiger Reserve [79]. These teams should rapidly deploy after a complaint to minimize the risk of retaliation.Another strategy to reduce human-leopard conflict could be randomly moving livestock around to different grazing areas to avoid repeat predation. Predation response team members could also work with a farmer to reduce tick fever losses, which, according to the respondents, account for far more livestock deaths than leopards, thereby decreasing financial losses [80]. They might also suggest that farmers monitor their flock regularly and immediately dispose of any carcasses to dissuade leopards from returning to a kill site and attacking more livestock.The government could also consider giving farmers financial incentives to conserve leopards since South African farmers complained of the lack of government subsidies during their interviews. It could consider subsidizing farmers who set aside land for conservation and implement predation prevention techniques, giving farmers an incentive to tolerate some livestock predation. More importantly, communities can and should be empowered to create their own solutions. Two ways to do this would be to have government officials and NGO staff support local leadership for conservation and offering local residents micro-loans to develop their own predation prevention methods.One important way to remove excuses for retaliatory killings is to make it easy to report them by, for example, creating hotlines and cell phone applications where local residents can report illegal killings anonymously. Wildlife officials also need to establish a positive presence in the community, similar to community policing officers. As an example, the Uganda Wildlife Authority created such a unit to liaise between wildlife enforcement officers and the community [71]. Wildlife officials could stop by every farm, introduce themselves, and learn about farmers’ struggles with human-wildlife conflict, or, alternatively, designate a community member to be a law enforcement liaison [80]. These efforts would reduce the distrust farmers generally have for conservation practitioners and allow for open discussion of ongoing community problems.Conflict resolution techniques could also reduce distrust and foster cooperation between local residents, government wildlife officials, and NGO staff. For example, the Kenyan Wildlife Service (KWS) has held community outreach meetings with local residents and researchers to discuss the human-wildlife conflict solutions they were implementing [81]. These forums help communities understand KWS’ actions and allow local residents to give feedback on whether or not these solutions are working. Similarly, participatory crime analysis workshops, where residents come together to discuss where crime occurs in their community and why to design tailored solutions, improve relations between people and local law enforcement by empowering local residents to protect their communities from crime [82]. These events can be a place to openly discuss conflicts and showcase common ground between parties, thus building trust [83].Our research has shown that locally designed and implemented solutions exist in the study site and that these solutions address the essential components for effective prevention of human-wildlife conflict and retaliatory leopard killings through key situational crime prevention techniques. These techniques focus on dissuading illegal behavior, negating its reward, or neutralizing triggers for it. What they do not currently do is consider positive incentives for behavior change nor specifically highlight the role of informal guardians in crime prevention. This gap was apparent during our research when respondents highlighted solutions like revenue sharing from tourism that encourage farmers to conserve leopards by giving them positive value. Another such example was the point person with leopard hunting dogs that farmers called when they experienced a predation event, an informal guardian for livestock that farmers set up themselves where formal government efforts failed. These gaps suggest the need for adapting crime prevention frameworks to better fit new disciplines in which they are applied, like conservation. We encourage researchers and practitioners to work together to do so (we provide one such adaptation based on this research in Table A1 in Appendix A; others also exist [53,84,85]). Environmental criminology has a lot to offer conservation when it comes to tackling crime problems like illegal leopard killings, but only when disciplines work together and with local communities to identify what works and why, will crime prevention tools be effectively leveraged to solve conservation problems. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the United Nations.

animals : an open access journal from mdpi

10.3390/ani13091436

PMC10177022

Over the years, the poultry industry has relied on the use of in-feed antibiotics as a growth-promoting agent and for the prevention of diseases. However, antibiotic use has brought about pathogens that are resistant to antimicrobials. To this end, spray-dried plasma (SDP), an animal blood by-product that is rich in protein-containing lipids, peptides, immunoglobulins, transferrin, and fibrinogen, is being explored as a replacement for in-feed antibiotics in poultry. We evaluated the immunological and biochemical profile of SDP in order to understand how it enhanced performance values when supplemented to a broiler diet. At the end of the four-week study, our findings demonstrated a decrease in the number of heterophils and an increase in immunoglobulin in circulation, with oxidative stress falling in the normal range. Bifidobacteria counts also increased in the SDP-supplemented treatment. This demonstrated that SDP supplementation prevented infection and caused an increase in immunoglobulin concentration required to support intestinal development and gut microbiota modulation.

Spray-dried plasma (SDP) contain a variety of functional proteins that play an immunomodulatory role. To evaluate the potential of SDP to stimulate the immune system, day-old Ross 708 male broiler chicks (200) were allocated randomly to five dietary treatments. Treatment 1 (CX) comprised chicks fed basal unmedicated corn–soybean meal (SBM) without the addition of SDP. Treatment 2 (MX) includes chicks fed unmedicated corn–SBM basal containing Bacitracin methylene disalicylate (BMD) at 0.055 g/kg diet. Treatments 3 (SDP1), 4 (SDP2), and 5 (SDP3) contained chicks given unmedicated corn–SBM basal, into which SDP was included at 10, 20, and 30 g/kg diet, respectively. On d 7, 14, and 21, chicks’ body weight and FCR were calculated. Additionally, leucocyte counts, oxidative status, and IgY concentrations were determined in blood. On d 23, fecal populations of selected indicator bacteria species were determined. Results showed that FCR for SP3 was superior (p < 0.05) to other treatments. Likewise, heterophil numbers decreased in MX and SDP treatments compared to CX. Circulating IgY concentration was higher for SDP dietary treatments (p < 0.05) compared to MX. In conclusion, dietary SDP at 30 g/kg enhanced immune surveillance by increasing circulating IgY levels, maintaining a normal oxidative state, and increasing gut Bifidobacteria, thereby improving chick growth performance.

1. IntroductionIncreasing broiler chicken performance while ensuring the health and wellbeing of the birds has been the primary goal of the poultry industry. Over the years, the industry has relied on the use of in-feed antibiotics as a growth-promoting agent and for the prevention of diseases [1]. However, the use of antibiotics has led to the emergence of pathogens that are resistant to antimicrobials. In a bid to use safer alternatives to antibiotics in the improvement of growth performance and immune response, feed supplements and additives such as plant extracts and biogenics have been used [2,3].Swine producers have consistently utilized spray-dried plasma (SDP) to enhance the growth performance, intestinal health, and survival of piglets [4]. Moreover, as a protein source in animal feed, SDP has been reported to promote livestock health [5]. Spray-dried plasma (SDP) is an animal blood by-product rich in protein and derived from the blood of healthy porcine or bovine animals through the separation of plasma from whole blood using the centrifugation method [6]. SDP contains active components such as amino acids, enzymes, lipids, peptides, immunoglobulins, transferrin, fibrinogen, and growth factors, which play a role in diverse biochemical and immunological processes [7]. It has also been suggested that the immunoglobulin-rich fraction in plasma may be responsible for the beneficial effects attributed to SDP [8]. Because the plasma is recovered from the blood that was collected from a healthy animal, the plasma is deemed safe. Moreover, the spray-drying process effectively destroys potential viral and bacterial pathogens that may be present [4]. To this end, the use of SDP in livestock feed is safe from a public health perspective [5].Extensive studies focused on evaluating the immune-enhancing and gut microbiota regulatory status of SDP in rats, and pigs have provided evidence of its influence as a growth promoter and in the suppression of inflammation [9,10]. In addition, it has been reported that regulation of the immune system is one of the many roles of the gut microbiota [11]. In a previous study, we demonstrated that porcine SDP supplementation at 30 g/kg diet and BMD antibiotic (at 0.055 g/kg diet) had similar effects in reducing intestinal Salmonella spp. colonization in broiler chickens [12]. However, the role of dietary SDP supplementation in enhancing the immune response and regulating broiler chicken gut microbiota has not been explored. Therefore, the current study was conducted in order to determine the effects of SDP at graded levels on pro-oxidant capacity, gut microorganisms, and immune response in broiler chickens. These data will provide an understanding of the mechanism by which SDP enhances the animals’ growth performance and health.2. Materials and MethodsThe animal care and use procedures were approved by the Institutional Animal Care and Use Committee (IACUC) of North Carolina Agricultural and Technical State University.2.1. Experimental Design, Diet, and Bird ManagementIn a 4-week experiment, day-old Ross 708 male broiler chicks (200) commercially sourced chicks were randomly allocated to five treatments in a completely randomized design (CRD). Treatment 1 (CX) consisted of chicks fed corn–soybean meal (SBM) basal without SDP. Treatment 2 (MX) consisted of chicks given corn–SBM basal into which Bacitracin methylene disalicylate (BMD; Zoetis Services LLC, Parsipanny, NJ, USA) was added at 0.055 g/kg diet. Treatments 3 (SDP1), 4 (SPD2), and 5 (SDP3) comprised chicks fed unmedicated corn–SBM basal into which SDP was included at 10, 20, and 30 g/kg diet, respectively (Table 1). Experimental diets were manufactured at the North Carolina State University Feed Mill (Raleigh, NC, USA) according to the feed formula (Table 1). Each treatment consisted of 4 replicate pens, with 10 chicks/pen fed ad libitum and allowed free access to water throughout the experiment. The pens contain a hanging feeder, a nipple drinker line, and fresh unused bedding. The SDP used in this study is a kind gift from APC Incorporated (Ankeny, IA, USA). In addition, the bird housing was set at a temperature of 92 °F from d 1 to d 7, and 87 °F from d 8 to d 21. Subsequently, it was reduced to 77 °F up to 28 d. Photoperiod consisted of continuous (23L: 1D) lighting at 30 lux from placement to 21 d; thereafter, it was reduced to 12L: 12D lighting up to 28 d. All experimental diets were formulated to meet or slightly exceed nutrient requirements based on the recommendations in the Ross broiler nutrition specification handbook [13]. The diets were fed as crumbles throughout the duration of the experiment.2.2. Growth PerformanceBodyweight (BW), body weight gain (BWG), and feed intake (FI) of chicks were recorded on d 7, d 14, and d 21, and the feed conversion ratio (FCR) was calculated.2.3. Blood and Plasma Collection and PreparationOn d 14 and d 25 of experiment, two birds were randomly taken from each pen, and blood was collected from the brachial (wing) vein using a sterile 23 gauge 1″ needle attached to pre-labeled sterile EDTA vacutainer tubes. The blood samples were then centrifuged at 1500× g for 10 min., and the plasma (supernatant) was collected and stored at −80 °C.2.4. Differential Leukocyte Count AnalysisOn d 14, a thin smear of each blood sample was created on glass slides that were subsequently stained using the HEMA 3 Wright-Giemsa staining kit (Fisher Scientific, Waltham, MA, USA) based on manufacturer’s instructions. The smear on each slide was allowed to dry, and then a drop of immersion oil was added onto the slide and viewed under the microscope (100× magnification, DME Side by Side Pathology 2×,). (Leica microsystems Inc., North Deerfield, IL, USA). Leukocytes (100 per slide) were counted, and the percentage of heterophil and lymphocyte was calculated in addition to the heterophil: lymphocyte ratio as previously described by [14].2.5. Evaluation of Phaseolus Vulgaris-P-Induced Cutaneous Delayed-Type Hypersensitivity (DTH)On d 25, the DTH analysis was performed on 2 birds/replicate pen, totaling 8 birds per treatment. With the aid of a constant tension micrometer caliper, the thickness of the toe web between the second and third digits of both feet was measured on d 24 before Phaseolus vulgaris (red kidney bean, PHAP) was injected. This data gave the initial pre-injection reading. Thereafter, PHAP (100 µL of 1 mg/mL; Sigma-Aldrich Inc., St. Louis, MO, USA) was injected into the right foot of the bird between the second and third digits. The left foot served as control and was injected with 100 µL of sterile PBS. At 24 h post-injection, the thickness of the toe web was measured with the aid of a micrometer.2.6. Assay of Total IgY ConcentrationBlood plasma was assayed for total IgY concentration. Plasma IgY concentrations were determined with the aid of a commercial sandwich ELISA kit (450 nm; E33-104, Bethyl Laboratory, Montgomery, TX, USA) based on manufacturer’s instructions on d 25. The immunoglobulin concentrations were determined relative to standard curve and expressed in nanograms per milliliter (ng/mL).2.7. Pro-Oxidant CapacityOn d 14, a reactive oxygen metabolites (d-ROMs) test (Diacron International s.r.l., Grosseto, Italy) measured using the FREE DUO system (Diacron International s.r.l., Grosseto, Italy) was employed to assess the pro-oxidant capacity of the plasma samples for each treatment. Based on the existing reference level, a pro-oxidant capacity >27.20 mg H2O2/dL was considered a high level of oxidative stress, as described by [15].2.8. Microbiological Analysis of Fecal MicrobiotaOn d 23, 10 g of fecal samples collected by covering the entire litter with sterile plastic bags for two hours were placed in zip lock bags on ice using forceps. Thereafter, the fecal samples were mixed with 90 mL of sterilized 0.1% peptone solution and homogenized using a stomacher at 250 rpm for 1 min. More so, the samples were serially diluted ten-fold in peptone solution. A total of 100µL was surface plated on deMan–Rogosa–Sharpe (MRS) (Remel Inc, Lenexa, KS, USA), Brain Heart Infusion (BHI) (BD, Sparks, MD, USA), modified Bifidobacterium Iodoacetate Medium-25 (mBIM-25) agar (Himedia, Kennett Square, PA, USA), and MacConkey agar (Remel Inc., Lenexa, KS, USA), to enumerate Lactobacillus spp., total bacterial count, Bifidobacterium, and E. coli, respectively. The plates were incubated for a period of 24 h at 37 °C for the total bacterial count and E. coli. Additionally, the MRS and BIM-25 plates were incubated anaerobically for 48–72 h before bacteria counts from all samples were carried out by the plate counting method.2.9. Statistical AnalysisGrowth performance, IgY ELISA concentration, differential leukocyte analysis, and DTH data collected were subjected to one-way ANOVA (Statistical Analysis Software (SAS) (2004) Version 9.2. SAS Institute Inc., Cary, NC, USA). On the other hand, fecal microbiota data were subjected to log10 transformation before analysis by one-way ANOVA. All data are presented as the mean ± SEM. Duncan’s multiple-range test was used to determine significant differences among means. Differences were considered statistically significant at p < 0.05.3. Results3.1. Growth PerformanceFrom d 1 to d 28, Average Body Weight (ABW), Average Weight Gain (AWG), and Average Feed Intake (AFI) were not significantly different (p > 0.05) for all treatments. However, the FCR was significantly influenced by the treatments (p < 0.05), with SDP3 having a lower FCR (1.248) value while CX, MX, SDP1, and SDP2 had higher FCR values. Suggesting better feed utilization at the SDP supplementation level of 30 g/kg diet compared to CX and MX (Table 2).3.2. Differential Leukocyte CountsTo investigate the possibility of SDP causing an increase in the percentage of granulocytic leukocytes in the peripheral blood of poultry as well as the percentage of immune cells, we carried out differential leucocyte count analysis. On d 14, the percentage of heterophils for all treatments was significantly different (p < 0.05), with SDP supplementation having similar values to MX, while CX had a much higher percentage of heterophils (12.25%). However, there were no significant differences (p > 0.05) in the percentage of lymphocytes as well as the ratio of heterophils to lymphocytes for all treatments (Table 3).3.3. Delayed-Type Hypersensitivity (DTH) ReactionThe delayed-type hypersensitivity reaction is an inflammatory response that mainly involves T cells [16]. The DTH response at 24 h post-phytohemagglutinin injection carried out on d 24 showed significant differences (p < 0.05) between all treatments, with MX having a higher response compared to CX (Figure 1). The DTH response of SDP treatments (i.e., SDP1, SDP2, and SDP3) was somewhat in between that of CX and MX.3.4. Concentration of Indicator Microorganisms in the Fecal of Broiler ChicksTotal bacteria count was similar for MX and SDP supplemented diet with the exception of SDP2, which had 6.88 Log10 CFU/g fecal content. E. coli counts were lower for SDP-supplemented treatments compared with MX, which had a value of 8.86 Log10 CFU/g fecal content. In addition, while the population of Lactobacillus was not significantly different (p > 0.05), the number of Bifidobacteria was significantly different (p < 0.05), with MX and SDP2 and SDP3 having a much higher population compared to SP1 and CX (Table 4).3.5. Plasma Total IgY antibody ConcentrationTo investigate the amount of circulating IgY antibodies in the plasma of chicken-fed SDP and MX diets, an ELISA assay was performed. Results demonstrated that SDP supplementation significantly increased (p < 0.05) the concentration of IgY in circulation compared to MX and was similar to CX (Figure 2).3.6. Pro-Oxidant CapacityThe effect of SDP supplementation on oxidative status in plasma is shown in Figure 3. SDP supplementation significantly influenced (p < 0.05) ROMs (reactive oxygen metabolites). SDP1 had the highest value for a pro-oxidant capacity at 29.5 mg Carr U, which is >27.20 Carr U and, therefore, indicates oxidative stress. Similarly, oxidant stress was indicated for MX and SDP2; however, oxidative stress was not observed in birds in the CX and SDP3 groups, where the ROMs were within the normal range (20.3 ± 4.08 and 22 ± 0.00 mg Carr U, respectively.4. DiscussionSpray-dried plasma contains biochemical components that influence metabolic processes as well as immune responses [7]. SDP can also serve as a suitable alternative to antibiotics in broiler chicken feed as it showed similar efficacy in reducing cecal Salmonella [13]. In the present study, we evaluated the immunological and biochemical profile of SDP in order to understand how SDP enhances performance values when supplemented to the diet. Our results demonstrated that SDP not only improved the FCR of broiler chickens over a period of 4 weeks, but it also increased the concentration of circulating antibodies and reduced the number of heterophils.In the present study, supplementing broiler diets with SDP improved the FCR of broiler chickens. However, previous studies using broiler chickens have reported an increase in other performance parameters. For example, Walters et al. [17] reported an increase in body weight at d 25 and d 42. Moreover, in challenge studies, SDP has been reported to reduce mortality and increase body weight during the starter and finisher phases compared to the control group [18]. The disparity in our findings may be a result of the difference in the duration of the experiment, suggesting that the SDP addition to the starter diet may have a long-term benefit on the growth performance of broiler chickens. Studies on SDP supplementation in the diet of piglets have shown that SDP strengthens intestinal barrier function and improves intestinal morphology [19]. While the improvement in growth parameters may be attributed to the improvement in intestinal development, other factors/co-factors may be responsible. In order to further understand how SDP enhances FCR in broiler chickens, we performed biochemical and immunological studies.The d-ROM test is a standard for measuring pro-oxidant capacity as it measures the blood concentration of hydroperoxides, which belong to the reactive oxygen metabolites group. These hydroperoxides are produced by the oxidation of molecules such as amino acids, glucosides, peptides, lipids, and proteins that are present in spray-dried plasma [7]. Our pro-oxidant test results demonstrate that the level of oxidative stress decreased with increasing SDP supplementation, with SDP3 falling within the normal range similar to CX, while MX, SDP1, and SDP2 showed a low level of oxidative stress. This suggests that an increase in the supplementation of SDP decreased the amount of hydroperoxide in the blood, thereby lowering oxidative stress in the system. The accumulation of free radicals in animal cells leads to oxidative stress, a condition that is detrimental to animal health and may result in reduced performance, disease, and death [20,21]. Interestingly, SDP3 had the best FCR among the treatments. In addition, immunological parameters such as circulating levels of heterophils decreased with SDP supplementation. Heterophils play an important role in innate immunity and resistance to infection [22]. In the present study, the diet supplemented with SDP and MX had a lower percentage of heterophils compared to the control, suggesting that there was no infection or inflammation present [23]. The interaction of cells involved in inflammatory responses and cytokine production at the site of antigen exposure influences delayed-type hypersensitivity (DTH). Therefore, the effective interaction of these cells and other factors required to trigger an immune response in response to stimuli leads to the detection of an intense DTH response [24]. In this study, MX had the most vigorous response and was similar to SP1 and SP3. Thus, our result demonstrated that SDP-supplemented treatments, and, to a larger extent, MX, can trigger an immune response to infection.In addition, immunoglobulin Y concentrations were higher in the SDP-supplemented treatment compared to MX. These higher IgY concentration levels may indicate the chick’s capacity to develop an efficient immune system sooner. In a study that carried out immunization using chicken immunoglobulin Y (IgY), chicken IgY was highly effective against a variety of intestinal pathogens and enhanced mucosal barrier integrity in different animals [25,26]. The gastrointestinal mucosa secretes immunoglobulins that help in the maintenance of its epithelial barrier by transporting them back into the lumen or by facilitating their removal [27,28]. In addition to neutralizing pathogens at mucosal surfaces, immunoglobulins also function in the regulation of gut microbiota [29]. This was reinforced by our fecal bacteria population result, which showed a statistical increase in the amount of Bifidobacterium, a beneficial microorganism of the gastrointestinal tract. Bifidobacterium can cause changes to the gut microbiota and modulate T regulatory cell functional metabolism, posing an immune checkpoint blockade to harmful pathogens [30]. This suggests that SDP’s ability to prevent infection and increase the concentration of immunoglobulin without elevating oxidative stress levels possibly plays a key role in supporting intestinal development and gut microbiota modulation.5. ConclusionsIn conclusion, our data demonstrated that SDP, and particularly SDP3, could serve as a dietary supplement in lieu of antibiotics in broiler chicken diets as it improved the feed conversion ratio and modulated gut microbiota. Moreover, the use of SDP was shown to be safe and did not cause the buildup of ROS that lead to oxidative stress. Most importantly, we determined that the decrease in heterophils and increase in immunoglobulin concentration supported intestinal barrier integrity and promoted an increase in the population of Bifidobacterium, which is beneficial for gut homeostasis. In conclusion, improvement in intestinal development was the key factor in the enhanced growth performance in broiler chickens.

animals : an open access journal from mdpi

10.3390/ani11123390

PMC8698110

Ultrastructural studies of cells and tissues are usually performed using transmission electron microscopy (TEM), which enables imaging at the highest possible resolution. The weak point of TEM is the limited ability to analyze the ultrastructure of large areas and volumes of biological samples. This limitation can be overcome by using modern field-emission scanning electron microscopy (FE-SEM) with high-sensitivity detection, which enables the creation of TEM-like images from the flat surfaces of resin-embedded biological specimens. Several FE-SEM-based techniques for two- and three-dimensional ultrastructural studies of cells, tissues, organs, and organisms have been developed in the 21st century. These techniques have created a new era in structural biology and have changed the role of the scanning electron microscope (SEM) in biological and medical laboratories. Since the premiere of the first commercially available SEM in 1965, these instruments were used almost exclusively to obtain topographical information over a large range of magnifications. Currently, FE-SEM offers many attractive possibilities in the studies of cell and tissue ultrastructure, and they are presented in this review.

The development of field-emission scanning electron microscopes for high-resolution imaging at very low acceleration voltages and equipped with highly sensitive detectors of backscattered electrons (BSE) has enabled transmission electron microscopy (TEM)-like imaging of the cut surfaces of tissue blocks, which are impermeable to the electron beam, or tissue sections mounted on the solid substrates. This has resulted in the development of methods that simplify and accelerate ultrastructural studies of large areas and volumes of biological samples. This article provides an overview of these methods, including their advantages and disadvantages. The imaging of large sample areas can be performed using two methods based on the detection of transmitted electrons or BSE. Effective imaging using BSE requires special fixation and en bloc contrasting of samples. BSE imaging has resulted in the development of volume imaging techniques, including array tomography (AT) and serial block-face imaging (SBF-SEM). In AT, serial ultrathin sections are collected manually on a solid substrate such as a glass and silicon wafer or automatically on a tape using a special ultramicrotome. The imaging of serial sections is used to obtain three-dimensional (3D) information. SBF-SEM is based on removing the top layer of a resin-embedded sample using an ultramicrotome inside the SEM specimen chamber and then imaging the exposed surface with a BSE detector. The steps of cutting and imaging the resin block are repeated hundreds or thousands of times to obtain a z-stack for 3D analyses.

1. IntroductionThe first transmission electron microscope (TEM) was invented by Max Knoll and Ernst Ruska at the Technische Hochschule zu Berlin in 1931 [1] based on research on electron motion in a magnetic field and the possibility of focusing the electron beam by Hans Busch [2]. This instrument created the opportunity to overcome the barrier of resolution occurring in light microscopy; however, the method of studying the internal organization of animal cells required much time and the development of specimen preparation methods and commercially available electron microscopes, which could be used in biological laboratories. In the 1940s, 1950s, and 1960s, scientists developed protocols for animal tissue fixation and preparation for TEM studies, which enabled images of satisfactory quality to be obtained [3,4,5,6]. As a gold standard, aldehydes were introduced for the fixation of proteins and osmium tetroxide for the preservation of lipids and contrasting of membranes [3,4,5]. The development of ultramicrotomy allowed the preparation of ultrathin sections that were essential for obtaining high-quality images in terms of resolution and focus [7,8]. The contrasting of sections using lead citrate and uranyl acetate improved the differentiation of the structures [9,10,11]. Moreover, significant progress occurred in the construction of the TEM, as electron guns, electromagnetic lenses, and power supplies became more stable and efficient. The simplification of TEM operation for use by unskilled operators was also important. From the end of the 1960s, transmission electron microscopy could provide high-quality, high-resolution images of animal specimens. The 70s and 80s of the last century was a period of intensive research on animal cells’ and tissues’ ultrastructure. Further development of transmission electron microscopy included introduction of cryo-techniques, immunochemistry, and digital imaging [12,13,14,15,16].The main weakness of transmission electron microscopy is the limited ability to analyze the ultrastructure of large areas and volumes of biological samples. Although the introduction of motorized stages and digital cameras with large sensors into TEM has created the opportunity for imaging of larger areas and analysis of serial sections, the acquisition of large-volume morphological information using transmission electron microscopy is extremely labor-intensive and frequently unsuccessful because of section deformation and damage. This limitation can be overcome by using a modern field-emission scanning electron microscope (FE-SEM) with high-sensitivity detection, which enables the creation of TEM-like images from the surface of resin-embedded biological specimens. Several FE-SEM-based techniques for two-dimensional (2D) and three-dimensional (3D) ultrastructural studies of cells, tissues, organs, and organisms have been developed during the 21st century [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]. These techniques have created a new era in structural biology and have changed the role of the scanning electron microscope (SEM) in biological and medical laboratories. Since the premiere of the first commercially available SEM in 1965 [33], these instruments were used in biological laboratories almost exclusively to obtain topographical information over a large range of magnifications. Currently, FE-SEM provides many attractive possibilities in the studies of cell and tissue ultrastructure.This article provides an overview of FE-SEM-based techniques developed for the ultrastructural studies of large areas (2D) and large volumes (3D) of resin-embedded animal specimens.2. Basic Principles of Image Formation in Scanning Electron MicroscopyThe SEM scans a focused electron beam over the surface of a sample. The electrons in the beam, known as the primary electrons, interact with the sample, producing various signals that can be used to obtain images showing the surface topography and material composition. When the primary electron enters a sample, it frequently travels a certain distance before contacting another particle. After colliding with this particle, the primary electron moves on a new trajectory, which is known as scattering. The entry of the electron beam into the specimen and the scattering events result in the formation of a teardrop-shaped reaction vessel (Figure 1).Secondary electrons (SE) are generated when the primary electrons extricate the specimen electrons. They have low energy and cannot escape from the deeper parts of the reaction vessel. Therefore, the SE detected in SEM originate exclusively from the surface or the near-surface area of the specimen (Figure 1). The SE produced in the deeper parts of the reaction vessel are absorbed by the sample. The shallow depth of origin of the detected SE makes them ideal for providing high-resolution topographical information. SE production increases slightly with the atomic number of elements; therefore, an SE signal can also be used to obtain TEM-like images from very smooth and flat surfaces of resin-embedded biological samples fixed using heavy metals. The advantage of SE imaging is that the primary electron produces several SE through multiple scattering events, which significantly increases the signal. SE can be detected using Everhart–Thornley, variable pressure SE, and in-column SE detectors.Backscattered electrons (BSE) are the original beam electrons that escape from the specimen owing to scattering (Figure 1). They have higher energies than SE. BSE emission is a function of the atomic number; therefore, the specimen area containing elements with higher atomic numbers produces a brighter signal. Consequently, BSE can be used to generate an image that indicates the differences in the chemical composition of the sample. The sample volume, from which BSE are generated, is significantly larger than the region that is a source of SE; therefore, BSE have poorer spatial resolution than SE. BSE can be detected using extra column diode-based and in-column detectors. Note that BSE participate to the same degree in the signals delivered by the SE detectors. Moreover, BSE induce the formation of SE.Since BSE provide information about material composition, they are used as a signal source for imaging of the cut surface of tissue blocks or the tissue sections mounted to the solid substrates in the majority of methods developed to study subcellular and cellular structures of animal tissue, organs, and even entire organisms in 2D and 3D modes [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32]. The effective, high-resolution ultrastructural imaging of biological samples using BSE signals is dependent on three aspects. First, the sample preparation procedure should result in the incorporation of a large quantity of heavy metals into the tissue to differentiate cell structures and increase the signal-to-noise ratio. Second, the electron optics of FE-SEM should enable stable imaging at very low acceleration voltages because the size of the reaction vessel is dependent on the acceleration voltage. At higher voltages, the electron beam penetrates deeper into the sample, BSE are emitted from a large volume, and the resolution is lower. Third, highly efficient BSE detectors are required to obtain good signal-to-noise ratios and reasonable acquisition times.Transmitted electrons are the primary electrons that pass through specimens if they are sufficiently thin (60–200 nm). The transmission of electrons through a sample depends on the atomic number; therefore, osmium fixation and heavy-metal-staining largely increase the contrast of the image. These electrons form images in TEM and can be detected by SEM using a scanning transmission electron microscopy (STEM) detector located under the ultrathin section (Figure 1).3. Multiscale Imaging of Large Sample AreasThe preparation of a resin-embedded tissue block for imaging in TEM frequently involves the cutting of semithin sections (0.5–1.5 μm), which are used to examine a large area of tissue with an optical microscope to locate the regions of interest for ultrastructural studies. Based on these observations, the block is trimmed to a size that enables the preparation of ultrathin sections (50–90 nm), which are placed on mesh grids and then contrasted with heavy-metal salts. Consequently, the investigated area has largely reduced overall dimensions and is divided into separate parts by grid bars, which additionally cover some portion of the sample area. Due to this separation, it is frequently difficult to recognize the histological context of TEM images. These problems can be overcome by the use of single-slot grids with supporting films, which enable obtaining an unobstructed image of the entire ultrathin section. The imaging of the entire ultrathin section or its large part at high resolution in standard TEM is frequently problematic because of the size of sensors in digital cameras and the requirement of a montage of many images into a larger one. A transmission electron microscope camera array (TEMCA) was constructed to increase the image acquisition efficiency in TEM [34], and this idea was recently developed as high-speed TEM [35].The detection of transmitted electrons in SEM using a STEM detector is a highly efficient method of imaging ultrathin sections [36,37,38,39]. The electron beam focused on a small spot scans the ultrathin section, and the image is formed by mapping, synchronously with the scan, the signal intensity below the sample. The advantages of this method include automatic imaging of large sample areas with a single frame of up to 24,000–36,000 pixels in each direction (available for many models of SEM), very low noise and high image quality, very high resolution, and significantly shorter acquisition time compared with other SEM methods of TEM-like imaging (Figure 2 and Figure 3). STEM imaging does not require any special staining of samples [38]; furthermore, the contrast in this method is higher than that in TEM because of the lower accelerating voltages (30 kV in SEM vs. 80–120 kV in TEM). Moreover, the transmission mode in SEM exhibits no chromatic aberration. The number of grids loaded simultaneously into the standard SEM holder is 6–12, whereas it is only 1 or 2 in the standard TEM holder. Many commercially available software packages dedicated to SEM provide automatic procedures for STEM imaging, such as overview imaging of all grids, autofocus, and automatic correction of astigmatism. The sections can also be transferred from SEM to TEM if required. Single-slot grids are particularly suitable for STEM detection because they enable an unobstructed digitalization of entire sections [39]. Carbon-coated formvar is frequently used as a support for the sections [39]. Recently, Dittmayer et al. [40] described a workflow to produce high-quality sections on large-slot grids coated with Pioloform films.The disadvantage of STEM detection as a method of large-area imaging is the necessity of preparing large ultrathin sections on slot grids that require special skills. Another problem is the focal change of the sample signal after longer exposure to the electron beam during image adjustment. This problem could be eliminated by sample pre-irradiation, which can be performed automatically using software macros [39].BSE detection enables the ultrastructural imaging of sections [17,18,19] placed on solid, electron-beam-impermeable, conductive supporting media, such as silicon wafers (Figure 4 and Figure 5, Supplementary Movie S1). In contrast to the collection of separate sections or very short section ribbons on grids for TEM, the ultrathin sections for imaging using BSE detection are frequently cut into long ribbons using a diamond knife with a large boat (jumbo knife), and these ribbons are attached to a silicon wafer (Figure 4A). A special manipulator is used to hold and move the wafer in a knife boat (Figure 6A). This technique enables the collection of numerous outsized ultrathin sections on a stable support. In contrast to imaging using the transmitted electrons, the semithin sections can be used for BSE imaging [18,19]. The advantages of semithin sections are a larger section area that provides more data about the histological context, compatibility with light microscopy, and a simple technique of cutting and collecting. The disadvantage is a frequently lower surface quality compared with the ultrathin section, and in some scenarios, higher susceptibility to charging. Reichelt et al. [18] and Rodiges et al. [19] demonstrated perfect results of the digitalization of semithin sections through BSE detection for the multi-scale imaging of animal tissues and for pathological diagnostics.Efficient, high-quality imaging using BSE requires a more intensive infiltration of tissue with heavy metals than for STEM and TEM imaging. Many protocols for sample fixation and en bloc contrasting for BSE imaging [24,41,42,43,44,45,46,47,48,49,50,51,52,53] have been proposed depending on the tissue properties and sample size (Table 1). The sections of tissues fixed with routine TEM methods and contrasted after cutting can also be imaged using BSE detection; however, they frequently require longer acquisition times; therefore, they are less useful for the digitalization of very large sample areas. However, these sections still facilitate better recognition of the histological context of the investigated structure compared with TEM imaging. According to our experience, intensive en bloc heavy-metal infiltration is a technique of choice for BSE imaging because it enables image acquisition at low dwell times, ensures a high signal-to-noise ratio, and eliminates the risk of section contamination during post-sectioning contrasting. The limitation of this procedure is the poor differentiation of chromatin structure and the different appearance of some cell components, such as secretory granules, compared with conventional TEM images.Silicon wafers provide perfect support for ultrathin and semithin sections because they are highly conductive. The glow discharge treatment of wafers is recommended to obtain a hydrophilic surface, which facilitates the collection of sections. In applications combining light and electron microscopy, indium-tin-oxide-coated glass coverslips or carbon-coated glass coverslips are used to collect ultrathin sections. Histological glass slides coated with 60–80 nm of carbon [18] or flat epoxy resin sheets coated with gold/palladium [19] can be used as a support for semithin sections.The ultrastructural imaging of resin-embedded biological samples using BSE detection requires a field emission source of electrons (Schottky emitter) that delivers a stable and relatively high current in the electron beam [54]. The beam current should be sufficiently high to provide acceptable image contrast and low noise at short dwell times, but it should not be excessively high. The increase in the beam current increases the acquisition speed; however, it negatively affects the image resolution [54]. The next requirement for SEM is to ensure very small beam diameters on the sample surface at low acceleration voltages. The diameter of the spot at which the electron beam hits the sample is a primary factor determining the resolution of SEM imaging, and a small volume of the reaction vessel ensured by the low acceleration voltage is a crucial factor for the resolution of BSE imaging (Figure 1). The efficiency of the detector is extremely important for high-quality BSE imaging with reasonable acquisition times. Both in-column detectors and retractable diode detectors have been successfully used to ensure a good image quality. Detection systems dedicated to biological samples are highly recommended. The scan generator should enable imaging with a single frame no smaller than 24,000 pixels in the x- and y-directions to decrease the stage movement and the necessity for image tilt montage. The final but very important component of the image acquisition system is software that allows automatic imaging of large sample areas. Research has demonstrated that the signal-to-noise ratio can be significantly improved by applying a negative bias voltage to the sample (beam deceleration) with a simultaneous increase in the acceleration voltage [54,55,56,57]. The increase in the recorded signal occurs owing to the re-acceleration of BSE in the bias field toward the detector. The deceleration retards the electron beam towards the sample and reduces the penetration of the primary electrons into the samples; therefore, a higher acceleration voltage does not result in an increase in the reaction vessel size [54]. For example, when the acceleration voltage is set to 5 kV and the deceleration voltage is set to −3 kV, the landing energy on the specimen surface is 2 keV.The imaging of large sample areas using a STEM or BSE detector is frequently performed in a hierarchical mode [58], from overview, low-resolution (200–15 nm/pixel) images of the entire sample through middle resolution (10–5 nm/pixel) images of large groups of cells and high-resolution (3–1 nm/pixel) images of small groups of cells or a single cell, to very high-resolution (>1 nm/pixel) images of cell parts and organelles (Figure 2, Figure 3, Figure 4 and Figure 5). Images with lower resolutions are used to locate targets for imaging with higher resolution. For BSE imaging, the pyramid of images is frequently based on digital macrophotography used for navigation and covers a scale from centimeters (length of section ribbons) to nanometers (Figure 4). It should be emphasized that the top of the pyramid of hierarchical imaging frequently contains images as large as 1 Gpixel and their montages. Moreover, several regions of interest can be digitalized with a very high resolution (multi-top pyramid). Large-area imaging enables the creation of a virtual ultrathin slide, which can be zoomed from the millimeter scale to the nanometer scale.The effective imaging of flat surfaces of resin-embedded biological samples has resulted in the development of 3D imaging techniques, including serial section imaging (array tomography), serial block-face imaging (SBF-SEM), and focused ion beam SEM (FIB-SEM). The last technique requires dual-beam SEM; therefore, it is not presented in this article.4. Array TomographyThe concept of array tomography was proposed by Micheva and Smith in 2007, mostly as a method of high-resolution, volumetric imaging of large numbers of antigens visualized using immunofluorescence [59]. The name “array tomography” was introduced because of the use of serial ultrathin sections to obtain 3D information. The imaging of ultrathin sections enabled the prominent improvement of resolution in the z-axis compared with confocal microscopy. Moreover, the cited authors demonstrated that the sections of acrylic resin-embedded tissue could be repeatedly stained with different antibodies, then contrasted with heavy metals and used for ultrastructural imaging in FE-SEM [59]. At the same time, Kasthuri et al. [60] reported the use of automatically cut and collected serial ultrathin sections using a tape ultra-microtome [61] for ultrastructural imaging in SEM to obtain 3D ultrastructural reconstructions. The automation of section preparation highlighted the requirement for the development of high-sensitivity BSE-based imaging systems and automated image acquisition software [27]. As an alternative to the tape ultramicrotome, which at this time was only a prototype, Horstmann et al. [22] described the use of serial ultrathin sections collected on a silicon wafer for ultrastructural volume imaging. The tape ultramicrotome was commercialized by Boeckeler Instruments, Inc., under an early adopters program in 2015. The last achievement in the development of array tomography as a method of 3D ultrastructural studies was the introduction of multi-beam SEM, which largely increased the image acquisition capability [62,63].The term array tomography, depending on application, comprises three different techniques: (i) fluorescence microscopy array tomography, which delivers volumetric, high-resolution data on the distribution of molecules and enables the detection of several antigens in the same section [59,64], (ii) electron microscopy array tomography, which enables the capturing of ultrathin sections for 3D ultrastructural studies [65], and (iii) correlative array tomography [66,67,68], which combines fluorescence imaging and electron microscopy imaging to obtain voxel-level associations between structure and chemistry. The new concept is the use of array tomography to locate targets for z-axis high-resolution imaging through FIB-SEM [69]. Electron microscopy array tomography should be subdivided into the method using a manual collection of sections (Figure 6A) and the method of using an automated tape-collecting ultramicrotome (ATUM). The first method is frequently limited to hundreds of sections in the form of ribbons attached to a silicon wafer, and the second method enables the collection of thousands of sections on a tape (Figure 6B–D). For the first method, the direct location of the section on the silicon wafer frequently eliminates the charging problem and ensures very high-quality imaging; however, it is limited in size owing to the manual collection of sections. These sections are imaged in conventional FE-SEM within a reasonable period, and the size of the data is not extremely large. Recently, new devices for the automatic collection of serial samples on silicon wafers or magnetic tapes have been described [70,71,72].The use of ATUM enables the automatic collection of thousands of ultrathin sections on tape while they are simultaneously cut with an ultramicrotome (Figure 6B,C). It operates by moving a plastic tape from a supply reel through a tape snout located in a water-filled diamond knife boat to a take-up reel. The sections are collected from the water on the surface to the tape. Glow-discharged Kapton tape is the most commonly used tape in ATUM. After section collection, the tape is cut into smaller strips, which are mounted using double-sided carbon tape to a silicon wafer with a diameter of 100 mm (Figure 6D). Section contrasting on tapes is possible if required. The Kapton tape is nonconductive; therefore, the wafers with tape strips must be coated with carbon to eliminate charging artifacts. However, this coating may be problematic for imaging using SE detection; therefore, the deposition of carbon on the tape before collecting the sections is advised in this scenario. Recently, conductive carbon nanotube-coated polyethylene terephthalate tape has been proposed to eliminate charging problems and increase the imaging quality [73]. Dedicated software is available for imaging serial sections in a semiautomatic manner. The acquisition of very large sets of sections may require several weeks using a single FE-SEM. The time required for the digitalization of large sets of sections can be dramatically reduced using multibeam SEMs, operating with 61 or more electron beams, and enabling the simultaneous imaging of numerous pixels [62].The most important advantage of array tomography is its non-destructive nature, which enables the reimaging of sections with different resolutions and regions of interest. It is possible to build archives of sections for further studies. The imaging procedure can begin with the digitalization of every n serial sample in an array to obtain a volumetric overview and select the region of interest. Array tomography facilitates a large versatility in terms of the size of the imaged area and resolution. Very high-quality imaging with a pixel size of 1 nm or below is easily achievable for samples placed directly on a silicon wafer. This method enables the study of archival samples fixed with the conventional protocol for TEM, because of the possibility of section contrast. However, note that the best results are obtained after the intensive infiltration of samples with heavy metals because of the lack of charging and section contrasting artifacts. Array tomography enables correlative light and electron microscopy studies, and their combination with 3D analyses [66,68,74]. The disadvantages of this method are the time-consuming preparation of sections for imaging and the alignment of sections into z-stacks.5. Serial Block-Face ImagingThe SBF-SEM technique is based on removing the top layer of a resin-embedded sample using an ultramicrotome with a diamond knife inside the specimen chamber of FE-SEM and then imaging the exposed surface using a BSE detector (Figure 7 and Figure 8). The steps of cutting and imaging the resin block are repeated hundreds or thousands of times to obtain a z-stack for 3D analysis. The first use of a microtome inside the SEM chamber was reported by Leighton in 1981 [75]; however, the system for automatic block cutting and imaging was constructed more than 20 years later by Denk and Horstmann [20]. This achievement was possible because of developments in SEM and computer technologies. The system of Denk and Horstmann was commercialized by Gatan, Inc., and called 3View (Figure 7). The ultramicrotome in Gatan 3View is mounted to a special chamber door, which replaces the standard door in many SEM models when SBF-SEM is used. It also includes a dedicated high-sensitivity BSE detector. In 2015, FEI Company (now Thermo-Fisher Scientific Inc., Waltham, MA, USA) introduced a microtome for montages on the regular SEM stage in a dedicated FEI microscope. A miniature ultramicrotome for montages on the SEM stage was demonstrated at the Microscopy and Microanalysis Congress in 2019 in Portland, USA, by ConnectomX Ltd. (Grove, Wantage, UK).The main advantage of SBF-SEM is the possibility of obtaining a well-aligned z-stack of thousands of images in a fully automatic manner. The setup procedure of SBF-SEM systems is easier and faster than the preparation for imaging in ATUM array tomography, which includes the collection of sections on tape, montage of tape strips on wafers, overview imaging, and selection of regions of interest [65,76]. The SBF-SEM results in a z-stack of images, which may require only a few alignments performed automatically. However, the success of the SBF-SEM method is critically dependent on sample fixation and embedding. The block of the resin-embedded sample must be conductive to avoid charging artifacts and must be resistant to damage by the electron beam (Figure 9). In contrast to section imaging, post-embedding contrasting and carbon coating of the cutting surface are not possible in SBF-SEM; therefore, both the signal-to-noise ratio and conductivity cannot be improved after resin polymerization. Thus, sample preparation must be conducted carefully. The protocol proposed by Deerinck et al. [41] has been successfully used to prepare many types of tissues, but the best results are obtained with the nervous tissue, which is rich in lipids. Adaptations of this protocol to other tissues include various modifications, such as the use of tannic acid and ruthenium red for contrasting collagen or desmosomes [45,77]. Hua et al. [47] proposed a modification that enables the fixation of large samples. Resin embedding and mounting to the pin holder are also important for obtaining satisfactory results [76]. Generally, the resin formulation should ensure the highest hardness after polymerization, and the block must be trimmed to remove the empty resin. For an electrical connection between the sample and stage, the bottom side of the block should contain tissue that is in contact with the sample holder pin, the conductive glue should be used for block fixation to the pin, and the block should be coated with a thin layer of gold.Three different strategies are used to eliminate the problems caused by tissue block charging during imaging. The oldest one is the use of a low vacuum in the specimen chamber (variable pressure SEM); however, this method induces a large decrease in the signal-to-noise ratio [78]. Despite these disadvantages, this method is commonly used. The second method is conductive embedding, which can be obtained by the addition of carbon particles to the resin or other modifications; however, its effectiveness is currently rather low [79,80]. The third is the focal charge compensation (FCC) proposed in 2017 by Deerinck et al. [81]. FCC is based on the application of nitrogen directly on the block face using a special nozzle during imaging. The dosing of gas is very low; therefore, the high vacuum of the specimen chamber is still maintained (10−4–10−3 mbar). The locally applied nitrogen gas molecules are ionized, contact the sample surface, and neutralize electrons, which charge the sample. FCC effectively reduces image artifacts (Figure 9A,B); thus, it even enables image acquisition from samples prepared without dense heavy-metal staining. The resolution of imaging with FCC is nearly the same as that without nitrogen application.The methods based on the mechanical cutting of samples to obtain 3D information have a lower spatial resolution along the z-direction compared with the x- and y-directions because of ultramicrotomy limitations. The microtome effectively removes sections with a thickness larger than 20 nm, whereas the lowest pixel size in the x–y plane is 3–5 nm. As a solution, a deconvolution technique was developed to obtain additional virtual layers in the sample by acquiring images at different primary beam energies and then processing the image stacks using a multi-energy deconvolution algorithm [82,83].The disadvantages of SBF-SEM are generally opposite to the advantages of array tomography: (i) SBF-SEM is destructive and the tissue cannot be re-examined, (ii) the region of interest is selected based on the first image in the z-stack; therefore, the block of tissue is examined blindly, and (iii) the tissue has to be stained en bloc without an alternative solution. The lateral resolution in SBF-SEM is lower than that in array tomography with sections mounted directly on silicon wafers.6. ConclusionsModern FE-SEM is a highly powerful and versatile tool for the studies of cell and tissue ultrastructure alone or in combination with the biochemical data provided by fluorescence microscopy. It enables TEM-like imaging of the cut surfaces of tissue blocks, which are impermeable to the electron beam, or ultrathin or semithin tissue sections mounted on the solid substrates, such as a silicon wafer or glass slide. This feature simplifies and accelerates ultrastructural studies of both large areas and volumes of biological samples. However, efficient, high-quality imaging using BSE requires a more intensive infiltration of tissue with heavy metals than for TEM imaging. The imaging of large sample areas in SEM can be performed using two methods based on the detection of BSE or transmitted electrons. The second method is limited to the ultrathin section; however, it provides very high-quality images of conventionally fixed samples. The imaging of large sample areas enables the creation of a virtual slide, which can be zoomed from the millimeter scale to the nanometer scale. Volume imaging techniques comprise AT and SBF-SEM. In AT, serial ultrathin sections are collected manually on a solid substrate or automatically on a tape using a special ultramicrotome. The imaging of serial sections is used to obtain three-dimensional information. SBF-SEM is based on removing the top layer of a resin-embedded sample using an ultramicrotome inside the SEM specimen chamber and then imaging the exposed surface with a BSE detector. This process is repeated to produce a digitized stack of aligned images, which allows to follow cell-to-cell arrangements or intracellular structures in the z-direction.

animals : an open access journal from mdpi

10.3390/ani12010092

PMC8749819

Economic analysis can be used before animal experiments to estimate the change in production that would be needed for experimental treatments to be as equally profitable as the control treatment. After the experiment, the results can be examined to assess if the production threshold was met. This approach was applied before and after an animal experiment that tested the effect of feeding dietary supplements on the milk production of dairy cows experiencing a heat event in climate-controlled chambers. Heat stress reduces the milk yield of cows, but the inclusion of supplements such as betaine and fat could lessen the impact. The pre-experimental economic threshold analysis showed that cows fed a diet containing betaine, fat, or betaine plus fat would need to produce 1%, 9% and 11% more milk, respectively, to be equally as profitable as the control diet. Results from the subsequent animal experiment combined with previously modelled projections of heat stress conditions showed that supplementing diets with fat or betaine, but not in combination could exceed the milk production threshold required to be as profitable as the control diet.

Ex ante economic analysis can be used to establish the production threshold for a proposed experimental diet to be as profitable as the control treatment. This study reports (1) a pre-experimental economic analysis to estimate the milk production thresholds for an experiment where dietary supplements were fed to dairy cows experiencing a heat challenge, and (2) comparison of these thresholds to the milk production results of the subsequent animal experiment. The pre-experimental thresholds equated to a 1% increase in milk production for the betaine supplement, 9% increase for the fat supplement, and 11% increase for fat and betaine in combination, to achieve the same contribution to farm profit as the control diet. For the post-experimental comparison, previously modelled climate predictions were used to extrapolate the milk production results from the animal experiment over the annual hot-weather period for the dairying region in northern Victoria, Australia. Supplementing diets with fat or betaine had the potential to produce enough extra milk to exceed the production thresholds, making either supplement a profitable alternative to feeding the control diet during the hot-weather period. Feeding fat and betaine in combination failed to result in the extra milk required to justify the additional cost when compared to the control diet.

1. IntroductionEx ante economic threshold analysis uses the prices of inputs and outputs to estimate the minimum extra benefit required to justify an investment, after allowing for an opportunity cost of capital [1,2]. In agriculture, this approach can be used to estimate the amount of extra output required to make investment in a new management practice economically equivalent to the current practice. For example, Henty and Griffith [3] used this approach to estimate the minimum improvement in milk production needed during heat stress events for an investment in shade infrastructure to return a net present value of 0 Australian Dollar (AUD) at a 10% real discount rate. In dairy nutrition research, the expected size of the production response from an experimental diet compared to the control diet is unknown prior to the experiment being conducted. Economic threshold analysis can be used before an experiment to assess whether the required production response, referred to here as the production threshold, from an experimental diet is likely to be biologically feasible. This information can be used to assess whether the idea in question warrants further investigation, and to enable prioritization and allocation of finite resources to experiments [2].The objective of this study was to demonstrate how ex ante economic threshold analysis could be applied in a research context, using the example of an experiment where heat stress was induced in dairy cows. Digestion and metabolism of feed generates heat that contributes to the total heat load on dairy cows and when this load cannot be dissipated due to environmental conditions, dry matter intake (DMI) and milk production (milk yield, milk fat and protein yield) are compromised [4,5]. The frequency and severity of heat stress events impacting the dairying regions of Australia are expected to increase in the future [6,7]. The animal experiment was designed to investigate whether supplementing dairy cow diets with betaine and/or fat (as canola oil) could alleviate the negative impacts of a moderate, short-term heat challenge on DMI, milk production and body temperature of the cows, compared to these same parameters in cows fed a control diet [8]. Adding betaine to the diet of animals has been proposed as a method to increase an animal’s resilience to a heat load by potentially reducing cellular damage and reducing the chance of endotoxins escaping the gut moving into the body [9]. This acts to reduce the potential effects of hot weather on the animal and decrease some of the negative impacts on milk production [10,11]. Fat has a lower heat increment than other feeds [12] and feeding high-fat diets to cows during hot weather has been shown to produce more energy-corrected milk than when cows were given lower fat diets [13,14]. While it was hypothesized in the animal experiment that milk production would increase by the dietary inclusion of either betaine supplement and/or additional fat, compared to the control diet, the expected amount of extra milk produced was unknown.The minimum change in milk production needed for each of the proposed experimental diets to be a profitable alternative to feeding the control diet was estimated by economic threshold analysis during the planning of the heat-stress experiment. Once the animal experiment was completed, the pre-experimental milk production thresholds were compared to the measured differences in milk production from feeding each experimental diet and the control diet. The comparison was made in the context of the Murray Dairy region of northern Victoria and southern New South Wales, Australia. The Murray Dairy region is a major dairy production area, with around 1200 dairy farms producing approximately 20% (1840 million litres) of Australia’s milk [15]. In the region, hot weather periods occur during spring, summer and early autumn, with on average 60 days above 30 °C and 21 days above 35 °C annually, and 74 days where the temperature-humidity index (THI) is 75 or more [7,16]. When THI exceeds 68, cows are likely to experience heat stress [17]. The frequency and severity of extreme temperature events in the Murray Dairy region is expected to increase and further exacerbate production losses in the future [18]. The post-experimental comparison was conducted to examine whether feeding each experimental diet to dairy cows during the annual hot-weather period in northern Victoria could be a profitable alternative to feeding the control diet. The control diet was designed to emulate diets commonly fed on Australian dairy farms during summer.Our paper reports (1) a pre-experimental economic analysis to estimate the milk production thresholds for an experiment where dietary supplements were fed to dairy cows experiencing a heat challenge, and (2) a comparison of the milk production thresholds to the results of the subsequent animal experiment. The detailed method, results, and in-depth discussion of the animal experiment are reported elsewhere [8].2. Materials and Methods2.1. Pre-Experimental AnalysisThe extra costs associated with feeding each experimental diet compared to feeding the control diet alone was used to establish the milk production thresholds. The diets proposed in the animal experiment were formulated as a mixed ration with only the supplement component changing. The three experimental diets were (1) BET—control diet plus 16 g betaine (trimethylglycine as a powder; Feedworks, Romsey, Victoria, Australia) per cow per day, (2) FAT—control diet plus 0.7 kg canola (Brassica napus L.) oil per cow per day, or (3) BET + FAT—16 g betaine (trimethylglycine) and control diet plus 0.7 kg canola oil per cow per day. The daily control diet offered to each cow was a total mixed ration (TMR) comprised of 7 kg dry matter (DM) lucerne hay (Medicago sativa L.), 6 kg DM pasture silage (predominantly perennial ryegrass, Lolium perenne L.), 5.0 kg DM grain mix (500 g/kg wheat grain (Triticum aestivum L.), 500 g/kg barley grain (Hordeum vulgare L.), 1.5 kg DM solvent extracted canola meal (Brassica napus L.), 0.2 kg DM of minerals and vitamins (Ca 134 g/kg, Mg 110 g/kg, P 60 g/kg, Zn 6.4 g/kg, Mn 2.4 g/kg, Cu 1.2 g/kg, I 80 mg/kg, Co 100 mg/kg, Se 24 mg/kg, Vitamin A 165 IU/g, Vitamin D3 24 IU/g, Vitamin E 800 mg/kg), 0.1 kg DM salt, and 42 mL of Bloat Drench (271 g/L alcohols, C12-15 ethoxylated; VicChem, Coolaroo, Victoria, Australia). The chemical compositions of the main dietary components are given in Table 1.The amount of canola oil fed was restricted to 0.7 kg/cow per day so that the dietary fat concentration did not exceed 70 g/kg DM to avoid a reduction in ruminal fermentation or voluntary DM intake [19]. The betaine dose rate (16 g/cow per day) was the average of dose rates used in experiments from published literature where there had been a positive production response or reduction in heat stress from feeding betaine [10,11,20]. The dose rate also aligned with that recommended by the manufacturer of 15 to 20 g/cow per day [21].For our pre-experimental economic analysis, all diets were evaluated on a DM offered basis as intake was yet to be measured. Marginal changes in the total DM offered with each diet, such as in conserved fodder or pasture that may be needed to achieve the required changes in milk production, were assumed to come from feed which would otherwise have been grown or fed, but not utilized. Therefore, there were no other costs in addition to the costs of the dietary supplements. It was assumed that the diets were fed every day for 212 days from 1 September, prior to the peak period for hot weather, to 31 March, the end of the hot weather period in northern Victoria. The benefits to milk production from feeding the experimental diets were assumed to be realized within the 212-day feeding period and that no carryover benefits occurred for the remainder of lactation. For each diet, the additional cost of the dietary supplement was calculated assuming an average price of AUD 8.50/kg for betaine and AUD 1.40/kg for vegetable oil, including delivery to northern Victoria, Australia [21,22]. The price of vegetable oil was used instead of canola oil, as this is a more commonly used component of livestock diets, and the two oils were deemed interchangeable as fat supplements. Compared to the control diet, this equated to an extra cost of AUD 0.14/cow per day for the BET diet, AUD 0.98/cow per day for the FAT diet, and AUD 1.12/cow per day for the BET + FAT diet. An 8% p.a. real opportunity cost of variable capital, equating to 4.65% for the 212-day period, was used to calculate the economic threshold (Equation (1)) using an average Victorian real milk (fat + protein) price of AUD 6.28/kg for the period 2013-14 to 2018-19 [23,24]. A litre of milk was assumed to contain 72 g of milk fat plus protein, referred to as milk solids in this analysis. Results were tested using a price sensitivity analysis where the average supplement and milk prices were adjusted by ±20%.MSe = ($c − $e × (1 + r))/$MS(1) where MSe equals the extra milk solids required over the 212 days from feeding the experimental diet for it to make an equal contribution to farm profit as the control diet, $c and $e are the costs of the control and the experimental diets, respectively for 212 days, r is the real opportunity cost of variable capital for the 212 days, and $MS is the price per kg of milk solids.While producers may implement dietary strategies for reasons in addition to immediate milk production responses, such as animal welfare, these factors were not accounted for in the economic analysis presented here.2.2. Animal Heat-Stress Experiment SummaryThe animal feeding experiment was conducted in 2018 at Agriculture Victoria Research, Ellinbank (38°14′ S, 145°56′ E), in accordance with the Australian code of practice for the care and use of animals for scientific purposes [25]. The animal experiment is reported in full elsewhere [8] but is described briefly here. Thirty-two multiparous, lactating Holstein-Friesian cows producing 18.6 ± 2.37 kg milk/day (mean ± standard deviation) with 566 ± 47.1 kg bodyweight, 216 ± 18.5 days in milk, 2.7 ± 0.79 parity and 101 ± 2.9 heat tolerance breeding value (DataGene, Bundoora, Victoria, Australia; 100 = national breed mean) were managed in four cohorts of eight cows. Each treatment was assigned at random to two cows within each cohort with treatment groups balanced for bodyweight, milk yield, days in milk and heat tolerance breeding value. Six of the eight cows in each cohort were assigned to six controlled-climate chambers for the heat challenge according to a row-column design. The remaining two cows were retained as spares. Cows had a 7-day covariate period, 14 days adaptation to treatment diets, 3 days of measurement pre-challenge, a 4-day heat challenge where cows were exposed to a moderate heat challenge in controlled-climate calorimeters [26], followed by a 7-day recovery period under ambient conditions.All cows were offered a common diet during the covariate period comprising 5 kg DM/cow of wheat and 15 kg DM/cow of lucerne hay fed in a paddock per day before being adapted to each treatment diet over a 14-day period. Treatment diets were offered in two equal amounts after the morning and afternoon milkings for 3.5 h each feeding period. Cows were milked twice daily at ~0600 h and ~1500 h and milk yield was measured for each cow at each milking using in-line milk meters. Samples of milk for composition analysis were collected on each day of the covariate period, one day of the pre-heat challenge period (the day closest to thermoneutral conditions) and on each day of the recovery period. During the heat challenge, milk yield was measured by collecting and weighing milk from individual cows at every milking. Milk samples were taken at every milking and milk fat and protein measured using a near infrared milk analyser (model 2000, Bentley Instruments, Chaska, MN, USA). Body temperature was recorded at 15-min intervals during the experiment using intravaginal loggers [27]. Rectal temperature was also measured during the heat challenge at 0600 h and 1500 h.Temperature-humidity index (THI) was used to define the climate conditions experienced by the cows during the experiment, which replicated those used by Garner et al. [27]. Conditions during the 4-day heat challenge were set to mimic diurnal patterns of previously recorded summer conditions in northern Victoria, with mild heat stress conditions of THI 80 from early morning to midday, increasing to moderate heat stress of THI 84 from midday until evening, and decreasing to mild heat stress conditions of THI 74 overnight.The results of the animal experiment were subjected to statistical analysis using the following linear model:y = μ + βycov + F + B + FB + C + K + ε where y was the outcome variable of interest, ycov was the same variable if available from the covariate period, F was a factor for the effect of added fat, B was a factor for the effect of added betaine and FB their interaction, C was an effect of chamber, K as an effect of cohort and ε a random error for individual animal. The model was applied using ReML software in GenStat 20 (VSN International Ltd., Hemel Hempstead, UK) with fixed effects for the covariate, cohort, chamber, and factorial fat by betaine diet treatments, and with animal as a random effect. Residuals were examined graphically to check distributional assumptions of normality and constant variance.2.3. Post-Experimental ComparisonResults from the animal experiment were combined with the future projections of the mean number of heat stress days (average daily THI greater than 75) predicted from 1 September to 31 March (212 days) for 2025 for the Murray Dairy region by Nidumolu et al. [7,18] to estimate the likely milk solids benefit from feeding each supplement diet for the full feeding period. Across the broader Murray Dairy region, the predicted mean number of consecutive heat stress days for 2025 ranged from three to five days, depending on the severity of the heat challenge [18]. The 4-day heat period of the animal experiment should therefore reasonably represent the type of heat challenges predicted to be experienced in the Murray Dairy region by 2025. Two discrete scenarios were analyzed.2.3.1. Scenario 1—Benefit Only on Days where THI ≥ 75The first scenario assumed that feeding each experimental diet would provide a benefit to milk solids production over the control diet only on days when THI was equal to or greater than 75. It was assumed that there were 91 days between 1 September to 31 March where THI was equal to or greater than 75, equal to the 2025 prediction for Tatura (36.4367° S, 145.2328° E) in northern Victoria by Nidumolu et al. [18]. For each of these days, the benefit was calculated as the mean extra milk production (kg milk solids/cow per day) measured during the heat challenge period of the animal experiment, minus the mean extra milk production (kg milk solids/cow per day) during the thermoneutral pre-challenge period of the animal experiment. Subtracting the increase in milk solids production during the pre-challenge period from that during the heat period corrected for the effects of feeding each experimental diet per se. On the remaining 121 days where THI was less than 75, it was assumed the milk solids production from cows fed each experimental diet was the same as that of cows fed the control diet.2.3.2. Scenario 2—Benefit for All 212 DaysThe second scenario aimed to capture all the benefits observed in the experiment of feeding each experimental diet over the control diet. On each of the 91 days where THI was greater than 75, the mean extra milk production (kg milk solids/cow per day) measured during the heat challenge period of the animal experiment was used. On the remaining days where THI was less than 75, each experimental diet was assumed to have a milk production response above the control diet equal to the mean extra milk production (kg milk solids/cow per day) measured during the pre-challenge and recovery periods of the animal experiment.3. Results3.1. Pre-Experimental AnalysisThe pre-experimental economic threshold analysis showed that feeding the BET diet over 212 days would need cows to produce 3.2 kg more milk solids per cow to provide the same contribution to farm profit as feeding the control diet (Table 2). This equated to 1 to 2% extra milk solids production for a herd being fed the control diet and averaging 1.7 kg milk solids/cow per day (23 kg milk/cow per day). This required increase was relatively small compared to that resulting from feeding the more expensive FAT diet which required an additional 23.1 to 51.9 kg milk solids/cow (6 to 14% across the herd) over the 212 days for the contribution to farm profit to be the same as the control diet. The combined BET + FAT was the most expensive needing an extra 26.3 to 59.1 kg milk solids/cow to be as profitable as the control diet.3.2. Animal Heat-Stress Experiment SummaryOnly the results from the experiment relevant for the economic analysis are presented here. Complete results from the animal experiment are reported in Williams et al. [8]. Milk solids production during the pre-challenge, heat challenge, and recovery periods of the experiment for the different diets are given in Table 3. Statistical differences in milk solids production were only detected by Williams et al. [8] for some treatments. However, the numerical differences between treatments are still useful for an economic analysis and these are reported here. Results from the animal experiment were constrained by the number of animals that completed the full heat challenge period. Raw data within the 3-day pre-challenge period were averaged. Day 2 of the heat challenge was taken to represent the average for the heat event period, a compromise between the heat challenge having an effect on the cows and retaining sufficient animals in the analysis to enable conclusions to be drawn (Williams et al. [8]). The average daily results for the full 7-day recovery period were used.During the heat challenge and recovery periods of the experiment, all cows fed the three experimental diets increased milk solids production compared to those fed the control diet (Table 3). Cows fed the BET diet produced an additional 0.07 and 0.05 kg milk solids/cow per day (~1.0 and ~0.7 kg milk/cow per day) more than cows fed the control diet during the heat and recovery periods, respectively, with no additional benefit recorded during the pre-challenge period (Table 3). Of the three experimental diets, the FAT diet had the greatest increase in milk solids production in all periods compared to feeding the control diet, with an average daily benefit of 0.31 kg milk solids/cow per day (~4.3 kg milk/cow per day) during the heat challenge (Table 3). The combined BET + FAT diet increased milk solids production in all periods compared to the control diet, with the greatest benefit of 0.11 kg milk solids/cow per day (~1.5 kg milk/cow per day) recorded during the heat challenge period.3.3. Post-Experimental Comparison3.3.1. Scenario 1—Benefit Only on Days where THI ≥ 75When benefits during hot weather only were accounted for (Scenario 1), the BET diet was the only diet to be a profitable alternative to the control diet where average supplement and milk prices were assumed. It was estimated that the BET diet would produce an extra 6.4 kg milk solids/cow over the 212-d feeding period compared to the control diet, exceeding the required 4.8 kg milk solids/cow pre-experimental threshold for the diet (Figure 1a). With high supplement price and low milk price conditions, the BET diet was unable to meet the required pre-experimental milk production threshold and was not a profitable alternative to the control diet.There was no milk production benefit attributed to the FAT diet over the control diet for Scenario 1 (Figure 1b). Feeding the FAT diet produced a similar amount of extra milk solids during the heat challenge period (0.32 kg milk solids/cow per day) as the pre-challenge period (0.31 kg milk solids/cow per day).For Scenario 1, the BET + FAT diet was estimated to increase milk production by 3.6 kg milk solids/cow compared to the control diet, a much smaller benefit than the required 39.4 kg milk solids/cow threshold with average supplement feed and milk prices (Figure 1c).3.3.2. Scenario 2—Benefit for All 212 DaysBoth the BET diet and the FAT diet were profitable alternatives to feeding the control diet under the assumptions of Scenario 2 (Figure 1a,b). The BET diet was estimated to increase milk production by 9.4 kg milk solids/cow over the 212-d period, with most of the benefit occurring on days where THI ≥ 75 (Figure 1a). Feeding the FAT diet was estimated to produce an extra 63.3 kg milk solids/cow over the 212-d period compared to feeding the control diet, exceeding the required 34.6 kg milk solids/cow under average supplement price and milk price conditions (Figure 1b). Despite the greater milk production benefits in Scenario 2, the BET + FAT diet still failed to be a profitable alternative to the control diet (Figure 1c). The findings of Scenario 2 were consistent across all supplement and milk price combinations tested.4. DiscussionThe milk production thresholds estimated in the pre-experimental analysis were biologically feasible, equating to a 1 to 11% increase in milk solids for the experimental diets to be equally as profitable as feeding the control diet. For example, a 7% increase in milk yield has previously been achieved by feeding cows a high fat diet during summer compared to a control diet [13]. After the animal experiment was completed, the ex ante milk production thresholds were compared to the experimental results using scenario analysis. This post-experimental comparison showed that the BET and the FAT diets have potential to be profitable alternatives to the control diet over the hot-weather period experienced in northern Victoria. We acknowledge that numerical, rather than statistically significant, differences in milk production [8] were used in this analysis, and that further research is required to confirm our findingsDespite generating the smallest milk production response of the three experimental diets, feeding betaine had the greatest potential to be a profitable alternative to the control diet. Cows fed the BET diet produced numerically the same amount milk as those fed the control diet during the pre-challenge period but produced on average 0.07 to 0.05 kg milk solids/cow per day more during the heat challenge and recovery periods, respectively. In the post-experimental economic analysis, accounting only for the heat-induced milk production benefits on days where THI ≥ 75 (Scenario 1), the BET diet exceeded the required milk production threshold under average supplement and milk price conditions. If supplement cost was high and milk price was low, the benefits of feeding the BET diet under hot weather conditions alone were not enough to justify the additional cost compared to the control diet. When the benefits recorded during the recovery period of the animal experiment were incorporated into the post-experimental comparison (Scenario 2), the BET diet was a profitable alternative to the control diet under all supplement and milk price conditions tested. The link between the benefits of feeding the BET diet and hot weather conditions means that if conditions vary from those used here for northern Victoria, based on the analysis of Nidumolu et al. [7,18], the profitability of this dietary strategy for managing heat stress will also likely change. This warrants further research under different climatic conditions to understand the industry-wide applicability of betaine as a dietary intervention strategy.Feeding extra fat alone required cows to produce additional milk solids on each day of the hot weather period to be a profitable alternative to the control diet. Cows fed the FAT diet during the animal experiment produced the greatest amount of milk solids of all treatments. The numerical increase in milk production was not limited to the heat and recovery periods, with a response in milk solids production also recorded during the pre-challenge period, a thermo-neutral period not associated with heat stress. This indicates a benefit for milk production from feeding extra fat per se, irrespective of heat stress conditions and was likely due to higher total energy intake during the pre-challenge period [8]. The post-experimental economic analysis showed that under the 2025 climate conditions predicted for northern Victoria by Nidumolu et al. [7,18], and accounting for the benefits on both hot days (THI ≥ 75) and on days where THI < 75 (Scenario 2), the FAT diet exceeded the required milk production economic thresholds under all price conditions tested. The milk production responses used in the post-experimental comparison of the FAT diet were similar for all days over the hot weather period (extra 0.31 kg milk solids/cow when THI ≥ 75, 0.29 kg milk solids/cow when THI < 75). Practically, this indicates that the benefits of feeding the FAT diet may be insensitive to changes in the frequency of hot days over the September to March period. Further research is required to confirm this result over a longer time than the duration of the animal experiment used here. Dietary supplementation with fat was associated with lower DMI during the heat-challenge period in the animal experiment [8] and this relationship may alter the milk response generated over longer-term exposure to periods of hot weather.Cows fed betaine and fat in combination did not produce enough extra milk solids to be a profitable alternative to the control diet. This finding held regardless of the scenario used or price conditions tested. During the animal experiment, cows fed the BET + FAT diet exceeded the milk solids produced by cows fed the control diet, but produced less milk solids than those fed the FAT diet. This result was unexpected, and Williams et al. [8] suggest this could be the result of an antagonistic interaction between betaine and canola oil for milk yield. With a combined lower cost and greater milk solids production, feeding the FAT diet was a more profitable option than the BET + FAT diet. However, further research is required to confirm and understand the interaction between the supplements over the longer-term.The post-experimental analysis used simplified assumptions and the numerical results from a single, short-term experiment to compare to the pre-experimental threshold estimates. A more nuanced approach may yield different conclusions. For example, the cumulative effects of heat stress during a heat challenge on the milk production of dairy cows have previously been reported. Garner et al. [27] showed that the energy corrected milk yield (kg/cow per day) declined by approximately 10% on the first day cows were exposed to a moderate heat challenge, compared to the pre-experimental energy corrected milk yield under ambient conditions. Energy corrected milk yield was reduced by 50% by the fourth day. Accounting for these cumulative effects of heat stress on the milk production of dairy cows as a heat event progresses, as well as the subsequent interaction with dietary intervention strategies, will be important in future analyses. It was also assumed in the post-experimental analysis that the size of the milk solids production benefit was uniform on all days where THI ≥ 75, regardless of the degree of heat stress experienced. However, there are differences in the number of days predicted to induce mild (THI ≥ 75), moderate (THI ≥ 78) and severe (THI ≥ 82) heat stress under future climate conditions [7]. The animal experiment used as a data source in our study was not designed to test the responses of cows offered different diets to different severities of heat stress, and responses may differ depending on the severity of the heat challenge. This is an important area for future work because the profitability of investments to reduce milk production losses from heat stress are influenced by the climatic conditions experienced, with the profitability of an intervention increasing with the severity of heat events [3]. Furthermore, accounting for interactions between dietary interventions and both the length and the severity of heat events will assist in expanding the economic analysis to other dairy regions beyond the northern Victoria example used here. If the full milk production benefits reported here can be achieved in a commercial farm context, dietary supplementation may offer an alternative or complementary strategy to longer-term investments for managing heat stress, such as shade infrastructure [3] and genetic improvement [26,28]. Changes to milk production were the focus of this analysis, but dietary supplementations are likely to impact ruminants in other ways, with changes in body temperature and dry matter intake also reported [8,12,20,29,30]. These additional considerations may have important, longer-term implications for animal production, welfare and profitability which should be accounted for in individual decision making.5. ConclusionsPre-experimental economic threshold analysis can be used in research prior to an experiment to identify the minimum production responses required for an experimental treatment to be a profitable alternative to the control treatment. The biological feasibility of achieving the production threshold can then help with experimental design and resource allocation. Comparing the estimated pre-experimental production thresholds to the results of the subsequent experiment can then provide a useful insight into the likelihood of achieving the required production benefits in a commercial farm context.

animals : an open access journal from mdpi

10.3390/ani13071156

PMC10093112

Domestic duck populations in China have abundant genetic resources, and this study is the first to evaluate domestic and wild Mallard populations. To explore the genetic diversity and ancestry of duck populations, we analyzed a sequence (619–667 bp) of mitochondrial DNA (mtDNA) in the control region. Two Muscovy duck breeds were included in this analysis, along with wild Mallard ducks, Eastern spot-billed ducks, and 22 domestic duck breeds. Significant genetic diversity was found among domestic duck breeds. These results not only provide valuable insights into the genetic makeup of domestic ducks but also provide clues to the possible origin of domestic ducks in China.

China has a rich genetic resource in its 34 domestic duck breeds. In order to detect the genetic diversity and explore the origin of these indigenous duck populations, the mitochondrial DNA (mtDNA) control region was sequenced and analyzed for 208 individual ducks, including 22 domestic breeds, wild Mallards ducks, Eastern spot-billed ducks, White Muscovy ducks, and Black Muscovy ducks. The haplotype diversity (Hd) was 0.653 and the average nucleotide diversity (Pi) was 0.005, indicating moderate genetic diversity. Sixty haplotypes were detected, and the maximum-likelihood (ML) phylogenetic tree and median-joining (MJ) network were generated from the sequence analyses. In this study, haplotypes from the Mallard duck (Anas platyrhynchos) were detected in most of the Chinese domestic duck breeds. In addition, the Eastern spot-billed duck (A. zonorhyncha) H8 haplotype was detected in two duck breeds. Only two haplotypes were found in Muscovy ducks, suggesting low genetic diversity within this population. The sequence and haplotype analyses revealed that both A. platyrhynchos and A. zonorhyncha contributed to the evolution of domestic ducks in China.

1. IntroductionWaterfowl resources are rich in China, and the Genetic Resources Committee has officially recognized these resources [1,2]. There are many possible origins of indigenous duck breeds in China [2,3,4]. The ducks may have originated from the domestication of the wild Mallard ducks (Anas platyrhynchos), the interbreeding of domesticated Mallard ducks and domesticated Eastern spot-billed ducks (A. zonorhyncha), or the domestication of wild hybrids of A. platyrhynchos and Anas zonorhyncha [3]. Previous studies of genetic diversity using amplified fragment length polymorphism (AFLP) of genomic DNA or analyses of genetic distance and relationships determined that domestic duck breeds evolved from both A. platyrhynchos and A. zonorhyncha [5,6,7]. Using microsatellite markers of genomic DNA polymorphisms, Chen et al. determined that the contribution to domestic duck breeds was greater for A. platyrhynchos than for A. zonorhyncha [8]. He et al. also used microsatellite DNA markers from 449 domestic ducks and wild Mallard ducks to demonstrate that all domestic duck breeds in China evolved from a single origin [9]. Similarly, Sultana et al. analyzed six Southeast Asian duck populations and determined that the ducks were domesticated from A. platyrhynchos [10]. Overall, these studies support the evolution of Chinese domestic ducks from both A. platyrhynchos and A. zonorhyncha. It was confirmed that the mitochondrial DNA (mtDNA) could more effectively evaluate the genetic diversity of poultry [11]. Unlike nuclear DNA, mtDNA acquires higher mutation rates that are 10 times than those of nuclear genomic DNA, and it is generally not recombined with the mtDNA of the parents [12]. Previous studies have confirmed that the control region has a higher mutation rate compared to the coding region [13]. In addition, it is estimated that mitochondrial control regions have a 2–5 fold higher evolution rate than mitochondrial protein-coding genes, making them more suitable for studying genetic variation among breeds [14]. In addition, a previous study using partial sequencing of mtDNA showed that nine domestic duck breeds along the Yangtze River evolved solely from A. platyrhynchos [15]. However, the major limitation of these genetic diversity studies was the limited number of domestic duck breeds and, more specifically, the lack of seven important duck breeds identified by the state, including the Fengtou duck (FT), Mawang duck (MW), Taiwan duck (TW), Zhongshan duck (ZS), Zongyang duck (ZY), White Muscovy duck (BF), and Black Muscovy duck (HF). This study performed the first comprehensive evaluation of domestic duck populations and wild Mallard ducks in China. We analyzed a 619–667 bp sequence of the control region of the mitochondrial DNA (mtDNA) of 22 domestic duck breeds, in addition to wild Mallard ducks, Eastern spot-billed ducks, and two Muscovy duck breeds. Sequence and haplotype analyses of the region revealed a high genetic diversity among the domestic duck breeds evaluated in this study. In addition, the identification of haplotypes shared across duck species suggested that Chinese domestic ducks originated from A. platyrhynchos and A. zonorhyncha.2. Materials and Methods2.1. Ethics ApprovalAll animal experiments were performed in accordance with the Regulations for the Administration of Experimental Animals issued by the Ministry of Science and Technology (Beijing, China). All experiments were approved by the Animal Care and Use Committee of Yangzhou University (approval code: 151-2014.).2.2. Collection of Blood Samples and DNA ExtractionAccording to the pedigree records, 208 blood samples were collected from unrelated duck individuals (Table 1, Figure 1). In addition, two Muscovy duck populations, the White Muscovy duck (n = 8) and the Black Muscovy duck (n = 8), were collected from Putian in Fujian Province and used as the reference group for the analysis. These blood samples (approximately 3 mL) were collected from the wing vein of the ducks in individual vacuum blood collection tubes. The genomic DNA was extracted using a genomic DNA extraction kit (TIANGEN, Beijing, China). RNA was purified to remove genomic DNA (gDNA), miRNA, and rRNA using the RNeasy Micro kit (Qiagen, Hilden, Germany), RNase-Free DNase kit (Qiagen), and RiboZero™ MagneticKit (Epicentre, Illumina, San Diego, CA, USA), respectively. Data collection and analysis were conducted at the College of Animal Science and Technology, Yangzhou University, Yangzhou, Jiangsu Province, China.2.3. Polymerase Chain Reaction (PCR) and DNA SequencingThe 667-bp (Mallard duck) or 619-bp (Muscovy duck) segment of the mtDNA control region was PCR-amplified using the primers 5′-CCTATGGTCCCGGTAATAAACA-3′ (F-Primer) and 5′-GATAACGCAGGTGTGTCCAG-3′ (R-Primer), which were designed according to the Beijing duck sequence (accession number NC_009684) in GenBank (Figure 2). The PCR assay was conducted in a 20 μL reaction mixture containing 2 μL of buffer (10×), 2.5 μL of MgCl2 (25 mM), 1.0 μL of dNTPs (10 mM), 1 μL of each primer (10 μM), 0.2 μL of Taq DNA polymerase (5 U/μL) (Takara Biomedical Technology (Beijing) Co., Ltd.), 1.0 μL of DNA template (100 ng), and of 11.3 μL of sterilized water. The reaction was brought to a final volume of 20 μL by adding double-distilled autoclaved water. PCR amplification was carried out with an Eppendorf Mastercycler (Eppendorf, Hamburg, Germany). The thermocycler was set as follows: 95 °C (5 min), 94 °C (45 s), 58 °C (45 s), 72 °C (45 s) for 35 cycles, and a final elongation step at 72 °C for 5 min. The last elongation step was extended to 8 min at 72 °C, and samples were held at 4 °C. The F-Primer and R-Primer were provided by Sangon Biotech Shanghai Co., Ltd. (Shanghai, China), for the direct bidirectional sequencing of target fragments within the PCR products using the ABI 3730 XL sequencing platform.2.4. Data AnalysisSequences of the mtDNA fragments were visualized by Chromas2 software and manually browsed to ensure the accuracy of the bases. Sequences were aligned by DNAStar-MegAlign V.7.10. The genetic diversity parameters of the populations, such as variable nucleotide sites, haplotype diversity, and nucleotide diversity, were calculated by DnaSP V.5.0 software [16,17]. The haplotype ratio was calculated as the ratio of the number of haplotypes to the sample size. The distances (Kimura 2-parameter) among all populations were calculated using Mega v.5.0 [18], and then the maximum-likelihood(ML) evolutionary tree was mapped out. Finally, the median-joining (MJ) network of the control region of the mtDNA haplotypes was drawn using Popart v.1.7 software [19].3. Results3.1. Nucleotide Composition of the mtDNA Region of 26 Duck PopulationsThe data of 208 ducks were analyzed, and the average composition of the sequences was 25.01% A, 15.71% G, 34.23% C, and 25.05% T (50.06% A + T, 49.94% G + C), showing little base bias. A total of 99 polymorphic sites (10 singleton variable sites and 89 parsimony-informative sites) were detected. As all insertions and deletions were removed from the analysis, the variable types were classified as transitions and transversions.3.2. Genetic Diversity of 26 Duck Populations in ChinaA total of 60 haplotypes were identified among 26 duck breeds (Table S1), including 2 haplotypes of Muscovy ducks (Table 2) and 58 haplotypes (Figure 3) of domestic ducks, wild ducks, and spot-billed ducks. The largest number of haplotypes (six) were detected in the Chaohu duck (CH), Ji’an Red Ma duck (JA), Jingxi Dama duck (JX), and Zongyang duck (ZY), with a haplotype proportion of 75.0%, while the lowest haplotype proportion (12.5%) was detected in the Sichuan Ma duck (SC) population (Table 3).The average haplotype diversity (Hd), nucleotide difference (K), and nucleotide diversity (Pi) in these 26 duck populations were 0.653, 3.104, and 0.005, respectively; the Zhongshan duck (ZS) had the highest (Hd = 0.929, K = 6.893, Pi = 0.0111), while Sichuan Ma duck (SC) and White Muscovy duck (BF) had the lowest (Hd = 0.000, K = 0.000, Pi = 0.000). 3.3. Phylogenetic Analysis of Chinese Duck PopulationsThe major H5 haplotype was shared by 61 ducks (30.77% of all ducks) from 18 domestic breeds and 3 wild Mallard ducks. The H8 haplotype was shared by one Mallard duck, one Zongyang duck, and one spot-billed duck (Table 3). The ML phylogenetic tree (Figure 4) of the 26 Chinese duck populations and the median-joining network chart (Figure 5) was constructed using all the mallards (A. platyrhynchos) and spot-billed ducks (A. zonorhyncha) in this study.4. DiscussionStudies on the origin and evolution of livestock breeds through the different mtDNA sequences also contribute to the understanding of the domestication process of species [20]. For the identification of the genetic diversity and phylogeny of different breeds and populations, mtDNA control region sequence analysis has proven to be the most valuable tool [21]. Some studies support the finding that natural selection affects the genetic diversity of populations, which can be measured by haplotype diversity (Hd) and nucleotide diversity (Pi) [22,23,24]. It has been shown that evaluating the impact of natural selection on genetic diversity through haplotype diversity (Hd) and nucleotide diversity (Pi) can help us better understand natural selection and the trends in population genetic diversity trends. In this study, the Hd and Pi of 26 Chinese domestic duck populations were determined to be 0.653 and 0.005, respectively. These results indicate that the domestic duck populations have a relatively large number of haplotypes and exhibit fewer nucleotide sequence differences. In addition, these values were slightly higher than the detected diversity values of four domestic duck breeds in Fujian Province (Hd = 0.645, Pi = 0.001) [25], but they were much lower than those of A. platyrhynchos (Hd = 0.987 and Pi = 0.008 in Kulikova’s study) [26]. In comparison with other animals, the Pi of Chinese ducks was higher than that of chickens (<0.001) [27,28,29] and swine (0.001) [30], but it was significantly lower than that of cattle (0.021) [31]. Haplotype and nucleotide diversity were highest in ZS ducks and lowest in SC ducks and BF Muscovy ducks.Studies on the genetic diversity and genetic structure of species are important components of biodiversity conservation [32]. Previous studies have also found that the SC and BF duck breeds may have experienced a long duration of high selective pressure for particular traits, resulting in low genetic diversity [33,34,35]. By contrast, the ZS population was more recently discovered and was bred in closed regions (Guangdong Province) without any artificial selection. Interestingly, the higher selection pressure might have led to some genotypes being better suited to specific environments [34]. Therefore, this selection increased the frequency of certain genotypes in the population, thereby reducing genetic diversity and even leading to the extinction of some genotypes. In addition, the between-population analyses showed a lower average nucleotide difference (K) than the within-population analyses in this study. Previous studies have found low genetic diversity across 24 domestic Chinese duck breeds and concluded that efficient measures should be adopted to protect domestic duck resources, especially for the JX and LW breeds [15,35]. The results presented here suggest the SC and BF breeds require more immediate attention because of their low genetic diversity.However, there is still no consensus in the field regarding the origin of domestic ducks in China [2,36]. In this study, the 192 domestic Chinese ducks presented 58 haplotypes, while the 16 Muscovy duck individuals conformed to two haplotypes. The H5 haplotype was shared by the majority of domestic and Mallard ducks, suggesting that most domestic duck breeds in China originated from A. platyrhynchos. The Eastern spot-billed ducks clustered into the H7–H10 haplotypes. Interestingly, the H8 haplotype was shared by one Mallard duck, one ZY duck, and one spot-billed duck, indicating that the partial domestic duck breeds in China might have originated from A. zonorhyncha. However, there were no other shared haplotypes between the domestic and spot-billed duck populations. Finally, our results suggested that domestic Chinese ducks evolved from A. platyrhynchos and A. zonorhyncha.5. ConclusionsSixty haplotypes were detected in most of the domestic Chinese duck breeds analyzed in this study. The sequence and haplotype analyses revealed that both A. platyrhynchos and A. zonorhyncha contributed to the evolution of domestic ducks in China. The Sichuan Ma duck (SC) and White Muscovy duck (BF) had the lowest Hd, K, and Pi. To summarize, their conservation requires more attention and prioritization of conservation funds in the future.

animals : an open access journal from mdpi

10.3390/ani11113249

PMC8614400

The selection for the rapid growth rate in broiler chickens that has been carried out over the years has negatively influenced their health and welfare status. In recent years, a number of reports have been delivered on the use of additives that improve broilers’ intestinal peristalsis and production results. The authors of this paper have proved that applying a mixture with 50% hops (manifesting strong antioxidant, antibacterial, and antifungal properties) may bring benefits to the quantity and quality of the final product. This may refer to the production performance, flock health status, and welfare of birds. The thematic scope of this research is currently of significant importance, as veterinary inspections pay particular attention to the quality of litter and the welfare of birds, and this motivates producers to improve breeding conditions, which will contribute to better production systems.

The objective of this study was to determine the influence of phytogenic product-supplemented, organic acid-supplemented, and prebiotic-supplemented diets on the production results, antioxidative status, and selected welfare indices in broiler chickens. A total of 1155 one-day old male Ross 308 broilers were randomly assigned to one of three treatment groups: Group C, no additives; Group A, supplemented with phytogenic supplement (50% hop); and Group P, supplemented with 65% organic acids and their salts, and 30% prebiotic complex. Health condition and production results were monitored during the entire experiment. After 42 days, 10 birds from each dietary treatment group were selected for blood sampling and slaughter analysis. The results obtained revealed that over the whole feeding period, none of the investigated additives significantly affected broiler performance indices. However, feeding the birds treatment-A increased the relative abundance of Bifidobacterium in caecal digesta compared to the other treatments, whereas feeding treatment-P increased the relative abundance of Lactobacillus compared to the control treatment. Overall, treatment-A was more effective at increasing relative abundance of Clostridia in birds at 42 days of age than treatment-P. Finally, there were no changes in blood levels of antioxidant indices or liver function indicators.

1. IntroductionLegal regulations state that since 2006, the preventive use of antibiotics and antibiotic growth promoters (AGP) in farm animals is banned in the European Union; therefore, studies on the use of plant preparations as drug alternatives in the diets of chickens have been carried out [1,2,3]. This is because, in the intensive broiler industry, the prevalence of compromised bird health, particularly that associated with the gut’s functional status, has increased due to the withdrawal of AGP. The functional status of the gut, which determines to a great extent not only the birds’ health but also the conditions of the chicken coop, depends on several factors, including the integrity of gut epithelium, dietary composition, and the commensal microbiota [4]. Overall, the barrier integrity in birds can be compromised by either a non-infectious method, such as feeding programs, dietary imbalances, dietary non-starch polysaccharides and protein levels, or feed contamination (e.g., mycotoxins), or by infectious factors including pathogen challenge [5,6,7]. These complex interactions between the host and the factors that affect birds’ health have created a great challenge for the poultry sector in developing reliable, cost-effective alternatives that are fully integrated with the production system in order to improve the birds’ health while maintaining good production results [8]. Moreover, consumer concern about the ways food animals are raised has created a move by producers and the industry towards transparency and has also led to demands for the use of antibiotic and AGP alternatives in the production setting [6,7,9].The market for feed additives has developed greatly in recent years and has adapted its offer to the needs of manufacturers. Feed additives, which are used in poultry diets, consist of probiotics, prebiotics, acidifiers, medium-chain triglycerides, and herbal extracts. Preparations based on herbs are increasingly popular due to consumer preferences and the search for natural products. In recent years, the term “phytogenics” has appeared in the market—meaning, preparations of plant origin obtained from herbs containing bioactive substances of secondary metabolism [10,11]. Secondary metabolites are the products of specialised metabolisms, which are indispensable for the survival of plants, and include the following compounds: glycosides, alkaloids, flavonoids, terpenes, essential oils, pectin, and plant mucoid fluids as well as organic acids, vitamins, and mineral salts. Phytogenics, due to their high content of biologically active components, appear to be very promising in terms of a prophylaxis for poultry. However, one of the factors that is of key importance is production profitability, which warrants that the application of the treatment must fulfil economic criterion to be successful, particularly in commercial conditions. In this regard, the recommendation is that phytogenics must be administered throughout the whole production period to be efficient for prophylaxis. According to calculations, the cost of the phytogenic prophylaxis is typically equivalent to the administration of antibiotics to a flock for a period of five days [12].The welfare indices, in particular the leg condition and walking ability of broilers reared in intensive conditions, is strongly influenced by the coops’ condition [13]. This association can be explained due to the direct influence of feed on the gut’s microbial composition and activity, and the utilisation of the feed’s ingredients by the birds, which determines, in consequence, the conditions of the chicken coop (by the quality of the litter) [14]. Birds reared under unfavourable conditions are exposed to an increased prevalence of walking difficulties, lesions, and/or depressed growth [15]. Currently, the economic loses in broiler production caused by leg weakness and footpad dermatitis (FPD) have been identified as one of the major challenges in the poultry production sector [16,17]. FPD has been defined as an important problem in the modern poultry industry and negatively affects the birds’ welfare and health status through compromised walking and feeding activity as well as compromised growth performance. For example, in the Netherlands, almost 38% of broilers have been reported as having severe footpad lesions, while 26% had mild footpad lesions, indicating, among other things, a problem of magnitude [18].The present study aimed to investigate the potential of using phytogenic ingredients from hops, liquorice, and gum arabic, or a mixture of a free butyrate acidifier and gluconic acid and a prebiotic complex (the prebiotic component in PreAcid is gluconic acid, 20% in the form of calcium gluconate) to improve broilers’ welfare. The assessment was based on the broilers’ walking ability, the skin condition of foot pads, and selected indices of antioxidant status and performance results in semi-commercial conditions.2. Materials and Methods2.1. Birds, Diets, and Experimental DesignThe trial was conducted on a total of 1155 one-day-old male Ross 308 broiler chickens. The birds were randomly assigned to one of three treatment groups (five repeated pens for each group, with 77 chicks per pen; n = 5): Group C, without any supplements; Group A, supplemented with a phytogenic product based on more than 50% hop ingredients in combination with a premix of liquorice, gum arabic, and other plant ingredients (13%); and Group P, supplemented with 65% organic acids and their salts (formic and lactic) and 30% prebiotic complex (the prebiotic component was gluconic acid, 20% in the form of calcium gluconate) [19].The diets were commercial, and each feeding was formulated to meet or exceed the requirements for Ross 308 broiler chickens. The amounts of supplement included in Group A were as follow: 400 mg/kg in the starter, 300 mg/kg in grower I, and 200 mg/kg in grower II and the finisher. The supplement amounts included in group P were: 3 g/kg in the starter, 2 g/kg in grower I, and 1 g/kg in grower II and the finisher. Overall, each treatment was repeated for five pens of 77 chickens each. Lighting and temperature in the building were maintained according to the Ross Management Guide [20]. The one-day-old chickens were vaccinated at the hatchery against infectious bronchitis, Gumboro disease, Marek’s disease, and Newcastle disease. All broilers were individually weighed (BW) throughout the experiment period on days 1, 10, 24, 35, and 42. Likewise, when diet changes took place, the total feed consumption for the respective feeding phases for the periods 1–10, 11–24, 25–35, and 36–42 days was determined for each group on a per pen basis. Based on these data, the feed conversion ratio (FCR) was calculated for the respective experimental periods (total weight of feed consumed/obtained gain). Mortality was recorded during the whole trial period. The European Yield Coefficient (EYC) was calculated for the whole period of rearing (42 days). EYC was calculated based on the following formula:EYC=BW×livability# (%)day of production×FCR (kgkg)#livability is characterized by the ratio of the final number of birds to the initial number. During the experimental period, the broilers were fed according to the following feeding program: starter, days 1–10; grower I, days 11–24; grower II, days 25–35; and finisher, days 36–42. The birds had ad libitum access to feed and water consumption. The composition of the diets was determined using the standard AOAC method [21] (Table 1).2.2. Litter AnalysisLitter samples were collected once every week from five different sites (two samples of the horns and around water lines, around feeder lines, and around the free area) within each coop and pooled and mixed before being measured (50 g). The litter moisture content was determined by the weight loss on drying. In brief, the pooled litter was weighed, then dried for 12 h at 105 °C in an electric drying oven (Model: WAMED, SUP–100 W5, Warsaw, Poland), and then weighed again. The decreased weight was recorded as dry matter (%).2.3. Gait-Scoring ObservationsAssessment of walking ability and foot pad condition. The birds’ ability to walk was scored on a six-point scale at the thirty-ninth day of life and was performed by qualified veterinarian staff. The following detailed descriptions of each gait score were applied, but not all the attributes of a score were necessarily identified in each bird [22]. One hundred and fifty birds were examined from each group, selecting randomly 30 from each pen, according to following scale.Gait score 0. The bird walked normally with no detectable abnormality; it was dexterous and agile. Typically, the foot was picked up and put down smoothly and each foot was brought under the bird’s centre of gravity as it walked (rather than the bird swaying). Often, the toes were partially furled while the foot was in the air. The bird should have been capable of balancing on one leg and walking backwards easily if necessary. It should also have been in full command of where it was going and been able to easily deviate from its course to avoid other birds.Gait score 1. The bird had a slight defect that was difficult to define precisely but would have precluded its use for breeding if gait had been the sole selection criterion at the standard of a pedigree breeder. For example, the bird may have taken unduly large strides which, although the observer may not have recognised the exact cause, produced an uneven gait.Gait score 2. The bird had a definite and identifiable defect in its gait, but the lesion did not hinder it from moving or competing for resources. For example, it may have been sufficiently lame in one leg to produce a rolling gait that did not seriously compromise its manoeuvrability, acceleration, or speed.Gait score 3. The bird had an obvious gait defect that affected its ability to move about. For example, the defect could take the form of a limp, a jerky or unsteady strut, or a severe splaying of one leg as it moved. The bird often preferred to squat when not coerced to move, and its manoeuvrability, acceleration, and speed were affected.Gait score 4. The bird had a severe gait defect. It was still capable of walking, but only with difficulty and when driven or strongly motivated. Otherwise, it squatted down at the first available opportunity. Its acceleration, manoeuvrability, and speed were all severely affected.Gait score 5. The bird was incapable of sustained walking on its feet. Although it may have been able to stand, locomotion could only be achieved with the assistance of the wings or by crawling on its shanks.The gait scoring was performed by two people, one gently herding and driving each bird with a light cane, and the other observing from a crouched position. The two assessors had different views of the bird, and agreement between them for each bird was required before the score was recorded. In some cases, it was necessary to assess a bird over several passes, each typically of about 2 m.2.4. Foot Pad Dermatitis ObservationFoot pad condition was visually defined five days before slaughter according to a five-step scale [23]. The veterinarians’ evaluation consisted of determining the degree of the lesions’ advancement on a scale of 0, 1, or 2 (Table 2). All birds involved in the experiment were subject to evaluation.2.5. Sampling ProceduresThirty chickens were chosen (10 birds from each treatment) for slaughter at the age of 42 days of life that had a body weight similar to the group mean. Blood samples were collected from the jugular vein for biochemical and antioxidant analysis.We aimed to evaluate the relative abundance of selected bacterial groups in caeca. On day 42 of age, a total of 10 chickens from each dietary treatment group were selected, representing the mean group by BW, and the luminal contents of both caeca from each bird were collected in sterile tubes and immediately frozen at −32 °C for further analysis. To determine the relative abundances of bacteria in the caecal samples, the following previously described protocol was applied [24]: briefly, the caecal digesta was thawed at 4 °C for four hours, and then, bacterial genomic DNA was extracted from digesta (approximately 200 mg) using the QIAamp Fast DNA Stool Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. The yield and purity of the isolated DNA were then estimated spectrophotometrically (Nanodrop, NanoDrop Technologies, Wilmington, DE, USA). The primer sets that were used for determining the respective bacterial populations are presented in the Table 3. The PCR conditions were applied as reported above for each respective bacteria group. The obtained PCR-products were separated by electrophoresis on a 2% agarose gel. PCR products were quantified using ImageJ 1.47v software for densitometry measurements, National Institute of Mental Health, Bethesda, MD, USA (NIMH), with the density of the bands for each of bacteria group expressed in relation to the density of the total bacteria primer product. Each sample was analysed in duplicate.Chemical analysis. Blood biochemical parameters such as enzymes (aspartate transaminase—AST, EC 2.6.1.1; alanine transaminase—ALT, EC 2.6.1.2; gamma-glutamyltransferase—GGT, EC 2.3.2.2) and electrolytes (sodium—Na, potassium—K, chloride—Cl, and calcium—Ca), glucose—GLU, magnesium—Mg, phosphorus—P, lactate (LAC), and total protein (TP) were determined using a Cobas INTEGRA 400 Plus biochemical analyser (Roche Diagnostics, Basel, Switzerland) based on spectrophotometry, turbidimetry, fluorescence polarisation, and ion-selective potentiometry methods. Samples of blood were collected in sterile tubes without an anticoagulant. To obtain serum, whole blood was centrifuged at 2000× g for 10 min at 4 °C and then placed in analyser racks.The level of vitamin C in the serum was determined using a LambdaBio-20 spectrophotometer (Perkin Elmer, Waltham, MA, USA). The following components were mixed thoroughly and centrifuged for 20 min: 0.5 mL of tissue homogenate, 0.5 mL of distilled water, and 1.0 mL of 10% trichloric acid. The obtained supernatant at 1.0 mL was combined with 0.2 mL of 2,4-dinitrophenylhydrazine-thiourea-copper sulphate reagent and incubated at 37 °C for 2 h. Subsequently, 1.5 mL of 65% sulfuric acid was added, and the mixed sample remained at room temperature for another 30 min. The change in colour of the sample was measured at 520 nm. Solutions of vitamin C standards (0.5–5 mg of vitamin C A92902 Sigma-Aldrich L-Ascorbic acid 99%) were treated similarly.Measurements for radical scavenging activity were performed by routine assay procedure [29] using a synthetic DPPH radical (1,1-diphenyl-2-picrylhydrazyl). Folin–Ciocâlteu reagent was used as an oxidizing reagent, and all the chemicals were purchased from Sigma-Aldrich Chemie GmbH (Munich, Germany) in the highest available purity.Glutathione (GSH) concentration was determined in the whole blood by means of the OxisResearch™ Bioxytech® GSH/GSSG—412™ test (Foster City, CA, USA). Before the analysis, the samples were frozen with the addition of M2VP (1-methyl-2-vinyl-pyridium trifluoromethanesulfonate) at a temperature of −80 °C. The released, reduced GSH was determined in accordance with the detailed instruction provided by the kit’s producer. The absorbance reading (λ412) and the measurement of reaction kinetics were performed using the microplate reader Synergy 4 (BioTek; Winooski, VT, USA). The results were calculated using Gen5 software (BioTek). GSH concentration was expressed in thiol groups (mmol-SH groups).The determination of the content of total phenols was performed as previously described [30].The samples were thoroughly mixed, and after 8 min, 2 mL of saturated sodium carbonate solution was added. The next stage of the analysis involved an incubation test at 40 °C for 30 min (until a stable characteristic blue colour had developed). The absorbance was measured at 765 and 735 nm against a blank sample (experimental material replaced with 0.5 mL ddH2O). Results were read using a calibration curve plot based on the absorbance of a gallic acid standard in the range of 0–0.5 mg/mL and expressed in mg gallic acid equivalent (GAE/mL) serum.2.6. Statistical AnalysisThe statistical analysis was performed using SPSS 23.0 for Windows. Analysis of variance was used to determine the influence of experimental factors on production results, antioxidative status, and the welfare of the broiler chickens. The normality of the variables’ distribution was checked using the Kolmogorov–Smirnov test. In turn, the Kruskal–Wallis test was used to determine the effect of treatment on body weight at each term of analysis, while the Mann–Whitney test was used to determine the differences between groups. Finally, the Chi-square test was employed to estimate the frequency of gait scoring and FPD in particular groups.2.7. Ethical StatementAll procedures in the present study were performed in accordance with the principles of the European Union and Polish Law on Animal Protection. This study was conducted by qualified veterinarians who performed all procedures that involved the handling of the birds. No action involving pain or suffering was practiced. This study was run in accordance with Directive no. 2010/63/EU and did not require the approval of the Local Ethics Committee based on the regulation of the Ethic Committee of November 2019 (resolution no. 174/2019).3. Results3.1. Performance ResponseOverall, there were no significant differences in the production results for the chickens that were given the treatments over the entire 42-day feeding period. Numerically, the best final body weight was achieved by chickens from group A (3.46 kg) and the lowest from group P (3.43 kg). Numerically, the best FCR result was also seen in group P, 1.49 kg/kg. The European Yield Coefficient (EYC)—which considers growth performance expressed by BW, feed efficiency expressed as the feed conversion ratio (FCR), the production duration over the 42 days, and the mortality rate during this period—was used to assess the economic efficiency of the chickens. The number of points that produced an effect in the treatment-A group was 532 points, group P—522 and group C—518.3.2. Walking Ability and Foot Pad ConditionThe data obtained regarding walking ability and leg conditions revealed that numerically, the best gait indicating walking ability was observed in the group of birds administered with treatment-P (the best walking ability was 13.3% vs. 10.7% and 5.3% for C and A treatments, respectively) (Figure 1). Where FPD occurred, the data indicated that there was no difference between either experimental treatment regarding a higher frequency of lesion occurrence (2%), but in the control group, it was numerically higher by 0.6 percentage points, yet the result has not been statistically proven (p = 0.275) (Figure 2).3.3. Determination of Litter Dry MatterThe assessment of dry matter content in the litter is presented in Figure 3. Overall, the dry matter content for each group decreased consistently over the 42-day period, with averages of 70% on day 1 and 32.3% on day 42.3.4. Bacterial Composition in the Caecal DigestaThe dietary treatments significantly affected the relative abundance of selected bacteria in the caecal digesta of the birds at 42 days of age (Figure 4). Feeding the birds treatment-A resulted in a significantly higher relative abundance of Bifidobacterium than in the other treatments (p = 0.023), whereas feeding the birds treatment-P increased the relative abundance of Lactobacillus compared to the control treatment (p = 0.045). Overall, it was evidenced that treatment-A was more effective at increasing Clostridia relative abundance in birds at 42 days of age than treatment-P (p = 0.028).3.5. Blood Indices of MetabolismThe metabolic profile parameters were determined for the blood of the birds at 42 days of life. It was shown that the activity of the determined liver enzymes, for example, AST and ALT, were not significantly different for the two treatments. The GGT activity was significantly (p < 0.05) higher in the P and A groups than in the control. We also evidenced significant changes in LAC activity as well as K and P concentrations in the blood. In the control group, all these parameters were significantly higher (p < 0.05) than in the remaining groups. Other biochemical parameters analysed, such as glucose, total protein, and Ca, Mg, Na, and Cl, did not differ between the treatments (Table 4).The analysed antioxidant indices in the blood showed that there were no differences in indices such as GSH (1.073, 1.175, and 1.14 mmol–SH; C, A, and P groups, respectively), percentage of DPPH (82.57%, 79.69%, and 83.34%), total polyphenols (2.30, 2.29, and 2.27 mg GAE/mL), and vitamin C (9.01, 8.75, and 8.43 mg/100 mL) in the birds as a result of the dietary treatments. Altogether, the blood-indices’ response showed that the use of the investigated feed additives did not disturb the oxidative status of the birds as assessed using blood markers.4. DiscussionThe incidence of leg problems such as FPD are currently considered leading welfare issues in poultry farming. FPD is characterised by inflammation and necrotic lesions (which can be superficial through to deep) on the plantar surface of footpads and toes [31]. The origin of FPD is quite complex as it is believed to result from several factors, but is particularly related to litter moisture, an imbalance in nutrition, and genetic susceptibility [16]. However, there is speculation that the primary contributor to a susceptibility to FPD is related to the quality of bedding material, which is closely related to its moisture content and microbial characteristics, particularly in the last period of birds’ growth [32]. Therefore, in the current study, both walking ability and FPD assessments were done in the last stage of the rearing period, when the birds are the most susceptible to locomotive disorders due to their high BW and poor litter conditions [17]. In this regard, there were no statistically significant differences in the walking ability and FPD parameters between the control and treatment groups. Despite this, however, a possible cross-link between the dietary treatments and litter quality could be implied due to the changes observed in bacterial composition of caecal digesta, which to a great extent contributed to the overall litter conditions. In line with this, the most effective treatment in the present study that caused a shift in the amounts of caecal bacteria was the treatment supplemented with phytogenics because in the birds fed with this treatment, the relative abundance of Clostridia was significantly higher than in the group fed the organic acid supplemented diet; the same was true regarding the relative abundance of Bifidobacterium. On the other hand, feeding the birds a diet supplemented with organic acids resulted in a significant shift in the relative abundance of Lactobacillus in the caecal digesta compared to control. The Lactobacillus species reside in the caeca and are present in great abundance, as well as playing a key role in the process of fermentation of the substrates that reach this environment. These bacteria are also responsible for the production of lactic acid, which manifests regulatory properties on gut function [33]. According to another report [34,35], some pathogens, including Clostridium perfringens, belonging to the group Clostridia, are known to be responsible for the skin inflammation that causes necrotic lesions in birds; however, due to the fact that in the present study, the investigated supplements did not compromise foot pad condition, walking ability, or the performance response of the chickens, we excluded the probability that the proportion of these pathogens increased along with Clostridia shift as a result of the applied treatments. Because the Clostridia group are involved in the metabolism of short-chain fatty acids, which are a major end-product of bacterial fermentation, they therefore determine the properties of the digesta and excreta [36]. This was also confirmed in our preliminary study [19], in which higher butyric acid concentration in the caecal digesta was evidenced as a result of organic acid supplementation. Given this, we may speculate that the changes in the abundance of key groups of bacteria in the caecal digesta, which are beneficial to the host, were responsible for the improved litter quality and skin condition of the birds’ legs. This is in line with our preliminary study, in which evidence was found that the supplementation of broiler diets with either phytogenic compounds or organic acids significantly affected caecal bacterial activity (assessment based on short-chain fatty acid production), indicating a possible association [19]. The beneficial role of the Bifidobacterium group in modulating digesta properties in farm animals has also been well reported [37]. The investigated additives’ possible mode of action in the present study could have been manifested through indirect action via short-chain fatty acids and lactic acid, which can penetrate the microbial membrane and dissociate into protons and anions, thus acidifying cytoplasm and making sensitive bacteria dysfunctional and unable to maintain optimal pH, which, in turn, results in changes in their abundance in the caecal environment [38]. In this study, we investigated the effect of phytogenic additives and organic acids on the microbiota composition within the caecal digesta instead of litter/faeces, as the indices for litter/faeces varies greatly depending on the coops’ conditions. According to Williams and Athrey [39], even cloacal swabs have been shown to be an unreliable source for reflecting the gut microbiota community and structure in birds due to inter-individual variation and high degrees of randomness. We link the possible action of both the investigated additives with their likely influence on microbiota composition in the caeca as was reported in our preliminary study regarding microbiota activity [19]. However, more insights in this regard are needed.The determination of the blood indicators of metabolism by laboratory analysis is an important procedure that assists in the diagnosis of various avian diseases and disorders. The response of the animal’s body is investigated on the basis of the enzymatic activity responsible for the hepatic profile of gamma-glutamyl transferase (GGT), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) as well as the electrolytes C, K+, Na+, and phosphor [38]. For the sake of animal health and welfare, increasing attention has now been given to oxidative stress parameters [13,40]. It is already known that the occurrence of dietary-induced oxidative stress is correlated with the induction of inflammation markers. This may indicate impaired immunity, which may affect skin quality, increase FPD risk, and consequently lead to a deterioration in the welfare and health status of the birds. Because blood indices reflect the physiological status of the birds, they can be used as indicators of the birds’ health, nutrient metabolism, and physiological status. Chickens that suffer from metabolic disorders can be diagnosed based on liver function blood markers, including ALT, AST, and GGT [31]. The lack of changes in AST and ALT levels for the treatments reported in this study may suggest that the supplements administered in the study (either P or A) did not cause metabolic changes that affected liver function when compared to the control group. In contrast, the observed changes in GGT activity were rather marginal for the birds’ physiological status since no significant differences in growth rate was observed. The increased GGT levels for the P and A treatments may also suggest intensified liver workload; however, no deterioration in the health status of the birds was observed, and the mortality levels remained within acceptable norms [41]. Liver load was not significant for the birds in the study groups because we observed a lower concentration of lactate in their blood. Blood lactate increases as a result of physical stress and severe tissue hypoxia, or respiratory failure. Rising levels can lead to metabolic acidosis. The lower value of lactate in the blood of these birds may also indicate a reduction in energy requirements related to physical activity. In the present study, lower lactate activity was confirmed in both experimental treatments and indicates that neither additive disturbed the oxygen supply. Oxygen delivery may have indicated reduced gait movement resulting from high body weight at 42 days of age in groups P and A. Broilers in commercial settings are exposed to a range of stressors. A growing body of information clearly indicates that excess, reactive oxygen and nitrogen species (ROS/RNS) production, and oxidative stress are major detrimental consequences of the most common commercial stressors in poultry production [41,42]. Our data indicate that the differences reported for the selected blood indices in the birds were likely not associated with any physiological disorders. This is also supported by the performance results reported in the present study, which showed no significant differences due to dietary treatments, indicating that the investigated additives rather did not affect feed taste or, in turn, feed intake [43]. Other reports also confirm that diets supplemented with either phytogenic ingredients or organic acids are well tolerated by the birds and do not compromise performance [44,45,46]. These findings might be considered beneficial since the predictable effect of dietary treatments is the key factor in determining potential dietary intervention for prophylaxis.5. ConclusionsThe application of phytogenic products and organic acids to the broilers’ diet had no compromising action on the condition of foot pad. Neither group of additives disturbed the blood indices of metabolism or selected indices for antioxidant status or compromised performance results. Both the investigated additives manifested effects that supported the proliferation, in the caeca, of bacteria that were beneficial to the host, which could contribute to the improvement of the litter quality during the last period of rearing.

animals : an open access journal from mdpi

10.3390/ani13101700

PMC10215121

Contagious yawning (CY) has garnered considerable interest in the fields of psychology and animal behavior, particularly as it relates to the potential connection this response has with empathy or emotional contagion. Recent reports have explored this association through the study of interspecific CY, whereby the detection of yawns from one species trigger yawning in a different species. While the evidence for interspecific CY is robust, links to empathy have been inconsistent. Here, we sought to explore this relationship more explicitly in humans by assessing how interspecific CY in response to images of yawns from common household pets relates to individual differences in empathic concern. The results provide further evidence for interspecific CY in humans, but self-reported empathic concern was a negative predictor of this response. We also found no sex difference in interspecific CY, though when comparing the sexes across the CY conditions, women reported a higher frequency of yawning in response to dog yawns, and men reported a higher frequency of yawning in response to cat yawns. Overall, these findings do not support a strong connection between interspecific CY and empathy or emotional contagion.

Interspecific contagious yawning (CY), whereby yawns from one species trigger yawning in different species, has now been reported across various taxa. This response to human yawning appears common among animals in captivity and has been interpreted as an empathic response towards human handlers/caregivers. A recent study found that humans also display interspecific CY, though this response was not modulated by proxies of empathic processing (i.e., phylogenetic relatedness or social closeness to the animals). Here, we explored this relationship more explicitly by assessing how interspecific CY to yawns from common household pets relates to self-reported empathic concern. Participants (N = 103) completed a survey measuring empathic concern and then reported on their yawning behavior following exposure to a control condition or yawning images either from domesticated cats or domesticated dogs. The results provide further evidence for interspecific CY in humans, but empathic concern was negatively predictive of this response. There was also no sex difference in interspecific CY, though when comparing the sexes across CY conditions, women reported a higher frequency of yawning in response to dog yawns, and men reported a higher frequency of yawning in response to cat yawns. Overall, these findings do not support a strong connection between interspecific CY and empathy or emotional contagion.

1. IntroductionContagious yawning (CY) is a distinctive pattern of behavioral coupling that occurs when the mere detection of yawns by others elicits the automatic and reflexive tendency for an observer to yawn. Yawning can be reliably elicited in this way in humans using images, videos, or audio clips of yawning [1,2,3,4,5], though there is considerable variability in this response [6,7]. While the first experimental demonstration of CY in a nonhuman animal was published less than two decades ago [8], evidence for CY is now present for a growing number of social vertebrates (for a review, see [9]). Various theories have been proposed pertaining to the biological significance of CY [10], with recent evidence indicating a role in facilitating motor synchrony and enhancing vigilance in groups [11,12].Independent of its ultimate function, one view that has gained considerable traction in the scientific community is that CY represents a primitive form of empathic processing, i.e., the tendency to yawn after detecting this action in another could reflect a form of emotional contagion. This idea, which could explain the distinctive individual differences in CY across studies, was initially proposed in a monograph by Lehmann [13] and further elaborated within the perception–action model of empathy [14,15]. Empirical support for a connection between empathy and CY was first published by Platek et al. [6], and this continues to be a prevailing view in the current literature [16]. However, systematic reviews of the studies examining the linkage between CY and empathy have shown that the totality of evidence is rather mixed and inconsistent [17]. Alternatively, CY could instead represent a simple feature of behavioral contagion or facial mimicry [18,19] that would show a statistical correlation to markers of empathy or emotional contagion without a causal relation. For example, one key issue in this debate is that at a proximate level CY requires the detection of yawns in others, and social attention and empathy are often difficult to disentangle [20,21]. Thus, while the use of CY as a potential indicator of empathy or emotional contagion is intriguing, a direct linkage remains unclear.Recently, interspecific CY, whereby yawns from one species trigger a member of a different species to yawn, has been examined in relation to empathy. The first evidence for interspecific CY came 15 years ago from a study showing that the overt display of human yawning was sufficient to induce a similar response among domesticated dogs [22]. This initial report garnered considerable attention and led to numerous follow-up studies designed to explore the interspecific CY/empathy connection. If this cross-species facial mimicry was enhanced by empathy, it had been predicted that dogs should respond more strongly to yawns from caregivers/owners where there was an existing social bond or attachment. This view was based largely on a 2011 study of chimpanzees, whereby an ingroup bias for CY was observed and interpreted as a sign of empathy [23]. The authors of this work suggested that since humans show ingroup biases for empathy for pain [24,25,26], similar biases in CY could be used as a measure of affective empathy. Overall, the evidence in support of this effect among dogs is mixed [27,28,29], and a recent study, including a meta-analysis of this literature, concluded that human-initiated CY in dogs was, in fact, not a signal of empathy [30].Studies on other captive species, however, have reported more consistent familiarity/ingroup biases when it comes to interspecific CY. In line with the view that CY may reflect a form of empathy or emotional contagion, familiarity has been shown to enhance both intra- and interspecific CY among nonhuman primates [23,31,32]. In one study on chimpanzees, subjects responded with CY to stimuli from familiar humans (i.e., researchers and husbandry staff) and ingroup conspecifics but not to outgroup chimps or unfamiliar members from other species [31]. A more recent and very similar study on captive red-capped mangabeys produced the same basic findings: subjects showed a higher CY response to familiar human caretakers and conspecifics compared to unfamiliar individuals from across three species [32]. In addition, one study on African elephants found that individuals responded with CY to yawns from familiar human handlers [33], though no comparisons were made with unfamiliar humans.To further examine this relationship, Gallup and Wozny [5] investigated whether humans also display interspecific CY. Using an online format, participants were shown a series of yawning stimuli from one of the following categories: fish, amphibians, reptiles, birds, non-primate mammals, great apes, and a compilation of common household pets (including an equal representation of domesticated cats and dogs). Given that prior studies have shown that people display greater empathy and attachment towards both more closely related species [34,35,36] and domesticated animals that are commonly kept as pets [37,38,39], it was predicted that interspecific CY would be enhanced by phylogenetic proximity and domestication/social closeness. The findings provided clear evidence for interspecific CY when compared to the control images, with 69% of participants reporting interspecific CY. Yet, while it was expected that yawns from great apes (chimpanzees, gorillas, and orangutans) and common pets (cats and dogs) would elicit the greatest response, this was not the case. Consequently, these findings did not support the purported connection between interspecific CY and empathy or emotional contagion in humans and instead suggest that the mechanisms governing CY are generalized. Importantly, Gallup and Wozny [5] included a task that ensured attention towards the yawning stimuli during testing. Given CY is a reflexive and automatic response, it seems that the detection of yawns, independent of the taxa represented, may be sufficient to generate this response. However, like previous works, overt measures of empathy or attitudes towards animals were not obtained in this experiment. In addition, subsequent exploratory analyses from this open-access dataset indicated that women may be more likely to yawn in response to yawns from a compilation of CY stimuli from domesticated cats and dogs. While sex differences in CY are rarely observed (for a review, see [40]), and an overall effect of gender was not found. In general, it has been argued that a female bias in CY would be consistent with links to empathy and/or emotional contagion [41]. Thus, future research could help clarify this picture by assessing whether this interaction is reproducible.The current study served as a direct follow-up to Gallup and Wozny [5], providing a more explicit examination of the relationship between interspecific CY and empathy in humans. We sought to examine whether validated measures of empathic concern and animal attitudes predicted interspecific CY to common household pets while controlling for attention. Empathic concern was selected because it has previously been shown to correlate with animal attitudes [42,43] and intraspecific CY in humans [44], while measures of cognitive empathy (i.e., reading emotions and perspective-taking) tend to be less correlated with animal attitudes and have more often failed to predict CY in humans [45,46,47]. Given that dogs have undergone a long period of domestication that involved distinct forms of social communication and cooperation with humans [48] and that people tend to perceive both human and dog facial expressions in a similar manner [49], we examined interspecific CY separately to domesticated cats and dogs. Moreover, based on exploratory analyses showing a potential female bias in interspecific CY to common household pets, we aimed to replicate this effect of participant gender. If empathy was related to interspecific CY, empathic concern should positively predict this behavior, and perhaps women would show a greater response [41]. If interspecific CY is generalized, due to it being more of an automatic and reflexive response, merely attending to the animal stimuli should elicit yawns independent of empathic concern or gender.2. Materials and Methods2.1. ParticipantsFollowing the methods of Gallup and Wozny [5], participant recruitment occurred online using Amazon Mechanical Turk (MTurk): https://www.mturk.com/ (accessed on 22 November 2022). This study was approved by the Institutional Review Board at SUNY Polytechnic Institute (IRB-2022-5), and informed consent was required prior to initiating the study. Several restrictions were applied to improve data quality. First, eligible MTurk workers were required to live in the United States (US) and have completed a minimum of 100 tasks, with a successful completion rate >95%. To address potential issues of reduced attention among MTurk respondents, we recruited a total of 180 participants (60 per condition) and included various attention checks [50,51]. To screen for inattentive respondents and bots [52], for example, an initial attention check question and a Completely Automated Public Turing test to tell computers and humans apart (CAPTCHA) were included within the initial demographic questions. Similar attention checks were also embedded within the empathic concern and animal attitudes survey measures (see below). Incorrect responses to these items excluded 24 participants. Moreover, given that visual detection of yawning is critical when attempting to elicit CY with videos or imagery [20,53], a dozen additional attention check questions were embedded as part of the task when reviewing the CY stimuli (see design and procedure below). Another 35 participants were excluded due to incorrect responses to these questions. An additional 17 participants were excluded because they did not complete the study, and one last participant was not included in the analysis because they chose not to identify their gender. This left a final vetted sample of 103, including 53 men and 50 women (age M ± SD: 34.7 ± 10.2).2.2. DesignThis study employed a between-subjects design. Like Gallup and Wozny [5], data collection occurred via Google Forms, and 24 pictures were used for each of the three conditions from online image searches (e.g., Google): domesticated cats, domesticated dogs, and control. For the two interspecific CY conditions (domesticated cats; domesticated dogs), twelve images were displayed of animals in mid-yawn (i.e., extended gaping of the jaw, head tilting, eye closure), while the other twelve images were of animals of the same or closely related breed/coloring clearly not yawning. Half of these images were previously used in Gallup and Wozny [5], including a compilation of six domesticated dog and cat images, each in one pet condition, while the other half were new. The same 24 images from the previous study were also used for the control stimulus, half consisting of open building windows and the other half depicting similar windows and frames that were shut. Maintaining the original aspect ratio, all images were then standardized to 7.62 cm in height. As in Gallup and Wozny [5], similar images of yawning and non-yawning animals or open and closed windows were paired side-by-side in randomized right/left positioning (note: all stimulus images are available upon request).2.3. ProcedureFollowing some initial demographic questions, participants completed the seven-item empathic concern scale of the Interpersonal Reactivity Index (IRI) [54] and the five-item version of the Animal Attitude Scale (AAS-5) [55]. Internal consistency was acceptable for the empathic concern measure (α = 0.72) but was too low for the AAS-5 to be included in subsequent analyses (α = 0.27).Next, participants were given the following instructions when viewing the CY stimuli for the condition they were assigned (dog or cat): “You are going to be presented with twelve pairs of images, one at a time, each depicting one animal that is yawning and one animal that is not. You need to review each image pairing and correctly identify the animal that is yawning by indicating whether it is the image on the left or the right. It is important to answer each question accurately” (p. 4, [5]). As we were not able to monitor or record the visual attention of participants during testing, this procedure ensured that each yawn was detected during trials. As we closely matched the animals for the yawning and non-yawning stimuli in each pairing, the participants could not rely on other stimulus characteristics to make their determination and thus were forced to review both images and correctly label the one with an animal yawning. For the control condition, instructional references were made to open and closed windows rather than yawning and non-yawning animals. Evaluation of the window images in the control condition served as a comparable task visually, but one that would not be expected to induce yawning. After viewing and responding to the entire stimulus compilation for the condition they were assigned, participants then self-reported on their yawning behavior (yes/no and how many times), which has been shown to be a valid measure of CY [3,56,57]. Lastly, based on the association between yawning and sleep and fatigue [58,59], participants indicated their sleep duration (in hours) the previous night and how tired they felt on a 10-point scale (1: not tired at all; 10: extremely tired).2.4. AnalysisThe final sample included 39 participants in the cat yawning condition, 27 participants in the dog yawning condition, and 37 participants in the control condition. A post hoc power analysis was performed using G*Power 3.1 [60] with a medium effect, revealing power of 0.784 and 0.811 to detect a main effect for the binary and frequency measures of yawning, respectively. A binary logistic generalized linear model (GLM) was run for yawn occurrence, while a Poisson loglinear GLM was run for yawn frequency. Each model included stimulus condition and participant gender as factors, and the output included stimulus × gender interactions. Participant age, prior sleep, current tiredness, and the self-reported trait measure of empathic concern were included as covariates. All analyses were conducted in jamovi [61]. 3. ResultsIn total, 57.6% of participants reported yawning in response to the interspecific CY stimuli, with an average of 2.88 ± 3.86 yawns/participant. In the control condition, 54.1% of participants reported yawning, with an average of 2.41 ± 3.16 yawns/participant. Table 1 includes the descriptive statistics for the non-yawning variables.For yawn occurrence (yes/no), prior sleep duration and current tiredness were both significant predictors (Table 2). Overall, participants who reported sleeping for fewer hours the previous night and being more tired at the time of testing were more likely to report yawning while reviewing the stimuli (ps < 0.01). There was only a marginal effect of stimulus condition (p = 0.059), with planned comparisons showing that participants in the cat yawning condition were significantly more likely to yawn compared to the control condition (p = 0.046; Figure 1). No other comparisons were significant (ps > 0.05). Empathic concern was not a significant predictor (p = 0.122), though the trend was in the opposite direction from what would be expected, i.e., individuals that scored higher on self-reported empathic concern were less likely to yawn (Figure 2). When assessing just the CY conditions, we found that participants that yawned scored significantly lower in empathic concern (Welch’s t66 = 2.23, p = 0.031). No effect was observed for the control condition (Welch’s t37 = 1.64, p = 0.118). Lastly, there was no main effect of participant gender (p = 0.851), nor was there a condition × gender interaction (p = 0.538). For yawn frequency, stimulus condition, prior hours of sleep, and current tiredness were all significant predictors (Table 3). Like the binomial model, participants who indicated sleeping for shorter durations the previous night and feeling more tired at the time of testing reported more yawns (ps < 0.01). In this case, however, participants in both yawning conditions reported more yawns compared to the control condition (main effect: p = 0.023; Figure 3), though this effect was marginal for the dog condition (cat: p = 0.009; dog: p = 0.054). No difference in the frequency of interspecific CY emerged between the cat and dog conditions (p = 0.645). Like the binomial model, there was a trend for empathic concern whereby individuals that scored higher on this measure reported fewer yawns (p = 0.098; Figure 4). When assessing just the CY conditions, we found a significant negative correlation between empathic concern and self-reported yawn frequency (B66 = −0.278, p = 0.003). No correlation was observed for the control condition (B37 = −0.193, p = 0.135). Lastly, there was no main effect of participant gender, but a significant stimulus condition × gender interaction emerged (p = 0.005). When comparing the sexes, women reported greater interspecific CY to images of dog yawns, and men reported greater interspecific CY to images of cat yawns (Figure 5). 4. DiscussionThe current findings provide further evidence of interspecific CY in humans, replicating previous research [5]. While the magnitude of this effect was relatively limited for yawning images of domesticated dogs, showing up only when examining CY frequency, there was a larger and significant effect observed within the cat yawning condition across both the binary and frequency models. This difference in interspecific CY when viewing images of cat and dog yawns was not expected given the coevolution of human–dog bonding and social communication [48,49,61]. Nonetheless, these results, coupled with the report of Gallup and Wozny [5], provide strong support that, in general, yawns from common household pets can reliably elicit CY among humans. A primary objective of the current study, outside of replicating interspecific CY in humans, was to more explicitly assess how this response varied as a function of individual differences in a self-reported measure of affective empathy. To date, prior studies in this area have only used proxies for empathy, focusing on features pertaining to the familiarity (e.g., human handlers/caregivers) or phylogenetic closeness of the interspecific stimuli. Here, we had participants complete the empathic concern subscale of the IRI [54] as well as the five-item version of the Animal Attitude Scale (AAS-5) [55] to assess how this predicted interspecific CY. Although the reliability of the AAS-5 was poor, the internal consistency of the empathic concern scale was acceptable and included in the analysis. Moreover, this measure of affective empathy towards humans has previously been shown to predict attitudes towards animals [42,43]. Despite people generally showing high levels of empathy towards domesticated cats and dogs [37,38,39], this measure was not a positive predictor of interspecific CY. In fact, participants with greater empathic concern were less likely to yawn during testing, casting doubt on the view that interspecific CY is strongly linked to features of empathy or emotional contagion [16,23,31,32]. Together, these results add to a growing number of psychological studies providing mixed support for a connection between CY and empathy when using self-reported trait measures (reviewed by [17]). Based on exploratory analyses of the dataset from Gallup and Wozny [5], a secondary objective of the current study was to further examine the role of gender in interspecific CY among humans. Consistent with most studies in the literature, the current results revealed no overall gender bias in interspecific CY. However, a significant gender × stimulus condition interaction emerged whereby women reported a higher frequency of CY to dog yawns, and men reported a higher frequency of CY to cat yawns. Given that this particular result was not anticipated, we can only speculate as to the mechanisms governing this effect. Prior work has shown, for example, that interactions with pet dogs tend to enhance oxytocin (OT) among female owners, while they lead to no change or decreases in OT among men [62,63]. Moreover, in contrast to human interfaces with dogs, a recent study including only women found no overall increase in OT following interactions with pet cats [64]. Thus, one potential explanation for the female bias in CY to dog stimuli could be differences in OT among participants. However, we are skeptical of this possibility for at least two reasons. First, while OT has been implicated in CY [65,66], studies that have manipulated OT via intranasal administration have failed to yield support for this connection [57,67]. Second, the studies measuring changes in human OT levels were undertaken following real-world interactions with bonded cats and dogs (i.e., pets and their owners), and it seems unlikely that the current procedures of merely viewing static images of unknown animals would elicit similar neurochemical changes. Nevertheless, this represents a potentially fruitful area for future research to explore.Consistent with similar online studies examining both intra- and interspecific CY in humans [5,7], indices of arousal/fatigue were the strongest predictors of self-reported yawning in the current sample. As expected, sleep duration the night before was negatively correlated with interspecific CY, while tiredness during the study was positively correlated with this response. These findings align with a large body of comparative evidence suggesting that circadian rhythms and internal temperature drive yawning behavior (reviewed by [9]). While the current research serves as a replication of interspecific CY in humans and furthers our understanding of the factors contributing to this response, this study has several limitations. First, no measure of pet ownership was captured from the participants. Prior studies suggest that just over half of US households have a pet, with the majority reporting dog and/or cat ownership [68], and thus further work could assess how living and bonding with a pet cat or dog contributes to interspecific CY in these conditions. Another limitation to this study was the relatively small sample of participants, which was reduced primarily due to attentional checks. However, the attentional measures obtained during testing remains a strength of this study, as they ensured (1) high quality data and (2) that visual attention was directed towards the yawning stimuli. That said, the manipulation was not robust, particularly for the binomial outcome. Though common in the literature, the use of one-time measures of yawning to assess the relationship to psychological traits could be considered a further limitation. While we included a self-report measure of affective empathy, we did assess the ability to engage in empathy. Follow-up studies in this area could explore this further while also examining the relationship between interspecific CY and cognitive measures of empathy (i.e., emotion reading and perspective-taking). In addition, further work in this area could attempt to improve the ecological validity of this online experiment, perhaps including live demonstrations of yawning in the laboratory or the use of dynamic (i.e., video) yawn stimuli to enhance the participant response. While self-report CY has proven to be a valid measure of this behavior in psychology [3,56,57], future studies could attempt to capture a combination of both self-report and objective measures of CY.5. Conclusions In summary, this study replicates recent findings providing evidence for interspecific CY in humans. By comparing separate conditions for common household pets, we were able to show a slightly greater CY response to yawns of domesticated cats compared to domesticated dogs. Furthermore, when explicitly examining how individual differences in empathic concern and gender influence this response, our findings do not support a strong connection between interspecific CY and empathy or emotional contagion.

animals : an open access journal from mdpi

10.3390/ani12030402

PMC8833337

The use of traps is key to the success of many wildlife management programs but the species trapped, type of trap used and its application will influence the impacts it has on animal welfare. Scientific assessment of the impacts of trapping on mammal welfare is necessary to justify the use of traps, aid trap selection, improve trap performance and develop international standards. The Sharp and Saunders humaneness assessment model was developed for the purpose of assessing the relative humaneness of a range of pest animal control methods and has been used to assess the welfare impacts of trapping on various mammal species. The model is based on the established Five Domains model, the structure of which represents the understanding that an animal’s welfare state arises due to the sum of its mental experiences which may include pain, breathlessness, thirst or fear, among many others. Here we make key recommendations for those wishing to apply the Sharp and Saunders model to scientifically assess the welfare impacts of traps. Consideration of these points will help optimize the value of information produced using the model to support ethical wildlife management practice and policy and retain social acceptance of management programs that involve trapping.

Scientific assessment of the impacts of trapping on mammal welfare is necessary to inform cost-benefit analyses of using traps in wildlife management, improve trap performance and trapping processes and develop international trap standards. The Sharp and Saunders humaneness assessment model was developed specifically for assessing welfare impacts in vertebrate wildlife management and has been used to assess the impacts of trapping various mammals. It is a specific version of the more general Five Domains model for welfare assessment which is based on the understanding that welfare state reflects the sum of the animal’s mental experiences. Our experience of applying the Sharp and Saunders model allows us to make key recommendations for those wishing to use it. First, the exact parameters of the trapping scenario to be assessed must be decided. Second, assessments should be based on published data, as well as integrating both scientific and practitioner expertise to provide rigorous and relevant outcomes. Third, conclusions about welfare impacts should be based on the appropriate indicators. As far as is possible, mental experiences should be inferred using animal-based indicators, and some representation should be provided of the scorers’ confidence in the data on which assessment is based. Careful consideration of these points will help optimize the value of information produced using the model for wildlife management decision-making.

1. Scientific Assessment of Animal Welfare Is Important to the Success of Wildlife Management ProgramsThe use of traps is integral to the success of many wildlife management programs around the world. Various types of traps exist, and they can be categorized as restraining or killing traps. Restraining traps, which are designed to hold but not kill the animal, include cage or box traps, snares, net traps, glue boards, leg-hold or foot-hold traps. Killing traps include body-grip traps, snap (or break-back) traps and electrocution traps. The choice of trap relates to the objective of the program and the trapping activity, the environment, the mammal species targeted and features of the population and individual animals [1,2,3]. The type of trap employed, and its specific application, will influence its impacts on the welfare of trapped individuals and other affected animals [4,5]. Robust scientific assessment of the impacts of trapping on mammal welfare is necessary for various reasons. Such information is needed to inform cost-benefit analyses of using traps in wildlife management, as well as to improve trap performance and trapping processes, and develop international trap standards [6,7]. Any welfare impacts must be considered alongside the effectiveness, cost, ease of use, human safety, non-target animal impacts and social acceptability of the trapping method [8]. In some programs, traps are used to kill unwanted wild animals, directly or indirectly (by restraining the animal for killing using another method). In such cases, the ultimate goal of welfare assessment is to identify and minimize negative impacts on the animal before irreversible loss of consciousness and death [9,10]. In other cases, certain restraining traps are used to facilitate management of valued animals through activities such as health monitoring and vaccination, translocation, collection of demographic and other research data and non-lethal population management using fertility control, (e.g., [11,12]). In these contexts, identifying and minimizing welfare impacts is additionally important because such impacts can also hinder achievement of the primary goals of the work, which may be conservation, disease management, research or monitoring [10].More broadly, explicit and genuine consideration of the welfare of managed wild animals is critical to maintaining societal acceptance of such activities, as sectors of the public become increasingly aware of, and concerned about, animal welfare and wildlife conservation [13,14,15,16]. Thus, transparent and rigorous, science-based systems for assessing the welfare impacts of mammal trapping are critical to the future success of wildlife management programs.2. The Sharp and Saunders Humaneness Assessment Model Is Based on the Five Domains Model and Facilitates Systematic, Holistic, Data-Based Assessments of Relative Welfare Impacts of TrappingThe Sharp and Saunders humaneness assessment model [17,18] was developed specifically for assessing welfare impacts in vertebrate wildlife management and has been used to evaluate the impacts of trapping of various mammal species (Table 1). The Sharp and Saunders model is a specific version of the more general ‘Five Domains model for welfare assessment’, which evolved from its original use in research, teaching and testing and is now widely applied to the systematic assessment of animal welfare states in a range of contexts, (e.g., [12,19,20,21,22,23]). 2.1. Animal Welfare and the Five Domains Model for Welfare AssessmentThe Five Domains model is based on the understanding that an animal’s welfare state reflects the sum of its various mental (affective) experiences at a particular point in time. In other words, animal welfare is now commonly considered to describe how the animal itself is experiencing its world and life [27,28]. While not all animals are considered capable of affective experiences (i.e., sentient), in some legal jurisdictions, mammals are explicitly deemed to be sentient and their welfare is protected to some degree e.g., European Union via the Treaty of Lisbon (2008), French Civil Code (2015), New Zealand Animal Welfare Act (2015), Australian Capital Territory Animal Welfare Act (2015) and OIE Global Animal Welfare Strategy (2017). Mental experiences arise due to processing of sensory information by the animal’s nervous system. Sensory receptors gather information about the outside environment (e.g., visual or olfactory signals about a predator or conspecific) and about the animal’s internal physical state (e.g., body water levels, tissue damage and respiratory function) [29]. Processing of this information by the nervous system, in a way that is specific to the species and individual, leads to generation of mental experiences, some of which matter to the animal [30]. Such ‘affective’ experiences are negative or positive, and this valence influences behavioral and physiological responses in predictable ways [31,32]. Importantly when considering the impacts of trapping, negative or unpleasant mental experiences such as thirst, hunger, pain, breathlessness and fear act as signals to the animal to respond, in a specific way, to try to alleviate or rectify the underlying problem [29]. Such negative experiences are detrimental to an animal’s current state of welfare (although they may have survival benefits in some situations). Unpleasant experiences that cannot be effectively rectified through behavioral and physiological responses, and so persist (e.g., thirst that cannot be slaked by drinking or persistent fear associated with inescapable restraint or capture), will have a greater detrimental impact on welfare state than short-lived experiences or those over which the animal has some control [27].However, mental experiences cannot be measured directly, so must be cautiously inferred from observable indicators of the animal’s physical or physiological state or its behavior, which is permissible because of our knowledge of the links described above. This understanding of animal welfare, and the relationship between physical state and mental experiences, is reflected in the structure of the Five Domains and thus Sharp and Saunders models. These models facilitate systematic organization of the observable/measurable evidence and require interpretation of that evidence in terms of the animal’s likely mental experiences [28]. Briefly, observable evidence of physical/functional states (welfare indicators) is collated in four Domains, which represent the animal’s nutritional and hydration status (Domain 1), its physical and sensory environment (Domain 2), its health and functional state (Domain 3) and its behavioral interactions with other animals including humans (Domain 4). In Domain 4, ‘agency’ refers to the animal’s engagement in voluntary, goal-directed behaviors, and negative impacts can arise when agency is restricted, for example when the animal is unable to escape confinement, restraint or close contact with humans or predators. The fifth Domain reflects the mental experiences likely to arise due to impacts in those four physical/functional Domains (Figure 1). Potential welfare indicators include measures of physical health and condition and behavioral, physiological and neurophysiological responses. They should be used to infer not only the occurrence but also the type (e.g., pain versus breathlessness versus thirst), intensity and duration of specific affective experiences that contribute to the animal’s overall welfare state.2.2. Application of the Sharp and Saunders model to Assess Trap ImpactsThe Sharp and Saunders model focuses the users’ attention on the negative impacts of wildlife management procedures such as trapping and poisoning [26]. The model can be used to assess the relative humaneness of both lethal and non-lethal traps by considering the impacts of any processes leading up to capture (non-lethal components; Part A) separately from the action that directly brings about death (Part B). Thus, Part A is used to assess all impacts of non-lethal trapping or the non-killing aspects of lethal trapping. In particular, non-killing aspects of lethal trapping include any impacts of the trap’s presence in the animal’s environment prior to capture [24]. Evaluation is usually undertaken by a panel comprising expertise of various kinds (see Section 3.2 below). In Part A, impacts are considered in each of the five Domains and the degree/intensity of impact (score) is assigned with reference to a set of impact scales (one for each Domain) to facilitate more consistent scores among panelists (e.g., Figure 2; [18]). The impact in Domain 5 arises from impacts in Domains 1 to 4 and is usually the highest of those scores. Ultimately, an overall impact score is assigned which usually represents the score assigned in Domain 5, as mental experiences are deemed to be most relevant to welfare state—a higher overall impact score (Domain 5) represents a greater number and/or intensity of unpleasant experiences for the animal. The total duration of the non-lethal impacts is then estimated (immediate to seconds, minutes, hours, days and weeks), and the intensity and duration are integrated using a scoring matrix to assign an overall grade ranging from 1 to 8 (Appendix A Figure A1; [18]). The higher the overall grade, the more intense and/or long-lasting are any welfare impacts associated with the non-lethal components of trapping. The welfare impact of a non-lethal/restraining trapping method can be represented entirely by its Part A score; Part B of the model need not be applied.For lethal trapping, Part B of the model is applied in addition to Part A, to evaluate the intensity and duration of welfare impacts associated with the killing method itself, including any handling required. Here, the intensity of suffering (integration of all unpleasant experiences occurring: No impact, Mild, Moderate, Severe and Extreme suffering) is assigned using an impact scale different from that used in Part A [see 18]. The duration of any such suffering is estimated as the time from which the method begins to have an impact (such as when a wolf becomes caught in a killing snare or a rat is struck by a snap trap) to the point at which consciousness is irreversibly lost; after this point, no affective experiences are possible. These components are combined using a separate scoring matrix to assign an overall score ranging from A to H (Appendix A Figure A2)—the higher the overall score, the greater (i.e., more intense and/or longer-lasting) any suffering associated with the killing method is deemed to be. Thus, any lethal trapping method can be ranked according to its Part A (non-lethal components; 1–8) and Part B (lethal components; A–H) scores. Likewise, if a lethal method (e.g., concussive blow to the head, cervical dislocation, injection of a lethal agent) is applied to the live-trapped animal, Part A and Part B assessments can be applied separately for the two components. If there are various options available for killing live-trapped animals, multiple assessments may be conducted to determine which killing method is the most humane. Box 1 provides an example of the relative welfare impacts of various methods of lethally controlling wild dogs in Australia, including evaluation of various restraining traps and subsequent options for killing. Alternatively, if the live-trapped animal is instead subjected to subsequent non-lethal procedures such as blood sampling, medical treatment or translocation, separate assessments may be conducted, again using Part A.Evaluations are often performed with reference to a standard operating procedure for trapping in the specific context and informed by review of the relevant scientific literature by all panelists as well as by panelists’ personal experience of the procedures under assessment. Generating panel-level outcomes has been achieved in various ways, usually through panelists undertaking independent scoring followed by development of some form of consensus through discussion (e.g., [18,24]) or by representation of the range of scores within the panel, (e.g., [33]). Aiming for group consensus through discussion of the likely welfare impacts in each Domain and overall should improve the reliability of the outcomes and encourage stakeholder acceptance of the results.Box 1An example of the relative welfare impacts of various methods of lethally controlling wild dogs in Australia.Summary figure (Figure 3) showing the relative welfare impacts of various methods of lethally controlling wild dogs (Canis familiaris) in Australia, including various restraining traps, using the Sharp and Saunders humaneness assessment model [18]. These assessments were undertaken on the protocols outlined in Standard Operating Procedures (SOP) and it was assumed that traps were checked every 24 h. Only impacts on the target animal were assessed and the scores represent consensus reached by the evaluation panel. The details of each assessment and a brief summary of the justification of the impact scores are shown in Table 2 and Table 3—see [18] for references.These results illustrate the versatility of the model, in that the relative welfare impacts of diverse control methods can be compared. Additionally evident is the importance of clearly defining the scenario to be assessed and the value of being able to assess the non-lethal (Part A) and lethal (Part B) aspects of trapping separately. For example, the impacts of capture/restraint in a padded foot-hold or leg-hold trap are moderate to severe when applied according to the SOP (i.e., current best practice). However, the method of killing the trapped animals has a profound effect on the overall welfare impact: killing by head shot causes much less additional welfare impact (suffering) than ingestion of strychnine poison. Confidence scores were not generated for these assessments.Figure 3Summary of the relative welfare impacts of various methods of lethally controlling wild dogs (Canis familiaris) in Australia, including various restraining traps, using the Sharp and Saunders humaneness assessment model. Reprinted from A model for assessing the relative humaneness of pest animal control methods. Second edition. (p. 124), by T. Sharp and G. Saunders, 2011, Australian Government Department of Agriculture, Fisheries and Forestry. Reprinted with authors’ permission.animals-12-00402-t002_Table 2Table 2Part A assessments of three types of restraining (non-lethal) traps used for wild dogs (Canis familiaris) in Australia. Summarized from data available at pestsmart.org.au (accessed 18 January 2022).TrapDomain 1Domain 2Domain 3Domain 4Domain 5DurationImpact ScorePadded foot-hold traps (e.g., Victor Soft Catch #3)Mild(No food/water for ≤24 h)Mild(Assumes fair weather and in shade)Mild–Moderate(Mostly minor skin lacerations; Some leg dislocations, tooth/mouth injuries)Moderate (Stress hormone levels high, struggling, disruption of natural behaviours)Moderate(Anxiety, fear, frustration due to restraint, pain from injuries and struggling)Hours5Padded leg-hold traps (e.g., ‘off the shelf’ padded Lanes Dingo trap)Mild(No food/water for ≤24 h)Mild(Assumes fair weather and in shade)Moderate–Severe(Heavy traps tend to catch higher on leg; Leg fractures and amputations, tooth/mouth injuries)Moderate(Stress hormone levels high, struggling, disruption of natural behaviours)Moderate–Severe(Anxiety, fear, frustration due to restraint, pain from significant injuries and struggling)Hours5–6Cage trapMild(Food bait but no water for ≤24 h)Mild(Assumes fair weather and in shade)Mild(Minor injuries; tooth, mouth, nose)Moderate(Some stress due to restraint, struggling, disruption of natural behaviours)Mild(Anxiety, distress due to restraint)Hours4animals-12-00402-t003_Table 3Table 3Part B assessments of three methods of killing restrained wild dogs (Canis familiaris) in Australia. Summarized from data available at pestsmart.org.au (accessed 18 January 2022).MethodSufferingDurationImpact ScoreShooting (head)Mild(Approach of human will cause some distress)Immediate—SecondsBStrychnineExtreme(Nervousness, stiffness, progressively more frequent and intense tetanic seizures, extensor rigidity, hyperthermia, death due to exhaustion or asphyxiation)HoursGLethal injectionMild(Approach of human will cause some distress; some pain associated with intramuscular injection)MinutesC2.3. Advantages of the Sharp and Saunders Model for Assessing Trap Impacts on WelfareThe Sharp and Saunders model facilitates systematic, data-based, transparent and holistic assessment of the welfare impacts of trapping and other wildlife management procedures [34]. Importantly, the structure of the model encourages users to interpret the observable/measurable data in terms of what it means to the animal itself i.e., the likely affective states experienced, which is consistent with a contemporary understanding of animal welfare. The systematic nature of the assessment allows those wishing to apply the outcomes to see exactly how the scores were generated and the data upon which those scores were based. In addition, the structure of the model encourages users to look for, and organize, evidence of a wide range of physical impacts and their associated mental experiences and to consider ways to specifically mitigate those impacts when developing novel or modified trapping methods [24,35]. The model also assists in identifying any gaps in knowledge that can be addressed by future research to improve subsequent welfare assessments [24,34].The model is versatile and can be applied to assessments of a range of trapping procedures and mammal species, and separate evaluations in Part A and B allow assessment of both lethal and non-lethal traps [18]. This versatility is demonstrated by comparisons of a diverse range of pest management methods, including fencing, scaring, habitat damage management, translocation and live release, shooting and vertebrate toxic agents (poisons), as well as various kinds of traps, in various vertebrate species [9,18,24,25,26]—see Box 1. The relative welfare impacts of a broader range of wildlife management activities could likewise be assessed, including fertility control [11], reintroduction [36], identification marking [5] and other research or management procedures [37]. However, this may require development of more generic reference scales for judging impacts in each of Domains 1 to 5. For example, it is feasible that reintroduction, medical treatment or supplemental feeding would have some beneficial effects for target or other animals, and thus conditions/states likely to lead to positive experiences may need to be integrated into impact scales (see Mellor and Beausoleil [28] for discussion of integrating negative and positive impacts into overall scores).The Sharp and Saunders model can be used prospectively to inform consideration of the ethical permissibility of proposed traps or new uses of existing traps as well as to retrospectively evaluate actual impacts of trapping to support selection of the most humane methods, i.e., the methods with the lowest welfare impacts [38]. However, it is important to note that, as with all assessments of welfare based on inference of mental experiences, the outcomes are inherently qualitative in nature. Ordinal, rather than interval, scales are deliberately used to assign impact scores in each Domain to ensure that the qualitative nature of the assessment is explicit. The use of interval impact scores would imply a degree of precision that is neither possible nor desirable in such assessments. For example, such scores could be used to inappropriately suggest the magnitude of differences between methods (e.g., welfare impacts of trap A are twice as bad as those of trap B) [34]. In addition, it can be difficult to compare the overall impacts of different types, intensities or durations of unpleasant mental experiences that influence welfare state. For example, is a trap that causes moderate pain for days preferable to one that causes extreme breathlessness for hours [34]? These are not questions that can be resolved by scientific investigations, and there is unlikely to be a universally acceptable answer. Thus, careful application of the model, taking the following key points into consideration, will allow qualitative comparison of welfare impacts associated with different options for achieving a particular wildlife management objective.3. Key Considerations for Applying the Sharp and Saunders Model and Applications of the Outcomes of Such AssessmentsOur collective experience of applying the model over the last 13 years leads us to highlight a number of key considerations for those wishing to use it. Careful application of these points will help optimize the value of the information produced when using the model for wildlife management decision-making and policy development.3.1. Determining the Right Trapping Scenario for AssessmentFirst, time should be allocated to decide on the exact parameters of the trapping scenario to be assessed; these can strongly influence the outcomes and their value for decision-making. For example, is it constructive to assess the ‘typical’ trapping event, the best- or worst-case scenario or devise an approach which integrates the likelihood of certain events occurring within the trapped population? A common approach has been to assess trapping applied according to a best practice standard operating procedure, (e.g., [24])—and see Box 1—but users should consider how well this will represent real-world applications and outcomes. To illustrate, while restraining traps are designed to be non-lethal, they sometimes cause the death of the trapped animal, for example through dehydration and/or exposure if traps are not checked regularly enough [39] or through suffocation if the animal’s muzzle is caught on a glue trap [40]. Likewise, if a killing trap is inappropriate for the species or set incorrectly, animals can be mis-caught, causing them to be restrained and potentially injured, but not killed [41,42]. Both such scenarios will result in different welfare impacts than when the trap operates optimally and when best practice procedures are followed [24,41]. Thus, it may be valuable in some cases to assess the range of possible welfare outcomes for a given method and to compare the likelihood and effects of divergence from ‘best practice’ in practical trapping situations or to examine the likely welfare gains that could be achieved by making specific changes to best practice. In all cases, the scenario to be assessed should include detailed information on the way the trap is presented in the environment and the procedures and conditions leading up to trap capture, as well as those that occur after capture in the case of live-traps.3.2. Assessments Should Be Robustly Evidenced, and Panels Include Diverse ExpertiseSecond, assessments should be conducted by a panel that includes both academic and practitioner expertise, to provide rigorous, relevant and credible outcomes. Importantly, assessments should be led by a facilitator familiar with the model to ensure a systematic and balanced process that adheres to the underlying principles of welfare evaluation using this framework. We have found that the best results are achieved by panels including those with expertise in the management techniques and the species involved, those with expertise in veterinary physiology and pathology, and, importantly, those with expertise in the general scientific principles of animal welfare evaluation and the specific application of the Five Domains/Sharp and Saunders models. For example, the panel for a recent assessment of the welfare impacts of rat management included experts in wildlife management, rodent management, rodent biology, animal welfare science and veterinary science and medicine [24]. Failing to include those with science/biology expertise may lead to misinterpretation of clinical signs of physiological disruption or under- or over-estimation of the significance of injuries sustained. In contrast, lack of practical management expertise can mean that results are irrelevant to control as it is carried out in the real world or fail to integrate accumulated knowledge about wild animal behavior or the probability of different outcomes in trapped populations (e.g., bodily location of trap strike), reducing the credibility and applicability of the findings.The panel should assess each method strictly as described according to the pre-defined scenario. Assessments should be informed by the relevant scientific literature, with panelists drawing on their own particular expertise to interpret the information available in terms of welfare impacts. Data published in the literature should be reviewed (and summarized if necessary) for panelists to read in advance of the assessments. Some of these data can be taken from studies specifically aiming to assess welfare impacts, but inevitably some data must be extrapolated from studies focused on the mode of action or efficacy of the method for achieving its conservation purpose, e.g., [4,33].As the panelists work through each assessment, they should discuss the likely welfare impacts, in the context of their experience and the literature, aiming to reach consensus by doing so. A useful approach may be for panelists to make their own individual assessments following the group discussion and then to reach consensus on a score or range of scores through further discussion [18,24]. Alternatively, individual panelist’s scores may be simply represented as the median and range of the group’s scores [33].Importantly, when making a Part A assessment, the facilitator should lead the panelists through consideration of each Domain and a detailed summary of the data (and their sources) and the thinking upon which the impact score has been based should be recorded. Likewise, a summary of data and justification of scores should be presented for Part B assessments. A way of indicating the panelists’ confidence in the scores produced using the available data is discussed below. 3.3. Inferences of Animals’ Mental Experiences Should Be Based on Appropriate Indicators and an Indication of the Panel’s Confidence in Their Scores Should Be PresentedThird, conclusions about welfare impacts should be based on the appropriate indicators. As far as is possible, trapped animals’ mental experiences should be inferred using animal-based indicators. These are indicators that represent the outcome of the animal’s perception and interpretation of what happens to it and thus provide the strongest justification for inferring mental experiences and overall welfare state. Examples include measures of pathology, physiology and behavior. In contrast, resource- and management-based indicators represent risks to the animal’s welfare (inputs) but do not provide direct evidence that the resources or management are, in fact, affecting the animal’s mental state [12]. Examples include trap mechanism, space within a cage trap or animal handling. Clearly, this approach is not always possible due to a lack of validated or practically measurable animal-based indicators. Alternatively, the relationship between the event/condition and the animal’s response is so well established that resource- or management-based indicators can be used in lieu of animal-based indicators [43].In all applications of the model, the onus is on the user to justify inferences of specific mental experiences for the taxon being assessed, as well as to demonstrate the validity of the indicators considered to reflect those experiences [43]. To illustrate, a particular challenge for assessing the duration of suffering and thus relative humaneness of kill traps has been validating indicators of loss of consciousness [25,33]. At a minimum, providing an indication of the panel’s certainty about particular inferences and thus conclusions about welfare impacts is recommended.Our experience suggests that in addition to presenting summaries of the data on which impact scores and durations are based, it is highly beneficial to also collect and present indications of the panelists’ confidence in their scores, particularly their scores for Domain 5: mental experiences (e.g., Table 4). For example, a wider range of impact scores among panelists usually arises when there is little information available about the effects of a particular trap and this can be represented as a lower confidence score [24]. This information allows richer interpretation of the outcomes of the assessment, as well as directing research efforts to improve future understanding where data are lacking [9].4. Concluding RemarksScientific assessment of the impacts of trapping on mammal welfare is necessary to support ethical wildlife management practice and policy and to retain social acceptance of management programs that involve trapping. The outcomes of welfare assessments using the Sharp and Saunders model can be used to develop relative rankings of the humaneness (i.e., welfare impacts) of different options for achieving management objectives and to explore ways to circumvent and/or mitigate welfare impacts and develop more welfare-friendly methods when there are no feasible alternatives [44]. Ultimately, the findings of welfare assessments should be used to inform and justify decisions about if, when and how to implement trapping activities for ‘ethical’ wildlife management [8,38]. In addition, the information gleaned from such assessments should be applied to develop international standards for trap approval [7]. Here, we have emphasized key considerations for optimal use of the Sharp and Saunders model for assessing the welfare impacts of mammal trapping; these considerations will apply equally to applications of the model more broadly in the fields of wildlife management and research. Assembling the right complement of expertise on assessment panels, including a knowledgeable facilitator and at least one expert in the application of the model itself, along with careful selection of the scenarios to be assessed and presentation of scores reflecting the panel members’ confidence in the underpinning data, will help optimize the value of the information produced using the model for wildlife management decision-making and policy development.

animals : an open access journal from mdpi

10.3390/ani11092719

PMC8469829

In the United States, castration is a common husbandry procedure utilized in the cattle industry. Despite castration being painful, it is commonly performed without the use of analgesia, one reason being the lack of available approved analgesics in the United States for use in alleviating pain associated with castration in cattle. Additionally, if pain mitigation is used, it is more often provided to older animals as there is a notion that younger animals experience pain to a lesser degree than older ones. The aim of this study was to characterize physiological responses to unmitigated surgical castration in calves of varying ages in terms of cortisol concentration, heart rate variability, and changes in eye temperature. Overall, our results indicate that the measured physiological responses to castration differed between age groups and changed over time post-castration. Younger calves showed a different response pattern than older calves for many of the variables measured suggesting that the response to castration-induced pain may be age-specific. For example, the youngest calves had lower cortisol and average eye temperature as compared to the oldest calves. Additionally, many variables showed a differential response to castration-induced pain, as compared with simulated castration, thus suggesting physiological indicators that could be targeted in future development and validation of analgesics for alleviation of pain associated with castration in cattle.

The objective was to characterize physiological responses to unmitigated surgical castration in calves of varying ages. Thirty male Holstein calves of three ages [<6 w (6W); 3 m (3M); 6 m (6M); n = 10] underwent a simulated castration treatment (SHAM) followed 24 h later by castration (CAST). For both treatments, heart rate variability, eye temperature, and cortisol were measured over time from treatment to specified end points to capture the acute response period. Interactions between treatment and age (p = 0.035) and time and age (p < 0.001) were noted for cortisol. The 6W calves had lower cortisol compared to 6M calves at SHAM and CAST. Cortisol of 6W calves decreased from peak to pre-treatment levels faster than 6M calves. An interaction between time and age was reported in squared differences of inter-beat-intervals (RMSSD; p = 0.02) and high-frequency power (HFP; p = 0.05), whereby both responses decreased in 6W calves during the sampling period which was not seen in 3M and 6M calves. Average eye temperature (AET) differed by age (p = 0.0018) whereby 6W calves had lower AET than 6M calves (p = 0.0013) regardless of treatment and time. The findings suggest that responses to unmitigated surgical castration seem to be mediated by the autonomic nervous system in an age-related manner.

1. IntroductionBovine castration is a painful procedure typically performed without analgesia in the United States [1,2], although stakeholder groups do encourage the adoption of techniques to minimize pain and distress associated with the procedure [3,4,5]. One of the limitations with providing pain mitigation for castration in the United States is the lack of available FDA-approved drugs to control pain associated with castration [1,2,6]. Although there are several approved analgesics for use to control pain in cattle available in other countries, in the United States, there is only one approved drug authorized for this use (e.g., transdermal flunixin meglumine, specifically for pain related to interdigital phlegmon, i.e., foot rot; [7]). Identifying reliable and repeatable methods of pain assessment in cattle is key to informing recommendations on analgesic use for alleviating pain from common management practices, such as castration.Pain is defined as an aversive sensation associated with actual or potential tissue damage, resulting in physiological, neuroendocrine, and behavioral changes that indicate a “stress” response [8,9]. Although castration is generally considered more painful in older calves, studies examining physiological and neuroendocrine responses to pain as a function of age are lacking. There are several physiological measures that can be used to measure the activation of these systems in response to a painful procedure such as castration in cattle. For example, blood cortisol concentration has been shown to increase in calves following castration (a selection: [10,11,12,13,14]). In addition, heart rate variability (HRV) is used to measure autonomic nervous system (ANS) regulation of cardiac function (e.g., sympathetic and vagal tone on the heart) and has been used to assess pain response in cattle associated with procedures such as castration with or without pain mitigation [15,16,17]. Another measurement used to assess pain in cattle in response to castration is infrared thermography (IRT), a noninvasive technique for visualization of a surface thermal profile. Eye temperature has been shown to be an effective tool to measure ANS activity in cattle related to castration [16,17,18,19,20].The aim of this study was to characterize acute physiological responses to surgical castration in calves of varying ages without pain control (i.e., unmitigated) in terms of cortisol concentration, HRV, and changes in eye temperature. This is critical information needed to support evidence-based recommendations of pain control for husbandry practices in the United States. We hypothesize that unmitigated castration would affect these indicators in an age-dependent manner.2. Materials and MethodsResearch reported here was part of a large study reported in a companion paper [21] and therefore the experimental design and sampling methods are as described in that study. This research was approved by the Institutional Animal Care and Use Committee at Kansas State University (Protocol #2831). This study was a component of a federally funded grant (USDA-CSREES NRI Award No. 2009-65120-05729) exploring differences in pain response in varying ages of cattle.2.1. Animals and HousingThirty male Holstein calves of <6 weeks (6W; 52 ± 9 kg), 3 months (3M; 89 ± 5 kg), and 6 months (6M; 139 ± 11 kg) of age (10 calves per age group) were enrolled in the study during Summer 2010. All study animals were housed at the Kansas State University Beef Cattle Research Center (BCRC; Manhattan, KS, USA). Calves came from one Kansas dairy herd and were acclimated at the BCRC for 10 d prior to study initiation. All study animals received a 4-way modified-live viral respiratory disease vaccine (Bovishield Gold, Pfizer, New York, NY, USA) and oxytetracycline (Noromycin 300 LA, Norbrook Laboratories Station Works, Newry, Co.Down N. Ireland; 9 mg/kg bodyweight IM).Pre-weaned 6W calves were housed in individual wire-panel enclosures (1.6 m × 5.3 m). All enclosures were under a roof and hutches were not provided. Calves did have contact with other calves across the wire-paneling. Calves were bottle-fed milk replacer (Maxicare, Land O’Lakes, Animal Milk Products Co. 039, Shoreview, MN, USA) and provided ad libitum water and starter ration (Herd Maker Supreme B90, Land O’Lakes, Animal Milk Products Co. 039, Shoreview, MN, USA). The 3M and 6M calves were housed by age group on outdoor concrete pads (9.8 m × 18.3 m). Enclosures had a partial roof and straw bedding. Calves were provided with water and grass hay ad libitum in addition to receiving a grain-based supplement provided at 3–4 kg/head/day. To assist with subsequent data collection, calves were restrained with a rope halter, head gate, and girth straps for 30 min daily during the acclimation period.2.2. Jugular CatheterizationJugular catheters were placed following the methods of Bergamasco et al. [21]. In brief, the area over the jugular vein was clipped and 70% isopropyl alcohol and povidone iodine swabs were used to disinfect the area. Prior to performing a small skin incision to facilitate placement of a 14 G x 130 mm extended use catheter (MILACATH®, MILA International, Florence, KY, USA), the site was infiltrated with 2% lidocaine injection (Lidocaine Hydrochloride Injection, USP (2%) (20 mg/mL), Hospira Inc, Lake Forest, IL, USA). The catheter was sutured in place using #3 nylon suture (Braunamid®, Braun, Bethleham, PA, USA). A heparin saline flush containing 3 USP units heparin sodium/mL saline (Heparin Sodium Injection, Baxter Healthcare, Deerfield, IL, USA) was used to maintain catheter patency.2.3. Experimental ProcedureEach calf was submitted to two experimental procedures: simulated castration (SHAM) and surgical castration (CAST). The CAST treatment was performed 24 h after the SHAM treatment for each calf. Within each age group, animals were blocked by bodyweight and scrotal circumference. Processing date and an order of processing within the day were randomly assigned to each animal to avoid confounding effects. Both SHAM and CAST procedures were conducted between 0600 and 1030 on each processing day at 45 min intervals. Each calf was restrained in the chute with head movement limited by a halter drawn close to a table attachment (For-most Livestock Equipment, Hawarden, IA, USA) for approximately 30 min in order for experimental procedures to be completed; additional measurements to what is reported in this manuscript were taken [21]. The same operator performed all surgical procedures. For both treatments, the scrotum was washed with chlorhexidine disinfectant. The SHAM and CAST experimental procedures followed Bergamasco et al., [21]. For SHAM, the testes were firmly grasped, and ventral traction was applied for approximately 20 s. For CAST, the lower one-third of the scrotum was cut with a sharp, disinfected scalpel and the testes and spermatic cords were exteriorized followed by manual traction until the spermatic cord and connective tissue ruptured. This time point is indicated as “treatment” in the study timeline. After the experimental procedures, calves were returned to their home pens with access to feed, water, and rest. Calves were monitored frequently during post-castration data collection for pain for 8 h following surgery, and then twice daily for 7 d. Per the animal care and use protocol, animals were checked for signs of excessive pain based on the evaluation of the attitude, gait and posture, appetite, lying and scrotal swelling. Pain mitigation was not provided for the experimental procedures.2.4. CortisolBlood samples were collected from jugular catheters immediately prior to the treatment (time 0) and again at 5, 10, 20, 30, 40, 50, 60, 120, 240, 480 and 720 min after castration or simulated castration. Halters remained on the animals to facilitate restraint during blood collection. Blood was drawn into serum clot activator tubes (Vacuette 6 mL Z Tubes, Greiner Bio-One, Kremsmünster, Austria) and was centrifuged at 1500× g for 10 min within 30 min of collection. The serum was pipetted off with transfer pipettes (Graduated 3 mL Transfer Pipettes Large Bulb, Samco Scientific, San Fernando, CA, USA), stored in 2 mL cryogenic vials (Fisherbrand Cryogenic Storage Vials, Fisher Scientific, Pittsburgh, PA, USA), and frozen at −80 °C prior to analysis for cortisol. Samples were analyzed within 2 months after collection. Serum cortisol concentrations were determined using solid-phase competitive chemiluminescent enzyme immunoassay and an automated analyzer system (Immulite® 1000 Cortisol, Siemens Medical Solutions, Los Angeles, CA, USA; [22,23]). The laboratory technician performing the analysis was unaware of age classification and treatment. The Cmax and the Tmax were observed directly from the data. The AUC was calculated using the trapezoidal method.2.5. Heart Rate VariabilityOn the day of each session (SHAM and CAST), approximately 1 h prior to Time 0, calves were haltered and fitted with heart rate monitoring equipment. Heart rate was recorded continuously using Polar heart rate monitors (S810i™, Polar Electro Oy, Helsinki, Finland). The left side of the animal was shaved and ultrasound gel (Ultrasound Gel, Medline Industries Inc., Mundelein, IL, USA) was applied to facilitate electrode contact with the thorax. The transmitter and receiver were fixed to the animal using an elastic belt and cohesive flexible bandages (Fisherbrand Cohesive Flexible Bandage, Fisher Scientific, Pittsburgh, PA, USA). Heart rate monitors remained on the calves while they were in the chute (approximately 30 min) during the treatment session to capture the acute responses to treatment and were then removed. At the end of each sampling period the stored data were downloaded onto a computer for analysis. The time points used for the HRV analysis included baseline (5 min before treatment; Base), early recovery (0–5 min after treatment; R05), middle recovery (5–10 min after treatment; R510), and late recovery (10–20 min after treatment; R1020) for both SHAM and CAST sessions. Equal time periods of 5 min were analyzed to fulfill recommendations for analysis of HRV [24]. Continuous recordings of R–R (interbeat) interval data are prone to measurement errors therefore, prior to analysis, a correction function within the Polar software (Polar Precision Performance Software; Version 4.03), set on default parameters, was used to correct for any artifacts (e.g., to eliminate ectopic heartbeats). Time domain parameters included heart rate (HR) and the square root of the mean squared differences of successive inter-beat-intervals (RMSSD).Frequency domain parameters included high-frequency power (HFP) (0.30–0.80 Hz), the low-frequency power (LFP) (0.04–0.30 Hz) and the LFP/HFP ratio, which were calculated using a parametric method based on an autoregressive model for frequency-domain analysis provided by Polar software (Polar Precision Performance Software; Version 4.03). The HFP and LFP were expressed in normalized units [as a percentage or proportion of total power (e.g., LFP/total power × 100 or HF/total power × 100) to account for inter-individual differences. The HRV data were analyzed by an operator blinded to treatment and age classification.2.6. Infrared ThermographyInfrared images of the eye region were collected at a consistent distance (approximately 0.5 m) and angle (90°) from the left side of the animal using an infrared camera (ThermaCAM® P65HS, FLIR Systems, Wilsonville, OR, USA). Maximum, minimum, and average temperature (°C) within the area of the medial posterior palpebral border of the lower eyelid and the lacrimal caruncle were recorded every 4 to 6 s throughout the entire treatment session while the calves were in the chute. The time points analyzed included immediately before treatment (PRE) and immediately after treatment (POST). Images were analyzed for changes in temperature using research-grade software (Thermacam Researcher Pro 2.8 SR-1, FLIR Systems, Wilsonville, OR, USA). Ambient temperature and relative humidity in the barn were recorded and entered into the infrared camera to ensure calibration for atmospheric conditions. The IRT data were analyzed by an operator blinded to treatment and age classification.2.7. Statistical AnalysisA general linear mixed model was fitted to each response variable included in the study. The LFP/HFP ratios were log-transformed to ensure variance stabilization, and normalized LFP and HFP were arcsine square root transformed for similar reasons. In all models for all response variables, the linear predictor included the fixed effects of treatments (namely the experimental sessions: SHAM vs. CAST), age, time points specific to each response and all 2- and 3-way interactions. The random effect of calf nested within the age group was included in the model to recognize calf as the experimental unit for age and the blocking factor for treatment. Moreover, the random effect of calf-by-treatment combination recognized calf as the experimental unit for treatment, and the random effect of calf-by-treatment-by-time combination accounted for technical replication, where applicable. Model assumptions were evaluated using studentized residuals and were considered to be appropriately met. The models were fitted using the GLIMMIX procedure of SAS (Version 9.2, SAS Institute, Cary, NC, USA), implemented using Newton–Raphson with ridging as the optimization technique. Results are presented in the original data scale (estimated least square means and corresponding standard errors or 95% confidence intervals). Relevant pairwise comparisons were conducted using Tukey–Kramer or Bonferroni adjustments, depending on the level of inference, to avoid inflation of Type I error rate due to multiple comparisons. p-values at or below 0.05 were used to determine statistically significant differences. p-values < 0.10 were considered indicative of marginal significance. 3. Results3.1. Cortisol ConcentrationsFor cortisol concentrations, 2-way interactions were identified between treatment and age (p = 0.035), between time and age (p < 0.0001) and between treatment and time (p < 0.0001). Therefore, results are presented by evaluating cortisol dynamics over time after SHAM and CAST for each age group. During SHAM, cortisol concentration decreased from peak levels to pre-treatment levels within 50 min post-treatment in 6W calves, whereas for 3M and 6M calves, cortisol concentrations returned to baseline values 120 min after the treatment (Table 1). During CAST, cortisol concentrations for 6W and 3M calves decreased from peak levels to pre-treatment levels within 120 min after treatment, whereas for 6M calves, cortisol concentrations returned to baseline values 240 min after treatment (Table 2).Regarding cortisol AUC, only a significant main effect of treatment was identified (p = 0.0071), whereby cortisol AUC was greater for calves in CAST (18,719 ± 1015 nmol/L) relative to SHAM (15,222 ± 1011 nmol/L) across all age categories. No evidence for differences in AUC was apparent between age groups (p = 0.1460).For Cmax, a 2-way interaction between age and treatment was identified (p = 0.0184), whereby, when subjected to SHAM, 6M calves had a lower Cmax (58.23 ± 20.03 nmol/L) than 3M calves (144.17 ± 7.61 nmol/L; p < 0.05) but Cmax was marginally different in 6W calves (129.51 ± 18.48 nmol/L; p = 0.0657). Additionally, during CAST, 6M calves had lower Cmax (83.36 ± 20.03 nmol/L) than 3M calves (134.90 ± 7.61 nmol/L; p = 0.033) and Cmax was marginally different from 6W calves (147.56 ± 18.48 nmol/L; p = 0.0953). The magnitude of difference in Cmax between 6M and 3M calves in SHAM was not as great as compared to CAST.Finally, for Tmax, a main effect of treatment was observed (p = 0.008), whereby Tmax was greater following CAST (26 ± 2 min) compared to SHAM (20 ± 2 min) for all age categories. No evidence for any differences in Tmax was apparent between age categories (p = 0.3156).3.2. Heart Rate VariabilityWhen HR was evaluated, evidence for a 2-way interaction between treatment and time was found (p = 0.0024; Figure 1). At baseline, none of the age groups showed any evidence for differences in HR between SHAM and CAST. However, at R05 and R510, HR was lower for CAST relative to SHAM across all age groups (p = 0.0009 and p = 0.012, respectively). In addition, a significant main effect of age was identified on HR (p = 0.0006); whereby, regardless of SHAM or CAST, 6W calves showed greater HR (105.7 ± 5.66 beats/min) than older calves (3M: 84.33 ± 5.39 beats/min; 6M: 71.52 ± 5.39 beats/min; p = 0.0280 and p = 0.0004 respectively). No evidence for differences in HR was apparent between 3M and 6M old calves (p = 0.2304).For RMSSD, there was a 2-way interaction between age and time (p = 0.0178), whereby for 6W calves, RMSSD decreased from the middle (R510; 68.8 ± 11.9 ms) to late (R1020; 48.8 ± 11.2 ms) recovery times (p = 0.0389) for both CAST and SHAM; this was not apparent in 3M and 6M calves.For HFP, a 2-way interaction between treatment and age (p = 0.0133) was identified whereby during SHAM 6M calves had lower HFP than younger (3M and 6W) calves (p = 0.0126 and p = 0.0282, respectively; Figure 2A). During CAST, 6M calves had lower HFP than 3M calves (p = 0.0048) whereas the difference between 6M and 6W calves was marginally significant (p = 0.0613). Additionally, a 2-way interaction between age and time (p = 0.0517) was observed, such that 6W calves showed a decrease in HFP from R510 to R1020 (p = 0.0008; Figure 2B). By contrast, 3M and 6M calves showed no evidence of any change in HFP over time.3.3. Infrared ThermographyFor average eye temperature (AET), a 2-way interaction was apparent between treatment and time (p = 0.029), whereby differences in AET were identified between SHAM and CAST both before treatment (PRE; SHAM: 34.87 ± 0.26 °C; CAST: 35.41 ± 0.26 °C; p = 0.0179) and immediately after treatment (POST; SHAM: 34.47 ± 0.27 °C; CAST: 35.26 ± 0.26 °C; p = 0.001), though the POST difference was of greater magnitude. In addition, age differences were also noted on AET (p = 0.0018), whereby 6M calves (36.1 ± 0.4 °C) had greater AET than 6W calves (33.8 ± 0.4 °C; p = 0.0013) regardless of treatment and time; 3M calves had an intermediate AET (35.2 ± 0.4 °C) that was not significantly different from the other age groups.Regarding minimum eye temperature, there was a marginally significant effect of treatment (p = 0.0523), such that, regardless of age, minimum eye temperature was greater during CAST (30.1 ± 0.5 °C) than during SHAM (29.1 ± 0.5 °C).4. DiscussionThis study characterized the acute physiological responses to unmitigated surgical castration in calves of varying ages, specifically cortisol concentration, heart rate variability, and changes in eye temperature, with the ultimate goal of supporting pain mitigation strategies for routine management practices like castration. Additional outcomes are reported in Bergamasco et al. [21], a companion paper that can be referred to for further information. Despite the fact that surgical castration is a painful procedure having both acute and chronic impacts, the use of pain control is not routinely practiced in the United States [1,2]. Surveys of producers and/or veterinarians exploring pain mitigation use for various management procedures including castration have demonstrated that analgesia use increases with cattle age [2,25]. Industry guidelines for castration procedures suggest performing the procedure at the youngest age possible to reduce stress [5]. Although responses to pain associated with castration have been shown to be age-specific [17], young calves do experience acute pain as evidenced by changes in physiological and behavioral responses to painful stimuli [8,10,21]. It is critical, as pain mitigation associated with painful husbandry procedures remains an important welfare consideration, that stakeholders understand the age-specific differences in pain response related to calf age in order to establish best practices for pain relief.Additionally, the United States Food and Drug Administration Guidance Document 123 for the development of effectiveness data for non-steroidal anti-inflammatory drugs recommends that validated methods of pain assessment are used to evaluate a drug indicated for pain relief in the target species [26]. Our findings on changes in IRT, time and frequency domain HRV responses, and increase in cortisol concentrations found during the castration procedure in this study support the acute pain response associated with surgical castration without analgesia and could be critical for future work developing and validating effective pain mitigation drugs for castration in calves.Cortisol is widely used to quantify acute distress associated with nociception in calves because the response magnitude (Cmax), the duration of response, and/or the integrated response (AUC) reportedly correspond with the predicted noxiousness of the animal husbandry procedure [10,11,27,28]. Current recommendations regarding the optimal age and weight at the time of castration in cattle are largely predicated on studies measuring plasma cortisol response [29]. It should be noted that cortisol concentrations may also be increased in response to the stress of handling alone [22] and therefore cortisol response to painful procedures should be considered in combination with other pain indicators. In the current study, a SHAM procedure was utilized in order to discriminate between a change in cortisol due to castration-inflicted pain from that of an overall stress response associated with handling. This design was particularly helpful in that it utilized an animal-specific benchmark to assess differences in the cortisol response to an actual painful procedure vs. simply a reaction to a potentially stressful experience with increased handling and processing.Stafford and Mellor [30] reported peak cortisol concentrations within 30–40 min post-surgical castration. In the current study, Tmax was increased following CAST (26 min ± 1.5) compared to SHAM (20 min ± 1.5) for all age categories. The establishment of adrenocortical function occurs approximately when calves reach their 2nd week of life [31]. Knowles et al. [32] reported that plasma cortisol concentrations were very high at birth, but that concentrations decreased to levels expected in adult cattle by 27 d of age. On the contrary in the present study, 6W calves showed lower cortisol concentrations compared to older calves, both at SHAM and CAST. These results are in agreement with Ting et al. [33] and might be related to the potential developmental stage of the adrenocortical function in younger calves, as reported for preterm and low-weight infants [34,35]. These infants showed a decreased baseline cortisol concentration that has been hypothesized to be related to an inability of the hypothalamus to “recognize” stress or a failure to secrete corticotropin-releasing hormone in stressful situations [34]. During the SHAM treatment, cortisol concentrations decreased from peak to pre-procedure levels faster (50 min post-treatment) in 6W than for older calves (120 min post-treatment). During CAST, cortisol concentrations of 6W calves decreased from peak to pre-procedure concentrations within 120 min post-treatment, as reported for the 3M calves, but still faster compared to 6M calves (240 min post-treatment). However, no evidence for age differences was noted on AUC and Tmax.Heart rate in cattle has been shown to increase during stressful events [36,37] and therefore it is used as a measure of relative stress and as an indicator of the intensity of a management procedure. It is possible to assess variation in the intervals separating consecutive heart beats using HRV measurements [15,38] and consequently obtain information on the relative involvement of the sympathetic and parasympathetic systems in the autonomic modulation of the cardiac response in livestock. Physiologic reactivity to painful stimuli is associated with intrinsic adjustments of the sympathetic and parasympathetic divisions of the ANS [39]. Studies have shown that the rise in HR as a result of pain is mostly related to an increase in sympathetic activity [40]. However, the activation of the sympathetic system may be followed by a rise in parasympathetic activity to play an antagonistic role and reestablish the homeostatic balance.Results from the present study show that HR decreased at CAST compared to SHAM, suggesting an increase in the vagal activity. It is widely accepted that the high-frequency band (HFP; 0.15 to 0.5 Hz) of the HRV represents vagal activity, while the low-frequency component (LFP; 0.04 to 0.15 Hz) is thought to reflect both sympathetic and vagal influences. Additionally, the RMSSD is the primary time domain measure of the HRV used to estimate the high-frequency beat-to-beat variations that represent vagal regulatory activity [24]. Furthermore, at acute recovery times during the CAST treatment, HR was significantly decreased compared to SHAM; also, HR decreased in CAST between early and later recovery times. This rapid decrease in HR may be due to an adaptive response of the parasympathetic nervous system (vagal nerve) to reduce the HR. Heart rate has been used in previous studies to determine responses to castration both in studies with or without pain mitigation [14,16,41,42,43]. Decreases in HR have been reported following surgical castration of calves [41] and ring castration of sheep [36]. Schwartzkopf-Genswein et al. [41] reported significantly lower HR at 15 and 30 min after castration compared with pre-castration rates that mirror the finding from the present study. It should be noted that as with cortisol concentration, heart rate alone may not be a reliable measure of pain response as it is a non-specific measure, responding to both positive and negative events similarly. Once again, the SHAM treatment was included in order to assess the impacts of handling alone, helping to identify the changes in response variables associated with a pain-specific response during CAST.Additionally, in the current study, mean HR was greater in younger animals compared to older animals, suggesting an increased sympathetic tone or, more likely, a physiologically higher heart rate than older calves due to increased metabolic rate [44]. Additionally, RMSSD was increased in younger calves between R05 and R1020, reflecting a shift toward more parasympathetic (vagal) dominance. The same increase was noted on HFP in 6W calves compared to older calves both in CAST and SHAM treatment, while 3M calves had an increased HFP at CAST compared to SHAM. Lastly, 6M calves have higher LFP/HFP ratio compared to younger calves. This is to be expected because normalized HFP is a contributing factor to LFP/HFP ratio. Stewart et al. [16] observed a significant increase in high-frequency power from baseline in calves castrated surgically without local anesthesia, and these data are in agreement with the results from the present study. This finding seems to support that an increase in RMSSD and HFP may imply an increase in parasympathetic activity probably associated with deep visceral pain as might occur when the spermatic cords are torn [45], as the parasympathetic nervous system acts to lower HR and carry noxious impulses from the pelvic viscera, including the testes. Von Borell et al. [15] provide a review of how heart rate variability can be used to explore autonomic nervous system function in livestock in response to stressors, such as painful procedures. Moreover, the results from the present study also indicate that calves of 6 weeks of age are capable of perceiving pain, as shown by their physiological response to painful stimuli.Infrared thermography is a noninvasive, contactless technology, which is commonly applied to measure the heat emitted from a surface and to display the temperature distribution as an image. Specific applications for IRT in the dairy and beef industries have been described, including an automated, non-invasive system for early diagnosis of infection in cattle [18]. Gloster et al. [46] concluded that eye temperature is a useful indicator of core body temperature and that it was not affected by ambient temperature. Furthermore, the role of the autonomic nervous system in controlling the eye temperature was confirmed by a drop in eye temperature that occurred following an infusion of epinephrine [47]. Epinephrine release associated with castration causes changes in sympathetic tone, so that the adrenergic effects on cutaneous blood flow are expected to induce modifications in skin temperature that can be quantified with a thermography camera. The results from the present study show a significant decrease in average eye temperature immediately after procedure both during SHAM and CAST, although the greatest magnitude of decrease was reported at CAST. A decrease in eye temperature observed following castration of calves without local anesthetic has been attributed to sympathetically mediated alterations in blood flow in capillary beds [16]. Conversely, the anticipation of exposure to various stressors may result in a rise in core body temperature, with concomitantly reduced peripheral temperature due to peripheral vasoconstriction, termed emotional fever or stress-induced hyperthermia [48]. Interestingly, the minimal eye temperature was greater at CAST compared to SHAM. When the activation of the parasympathetic nervous system occurs, it can be expected to lower cardiac output and blood pressure, resulting in vasodilation, and an increase in eye temperature. However, the detailed mechanism for the increase that was noted in the eye temperature is still unknown.5. ConclusionsIt is evident, as supported by this study and others, that surgical castration causes changes in physiological outcomes in cattle indicating an acute pain response. Concerns for animal welfare are increasing awareness for the need to use pain mitigation during management procedures such as castration. However, analgesia is more often provided to older animals, likely in part due to the notion that younger animals experience pain to a lesser degree than older counterparts. This study adds to the current body of literature that demonstrates younger animals do in fact experience pain in response to castration but that depending on the physiological measure, the magnitude and direction of the response is age-specific and may differ from that of older animals. Cortisol, IRT and HRV have the potential to be used in combination with each other and with other physiological and behavioral parameters to effectively assess pain and thus substantiate the need for pain mitigation during management procedures in cattle of various ages. This information will be critical as the cattle industry continues to focus on developing effective methods of pain mitigation for painful husbandry procedures such as castration.

animals : an open access journal from mdpi

10.3390/ani13050771

PMC10000183

Wild meat is a primary source of protein for many rural populations and its illegal harvest can threaten worldwide wildlife conservation efforts. Near tropical forests, wild meat can be an alternative to domestic meat consumption for different reasons (economy, access, culture, etc.). We aimed to understand drivers of illegal activities affecting wildlife in a hotspot of biodiversity (Sebitoli, Kibale National Park, Uganda) hosting threatened species (including chimpanzees) and bordered by agricultural landscapes with high human densities. Illegal activities were mapped, and individual interviews were conducted. We highlighted that being a young man coming from districts contiguous to Kibale National Park particularly raises the odds of consuming wild meat. This study might help to identify and recommend sustainable social and environmental alternatives to manage biodiversity.

The African tropical forests host an inestimable number of resources, including food, medicine, vegetal and animal species. Among them, chimpanzees are threatened with extinction by human activities affecting their habitats, such as forest product harvesting, and/or more directly, snaring and trafficking. We aimed to better understand the spatial distribution of these illegal activities, and the reasons for setting snares and consuming wild meat in an agricultural landscape (subsistence farming and cash crops) densely populated near a protected area (Sebitoli, Northern part of Kibale National Park, Uganda). To carry out this study, we combined GPS records of illegal activities collected with group counts (in total, n = 339 tea workers, 678 villagers, and 1885 children) and individual interviews (n = 74 tea workers, 42 villagers, and 35 children). A quarter of illegal activities collected (n = 1661) targeted animal resources and about 60% were recorded in specific areas (southwest and northeast) of the Sebitoli chimpanzee home range. Wild meat consumption, which is illegal in Uganda, is a relatively common practice among participants (17.1% to 54.1% of respondents depending on actor types and census methods). However, consumers declared that they eat wild meat unfrequently (0.6 to 2.8 times per year). Being a young man coming from districts contiguous to Kibale National Park particularly raises the odds of consuming wild meat. Such an analysis contributes to the understanding of wild meat hunting among traditional rural and agricultural societies from East Africa.

1. IntroductionBesides high human, developmental, and financial investments in wildlife conservation worldwide [1,2,3] and access limitation to protected areas (PAs) [4], illegal wild meat hunting (here, defined as poaching) continues to threaten terrestrial mammals’ extinction [5]. Wild meat (a wild animal killed for consumption, which is different from bushmeat killed for trade [6]) serves as a vital resource in many rural, lower-income regions of sub-Saharan Africa, and near tropical forests it is often a primary source of protein for rural populations [7]. Its consumption tends to prevail in areas with greater biodiversity indices, which also frequently experience higher poverty and food insecurity [8,9]. Nevertheless, the dependence of households on wild meat is particularly lacking documentation in East Africa [10].Recent increases worldwide in illegal wildlife harvesting are less related to local use and traditions in provenance countries [11] than to an increase in wealth and a decrease in resources in countries importing harvested species or body parts (sent from Africa to Asia; [12]). Therefore, wild meat hunting and poaching (hunting becomes poaching when the practice does not follow the law/rules) can respond to different motivations such as harvesting wild meat for consumption or trade (locally and abroad [13]). To regulate both practices, most countries require hunters to have hunting permits and refer to quotas. Following international guidelines, countries also regulate subsistence hunting through laws that restrict hunting certain species (protected or non-protected), during certain time periods (months, day/night, seasons, etc.), and using specific weapons/tools (riffles, snares, traps, dogs, etc.). However, in practice, it is not this simple: reforming and adapting the regulatory framework of local communities hunting for food is necessary [14], conservation militarization is questionable [15], and development projects aiming to reduce wild meat hunting and/or poaching and maintain food security need a large, adaptable, and clear framework [16,17,18].More research on the prevalence of wild meat consumption and its drivers has been conducted in West Africa and Central Africa than in East Africa. Additionally, across Africa, cultural, sociopsychological, and sociodemographical factors driving wild meat consumers’ behaviors remain understudied [19,20,21]. Some studies show that wildlife illegal harvesting is determined by non-exhaustive ecological, social, and economic drivers such as the frequency of game species, poverty, countries/areas of provenance, ethnic groups, cultural values/beliefs, revenge from crop-raiding animals, lack of access to alternative incomes, heath issues, the distance to local markets, and/or the frequency of patrols [10,22,23,24,25,26,27,28,29,30]. Species abundance and wild meat affordability are often cited as main predictors of harvest levels [31,32], variables such as taste or health issues are often considered secondary factors for wild meat food choices [33], and remoteness or landscape characteristics have been rarely investigated [34]. Indeed, the driving factors of wild meat consumption are complex and variable [16]. In the end, motives to consume legal or illegal wild meat are distinct, mainly relying on case-by-case studies [35], and are more likely to result from a combination of factors, with some being more prevalent than others depending on each case. Consequently, different responses should target different factors [36], which is our aim here. In East Africa, and specifically, in Uganda, wild meat hunting is forbidden within parks and their surrounding areas [37]. The continued existence of a global market for trade in natural resources acquired through illegal means, as well as a lack of data and research for decision and policy makers to implement successful wildlife management and regulation, has been noticed [38]. Between 2005 and 2009, over 71% of households interviewed on their incomes (cash or subsistence) that were located around two protected areas and one unprotected area reported having participated in wild meat hunting [39]. Such practices were related to different reasons in different areas [28,40,41]. Yet, conservation strategies used over the last two to five decades in/around a Ugandan National Park seem to have been effective for protecting the park and animals living within [41].Here, we aim to add a new wild meat consumption case in East Africa to the literature, and identify the drivers of wild meat consumption around a PA in order to design better local alternatives and interventions [16,42,43] to reduce pressure on wildlife in a context where both human density and animal species diversity are high. We focus on the Sebitoli area, located in the extreme north of Kibale National Park (hereafter, Kibale NP), southwestern Uganda. Besides a combination of international and national policies trying to conserve different wildlife forms [37,44], the illegal activity index is high in the area [41]. Chimpanzees (Pan troglodytes schweinfurthii; CITES and IUCN) are endangered. In the Sebitoli community they are indirectly victims of illegal snare injuries [45]. Indeed, snares are likely directed at small game meat (duikers, bush pigs, etc.) not at primates. However, it threatens primate conservation and a non-negligible part of tourism income [46]. Additionally, over the course of the study a chimpanzee was found dead and smoked in a surrounding household (august 2016). Therefore, to estimate the threat animals face in the Sebitoli area, we aimed to obtain a spatial understanding of illegal activities targeting wildlife (objective 1), to identify the targeted species and evaluate the frequency of wild meat consumption (objective 2), and to discuss, in general, drivers of domestic and wild meat consumption in the area (objective 3) with six-year-old data. This study will contribute to the PA’s management plan, help identify and compare wild meat consumption drivers with those identified in other studies, and add to the understanding of this practice in Uganda in particular, and in East Africa in general.2. Materials and Methods2.1. Study Area and InhabitantsKibale NP (795 km2) was established in 1993 and is currently under Uganda Wildlife Authority (UWA) management. It covers three districts: Kyenjojo, Kabarole, and Kamwenge. Local communities can access the park by request to the UWA through resource use agreements [47]. The area of Sebitoli (25 km2) is densely populated with humans and wildlife [48,49,50]. Landscapes are anthropogenic, combining tea, plantain, and eucalyptus plantations; small-scale food gardens; a busy tarmac road [51]; and a high human population density that tripled between 1959 and 1990 and is currently as high as 293 inhabitants/km2 in subcounties including Kibale NP [52]. As wild animals from the park (especially elephants, baboons, and chimpanzees) raid crops in local communities’ gardens (maize, cassava, bananas, etc.), human–wildlife conflicts around Kibale NP are recurrent [53,54].Batooro and Bakiga are the main tribes in the area [48], and Batooro are particularly represented in the north of the park [52]. Batooro are rare meat eaters and their feeding taboos regarding meat consumption can favor the conservation of some wild species, including primates [55]. Additionally, there is not a large wild meat market around Kibale NP: most demands are minimal because most illegally harvested meat (bush pigs and small antelopes) are consumed locally and are not supplied to large external markets [56].Outside of the park, inhabitants practice subsistence farming, and some families work in the large tea plantations [50]. Migrant workers from other districts (not contiguous to Kibale NP) or other countries such as Rwanda can comprise up to 40–60% of the tea workforce, depending on the size of the tea concessions [57]. Tea plantations cover a particularly large surface outside of the park and play a buffering role between the park’s edge and palatable crops (maize, potatoes, millet, etc.). However, few studies have focused on tea workers’ way of life [58].2.2. Snare Removal and Illegal Activity Patrols in the Sebitoli AreaThe Sebitoli area benefits from the presence of UWA rangers patrolling the park to monitor illegal activities and interacting with the local communities neighboring the park about wildlife conservation. It also has benefitted from the presence of the Sebitoli Chimpanzee Project (SCP) since 2008.In addition to UWA patrols, research teams studying chimpanzees in Kibale NP also run snare removal projects (Kanyawara Snare Removal Project; Ngogo Chimpanzee Project; and SCP). At SCP, three local community members were recruited for this task beginning on May 29th, 2015 (study period: 29 May 2015–30 November 2016, 271 patrol days). Using transects and their knowledge of the park, they patrolled the Sebitoli chimpanzee home range 5/7 days for 6–8 h/day using GPS to record illegal activity locations where the targeting of natural resources occurred (snaring, tree cutting, charcoal burning, etc.) and datasheets to record illegal activity characteristics (date, hour, GPS location, type and oldness of activity, amount of evidence, etc.). They disactivated and confiscated any evidence (snares, spears, cables, etc.) that they brought back to the Sebitoli research site, where they were stored. This spatial information was later mapped by SB.2.3. Individual Interview and Group CountWe aimed to compare the legal and illegal animal protein access of tea workers and villagers since both populations live and work within a close distance of the chimpanzee home range and seem to experience sociodemographic, economic, and cultural differences [58]. Children were also included in our sampling, as they represent an important proportion of the local populations [59] and are sensitive to environmental degradation, such as species loss [60].Group counts and individual interviews were carried by the first author (SB) and a local translator. The two methods were designed to be complementary to estimate wild meat consumption in the area. In group counts, votes (via boxes and hands) were opportunistically set up before other SCP activities and described wild meat consumption in a quantitative way from a large sample of respondents (no sociodemographic characteristics identified outside of the actor type—villager, tea worker, and children). In semi-structured interviews, wild meat consumption was qualified with more details and time from a smaller sample of respondents (including sociodemographic characteristics and the discourse of participants).2.3.1. Individual InterviewsInterview ProcessTea workers, villagers, and children had the context of the survey verbally explained to them (e.g., a postdoctoral research study on domestic and wild meat consumption, and relationships with wildlife), the anonymity of their name and the possibility to withdraw from the survey at any time were guaranteed, and both SB and interviewees signed a consent form (Figure A1) after each participant agreed to participate in the survey. A 30 min semi-structured interview was then administered by SB and a translator (n total interviewees = 151; see Figure A2 for detailed questions). Questions focused on: (1) domestic animal protein consumption (frequency, buying location, and transport means); (2) wild meat consumption and illegal activities knowledge (frequency of consuming wild meat, and if respondents did not eat bushmeat they were asked the frequency they come across with it, number of poachers known, targeted species, and reasons to eat wild meat); (3) relationship with wildlife (chimpanzee knowledge and crop-raiding levels); and (4) sociodemographic characteristics. Most questions were closed, but questions related to wild meat consumption were more open and allowed for free speech. In this case, responses were later classified by SB into the main categories.Tea WorkersA tea company granted us access to tea worker camps and allowed us to carry out interviews with their employees during working hours between May and July 2016. SB came to an agreement with 8 tea estate managers on when to carry out interviews at tea camps located around the Sebitoli chimpanzee home range (average distance of 309 m, range: 151–576 m from Kibale NP). In the morning, interviewers came to a tea camp, presented the survey to tea workers, and interviewed workers who volunteered one after the other in an isolated place near the tea fields. A total of 74 workers were interviewed (4 to 11 workers interviewed per camp).The number of tea workers working at each camp varied between 62–178 workers (median: 143). Some workers reside in villages, whereas some reside at the tea camps for free, with a proportion varying between estates (18.6–88.7%; median: 49.6%) (Rwenzori Commodities Ltd., Fort Portal, Uganda, unpublished data). At the tea camps, workers benefit from individual bedrooms where they can occasionally host their family and share common cooking fireplaces and sanitary installations with other workers staying at the camp. Additionally, the tea company provides free lunches to tea workers.Villagers and ChildrenFour village chiefs (LC1) and two head teachers (primary schools) located around the Sebitoli chimpanzee home range granted us permission to interview local inhabitants and children, choosing days when they would be available between October and November 2016. The same voluntary selection process was applied as for the tea workers. A total of 42 adults (10–12 per village) living around Kibale NP (average distance of 300 m, range: 2–572 m from the park’s edge) were interviewed at their household in four villages during their daily activities. Additionally, 35 children (15–20 per school) living at the park’s edge were also interviewed in two schools (average distance of 2000 m, range: 1349–2668 m from Kibale NP) during class hours.2.3.2. Group CountsWe took advantage of a chimpanzee awareness presentation on chimpanzee biology and behavior conducted by SCP that is a regular program of the project to ask questions about wild meat consumption. The program was conducted in 11 tea worker camps, 20 villages, and 14 schools located around the Sebitoli chimpanzee home range between June and December 2016. More than 3900 persons attended presentations, mainly in schools (n = 2722 persons, median: 181 pers/school, range: 22–598; 1 nursery school, 8 primary schools, 1 high school, and 4 institutes/vocational schools), villages (n = 739 persons, median: 49 pers/village, range: 15–99) and tea worker camps (n = 456 persons, median: 41 pers/camp, range: 25–64).No specific sampling method was used to select respondents. After setting up an appropriate time with tea managers, LC1s, and head teachers, the aim of the presentation was explained to the people who came, and it was explained that on a voluntary and anonymous basis, the SCP team would like to ask them two questions about wild meat consumption. Informed consent was verbally given by adults and children, and only people who volunteered to participate contributed. A total of 2902 persons (Figure 1) answered: “Do you eat bushmeat?”, and if yes, “Do you eat bushmeat more than once a month?”. Two different methods to collect responses were used for the different age groups. The ballot-box method was used with adults (villagers and tea workers): two boxes were set out of sight and participants dropped a paper (one paper yes, and one paper no) in each of these boxes to answer questions (Figure A3). This method has the advantage of reducing the social desirability bias [61] and the disadvantage of showing a trend without the possibility to link individual behavior and explanatory variables [62]. With children in schools, hand votes were used to respond to the same two questions. This method has the advantage of reducing children’s confusion that can occur with two sets of boxes, but has a social desirability bias due to other pupils’ presence (teachers were asked to leave the classroom) during votes.2.4. AnalysesData acquisition types, methods, sample sizes, research questions, and analyses are presented in Figure 1 and Figure A4 to help synthesize the key information.2.4.1. Geospatial AnalysesIllegal activity locations were recorded by the SCP snare removal team during their patrol. SB took the locations of each tea camp, village, and school and georeferenced them all using GPS coordinates (GPS Garmin 64s; ArcGIS 10.2; geodesic system–WGS 84; cartographic projection–UTM 36 N). To facilitate spatial analysis, the Sebitoli chimpanzee home range was divided in four areas of relative equivalent size (northwest (NW) 7 km2, northeast (NE) 6 km2, southwest (SW) 8 km2, and southeast (SE) 6 km2). Euclidean distances of illegal activities to the edge and to the road were assessed to evaluate which border of the forest was more at risk. The Euclidean distance was also calculated between participants’ residences and the sellers of domestic animal protein by using the centroids of villages/trading centers/camps to evaluate their accessibility.2.4.2. Statistical AnalysesWe used bi-/multivariate analyses and regression models to estimate the relationships between variables depending on key research questions (Figure A4).With the GPS sample (n = 1661), the relationships between types of illegal activities, Euclidean distances to the road and the edge, and the sides of the park (Figure A5) were assessed though a multiple logistic regression. In the group count sample (n = 2902), we used a chi2 test to estimate the relationship between actor type and wild meat consumption (Test 1, Figure A4). In the individual interview sample (n = 151), five distinct analyses were conducted:(i) A chi2 test was used to estimate the relationship between actor type and wild meat consumption (Test 2, Figure A4).(ii) To assess which variables were associated with wild meat consumption, we carried out a multiple generalized linear regression model. Given the distribution of the response variable (frequency of consuming wild meat, i.e., a count variable) and the hierarchical structure of the data (151 individuals nested into 28 residency locations), we estimated the following random intercept multilevel (i.e., mixed) Poisson model:yij~Poisson μij for i=1,…,151andj=1,…,28lnμij=β0+β1xi1+⋯+βkxik+(eij+μ0j) where i denotes the individuals, j is the residency locations, β is the parameters estimated by the maximum likelihood (Laplace approximation), xik is the value of the kth covariate for individual i, and eij and uoj are the two residual components at both the individual and location residency levels. Adding location residency as a random effect allowed us to account for between-location heterogeneity while controlling for within-location spatial dependence. Multicollinearity among regressors was previously verified through VIF values. Covariates with a VIF > 2 were removed, leading to the exclusion of the side of the park and ethnicity variables. The final model includes the five following regressors: sex, age, actor type (children, tea workers, or villagers), provenance (contiguous/non-contiguous to Kibale NP, Kabale, and Rwanda) and number of people in the household.(iii) We used the Wilcoxon test to assess the relationship between wild meat consumption and salary (Test 1, Figure A4), as well as for Euclidean distance to market and actor type (Test 2, Figure A4).(iv) The relationship between the percent of salary respondents spent on food (response variable) and actor type was estimated through a beta regression. Beta regression is a class of model used when the response variable is beta-distributed, which is commonly true when it is between 0 and 1 for such proportions. As other generalized linear models, beta regression relates the mean response to the regressors through a link function.(v) To assess whether the Euclidean distance to the market was equivalent in function of the side of the park respondents reside in, we used a Kruskal–Wallis test.In all the GLM presented in this study, coefficients were exponentiated (i.e., leading to odds ratios (OR)) for an easier interpretation of elasticities. R software [63] with betareg and lme4 packages [64,65] was used to perform statistical analyses.3. Results3.1. Illegal Activity Spatial DistributionBetween May 2015 and November 2016, the SCP georeferenced a total of 1661 illegal activities (9 different types, targeting animal and vegetal species) inside (n = 1436) and outside (n = 255) the park in the Sebitoli area (Figure 2). Overall, illegal activities targeted more vegetal (74.7%) than animal resources (24.7% including 24.6% of snares), and were found closer to the park’s edge (mean: 165.2 m) than the road (mean: 1529.2 m) (Figure A5). Animal resources were more likely to be found in the NE (concentration of 1/3 of snares collected in the area during the study period). On that side, the tea company had eight tea camps/factories at the border of the forest. The logistic regression confirmed that animal resources have fewer chances to be found in the northwestern (OR = 0.28, 95% CI = [0.19–0.41], p < 0.001) and southern (OR = 0.33, 95% CI = [0.23–0.49], p < 0.001) sides of the park compared to the NE side and the area close to the road (OR = 0.83, 95% CI = [0.72–0.96], p < 0.05; Figure A6).3.2. Wild Meat Consumption in Group CountsWild meat was said to be eaten by 27.3% of overall participants according to the group counts (n = 2902, Table 1). It was reported to be more consumed among villagers (47.9%) than tea workers (31%) or children (19.3%), and this difference was significant (χ2 = 209.08, df = 2, p < 0.001). Additionally, the proportion of children eating wild meat more regularly throughout the year was more important (38.9%) compared to villagers (31.2%) or tea workers (26%).3.3. Wild Meat Consumption in Individual Interviews3.3.1. Descriptive StatisticsAccording to individual interviews, meat consumption is generally low in the Sebitoli area. About 60% of respondents do not eat domestic animal protein (fish, chicken, goat, beef, and pork; Figure A7), eat it less than once a month, or eat it once a month when it is accessible (reduced time, and diversity of locations and choices; Figure A7). Additionally, even if about 40% of respondents in individual interviews declare to eat wild meat, it happens rarely: 2.8 times/year for tea workers, 1.3 times/year for villagers, and 0.6 times/year for children (Table 2).Individual interview results differ from group counts. Here, more tea workers declare to eat wild meat (54.1%) compared to villagers (35.7%) and children (17.1%), and this difference is significant (χ2 = 13.975, df = 2, p < 0.001). From the interviews, the main reasons to eat wild meat were that its cheaper in price compared to domestic meat (43.7%), unknown motives (20.5%), culture/tradition (19.9%), taste (9.9%), and other motives (medicine, distance to access domestic meat, and revenge against crop-raiding animals–6%). Participants in interviews mainly consumed bushbuck (24.5%; Tragelaphus scriptus), red duiker (15.9%; Cephalophus harveyi), bush pig (12.6%; Potamochoerus porcus), and edible rat (6.6%; unknown species). Primates were cited by 5.3% of respondents (seven out of eight were children). Additionally, 26.5% of respondents did not know which species were consumed.3.3.2. Results of the Mixed Poisson Model of Wild Meat Consumption FrequencyThe mixed Poisson model (location residencies as a random variable) exhibited four significant variables (Table 3). Among them, the strongest association was sex. On average, males consume 3.08 (95% CI = [1.83–5.19]) times more wild meat than females, whereas all other variables held constant. Respondents coming from the Kabale district consume 0.50 (95% CI = 0.29–0.85) less than those coming from districts contiguous to Kibale NP. Regarding continuous variables, age was negatively associated with wild meat consumption (IRR = 0.96, 95% CI = [0.94–0.99]; meaning that a one-year increase in age was associated with 0.04% less meat consumption), whereas the number of people sustained in the household had a positive effect (IRR = 1.13, 95% CI = [1.03–1.23]). The regression’s interclass correlation coefficient (ICC = 0.43) indicates a strong effect of residency locations in the model, but no clear patterns appear between tea camps and village residencies. In summary, being a young man coming from districts contiguous to Kibale NP particularly raises the odds of consuming wild meat when all other covariates are kept constant and controlled for within-residency location dependence.3.3.3. Complementary Factors from Individual Interviews (Adults Only)This section offers complementary data analyses from variables that were not available for the entire individual interview sample. Indeed, for some questions (salary and market location for domestic meat) only adults were able to answer. Yet, this information is useful when discussing our results.Revenues and Percentage Spent on FoodThe difference between adult tea workers’ (n = 74) and villagers’ (n = 42) revenues were not significant (Wilcoxon test W = 1496, p = 0.75; Figure A4). Villagers tend to spend more money on food than tea workers (97.6% vs. 13.5% had a self-sufficient garden and most tea workers lived in camps without their family where they were offered free lunches) while their main occupation was subsistence agriculture (61.9%), but this relationship was not significant according to the beta regression model (Figure A4).Geographical Distance to Domestic MeatThree market centers, Fort Portal, Kaswa, and Ntoroko, with distances of 16 km, 4.4 km, and 2.5 km from the forest edge, respectively, attracted domestic meat buyers (Figure A8 and Figure A9). However, to access domestic animal protein, adult respondents travelled a longer or shorter distance. Significant differences existed in the distance to access domestic meat yearly (Kruskal–Wallis chi-squared test= 26.434, df = 3, p < 0.001), and this difference was stronger between the NE and SW (respective mean: 25.6 km and 32.8 km, pairwise comparison p < 0.001) than the NE and SE (respective mean: 25.6 km and 20 km, p < 0.01), and the SE and SW (p < 0.05). No significant difference existed between the NW (mean: 24.5 km) and other areas of Sebitoli. Additionally, a significant difference existed between tea workers’ and adult villagers’ Euclidean distance to access/buy domestic meat (Wilcoxon test W = 1191.5, p < 0.05; villager mean: 21.5 km; tea worker mean: 29.2 km).4. DiscussionOver an 18-month study period, 24.6% of overall censused illegal activities consisted of targeting animals (snares), putting them at risk to be trapped, especially in the northeastern and southwestern sides and at the forest edges. Results estimating wild meat consumption among the three groups differed between group count and individual interview methods. Children were always consuming wild meat the least, but different results regarding adult and tea worker consumption occurred depending on the methods used. Results from the individual interview method may be more reliable as more time and attention were taken in collecting information.4.1. Spatial-, Social-, and Species-Related Factors Affecting Wild Meat ConsumptionIn the Sebitoli area, the park is probably accessed from its edges (rather than the road) to extract resources. This is common in such a conservation context [41,66,67], especially for forest products [41,68,69]. Surprisingly, the setting of snares occurred close to the forest edge, where chances of being caught are likely higher. More snares were collected in the NE (Figure 1) where crop-raiding impacts are of a medium level relative to other locations around the park [53], but they were also collected in the SW of the Sebitoli area where group count participants mostly declared to eat wild meat, distances are longer to access domestic meat, and crop-raiding impacts are high [53]. This edge effect supports the belief that large PAs provide a good option to prevent species losses [70,71]. The stability of illegal activity hotspots (e.g., location in previous years) may be a good predictor for illegal activities in PAs [72], which can be helpful to establish de-snaring strategies [73].Our interview methods highlight that between 17.1% and 54.1% of our samples declared to eat wild meat in the Sebitoli area, but more on an occasional (a few times a year) than regular basis. These proportions are lower than the 71% respondents from the three Ugandan sites [39], and are close, on average, to the 31% of households wanting to access wild meat around Kibale NP [53]. Regarding our results, we reviewed common but non-exhaustive motives found in previous research that should be considered if trying to increase the sustainability of wild meat consumption at a study site:Price: In interviews, wild meat appearing to be cheaper than domestic meat (or free if the harvester) is a common pattern among poor African households that can explain its consumption frequency compared to other meats [26,74,75]. As in other studies, our results highlight that respondents’ frequency of eating wild meat increases with the number of people in the household [20,76], as eating wild meat can decrease spending on food [77], and it is an important factor in poverty reduction in rural areas [22,78,79,80].Taste: Species preferred as wild meat, such as bushbuck, red duiker and wild pig, are known to be hunted for wild meat in Kibale [81]. None of them are threatened with extinction [82]. Wild meat is believed to be “sweeter” than livestock meat and several people believe that it contains fewer chemicals than other meat around the park [83]. Across Africa, wild meat is generally preferred to domestic meat, with arguments of higher quality and better taste [26,84]. However, sometimes, the wild meat sold is misrepresented as another species of wild meat [85]. Ungulates constitute more than the majority of all hunted animals in West, Central, and East Africa [27,31,86], and among wild meat, antelopes are often cited as a preferred species [10]. Compared to other meats, ungulates have a superior quantity of meat with less fat [26,87], and a greater amount of edible protein per unit of live weight than domestic animals (Ledger, 1967, cited in [26].) Bush pigs are also advantageous animal protein sources, representing an important quantity of meat with a low level of total fat [88].Remoteness of domestic meat: The northeastern part of the Sebitoli area experiences more snares targeting wildlife and it is also where the distance to access domestic meat is one of the longest. Most studies and field programs on wild meat consumption focus on individual preferences and/or the role of sociodemographic variables in such behavior. Remoteness and, more generally, landscape characteristics, which can limit access to marketed domestic animal protein, are less frequently used to account for wild meat consumption despite it being a relevant factor linked to dependence on hunting for subsistence [34,89,90,91].4.2. Increasing Human Livelihoods to Promote Wildlife ConservationIn the Kabarole district (connected to Kibale NP), animal protein contributes a small percentage of the total protein intake, as families seem to strive to eat and buy fish or red meat once a week and chicken (the most expensive animal protein) is eaten more rarely [92]. The Batooro and Bakiga are mainly subsistence-level agriculturalists [93], and have many taboos regarding meat; it is considered “pure” when it is coming from ruminants (or, generally, animals eating plants), but “impure” and inconsumable when coming from carnivore and omnivore animals [55].Relative to other foods, meat and fish have a high “diet impact ratio’’ (e.g., a high environmental impact per calorie of food supplied [94]) and the amount of meat consumed through the world is not predicted to decrease [95]. Therefore, increasing domestic animal production is not necessarily the most suitable perspective for biodiversity conservation [18,96,97,98]. In Uganda, increasing the supply of domestic animal protein and reducing its price locally could be a viable policy option to reduce wild meat quantity demand and hunting pressure on Kibale NP’s wildlife, as it was also suggested in other locations [18]. As for tea workers specifically, implementing lunches provided by the tea company with meat (for example, once a week) may be an incentive for wild meat consumption as well. Therefore, changes in diet would have to be followed up carefully as some studies suggest that a substitution away from wild meat can mean a significant increase in fish consumption [18,79], which could also cause biodiversity conservation issues.New ways to involve and sustain local communities’ needs while contributing to wildlife conservation are developing. For example, mini/micro-livestock (the rearing of small wild mammals [99]), eating insects [100], or sun-dried meat in times of food shortage [101] can enhance the animal product supply in sub-Saharan Africa. Regarding the taste issues mentioned earlier, meat should be produced locally to be positively perceived [19], and it could be sold in the Sebitoli area through community markets for conservation [102]. Other means, such as reducing the impact of crop raiding with the use of wildlife deterrents to avoid risks of reprisals from agriculturalists [40,57,73,92], incentivizing poachers [103,104], and developing programs about wildlife conservation and/or spiritual associated knowledge may be effective [20,105,106,107,108,109,110] at strengthening and combining the positive support and sustainable attitudes toward wildlife.5. ConclusionsA combination of factors leads to illegal hunting and consumption of wildlife in the Sebitoli area. The driving factors are spatial (proximity to park’s edges) and social (sex, age, provenance, and household size). The wild meat hunting crisis is a fundamentally distressing problem to address because it is intimately tied to human development challenges such as food insecurity, emergent disease risks, and land-use changes [111]. Therefore, efforts to promote biodiversity conservation need to be integrated, conjoint, multilevel, interdisciplinary, progressive, and sustained for the long term.

animals : an open access journal from mdpi

10.3390/ani11061758

PMC8231097

The wild boar population decreased drastically in Eastern Europe after the emergence of a viral disease called African Swine Fever. We studied how the gray wolves’ diet changed in two regions in Belarus during this situation. Wolves mainly hunted wild boar, elk, red deer, roe deer and beaver. The decrease in the wild boar population caused changes in the diet of wolves, but only in Vitebsk region. After the decrease in the wild boar population, wolves in this region hunted wild boar less, but they hunted roe deer and red deer more. The more the wolves consumed wild boar, the less they consumed both deer species (roe deer and red deer). Moreover, the more the wolves consumed elk, the less they consumed beaver. In another region, Grodno, no changes in the wolves’ diet were found.

After the emergence of African swine fever (ASF), the wild boar population numbers fell drastically in Eastern Europe. This situation made it possible to verify the changes in the wolves’ diet that occurred. The material collection was carried out in two regions, Grodno and Vitebsk, in Belarus. In total, 19 species/groups of prey were observed in the gray wolf diet, but the most important were wild boar, elk, red deer, roe deer and beaver. The decrease in the number of wild boar caused changes in the diet of wolves but only in Vitebsk region, where wolves’ diet before the ASF epidemic outbreak consisted mainly of elk and wild boar. After the decrease of wild boar numbers, wolves still mainly hunted elk, but other types of prey included roe deer, red deer and beaver. We found a negative correlation between wild boar and both deer species (roe deer and red deer) in the wolves’ diet. Moreover, the more the wolves consumed elk, the less they consumed beaver. In our opinion, only intensive hunting of wolves by humans can explain the resulting dietary fluctuations between elk and beaver, as well as the fact that wolves did not turn to other food sources.

1. IntroductionGray wolf (Canis lupus) is the most common large carnivore in Europe. The population dynamics of this species can significantly affect the dynamics and distribution of certain groups of ungulates in ecosystems [1,2]. This species is a carnivorous opportunist with a wide choice of prey [3,4,5,6,7]. We can talk, however, about a dietary preference when a predator hunts a given species of prey disproportionately to its abundance in the environment [8]. Former studies on the gray wolf’s diet show a clear preference for some species of prey [9,10]. Moreover, preferences for selected age classes of prey have also been found: young ungulates under the age of one year often fall prey to wolves [11,12,13]. The gray wolf’s main prey species are red deer (Cervus elaphus), elk (Alces alces), wild boar (Sus scrofa) and livestock [7,9,14]; however, depending on the availability of prey, gray wolves sometimes feed on fish and rodents, but large and medium-sized animals usually predominate in their diet.As a consequence of prey density differences, a wolf’s diet can vary significantly between regions and habitat types. Wolves mainly hunt elk and deer in Scandinavia and northern North America [15,16,17,18]; red deer, wild boar, roe deer and beaver in Central Europe [19,20,21]; livestock in Greece [22]; and livestock, wild boar and roe deer in Italy [23,24]. When a given species dominates in the gray wolves’ diet, a decrease in its density significantly changes the composition of the wolves’ diet. Such a relation was found for wild boar, whose fall in numbers caused it to be replaced by medium-sized ungulates of the deer family [25]. This was also confirmed for beaver, for which a decline in the number of ungulates caused an increase in predation on this species [26,27]. Diet fluctuations have been observed even without significant fluctuations in the number of individual ungulate species [28]. Within one region, the proportion of wild boar in the diet can vary significantly [29].Depending on local conditions, wild boar might be second- or third-order prey, and the share of this species in the biomass consumed by wolves can exceed 20% [19,30]. Wild boar may also be the most important prey when a decrease in the numbers of other ungulate species is observed [23]. However, after the emergence of African swine fever (ASF), which is a viral swine disease [31], the wild boar population numbers changed drastically in Eastern Europe. The virus appeared in 2013 in Belarus and caused (mainly as a result of deliberate intensive hunting) a drastic decline of the wild boar population, which is one of the gray wolf’s main prey [32,33]. The number of wild boar fell by 90% over a short period of time [34]. This situation, which is unique to Europe, made it possible to verify changes in the wolves’ diet as a result of a drastic decline in food availability. A recent study in Estonia showed a shift in the wolves’ winter diet to a higher proportion of roe deer and other less typical food sources after the emergence of ASF [35]. We chose two regions in Belarus whose densities of ungulates and species compositions differed [36]. We hypothesized that due to the drastic wild boar population decline, the wolves’ diet would shift to other ungulate species and other less-hunted animals.2. Materials and Methods2.1. Study AreasThe material collection was carried out in two regions, Grodno and Vitebsk, both of which are in Belarus (Figure 1). These regions were chosen because the abundance of wolves’ prey (mainly roe deer and red deer) differs between them [36]. Roe deer density was over four times higher in the Grodno region than in Vitebsk (18.7 and 4.3 ind./10 km2, respectively). Red deer was not noticed in the Vitebsk region, while in the Grodno region, the density was estimated to be 5.1 ind./10 km2. The Vitebsk region is located in north-eastern Belarus and covers two districts: Gorodok and Vitebsk (55°23′ N 30°16′ E). The study area was characterized by vast forests and swamps with a dense network of natural water bodies: rivers, streams and glacial lakes. Forest complexes covered about 72% of the area. Broad-leaved tree species prevailed (52%), and coniferous tree species covered about 44% of the area. The remaining 28% of the region was open areas, mainly agricultural lands (meadows and crop fields). The Grodno region is located in western Belarus and contains Mostovsky and Grodno districts (53°25′ N 24°48′ E). The region was mainly open areas (63%). Forest complexes that were dominated by pine (Pinus sylvestris) (59%) and birch (13%) stands covered about 37% of the region.2.2. Population TrendsBased on official data [34], wild boar population numbers in Belarus underwent a drastic decline between 2013 and 2014 (Table 1). From 2010 to 2013, the population grew and reached 80,000 individuals. In 2014, only 8600 individuals were registered in official data. Subsequently, the population numbers of wild boar constantly decreased to 2400 individuals in 2019. Data for 2020 were not yet available. Wild boar hunting almost doubled between 2010 and 2013, and from 2014, the harvest numbers exceeded the estimated population size.The beaver population trends show that the population density was rather stable (Table 2). In the Vitebsk region, a slight decrease was observed between 2010 and 2019 (from 13.3 to 10.1 ind./10 km2). In the Grodno region, the beaver density was much lower than in Vitebsk region, but only slight fluctuations were observed during the studied period.The official data did not cover ungulates in the study sites or regions, and the population trends were not clearly known. Nevertheless, changes in the ungulate population density could be derived from the census conducted by Yanuta (unpublished data), for which typical winter track counts on transects (34–37 km for each year) were conducted, according to Priklonski [38]. The ungulate population trends could be regarded as stable, except for wild boar (Table 3). In both regions, wild boar density declined after the ASF outbreak (years: 2015–2018). Other ungulate populations were stable, and a slight increase of red deer in Grodno was observed.2.3. Sample Collection, Elaboration and StatisticsWe collected 237 gray wolf fecal samples from both regions between 2010 and 2020: 116 in the Vitebsk region and 121 in the Grodno region (Table 4). We tried to collect a similar number of samples for particular periods of the year. The samples were collected during both vegetation and non-vegetation periods in both regions: April to October (summer) and November to March (winter). The collection of excrement was carried out where wolves had marked their territory, which was revealed in the winter. No more than two feces samples were collected from each site marked by wolves during a given season, to assure collection of samples from various wolf packs. The samples were placed in plastic bags with detailed labels. Subsequently, the samples were immediately analyzed or frozen.The identification of prey species in the fecal samples was performed according to Jędrzejewska and Jędrzejewski [39], Pucek [40] and Debrot et al. [41]. A washed sample was dried, and the mass of the residues was weighed. The relative amount of prey was calculated from the dry mass of the undigested residues in the sample. To estimate the biomass of individual prey, the weight of the residues of prey extracted from the feces was multiplied by digestibility coefficients [42]. Transverse hair microstructure analysis was used to identify taxonomic groups [43].As the percentage of each prey species varied significantly between samples, we calculated the mean percentage of each prey species from all samples from each year in a given region. In further analyses, we included the main prey species: wild boar, red deer, roe deer, elk, beaver and others (as a sum of the percentage of all other prey). To show the changes in the diet composition, we calculated the mean percentage of each prey in the wolves’ diet for the three years before the ASF epidemic outbreak (2010–2013) and for three (for the Vitebsk region, 2017–2020) or two (for the Grodno region, 2017–2019) years after the epidemic outbreak. We compared the wolves’ diet between periods using the Z-test, with which the frequency of each species in the samples was analyzed. Each pray species was compared separately for the given region. We also built 12 linear regression models (6 for each region); for each model, the dependent variable was the percentage of each prey, and the explanatory variable was the year. The models were built to show the fluctuations in the percentage of each prey in the wolves’ diet for each year. We also ran a Pearson’s correlation matrix, which tested the relations among the percentage of prey species in the wolves’ diet. All statistics were calculated in SPSS software.3. ResultsIn the wolf diet in the two regions, we observed 19 species/groups of prey in total. Apart from wild boar, elk, red deer and beaver, other less important species/groups were observed: red fox (Vulpes vulpes), muskrat (Ondatra zibethicus), white hare (Lepus timidus), brown hare (Lepus europaeus), rodents, birds, domestic dogs and livestock (cows and sheep).In Grodno, before the ASF epidemic outbreak, the main prey species were red deer, roe deer and beaver (27.0%, 22.9% and 19.4%, respectively). Wild boar and elk were less hunted by wolves (10.3% and 11.9%, respectively). This distribution did not change much after the ASF epidemic outbreak. Red deer, roe deer and beaver were still the main prey species (32.7%, 18.8% and 20.9%, respectively), but elk was more hunted (17.4%), while the share of wild boar in the wolves’ diet decreased slightly (7.4%) (Figure 2). The proportion of each species in the diet did not differ statistically between the periods (p > 0.05).In the Vitebsk region, the wolves’ diet before the ASF epidemic outbreak was dependent mainly on elk and wild boar (35.6% and 28.4%, respectively). Other species were less hunted, mainly beaver (17.4%) and roe deer (11.6%), and red deer was killed only extremely rarely (1.1%). After the ASF epidemic outbreak, the diet composition changed significantly. Wolves still mainly hunted elk (34.7%), but other prey species were roe deer, beaver and red deer (24.2%, 15.4% and 14.2%, respectively). The proportion of wild boar in the diet dropped to 5.9% (Figure 2). The Z-test showed a statistically significant lower proportion of wild boar after the ASF epidemic outbreak (p < 0.05). Roe deer and red deer proportions were statistically higher in the period after the ASF epidemic outbreak (p < 0.05).In the Grodno region, no change was found in any prey species in the wolves’ diet over time. All the regression coefficients were statistically non-significant (p > 0.05). In the Vitebsk region, we found three statistically significant relations with year (Figure 3). The share of wild boar in the wolves’ diet decreased over time (F = 13.78, p = 0.005), which explained over 60% percent of the variance of wild boar in the gray wolves’ diet (R2 = 0.605). A percentage increase of two other species in the wolves’ diet was found over time for which the explanatory power was higher: red deer (F = 7.82, p = 0.021, R2 = 0.465) and roe deer (F = 8.96, p = 0.015, R2 = 0.499). No other changes over time were found.We found no correlation between prey species in the wolves’ diet in the Grodno region. In the Vitebsk region, we found a strong correlation between the percentage in the wolves’ diet of wild boar and red deer (r = −0.767, p = 0.006), and between wild boar and roe deer (r = −0.714, p = 0.014). Elk did not correlate with wild boar (p > 0.05), but it strongly correlated with beaver (r = −0.907, p = 0.000). All relations were negative (Figure 4).4. DiscussionAccording to our hypothesis, along with the decline of the wild boar population, the wolves increasingly hunted other ungulate species, such as roe and red deer. This was observed in the Vitebsk region, where these species were much less important in the wolves’ diet before the ASF epidemic outbreak. The strong correlation between wild boar and roe and red deer confirmed this finding (Figure 4). This result was in line with other studies that reported a change in the proportion of ungulates (roe deer, red deer and wild boar) in the diet of the gray wolf. In relatively natural ecosystems with an abundance of ungulates of the deer family, the share of wild boar in the wolf’s diet is usually low [9,14]. Ansorge et al. [44] observed that when roe deer and red deer numbers increased, wolves usually hunted them more than wild boar. However, when the numbers of ungulate prey species fell, wolves instead hunted medium-sized ungulates, mainly wild boar [23,25]. Thus, in some regions, the wolves’ diet may be based on wild boar [45,46,47], and this species’ offspring may even seasonally dominate in the diet, despite the high density of other ungulate species [48]. In contrast to the results in the Vitebsk region, the wolves did not show any response to the decline in the wild boar population in the Grodno region. This was an effect of the low share of this species in the wolves’ diet before the ASF epidemic outbreak (only 10.3%). The observed changes in the wolves’ diet were usually related to the decline of prey species that are an essential group exploited by wolves.Contrary to our hypothesis, the decline in the wild boar population did not cause an increase of other prey in the wolves’ diet, except for ungulates. Wolves did not hunt domestic animals more often, nor did they turn to any other food source. This result was surprising, as some authors have indicated that the most important buffer victims that compensate for a deficit of ungulates are medium-sized wild mammals (mainly hares, but also raccoons, dogs and beavers) and domestic animals (mainly cattle and dogs) [47,49,50,51,52]. Similar trends were observed during monitoring of the diets of wolves in Belarus in the 1990s. When a large number of ungulates were observed (elk, roe deer, wild boar), they constituted 88% of the food biomass in the gray wolves’ diet, while domestic animals constituted only 4%. A rapid decline in the ungulate population resulted in significant changes in the diets of wolves. Large ungulates fell to 32% of the biomass in the diet, and domestic animals accounted for 38% of the dietary biomass [52]. Moreover, the wolves ate roe and red deer, whose density in the Vitebsk region was low (3.8 and 0.0 ind./10 km2, respectively) after the ASF epidemic outbreak (Table 3). In our opinion, this fact should be considered in conjunction with another observed phenomenon, namely the relation between the share of beaver and elk in the wolf’s diet.Another phenomenon observed in this study was the relation between the proportion of elk and beaver in the wolves’ diet in the Vitebsk region. The greater the share of elk, the smaller the proportion of beaver. Our results were consistent with previous studies in the Vitebsk region, where wolves increasingly hunted beavers after a decrease in the numbers of elk [52]. Wolves mainly hunt beavers in spring, after the ice cover has subsided, and in autumn. In these two periods, beavers spend more time on land and are therefore more vulnerable to predation [53]. Wolves’ interest in beavers decreases in summer, when predators mainly hunt young ungulates [54]. However, some studies have indicated all-year-round consumption of beaver [18,19,55]; this was explained by mild winters, when a lack of hard ice cover makes beavers more vulnerable to predation. Another factor that makes beavers more vulnerable to wolves is lower water levels (e.g., due to drought or low rainfall). In such conditions, beavers spend more time on land or in shallow water, thus making them more exposed to wolves [19]. In our opinion, however, these weather-related reasons are unlikely to have affected our results, as a similar phenomenon was not noticed in the Grodno region. Both study areas were relatively close to each other, which suggests similar conditions in winter and similar rainfall in the individual years of the study. Thus, why was the beaver’s proportion in the wolves’ diet not related to the proportion of elk in Grodno?The share of beavers in the wolves’ diet was similar in both regions, with slightly higher values in the Grodno region than in the Vitebsk region (Figure 2). These regions, however, significantly differed in terms of the proportion of elk in the wolves’ diet: in Vitebsk, elk was the most important prey, but in Grodno, it was much less important (Figure 2). In our opinion, these differences could explain the fluctuating proportions of beaver and elk in the wolf’s diet. Wolves prefer to hunt elk calves [16]. Female moose with calves are less mobile, which may make them easier to hunt [56]. Elk, however, show various behavioral adaptations to the presence of wolves, including increased vigilance and aggressive behavior towards predators [15,57,58]. Thus, the effectiveness with which wolves hunt elk may depend on the size of the wolf pack [59]. Beavers can be an important part of the wolf’s diet [5], although consumption of this species may be related to its density. Romański [60] and Moayeri [61] indicated that beaver hunting may also result from the specialization of an individual or family group. Beavers can be attractive prey for lone wolves or small packs, because they are an easier food source compared to ungulates [62,63]. Taking the above into account, we speculate that wolf pack structures have been affected by the results of hunting. Smaller groups or a larger proportion of smaller packs has probably resulted in an increased interest in beavers. We did not possess data on the intensity of wolf hunting by humans; however, we believe that wolf hunting has increased after the wild boar population decline. We speculate that wolves hunted deer more intensively after the ASF epidemic outbreak, and this probably encouraged hunters to exploit wolves more in the Vitebsk region. This speculation could be confirmed by the lack of an increase in livestock depredation in the Vitebsk region after the ASF epidemic outbreak. Such a phenomenon should theoretically occur when a decrease of one of the gray wolf’s main prey is observed [7]. However, less predators effectively means less depredation of farm animals [64]. Increased wolf hunting could also explain the lack of other prey in the wolves’ diet after the ASF epidemic.5. ConclusionsOur study has shown that the drastic decline in the wild boar population after the ASF outbreak triggered a significant change in the diet of the gray wolf. However, this effect was dependent on the importance of wild boar in the wolves’ diet. When this species constituted 10% of their diet, the decrease in its number in the environment did not make any difference for wolves. Wild boar in the diet was replaced by roe and red deer, while the secondary effect of the ASF epidemic outbreak was a fluctuation of elk and beavers in the wolves’ diet. In our opinion, only intensive killing of wolves by humans could explain the resulting dietary fluctuations (between elk and beavers) and the lack of an increase in other prey (including livestock) in the wolves’ diet in the Vitebsk region.

animals : an open access journal from mdpi

10.3390/ani13101645

PMC10215771

Wenchang chickens are the only chicken breed listed in the ‘animal genetic resources in China (poultry)’ in Hainan Province and are famous for their excellent meat quality. Protection of this genetic resource may ensure poultry husbandry’s sustainable and successful development. For more effective conservation, development, and utilization of this genetic resource, we investigated the diversity, degree of inbreeding, and runs of homozygosity (ROH) patterns for Wenchang chickens using whole-genome sequencing data. Our analysis showed that the genetic diversity of Wenchang chickens was relatively high. Selection signal analysis of Wenchang chickens based on ROH found some candidate genes that were putatively associated with meat quality traits and stress resistance traits, such as disease resistance and heat tolerance.

Wenchang chickens, a native breed in the Hainan province of China, are famous for their meat quality and adaptability to tropical conditions. For effective management and conservation, in the present study, we systematically investigated the characteristics of genetic variations and runs of homozygosity (ROH) along the genome using re-sequenced whole-genome sequencing data from 235 Wenchang chickens. A total of 16,511,769 single nucleotide polymorphisms (SNPs) and 53,506 ROH segments were identified in all individuals, and the ROH of Wenchang chicken were mainly composed of short segments (0–1 megabases (Mb)). On average, 5.664% of the genome was located in ROH segments across the Wenchang chicken samples. According to several parameters, the genetic diversity of the Wenchang chicken was relatively high. The average inbreeding coefficient of Wenchang chickens based on FHOM, FGRM, and FROH was 0.060 ± 0.014, 0.561 ± 0.020, and 0.0566 ± 0.01, respectively. A total of 19 ROH islands containing 393 genes were detected on 9 different autosomes. Some of these genes were putatively associated with growth performance (AMY1a), stress resistance (THEMIS2, PIK3C2B), meat traits (MBTPS1, DLK1, and EPS8L2), and fat deposition (LANCL2, PPARγ). These findings provide a better understanding of the degree of inbreeding in Wenchang chickens and the hereditary basis of the characteristics shaped under selection. These results are valuable for the future breeding, conservation, and utilization of Wenchang and other chicken breeds.

1. IntroductionWenchang chickens are a typical native broiler breed in China, mainly in Hainan province, the southernmost part of the country. Being the descendants of those lines who have lived a long time under environmental conditions with high humidity and temperature has resulted in Wenchang chickens performing well when undergoing heat stress and exposure to zoonotic diseases. In addition, Wenchang chickens are well known for their excellent meat quality and high prolificacy. Due to their superior meat quality, Wenchang chicken possesses a large share of the poultry market in Hainan province, and its production yields are sold in Southeast Asian countries. However, the introduction of commercial breeds put Wenchang chickens in danger due to their subpar performance in terms of growth traits and feed conversion ratios [1]. As a result, the number of Wenchang chickens has been decreasing in the past few decades, which may lead to occurrences of inbreeding. The deficient control of inbreeding may give rise to weaknesses in the genetic variability and genetic diversity of Wenchang chickens, which have a detrimental effect on conserving this valuable genetic resource and the sustainable development of poultry husbandry. In addition, inbreeding may also increase the probability of genetic drift and the frequency of autozygosity for deleterious alleles, consequently reducing performance for individuals in the Wenchang population. Therefore, to better preserve the genetic diversity and utilization of Wenchang chicken, we must find an effective way to characterize and understand inbreeding and autozygosity in this valuable genetic resource.Runs of homozygosity (ROH), the continuous homozygous segments in an individual genome, are common in human and animal populations [2]. ROH segments are identical haplotypes transmitted from parents to offspring, so these segments can be hereditary in a population and provide information about the population’s history and demographic evolution. Long ROH are generally due to recent parental relatedness, while shorter ROH indicate more ancient common ancestors in the pedigree [3]. Thus, detecting ROH can estimate the whole genome’s inbreeding level, which can be used to improve mating systems and minimize inbreeding [4]. The assessment of whole-genome inbreeding based on ROH is widely used and effectively distinguishes between recent and ancient inbreeding. In addition, the inbreeding coefficient (F) estimation method by ROH is suitable for large populations [4].Besides consanguineous mating and population size reduction, selection pressure can also result in long homozygous regions along the genome [5]. Studies have claimed that natural and artificial animal selection has resulted in breeds with extensive phenotype variation [1,6]. Using selection signatures to identify regions in the genome under selective pressure may help us determine harbored genes and variants that modulate important animal phenotypes. Nowadays, more and more researchers perform ROH to reveal genetic mechanisms of important traits, and it has been widely used in a variety of animals, such as cattle [7], pigs [8], and sheep [9]. For example, Li et al. detected ROH in Hu sheep based on sequencing data and identified selected genes within the ROH islands relevant to agricultural economic characteristics [10]. However, these types of studies are seen less commonly in chickens, especially Chinese indigenous chickens such as the Wenchang chicken.Therefore, the present study aimed to estimate the diversity and detect ROH patterns in Wenchang chicken populations, observe the degree of inbreeding in Wenchang chickens, and identify candidate genes related to breed-specific traits of Wenchang chickens from within ROH islands. The results of this research contribute to our understanding of inbreeding in Wenchang chickens and help elucidate how artificial or natural selection affects the distribution of functional variants at the whole-genome level.2. Materials and Methods2.1. Animals and GenotypesIn this study, 235 individual Wenchang chickens were sampled from three local conservation farms in Hainan province, China (Table S1). All selected individuals’ genomes were sequenced using the Illumina Nova Seq platform (Illumina, San Diego, CA, USA) and 150-base pair (bp) paired-end sequencing [11]. For details, see Table S1. Raw data from re-sequencing were filtered using the fastp software (a FASTQ data pre-processing tool) with the default parameters. After filtering, the remaining reads were aligned to the chicken reference genome (bGalGal1.mat.broiler.GRCg7b, https://www.ncbi.nlm.nih.gov/assembly/organism/9031/latest/) (accessed on 1 August 2021) using Burrows–Wheeler Aligner (BWA, version 0.7.17) [12]. GATK4 software (version 4.1.6.0) [13] was used for the single nucleotide polymorphism (SNP) calling of each individual. STITCH [14], a software program that can perform this task using ultra-low coverage data, was used for the imputation of the missing genotypes. In addition, SNPs were filtered using PLINK (v1.9) [15]. SNPs were retained for further analysis using the following criteria: (1) minor allele frequency (MAF) ≥ 0.05; (2) missing rate ≤ 0.1; (3) calling rate ≥ 0.9; (4) the calling quality ≥ 30. After that, the remaining SNPs were used for further downstream analysis.2.2. Genetic Diversity and Linkage Disequilibrium (LD) AnalysisThe estimation of the genetic diversity of Wenchang chickens was performed using some indices, including expected heterozygosity (HE), observed heterozygosity (HO), the percentage of polymorphic loci (PN), the minor allele frequency (MAF), and nucleotide diversity (pi). HE, HO, PN, and MAF were calculated using PLINK (v1.9) [15], and pi was calculated using VCFTOOLS (version 0.1.16) [16]. The squared correlation (r2) between pairwise SNPs served as a measure of the linkage disequilibrium (LD) decay using PopLDdecay with default parameters [17].2.3. Identification of ROHROH on all autosomes of each individual were identified using PLINK (v1.9) [15] with a sliding window. In addition, according to previous studies [18,19], specific criteria were applied based on the following: (1) Each sliding window should contain 50 SNPs across the genome; (2) due to genotyping error, up to five SNPs with missing genotypes and one SNP with a heterozygous genotype were allowed for each ROH; (3) each ROH should have a sequence of more than 50 consecutive SNPs; and (4) only detect segments with ROH length greater than 100 kilobases (kb). ROH extracted from sequence data were further classified into three length categories: short ROH (<1 megabase (Mb)), medium ROH (1–2 Mb and 2–3 Mb), and long ROH (>3 Mb). In addition, to further understand the influence of different parameters, we use the detection method of adding 50 SNPs at a time to verify the influence of different parameters.2.4. Assessment of Inbreeding CoefficientsThree methods (FROH, FHOM, and FGRM) were used to estimate the inbreeding coefficients of Wenchang chicken populations. First, the FROH was calculated according to the method proposed by McQuillan et al. [20], which was defined as the ratio of the total length of ROH to the total length of the genome covered by the analyzed SNPs or sequences. Formula (1) was used as follows:(1)FROH=∑LROH∑Lauto where ∑LROH is the total length of all the ROH for one individual across the genome, and ∑Lauto is the total length of the autosomal genome covered by the analyzed SNPs, which was 900 Mb in our study. This length is consistent with the chromosome length of chickens reported in previous studies [21].Second, FHOM was calculated using PLINK (v1.9) [15] to assess the number of observed and expected autosomal homozygous genotypes for each sample [7]. Finally, the genomic inbreeding coefficient of each individual was evaluated from the genomic relation matrix (FGRM) according to the previous method proposed by VanRaden [22]. FGRM coefficients were estimated using the option “--ibc” from the GCTA software. The genome relation matrix is obtained and used by us to calculate the diagonal of the matrix to calculate each individual’s FGRM value. The formula is as follows:(2)FGRMj=Gjj−1, where FGRMj is the genome inbreeding coefficient of each individual, and Gjj is the diagonal element of the genome relation matrix [21].2.5. Identification of Candidate Genes within ROH IslandsPLINK (v1.9) was first utilized to identify ROH islands with the command “--homozyg” [23]. Secondly, the frequency of each SNP that appeared in an ROH was calculated, and the percentage of SNPs that existed within an ROH was subsequently estimated. Thirdly, the top 1% of SNPs were defined as candidate SNPs, and the genes underwent further identification. Finally, Gene Ontology (GO) terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were analyzed to identify the functions of candidate genes further using DAVID (v6.7, https://david.ncifcrf.gov/) (accessed on 26 December 2022) [24]. Only the p-values with less than 0.05 of these enriched GO terms and KEGG pathways were considered significant and listed in this study.3. Results3.1. SNP IdentificationAfter quality control and filtration, 16,511,769 SNPs were obtained from the whole genomes of 235 Wenchang chickens. To further understand the distribution characteristics of these SNPs, we first categorized them into their functional classes. As a result, 73,238 (28.60%) SNPs were categorized as nonsynonymous, and 182,812 (71.40%) SNPs were categorized as synonymous. A total of 4,857,664 (29.30%) SNPs were obtained in intergenic regions. For the SNPs found in gene regions, a total of 257,930 (1.56%) SNPs were in exon regions, and 694,328 (4.19%) SNPs were in untranslated regions (Figure S1). The R package (v4.1.3) “CMplot” then calculated and visualized the density distributions of the SNPs found in Wenchang chicken on each chromosome. This result is depicted in Figure S2. The distribution of SNP has about one SNP site per 17.47 kb. Chromosomes 1 and 2 displayed the largest number of SNPs, with 3,424,555 and 2,584,382, respectively, while chromosome 6 had the highest densities of SNPs. Since a region with a high density of SNPs was contained in chromosome 6, this suggests that chromosome 6 may be an important target for further research. 3.2. Genetic Diversity and LD AnalysisThe results of the genetic diversity indices are shown in Table 1. HE (0.24) was found to be slightly higher than HO (0.23). The value of PN was 0.83, and the pi value was 0.0043 (Figure 1A). The average value of MAF was 0.17, varying from 0.01 to 0.50. The MAF of more than 10.67% of the SNPs was higher than 0.40, and the MAF of 43.90% of the SNPs was lower than 0.10 (Figure S3). LD analysis was measured with r2 values and could provide further information on the overall diversity level of the Wenchang chicken population. Overall, the LD value decreased with the increased distance between SNPs, and the decay was rapid in Wenchang chicken (Figure 1B). In brief, these results indicated that most Wenchang chickens displayed high genetic diversity.3.3. Genomic Distribution of ROHA total of 53,506 ROH segments were identified in 235 Wenchang chickens through ROH detection on all autosomes (Figure 2). The overall mean length of detected ROH was 53.53 ± 9.47 Mb, with a mean number of 227.69 ± 33.65 ROH per animal. The coverage of the identified ROH segments per chromosome ranged from 2.67% to 8.88% in Wenchang chickens. The chromosomes of 33 (61.18%) and 31 (41.15%) had very high ROH coverage. Summary statistics of the numbers of ROH segments across different length classes are shown in Table 2. The average size of each segment was 0.2351 Mb, ranging from 0.10 Mb to 4.74 Mb, and the longest fragment was found in chromosome 2 (which contains 43,925 SNPs). In terms of the identified segments, we can find that the majority (98.17%) of the whole ROH length was made up of the short segments (<1 Mb), while the long ROH segments (>3 Mb) accounted for just 0.03% of the whole ROH length, indicating that ROH covered the highest proportion of the genome (88.53%) in Wenchang chickens. The influence of different detection parameters on the results is shown in Table S4. With the increase of SNP in a single test, the number of short fragments decreased until the number of short fragments (ROH < 1 Mb) accounted for 98.47%, while the number of long fragments was still 0.03%. 3.4. Inbreeding CoefficientsThe results of the inbreeding coefficients calculated by different methods are shown in Table 3. The mean value of FROH in the 235 Wenchang chicken sample population was 0.0566, with a range of 0.0267 to 0.0888, and the value of FHOM ranged from 0.0281 to 0.1527, with a general mean of 0.05614. The values of FGXM ranged from 0.02168 to 0.12298, with a general mean of 0.05999. The inbreeding coefficient values obtained using the three methods were roughly the same and remained low, indicating that the level of inbreeding in the Wenchang chicken population was relatively low. Furthermore, the inbreeding coefficients estimated based on the different physical lengths of the ROH fragments varied greatly, of which, FROH (<1 Mb) was significantly larger than FROH (1–2 Mb), FROH (2–3 Mb), and FROH (>3 Mb). Furthermore, a strong correlation (0.94) was found between FROH (<1 Mb) and FROH (All), while the weakest correlation (0.15) was found between FROH (>3 Mb) and FROH (All) (Figure 3). This result indicated that short ROH fragments might play a major role in calculating FROH.3.5. ROH-Based Selective Signal AnalysisROH hotspots or islands were defined as the genomic regions with the highest frequency of ROH occurrence. In total, 19 ROH islands, 77,975 SNPs, and 393 genes under selection were detected across the 39 autosomes based on their occurrence in the top 1% of the ROH islands as possible regions for candidate genes (Figure 4 and Table S2). The length of these genomic regions ranged from 574.6 bp on Chromosome 2 to 0.9 Mb on Chromosome 1. Notably, the genomic region with a length of 1.7 kb located on Chromosome 8 contained 91 genes, which might be the most relevant region for functional expression in the Wenchang chicken population. The functional annotation of the identified genes was analyzed, and the results are depicted in Figure 5. In total, 25 GO terms and one pathway were significantly enriched (Table S3). GO clustering analysis revealed that the genes that contained SNPs were significantly enriched for metabolic processes (GO:0005975-carbohydrate metabolic process; GO:0006004-fucose metabolic process), activity processes (GO:0004556-alpha-amylase activity; GO:0046922-peptide-O-fucosyltransferase activity), development processes (GO:0035987-endodermal cell differentiation; GO:0098609-cell adhesion), binding processes (GO:0005178-integrin binding; GO:0005518-collagen binding; GO:0000049-tRNA binding; GO:0003725-double-stranded RNA binding); and collagen binding (GO:0005518). Genes containing SNPs were significantly enriched for the KEGG pathway “ECM-receptor interaction (GGA04512)”.4. Discussion4.1. Genetic Diversity of the Wenchang Chicken Sample PopulationNumerous factors affect genetic diversity, and understanding genetic diversity is essential for developing conservation and sustainable livestock management programs. Our study used several indices to estimate the genetic diversity of Wenchang chickens. Heterozygosity, including observed heterozygosity (HO) and expected heterozygosity (HE), one of the most widely used genetic diversity parameters, was first carried out to estimate the Wenchang chicken population. In comparison with other chicken breeds reported in previous research, the heterozygosity (HO) and expected heterozygosity (HE) values of Wenchang chicken estimated in our study are relatively higher when compared to Italian local chicken breeds (HO = 0.1626 ± 0.200) [23], Swedish chicken breeds (HO = 0.225 ± 0.023) [24], Guangzhou chickens (HE = 0.2114), Huiyang chickens (HE = 0.2376), and Commercial broilers (HE = 0.2337) [25]. In addition, many studies have also extensively used the MAF index to evaluate genetic diversity [26]. The distribution of MAF may provide valuable insights into the distinction between common and rare variants in the population. A higher proportion of one population’s low MAF values may indicate high genetic diversity [27]. In our study, the average value of MAF in Wenchang chickens was 0.17, and the proportion of MAF values less than 0.1 was 43.90%. These results indicated that the genetic diversity level of Wenchang chickens was high, consistent with heterozygosity. Likewise, the pi value in Wenchang chickens was higher than other chicken breeds reported in previous studies, such as Guangzhou chickens (0.00199), Beijing chickens (0.00216) [26], and Dongzhongai chickens (0.00332) [28]. The genome’s high pi value also provides evidence for the fast-decaying LD of Wenchang chickens.LD, an important genetic phenomenon, is a non-random association of alleles at different loci. The pattern of LD decay between genetic markers can provide valuable views on a population’s history and evolution. Compared to other chicken breeds, the LD extension of the Wenchang chicken population was much smaller than commercial chicken breeds, such as White Leghorn [28], and was at a low level among some Chinese indigenous chicken breeds, such as Xichuan black bone chicken [29], Xianju chickens, and Taihe chickens [26]. The degree of LD decay rapidity indicates high genetic diversity and is associated with processes such as migration, selection, and genetic drift in one population. Commercial chicken breeds have undergone strong artificial selection for economic traits, which may result in a low LD decay rate and diversity decline. Altogether, Wenchang chickens displayed a high level of genetic diversity, which indicated that the protection strategy for Wenchang chickens used on conservation farms was effective. However, a variety of dynamic factors influence genetic diversity. Therefore, genetic variation and structure should be continuously monitored in the future to prevent the rapid decline of diversity, which is significant for the sustainable development of the poultry industry. 4.2. Characteristics of the Identified ROHAnalysis of the distribution of ROH across the genome and the number of ROH classified by their physical lengths can provide valuable information on the genetic history and inbreeding of populations [30]. Due to the occurrence of recombination events, the length of the ROH decreased over time. Hence, longer ROH segments indicate that inbreeding events were recent, while shorter ROH segments were remote [31,32]. In our study, the distribution of ROH segments across the genome of the Wenchang chicken sample population was mostly short segments, while the percentage of long segments was much lower, especially longer than 3 Mb. This finding suggested that ancient and contemporary inbreeding events might impact the Wenchang chicken population, but this population had a low level of inbreeding, and ancient ancestors were the main inbreeding event-affected group [30].Additionally, some researchers suggested that, in comparison to SNP chips, the higher resolution of whole genome sequence data may lead to the identification of ROH shorter than 1 Mb [33]. In our research, the length of identified ROH segments shorter than 1 Mb was predominant, consistent with previous research on the analysis of other chicken breeds based on whole-genome sequencing data [3]. These short ROH segments may reflect ancestral relationships and more ancient inbreeding events in the Wenchang chicken sample population. In our previous research, the Wenchang chicken sample population displayed a comparatively higher genetic diversity than other chicken breeds; the results of the distribution of ROH segments reconfirmed that finding [11]. At the same time, we verified the influence of different detection parameters on the results. With the increase in the number of single SNPs detected, the number of short fragments detected decreased, but there was almost no effect on the number of long fragments, and the change was not noticeable. Short segments still accounted for the largest proportion of detected SNPs; however, this did not affect the results and conclusions of this experiment.Each individual’s average length of ROH segments identified across the genome in the Wenchang chicken sample population was 53.53 Mb. This result was much smaller than a previous report on a commercial broiler line, in which the length of ROH was 130.9 Mb on average [33]. Commercial breeds have suffered strong artificial selection pressure for genetic improvement on traits of economic interest. Zhang et al. [34] pointed out that the reason for the difference between Chinese indigenous chicken breeds and commercial breeds on the value of FROH might be the difference in selection pressure they have been undergoing. Thus, in comparison with commercial chicken breeds, lower inbreeding events and higher genetic diversity may exist in the Wenchang chicken sample population, consistent with the results of genetic diversity and LD analysis. Moreover, the total length of ROH varied among individuals in the Wenchang chicken sample population, which indicated that ROH differs among individuals, consistent with previous studies [3]. Some scholars have suggested that this difference may be attributed to the lower length threshold used to detect ROH and the higher density of SNPs used to perform analysis [19,34]. 4.3. Inbreeding CoefficientsEstimating inbreeding coefficients based on pedigree data (FPED) has been widely used in previous studies. However, FPED may not be considered an accurate estimation of true inbreeding degree because of many limitations, such as errors in pedigree records and the fact that the coefficient does not reflect the random nature of Mendelian sampling and recombination [35]. Many studies have implemented and attested that the estimation of the inbreeding coefficient based on ROH fragments is feasible without pedigree information [7,36,37]. Estimating the inbreeding coefficient based on ROH does not depend on the allele frequencies or the pedigree’s incompleteness. As a result, FROH is typically less affected by external factors and is more precise in estimating the degree of inbreeding than other methods [36]. The average value of FROH estimated in the Wenchang chicken sample population was 0.0566, and the value of FROH (<1 Mb) was significantly larger than others. A similar result was found in other chicken breeds [3]. FGRM and FHOM were also calculated in our study to verify the correctness of the inbreeding coefficient of the Wenchang chicken sample population. Comparisons with the values among the inbreeding coefficients calculated by these three methods were similar (approximately 0.05), indicating that the inbreeding degree in the Wenchang chicken sample population was again low. Our results also implied that the estimation of FROH based on ROH lengths was reasonably accurate in predicting the genomic inbreeding coefficient, consistent with the conclusion of previous studies [38].4.4. Candidate Genes within ROH IslandsROH islands may represent regions of the genome that have undergone natural or artificial selection. In our study, 19 genome regions with a high frequency of ROH occurrence were identified. After annotation, we found that some GO terms were related to digestion. For instance, GO:0004556-alpha-amylase activity has been demonstrated to be helpful in the digestion of starch in a corn diet. Some researchers added alpha-amylase to broiler diets and found some influence on the growth rate and the development of digestive organs [39,40]. Hence, these GO terms may relate to the growth and development traits of Wenchang chickens. In addition, the GO:0005518-collagen binding was also enriched in our study. Collagen in muscle is associated with the toughness of meat and can affect the maturation of connective tissue and the tenderness of meat [41]. Thus, this GO term may be associated with meat-quality traits. Wenchang chickens are well known for their juicy and tender meat and are well-received by consumers. The significant KEGG pathway enriched in our study is Extracellular matrix-cell interactions -receptor interaction (GGA04512), which has been reported to play a potentially central role throughout the ovulation cycle [42]. These results could putatively explain some reproduction traits of the Wenchang chicken breed.For the candidate genes within ROH islands identified in Wenchang chicken, we found some interesting genes that may be related to economically important traits. Among these genes, the AMY1a gene has been reported to be strongly associated with growth performance, feed intake, and body shape traits in chickens [43]. Likewise, we found that some genes may influence stress resistance in Wenchang chickens. An example of this would be the THEMIS2 gene, which has been previously implicated in disease resistance based on whole-genome sequencing data in chickens [44]. The PIK3C2B gene has been reported to play an important role in the adaptation mechanisms of ducks to heat stress [45]. Wenchang chickens are produced in the Hainan province of China, where the temperature is relatively high throughout the year and has driven the evolution of strong heat tolerance over time in the local livestock and poultry populations. We also identified some genes related to meat traits in our Wenchang chicken sample population. The MBTPS1 gene is associated with meat quality parameters such as shearing force [46]. The DLK1 genes are involved in fat development and differentiation, affecting muscle growth and meat tenderness [47]. EPS8L2, a family of eps8-related proteins, is a new protein family responsible for functional redundancy that leads to actin remodeling in RTK-activated signaling pathways. Related research indicates that EPS8L2 may also play an important role in muscle formation [48]. Fat deposition often has an impact on meat quality and flavor. The LANCL2 gene, which was enriched in the Wenchang chicken sample populations, was reported to be involved in the process of trans-activation of downstream lipogenic genes mediated by PPARγ [49]. Thus, this gene may be linked to fat deposition and indirectly affect the meat quality trait of Wenchang chickens. In brief, the regions identified in this study may help explain the genetic mechanisms underlying the favorable qualities of Wenchang chickens.5. ConclusionsIn summary, in this study, we detected the ROH across the genome of a Wenchang chicken sample population and calculated the inbreeding coefficient to investigate the degree of inbreeding. We also identified the candidate regions within ROH islands that contain genes related to the economically important and identifying characteristics of Wenchang chickens. Our findings demonstrated that historical inbreeding events had little impact on the Wenchang chicken sample population, which displayed a relatively low level of inbreeding. Based on the enrichment analysis of identified candidate regions within ROH islands, we found some genes were related to the economically important traits of Wenchang chicken, such as body shape, meat quality, disease resistance, and heat tolerance. Overall, our research provides evidence for a better understanding of the genetic mechanisms controlling Wenchang chicken characteristics and provides insight into inbreeding events for preservation strategies and utilizing Wenchang chickens in the future.

animals : an open access journal from mdpi

10.3390/ani11102867

PMC8532886

A relationship exists between a female’s early nutritional environment and her ability to produce milk when she lactates as an adult. Colostrum is the first milk available to neonates after birth. We hypothesized that differing levels of colostrum stimulate differences in very early mammary development. Despite differences in weight at 24 h and 7 days, mammary morphological development and DNA content was not found to be different between gilts fed a high versus low dose of colostrum. The rate of mammary gland protein and DNA synthesis over the first week was not different between the groups. Circulating levels of amino acids were determined after 24 h of colostrum feeding, and levels of circulating lysine were found to be related to average daily gain and mammary DNA synthetic rate. Moreover, the level of lysine was related to a lower ratio of DNA to protein synthesis, suggesting that higher lysine favored cell division versus differentiation (by leaving the cell cycle). Further studies are needed in this area.

Perinatal nutrition affects future milk production. The number of mammary epithelial cells affect milk production capacity. Therefore, it was hypothesized that the level of colostrum intake affects the proliferation rate and the total number of mammary epithelial cells in the gland. The ratio of newly synthesized protein to newly synthesized DNA reflects the relative amount of cellular differentiation to cell division. The study objective was to determine the relationship between the level of colostrum intake and 24 h-level of circulating amino acid, glucose and insulin with mammary parenchyma histological features, cell division and protein synthesis over the first week postnatal. One of two standardized doses of a homogenate colostrum sample, 10% (n = 8) and 20% (n = 8) of birth bodyweight, was fed to gilts over the first 24 h postnatal. Gilts were administered deuterium oxide immediately after birth and daily to label newly synthesized DNA and proteins. Gilts were euthanized on postnatal day seven, and DNA and protein were isolated from mammary parenchyma. DNA and protein fractional synthesis (f) and fractional synthetic rate (FSR) were calculated using mass isotopomer distribution analysis. The ratio of protein f and FSR to DNA f and FSR were calculated and used to indicate the relative amounts of differentiation to cell division. Mammary morphological development was also analyzed by measuring the parenchymal epithelial area and the stromal and epithelial proliferation index on postnatal day seven. Colostrum dose was not related to any of the variables used to evaluate mammary development. However, plasma lysine levels at 24 h postnatal were positively related to average daily gain (ADG; r = 0.54, p = 0.05), DNA f (r = 0.57; p = 0.03) and DNA FSR (r = 0.57; p = 0.03) in mammary parenchyma. Plasma lysine was inversely related to the ratio of protein to DNA f and FSR (r = −0.56; p = 0.04). ADG was related to the parenchymal epithelial area and DNA and protein f and FSR (p < 0.05). These relationships support the idea that the nutritional environment affects early mammary development and that higher lysine levels in the perinatal period favored a greater degree of cell division versus differentiation in mammary of neonatal pigs and thus, warrant further investigations.

1. IntroductionThe first days postnatal are a critical period of metabolic-nutritional programming in pigs. Of particular interest to this study is the relationship between early nutritional environment and future lactation performance. In swine, greater colostrum intake by gilts was related to earlier puberty and better lactation performance (as sows) than the low colostrum intake counterparts [1]. Studies in sheep showed the nutrition of ewes during pregnancy affected the yield and composition of milk produced by the offspring [2,3]. The preweaning growth rate of heifers was positively related to their milk production as cows [4]. Heifer dairy calves fed two liters of colostrum produced less milk in their first and second lactations than calves who were fed four liters [5]. The lower milk production was related to greater rates of morbidity and lower body weights of calves. Heifer calves fed restricted versus ad libitum intake of milk replacer had less mammary gland mass, mammary parenchyma, fat pad mass, and lower expansion of epithelium into the adjacent stromal tissue [6].The number of mammary epithelial cells is highly correlated to milk production [7]. Mammary cell number is established during the development of the gland, which begins in utero [8]. The development of mammary glands, from birth to the peripubertal period, in pigs is characterized by ductal elongation and formation of a lumen [9]. The pattern of ductal branching in swine is similar to the human breast, which is characterized by terminal ductal lobular units (TDLU). Prior to the onset of puberty, mammary branching and organization increases in complexity from primarily TDLU-1 to TDLU-2 through the proliferation of lobular buds and ductal elongation.Knowing that nutrition in early life impacts future milk production and that milk production is determined by the number of mammary epithelial cells led us to hypothesize that the level of colostrum intake during the first 24 h postnatal affects the proliferation rate of mammary epithelial cells in the gland, which in the long term would affect the number of milk-producing epithelial cells. Stem and progenitor cells populate tissues by asymmetric cell division [10]. The proliferation and self-renewal of stem-progenitor cells are balanced in the tissue by daughter cells exiting the cell cycle and beginning the differentiation process. As cells differentiate, they become more specialized, and this specialization is marked by protein synthesis [11]. In the gland, at any point in time, there is an array of cellular states as cells progress towards differentiated states [12]. Accounting for these parameters, we posited that the ratio of newly synthesized protein to newly synthesized DNA could be used as an indicator of the relative amount of cellular differentiation to cell division. The overall goal of this study was to test this hypothesis and determine the relationship between the level of colostrum intake and 24 h level of circulating amino acid, glucose and insulin with mammary parenchyma histological features, as well as cell division and protein synthesis over the first postnatal week. For the described studies, one of two standardized doses of a homogenate colostrum sample, 10% and 20% of birth bodyweight, was fed to study animals [13]. Gilts were given a bolus of heavy water (deuterium oxide) immediately after birth and daily for seven days to label newly synthesized DNA and protein over this time period. DNA and protein were isolated from mammary parenchymal tissue. DNA and protein fractional synthesis (f) and fractional synthetic rate (fraction per day, FSR) were calculated using mass isotopomer distribution analysis (MIDA) [14,15,16,17]. The ratio of protein f and FSR to DNA f and FSR were calculated and used to indicate the relative amount of differentiation to cell division. Mammary morphological development was also analyzed by measuring the epithelial area of parenchymal tissue and the stromal and epithelial proliferation index on postnatal day seven.2. Materials and Methods2.1. Animals and Study DesignAnimals and Study Design (Figure 1). Prior to the start of this study, all animal procedures were reviewed and approved by the Institutional Animal Care and Use Committee (Protocol # 1907001920). Animals used for the study were born at the Purdue University Animal Sciences Research and Education Center Swine Farm to eight York (×) Landrace multiparous (3.25 ± 1.67 parities) sows bred to terminal sire Duroc boars. Sixteen animals (n = 16) used for the investigations were part of a larger study [13]. Animals were selected at birth from across eight different litters (n = 2 gilts/litter; birth litter size was 12.3 ± 1.3 live piglets/sow). At birth, gilts were towel-dried, weighed and assigned to one of two study treatments (1 gilt/litter/treatment); colostrum fed at a rate of 24 h intake of 10% of birth body weight (BW; COL10) or 20% of BW (COL20). Gilts that were selected weighed between 1.2 and 1.8 kg and treatments were matched across litters by weight. All gilts were bottle-fed with a pooled colostrum sample every 2 h from birth until 24 h of age. Between feedings, they were returned to a nursery area, which was maintained at 40 °C. Nursery temperature was chosen to prevent any risk of hyperthermia at this age, as this temperature is the same as the offsprings’ in utero environment.A blood sample was collected from each gilt at 24 h postnatal, via jugular venipuncture, using a 22 gauge × 2.5 cm needle into a 2 mL potassium-EDTA coated vacutainer tube (BD367841, BD, Franklin Lakes, NJ, USA). At 24 h, body weight, body temperature and blood samples for plasma preparation were collected from each gilt to measure blood glucose, insulin, total protein and amino acid. Plasma was prepared by immediately centrifuging blood at 2000× g for 15 min (E8 Centrifuge, LW Scientific Inc., Lawrenceville, GA, USA).Gilts were returned to their birth dam to be nursed for the remaining days of the study. Nursing litter size was standardized to 12–14 piglets per sow. Body weights were recorded daily for each gilt. Gilts in the COL10 and COL20 groups were euthanized on day seven postnatal. Animals were euthanized using CO2 inhalation, and then mammary tissue was collected by doing an incision longitudinally along both sides of the left mammary chain. Additionally, six gilts weighing between 1.2 and 1.8 kg were identified at birth, immediately euthanized and used to evaluate mammary morphology at birth.The whole mammary chain was removed by dissecting through the subcutaneous tissue. For histology, a square of 1.5 × 1.5 cm of skin around the nipple was sectioned and placed in 10% buffered formalin. After 24 h, histology samples were transferred to 1X PBS. Mammary tissue used for MIDA analysis of DNA and protein synthesis was collected from the thoracic mammary glands, and mammary parenchyma was grossly dissected to remove muscle, skin and other tissue not associated with gland parenchyma, snap-frozen in liquid nitrogen and stored at −80 °C.A sample size power analysis was performed prior to the start of the study. The power of the study with six animals per treatment with an alpha-error of 0.05, 1.5-fold difference between treatments and 0.25 standard deviation was 0.95. If the difference dropped to 1.4-fold, the power of the study was 0.8 with six animals. Since we anticipated the potential of loss of piglets, the study was begun with eight animals per treatment. Following tissue and plasma collection, all researchers were blinded to treatment during the experimental analysis portion of the study. The treatment groups were revealed for data compilation and statistical analysis on the effect of treatment. 2.2. Colostrum Sample and AnalysisApproximately 50 mL of colostrum was collected from multiple sows (~250) over the course of 7 mo. Colostrum collection was done manually during active farrowing when oxytocin levels are naturally high. Following collection, colostrum was frozen and stored at −80 °C until the day prior to the start of the study. A homogenate-pooled sample was prepared following overnight thawing of colostrum at 4 °C. Piglets were fed this homogenate sample, and several aliquots were collected and stored at −80 °C for subsequent composition analysis. Colostrum composition was analyzed for percent fat, protein, and insulin concentration. Fat percentage was determined using the creamatocrit approach by centrifuging homogenate samples at 12,000× g for 10 min in a non-heparinized hematocrit tube (3 tubes per sample). Fat percentage was calculated as the ratio of the length of fat to total sample length measured with a caliper and then multiplied by 100.The protein content of colostrum samples was measured using the Bradford Assay Kit (Pierce Coomassie Plus Assay Kit, Thermo Fisher Scientific, Waltham, MA, USA). Samples were diluted at 1:100 in phosphate buffer manufacturer’s instructions were followed. A plate spectrophotometer (Sparks 10M multimode microplate reader, Tecan) was used to analyze absorbance at 495 nm wavelength.Colostrum composition was analyzed for percent fat, protein and insulin concentration. Fat percentage was determined using the creamatocrit approach by centrifuging homogenate samples at 12,000× g for 10 min in a non-heparinized hematocrit tube (3 tubes per sample). Fat percentage was calculated as the ratio of the length of fat to total sample length measured with a caliper then multiplied by 100. The protein content of colostrum samples was measured using a Bradford Assay Kit (Pierce Coomassie Plus Assay Kit, Thermo Fisher Scientific; Waltham, MA, USA). Samples were diluted at 1:100 in phosphate buffer, and the manufacturer’s instructions were followed. A plate spectrophotometer (Sparks 10M multimode microplate reader, Tecan Trading AG, Mannedorf, Switzerland) was used to analyze absorbance at 495 nm wavelength. Colostrum insulin was analyzed in duplicate samples using a porcine insulin ELISA kit (cat no. 10-1200-01; Mercodia AB; Winston Salem, NC, USA). Insulin was measured in both homogenate and skimmed colostrum samples. Intraplate variation was 4.75%. 2.3. Neonate Plasma2.3.1. ProteinPlasma protein was measured in duplicate using the Bradford Assay (Bio-Rad Laboratories, Inc., Hercules, CA, USA) following manufacturer instructions. Prior to analysis, plasma was diluted 1:100 with phosphate-buffered saline. Intraplate CV was 3.65%.2.3.2. InsulinPlasma insulin was analyzed in duplicate samples using a porcine insulin ELISA kit (cat no. 10-1200-01; Mercodia AB; Winston Salem, NC, USA), following manufacturer instructions. Intraplate variation was 4.75%.2.3.3. GlucosePlasma glucose was determined using Autokit Glucose (Fujifilm Wako Diagnostics USA Corporation, Mountain View, CA, USA) following manufacturer instructions. Intraplate CV was 4.84%.2.3.4. Free Amino AcidsFree amino acid content of neonate plasma was analyzed using liquid chromatography-tandem mass spectrometry (LC/MS-MS) in Purdue University’s Bindley Biosciences Metabolite Profiling Facility. Briefly, 10 μL of amino-butyric acid at a concentration of 1 μg/uL and 25 μL of 100% trichloroacetic acid (TCA) solution were added to 100 µL of plasma. Samples were incubated for 10 min at 4 °C followed by centrifugation at 14,000× g for 10 min. The supernatant was collected and stored at −20 °C until analysis. Just prior to liquid chromatography, 100 µL of acetonitrile (ACN) was mixed with 100 µL of supernatant. Liquid chromatography was performed using Intrada Amino Acid 3 μm, 2 × 150 mm column (Imtrakt USA, Portland, OR, USA) connected to an Agilent 6470 QQQ LC-MS/MS system (Agilent, Santa Clara, CA, USA). Acetonitrile with 0.3% of formic acid and acetonitrile with 100 mM ammonium formate solution (20:80 v/v) were used as mobile phases.2.4. Histological Analysis of Mammary Gland DevelopmentAll tissue preparations for histological analysis were done by the Purdue University Histology Research Laboratory. Mammary tissues were fixed in 10% neutral buffered formalin for 24 h and transferred to PBS until processing for paraffin embedding. Paraffin processing was done in a Sakura Tissue-Tek VIP6 tissue processor for dehydration through graded ethanols, clearing in xylene and infiltration with Leica Paraplast Plus paraffin. After processing, tissues were embedded in Leica Paraplast Plus paraffin. Tissue sections were taken at a thickness of 4 µm using a Thermo HM355S microtome. Sections were mounted on charged slides and dried for 30–60 min in a 60 °C oven. After drying, all slides were deparaffinized through 3 changes of xylene and rehydrated through graded ethanols to water in a Leica Autostainer XL. For hematoxylin and eosin (H&E) staining of tissues, the Leica Autostainer XL was used. Tissue sections were stained in Gill’s II hematoxylin, blued and counterstained in an eosin/phloxine B mixture. Finally, tissues were dehydrated, cleared in xylene and cover-slipped in a toluene-based mounting media (Leica MM24).H&E-stained tissues were used to measure the proportion of epithelial tissue within the parenchymal compartment. First, ImagePro Plus 5.1 (Media Cybernetics) was used to capture histological images in conjunction with a Nikon Eclipse 50i microscope (Nikon Inc., New York, NY, USA; Evolution MP, Media Cybernetics Inc., Rockville, MD, USA). Multiple images of H&E stained tissue were captured at 10× magnification to encompass the entire parenchymal area of the gland for each animal. The parenchymal area was defined for this study as the epithelial cells of the terminal ductal lobular units (TDLU) and associated ducts along with intralobular and interlobular stroma. To create a panorama of the entire parenchymal area of the cross-section, images were merged into a single image using Adobe Photoshop (V 22.1.0, Adobe). ImageJ was used to measure the area in the tissue section (Figure 2). The “Draw/Merge: Trace” tool was used to first select parenchymal tissue and calculate the area, then to trace and calculate the entire epithelial area of TDLU (epithelium plus lumen) and finally to trace around the lumen and calculate that area. The ratio of epithelium within parenchyma was calculated by subtracting the lumen from the epithelial area of the TDLU and then dividing this by parenchyma area, and this was defined as parenchymal epithelial area (PEA).Tissue sections were also immunostained with KI67 to mark proliferating populations of cells. After deparaffinization, antigen retrieval was done with a TRIS/EDTA pH 9.0 solution in a BioCare decloaking chamber (Pacheco, CA, USA) at a temperature of 95 °C for 20 min. Slides were cooled for 20 min at room temperature and transferred to TRIS buffer with Tween 20 detergent (TBST). The rest of the staining was carried out at room temperature using a BioCare Intellipath stainer. Slides were incubated with 3% hydrogen peroxide in water for 5 min. Slides were rinsed with TBST and incubated in 2.5% normal goat serum for 20 min. Excess reagent was blown off, and Ki67 primary antibody (Cell Marque, 275R-16, Rocklin, CA, USA) was applied at a dilution of 1:100 (0.364ug/mL) for 30 min. The negative control slide was stained with Rabbit IgG (Vector Labs, I-1000, Burlingame, CA, USA) at a concentration of 1:5000 (1 µg/mL) for 30 min. Slides were rinsed twice in TBST, and a goat anti-rabbit secondary antibody (Vector Labs, MP-7451) was applied for 30 min. Slides were rinsed twice in TBST, and Vector ImmPACT DAB (Vector Labs, SK-4105) was applied for 5 min. Slides were rinsed in water and transferred to a Leica Autostainer XL(Wetzlar, Germany)or hematoxylin counterstain, dehydration and cover-slipping. Five images per gilt were taken at 200× magnification. Sections of jejunum tissue were used as a positive control for the specificity of KI67 staining for proliferating populations of cells. To determine the proliferation index of mammary epithelial cells and proliferating intralobular stroma cells in parenchymal tissue, an ImageJ plugin called Cell Count by GNU General Public License was utilized. As above, the parenchymal area was defined for this study as the epithelial cells of the TDLU with ducts and associated intralobular and interlobular stroma. The proliferation index of epithelial and stromal cells within parenchymal tissue was determined. All epithelial cells were positively stained for KI67, the five sections were counted, and the epithelial cells without staining were counted. Similarly, intralobular and interlobular stromal cells that were immunostained for KI67 were counted, and all cells not stained were counted. The total number of each cell type was determined, and then the number of proliferating epithelial or stromal cells was divided by the total of each type to determine the percent of proliferating cells.All research assistants that analyzed histomorphology were blinded to treatment and day and trained by one individual on the approach to conducting analyses. For each animal, three research assistants analyzed histomorphic features, and data across the three researchers were averaged for final counts. Tissue was available for all animals that survived to postnatal day 7 of COL10 (n = 7). However, the quality of tissue collected for one COL20 piglet was not representative of parenchyma, so only six animals in this treatment were used for histological analysis. To determine relative changes in the proliferating index of epithelial and intralobular stromal cells in the parenchyma and parenchymal epithelial area between birth and postnatal day 7, tissue from the baseline group (n = 6) of gilts was also analyzed. 2.5. Mass Isotopomer Distribution Analysis (MIDA) of DNA and Protein Synthesis (f) and Fractional Synthetic Rate2.5.1. Metabolic Labelling with Deuterium Oxide (D2O)In order to obtain good quality data for mass isotopomer distribution analysis for proteome studies, greater than 2.5% steady-state enrichment of heavy water-deuterium oxide (D2O) is recommended [14]. To achieve this, the doses of D2O used were adapted from Lam et al. [17]. Metabolic labeling of newly synthesized deoxyribose and amino acid molecules was begun immediately after birth by administering a deuterium oxide bolus to animals via intraperitoneal injection (IP) of 0.9% NaCl in D2O (Millipore Sigma, Burlington, MA, USA), at 20 mL/kg of BW. A second bolus was given to animals 4 h later, using the same dose and route of administration. At 24 h postnatal, and then daily until postnatal day 7, piglets were orally gavaged with 10 mL/kg of BW deuterium oxide at 0600 h to maintain enrichment. Six unlabeled animals were administered equivalent doses of saline. Tissue and blood of the unlabeled group were used to determine the baseline mass isotopomer distribution. At birth and postnatal days 1, 3, 5 and 7, blood samples were collected by jugular puncture and plasma was isolated for analysis of D2O enrichment.2.5.2. Determination of D2O Enrichment in Plasma Using Gas Chromatography-Tandem Mass Spectrometry (GC-MS/MS)To determine the percent of body water that was D2O, the approach of base-catalyzed exchange of hydrogen (deuterium) between water and acetone was used [18], with modifications by Purdue University’s Bindley Biosciences Metabolite Profiling Facility. Briefly, piglet plasma (20 μL) was mixed with 2 μL of 10 N NaOH and 4 μL of 5% (v/v) acetone in acetonitrile. The sample and standard curve (ranged from 0.5 to 16% of D2O in H2O) mixtures were incubated at room temperature overnight, after which the acetone portion was extracted by the introduction of 500 μL of chloroform and 0.5 g of anhydrous sodium sulfate (Na2SO4). Samples were centrifuged at 14,000× g for 1 min to precipitate the Na2SO4. The chloroform solution was transferred to a vial, and acetone was measured using GC/MS in Purdue University’s Metabolomics Core. Gas chromatography was carried out using an Agilent Select FAME column (CP7419, Agilent, Santa Clara, CA, USA) attached to a TSQ 8000 triple quadrupole GC-MS/MS (Thermo Fisher Scientific, Waltham, MA). The intensity of acetone was measured at 58 and 59 m/z and was used to calculate D2O percentage in plasma.2.5.3. Isolation of DNA from Tissue and DNA HydrolysisThe procedures for DNA isolation and hydrolysis described were modified from the approach described by others [19] in the following way. Approximately 25 mg of mammary tissue from the parenchymal area was used for DNA isolation. Tissue was ground using a tissue homogenizer in DNA extraction buffer from the gMAX Mini Genomic DNA Kit (IBI Scientific, Dubuque, IA, USA). DNA was eluted from the column using 100 µL of TE buffer (tris-HCl + EDTA), and the concentration was measured with a nanodrop system. DNA was diluted with hydrolysis buffer (20 mM Tris-HCl, 100 mM NaCl, 20 mM MgCl2, pH 7.9) to 1 µg in a total volume of 50 µL, and 50 µL of hydrolysis enzyme cocktail [benzonase (E-1014); phosphodiesterase I (P-3243) and alkaline phosphatase (P-7923); MilliporeSigma, Burlington, MA, USA) was added. The hydrolysis reaction was carried at 37 °C for 6 h. Then, samples were dried overnight and stored at −20 °C until analysis.2.5.4. Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) Analysis of Adenosine Isotopomer DistributionD2O labels the deoxyribose moiety of dNTPs in replicating DNA through the de novo nucleotide synthesis pathway. The isotopic enrichment of the purine deoxyribonucleoside adenosine is then determined by LC-MS/MS. Briefly, samples were reconstituted in 100 µL of 5% MeOH/95% 5 mM ammonium formate. Molecule separation was carried out with 5 mM ammonium fumarate and 100% methanol as mobile phases in a Waters Atlantis T3, 3 μm, 2.1 × 50 mm column (186003717, Waters Corp., Milford, MA, USA) connected to an Agilent 6470 QQQ LC-MS/MS system (Agilent, Santa Clara, CA, USA). Multiple reaction monitoring (MRM) of the ribose portion of adenosine (dA) was measured based on the parental and product ions 251 → 117 m/z (M0). Ion combinations for M+1 and M+2 were identified and measured based on the identifications of 252 → 118 m/z and 253 → 119 m/z, respectively.2.5.5. Protein HydrolysisPreparation of protein hydrolysate for measuring global protein synthesis was done as described [15] with some modifications. Briefly, approximately 25 mg of parenchymal mammary tissue were placed in a 5 mL amber glass vial (Fisherbrand, Thermo Fisher Scientific, Waltham, MA, USA), and 1 mL of 6 M HCl was added under the fume hood. Samples were homogenized using the Fisherbrand 150 handheld tissue homogenizer (Thermo Fisher Scientific, Waltham, MA). The probe of the homogenizer was washed with sterile water between samples. Caps were placed in vials and incubated at 120 °C in a forced air oven (Model 414004-576, VWR International, West Chester PA, USA) for 24 h. Following incubation, samples were transferred to a 1.5 mL tube and centrifuged at 14,000× g for 10 min. The supernatant was transferred to a 1.5 mL tube and dried in a savant SPD 2010 speedvac concentrator (Waltham, MA, USA) overnight. The dried samples were stored at −20 °C until amino acid extraction.2.5.6. Amino Acid Extraction LC/MS Analysis of Isotopomer Distribution of AlanineDried protein hydrolysates were reconstituted by adding 300 µL of PBS and vortexing the samples, and 100 µL was transferred to a new 1.5 mL tube. Twenty-five µL of TCA (trichloroacetic acid, saturated solution, 1000 mg of TCA + 700 µL H2O) was added and samples vortexed to mix. Samples were then centrifuged at 14,000× g for 10 min, and 50 µL were transferred to a new tube, being careful to avoid black precipitate. Then 50 µL of acetonitrile was added, and samples were mixed well by vortexing. One hundred µL of this extract was used for LC/MS analysis of alanine.The method used to determine the isotopomers of alanine was developed by Purdue University’s Metabolite Profiling Facility, Bindley Bioscience Center, through modification of the methods used to measure amino acids. In this method, an Intrada Amino Acid column was used for the liquid chromatography (LC), followed by a quadrupole mass spectrometer (MS). Alanine is retained to ~11.5 min of the run, and the mass spectrometry returns a precursor ion of 90 m/z and a product ion of 44 m/z. The fragment of 44 m/z (with chemical formula C2H6N) contains four hydrogens that can potentially be replaced by deuterium during the synthesis process. The precursor (alanine, C3H7NO2) and product (C2H6N) will increase mass equally as deuterium is added to the molecule. For this method, the LC/MS machine and software is programmed to measure the intensity/area of the peaks of molecules with precursor → product ion pair of 90 m/z → 44 m/z; 91 m/z → 45 m/z; 92 m/z → 46 m/z; 93 m/z → 47 m/z; and 94 m/z → 48 m/z; in order to measure the intensity/area of isotopomer (M) with no heavy isotopes (M0), one (M+1), two (M+2), three (M+3) and four (M+4), respectively. Supplementary Materials Information S1 shows the distribution of alanine M0, M+1, etc., in a sample from an unlabeled animal (blank) and eight samples from D2O labeled animals, with corresponding LC/MS spectra of samples.2.5.7. Mass Isotopomer Distribution Analysis (MIDA) of Adenosine and Alanine for Calculation of the Fraction (f) of DNA and Protein Newly Synthesize and the Fractional Synthetic Rate (FSR) of DNA and ProteinTo determine the percent of newly synthesized DNA and proteins, the percent of enrichment (p) was calculated as the mean percentage of D2O enrichment from postnatal day one to day seven in each piglet. LC-MS/MS analysis of adenosine and alanine provided the intensities for M0 to M+2 and M0 to M+4, respectively. The percentage of M0 (%M0) at birth and day seven (%M07) was calculated as the intensity M0 isotopomer divided by the total of the intensities for all isotopomers measured for adenosine and alanine. The difference between %M0 and %M0t is defined as EM0t [14]. EM0t = %M0t − %M0(1)EM0t can be defined as the modifications on %M0 after a specific time (t) of exposure to D2O and is, in part, determined by the synthetic ratio (balance between synthesis and degradation) of DNA, as tissue cells are in a constant turnover of division and death. Each cell has a different turnover based on its biological functions and properties. Moreover, the speed of synthesis of dA, or FSR, can be modified by physiological events or experimental treatments. The %M0* (max %M0 when 100% of the DNA are de novo synthesized) can be calculated based on %M0, p and n. Where p is the probability of being labeled, which is the percent enrichment of D2O, and n is the number of hydrogens in a molecule that can be labeled (‘replace’) hydrogen. The following formulas for DNA synthetic rate were adapted from [16,17]:%M0* = (1 − p)n* %M0(2)After we know the %M0*, we are able to elucidate EM0*, similar to EM0t. EM0* = %M0* − %M0(3)EM0* is the %M0 when 100% of the DNA is turned over.Once EM0t and EM0* are determined, fractional synthesis (f) is calculated. f = EM0t/EM0*(4) FSR = −ln (1 − f)/t(5)2.6. Statistical AnalysisAll statistical analyses were performed in SAS (version 9.4; Cary, NC, USA). Amino acids that were below the detectable limit were assigned a value that was one-fifth the highest value, which was below the lowest value. Procs Mixed model was used with treatment run as the class. Normality was checked by running the residuals in the model. All statistical models included treatment and birthweight. Birthweight was excluded from the model if not significant (p < 0.05). The least-square means were performed using the Tukey-Kramer method, with a pairwise comparison. p < 0.05 was considered significant, with p-value > 0.05, but ≤ 0.1 discussed as a tendency. Bodyweight and average daily gain were run with Procs Mixed with treatment and day as fixed effects with the day as a repeated measurement. Treatment by day interactions and birthweights were included in the model and removed if not significant (p < 0.05). Proc corr was used for correlations.3. Results3.1. Colostrum Composition and the Effect of Colostrum Intake on Neonate’s Plasma Insulin, Glucose, Protein and Amino Acid ConcentrationsThe homogenate colostrum sample that was fed to neonates was 10.1% fat and 9.8% protein, and the insulin concentration was 289 milli-international units per liter (mIU/L) in whole colostrum and 312 mIU/mL in skimmed colostrum. The birth weight of COL10 animals was not different from COL20 (Table 1). The COL20 animals gained significantly more weight over the 24 h colostrum feeding period than COL10 gilts (p = 0.03). From day two postnatal to study completion on postnatal day seven, there was no significant difference in average daily gain (ADG) between COL10 and COL20 animals. The final weight of COL20 was numerically greater than COL10 (Table 1).Plasma insulin concentration was not different between COL10 and COL20 at 24 h after birth (1.84 vs. 1.65 mLU/L; p > 0.05), and glucose concentration at 24 h postnatal was not different between groups (Table 2).There was no difference in total protein concentration between groups; however, the concentration of several amino acids was different between the treatments (Table 3). COL20 piglets had greater concentrations of all branched-chain amino acids (BCAA) (Ile, Leu, Val) than COL10 piglets at 24 h postnatal (p < 0.05; Table 2). COL20 piglets also had greater concentrations of the essential amino acids (EAA) Met and Phe (p < 0.05). While COL20 piglets had a trend to have greater Thr at 24 h of postnatal (p = 0.07), there was no difference between treatments for the EAA of Arg, His, Lys or Trp (p > 0.05). For non-essential amino acids, COL20 piglets had greater concentrations of Asp, Gln and Pro compared to COL10 piglets at 24 h postnatal (p < 0.05). There was no difference in Ala, Asn or Cys between the treatments 24 h postnatal.3.2. Effect of Day and Colostrum Intake on Proliferation Index of Epithelial and Stromal Cells in Parenchyma and Proportion of Epithelial Tissue in Parenchymal Component of the Mammary GlandThere was no difference (p > 0.05) between COL10 and COL20 treatments in the proportion of epithelial tissue in the parenchymal compartment (PEA) of gilt mammary glands, nor was there a difference between baseline levels on day zero with the epithelial area on day seven. Both epithelial and stromal cells were immunostained for KI67 in the mammary parenchymal area, with KI67 positive stromal cells in both the intralobular and interlobular stroma (Figure 3). There was no difference in the proliferation index of epithelial (percent proliferating) or stroma cells in mammary parenchymal tissue between COL10 and COL20 piglets (p > 0.05; Figure 4). The epithelial cell proliferation index was greater in mammary tissue isolated from animals at birth versus the rate of proliferation in mammary tissue of COL10 and COL20 animals, which was isolated on postnatal day seven. Meanwhile, there was no significant difference in stromal cell proliferation rate nor PEA in the tissue between birth and day seven.3.3. Effect of Colostrum on Amount (f) and Rate (FSR) of Mammary DNA and Protein Synthesis over the First Week PostnatalAnalysis of percent deuterium oxide in piglets’ plasma across the seven days of labeling indicated that the dosing regimen achieved the goal of greater than 2.5% D2O. The average enrichment (mean ± standard deviation) of COL10 animals from day one to seven postnatal was 3.48% ± 1.21, and COL20 averaged 3.38% ± 0.82 (Figure 5). The yield of DNA per unit of mammary parenchymal tissue, although numerically greater in COL20 (0.66 ± 0.08 ng DNA/mg tissue) treated gilts, was not different (p = 0.2) from COL10 (0.55 ± 0.21 ng DNA/mg tissue) animals (Table 4).There were no differences between the fraction (f) of newly synthesized DNA over the seven days of labeling in mammary parenchyma between COL10 (29 ± 5%) and COL20 (29 ± 5%) piglets. The fractional synthetic rate (FSR) of DNA was approximately 5%, indicating approximately 5% of the cells were turning over each day in parenchymal tissue (p > 0.05). The f of the new protein in parenchyma over the seven days of labeling and the FSR of protein was not different between the treatments, with the f at 70% in COL10 and 68% in COL20 gilts and FSR at 18% for both groups (Table 4). The ratio of protein f to DNA f reflected the amount of newly synthesized protein synthesized per new DNA complement (i.e., new protein per new cell synthesized) over the seven days and was approximately 3:1 for both treatments. The ratio of protein FSR to DNA FSR was also not different between treatments and was 4.36 ± 0.99 for COL10 and 4.47 ± 0.99 for COL20.The relationship between mammary morphological features and mammary DNA and protein f and FSR were investigated (Table 5). A relatively strong relationship (r = 0.86; p < 0.0001) between percent KI67 labeled stroma, and epithelial cells was evident on postnatal day seven. PEA was related to percent KI67 labeled stroma (r = 0.75; p < 0.001) and KI67 labeled epithelial cells (r = 0.66; p = 0.01). The percent of KI67 labeled epithelial cells on postnatal day seven was positively (p < 0.05) associated with protein f (r = 0.61) and FSR (r = 0.63). There was a trend for a relationship between parenchymal epithelial area and the fraction of newly synthesized DNA (r = 0.48; p = 0.09), as well as a trend for inverse relationships between parenchymal epithelial area and the ratio of protein to DNA f (r = −0.49; p = 0.09) and FSR (r = −0.48; p = 0.1). Consistent with the relationship between f and FSR, these showed strong (p < 0.001) correlations with each other within DNA and protein. The positive relationships between protein and DNA f and FSR were also strong (p < 0.001). There were also strong negative relationships between DNA f and FSR with the ratio of protein/DNA f and FSR. The relationship between protein f and FSR and this ratio was also negative.Correlation analysis was run to determine if growth and morphology variables of all animals were related to markers of perinatal nutritional environment such as birthweight, growth variables (i.e., average daily gain, crown-rump length), plasma glucose, insulin, protein and amino acids at 24 h after birth. No relationship was found between birthweight and the f or FSR of DNA and protein. There was also no relationship between birthweight and histomorphic features of gilt mammary glands on postnatal day seven. However, average daily gain was correlated with PEA (r = 0.54; p = 0.05) and f and FSR of both DNA and protein (p < 0.01; Table 4). There was also a significant relationship between DNA f (r = 0.60; p = 0.02) and FSR (r = 0.61; p = 0.02) and crown-rump length on day seven (Table 5). Ongoing analysis indicated growth of other tissues, including longissimus dorsi muscle from the same animals that mirrored that of the mammary parenchyma.There was no relationship between any mammary variables and plasma levels of glucose, insulin or protein at 24 h postnatal. The relationship between individual plasma amino acid levels and mammary variables was found only for plasma lysine and glutamine, and so only these amino acids are listed in Table 5. Plasma lysine level at 24 h postnatal was positively related to mammary DNA f and FSR (r = 0.57; p = 0.03 and r = 0.57; p = 0.03, respectively, Table 5). There was an inverse relationship between lysine levels and the ratio of protein to DNA f (r = −0.56; p = 0.04) and FSR (r = −0.56; p = 0.04) and lysine levels. Plasma lysine levels were also positively correlated with average daily gain across the seven days (r = 0.54, p = 0.05). Plasma glutamate levels were negatively associated with the parenchymal epithelial area (PEA; r = −0.55, p = 0.05), and there was a tendency (p ≤ 0.1) for a positive relationship between plasma glutamate and the ratio of protein to DNA f (r = 0.47) and FSR (r = 0.48).4. DiscussionThe data collected supports the relationship between factors indicative of perinatal nutritional environment and mammary growth and development over the first week postnatal. In particular, plasma lysine level at 24 h postnatal was positively related to average daily gain, the fraction of newly synthesized DNA (f) in mammary parenchymal tissue over the first seven days postnatal, and the fractional synthetic rate of DNA in mammary parenchyma. Plasma lysine was also inversely related to the ratio of protein to DNA f and FSR. This relationship, as posited in the introduction, may reflect that higher lysine levels favored a greater degree of cell division versus cells leaving the cell cycle and differentiating. The relationships between nutritional environment and mammary development were found despite the fact that colostrum dose was not related to any of the variables used to evaluate mammary development. There may not be an effect of colostrum dose on variables measured. In light of this possibility, it is interesting to note that the amount of DNA isolated per unit of mammary parenchymal tissue was numerically higher in COL20 versus COL10 animals. This finding suggests that the level of colostrum intake may affect the number of cells in parenchyma. Analysis of DNA content at an earlier time point is needed to determine this. Moreover, future studies using tools like single-cell RNA-seq would help in understanding whether the amount of colostrum consumed affects the developmental program of subpopulations of cells within the gland. The lack of an effect may also have been related to the study design. Different doses of colostrum resulted in COL20 animals weighing significantly more after the 24 h of colostrum feeding, and these differences were maintained to postnatal day seven [13]. However, returning piglets to birth litters likely had unmeasured impacts on perinatal nutrition. One piglet in each group died by crushing, and the growth rates were highly variable after return to litters. Bottle feeding and returning neonates to litters where they competed for access to milk, likely differentially stressed animals and contributed to piglets’ nutritional environment. Moreover, sow milk quality likely varied across litters. With this in mind, circulating levels of nutrients and gross measures of growth were used as corollaries of the postnatal nutritional environment.Variables used to assess mammary development over the first week postnatal correlated with each other, supporting the potential for underlying relationships. PEA reflects the relative amount of epithelial tissue in the gland and was positively correlated with the percent of epithelial and stromal cells proliferating on postnatal day seven, as well as the average daily gain. Percent of proliferating epithelial cells was also related to protein synthesis over the first week postnatal. The different degrees of mammary development between animals in our study were most likely attributed to the postnatal environment and not prenatal factors, such as mammary histomorphology and DNA and protein f and FSR variables were not correlated with gilt birthweight. Whereas PEA and f and FSR of mammary DNA and protein were correlated with average daily gain. Moreover, crown-rump length on postnatal day seven was correlated with the fraction of newly synthesized DNA and DNA FSR. Although it is important to note that the nutritional environment in late gestation carries over to postnatal growth performance, as discussed below, in regard to lysine. Since gilts were matched by birthweight when assigned to treatments, the relationship to these variables would be minimally expected.The relationship between average daily gain and mammary growth metrics supports mammary growth as isometric to body growth in the first postnatal week. The relationship between average daily gain to the mammary parenchymal epithelial area is particularly intriguing with regard to our hypothesis. This relationship likely indicates that the more adequately nourished the gilt is, the more resources are partitioned to growing the secretory component of the gland. A higher proliferative index of epithelial cells was found in mammary tissue isolated at birth compared to tissue from COL10 and COL20 animals on postnatal day seven. The proliferative population at birth may be particularly sensitive to nutritional growth cues in the gilt’s postnatal environment. Studies of mature pigs support that the nutritional environment affects mammary development [20,21].As an essential amino acid, lysine availability is limiting to porcine growth. Studies of the effect of lysine availability on suckling neonates through maternal milk found diets with a 20% deficiency in lysine content had reduced litter growth by 8–10% [20]. At the same time, litter growth increased by 2.35 times with a 2.90-fold increase in amino acids in the sow’s diet. Improved growth was directly related to the increased intake of lysine and other amino acids by neonates [20]. Thus, our finding that plasma lysine levels at 24 h were related to the average daily gain of neonates is consistent with its availability to neonates limiting growth. Currently, it is not known what led to the varying concentrations of circulating lysine across gilts, as levels of lysine were not related to the 24 h dose of colostrum. Increasing lysine content of sow diet in late gestation increased the total number born alive and birth weight of piglets [22]. Similarly, increasing the lysine and fat content of sow diets in late gestation diets improved the overall performance of litters [23]. It also increased sow colostrum production. Although all sows in the present study were on the same diet, the efficiency of absorption of lysine by sows and placental lysine transfer to gilts during their late fetal growth may be a factor. Further research in this area is needed, as lysine levels at 24 h postnatal were also positively related to the fraction of newly synthesized DNA and the fractional synthetic rate of DNA in mammary parenchymal tissue over the first seven days postnatal. At the level of the cell, nutrients, including amino acids, regulate gene expression [24]. The mTORC1 signaling pathway in cells functions to integrate nutrient availability, growth factor signaling and developmental cues to regulate growth [25]. The production of proteins, lipids and nucleotides need to increase for cells to grow and divide, while catabolic pathways such as autophagy need to be suppressed. mTORC1 regulates all of these processes [25]. Amino acids and positive cellular energy status activate mTORC1, whereas lack of amino acids or energy inhibits its activity [26]. Thus, the association of lysine with the total amount of DNA synthesized and DNA synthetic fractional synthetic rate likely reflects that it is indicative of nutrient-amino acid availability that can be used for cellular growth, and in particular, nucleotide synthesis. The body does not store amino acids, so muscle protein is mobilized to produce free amino acids [24]. Protein undernutrition decreases the plasma level of most essential amino acids and causes adjustments in physiological functions, with a primary consequence of feeding a low protein diet resulting in the inhibition of growth [24]. The lower levels of essential amino acids and lower growth of COL10 gilts relative to COL20 animals likely reflect adaptations of COL10 animals to the undernutrition they experienced over the first 24 h of the experiment. Furthermore, lower in COL10 animals, there were several non-essential amino acids, including glutamine. For maximal growth performance, pigs require dietary glutamine [27]. Studies of neonatal pigs found that nearly all glutamate and glutamine feed was metabolized by the gut, so that glutamate and glutamine in the body must derive almost entirely from synthesis de novo [28]. Glutamine and glutamate are precursors and products of each other, with reactions catalyzed by glutamine synthetase and glutaminase. Circulating glutamate levels appear to be resistant to large variations, with levels relatively constant in experimental manipulations of diet and energy [29]. Part of this persistence may be due to the sensitivity of hepatic glutaminase to metabolic state and its transcriptional stimulation induced by starvation. Meanwhile, there is little effect of metabolic state on glutamine synthetase [29]. This may explain the significantly lower levels of glutamine in COL10 versus COL20 animals. Moreover, although seemingly counterintuitive, the negative relationship between glutamate and PEA may reflect that lower energy in the postnatal environment negatively impacts mammary epithelial expansion.In the present study, the effect of returning piglets to the birth litter after 24 h of bottle feeding was not controlled, other than by matching treatments by litter. Differences in milk composition between sows and competition between piglets could influence developmental trajectory in the mammary tissue. In addition, this study did not consider other bioactive factors in colostrum, like fatty acids or hormones. Future studies aiming to understand the programming effects of colostrum need to control for maternal effects on piglet development after colostrum feeding by returning piglets used in the study to a common sow. This approach would limit competition of study piglets with established piglets that have experience nursing from the dam. Moreover, to control for nourishment versus bioactive factors in milk, future studies should use a nutrient-matched formula that represents the energy provided by colostrum but does not contain bioactive components. Finally, future studies should include more time points for sample and tissue collection to increase the understanding of the mechanisms involved in neonatal programing. 5. ConclusionsOverall, this study found plasma lysine levels at 24 h postnatal were positively related to average daily gain and DNA synthesis in mammary parenchyma over the first week postnatal. This relationship supports that the nutritional environment affects early mammary development. Moreover, data support the potential that higher lysine levels in the perinatal period favored a greater degree of cell division versus differentiation in mammary parenchyma of neonatal pigs. Further investigations are needed to determine if manipulating the level of lysine provided to neonatal pigs affects postnatal mammary development and distribution of cell types in the parenchyma.

animals : an open access journal from mdpi

10.3390/ani11123595

PMC8698004

The present study aimed to evaluate the accuracy of the leg volume obtained by the Microsoft Kinect sensor to predict the carcass composition of twenty-two male light lambs. The carcasses were divided into eight cuts, which were grouped according to their commercial value into high-value, medium value, and low-value. Several linear, area, and volume leg measurements were performed to predict cut and carcass composition. The leg volume determined by 3D image reconstruction using Microsoft Kinect sensor and Archimedes principle shows the higher correlations values with cuts and carcass. Additionally, it was observed that the models, which include the leg volume obtained by the Kinect sensor, are very good in predicting the weight of the medium value and leg cuts (R2 of 0.763 and 0.829, respectively). Thus, the results of this study confirm the good ability to estimate cuts and body traits from light lambs with volume measurements, particularly those obtained with the Kinect 3D sensor.

This study aimed to evaluate the accuracy of the leg volume obtained by the Microsoft Kinect sensor to predict the composition of light lamb carcasses. The trial was performed on carcasses of twenty-two male lambs (17.6 ± 1.8 kg, body weight). The carcasses were split into eight cuts, divided into three groups according to their commercial value: high-value, medium value, and low-value group. Linear, area, and volume of leg measurements were obtained to predict carcass and cuts composition. The leg volume was acquired by two different methodologies: 3D image reconstruction using a Microsoft Kinect sensor and Archimedes principle. The correlation between these two leg measurements was significant (r = 0.815, p < 0.01). The models to predict cuts and carcass traits that include leg Kinect 3D sensor volume are very good in predicting the weight of the medium value and leg cuts (R2 of 0.763 and 0.829, respectively). Furthermore, the model, which includes the Kinect leg volume, explained 85% of its variation for the carcass muscle. The results of this study confirm the good ability to estimate cuts and carcass traits of light lamb carcasses with leg volume obtained with the Kinect 3D sensor.

1. IntroductionThe evaluation of carcass characteristics is a fundamental process for attributing the quality and value of the animal at slaughter. Over the last three decades, several approaches for carcass grading systems supported by objective measurements have been developed for cattle [1,2], pigs [3], and sheep [4,5]. In general, these works aimed to classify the carcass quality based on non-destructive image analysis techniques by introducing consistency, accuracy, credibility, and confidence in the value assessment of the carcass [1,6]. The ultimate aim is to replace the subjective evaluation grounded on standards and move towards advanced value-based payment systems [7]. These evaluations techniques are also relevant for the classification of light lamb carcasses produced in Mediterranean countries. In these regions the slaughter of lambs from dairy breeds at 4 to 6 week of age with a low body weight ranging approximately between 10 and 11 kg is one of the most widely used production system [8]. These light carcasses are from different local breeds and particularly those with Protect Denomination of Origin (PDO) and Protected Geographical Indication (PGI) quality labels are regarded with high edible value. “Borrego Terrincho” and “Cordeiro Mirandês” with PDO [9,10] or “Lechazo de Castilla y León” and “Ternasco de Aragón” with PGI [11] are some examples of light lamb products with quality labels that can be found in Portuguese and Spanish markets. Unlike heavier lamb carcasses for light carcass weights (carcass weight < 13 kg), there is no conformation assessment, and therefore they are tab penalized due to their naturally poor morphology [12].Video image analysis (VIA) has been one of the most researched technologies and commercial solutions for carcass assessment of beef [1,13], pork [14], and lamb [15,16]. The VIA equipment used in these studies has been tailored for large industrial slaughterhouse plants that mainly use color and dimensional data obtained from 2D images of lateral or dorsal views of carcasses. From these images, data are extracted to estimate yield, conformation, and EUROP fat and conformation scores [3]. However, some work has been developed in recent years using 3D sensors to obtain information on carcasses [17,18] and live animals [19,20]. Some of these works use 3D sensing devices such as stereoscopic and time-of-flight cameras. Advanced depth imaging with low-cost sensors such as Microsoft Kinect is increasingly used in animal science [21,22]. The latter sensor can provide 3D data from its infrared and RGB color (Red, Green, Blue) images, representing a flexible objective technology that can be applied in predicting carcass composition, cut distribution, and lean yield prediction of carcasses. Despite the potential of this technology, there is a lack of information on its application to lamb carcasses. In this regard, this preliminary study aimed to evaluate the accuracy of leg volume obtained by the Microsoft Kinect sensor to predict the composition of light lamb carcasses.2. Materials and Methods2.1. Animals and CarcassesThe trial took place at the animal facilities of the University of Trás-os-Montes and Alto Douro (UTAD) at Vila Real (Portugal), and all the handling was performed according to the Portuguese law on animal welfare in experimental research. The protocol was approved by the ORBEA (Animal Welfare Body) of UTAD (669-e-DZ-2018). The trial was performed on twenty-two Churra da Terra Quente male lambs, weighing 17.6 ± 1.8 kg. After slaughter, the carcasses were obtained and then refrigerated at 4 °C for 24 h. After this period, the cold carcass weight (CCW) was recorded.2.2. Leg Area and Leg Linear Carcass MeasurementsCarcasses were measured in two ways. First, the perimeter measure of the hindquarter was obtained from the entire carcass. Then, the carcasses were split along the spine, and the left side was used to perform the remaining measurements. For this, the procedure using image analysis proposed by Batista et al. [23] was used. Briefly, the carcass measurements were recorded from photographic images of the left outer side. The images were obtained with a digital camera (Nikon D3100, Mitsubishi, Tóquio, Japan) with an 8-megapixel sensor positioned at 3 m from the carcasses and under a constant standard artificial light. The acquired images were analysed with the Fiji software (ImageJ 1.49u, National Institutes of Health, MA, USA) [24] to calculate the measurements of leg area, leg length, leg perimeter, hind quarter perimeter, and three widths (thinnest and largest leg width and minimum waist width).2.3. Carcass Cuts and CompositionAfter obtaining carcass measurements, the half-carcasses were divided into eight cuts: neck, shoulder, breast, anterior rib, rib, loin, chump, and leg, as described by Santos et al. [25]. Following the methodology proposed by Rodrigues et al. [26], the cuts were split into three groups according to their commercial value: high-value group (HVC), which included the leg, chump, and loin; medium value group (MVC) that included rib and shoulder; and low-value group (LVC) in which the breast, anterior rib, and neck were included. The leg (Figure 1) volume was then acquired by two different methodologies: 3D image reconstruction using a Microsoft Kinect sensor and Archimedes principle. After that, all cuts were dissected into muscle, fat (which includes subcutaneous and intermuscular fats), and bone, according to the methodology proposed by Panea et al. [27]. All dissection work was performed in a room under a controlled environment.2.4. Leg Volume with Kinect 3D ImageA Microsoft Kinect 2.0 sensor (Microsoft, NM, USA) was used to acquire the leg volume by 3D image reconstruction. This sensor incorporates an RGB camera and a depth infrared sensor, which is the main feature used in the study. The minimal computer hardware requirements needed for taking 3D images with the Kinect sensor are Windows 8 operational system, Dual-core 3.1GHz processor, 4GB RAM, and a 3.0 USB port. The Kinect Fusion Explorer program, included in the Windows Software Development Kit provided by Microsoft, was used to scan the leg and build the 3D image model. This software allows choosing the maximum and minimum distances from the depth sensor to the object used to capture the image. It also allows capturing the object colors and choosing how detailed the final model will be. The program configuration will vary due to differences in luminosity and reflection of the light in the surrounding environment and the scanned object. Therefore, there are no standard configurations to run the program as they will be different for different objects. Furthermore, the computer specifications also affect the image acquisition performance: the better the computer, the more detailed and smoother is the image generated.For the 3D model generation, the cut was hung in a structure that allowed the Kinect sensor to be moved around it undisturbed. The camera was held at chest height and slowly moved around the cut, so the program had time to construct a model with the most detail possible. All the models were exported in STL format. After that, models were imported into the Autodesk Meshmixer program to determine the leg volume. For that, the two steps were performed. First, the background was segmented from the leg region using an edition tool of the Meshmixer toolbar. After that, leg measurements were acquired. For that, the manually acquired leg length, in centimeters, was input in the Z-axis of the Meshmixer toolbar to serve as a reference for px to metric units’ transformation. Then, the program automatically determined the leg volumes.2.5. Leg Volume with Archimedes PrincipleThe Archimedes principle was also used to acquire the leg volume, as this is the standard method used to acquire that measurement. First, a 5 L container was filled with water until the superficial tension was reached. The container was large enough to fit the entire leg. This container was placed inside another container. After that, the cut was slowly inserted into the water, and the outside container captured excedent water. It was assumed that the water density value is 1 kg/L (one kilogram per liter) and that one liter of water is equivalent to 1000 cm3. With these assumptions and based on the weight of the spilled water, the leg volume was calculated. Special care was taken with measuring the spilled water. For this, a precision balance (Precisa LT 6200C, Precisa, Livingston, UK) with a resolution of 0.1 g was used, and all procedures for weighing the water were kept constant.2.6. Statistical AnalysisA descriptive statistical analysis was performed. Mean, standard deviation, maximum and minimum value, and coefficient of variation were obtained for the weight of cold carcass, cuts, and carcass composition, and for the measurements of carcass and leg. A correlation analysis was performed to examine the relationship between cut and carcass composition and all carcass and leg measurements. Additionally, a correlation analysis was performed between the leg volumes obtained with the Archimedes principle and with the Kinect sensor. Additionally, a multiple regression analysis that included the CCW and the carcass and leg measurements was performed. The best equations were chosen based on the precision of the prediction model, measured by the coefficient of determination (R2), and the residual standard deviation (RSD). As an indicator of the overall prediction ability, the models were also evaluated for the ratio of prediction to deviation (RPD), which is calculated as the ratio of standard deviation (sd) values to the RSD of the multiple regression (RPD = sd/RSD). All statistical procedures were carried out using the JMP software version 15 (SAS, Cary, NC, USA) [28].3. Results and Discussion3.1. Cold Carcass Weight, Cuts, and Carcass CompositionTable 1 summarizes the descriptive statistics (mean, standard deviation, minimum, maximum, and coefficient of variation) for cold carcass weight, cuts, and carcass composition. The carcasses show a small weight range (CCW between 6.85 and 9.91 kg) and a reduced variation (CV = 10.1%). These characteristics were observed in other studies with light carcasses with CV between 4 and 13% [25,29,30]. As previously mentioned, light carcasses are closely associated with the traditional lamb meat production in Mediterranean countries [8,12] and are often linked to PDO and PGI quality labels, which present in their specification and narrow carcass weight ranges [13], which explains the low variation for this type of carcass. Regarding muscle and fat from cuts and carcass, as expected, the variation is higher for fat (CV between 19.5 and 34.8%) than for muscle (CV between 7.6 and 12.2%). Still, regarding fat, a more significant variation is observed in LVC than in HVC cuts (CV = 34.8 vs. 16.5%, respectively). These results are in line with what was pointed out by [31] that the cuts with the highest fat content are those included in the LVC, such as the breast, with 42.1%, and the leanest cut was the leg.The mean, standard deviation, minimum, maximum, and coefficient of variation of leg measurements obtained through VIA are presented in Table 2.All measurements show a reduced variation (CV between 4.2 and 9.1%), which agrees with the reduced variation of CCW of the studied carcasses.3.2. Correlation between Measurements and Composition of Cut and CarcassTable 3 shows the correlation values of the leg measurements with cut and carcass weight and composition traits. In general, there is a significant correlation between the different leg measurements and cut and carcass traits. However, only the leg volume measurements were significantly correlated with all cuts and carcass traits (r between 0.417, p < 0.05 and 0.835, p < 0.01). Additionally, the correlation values of leg volume (Archimedes and Kinect 3D) with cuts and carcass traits show a very similar pattern, which reflects the relationship between these two leg measurements (r = 0.815, p < 0.01). The leg length is the measurement that shows the smallest correlations with cuts and carcass traits (r between 0.084, p > 0.05 and 0.450, p < 0.05). The leg area measurement presents intermediate correlation values with only two non-significant correlations (r = 0.337 and 0.413, p > 0.05 for leg muscle and LVC muscle, respectively). Additionally, the correlation values are generally less significant with the fat trait, whereas the correlation values are comparable for the cut and carcass weight and muscle. The value of the measurements obtained in the leg, in general, have been observed by other authors who have used linear and area measurements to predict cuts and lean meat variation of carcasses [23]. In this work, which studied light carcass, although there is no simple correlation between the measurements obtained in the carcass and the cuts, it is possible to observe that area and perimeter of the leg are included on the HVC, MVC, and LVC cut weight prediction models, whereas for models’ prediction of lean meat weight always include leg area measurements.Multiple regressions were studied with each of the leg measurements and the CCW. The equations that best explain the weight, muscle, and fat composition of the cuts and carcass are presented in Table 4. Carcass weight is extensively used in studies to predict carcass composition as it is an accessible variable and shows to be an informative predictor for primal cut variations [32,33].The leg volume measurements are the most used in multiple regressions with CCW to estimate cuts and carcass traits. Of the fifteen models presented, only those estimating LVC and fat carcass did not include leg volume measurements. It is also observed that for the leg volume measurements, the one obtained by the Kinect 3D sensor was the most used (9 out of 13 models). The models that include leg volume measurements obtained by the Kinect sensor are shown to be very good in predicting the weight of the MVC and leg cuts (R2 of 0.763 and 0.829, RDP of 2.0 and 2.3, respectively). In turn, for the HVC, the model that includes the Archimedes volume measurement and the CCW presents a very good prediction (R2 = 0.817, RDP = 2.2). These values are very close to what was observed for the model that included the leg volume measurement obtained by the Kinect sensor (R2 = 0.804, RDP = 2.2; data not shown), which reinforces its ability to predict cut traits. The results of cuts muscle estimation models are more modest, and only the MVC muscle estimation using the leg volume measurement with Kinect 3D has good prediction capacity (R2 = 0.723, RDP = 1.8). For the carcass muscle estimation model with the Kinect leg volume measurement, it was possible to explain 85% of its variation with the model classified as very good (RPD = 2.4). Regarding the cuts of fat, all models were classified as poor or fair (R2 between 0.433 and 0.577, RPD between 1.3 and 1.5). In turn, for the carcass fat estimate, the model showed higher capacity but with the leg area measurement included in the model (R2 = 0.742; RDP = 1.9).Two-dimensional video image analysis is one such technique that has been used for carcass evaluation of different species. Although 2D shape information can be useful, 3D information is preferable to ensure more accurate weight estimates [35] and, therefore, better carcass quality prediction. There is a growing interest in developing prediction models of carcass and meat quality traits using 3D measurements, for example, computed tomography (CT) and other image-based approaches [5,7]. With CT, most works target pigs [36,37]. This priority is understandable for the economic expression of this species. Despite this, contributions to predicting the body and carcass composition of small ruminants were also made with CT using 3D measurements [38,39]. In general, the results are promising, and it is expected that with the equipment progress and with advances in 3D carcass modelling software, it will be possible to speed up all procedures to obtain accurate information based on 3D images both in vivo and in the carcass [39]. In addition to CT, other techniques are also evolving to obtain 3D images, such as dual-energy X-ray absorptiometry—DXA [40]. Despite the enormous value of CT and DXA, its cost and complexity of use are substantial limitations as an aid to predict carcass traits that would maximize carcass value. However, the potential of three-dimensional (3D) image reconstruction has not yet been largely explored, especially for lamb carcass assessment.The low-cost Microsoft Kinect depth sensor has been used as a tool for rapid, reliable, objective, and non-invasive measurements in animal science [12], demonstrating potential as a carcass measurement device. Therefore, the present work evaluated the feasibility of using the Microsoft Kinect sensor to obtain the volume (3D reconstruction) of twenty-two light lamb legs. The volumes obtained were then assessed for their significance in predictive models of carcass and cuts traits. Nine out of fifteen models analyzed required the use of leg volume obtained through 3D reconstruction, along with cold carcass weight, as an independent variable, with the majority of the models presenting an R2 of 0.7 or higher. Such positive results expose the potential of using a Kinect sensor to predict light lamb carcass composition. However, some constraints with the proposed method must be overcome to take full advantage of the Kinect sensor capability in estimating carcass traits. The constraints rely on reduced variables, variation and sample size, limiting predictive capabilities. Furthermore, the method is not fully automated, requiring a careful scan of the cut by manually moving the sensor and the manual input of the leg length for unit calibration. Therefore, the possibility of obtaining leg volume measurements automatically is an attribute that must be followed to improve the capacity of this technique. In this way, it will be possible to make the image capture and analysis procedure faster, and in this way to overcome time as a constraint even in its potential application to breeds that give rise to carcasses with a quality label, not posing the critical problems in the chain speed of a large commercial abattoir [41] as this type of animal is generally linked to local slaughterhouses [42].4. ConclusionsThis study confirms the feasibility of the Kinect 3D images to predict light lamb carcasses composition from leg volume. Accordingly, the following steps of this research should include larger sample size and focus on the automation of both acquisition and analysis of 3D images in order to produce a more reliable, fast, and practical method of lamb light carcass assessment; and with light carcasses from different breeds and using the Kinect 3D as a tool to find a benchmark for quality of that type of carcass.

animals : an open access journal from mdpi

10.3390/ani13091553

PMC10177108

Tibetan sheep are the characteristic sheep breed on the Qinghai-Tibet Plateau, characterized by good adaptability to the hypoxic conditions, delayed maturation, and low fecundity. The epididymis is a male reproductive organ well known to be responsible for sperm transport, storage, and maturation, which is crucial for male fertility. To clarify the dynamic gene expression patterns and their potential contribution during sperm maturation of Tibetan sheep, in this study we characterized the comprehensive transcriptional profiles in the three epididymal areas (caput, corpus, and cauda) of Tibetan sheep using RNA sequencing. The results revealed that numerous genes are present in Tibetan sheep epididymis in a stage-region-dependent manner, showing more dramatic changes in gene expression in various epididymal areas of post-pubertal Tibetan sheep. These genes perform some sort of function in reproduction, development and morphogenesis, and immune privilege to facilitate the maturation of spermatozoa and to provide the microenvironment required for sperm development and maturation. This study provides new insights into the molecular mechanisms by which genes are regulated during post-testicular sperm development.

While traveling through the epididymis, immature sheep spermatozoa undergo a sequence of processes that ultimately give them the capacity to swim and fertilize an egg. Different gene expression patterns may be found in the epididymal caput, corpus, and cauda, conferring variant or unique biological roles during epididymis development and sperm maturation. To search for candidate genes associated with ovine sperm maturation and assess their possible modulating mechanisms, we characterized gene expression in each epididymal segment derived from pre- and post-pubertal Tibetan sheep by RNA sequencing. Compared with pre-puberty, 7730 (3724 upregulated and 4006 downregulated), 7516 (3909 upregulated and 3607 downregulated), and 7586 (4115 elevated and 3471 downregulated) genes were found to be differentially expressed in the post-pubertal caput, corpus, and cauda epididymis, respectively, and real-time quantitative PCR verified the validity of the gathered expression patterns. Based on their functional annotations, most differential genes were assigned to the biological processes and pathways associated with cellular proliferation, differentiation, immune response, or metabolic activities. As for the post-pubertal epididymis, 2801, 197, and 186 genes were specifically expressed in the caput, corpus, and cauda, respectively. Functional annotation revealed that they were mainly enriched to various distinct biological processes associated with reproduction (including the caput binding of sperm to the zona pellucida; fertilization in the caput and corpus; and meiosis in the caput and cauda) and development (such as cell differentiation and developmental maturation in the caput; cell proliferation and metabolism in the corpus; and regulation of tube size and cell division/cell cycle in the cauda). Additionally, we focused on the identification of genes implicated in immunity and sperm maturation, and subsequent functional enrichment analysis revealed that immune-related genes mainly participated in the biological processes or pathways associated with the immune barrier (such as JAM3 and ITGA4/6/9) and immunosuppression (such as TGFB2, TGFBR1, TGFBR2, and SMAD3), thus protecting auto-immunogenic spermatozoa. Additionally, sperm maturation was mostly controlled by genes linked with cellular processes, including cell growth, proliferation, division, migration, morphogenesis, and junction. Altogether, these results suggest that most genes were differentially expressed in developmental epididymal regions to contribute to microenvironment development and sperm maturation. These findings help us better understand the epididymal biology, including sperm maturation pathways and functional differences between the epididymal regions in Tibetan sheep and other sheep breeds.

1. IntroductionSpermatozoa in mammals are unable to fertilize an egg once they leave the testis because they are immobile and lack the ability to undergo capacitation. Spermatozoa only fully develop when traveling through the epididymis, including their acquisition of motility and fertilizing capacity [1]. The epididymis joins the testis and vas deferens by a single, highly coiled, convoluted tubule. Principal, basal, narrow, transparent, and halo cells, among others, make up the epididymal epithelial lining, which regulates sperm maturation prior to fertilization, exhibiting a variety of biological roles such as promoting structural development, sensing and responding to changes in their surrounding microenvironment, and communicating with spermatozoa [2]. Before being released during ejaculation, spermatozoa are transported sequentially through the caput, the corpus, and finally the cauda [3,4]. Since spermatozoa tend to be synthetically inactive, maturation relies on the spermatozoa interacting with proteins made and secreted by particular parts of the epididymal epithelium [2]. Exposure to the luminal environment of the epididymis, rather than events inherent to germ cells, is responsible for spermatozoa maturation [5,6]. Although the epididymis seems to be a simple tubule, each segment has its own unique gene expression profile and physical characteristics, which establish and maintain the constantly changing luminal environment required for sperm maturation [7]. Segment-specific secretion of proteins into the luminal fluid of the epididymal epithelium determines gene expression patterns that directly or indirectly impact sperm maturation, transport, and storage [8,9]. The distinct patterns of gene expression in each of these areas have been linked to physiological processes that play a role at various stages of sperm development [10]. Therefore, it is crucial to comprehend epididymal sperm maturation by finding and understanding the role of segment-specific genes. Many genes have been reported to participate in the development and function of mammalian epididymis, such as aquaporins, beta defensins, claudins, cadherins, lipocalins, and the glutathione peroxidase family [11,12]. For instance, GPX5, an important member of the glutathione peroxidase family, is specifically expressed in the mammalian epididymis and functions to protect sperm from reactive oxygen species and lipid peroxidation damage during maturation [13]. Peroxiredoxins (PRDXs) are involved in the protection of sperm function and DNA integrity during epididymal maturation to ensure male fertility [14].Overall, understanding the maturation of sperm requires a full understanding of the expression of genes and the biological activities they serve in the epididymal segment. Both extrinsic (such as nutritional levels) and intrinsic (such as androgens) factors can affect the production, development, maturation, and quality of sperm through controlling the expression of genes [15,16,17,18]. In order to produce the appropriate luminal environment for the functional development and protection of spermatozoa, greater emphasis must be placed on the regional expression of epididymal genes. Unfortunately, very few reports in this context are available on sheep. Recognizing the regulatory and functional differences in genes between the various parts of the sheep epididymis requires a systematic approach to understanding their transcriptional patterns. Throughout the Qinghai–Tibet Plateau, the remarkable domestic sheep breed known as Tibetan sheep (Ovis aries) is found mostly at elevations of 3000 m above sea level [19]. Due to their long-term, excessive dependency on grazing without supplemental food, Tibetan sheep have late sexual maturation (around 1 year old) and a low reproduction rate (including seasonal estrus, initial mating aged 2.5 years old, lambing once per year, and having only 1 lamb at a time,). It is crucial to the study of sheep reproduction to get a deeper understanding of the processes governing epididymal development in male Tibetan sheep. However, the gene profiles in developmental Tibetan sheep epididymis and their function in each epididymal region are unknown. Therefore, we investigated the dynamic gene expression profiles in developmental Tibetan sheep epididymis (caput, corpus, and cauda) using high-throughput RNA sequencing (RNA-seq) technologies and then assessed their potential functions during sperm maturation. The findings of this study have significant implications for understanding the mechanisms regulating ram sperm maturation and, ultimately, for enhancing the quality of semen and, by extension, male reproductive success.2. Materials and Methods2.1. Ethical StatementAll experiments were conducted in accordance with the National Laboratory Animal Welfare instructions (2006-398) and were authorized by the Ethics Committee of the Laboratory Animal Center at Gansu Agricultural University (GSAU-Eth-ASF2022-008).2.2. Sample Collection and ProcessingThe Xike Tibetan Sheep Breeding Base (Xiahe, China) offered 8 male Tibetan sheep descended from the same father at 2 reproductive stages: pre-puberty (3 months old, 3M; n = 4; weighed 9.23–10.18 kg; scrotal circumference: 8.4–8.9 cm) and post-puberty (1 year old, 1Y; n = 4; weighed 34.25–36.73 kg; scrotal circumference: 17.6–18.3 cm). All Tibetan sheep were from the same farm under traditional grazing management which meant that the sheep grazed in a fenced pasture all year round with free access to food and water without feed supplements. All animals were sacrificed with an overdose of sodium pentobarbitone (intravenous). Following incision of the scrotum, the testis and epididymis were exposed, and the middle segment from each right epididymal region (caput, corpus, and cauda) was harvested immediately, rinsed in PBS to remove excess blood, quickly placed into a centrifuge tube containing RNA protective agent, immediately put into a liquid nitrogen tank, transported back to the laboratory within 1 h, and stored in a −80 °C refrigerator for RNA extraction.2.3. RNA Extraction, Library Preparation, and SequencingTotal RNA was isolated from frozen epididymal tissue following the manufacturer’s instructions using a Trizol kit (Invitrogen, Carlsbad, CA, USA). RNase-free agarose gel electrophoresis, the Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA), and a NanoDrop Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) were used to evaluate the RNA quality. All RNA samples had good integrity (RNA integrity number (RIN) > 7.5). The mRNA was purified and fragmented to approximately 200 bp from the isolated total RNA after removal of rRNA, followed by being subjected to first- and second-strand cDNA synthesis. Final cDNA libraries were made by purifying and enriching cDNA fragments that had previously undergone end repair and adaptor ligation before being sequenced using an Illumina HiSeq 2500 platform (Gene Denovo Biotechnology, Guangzhou, China). Finally, we constructed 24 cDNA libraries (4 replicates per group), including pre-pubertal caput (i.e., Cp_3M-1, Cp_3M-2, Cp_3M-3, Cp_3M-4), corpus (i.e., Cr_3M-1, Cr_3M-2, Cr_3M-3, Cr_3M-4), and cauda (Cu_3M-1, Cu_3M-2, Cu_3M-3, Cu_3M-4), as well as post-pubertal caput (i.e., Cp_1Y-1, Cp_1Y-2, Cp_1Y-3, Cp_1Y-4), corpus (i.e., Cr_1Y-1, Cr_1Y-2, Cr_1Y-3, Cr_1Y-4), and cauda (i.e., Cu_1Y-1, Cu_1Y-2, Cu_1Y-3, Cu_1Y-4).2.4. Differential Expression Analysis and Data ProcessingHigh-quality clean data were obtained by filtering low-quality reads and adaptors of all the samples for subsequent analyses. Reference genome assembly Oar v4.0 for sheep was used to align the clean reads utilizing HISAT default parameters. The expression abundances of genes were measured in FPKM using StringTie [20]. DESeq2 was used for group-specific differential expression analysis. Parameters of absolute fold change (FC) ≥ 2 and false discovery rate (FDR) below 0.05 were used to identify differentially expressed genes (DEGs).2.5. Evaluation of Functional EnrichmentIn order to better understand the biological functions of DEGs, we conducted a functional enrichment analysis utilizing the GO (http://geneontology.org/ (accessed on 17 September 2022)) and KEGG databases (http://www.genome.jp/kegg/ (accessed on 17 September 2022)).2.6. Quantitative Real-Time PCR as an RNA-Seq Validation ToolTo evaluate the repeatability and reproducibility of RNA sequencing data, a quantitative real-time PCR (qPCR) study was carried out on DEGs. Specifically, cDNA was reverse-transcribed from isolated total RNA using the EvoM-MLV RT Kit (Accurate Biotechnology, Hunan, China), and qPCR assay was carried out using the SYBR Green qPCR Kit (Accurate Biotechnology, Hunan, China). At least three independent repetitions of each experiment were performed. Table S1 contains information on the primers used. Using a 2−ΔΔCt method, the expression levels relative to the internal reference gene (β-actin) were calculated.3. Results3.1. Overview of Transcriptome DataIn all, 187.57 Gb of clean data were generated by the sequencing. The total size of the clean data for each sample was 6.39 Gb, and the Q30 base percentage was 90.13% or higher. Each sample’s clean reads were sequence-aligned to a predefined reference genome, and the alignment accuracy ranged from 87.11% to 89.5% (Table S2).3.2. Differentially Expressed Gene Analysis and Functional Analysis3.2.1. Identification and Functional Analysis of Differential Expression Genes in Caput Epididymis before and after PubertyIt was found that there was high repeatability across four independent samples from each group, as seen by the sample correlation heatmap derived from the mRNA expression profiles (Figure 1A). A total of 7730 transcripts were expressed differentially with 3724 upregulated and 4006 downregulated in the 1-year-old group (Figure 1B,C). The predominant enriched GO terms of these genes, in the biological process, were associated with metabolism, growth and development, reproduction, biological adhesion, and immune function; in the cellular component, were associated with extracellular region, membrane part, and cell junction; in the molecular function, they were associated with binding, catalytic activity, and transporter activity (Figure 1D). The KEGG analysis revealed that the majority of these differential genes were enriched in metabolic pathways (such as lipid, amino acid, and energy metabolism), cellular processes (such as cellular growth, death, and motility), and organismal systems (such as development, immune, and endocrine systems) (Figure 1E).3.2.2. Identification and Functional Analysis of Differential Expression Genes in Corpus Epididymis before and after PubertyThe sample correlation heatmap indicated high repeatability of the data from each group (Figure 2A). Compared with the 3-month-old group, the 1-year-old group had 7516 DEGs (3909 upregulated and 3607 downregulated) (Figure 2B,C). According to the functional analysis of GO annotation, these genes were shown to be involved in a wide variety of biological activities, including cellular function, metabolism, growth and development, reproduction, and immunity; in terms of the cellular component, GO terms associated with membrane, organelle, complex macromolecules, outside the cell, and cell junction were the most represented; in terms of molecular function, among the most enriched GO terms were “binding,” “catalytic,” “transporter,” and “molecular function regulator” (Figure 2D). Analysis of these DEGs revealed that they were abundant in metabolic pathways such as lipid metabolism, amino acid metabolism, and energy metabolism, with biological processes such as cell proliferation, death, and motility, as well as pathways associated with development and immune signaling (Figure 2E).3.2.3. Analysis of Pre- and Post-Pubertal Cauda Epididymal Gene Expression Differences and Their Functional SignificanceCorrelation heatmaps based on gene expression patterns demonstrated high repeatability across four samples from each group (Figure 3A). Differential expression analysis identified a total of 7586 DEGs (4115 elevated and 3471 downregulated) (Figure 3B,C). GO analysis indicted that metabolism, biological adhesion, growth and development, reproduction, and immune function were the main biological processes enriched by genes; with respect to cellular component, almost all gene expression was found within the organelle, outside the cell, membrane part, and cell junction; with respect to molecular function, most genes served some sort of binding, catalytic, or transporter function (Figure 3D). The majority of genes were enriched for their participation in cell proliferation, cell motility, development, immunity, and metabolism (such as lipid, energy, amino acid, and carbohydrate metabolism) (Figure 3E), as determined by the KEGG database.3.2.4. Region-Specific Gene Expression and Functional AnalysisTo compare the differences in gene expression and their biological functions across epididymal regions, we searched for genes that were specifically expressed in a single region. Only transcripts with an average FPKM value greater than 1 were considered expressed. For the pre-pubertal epididymis, we found that 87 and 190 genes were specifically expressed in the caput and corpus, respectively, but showed very low expression for the vast majority of genes (FPKM < 2); for the post-pubertal epididymis, 2801, 197, and 186 genes were specifically found in the caput, corpus, and cauda, respectively (Figure 4A). The subsequent GO annotation based on biological processes showed that the genes in the post-pubertal caput were mainly involved in processes associated with reproduction (such as meiosis-associated processes, sperm attachment to the zona pellucida, and the negative control of fertilization) and growth and development (such as multicellular organism development, cell development, cell differentiation, developmental maturation, and organ growth/development); the genes in the post-pubertal corpus were significantly enriched in biological processes including fertilization, sexual reproduction, the reproductive process, cell proliferation, the cellular metabolic process, and carnitine (a nutritional aid for improving sperm motility) transport; the genes in the post-pubertal cauda were significantly enriched in processes associated with reproduction (such as the reproductive process and meiosis), development (regulation of tube size, cell division, mitotic recombination, and cell cycle), and protein secretion (Figure 4B).3.2.5. Screening and Functional Analysis of Epididymal Immunoprotection-Related GenesThe intraluminal compartment of the epididymis is immunoprotected, similar to the testis. To investigate the potential activities of genes involved in sustaining immune function during epididymal development, we examined the immune-related DEGs in further detail based on the results from the GO and KEGG functional analysis. We discovered 428 (138 elevated; 290 downregulated), 385 (140 upregulated; 245 downregulated), and 411 (183 elevated; 228 downregulated) immune-related differential genes in the epididymis caput, corpus, and cauda, respectively (Figure 5A). Heatmap analysis of these genes revealed remarkable repeatability and gene expression levels across age groups (Figure 5B). See Table S3 for a complete list of these genes and their expression levels. Among these, 190 DEGs were co-expressed in all epididymal regions detected (Figure 5C). In order to better understand and interpret the possible biological roles of the genes, the KEGG database was used to determine the gene–pathway associations for all the immune-related genes across the various comparison groups. These genes were found to be significantly enriched in pathways related to the immune barrier (such as the tight junction, gap junction, adherens junction, ECM–receptor interaction, and modulation of actin cytoskeleton) and immunosuppression (such as cytokine–cytokine receptor interaction, B/T cell receptor signaling pathway, and Toll-like receptor signaling pathway) (Figure 5D).3.2.6. Analysis of DEGs Associated with Sperm MaturationTo probe the potential functions of the genes implicated in sperm maturation, we conducted further analysis of the genes encoding eight kinds of sperm-maturity-related protein families: glutathione peroxidases (GPXs), β-defensins (DEFBs), lipocalins (LCNs), interleukins (ILs), Toll-like receptors (TLRs), transforming growth factors (TGFs), peroxiredoxins (PRDXs), and aquaporins (AQPs). In total, we obtained 66 (41 upregulated and 25 downregulated), 57 (35 upregulated and 22 downregulated), and 55 (31 upregulated and 24 downregulated) differential genes in the caput, corpus, and cauda (Figure 6A), respectively. Of these, 27 genes were co-expressed in all epididymal segments (Figure 6B), with upregulated expression in all post-pubertal segments for most genes (Figure 6C). Interaction network analysis for co-expressed protein-coding gene sets showed that TGFs (TGFB1/2) and their receptor (TGFBR2) were in the more central position of the network (Figure 6D). Further functional annotation revealed the participation of these genes in various cell events, including cell growth, proliferation, division, migration, morphogenesis, and intercellular junction (Figure 6E).3.3. qPCR ValidationIn order to ensure the validity of the RNA-seq data, qPCR analysis was performed on 12 randomly chosen DEGs. In general, the qPCR data and RNA sequencing data showed similar expression patterns of elevation or reduction in mRNA between age groups (Figure 7), suggesting the reliability of the transcriptome data.4. DiscussionThe characterization of epididymal transcriptome is an attractive topic because the epididymis is a highly regional organ whose core function is to perform all the biochemical changes responsible for post-testicular sperm maturation and storage. To understand the establishment of epididymis development and sperm maturation, a complete and broad-scale transcriptional perspective of the epididymis and its developmental process is required. Here, we obtained 187.57 Gb of clean sequencing data during postnatal development of Tibetan sheep epididymis, allowing a more comprehensive study of ovine epididymal transcriptome. These findings provide the groundwork for comprehending the environment of sperm maturation by shedding light on the landscape of the epididymal transcriptome and a detailed investigation of transcriptomic alterations during epididymal development.RNA-seq was performed on pre-pubertal and post-pubertal epididymis tissues (caput, corpus, and cauda) to discover the possible genes involved in epididymal development and sperm maturation. We discovered a total of 23,463 transcripts in developmental epididymis tissues. Of these, 7730 (3724 upregulated; 4006 downregulated), 7516 (3909 upregulated; 3607 downregulated), and 7586 (4115 upregulated; 3471 downregulated) transcripts were variably expressed in the caput, corpus, and cauda of the developing epididymis, respectively. Each segment of the epididymis exhibits different gene expression patterns as well as diverse morphological characteristics, which lend this obviously simple tubule a higher level of functional complexity than could be anticipated [21]. Thus, we also analyzed the regional distribution of the transcripts in the caput vs. corpus, caput vs. cauda, and corpus vs. caput. We found that DE mRNAs were more abundant in the cauda epididymis at pre-puberty and more abundant in the caput epididymis at post-puberty. Interestingly, the differential genes in the corpus and cauda epididymis were occupied by the caput epididymis. The gene expression profiles in the corpus and cauda at post-puberty were not similar at all, but both differed from the caput to a similar degree. This is in contrast to previous studies in the human epididymis. James et al. discovered that the gene expression profiles in the corpus and cauda were remarkably similar [22]. This difference might be due to the fact that each species seems to have developed its own sperm maturation strategy. Studies in rodents confirm that the proximal epididymis (initial segment and caput) is the segment most responsive to intraluminal factors proposed to control downstream gene expression [23]. Our results likewise confirm that the corpus epididymis is the most active region for gene expression.According to a GO and KEGG database-based functional bioinformatics study, these divergent genes mostly function in processes and pathways associated with cell growth, reproduction, immunity, and metabolism. We found that the most enriched term for differential genes in the developmental caput, corpus, and cauda was cellular process, which falls under the biological process domain. We further discovered in the caput that upregulated (higher expression in the 1Y group) genes such as GPX3 and IL13RA1 were engaged in the cellular process term. GPX3 is involved in the protection of mature sperm from reactive oxygen species damage and catalyzes the degradation of peroxides to which sperm are exposed during the maturation process [24]; multiple processes rely on IL, including the development, activation, proliferation, and control of immune cells [25]. These activities in the caput epididymis work together to ensure proper sperm development at an early stage. The upregulated (relatively higher expression in the post-pubertal cauda) genes that were enriched in the cellular process GO term, such as GPR50, CST3, and LCN2, were reported to be correlated with the following cell components: GPR50 was associated with plasma membrane [26]; CST3 is a member of the cysteine protease inhibitor family [27]; proteases and protease inhibitors serve a functional role in protein processing at sperm plasma membrane levels during maturation [28]; and the LCN family is involved in lipid transport, which has implications for a wide range of biological processes such as the immune response, sperm maturation, and storage [29]. The term “cellular process” is used to describe any activity that occurs at the cellular level [30], which may or may not involve a single cell. For example, cellular communication, which involves several cells yet occurs at the cellular level. The upregulated genes in the corpus such as PRDX6 were identified as playing a role in epididymal cellular processes. The antioxidant effect of PRDX6 is important for the protection of sperm during epididymal maturation [31]. These roles of the upregulated genes enriched in the cellular process term, as indicated above, may be relevant for the proposed function of the epididymis such as the establishment of a luminal environment to promote sperm motility and ovum recognition.Pseudo-stratified epithelia line the epididymal duct and perform several crucial functions, including forming tight junctions and enforcing a physical barrier. Sperm, critically, have immunogenic properties [32]. The testicular barrier and immunosuppressive factors [33,34] have been shown to sustain a non-inflammatory local steady state, which is necessary for reproductive function. Nevertheless, the epididymis has a less stringent blood–tissue barrier than the testis, and while a variety of immunosuppressive factors have been found in the testis, such mechanisms for epididymal immune privilege have not yet been described [35,36]. The blood–epididymis barrier (BEB) is a membrane structure between the immune system and sperm that blocks interactions between sperm autoantigens and the immune system. It also functions to establish and develop the epididymal luminal environment, which shields developing spermatozoa from the immune system and is beneficial for normal sperm maturation [37]. The fully functional BEB is composed of anatomical, physiological, and immunological barriers [35]. The anatomical barrier is made up of tight junctions to restrict molecules and cells from entering and exiting the lumen, while the immunological barrier consists of different immune components inside and outside the tubule/duct [35]. Antigenic spermatozoa are shielded from an immunological response during sperm maturation, a process that is known to rely heavily on immune processes [38]. Hence, we analyzed immune-related differential genes in the pre- and post-pubertal caput, corpus, and cauda epididymis and identified a total of 725 genes. Overall, most immune-related genes are expressed in the caput epididymis, followed by the cauda and corpus, with most exhibiting significant downregulation in the post-pubertal caput, corpus, or cauda epididymis. Functional analysis suggested that significant enrichment of these genes was in pathways directly related to the make-up and function of the BEB [37,39], such as the tight junction, gap junction, adherens junction, ECM–receptor interaction, and regulation of actin cytoskeleton, as well as being associated with systemic immunosuppression [34], such as cytokine–cytokine receptor interaction and B/T cell receptor and Toll-like receptor signaling pathways. It is well recognized that the caput epididymis provides a tolerogenic environment, while the cauda epididymis favors pro-inflammatory conditions. Protecting spermatozoa from autoimmunity and defending them from pathogenic harm both rely on a well-controlled immune environment [40]. Thus, the reason for the higher number of immune-related genes in the caput and cauda epididymis of Tibetan sheep might be to protect spermatozoa from autoimmunity and defend them against pathogenic damage.Additionally, we analyzed the expression patterns and functional characterization of eight known gene families (GPXs, DEFBs, LCNs, ILs, TLRs, TGFs, PRDXs, and AQPs) implicated in sperm maturation to identify potential candidates and their biological roles during maturation of spermatids of Tibetan sheep. Antioxidant activities can be found in genes belonging to the GPX family, and individual GPXs have specialized biological roles. To date, eight genetic isoforms of GPXs (GPX1-8) have been reported in mammals, and these genes are widely distributed in different organs of mammals [41,42]. Among the family of GPXs, the activity of GPX1-4 is dependent on selenium, of which GPX4 is an important structural protein that is very abundant in sperm mitochondria (about 50% of sperm mid-segment proteins), while when it is absent it leads to male infertility [43]. Another member of the GPX family, GPX5, which can protect sperm DNA from oxidative damage, is secreted by the principal cells within the epididymal epithelium and is highly expressed in the caput epididymis [42]. Similarly, we found that GPX5 exhibits high expression levels only in the caput epididymis, which was consistent with previous research in Small-Tail Han sheep that found a decreasing trend of GPX5 expression from the caput to cauda epididymis [3], suggesting that GPX5 is involved in the early stages of sperm maturation.The epididymal microenvironment, including the BEB function, is maintained by the secretion of cytokines [1]. We evaluated the expression profiles of interleukins, Toll-like receptors, and transforming growth factors and found that the majority of all genes are downregulated in the post-pubertal epididymis, suggesting that most genes have an immunoprotective effect during the progressive maturation and motility of spermatozoa. DEFBs are a class of small cysteine-rich cationic peptides that serve antibacterial and anti-inflammatory activities and combine with sperm to control sperm maturation in the epididymis to ensure environmental stability [44]. Many DEFBs have been found in mammals where they form 4–5 syntenic gene clusters [45]. Almost all of the DEFBs are highly abundant in the epithelial cells in the male reproductive tracts, especially the epididymis [46]. Zhang et al. [47] showed that DEFB42, DEFB23, and DEFB26 are significantly more abundant in the caput and corpus epididymis, and both regions are known to be major contributors to sperm maturation. We found their homologs, DEFB33/36/107A/109/110/112/113/115/119/125/130/134, in the caput and corpus epididymis, and all of them showed highly upregulated expression levels at post-puberty, suggestive of roles in immunoprotection and maturation of spermatozoa. The lipocalin family participates in lipid transport and other biological functions, such as sperm maturation and storage and immune response [29]. Although Thimon et al. [11] found evidence of the existence of eight lipocalins in human epididymis, only LCN2 was abundant in the caput, whereas LCN6/8/10 were abundant in the corpus. This is not quite consistent with our study, where LCN6/8/10 were found to be highly present in the caput epididymis of Tibetan sheep, but is consistent with a study on pigs [48]. Notably, an interaction network among genes revealed that two TGFb isoforms (TGFB1 and TGFB2, well-known immune-homeostasis-related genes) and the receptor TGFBR2 were in the more central position of the network, participating in various cellular biological processes (such as cellular growth, proliferation, division, migration, morphogenesis, and junction) based on functional annotation. These results strongly suggested that these genes have multiple roles during sperm maturation and microenvironmental homeostasis in the epididymis of Tibetan sheep, but the mechanism by which they function in sperm maturation needs to be explored thoroughly.5. ConclusionsIn summary, we determined that the three conventional areas (caput, corpus, and cauda) of the epididymis of Tibetan sheep demonstrate significant variations in the expression of genes linked to their functions. Most genes exhibited stage–region-dependent expression changes and very active gene expression in the caput epididymis. Differential genes in each epididymal region of development participated in a wide variety of biological activities as well as related pathways, including growth and development, cell events, reproduction, metabolism, and immune function. The pre-pubertal epididymal regions showed very few and variable changes in gene expression, whereas changes in gene expression for the post-pubertal epididymal regions were dramatic, and specific genes were characterized by differential functions executed in the caput, corpus, and cauda epididymis for creating a microenvironment suitable for sperm development and storage as well as facilitating the maturation of spermatozoa. These results shed light on the spectrum of transcriptome changes that occur during epididymis development in Tibetan sheep, providing a foundation for further studies of the mechanism underlying sperm maturation in sheep.

animals : an open access journal from mdpi

10.3390/ani11113134

PMC8614323

Agonistic behavioural interactions play a decisive role in the competition for food, space, mating opportunities, and establishing social rank. We propose the use of the number of bites on the pelt of red deer as an index of agonistic interactions between group members. Using behavioural data from a 14-year time series of a captive population of Iberian red deer (Cervus elaphus), we found that deer that were of higher social rank, heavier, living in smaller groups, or under no heat stress conditions suffer less pelt bites than those of lower social rank, lighter, living in bigger groups, or under heat stressing conditions. Hinds that gave birth earlier in the parturition period suffered less pelt biting than those that gave birth around the peak of the parturition season. Pelt biting is useful to identify management situations in which deer welfare could be at stake.

Agonistic behavioural interactions play a decisive role in the competition for food, space, mating opportunities, and establishing social rank. We used pelt biting (number of bites on an animal’s body) as a proxy for assessing the intensity of agonistic animal interactions and how it responded to social, population, and heat stress factors. We modelled a 14-year time series of pelt biting records and observational data of agonistic interactions on a population of captive Iberian red deer (Cervus elaphus). We found that (i) the higher the social rank of deer, the lower the number of pelt bites received; (ii) increasing heat stress conditions caused deer to suffer more pelt bites; (iii) males received more bites than females; (iv) the heavier the deer, the lower the number of bites on their bodies; (v) the bigger the group, the more bites exhibited on its members; (vi) deer 5–6 years old suffered greater rate of pelt biting than younger or older deer; and (vii) hinds that gave birth earlier in the parturition period suffered less pelt biting than those that gave birth around the peak of the parturition season (p < 0.01 for all effects). Pelt biting is useful to predict management situations in which deer welfare could be at stake.

1. IntroductionAnimal societies are driven by a complex network of interactions between group members, and one type of interaction facilitates the establishment of the social rank between individuals, such as agonistic interactions [1]. Social rank and group structure are the expression of the dominant/subordinate role of the group members, which reflects the individual skills required to gain access to limited resources, for example, space, food, and mating opportunities [2,3,4]. Social rank within a group is dynamic [1,5,6], as it is affected not only by the outcome of continuous social interactions but also environmental factors that impact on resource availability [7,8]. Group size affects the intensity of agonistic interactions and the type and stability of the social structure. Small groups tend to have linear ranks and are more stable than bigger groups [9], since aggressiveness increases with group size, because the availability of resources per individual declines [10]. Individual traits, such as sex, size, and age, are instrumental in the outcome of social interactions and group structure [11,12].External abiotic stressors, such as heat [13], which affect aggregation patterns, comfort, or hormone levels, can also determine aggression activity [10,14]. Global climate models predict an increase in air temperature in the next few decades [15], and heat stress has been demonstrated to be an important stressor in domestic ruminants and in farmed deer [13,16]. It is recognised that good animal welfare and optimal levels of production are closely related, but there are benefits of improving animal welfare that extend beyond production gains and minimum legal requirements [17]. Animal aggression in farms is one of the main factors affecting livestock welfare, and it has direct repercussions on the economic viability of the exploitation [18].Ungulate species display a repertoire of agonistic behavioural interactions to establish their social rank. These behaviours range from vocalisations, displays of dominance or submission, butting, biting, kicking, chasing, and changes in the distance between individuals, and some can escalate to aggressive behaviours that can lead to injuries and even death [6,19,20,21,22]. Recording animal interactions in large groups is a time consuming task [23]. In the last decade, automatic devices have eased the collection of data for monitoring animal behaviour, specially spatial movement, from which social interactions can be inferred [21,24]. However, these devices are expensive and require complex logistics for their deployment.In red deer, one aggressive behaviour is biting the pelt of their peers [20,25]. The result of this behaviour can be observed on the animal´s pelt as a number of bite-size patches of fur that have been plucked. We assessed whether pelt biting can be efficiently used as a practical indicator to record agonistic interactions in captive red deer and how this behaviour is modulated by sex, body size, age, social rank, population density, and environmental stressors. We also tested some social predictions on animal interactions and discussed the potential applications of pelt biting as a useful indicator of social and environmental distress in deer farms.2. Materials and Methods2.1. HypothesesTwo groups of hypotheses were tested on the response of pelt biting, related to (i) the hierarchical rank of the deer and related body traits, which correlate positively with successful outcome in competition for resources, and (ii) the abiotic environmental stress where they live (Table 1). It is hypothesised that males individuals (H4) [26], younger individuals (H3), lower hierarchical rank individuals (H1), smaller body size (i.e., body weight, H2) individuals, and those exposed to heat stress (H6) suffer higher rates of pelt biting (Table 1).Sackett et al. [27], in a colony of captive pigtail monkeys (Macaca menestrina), found that mothers carrying female foetuses were more frequently harassed and bitten than those carrying male foetuses (H5, Table 1). They hypothesised that this was related to mother´s hierarchical status, contingent upon foetus sex-related hormone levels. Namely, during the second half of pregnancy, foetal male gonads begin secreting testosterone and producing a sharp increase in transplacental maternal circulating testosterone [28], which might induce maternal behavioural changes and produce chemicals that are exudate, making it possible for peers to smell them and so identify their condition.2.2. Pelt BitingIn our farmed red deer, we observed body biting between conspecific deer as a direct consequence of agonistic behaviour. Deer bit their peers in four main situations: when (i) deer were spatially very close together, (ii) competing for food at feed bins, (iii) a deer actively approached another and bit it, and more rarely, (iv) as a retaliation to an aggression. In general, biting behaviour was performed as a single bite, and the bitten deer reacted immediately by moving a few steps away, whereupon no more interactions took place. Normally, a bite produced a conspicuous hairless patch on the pelt of the bitten animal, as a consequence of hair removal by plucking action (i.e., hair removal from the root). In some cases, when the biting strength was light, no hair was apparently plucked. Due to these observations, we undoubtedly discard the theory that hairless patches on the pelt of our deer were a consequence of self- or allo-grooming, ecto-parasites, mycosis, or mineral deficiency. A single bite produces a conspicuous bald area, no larger than 5 cm2; multiple bites can overlap and produce larger bitten surfaces on the deer pelt. Just after biting, the exposed skin is pink-pale in colour and, after a few days, turns grey. The bald area can be easily identified even when hair is re-growing, as the new hair is lighter in colour in comparison with the surrounding hair (Figure 1).Pelt biting monitoring was performed in each deer once a week by visual inspection of the right side of the deer when they were handled for routine health, body weight, and condition assessment. The reason behind assessing pelt biting on the right side of the animal was for practical convenience, as this was the side visually exposed when the animal was in the handling pen. Pelt biting was carried out by the same observer (AJG) across the duration of the study. We categorised the extension of pelt biting into five ordinal main classes as follows (Figure 1). Class 1, no bites to very few bites, [0–3%] of pelt surface bitten (pelt surface is defined as the body surface with the exception of head, backstrap, lower parts of limbs, and lower parts of abdomen and genital area); class 2, few bites on shoulder, flank, lower rump, and upper and middle parts of the haunch, (3–10%] of pelt surface bitten; class 3, frequent bites on body parts described in class 2, bites start to appear on the neck, (10–40%] of the pelt surface bitten; class 4, abundant biting on body parts of classes 2 and 3, bites start to overlap, producing continuous bald areas, bites start to appear on the shanks of fore and rear limbs, (40–70%] of the pelt surface bitten; class 5, great extension of the surface of the pelt bitten, bites overlapping, abundant bites on shanks, >70% of the pelt surface bitten. These five main classes were further divided into ordinal quartiles categories (e.g., 2.25 = class 2 plus 25% of the area comprised between class 2 and class 3). The percentage of pelt surface bitten to produce this classification was calculated using ImageJ software [29] applied to red deer pictures taken on lateral view and representative of each of the pelt biting classes.2.3. Data and AnimalsData collection was carried out at the University of Castilla-La Mancha (UCLM) deer farm experimental facilities (38°57′32.8′′ N 1°52′51.8′′ W, Albacete, Spain) between 2006 and 2019. The climate was continental Mediterranean, cooler summers and greater variation in seasonal temperatures than the typical Mediterranean climate, bordering a cold semi-arid climate (annual mean min and max temperature = 5.9 °C December and 24.3 °C July; min–max rainfall = 12 mm July, 42 mm October; http://crea.uclm.es/siar/datmeteo/ (accessed on 21 April 2021). The study used 427 red deer females and 424 males (Table 2).Females’ age ranged between 1 and 21 years old (mean = 4.7, Q1 = 1.1, Q3 = 7.1) and males’ age between 1 and 15 (mean = 1.7, Q1 = 0.4, Q3 = 2.2). Mean body weight in females was 85.7 kg (Q1 = 71, Q3 = 105) and 93 kg in males (Q1 = 48, Q3 = 129). Females and males were split into 2–6 groups, depending on the number of deer in the farm; each group was allocated to different fields of size between 0.6 and 1.2 ha (mean density in fields = 25 deer/ha). Deer relied entirely on supplementary feed, as the amount of grass provided by the fields was negligible. The base diet year-round was a well-balanced mixture of chopped alfalfa hay and orange pulp, supplied ad libitum three times a week, and between March and October, this was supplemented with pelleted feed. Feed was presented to deer on both-side access 14 m long belt feeders to minimise aggressions during feeding [13]. Animals had free access to water at all times. Similarly, males were kept in separate groups, except during the rut, during which some stags were brought into the females’ groups for mating.On a weekly basis, deer were driven from the fields to a nearby handling facility, where they were weighed, their condition was monitored, and the number of bites on the pelt of each animal was classified as detailed in the previous sections. As a result, 40,159 animal monitoring events with information on pelt biting were achieved throughout the study (Table 2).Animals were daily attended by qualified personnel, and an expert deer veterinarian (AJG) looked after the animals on a weekly basis. The farm complied with Spanish animal welfare legislation, and the monitoring procedure did not require an animal experimental license.2.4. Hierarchy RankBetween April and October of 2017 and 2018, agonistic interactions between 36 adult hinds (2017: group 1 = 9 deer, group 2 = 16 deer; 2018: group 1 = 16 deer, group 2 = 19 deer) were recorded using direct observations to estimate the linear hierarchy among animals.Animal agonistic interactions (head butting, boxing, pelt biting, kicking, pushing, chasing, walking/running away, spatial displacement, visual threat) were recorded by one of the co-authors (ML-Q), with the aid of a pair of binoculars (8 × 42) and a telescope (20 × 60) from the top of a 4 m tall tower located at a vantage point in a plot where the hinds and their calves grazed. Observations took place between 08:00 h and 12:00 h, for a total of 521 h for 133 days, comprising 5067 animal interactions. Behavioural interactions were carried out by continuously scanning all animals (i.e., sampling as defined by Martin and Bateson [23]) and recording the type of behaviour, together with the identity of the pair of hinds involved and identifying which one was the aggressor and aggrieved. The frequency of interactions was generally low, which allowed the observer to record most of them, with some interactions missing during events of exceptional high activity.To calculate the hierarchical rank of our animals, we used the Combi1 index (Equation (1)), together with the algorithm I&SI that minimises inconsistencies and ties in the calculation of the rank, implemented in Domicalc software [30] Combi1 index = [Di/(Di + Si)] + Di − Si,(1) where, Di and Si are the number of domination and subdomination events, respectively, of individual i over the rest of individuals in the group.2.5. Heat Stress IndexWe used meteorological data from the Spanish Ministry of Agriculture, Food, and Environment, supplied by SIAR regional service of Castilla-La Mancha (available at http://crea.uclm.es/siar/datmeteo/ (accessed on 21 April 2021) of July and August for the period 2006–2019. Data came from the meteorological station of Albacete (38°56′56.5′′ N 1°53′53.3′′ W), 2 km from the UCLM experimental deer farm and located at the same altitude. We used daily mean records across the study period of air temperature, relative humidity, wind speed, and global solar radiation to produce an index of heat stress [31,32] that has been used efficiently to assess thermal stress on red deer in outdoor conditions [13], THIWS = 4.51 + 0.8 × T + 0.01 × Hr × (T − 14.4) + 46.4 − (1.992 × W) + 1.887 × SR,(2) where T is the daily mean temperature (°C), Hr is the mean relative humidity (%), W is the mean wind speed (m∙s−1), and SR is the accumulated solar radiation over a 24 h circadian period (MJ∙m−2).2.6. Statistical AnalysisAs an exploratory hypotheses testing approach, we used GAM models (Generalised additive mixed models), implemented in the “gam” function of the mgcv R package [33]. The model showed that the GAM smooth relationships were in fact generally quite simple and could be well-described by simple polynomial functions. Consequently, we used linear mixed models with polynomial functions equivalent to those obtained by GAM models, implemented in the package lme4 [34] in R software version 3.4.1 [35]. For model selection, we used p-values against measures based on information theory, such as ΔAIC or BIC [36], as the objective was to identify the main drivers of the dependent variables. We proceeded by first fitting full models that included the explanatory variables and the pertinent interactions and then reducing the terms of the model using backward elimination by removing the non-significant fixed-effects interactions, one at a time, following the principle of marginality: the highest order interactions were tested first, and if they were significant, then the lower order effects were not tested for significance. Significance of the terms in the model was assessed using the R function lmerTest [37], which approximates degrees of freedom via Satterthwaite’s method, as in linear mixed-effects models, degrees of freedom are difficult to define appropriately [38]. The variance explained by the linear mixed model was represented as R2 marginal (variance accounted for by the fixed effects R2LMM(m)) and R2 conditional (variance accounted for by random and fixed effects; R2LMM(c)), following a method developed for linear mixed-effects models [39]. Calf ID, hind ID, and year were fitted as crossed random effects in the models. In order to not over-parameterise the models, we limited the number of interactions terms fitted. Graphics were constructed using ggplot2 R package based on the grammar of graphics [40].3. Results3.1. Pelt Biting DescriptionThe distribution of pelt biting classes was clearly biased towards the lowest classes (median = 1.25, mean = 1.50, Figure 2). The maximum class recorded was 4.75; this was exceptional, only recorded once in three different animals. These animals had no hair on most of their body, except head, lower parts of limbs, and underneath parts of the body. Frequency of pelt biting classes <1.5 was higher in males than in females, while it was more likely to find females exhibiting pelt biting classes >1.5 (Figure 2). Pelt biting was more intense for the period 2006–2011 compared to period 2012–2019, coinciding with a reduction in the number of deer in the farm that took place in 2011 (Figure 3).3.2. Hypotheses TestingThere was a significant quadratic relationship between hierarchical rank and pelt biting in the observed group of 36 hinds; the higher the social rank, the lower the number of pelt bites after controlling for the effects of hind age and hind body weight (Table 3, Figure 4), which supported H1 (Table 1). Pelt biting responded to hind body weight and hind age, as predicted in hypotheses H2 and H3; the heavier and older the hind, the fewer bites on their coats, although the response was only significant for body weight (body weight coefficient = −0.005, SE = 0.0018, p = 0.009; age coefficient = −0.008, SE = 0.0044, p = 0.068, Table 3).The analysis of the pelt biting time series provided some support for hypotheses H2, H3, H4, and H6 after controlling for group size (Table 1, Table 4). The bigger the group, the larger the number of pelt bites exhibited on deer of both sexes (p < 0.001, Figure 5). We found a significant quadratic relationship between pelt biting and age in both sexes. The maximum number of pelt bites was observed in deer 5–6 years old; younger animals, but especially those in the oldest age classes, exhibited fewer pelt bites than 5–6 year-old deer (p < 0.001, Figure 5). The heavier the stag, the lower the number of bites on their coat; this pattern contrasted with that observed in hinds, which showed a peak in the number of pelt bites in hinds of average body weight, as compared to those lighter but especially to those heavier (p < 0.001, Figure 5). Hinds that gave birth earlier in the parturition period suffered less pelt biting than those that gave birth around the peak of the season (p < 0.001, Figure 5). It was clear that males and females suffered higher intensity of pelt biting as heat stress increased (H6) (p < 0.001, Figure 5). There was evidence that males suffered higher rates of pelt biting than females (H4) (male coefficient = 0.18, SE = 0.038, p < 0.001). Sackett et al.’s [27] hypothesis (H5) was not supported by our data; as a matter of fact, hinds carrying male foetuses were those that exhibited greater number of pelt bites than those carrying female foetuses, which was contrary to prediction (Table 5, Figure 6).4. DiscussionThe analyses clearly supported our hypotheses, except H5. On average, deer that were of higher hierarchical social rank (H1), heavier (H2), or older (H3) were bitten less by their peers than those that were of lower rank, lighter, or younger, respectively. Males were bitten more by same-sex peers than were females (H4), and biting behaviour increased in years of high heat stress (H6). All this being the case after controlling for group size, which had the expected significant effect, the larger the group, the higher the activity of pelt biting between its members. In addition, we found that hinds that gave birth earlier in the birth season suffered less biting than those giving birth at the peak of the season or later.These results are relevant for deer farming, as they can be used to identify situations of social stress and put in place measures to minimise it in order to improve animal welfare and probably production. One of the most obvious measurements to control social stress in animal farms is by minimising spatial crowding [17,41]. Spatial restrictions can cause animal discomfort, even if they do not impose serious deprivations or injury [42]. Furthermore, appropriate spatial space facilitates natural social interactions that are beneficial to animal welfare [43]. Animal density is dynamic over time, as clustering level changes depending on the aggregation pattern of the animals [44,45]. For example, aggregation increases at feeding points in farms (feed bins, feedlots) but also in the wild, at points where food is scant and spatially concentrated (snow craters grazing in reindeer [4]). In an observational experiment carried out between 2017 and 2018, in the same facilities of this study and under similar population conditions, we observed that pelt biting rates were the same at the feeding belts as away from them (number of pelt biting observations at feeding belt = 350, away from feeding belt = 346, unpublished data). This suggests that, in our case, space restrictions impose conditions favourable to develop pelt biting behaviour. In contrast, in a nearby deer farm in which densities varied between 1.5–2.5 deer/ha and grazing was the main food resource, pelt biting was also observed, but it was an uncommon behaviour, as only a few individuals displayed bites on their pelt (Lagunes farm, Ciudad Real, Spain, unpublished data). In both farms, cases of extreme pelt biting were observed to be directed towards particular individuals of a group, and in one case, towards a hind and its 1-year-old calf, especially when individual animals were incorporated into an existing group. Similarly, in dairy goats, some individuals display consistent aggressive/submissive behaviour that identify them [46]. This points out that pelt biting can be useful to identify individuals that are being bullied and apply corrective measures consequently (e.g., move the bullied animal to a new group structure; identify and remove the bully animal from the group). In deer hunting ranches, where density reaches up to 0.4 deer/ha and food supplementation takes place when graze is scarce, pelt biting is infrequent, based on observations over four thousand legally culled red deer to which we had access to take samples for research purposes over several years [26,47], and we never found a case in which pelt biting exceeded class 1.7. This is expected, as free-ranging conditions hardly create favourable conditions for bullying; population density is generally low, sufficient space precludes oversized groups, and graze is widely dispersed and therefore difficult to defend and of low energy reward value per food item [48], although there are exceptions [4,22,49].Hierarchical Rank, Age, Body Weight, Date of Birth and Heat StressWe predicted that animals with greater competitive abilities in agonistic encounters (those that were of higher hierarchical rank, heavier, and older) should receive less biting. Social rank has been found to be positively related to age and body weight in adult males and females of ungulate species [6,19,50,51,52]. On average, animals that are of higher hierarchical rank, heavier, or older are less likely to be challenged [53], and so are less bitten, than animals that are of lower rank, lighter, or younger. In contrast, animals at the bottom of the hierarchy might be aware of their limited competitive capabilities, and so respond by avoiding encounters that might result on being bitten when space is lacking. In a review, Miranda de la Lama and Mattiello [54] describe that goats attain their position in the social linear hierarchy at a very young age and tend to maintain this for a long time, independent of changes in the physical environment, although the introduction of new members to herds can increase aggression and, on occasion, alter the social hierarchy of the group. In goats, rank is very important for gaining access to resources; high-ranking goats gain access to more food than do low-ranking individuals [54].Studies on wild red deer have shown that low-rank individuals leave their feeding station when peers of higher rank approach [55]; hinds that do not retaliate in fighting received less severe agonistic interactions than those that retaliate and lose [52]; in agonistic interactions between stags, 9% involved low-ranking individuals, 33% intermediate-ranking, and 58% top-ranking animals [26,52,56]. This could explain the quadratic effects between pelt biting and hierarchical rank, body weight, and age found in our hinds (Figure 4 and Figure 5). Namely, animals at the bottom of the hierarchy were bitten less, but those animals who had an average social rank, body weight, and age, therefore with some chances of winning, were keen to get involved in agonistic encounters to keep or increase their hierarchical rank, thus making them more likely to be bitten. This pattern was significantly more pronounced in females than in males for variable body weight, which could be explained by the fact that, in polygynous species, the reward of reproductive fitness is higher in males than in females [57]; thus, males are more prone than females to be involved in agonistic encounters and so be injured [26,52].Caution is needed when assessing sexual differences in pelt biting in farmed deer, because they are probably affected by management. For example, in our setting, and in many deer farms, antlers are cut when they become hard to protect deer and staff personnel [58]. Males may have an average number of bites greater than females, because with no antlers, they are not able to antler fight and must resort to other agonistic behaviours, such as biting. On the other hand, males may have a lower number of bites than females due to receiving a more static management of social grouping than females. This is a consequence of males being kept in smaller and more stable groups, as their main use is mating in large harems [59]. On the other hand, female groups are more dynamic, as their management requires group reorganisation depending on breeding stage, age, and sales, which might lead to increasing agonistic behaviour to re-establish hierarchical relationships when group composition changes. This is supported by observations in wild red deer, where the number of escalated fights was greater when the animals meet unfamiliar conspecifics compared to the number of escalated interactions within same group members [52]. Furthermore, in reindeer, the reorganisation of groups led to increased fighting in order to establish a new hierarchy [60].Our results on hypothesis H5 were contrary to the prediction [27]. It is not clear why our females carrying male foetuses suffered greater biting than those females with female foetuses, unless carrying one sex changes the aggressiveness of the mother, making her more likely to challenge higher-ranking mothers and thus suffer long-term retaliation. This result may be indirect evidence that deer can detect the sex of pregnant hinds, as was suggested in pigtail monkeys [27].Heat stress increased the number of pelt bites in both sexes; this adds to further evidence that heat stress affects their behaviour and growth [13,61], and it should be taken into account as an important stressor, especially in farm conditions, when there are limited opportunities for heat abatement.In seasonal breeders, such as red deer, timing calving to the period that concentrates vegetation growth is an advantageous strategy to meet the high nutritional demands of lactation and related reproductive fitness traits [62,63]. Births of male calves, the most costly sex, tend to be earlier than births of female calves (Cervus dama [64]; Cervus elaphus [51]). Early births in farmed red deer increases milk yield, milk energy, milk fat content, and calf growth and reduces body weight losses of hinds during lactation [65,66]. Early births are generally related with mothers being in good body condition [67]; this can be achieved by earlier exposure to spring green-up, being more efficient or competitive at grazing during the months before births, which might imply they were high rank animals that facilitated access for food resources [4,63,68]. We found that females that produced early births were less bitten than females giving birth later in the season; this suggests that they were high rank animals, which supports the findings of the previous studies commented on above.5. ConclusionsPelt biting responded to individual animal and social traits and could be useful as a conspicuous visual index to assess animal welfare in farm conditions. Because it is affected by a number of factors that make comparisons difficult between populations living under different conditions, it should be used as a monitoring tool to detect how social environment and management affect animal welfare within a population and to enable an according response with corrective measures. There is room to consider integrating pelt biting into automated image analysis systems, which are already being used to assess animal welfare, e.g., heat–cold stress, limping, and growth [69,70]. The value of our pelt biting index in wild populations is, however, limited, as pelt biting activity is expected to be low in the wild, but it should not be discarded as a diagnostic index in the animal welfare toolbox of game keepers and deer population managers.

animals : an open access journal from mdpi

10.3390/ani13101618

PMC10215599

Understanding the mechanisms of heat stress is increasingly important due to global warming. However, these mechanisms are poorly understood. Here, we aimed to investigate the mechanisms underlying heat stress in rats. We identified key genes associated with rectal temperature and adrenal levels of dopamine, norepinephrine, epinephrine, and corticosterone in heat-stressed rats. In particular, methyltransferase 3, poly(ADP-ribose) polymerase 2, and zinc finger protein 36-like 1 were found to be associated with the heat stress response. These genes may be candidate genes involved in the regulation of heat stress. Our findings provide new insights into the molecular mechanisms driving heat stress in rats and might help guide future research into heat stress in mammals.

Heat stress has been a big challenge for animal survival and health due to global warming. However, the molecular processes driving heat stress response were unclear. In this study, we exposed the control group rats (n = 5) at 22 °C and the other three heat stress groups (five rats in each group) at 42 °C lasting 30, 60, and 120 min, separately. We performed RNA sequencing in the adrenal glands and liver and detected the levels of hormones related to heat stress in the adrenal gland, liver, and blood tissues. Weighted gene co-expression network analysis (WGCNA) was also performed. Results showed that rectal temperature and adrenal corticosterone levels were significantly negatively related to genes in the black module, which was significantly enriched in thermogenesis and RNA metabolism. The genes in the green-yellow module were strongly positively associated with rectal temperature and dopamine, norepinephrine, epinephrine, and corticosterone levels in the adrenal glands and were enriched in transcriptional regulatory activities under stress. Finally, 17 and 13 key genes in the black and green-yellow modules were identified, respectively, and shared common patterns of changes. Methyltransferase 3 (Mettl3), poly(ADP-ribose) polymerase 2 (Parp2), and zinc finger protein 36-like 1 (Zfp36l1) occupied pivotal positions in the protein–protein interaction network and were involved in a number of heat stress-related processes. Therefore, Parp2, Mettl3, and Zfp36l1 could be considered candidate genes for heat stress regulation. Our findings shed new light on the molecular processes underpinning heat stress.

1. IntroductionHeat stress is defined as the accumulation of the body’s non-specific responses to a high-temperature environment [1]. Heat stress response relies on the regulatory activities of the hypothalamic–pituitary–adrenal (HPA) axis, hypothalamic–pituitary–thyroid (HPT) axis, and hypothalamus–pituitary–gonadal (HPG) axis, which helps the body defend against heat stress stimulation [2]. Previous studies have found that acute or chronic heat stress can alter physiological and biochemical signs at the cellular, systemic, and organismal levels in animals [3,4], involving neuroendocrine regulation [5], oxidative stress responses [6,7], and basic metabolism [8]. However, the molecular mechanisms underlying heat stress remain unclear. Heat stress has become a big concern as global temperatures rise, endangering animal survival and health. Therefore, there is an urgent need to illustrate the regulatory mechanism underlying the heat stress reaction. The development of next-generation sequencing technology [9] has made it possible to identify heat stress-responsive genes and transcripts more accurately and rapidly at the transcriptome level. This technology facilitates in-depth analysis of the molecular mechanism of heat stress [10,11].The liver and adrenal glands are critical tissues for heat stress studies. The liver is a crucial metabolic organ that regulates numerous biological processes in response to heat stress [12]. The adrenal glands are involved in the creation and release of several hormones associated with heat stress, such as glucocorticoids, which maintain equilibrium in the body [13]. Furthermore, heat stress causes transcriptional, metabolic, and protein alterations in the adrenal glands [13] and liver [12,14] by stimulating the HPA, HPT, and HPG axes. Activation of the HPA axis is particularly essential. In response to stressors, secretagogues are released into the anterior pituitary, prompting the secretion of adrenocorticotropic hormone. This activates adrenocortical cells, which then manufacture and release glucocorticoids, such as corticosterone. The corticosterone level is a typical metric used for assessing heat stress reactions. Heat stress can also stimulate the sympathetic-adrenal medulla system and enhance catecholamine release, causing the levels of catechol hormones such as adrenaline, norepinephrine, and dopamine to rapidly rise. Masaki et al. [15] found that norepinephrine and epinephrine concentrations in human plasma increase in response to heat stress.To investigate potential genetic markers associated with heat stress response, the gene expression data from adrenal glands and liver tissues [9] along with 14 phenotypes were used to perform weighted gene co-expression network analysis (WGCNA) [16]. The 14 phenotypes included rectal temperature (Tc), levels of corticosterone (adrenal_CORT), dopamine (adrenal_DA), epinephrine (adrenal_E), and norepinephrine (adrenal_NE) in the adrenal glands, as well as levels of corticosterone (liver_CORT), dopamine (liver_DA), epinephrine (liver_E), and norepinephrine (liver_NE) in the liver. We also examined catalase (blood_CAT), lactic acid (blood_LA), adrenocorticotropic hormones (blood_ACTH), growth hormones (blood_GH), and prolactin (blood_PRL) in the blood tissues. We screened out functional modules related to heat stress by WGCNA analysis and combined functional enrichment analysis, protein–protein interaction (PPI) networks, and short time-series expression miner (STEM) analysis, as well as conducted a phenome-wide association study (Phe-WAS) to identify key genes in heat stress response.2. Materials and Methods2.1. Data SourcesThe data used in this study were derived from our previously published data [9]. Twenty rats were used to build thermal animal models under various environmental conditions. Five rats were housed at 22 ± 1 °C (control group, CT; n = 5), and three groups of five rats were placed under three different environmental conditions: 42 °C for 30 min (H30, n = 5), 60 min (H60, n = 5), or 120 min (H120, n = 5). In a preliminary experiment, we assessed thermosensitivity in different tissues of rats treated for varying lengths of heat stress time (30 min, 60 min, and 120 min) using receiver operating characteristic (ROC) analysis. We found that the area under the curve (AUC) was greater than 0.80 for different heat stress times, indicating that the rats were in a state of heat stress at these time points. However, it is possible that the rats experienced different degrees of heat stress at different times. Therefore, we chose these three time points for our experiment. The Tc of rats was detected using the electronic clinical thermometer with a precision of ±0.1 °C (MC-347, Omron Corporation, Kyoto, Japan). The adrenal glands, liver, and blood samples were collected from each rat. Biochemical indicators in the adrenal glands, liver, and blood tissues were examined, and gene expression levels in the adrenal glands and liver tissues were analyzed. Briefly, the concentrations of dopamine, noradrenaline, epinephrine, and corticosterone in the adrenal glands, liver, and blood tissues were determined using enzyme-linked immunosorbent assays. The Trizol method was used to extract RNA from the adrenal glands and liver, and RNA quality was evaluated via NanoDrop 2000, as described by Dou et al. [17]. On the Illumina® HiSeq 2000 platform (San Diego, CA, USA), an RNA sequencing (RNA-seq) library was built and sequenced, yielding 150 bp paired-end reads. Each sample’s gene expression data were normalized using the reads per kilobase of exon model per million mapped reads (FPKM) method. Differential expression analyses were performed using a t-test. Genes in the liver and adrenal glands that met the criteria p ≤ 0.01, false discovery rate (FDR)-adjusted p = 0.05, and |fold change (FC)| > 2 were regarded as differentially expressed genes (DEGs) [18]. The RNA-seq datasets were released on the Sequence Read Archive at the National Center for Biotechnology Information (BioProject accession number PRJNA690189).2.2. WGCNAWGCNA (v1.12.0), implemented in the R program, was used to construct a gene co-expression network for all genes identified in the adrenal glands and liver tissues [19]. The expression levels of all genes in the 40 samples, including 20 adrenal gland samples (n = 5 for each treatment group: CT, H30, H60, H120) and 20 liver samples (n = 5 for each treatment group: CT, H30, H60, H120), were calculated using Pearson’s correlation matrices. The formula amn=cmnβ was used to establish a weighted adjacency matrix, where amn is the adjacency between genes m and n, cmn is Pearson’s correlation coefficient, and β is the soft-power threshold [19]. To evaluate the connectivity of each gene in the network, the weighted adjacency matrix was transformed into a topological overlap measure (TOM) matrix. A clustering dendrogram of the TOM matrix was created using the average hierarchical clustering technique. Genes with comparable patterns of expression were clustered into the same modules using the dynamic hybrid cutting method. Furthermore, using Pearson’s correlation matrices, we correlated the eigengenes of the modules (the first principle component of the corresponding expression matrix) with 13 bioindicators including adrenal_CORT, adrenal_DA, adrenal_E, adrenal_NE, liver_CORT, liver_DA, liver_E, liver_NE, blood_CAT, blood_LA, blood_ACTH, blood_GH, blood_PRL, and Tc. We choose the modules with an absolute value of correlation greater than 0.80 and a significance level less than 0.05 as functional modules. Considering that Tc is a simple and effective measure of heat stress, we subsequently selected two modules (black modules in adrenal glands and green–yellow modules in liver tissues) that were significantly related to Tc for further analysis. The visualization was completed by the WGCNA package in R software, including Figure 1A,B and Figure 2A,B.2.3. Functional Enrichment Analysis of Functional Modules and Mining of Key GenesGenes in each module were submitted to gene ontology (GO) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis using the “clusterProfiler” in the R package to discover the biological activities and signaling pathways associated with each module [20]. The genes with the top 0.1% intramodular connectivity in significant modules were regarded as hub genes. In functional modules, genes that were identified as both hub genes and DEGs were considered key genes. Figure 1C and Figure 2C were generated using clusterProfiler and ggplot2 packages in R software. Figure 3D,E was generated by using the ggplot2 package in R-4.2.2 software, showing key genes and significant Go terms that key genes involved.2.4. Short Time-Series Expression Miner (STEM) Analysis of Key GenesSTEM v1.3.13 software [21] was used to identify key genes exhibiting the same expression trend. Because there are just a few time points per dataset, the STEM clustering method selects a set of distinct and typical timed expression profiles, known as model profiles. Each gene was assigned the closest model profile by using the coefficient of correlation. Next, the computed p-value based on a hypergeometric distribution was applied to identify which model profiles were assigned significantly more genes. The clustering parameters were set as follows: a maximum of 50 model profiles, a maximum unit change between time points of 2, and a minimum correlation for clustering comparable profiles of >0.7 [22]. Figure 3A,B was built with the STEM v1.3.13 program.2.5. Downstream Bioinformatics AnalysesTo explore functions of the significant clustered genes identified via STEM analysis, we performed protein–protein interaction (PPI) network analysis (STRING, https://string-db.org/ (accessed on 15 February 2023) and a phenome-wide association study (Phe-WAS, https://atlas.ctglab.nl/ (accessed on 6 December 2022). The PPI network was investigated using four types of evidence (experimental, text mining, co-expression, and databases). The Phe-WAS was a study strategy of 3302 human phenotypes designed to discover associations between a particular SNP or gene and a wide range of traits. This strategy has proved to be effective in both recovering previously discovered genotype–phenotype relationships and finding new ones [23,24]. Figure 3C,F is drawn using STRING and Figure 4A–F is drawn using R.3. Results3.1. Summary of RNA-Seq DataIn total, 40 RNA-seq datasets (20 RNA-seq datasets from the adrenal and 20 RNA-seq from the liver) were obtained. After quality control and alignment analyses, 5881 and 8472 genes were identified in the adrenal glands and liver tissues, respectively [9]. In the adrenal glands (CT vs. H30, CT vs. H60, and CT vs. H120) and liver tissues (CT vs. H30, CT vs. H60, and CT vs. H120), 1501 and 1310 DEGs were identified, respectively, based on a threshold of p < 0.05, FDR-adjusted p = 0.05, and |FC| > 2.3.2. Gene Co-Expression Modules Associated with Phenotypic TraitsBy employing WGCNA, nine gene modules were identified in the adrenal glands samples, including eight co-expression modules and one module containing the remaining uncorrelated genes (Figure 1A,B). Six modules were identified in the liver samples, including five co-expression modules and one module containing the remaining independent genes (Figure 2A,B). In adrenal glands, the black module was significantly (p < 0.05) negatively correlated with rectal temperature and adrenal_CORT, with correlations of −0.9 and −0.78, respectively. In liver tissues, the green-yellow module was significantly (p < 0.05) positively correlated with adrenal_DA, adrenal_E, adrenal_NE, adrenal_CORT, and Tc, with correlations of 0.74, 0.78, 0.83, 0.69, and 0.77, respectively.3.3. Functional Enrichment AnalysisNext, we investigated the biological function of the genes in the black and green-yellow modules. The GO terms and KEGG pathways with the top 10 black module counts are shown in Figure 1C and Figure 2C. GO enrichment analysis showed that genes in the black module were significantly involved in ncRNA metabolic processes (GO:0034660), ribonucleoprotein complex biogenesis (GO:0022613), ncRNA processing (GO:0034470), mRNA processing (GO:0006397), RNA splicing (GO:0008380), ribosome biogenesis (GO:0042254), Rrna metabolic processing (GO:0016072), Rrna processing (GO:0006364), tRNA metabolic process (GO:0006399), and mitochondrial gene expression (GO:0140053) in biological process (BP); mitochondral matrix (GO:0005759), organelle inner membrane (GO:0019866), mitochondrial protein complex (GO:0098798), nuclear speck (GO:0016607), ubiquitin ligase complex (GO:0000151), spliceosomal complex (GO:0005681), organellar ribosome (GO:0000313), mitochondrial ribosome (GO:0005761), and organellar large ribosomal subunit (GO:0000315) in cellular component (CC); and catelytic activity acting on RNA (GO:0140098), transcription coregulator activity (GO:0003712), coenzyme binding (GO:0050662), transferase activity, transferring one-carbon groups (GO:0016741), ligase activity (GO:0016874), catalytic activity, acting on a tRNA (GO:0140101), helicase activity (GO:0004386), single-stranded DNA binding (GO:0003697), aminoacyl-tRNA ligase activity (GO:0004812), and ligase activity forming carbon-oxygen bonds (GO:0016875) in molecular function (MF). The KEGG pathway analyses confirmed that genes in the black module were significantly enriched in thermogenesis (rno04714), RNA transport (rno03013), biosynthesis of cofactors (rno01240), spliceosome (rno03040), carbon metabolism (rno01200), peroxisome (rno04146), RNA degradation (rno03018), ribosome biogenesis in eukaryotes (rno03008), nucleotide excision repair (rno03420), and fatty acid metabolism (rno01212). GO enrichment analysis showed that genes in the green-yellow module were significantly involved in posttranscriptional regulation of gene expression(GO:0010608), covalent chromatin modification (GO:0016569), organic cyclic compound catabolic process (GO:1901361), RNA catabolic process (GO:0006401), mRNA catabolic process (GO:0006402), response to endoplasmic reticulum stress (GO:0034976), regulation of DNA-templated transcription in response to stress (GO:0043620), regulation of transcription from RNA plolymerase 2 promoter in response to stress (GO:0043618), bile acid and bile salt transport (GO:0015721), negative regulation of transcription from RNA polymerase 2 promoter in response to stress (GO:0097201), and ubiquitin-binding (GO:0043130). In the KEGG pathway analysis, no significant pathway was found.3.4. Identification of Hub Genes in the Functional Modules and Mining of Key Genes Associated with Heat StressOne pattern (profile 9), containing 17 key genes, was significantly enriched (p < 0.05) in the adrenal glands. Another pattern (profile 49), containing 13 key genes, was significantly enriched (p < 0.05) in the liver.Seventeen key genes from the black module were considerably enriched in profile 9, and 13 key genes from the green-yellow module were strongly enriched in profile 49, as shown in Figure 3A,B. This indicates that key genes in the same module exhibited the same pattern of expression changes, suggesting that the module was accurately identified. In Figure 3D,E, we picked out significant Go teams in which the key gene was involved. Mettl3 was related to RNA splicing and metabolism, methylation, and histone modification, Parp2 was related to DNA repair, and Zfp36l1 was involved in endoplasmic reticulum stress and transcriptional regulation under stress. In our view, these processes are all related to the presence of heat stress. Then, we performed protein–protein interaction analysis of gene sets containing key genes from the black and green-yellow modules, as shown in Figure 3C,F. In the protein–protein interaction networks, Mettl3 and Parp2 from the black module and Zfp36l1 from the green-yellow module had central regulatory roles. Thus, we selected these three genes for subsequent analysis and in-depth discussion.3.5. Phe-WAS of Key Genes Associated with Heat Stress in HumansTo further annotate the functions of the candidate genes in mammals, we performed Phe-WAS of human orthologs of Parp2, Mettl3, and Zfp36l1 across 3302 human phenotypes (https://atlas.ctglab.nl/ (accessed on 6 December 2022)). The levels of expression of Parp2 and Mettl3 were significantly downregulated in response to heat stress, whereas Zfp36l1 was significantly upregulated (Figure 4A,C,E). Parp2, Mettl3, and Zfp36l1 were significantly associated with immunity, endocrine function, and metabolism (Figure 4B,D,F).4. DiscussionHeat stress is a well-known phenomenon that triggers a range of bodily responses, including alterations in hormone concentrations such as dopamine, norepinephrine, epinephrine, and cortisol. Dopamine, norepinephrine, and epinephrine are all catecholamines that play significant roles in the body’s response to heat stress. Gruntenko et al. found that dopamine levels significantly rise when Drosophila suffers from heat stress [25]. Alvarez et al. also found that the norepinephrine and epinephrine levels in the blood significantly increase when cattle are exposed to heat conditions [26]. This result is consistent with the present results (attached Figure 1, Figure 2 and Figure 3). Lyte et al. suggested that catecholamine has been potentially used as an indicator of heat stress [27]. Catecholamines can increase blood flow to the skin and stimulate sweat production [28], which can facilitate heat loss and help maintain a normal core body temperature [29]. The catecholamines induce diverse physiological responses across different organs, such as reduced visceral function and digestive inhibition, enhanced cerebral blood flow, improved pulmonary gas exchange efficiency, the breakdown of glycogen to release glucose reserves, vasodilation in muscles, and elevated heart rate [30]. Although these processes can assist in resisting heat stress, they may also trigger an overactive sympathetic nervous system, which can lead to increased thermogenesis. As a result, the exact mechanism by which catecholamines operate during heat stress remains unclear.Cortisol, an important glucocorticoid, has been shown to play the most important role in the process of heat stress in animals. When the animals suffered from heat stress, the HPA axis gets activated, which consequently increases plasma glucocorticoid concentrations in blood [31]. In the current results, cortisol showed a significant upward trend after heat stress (attached Figure 4). Cortisol has multiple functions that help animals tolerate stress. They act as vasodilators, aiding in heat dissipation and increasing blood flow [32]. Additionally, they stimulate proteolysis and lipolysis [32], which supply energy to the animal when food intake is reduced. These physiological adjustments enable the animal to cope with stress more effectively. Following exposure to heat stress, our study observed an increase in levels of dopamine, norepinephrine, epinephrine, and cortisol in the adrenal glands, while no similar increase was observed in liver tissues. With the increase in heat stress time, these hormones first increased and then decreased to the control level. This result may be related to the heat stress period. Chan et al. found that chronic heat stress does not increase dopamine, epinephrine, and norepinephrine levels in blood in rats, but is similar to that of the control group [33]. For cortisol, in response to acute heat stress, the activation of adrenocorticotrophin release in the hypothalamus is responsible for the initial increase in cortisol [34]. The later the return to normal, despite the continued heat shock, shows another reaction, most likely negative cortisol feedback [35].To investigate the genetic and biological underpinnings of heat stress, we employed rats as model animals and implemented WCGNA. In our study, we chose two modules (black modules in the adrenal glands and green-yellow modules in the liver tissues) as functional modules. In the adrenal glands, the black module was significantly (p < 0.05) negatively correlated with Tc and adrenal_CORT, with correlations of 0.9 and −0.78, respectively. In the liver tissues, the green-yellow module was significantly (p < 0.05) positively correlated with adrenal_DA, adrenal_E, adrenal_NE, adrenal_CORT, and Tc, with correlations of 0.74, 0.78, 0.83, 0.69, and 0.77, respectively. However, in the enrichment analysis results, we could not uncover any pathways associated with the manufacture and secretion of the above hormones, indicating that the above functional modules may not be directly related to hormone secretion and activity. The process of heat stress is intricate; hormonal regulation and the function of modules may both take place at the same time. For instance, the black module affects thermogenesis, whereas the green-yellow module participates in endoplasmic reticulum stress. The top 0.1% of genes in the black and green-yellow modules were identified as hub genes, and the differentially expressed hub genes were defined as key genes. In the black module, 17 key genes were discovered, and in the green-yellow module, 13 key genes were discovered. Using STEM to cluster the key genes of the two modules, it was discovered that key genes in the same module exhibited comparable expression changes in response to changes in the duration of heat stress, indicating that the modules of key genes exhibited a consistent pattern of change. In the protein–protein interaction network, Parp2, Mettl3, and Zfp36l1 were located in the hub region. Heat stress is known to have many effects on cells, including DNA damage [36], RNA splicing [10,37], and modification [38]. GO enrichment analysis of the black and yellow-green modules revealed that Parp2 was involved in DNA repair, Mettl3 was mainly involved in post-transcriptional regulation, such as RNA splicing modification and catabolism, and Zfp36l1 was involved in transcriptional regulation during stress (Figure 3D,E).As we see in the results of the enrichment analysis (Figure 3D), Parp2 is involved in DNA repair. Parp2 catalyzes the synthesis of poly(ADP-ribose) (PAR) using NAD+ as a substrate [36,37], recognizes damaged DNA, and synthesizes long, branched PAR chains covalently attached either to themselves or to acceptor proteins that activate the base excision repair machinery [39,40]. Ame et al. [41] found that purified recombinant mouse Parp2 acts as a damaged DNA-binding protein in vitro and catalyzes the formation of PAR polymers in a DNA-dependent manner. Parp2 is mainly detected at a single DNA nick site and exhibits a low level of binding to undamaged DNA and double-strand breaks, according to Maria et al. [42]. In plants, stressors such as heat, light, and drought activate PARP, causing NAD+ breakdown and ATP consumption [43]. At present, many studies have demonstrated that heat stress could cause cell DNA damage [44,45]. Therefore, we initially thought that Parp2 could be involved in heat stress regulation by mediating DNA damage repair during heat stress. However, when we looked at the expression of this gene in response to heat stress, we found that this may not be the case. Our results discovered a substantial drop in Parp2 gene expression in rats after heat stress (Figure 4A). Heat stress may cause the inhibition of DNA damage repair by reducing Parp2 expression, rather than cells reducing heat stress-induced damage through Parp2-mediated DNA damage repair. The current studies have shown that heat stress can inhibit key components of virtually all repair systems, including the base excision repair system [46,47,48] and nucleotide excision repair system [49,50]. So how does Parp2 play a role in the regulation of heat stress processes? Some recent studies may shed some light on this, such as DNA methylation. Tomer et al. [51] found chicks injected intraperitoneally with Parp inhibitors showed lower body temperature, improved response to chronic heat stress, and exhibited resilience to heat stress. They found Parp inhibitors could reduce systemic DNA methylation by inhibiting the activity of methyltransferase, especially by reducing the methylation of corticotropin-releasing hormone (CRH) introns, resulting in decreased CRH expression. These results are similar to ours, and the module in which Parp2 is located also shows significant correlations with body temperature and cortisol in the adrenal glands. The second study shows the improvement of energy utilization efficiency and the reduction in oxidative stress. Marc et al. [43] also found that Parp inhibitors could reduce NAD+ breakdown and consequently energy consumption, oxidative stress, and promote plant regeneration and repair in heat stress. This evidence shows that when exposed to heat stress, reducing the expression of the Parp2 gene can help the body resist such stress.Mettl3 is the catalytic component of the N6-adenosine-methyltransferase complex, which is involved in multiple processes, including RNA metabolism, gametogenesis, and DNA damage repair. Mettl3 may also be involved in post-transcriptional regulation mechanisms, such as RNA truncation and modification, according to the results of our GO enrichment analysis (Figure 3D). In recent years, heat stress has been linked to RNA m6A methylation. Zhou et al. [38] discovered that nuclear YTHDF2 preserves the 5’UTR methylation of stress-induced transcripts, increases 5’UTR methylation in the form of m6A, and promotes cap-independent translation initiation, providing an alternative translation mechanism for selective *mRNA translation under heat shock stress. Mettl3 as the m6A “writers” is closely related to RNA m6A methylation. Yu et al. found that Mettl3 expression was significantly downregulated and YTHDF2 was significantly upregulated due to heat stress, and Mettl3 knockdown resulted in significant upregulation of a variety of heat shock proteins, including HSPA1b, HSPA9, and HSPB1 [52]. Our results also prove that the expression of the Mettl3 gene exhibited a downward trend after suffering from heat stress. Therefore, our understanding is that Mettl3 has the potential to modulate heat stress protein expression through m6A, which could impact the heat stress response. In addition, heat stress causes changes in the activity of various antioxidant enzymes in the body, such as catalase, glutathione peroxidase, and oxide dismutase, disrupting the body’s oxidant and antioxidant balance. This results in oxidative stress, the overproduction of reactive oxygen and nitrogen species, and damage to cellular DNA. Mettl3 is involved in the repair of UV-induced DNA damage via the RNA m6A modification pathway, according to Yang et al. [53]. In Mettl3-knockout cells, UV-induced cyclobutane pyrimidine adduct repair is delayed, and UV sensitivity is enhanced.Zfp36l1 is a zinc lipoprotein that plays a role in the degradation of nuclear-transcribed mRNA. In our research, Zfp36l1 was related to response to endoplasmic reticulum stress (GO:0034976) and regulation of DNA-templated transcription in response to stress (GO:0043620), according to the results of the GO enrichment analysis (Figure 3E). When heat stress occurs, it is accompanied by a response to endoplasmic reticulum stress through oxidative stress, resulting in the accumulation of unfolded or misfolded proteins [54]. This buildup initiates a signaling cascade known as the unfolded protein response (UPR), which is a cellular defense mechanism meant to restore ER equilibrium. The UPR response to heat stress includes the overexpression of heat shock proteins, which aid in the refolding of denatured proteins, as well as the triggering of death if the cell damage is too severe to repair (apoptosis). In reaction to oxidative stress, polo-like protein kinase 3 (Plk3) has been postulated to trigger apoptosis. For example, Deng et al. [55] discovered that plk3 played a role in oxidative stress-induced DNA damage and apoptosis, and they found that overexpressing plk3 reduced cell vitality and increased apoptosis, whereas silencing plk3 weakened the apoptotic response. Meanwhile, in another study, Zfp36l1 was able to degrade the mRNA of plk3 through the AU-rich elements in the 3’UTR region [56]. In our research, the expression of the Zfp36l1 gene exhibited an upward trend after suffering from heat stress. Therefore, we hypothesize that overexpressed Zfp36l1 may participate in oxidative stress and endoplasmic reticulum stress by degrading plk3, reducing autophagy under heat stress. In addition, in human research, the orthologs of Zfp36l1 and Zfp36l1 have been shown to influence variation in HSPA1A and HSPA1b (two important heat stress proteins) expression [57]. This may be another important pathway for the Zfp36l1 gene to participate in heat stress.We propose that Parp2, Mettl3, and Zfp36l1 are candidate genes involved in the regulation of heat stress. However, this requires validation in future functional studies.5. ConclusionsIn this study, we performed WGCNA of RNA-seq data from adrenal glands and liver tissues to detect key genes and pathways involved in the heat stress response in rats. The black modules in the adrenal glands and green-yellow modules in the liver tissues were significantly correlated with phenotypes under heat stress, including Tc and hormone levels in the adrenal glands. Based on intramodular connectivity, differential expression analysis, PPI, and enrichment analysis, we believe that three key genes (Parp2, Mettl3, Zfp36l1) play an important regulatory role in the response to heat stress. Our results provide new insights into the molecular mechanisms underlying heat stress and demonstrate that integrative analyses of omics data are promising strategies for illustrating the genetic architecture underlying complex traits and diseases.

animals : an open access journal from mdpi

10.3390/ani11123436

PMC8698006

Improvements in dairy cattle feed efficiency have substantial effects on economic efficiency and can reduce environmental impacts through lower feeding costs and fewer emissions associated with dairy farming. The efficiency of an animal for converting feed into products is influenced by genetic, physiological, and environmental factors that result in individual variations. The utilization of feed efficiency indexes aims to identify and select animals with great economic value in a production system. Associations between morphometric indicators, hormone concentrations, and blood parameters may assist in the identification of differences in the efficiency of feed utilization and in understanding the physiological bases linked to animals’ metabolic responses, thus helping to identify more efficient animals. In our study, it is unlikely that measurements of blood, rumen, or morphometric indicators, per se, will be useful in the early identification of more efficient animals. Understanding the underlying physiological basis for improved feed efficiency in dairy heifers requires further investigation.

The objectives of this study were: (1) to evaluate feed efficiency indexes and their relationships with body measurements and blood and ruminal metabolites in the pre-weaning period; (2) to determine if such measurements can be used as feed-efficiency markers during the pre-weaning period. Holstein–Gyr heifer calves (n = 36), enrolled between 4 and 12 weeks of age, were classified into two residual feed intake (RFI) and residual body weight gain (RG) groups: high efficiency (HE; RFI, n = 10; and RG, n = 9), and low efficiency (LE; RFI, n = 10; and RG, n = 8). Calves were fed whole milk (6 L/day) and solid feed ad libitum. Body developments were measured weekly and feed intake (milk and solid feed) daily during the whole period. Blood samples were collected at 12 weeks of age and analyzed for glucose, insulin and β-hydroxybutyrate (BHB). Samples of ruminal content were collected on the same day and analyzed for pH, NH3-N, and volatile fatty acids (VFA). Among the growth characteristics, only the initial hip width differed between the RFI groups, and withers height differed between the RG groups. Concentration of BHB was greater and glucose: insulin ratios tended to be greater in LE-RG animals. Butyric acid proportions were similar among RFI groups, but tended to be greater for HE-RG than for LE-RG. Overall, correlation coefficients between RFI or RG and blood, rumen, or morphometric markers were low. Thus, it is unlikely that measurements of metabolic indicators, per se, will be useful in the early identification of more efficient animals. Understanding the underlying physiological basis for improved feed efficiency in dairy heifers requires further investigation.

1. IntroductionImprovements in dairy cattle feed efficiency (FE) have substantial effects on economic efficiency and the reduction of environmental impacts through lower feeding costs and less emissions associated with dairy farming. To increase the biological and economic efficiencies of cattle production, animal selection strategies need to focus on improving FE without compromising performance.A 5% improvement in FE has an economic impact eight times greater than a 5% increase in average daily gain (ADG) [1]. In addition, breeding for improved residual feed intake (RFI) can enhance feed efficiency without increasing the animal’s mature size [2]. This has obvious positive ramifications for the improvement of the FE of growing and adult cattle. However, there is limited published information on phenotypic RFI and the residual body weight gain (RG) of dairy heifers, as well as on its potential impacts on milk production and the biological factors contributing to variation in these traits.Residual feed intake is the most-used index of FE [3], and it is defined as the difference between realized and predicted intake, using a linear regression of individual intake as a function of mean metabolic body weight (BW0.75) and ADG. This index is independent of growth rate and body weight (BW). Another measurement of efficiency is RG, which is similar to RFI, but instead of regressing feed intake on BW and ADG, as for RFI, RG is obtained from the regression of ADG on feed intake and BW [4].Since variability in RFI has been acknowledged, there has been abundant research assessing its underlying physiological mechanisms [5,6]. Evidence shows that no single physiological mechanism is responsible for the observed variability [5]. Theoretically, every physiological step that affects the conversion of feed gross energy to animal products could be associated with the observed variability in RFI.Associations between hormone concentrations and RFI have been studied [7,8], and blood parameters may assist in the identification of differences in the efficiency of feed utilization and in understanding the physiological bases linked to the animal’s metabolic response, thus helping to identify more efficient animals. Circulating the blood metabolites in growing heifers, for example, is responsible for contributing approximately 35% of the variation in RFI [7]. Likewise, studies have been done to evaluate differences in ruminal parameters between high and low RFI animals and their possible role as FE markers. However, there is limited published information on the repeatability of RFI and its associated traits at different phases of the production cycle, which is ultimately essential for wide-scale adoption by producers.Therefore, the objectives of this study were: (1) to evaluate FE indexes and their relationships with body measurements, blood, and ruminal metabolites in the pre-weaning period; (2) to determine if such measurements can be used as FE markers during the pre-weaning period. Our hypothesis was that there are differences in RFI and RG that explain better feed utilization by some animals, and that such measurements can be used as feed-efficiency markers.2. Materials and MethodsThe experiment was conducted at the Embrapa Dairy Cattle Experimental Farm, located in Coronel Pacheco, Minas Gerais, Brazil.2.1. Animals, Housing, Management, and Feed Efficiency GroupsThe data presented in this paper is part of a study that classified Holstein × Gyr F1 crossbred heifer calves in high and low efficiency groups using 2 FE indexes (RFI and RG). Detailed descriptions of the methods, performance data, calculation of indexes, and how the animals were classified in high and low efficiency groups are provided in the manuscript [9]. Briefly, 36 heifer calves (BW at birth = 32.4 ± 6.6 kg, mean ± SD) were removed from their dams immediately after birth, weighed, and had their umbilical cords immersed in 10% iodine solution. Colostrum was fed within 6 h of birth (10% of birth BW; >50 g of IgG/L). Passive immunity transfer was assessed using total serum protein. Blood samples were collected via jugular venipuncture within 48 h of birth and centrifuged at 1800× g for 10 min at room temperature (22–25 °C), and total serum protein was measured using a refractometer (Serum protein REF-301, Biocotek, Beilun, Ningbo, China).Heifers were housed in sand-bedded individual pens (1.25 × 1.75 m, tethered with 1.2-m long chains), which were allocated in a barn with open sides and end-walls. Transition milk was fed until 3 days of age, followed by whole milk thereafter until an abrupt weaning at 82 days of age. The volume of milk supplied during the pre-weaning period was fixed at 6 L of milk/day, divided into 2 equal meals of 3 L of milk each, at 0700 and 1400 h. Water and solid feed (starter and chopped hay) were offered in buckets for ad libitum intake (10% orts of solid feed) throughout the experimental period. Solid feed (as-fed) was composed of 95% starter (Soylac Rumen 20% Flocculated, Total Alimentos, Três Corações, Brazil) and 5% chopped Tifton 85 hay (Table 1).Milk and solid feed were measured between 4 and 12 weeks of age. Feed was weighed using a bench scale (9094 plus, Toledo®, São Bernardo do Campo, São Paulo, Brazil.Solid intake was calculated by the difference between offers and refusals measured daily. Daily milk intake was calculated as the sum of the differences between offered and refused amounts at morning and afternoon feedings. Heifers were weighted at birth, at 3 days of age (i.e., when the experimental diet was first offered), and weekly thereafter, always before the morning milk feeding.The residual feed intake and residual gain were calculated over 56 days of observation. Intake and performance were evaluated from 25 days to 80 days of age. The total DMI was obtained from the sum of milk DMI and solid feed DMI (offered amount minus refusals on DM basis).The average daily gain was calculated as the linear regression coefficient of BW (PROC REG; SAS Inst. Inc., Cary, NC, USA), composed of nine BW measurements per heifer at 7 days intervals. Metabolic body weight (BW0.75) was calculated using the BW at day 23 of the test. Feed efficiency was measured using the relationship between mean daily total DMI and ADG [10].Linear regressions were used to estimate RFI and RG [3], where RFI and RG were calculated as the difference between realized and predicted total DMI and ADG, respectively, as follows:Yj = β0 + β1(BW0.75j) + β2(ADGj or total DMIj) + ej, where:Yj = is the standardized total DMI (RFI) or ADG (RG) of heifer j;β0 = is the intercept, β1 is the regression coefficient for BW0.75;β2 = is the regression coefficient for ADG (RFI) or total DMI (RG);ej = is the error term for heifer j.Heifers were classified into two RFI and RG groups: high efficiency (HE; RFI < 0.5 SD below the mean (n = 10) and RG > 0.5 SD above the mean (n = 9)), and low efficiency (LE; RFI > 0.5 SD above the mean (n = 10) and RG < 0.5 SD below the mean (n = 8)) (Figure 1). The remaining animals were classified as intermediate and were not included in subsequent analyses.2.2. Morphometric MeasurementsMorphometric measurements were performed weekly (between weeks 4 and 12) before morning milk feeding and after weighing, in a flat location that allowed the animals to remain with their limbs well-set. Withers height (the distance from the withers to the ground) and hip height (the distance from ileosacral tuberosity to the ground) were measured using a hipometer (Walmur, Porto Alegre, Brazil). Hip width (the distance between the two iliac tuberosities), and heart girth (measured immediately and caudally to the front limbs) were evaluated with a measuring tape (Bovitec, São Paulo, Brazil). The variation of each body measurement was calculated as the difference between the final (12th week) and initial (4th week) values.2.3. Blood Collection and AnalysesBlood samples were obtained by jugular puncture 3 h after the morning milk feeding and on the same day of ruminal fluid collection (12th week), during the RFI and RG evaluation period. Samples collected into untreated tubes were used for the dosage of insulin and BHB, and those from sodium fluoride-treated tubes were used to determine glucose concentration (Vacutainer, Becton Dickinson, São Paulo, Brazil). Tubes were placed in crushed ice until centrifugation at 1800× g for 10 min at room temperature (22–25 °C). Plasma or serum aliquots were stored at −20 °C until analysis.Plasma glucose was measured on a microplate Spectrophotometer EON (Biotek Instruments Inc., Winooski, VT) using an enzymatic colorimetric method (Kovalent do Brasil Ltd.a., Rio de Janeiro, Brazil). Serum insulin was analyzed using a bovine ELISA kit (Mercodia, Uppsala, Sweden) and serum BHB was determined using an enzymatic kinetic kit RANBUT—Ref.: RB 1007 (RANDOX Laboratories—Life Sciences Ltd., Crumlin, UK) and spectrophotometry (Automatic System for Biochemistry, Model BIOPLUS BIO 2000®, Bioplus Produtos para Laboratórios Ltd.a, Barueri, Brazil).2.4. Rumen Variables and AnalysesRuminal fluid samples were obtained at the end of the experimental period (12th week). Samples of approximately 50 mL were collected 3 h after milk feeding using a stomach tube technique [11,12,13]. The liquid was double-filtered through cheesecloth, and the pH was measured immediately after collection using a portable potentiometer (DM-2-Digimed, São Paulo, Brazil).Rumen content samples (5 mL) were acidified with 1 mL of sulfuric acid (500 mL/L) and stored at −20 °C until the analysis of ruminal NH3-N concentration, which was quantified after the distillation of Kjeldahl with magnesium oxide and calcium chloride, according to method 920.03 [14]. Samples were centrifuged at 1800× g for 10 min at room temperature (22–25 °C) for the measurement of volatile fatty acids (VFA) concentration by high-performance liquid chromatography (Waters Alliance e2695 Chromatograph, Waters Technologies do Brazil LTDA, Barueri, Brazil).2.5. Statistical AnalysesData were analyzed using SAS version 9.0 (SAS Institute Inc., Cary, NC, USA). In all models, the normality and homoscedasticity of the standardized residues were evaluated graphically and using the Shapiro–Wilk and Bartlett tests, respectively.Total DMI, water intake, and performance were analyzed using the linear mixed model (PROC MIXED), including the calf as the random component of the model, and the feed efficiency group, week, and their interaction as the fixed components (efficiency group × week interaction). The birth weight and total serum protein were used as covariates and included in the fixed-effect model.Variables with a single measurement during the study (blood, rumen, and morphometric variables) were analyzed including the calf as the random term, and the efficiency group as a fixed variable (PROC GLM).The PROC CORR procedure was used to assess the correlations between the response variables and RFI and RG. For all the variables analyzed, significance was declared when p ≤ 0.05, and a tendency declared when 0.05 > p ≤ 0.10.3. ResultsThe residual feed intake ranged from −0.14 to 0.13 kg/day (p < 0.01) for HE and LE, respectively, equating to a difference of 0.27 kg of DMI/day between the groups ranked high and low RFI. The residual weight gain ranged from 0.05 to −0.07 (p < 0.01), equating to a difference of 0.12 kg of ADG between the groups ranked high and low efficiency for RG.At the start of the experimental period (week 4), BW ± SD was 46.78 ± 5.81 kg in the RFI test and 45.36 ± 6.98 kg in the RG test. At the end of experimental period (week 12), BW was 101.68 ± 12.89 kg and 99.49 ± 12.68 kg for RFI and RG tests, respectively. The effect of phenotypic classification by RFI and RG on feed intake, feed efficiency, and performance was evaluated in previous work [9], and the data is summarized in Table 2.Differences were not detected between groups in the RFI test for withers height, hip height, and heart girth (p > 0.05; Table 3), and between groups in the RG test for hip height, hip width, and heart girth (p > 0.05; Table 3). Among the RFI groups, HE-RFI had greater initial hip width than LE-RFI (22.3 and 20.6 cm, respectively; p = 0.03). Moreover, the HE-RG group had greater variation in withers height than LE- RG (15.2 and 12.7 cm, respectively; p = 0.01).Morphometric measurements was not correlated with the feed efficiency indexes (p > 0.05; Table 4).No significant differences in glucose, insulin, BHB concentrations, and glucose:insulin ratios were observed among the groups of RFI (Table 5). For the RG tests, differences were detected between BHB concentrations (p = 0.01) and tended for the glucose:insulin ratios (p = 0.07). The LE-RG group had greater BHB concentrations than the HE-RG group (0.38 and 0.28 mmol/L, respectively), and a tendency for a higher glucose:insulin ratio (1.69 and 0.83 mg/μU, respectively). There were no significant correlations between blood variables (glucose, insulin, glucose:insulin ratio, and BHB) and RFI, total DMI, and feed efficiency. Significant correlations were observed between blood insulin and glucose concentrations for the RFI (r = 0.46; p = 0.05) and RG groups (r = 0.61; p = 0.02), respectively. There were negative correlations of −0.54 (p = 0.03) between RG and BHB and −0.68 (p = 0.01) between RG and the glucose:insulin ratio.Residual feed intake and RG had no effect (p > 0.05) on rumen fermentation (Table 6), except for NH3-N and butyrate concentration. The HE-RG group had greater NH3-N concentrations than LE-RG (10.42 and 6.51 mg/dL, respectively), as well as higher butyrate concentrations (3.99 and 3.05 µmol/mL, respectively). Rumen fermentation variables were not correlated with the feed efficiency indexes (p > 0.05).4. DiscussionThis study investigated feed efficiency indexes and their relationships with body measurements, blood, and ruminal metabolites during the pre-weaning period, as well as whether such measurements can be used as feed efficiency markers during the pre-weaning period.The average daily gain and morphometric measurements were similar between HE and LE calves during the RFI and RG test period, similar to those reported by another study [15]. Others have indicated that morphological traits do not differ according to RFI ranking, consequently not altering the morphometric pattern of the animals selected for RFI [7,16,17]. The variation in withers height was greater for HE-RG than for LE-RG, demonstrating a greater growth in the former group. This agrees with results reported by another study [4], which reported high correlations between RG and body growth rate (r = 0.70).In the present study, glucose concentration was not different among RFI-divergent animals. A recent study with pre-weaning dairy heifers also did not observe differences in glucose or insulin between HE and LE RFI groups [18]. In addition, a study with growing Nellore cattle also did not observe differences in glucose concentration between RFI groups, but HE-RFI animals had a lower glucose:insulin ratio and higher insulin concentrations than LE-RFI animals [8]. In this work, the greater satiety of HE-RFI animals was associated with a high concentration of insulin in the blood.In the present experiment, HE-RG heifers tended to have lower glucose:insulin ratios, indicating a greater secretion of insulin per unit of blood glucose by these animals. Insulin is an important hormonal regulator of the metabolism and an inhibitor of hepatic gluconeogenesis that reduces the hepatic absorption of some glucose precursors, directing the flow of glycogenic nutrients to muscle and adipose tissues [19]. Greater insulin concentrations promote protein and lipid synthesis and body weight gain, which may explain the greater withers height variation in the HE-RG group.Heifers classified as HE-RG tended to have a lower BHB concentration, but this metabolite was not different among RFI groups. In our study, total DMI did not differ among HE-RG (1535 g/day) and LE-RG (1594 g/day) [9], but HE-RG tended to have greater ruminal butyrate concentrations. In young animals, BHB is produced when rumen butyrate is metabolized by the rumen epithelial cells. The high rumen pH observed in the HE-RG group (6.11) may have impaired passive diffusion across the rumen epithelial membrane, resulting in a lower butyrate absorption rate and greater accumulation in the rumen, resulting in a lower concentration of BHB.There is evidence of differences in rumen digestion between RFI-divergent cattle [20]. However, this was not observed in the present experiment. In accordance with that, research conducted on beef bulls under a high-concentrate diet (rolled barley, 860 g/kg DM) reported no differences in rumen pH and VFA proportions for divergent RFI groups [21]. On the other hand, a previous study using a diet comprised purely of grass silage [22] reported that HE-RFI cattle tended to have greater ruminal propionate concentrations and lower acetate:propionate ratios. This agrees with results reported by other studies [11,23], also feeding high-fiber diets. A lower acetate:propionate ratio in HE-RFI cattle is consistent with greater energy efficiency and lower methane production. It seems, therefore, that differences in the rumen fermentation profile are evident in high-fiber diets but not in high-concentrate diets, such as those fed to pre-weaning calves.The concentration of NH3-N was greater in HE-RG calves, in comparison to LE-RG calves, indicating a better rumen fermentation efficiency. In this group, rumen ammonia levels were similar to those observed by previous studies [24]. On the other hand, the LE-RG group presented values below those considered optimal, suggesting a lower ruminal efficiency. This better efficiency in rumen fermentation may have resulted in an increase in microbial protein production and a consequent increase in the flow of metabolizable protein to the intestine. Pre-weaning calves are largely dependent on intestinal digestion for nutrient absorption, and this appears to be the factor responsible for the difference in RG between groups in this current experiment.In this study, we evaluated the feed efficiency divergence in pre-weaning dairy heifers. In order to produce results that are useful for the dairy industry, we designed an experiment following the feeding practices adopted in commercial dairies (i.e., set amount of milk and ad libitum starter). Although this practice represents common rearing conditions, it limits the understanding of the separate effects of milk and starter. Therefore, it is not possible to determine if the variation observed for certain variables results from milk or starter intake, nor is it possible to study the interactions among them. However, as the milk supply was fixed at 6 L/day for all animals during the trial, the total DMI variation among heifers would be related to the difference in solid feed intake.5. ConclusionsThe RFI and RG groups had no negative effects on relevant metabolic and, morphometric measurements or metabolic and performance characteristics. Consequently, such feed-efficiency divergence tests are applicable to the selection of more efficient pre-weaned heifers. However, there was no evidence of strong associations between the blood or rumen metabolites and the RFI and RG tests. In this study, the measurements of metabolic indicators, per se, were not useful in the early identification of more efficient animals. Understanding the underlying physiological basis for improved feed efficiency in dairy heifers requires further investigation.

animals : an open access journal from mdpi

10.3390/ani13050954

PMC10000206

Probiotics have been considered as alternatives to the antibiotics currently used to control diseases caused by different microorganisms, which show high prevalence and losses in aquaculture. In this study, the possible utility of three probiotics (Enterococcus gallinarum L1, Vagococcus fluvialis L21 and Lactobacillus plantarum CLFP3—effective against vibriosis or lactococosis in sea bass or rainbow trout) was investigated for the biocontrol of saprolegniosis in rainbow trout. For this purpose, both in vitro inhibition studies and competition for binding sites against Saprolegnia parasitica and in vivo tests with experimentally infected rainbow trout were carried out. Although the three probiotics showed inhibitory capacity and reduced the adhesion activity of S. parasitica cysts to cutaneous mucus in vitro, none of the three bacteria showed in vivo protection through either water or feed. The obtained results show the importance of selecting the most appropriate probiotic and its mechanism of action depending on the species of fish and the disease to be prevented.

Previous studies have demonstrated that the strains Enterococcus gallinarum L1, Vagococcus fluvialis L21 and Lactobacillus plantarum CLFP3 are probiotics against vibriosis or lactococosis in sea bass or rainbow trout. In this study, the utility of these bacterial strains in the control of saprolegniosis was evaluated. For this purpose, both in vitro inhibition studies and competition for binding sites against Saprolegnia parasitica and in vivo tests with experimentally infected rainbow trout were carried out. In the in vitro tests, the three isolates showed inhibitory activity upon mycelium growth and cyst germination and reduced the adhesion of cysts to cutaneous mucus; however, this effect depended on the number of bacteria used and the incubation time. In the in vivo test, the bacteria were administered orally at 108 CFU g−1 in the feed or at 106 CFU ml−1 in the tank water for 14 days. None of the three bacteria showed protection against S. parasitica infection either through water or feed, and the cumulative mortality reached 100% within 14 days post infection. The obtained results show that the use of an effective probiotic against a certain disease in a host may not be effective against another pathogen or in another host and that the results obtained in vitro may not always predict the effects when used in vivo.

1. IntroductionSaprolegniosis is a fungal disease, the main pathogenic agent of which is Saprolegnia parasitica, that can cause heavy economic losses in inland fish farms due to depletion of fish stock and eggs [1]. This disease affects freshwater fish at any stage of development and is characterized by skin lesions such as patches with a cotton wool-like appearance. Current prevention and treatment measures for saprolegniosis rely on the use of chemicals products that may cause damage to the environment [2,3] and humans [4]. In addition, none of the available compounds offers enough protection after the rearing period [5]. Recently, metal-based nanoparticles have attracted attention as a potential material for prevention and control of saprolegniosis [6,7]. Among the biological control alternatives to chemotherapy is the use of plant extracts [8,9,10,11,12] or probiotics [13,14].In this sense, a number of in vitro studies have evidenced that some bacterial strains can inhibit Saprolegnia spp., including Pseudomonas fluorescens [15,16,17,18,19], Alteromonas sp., Pseudomonas alcaligenes, Pseudomonas saccharophila, Aeromonas caviae and Aeromonas eucrenophila [20], Serratia marcescens [21], Bacillus subtilis [22], Aeromonas media [23], Aeromonas sobria, Pantoea agglomerans, Serratia fonticola, Xhantomonas reflexus and Yersenia kristensenii [17,18], Lactobacillus plantarum [24] and Pseudomonas aeruginosa [25]. However, their action in vivo has only been studied in eel Anguilla australis [13], silver perch Bidyanus bidyanus [14] and eggs of rainbow trout Oncorhynchus mykiss [26] or eggs of gourami Osphonemus gouramy [27].In previous studies by our research group, various bacterial isolates with an in vitro ability to inhibit the growth of S. parasitica were obtained [17]. Their harmless nature with respect to rainbow trout (Oncorhynchus mykiss) and their ability to adhere the cutaneous mucus and reduce the adhesion of zoospores and cysts of S. parasitica were also demonstrated [18]. It was also found that of the fifteen investigated isolates, under experimental conditions, two isolates of Pseudomonas fluorescens reduced infection by S. parasitica in rainbow trout when these bacteria were added to the water in the tanks [28], and the mode of action of these bacteria was likely associated with the production of siderophores [29].Enterococcus gallinarum L1 (obtained from Dicentrarchus labrax gut content) and Vagococcus fluvialis L21 (obtained from Solea solea gut content) demonstrated high levels of protection against Vibrio anguillarum in European sea bass (Dicentrarchus labrax) after an experimental challenge when these bacteria were added to feed for 20 days [30,31]. In the same way, Lactobacillus plantarum CLFP-3 (obtained from O. mykiss cutaneous mucus) added to the diet of rainbow trout for 36 days was demonstrated to improve protection against Lactococcus garvieae [32]. In the present study, the in vitro inhibitory activity of these probiotic bacteria against S. parasitica was determined, and in vivo tests with experimentally infected rainbow trout were carried out in order to investigate their potential use for the biocontrol of saprolegniosis.2. Materials and Methods2.1. Collection of Isolates Used in the StudyThree probiotic bacteria (Enterococcus gallinarum L1, Vagococcus fluvialis L21 and Lactobacillus plantarum CLFP3) previously obtained by our research group that were stored at −80 °C in our laboratory facilities were used in this study (Table 1). The first two have shown their efficacy against infection by Vibrio aguillarum in sea bass (Dicentrarchus labrax), and L. plantarum reduced mortality in rainbow trout (O. mykiss) infected with Lactococcus garvieae.After thawing, the bacterial isolates E gallinarum L1 and V. fluvialis L21 were cultured overnight at 20 °C in 3 mL of brain–heart infusion broth (BHIB, Pronadisa, Condalab, Madrid, Spain), while L. plantarum CLFP3 was cultured in Man, Rogosa and Sharpe broth (MRS, Pronadisa). An aliquot of 500 μL of the culture was inoculated into 25 mL of BHIB or MRS broth and incubated at 20 °C and 200 rpm on a rotary orbital shaker (Innova® 44, New Brunswick Scientific, Edison, NJ, USA) until the middle of the exponential growth phase (based on previously established growth curves). Bacteria were pelleted and washed twice by centrifuging at 1000× g for 15 min and resuspended in sterile saline solution. The number of bacteria was adjusted to the desired concentration according to the experiment in which they were used (see below) by counting in a hemocytometer chamber (Improved Neubauer Type, Albert Sass, Germany).The strain Saprolegnia parasitica TRU12 isolated from a wild brown trout (Salmo trutta) with saprolegniosis [33] and maintained refrigerated in our laboratory was used. It was cultured for 3 d at 20 °C on glucose peptone (GP) agar; then, autoclaved half hemp seeds of Cannabis sativa were placed on the edges of the colony, followed by an additional 24 h of incubation at 20 °C. To obtain zoospores, the half hemp seeds colonized by hyphae were placed in Petri dishes with filtered autoclaved river water. After 36–48 h at 20 °C, the water was filtered through Whatman 541 cellulose filter paper. The number of zoospores was then estimated using a Hawksley Cristalite BS 748 counting chamber.2.2. In Vitro Inhibition Assays of the Probiotic Bacteria against S. parasiticaTo determine the in vitro inhibition of the probiotic bacteria strains against S. parasitica, different assays were performed. All these experiments were conducted following the procedures described in a previous study [17].2.2.1. Inhibition of Hyphal Growth on Solid Media (Plate Assay)This test was performed in with BHI agar (Cultimed, Panreac Química SLU, Barcelona, Spain), whereby each bacterium was streaked twice. This assay was performed in triplicate with incubations at 3, 5 or 7 d at 20 °C. Then, a 3 mm diameter block of agar with young hyphal tips of S. parasitica was placed between the 2 bacterial streaks and incubated for 3 d at 20° C, measuring the diameter of the S. parasitica colony. A negative control plate was used without bacterial streaks under the same conditions.2.2.2. Inhibition of Hyphal Growth from Colonized Hemp Seeds and Cyst Germination in Liquid MediaBacterial dilutions, hemp seeds colonized by Saprolegnia and zoospore suspensions were prepared following the methods described above. The bacterial concentration was adjusted to 2 × 105 cells mL−1, and, taking this as the first dilution, four 10-fold dilutions of bacterial suspension were prepared. The concentration of zoospores was adjusted to 4 × 104 zoospores mL−1. For the inhibition of hyphal growth, duplicate wells of a 24-well tissue culture plate (Falcon, Corning, Glendale, AZ, USA) filled with 1 mL of each bacterial dilution, 1 mL of BHIB and a half hemp seed colonized by S. parasitica were dispensed into each well.For the inhibition of zoospores/cyst germination, the same procedure was followed, with the only difference being that 0.5 mL of the zoospore suspension (4 × 104 zoospores mL−1) was added in place of a colonized hemp seed, and 0.5 mL of each bacterial dilution was used instead of 1 mL.In both assays, plates were incubated for 3 d at 20 °C, and bacteria and S. parasitica zoospore negative controls were included on each plate. The presence or absence of macroscopic or microscopic hyphal growth and germination of cysts were observed by a Nikon Diaphot inverted microscope.2.2.3. Assay to Evaluate Fungicidal EffectsThe initial step was the same as that for the hemp seed test, but the plates were placed in an incubator at 20 °C for 1, 2 or 3 d. Following each incubation period, the hemp seeds from the wells where the growth of S. parasitica was inhibited were removed, washed 3 times with sterile distilled water and transferred to a Petri dish containing 20 mL of filtered and autoclaved river water with streptomycin (2 × 102 mg L−1) and penicillin (2 × 105 UI L−1). The plates were incubated for 20 d at 20 °C. To avoid false-positive results due to inhibitory effects of viable bacteria in the mycelium, only the results with no bacterial growth in water samples collected after 24 h were accepted. A negative control was achieved by incubating a hemp seed colonized by S. parasitica in water both with and without antibiotics. The bacteria were considered to have a fungicidal effect if no mycelial growth was observed after 20 d.2.2.4. Inhibition Assays with Bacterial Culture SupernatantsThe supernatants used to perform this assay were prepared according to Lategan et al. [34] using BHIB for L1 and L21 or MRS for CLFP3 strains. This supernatant and three 2-fold serial dilutions were tested to analyze the ability of the supernatants to inhibit hyphal growth (hemp-seed test) and to evaluate the capacity of the supernatants to inhibit cyst germination (cyst test). In the hemp-seed test, 2 mL of each supernatant dilution and a half hemp seed colonized by S. parasitica were dispensed in duplicate into each well of a 24-well tissue culture plate. In the cyst test, 1 mL of each supernatant dilution and 1 mL of the zoospore suspension (4 × 104 zoospores ml−1) were added to each well. Incubation times and observation procedures were the same as those described in Section 2.2.2. Negative controls with only culture medium were included in both assays.2.3. Adhesion to Skin Mucus of the Probiotic Bacteria and Inhibition of S. parasitica Cyst AdhesionThe adhesion capacity of the probiotic strains L1, L21 and CLFP3 to cutaneous mucus of trout and its potential to hinder the adhesion of S. parasitica cysts under conditions of exclusion, competition and displacement were investigated using the methods described by Carbajal-González et al. [18].Cutaneous mucus was obtained from four male brown trout (Salmo trutta) with an average body weight of 196.7 ± 39.5 g from a hatchery belong to the Regional Government of Castile and Leon by scraping the surface with a plastic spatula. The mucus was centrifuged twice at 12 000× g for 5 min at 4 °C to eliminate particulate and cellular debris. The protein concentration was adjusted to 0.05 mg ml−1 in PBS using Bradford reagent (Sigma-Aldrich, Merck Life Science, Madrid, Spain). The resulting mucus suspension was finally sterilized by UV light exposure for 30 min and stored in aliquots at −20 °C until use.Bacteria and cysts of S. parasitica were stained with Syto 9® green fluorescent nucleic acid stain (Invitrogen, Fisher Scientific, Madrid, Spain) at a concentration of 1 μL per 109 bacteria in 1 mL and 1 μL per 105 cysts in 1 mL.2.3.1. Adhesion to Skin Mucus of the Probiotic BacteriaThis assay was performed in 96-well black polystyrene plates (Costar, ImmunoChemistry Technologies, Davis, CA, USA) coated with cutaneous mucus. In the adhesion assay with the probiotic strains, wells without coating or coated with BSA (0.05 mg ml−1, Sigma-Aldrich) or mucin from swine stomach type II (MSS; 0.05 mg ml−1, Sigma-Aldrich) were used as controls for non-specific adhesion. To coat the plates, 25 μL of mucus, BSA or mucin and 75 μL of coating buffer (16.8 g sodium hydrogen carbonate, 21.2 g sodium carbonate per liter, pH 9.6) were added to each well and left overnight at 4 °C. Subsequently, the wells were washed with PBS-Tween 20 (0.1%), the fluorescently labelled bacterial solution (25 μL of 109 bacterial cells ml−1) was added and plates were centrifuged at 163× g for 12 s to promote the adherence of bacteria to the mucus. Bacterial fluorescence was measured with a spectrophotometer (Synergy HT Multi-Detection Microplate Reader, Bio-Tek®, Winooski, VT, USA) at 485 nm excitation and 535 nm emission after 30 min of incubation in darkness and at room temperature (time 0). Non-attached bacteria were removed by washing with saline solution (50 µL/well), and plates were placed upside down on absorbent paper and centrifuged to remove any remaining liquid. Then, 50 µL of saline solution was added per well, and a new measurement was performed. Adhesion was expressed as the percentage of fluorescence recovered in the second measurement relative to the fluorescence at time 0.2.3.2. Inhibition of S. parasitica Cyst AdhesionSeven serial 10-fold dilutions were performed from non-stained bacteria (2 × 109 cells mL−1) in saline solution. The cysts were stained using the method mentioned previously using a concentration of 105 cysts mL−1. In the exclusion test, each dilution of the bacterial isolates was added to the wells, and after being incubated and washed, the labelled cysts were added. In the competition test, both bacterial dilution and stained cysts were added simultaneously, and in the displacement test, the labelled cyst suspension was first added, incubated and washed; then, each dilution of the bacterial suspension was added. Plates were washed with saline solution after 60 min of incubation at room temperature for each condition; then, the fluorescence was recorded. Wells with saline solution instead of bacterial cells were used as controls in this assay. The percentage of cysts bound to these control wells was considered the reference value (100%), and the percentage of cyst adhesion in the presence of bacteria was compared to this reference value.2.3.3. Statistical AnalysisIn the adhesion assays, statistical analyses were carried out using SPSS for Windows version 26 (IBM SPSS Statistic, Armonk, NY, USA) by the Student’s t-test (p ≤ 0.05). Data were shown as the mean ± standard deviation of three separate trials.2.4. Pathogenicity for Rainbow TroutThe probiotic bacterium CLFP3 (L. plantarum), which has previously been used successfully to prevent lactococcosis in trout [32], was obtained from rainbow trout. However, the L1 (E. gallinarum) and L21 (V. fluvialis) bacteria were obtained from marine fish, and although they were previously reported to be non-pathogenic for sea bass [30,31], their potential pathogenicity for rainbow trout is unknown. Therefore, prior to being used in in vivo tests, experimental inoculations with L1 and L21 isolates were conducted to verify their lack of pathogenicity for rainbow trout.Rainbow trout (O. mykiss) from a commercial fish farm with an average body weight of 37.24 ± 7.56 g were acclimatized for 10 days in a 120 L tank and observed daily to ensure that they did not show clinical signs of disease. Trials were conducted using groups of 20 fish kept in 40 L tanks filled with chlorine-free well water with a renewal rate of 1.5 L per h at 12 °C with constant aeration and a photoperiod of 12/12 h. Effluent water was disinfected with ozone. The experimental protocol was approved by the Subcommittee for Experimentation and Animal Welfare of the University of Leon, Spain (protocol number ULE_04_2015).The bacterial isolates were grown in BHIB as described in Section 2.1 to obtain a suspension of 107 cells mL−1. Fish were anaesthetized with tricaine methanesulfonate (MS-222, 50 mg mL−1) and injected with 0.1 mL intraperitoneally (ip) and intramuscularly (im) in two separate groups of 20 trout each. Control groups of 20 trout were inoculated with 0.1 mL of saline solution at the same time points. Fish were observed for 10 days and euthanized by overdosing them with MS-222 (100 mg mL−1) at the end of the observation period. They were subsequently examined for evidence of disease (i.e., gross lesions). Additionally, loopfuls of material from the kidney, liver and spleen were spread over plates of BHI agar with incubation at 20 °C for 3 days to determine the presence or absence of the inoculated bacterial isolates in the fish. The recovered bacterial isolates were examined by means of Gram-stained smears and basic biochemical tests such as catalase and oxidase tests.Temperature and other physicochemical parameters of water were measured daily, and no significant differences were observed between the different groups (temperature, 12.47 ± 0.24 °C; pH 7.92 ± 0.06; dissolved oxygen, 9.72 ± 0.16 mg L−1; nitrites (NO2), 0.034 ± 0.005 mg L−1; and non-ionized ammonia (NH3), 0.059 ± 0.005 mg L−1).2.5. Biocontrol of Saprolegniosis by Adding Bacteria to Water or Feed2.5.1. Biocontrol by Adding Bacteria to WaterFor each of the tested bacteria, 60 rainbow trout with an average body weight of 28.79 ± 10.01 g in three 40 L tanks were used: 20 rainbow trout for treatment with the probiotic and 2 as controls for infection and ami momi treatment, respectively. Fish were infected with S. parasitica TRU12 following the method described previously by Fregeneda-Grandes et al. [33]. Briefly, the fish were slightly scarified on the skin using a variation of ami momi treatment [35]. Groups of 10 fish were shaken for 2 min in a cylindrical stainless-steel colander with a mesh of 7 mm before adding a suspension of S. parasitica zoospores to the water to obtain a final concentration of 3 × 102 spores mL−1 in the tank water. At the same time, a bacterial suspension of 106 cells ml−1 in PBS was added. During the next 48 h, the flow of water was halted and then resumed. The bacterial treatment was repeated every 24 h for 14 days, halting the flow of water for 6 h. The bacterial suspension was replaced by the same volume of sterile PBS in the infection control tank, while neither bacteria nor zoospores were added to the tank to control for death caused by ami momi treatment. The fish were monitored under observation for 14 days to detect the emergence of signs of saprolegniosis.2.5.2. Biocontrol by Adding Bacteria to FeedTo conduct these experiments, three groups of 20 rainbow trout with an average body weight of 29.29 ± 10.33 g were maintained in 40 L tanks: one tank for the bacterial treatment and two tanks as controls. Fish were fed for 14 days with T2 Optiline 1P (Skretting®, Burgos, Spain) feed containing 108 bacteria g−1 feed, with a daily administration of 2% of tank biomass, followed by a fast of 48 h and 24 h before and after infection with S. parasitica, respectively. After infection, the feed with the bacteria continued to be administered for another 10 days. To prepare the feed with the probiotics, 10 mL of a bacterial suspension with a concentration of 109 cells mL−1 was added to 90 g of feed, and the mixture was homogenized with an electric hand mixer (HM 3100, Braun GmbH, Germany) for 1 min. Subsequently, the feed was allowed to dry for 30 min at room temperature in a laminar flow cabinet. Infection with S. parasitica zoospores was performed as described above, and water flow was halted for the subsequent 24 h. The fish were maintained under observation to detect the emergence of signs of saprolegniosis.The origin of the fish, their acclimation, experimental conditions and the quality of the physicochemical parameters of the water during these biocontrol tests were similar to those detailed in Section 2.4.2.5.3. Statistical AnalysisSurvival analysis was performed by the Kaplan–Meier method, and the survival probability of the groups treated or not with the probiotics was compared using the log-rank test, considering that the differences were statistically significant at p ≤ 0.05. All calculations were performed using SPSS for Windows version 26.3. Results3.1. In Vitro Inhibition Assays of the Probiotic Bacteria against S. parasitica3.1.1. Assays with solid mediumThe results of the plate assay in solid medium are shown in Table 2. On the negative control plates, the colony of S. parasitica reached more than 5 cm in diameter (Figure 1). All three bacteria showed a higher mycelial growth inhibition capacity when S. parasitica was tested on BHI medium with bacterial strains previously grown for 7 days. L. plantarum CLFP3 showed the highest fungistatic activity since it was already observed with a 3-day bacterial culture.3.1.2. Assays with Broth Medium (Hemp Seed Test and Cyst Test)Both E. gallinarum L1 and V. fluvialis L21 isolates partially inhibited the growth of the mycelium of S. parasitica with the dilution containing 2 × 104 bacteria mL−1 and with higher concentrations (Figure 1). On the contrary, L. plantarum CLFP3 did not inhibit the growth of the mycelium with any of the concentrations used.E. gallinarum L1 and V. fluvialis L21 also inhibited cyst germination at concentrations greater than 4 × 103 bacteria mL−1, but L. plantarum CLFP3 only controlled germination with the highest bacterial concentration of 4 × 105 mL−1.3.1.3. Fungicidal Effect of the BacteriaThe three isolates showed to be lethal for S. parasitica, but the effect varied according to the concentration and incubation time of the bacteria. A lethal effect on the mycelium was achieved at lower bacterial concentrations as the incubation time increased. Thus, with one or two days of incubation, a lethal effect was observed only with the highest concentration (2 × 105 bacteria mL−1), while with 3 days of incubation, a lethal effect was observed up to the third dilution (2 × 103 bacteria mL−1). V. fluvialis L21 was the bacterium that presented the greatest fungicidal effect. An atrophied and small mycelium was observed after a period of incubation if a lethal effect was produced, while the mycelium grew and produced zoospores on the control plate and on the plate with a non-lethal concentration of bacteria.3.1.4. Assays with SupernatantsCulture supernatants only inhibited mycelium growth when they were undiluted, as well as in the case of E. gallinarum L1 with the first double dilution (1:2). Regarding the inhibition of cyst germination, none of the supernatants of the three bacteria was inhibitory.3.2. Adhesion to Skin Mucus of the Probiotic Bacteria and Inhibition of S. parasitica Cyst AdhesionTable 3 summarizes the results for the adhesion of the three probiotic strains to cutaneous mucus of brown trout. The adhesion was generally low (between 14.60% for V. fluvialis L21 and 7.54% for L. plantarum CLFP3), but it was higher than the adhesion of bacteria to bovine serum albumin and swine gastric mucus, although there were no significant differences between the adhesion of cysts to the mucus and that observed on the other substrates (p > 0.05, Student’s t-test).The results of the exclusion, competition and displacement of S. parasitica cysts are shown in Table 4. In general, the three bacteria required lower concentrations to reduce the adhesion of cysts in the competition test (bacterial isolates and cysts were added at the same time). On the contrary, for the displacement of the cysts (bacterial isolates were added after the cyst), higher bacterial concentrations were required. L. plantarum CLFP3 was the bacterium that best reduced the adhesion of S. parasitica cysts in the three tests, despite presenting lower percentages of adhesion to mucus than L1 and L21 isolates (Table 3).3.3. Pathogenicity for Rainbow TroutBoth E. gallinarum L1 and V. fluvialis L21 isolates were non-pathogenic for rainbow trout. No signs of disease or lesions were observed in any of the inoculated fish, and no mortality occurred during the days that the fish were kept under observation.In the microbiological study of the internal organs of rainbow trout, only the injected strain was reisolated in two fish inoculated with E. gallinarum L1. In one of the fishes, which was inoculated intramuscularly, it was reisolated from the kidney, and in the other, which was injected intraperitoneally, it was detected in the spleen.3.4. Biocontrol of Saprolegniosis by Adding the Bacteria to Water or FeedNone of the three probiotic strains, i.e., E. gallinarum L1, V. fluvialis L21 or L. plantarum CLFP3, was able to prevent S. parasitica infection in rainbow trout when added to tank water or administered through feed. Between 24 and 48 h after infection, the first macroscopic S. parasitica lesions were observed, while the first deaths were noted on day 3 post infection, reaching 100% mortality between days 6 and 9 post infection both in the groups treated with the probiotic strains and in those not treated (Figure 2), with no statistically significant differences observed. In the control groups (not infected with S. parasitica), 100% survival was observed, except in the experiment with the E. gallinarum L1 strain, in which one trout died 24 h post infection, probably due to the ami momi treatment.4. DiscussionPrevious studies have shown that E. gallinarum L1, V. fluvialis L21 and L. plantarum CLFP3 can be used as potential probiotics against vibriosis or lactococosis for sea bass or rainbow trout. In the present work, the possible utility of these bacteria in the control of saprolegniosis was studied. For this purpose, both in vitro inhibition studies and competition for binding sites against S. parasitica and in vivo tests with experimentally infected rainbow trout were carried out.The three bacteria inhibited the growth of S. parasitica in solid medium, with the most pronounced inhibition in the case of L. plantarum CLFP3, since it appeared after three days of incubation of the bacteria. The greatest inhibition was observed in bacterial cultures that already had 7 days of growth. Our results agree with those obtained by Hussein and Hatai [20], although they found differences depending on the culture medium used; they observed the greatest inhibition when they used BHI agar with longer incubation times.Regarding the inhibition of S. parasitica in liquid culture with hemp seed, the strains E. gallinarum L1 and V. fluvialis L21 prevented mycelial growth, while L. plantarum CLFP3 was unable to control it. However, the three bacteria inhibited the germination of cysts, although L. plantarum CLFP3 required higher concentrations. In all cases, the concentration of bacteria necessary to inhibit mycelial growth was greater than that required to prevent cyst germination. These data agree with those obtained by Bly et al. [16], who found that small amounts of different bacteria of the Pseudomonas genus prevented the germination of cysts. Carbajal-González et al. [18] suggested that the need for lower concentrations of bacteria to inhibit cysts may be due to the fact that a hyphal thallus already exists in hemp seeds, while the cysts have to germinate and form the mycelium. Inhibition of cyst germination in vitro could indicate that these bacteria hinder germination on the external surface of the fish, thereby helping to control S. parasitica infection. In addition, the three bacterial isolates were lethal to S. parasitica. This fungicidal effect is of interest since the three bacteria are able to inactivate S. parasitica and help control infection, acting not only as fungistatics.None of the supernatants of the three bacteria inhibited the germination of S. parasitica cysts, in agreement with [16]; however, in our case, the supernatants did stop the growth of the preexisting mycelium in the hemp seed. The greater inhibition exhibited by E. gallinarum L1 may be due to the fact that this bacterium produces lactic and acetic acid, as well as small amounts of ethanol [31].Epithelium surface colonization and adhesion, interfering with the pathogen’s adhesion, are two positives for the selection of candidate probiotics [36]. Adhesion to the cutaneous mucus of brown trout for the three bacterial strains tested in the present work was found to be low. The limited adhesion of bacteria to fish mucus agrees with results reported by other authors, with greater adhesion on the intestinal mucus of rainbow trout and other fish species than on cutaneous mucus [37,38]. Comparing the adhesion reported by Sorroza et al. [30,31] for E. gallinarum L1 and V. fluvialis L21 to the intestinal mucus of various marine fish and to the cutaneous mucus of brown trout obtained in the present study, both bacteria presented a higher percentage of adhesion to intestinal mucus. These results agree with those of Bálcazar et al. [38] who, using three bacteria from the intestinal microbiota of rainbow trout, observed a greater adhesion to intestinal mucus than to cutaneous mucus. The greater adhesion of E. gallinarum L1 and V. fluvialis L21 bacteria to intestinal mucus may be related to the origin of these bacteria, since they were isolated from intestinal mucus; however, L. plantarum CLFP3, which was originally isolated from cutaneous mucus, presented percentages of adherence to skin mucus lower than the other two bacteria. Chabrillón et al. [37] indicated no specific adhesion of bacteria to a specific host or substrate but that the adhesion capacity depends more on the strain. This is in agreement with the data obtained in the present study for E. gallinarum L1, V. fluvialis L21 and L. plantarum CLFP3, since there were no significant differences in adhesion to the different substrates used, although adhesion to skin mucus was greatest.Although adhesion to skin mucus was low, the three bacteria isolates significantly reduced adhesion of S. parasitica cysts, although higher bacterial concentrations were required for displacement of S. parasitica cysts. This agrees with the results obtained by Carbajal-González et al. [18], who indicated that the greater need for bacteria for displacement may be due to a high adhesion of S. parasitica cysts to mucus, making it difficult for bacteria to subsequently eliminate them. The lowest concentrations of bacteria to significantly reduce attachment of S. parasitica cysts occurred under competitive conditions.An essential characteristic of a potential probiotic is its safety and lack of pathogenicity. It has previously been shown that L. plantarum CLFP3 is safe for rainbow trout [32] and that E. gallinarum L1 and V. fluvialis L21 are safe for sea bass [30,31]. In the present study, found that L1 and L21 are also non-pathogenic for rainbow trout. The genera Lactobacillus, Enterococcus and Vagococcus are included in the group of lactic acid bacteria, which are mostly non-pathogenic and considered part of the commensal microbiota in the gut of several fish species. Thus, González et al. [39] reported the presence of L. plantarum and V. fluvialis as a part of the microbiota of several species of freshwater fish, including rainbow trout, and Petersen and Dalsgaard [40] isolated various Enterococcus spp., including E. gallinarum, from the intestine of fish from integrated chicken-fish farms. However, Osman et al. [41] found signs of septicemia in Nile tilapia (Oreochromis niloticus) associated with two isolates of E. gallinarum. In addition, other species of these genera, such as V. salmoninarum, have been isolated from diseased fish and are capable of causing mortalities up to 50% in rainbow trout farmed at low water temperature [42]. Regarding the genus Enterococcus, three species (E. faecalis, E. faecium and E. hirae) have been associated with enterococcosis/streptococcosis both in freshwater and marine fish [43,44]. Another Enterococcus species, E. seriolicida, initially described as pathogenic for yellowtail (Seriola quinqueradiata) in Japan [45] was later shown to be really identical to Lactococcus garvieae [46].Although together, the three probiotic strains used in the present study achieved acceptable results in vitro, inhibiting the growth of S. parasitica and decreasing the ability of cysts to adhere to the cutaneous mucus, none of the three strains showed effective results in vivo to prevent experimental saprolegniosis in rainbow trout either through the water or administered in feed. Nevertheless, the beneficial effects against S. parasitica, even if limited to in vitro responses, must be considered since the combination of these bacteria with other strains could ensure a sufficient synergistic effect, with the capacity to prevent the presentation of the disease. Gram et al. [47] pointed out that the selection and application of probiotics must be tested for each individual host–pathogen combination and that in vivo activity cannot be predicted based on in vitro testing. These authors found that a strain of Pseudomonas fluorescens AH2 was strongly inhibitory against Vibrio anguillarum in vitro and that it was also able to reduce mortality in rainbow trout against V anguillarum infection via addition to the tank water [48]. Later, they found that this probiotic strain also inhibited A. salmonicida in vitro but was unable to prevent furunculosis in salmon (Salmo salar) in an in vivo model [47].On the contrary, Nurhajati et al. [24] managed to prevent infection by S. parasitica in catfish (Pangasius hypophthalamus) by adding different concentrations of L. plantarum FNCC 226 to the water and found that the effect was dose-dependent, with higher concentrations of S. parasitica necessary to achieve inhibition. The best concentration treatment was between 4.2 × 105 cfu mL−1 and 8.4 × 105 cfu mL−1 of L. plantarum FNCC 226, which can inhibit infection with a concentration up 4 × 107 zoospores mL−1 of S. parasitica. In our case, the concentration of L. plantarum CLFP3 used was very similar (106 bacteria mL−1), but the concentration of zoospores was much lower (3 × 102 zoospores mL−1), although the results are not necessarily comparable since a different experimental infection method was used, as well as different species and sizes of fish.In previous works by our research group using the same experimental conditions, we observed that two Pseudomonas fluorescens strains (LE89 and LE141) were able to prevent infection by S. parasitica in rainbow trout when added to tank water but not when administered with feed [28]. Subsequently, we investigated the possible mechanisms of action of these two isolates and verified that they seems to be related to activity of competitive exclusion and to the production of some siderophores [29]. In addition, Lategan et al. [13,14] found that administration of Aeromonas media strain A199 in water significantly decreased the incidence of saprolegniosis in eel (Anguilla australis) and silver perch (Bidyanus bidyanus) and that the substance associated with the inhibitory activity was indole production [34]. In another study [49], the application of supernatant from a culture of a bacterium identified as Burkholderia sp. Reduced the rate of infection by Saprolegnia sp. in grass carp (Ctenopharyngodon idella). The substance involved with that antifungal activity was thermostable and was identified as 2-pyrrolidone-5-carboxylic acid, a derivate of glutamic acid.Competitive exclusion is generally considered a major strategy by which probiotic bacteria prevent the adhesion of pathogenic microorganisms to fish mucous membranes. Although in vitro studies have frequently reported this effect, studies with animals have failed to reproduce the same results with in vivo models on many occasions [50], which could partly elucidate the results obtained in the present study.Probiotics have been shown to have the capacity to increase innate and adaptive immunity of fish, with the effects exerted on the fish innate immune system as the main desirable characteristics [36]. Pérez-Sánchez et al. [51] investigated the expression of immune-related genes in rainbow trout feeding with L. plantarum CLFP3 (106 CFU g−1 f for 36 days) and found that mRNA levels of IL-10, IL-8 and IgT were significantly higher in the L. plantarum group compared to the control group after Lactococcus garvieae infection, suggesting that protection conferred by L. plantarum was mediated by the stimulation of the immune response.In the present study, the possible mechanism of action of the three isolates used was not investigated, so we do not know whether these isolates could increase the immune response of rainbow trout under the tested conditions. The mechanisms by which probiotics stimulate the immune system are not yet well understood, but it is known that factors such as the type of strain, dose, duration and mode of administration or environmental conditions can affect the immunomodulating potency of probiotics [52]. In this work, a dose of 106 bacteria mL−1 in a water tank or 108 bacteria g−1 for two weeks was used, which may not have been sufficient to stimulate the immune system. However, in the studies in which the usefulness of probiotic strains in the prevention of saprolegniosis has been confirmed, their beneficial effect seemed to be related to the production of inhibitory substances or competition for iron and not to the increase in the immune response [19,29,34,49].5. ConclusionsThe obtained results show that the use of an effective probiotic against a certain disease in a host may not be effective against another pathogen or in another host and that the results obtained in vitro may not always predict the effects when used in vivo. These results highlight the importance of selecting the most appropriate probiotic and its mechanism of action depending on the species of fish and the disease to be prevented.

animals : an open access journal from mdpi

10.3390/ani11092755

PMC8470241

In this work, for the first time, we studied some aspects of the development of the visual system of the annual killifish (Nothobranchius guentheri) in the embryonic and postembryonic periods. The morphoanatomical features of the eye during the growth and maturation of the fish are determined. In addition, the normal histological structure of the main elements of the eye is studied. Age-related changes in the elemental composition of the lens of the eye have been studied. The data obtained make it possible to assert the important role of the visual system for the survival of fish in shallow ephemeral pools. In such environmental conditions, the efficacy of the visual system is a factor in survival and evolution.

In this, work some aspects of the development of the visual system of Nothobranchius guentheri at the main stages of ontogenesis were described for the first time. It was possible to establish that the formation of the visual system occurs similarly to other representatives of the order Cyprinodontiformes, but significantly differs in terms of the individual stages of embryogenesis due to the presence of diapause. In the postembryonic period, there is a further increase in the size of the fish’s eyes and head, to the proportions characteristic of adult fish. The histological structure of the eye in adult N. guentheri practically does not differ from most teleost fish living in the same environmental conditions. The study of the structure of the retina showed the heterogeneity of the thickness of the temporal and nasal areas, which indicates the predominant role of peripheral vision. Morphoanatomical measurements of the body and eyes of N. guentheri showed that their correlation was conservative. This indicates an important role of the visual system for the survival of fish in natural conditions, both for the young and adults. In individuals of the older age group, a decrease in the amount of sodium (Na) and an increase in magnesium (Mg) and calcium (Ca) were found in the eye lens. Such changes in the elemental composition of the lens can be a sign of the initial stage of cataractogenesis and disturbances in the metabolism of lens fibers as a result of aging. This allows us to propose N. guentheri as a model for studying the structure, formation, and aging of the visual and nervous systems.

1. IntroductionThe visual system of vertebrates is one of the most conservative, since the vitality of the organism and its interaction with the environment, such as reproduction, migration, search for food, and physical activity, depend on its work [1]. Elements of the visual system of fish are laid down at an early stage of embryogenesis. The quality of fish vision depends on many factors, such as the density of the retinal photoreceptors, types of photoreceptor cells, the size of the eye, the lens, and the vitreous body [2,3]. It should be noted that the eye is not the main sense organ for all groups of fish, nor for this class of vertebrates as a whole; thus, fish have wide variations in eye size and shape.The Nothobranchiidae family of fish is a large group, mainly living in the shallow ephemeral pools of northern Africa, in connection with which they have formed a unique mechanism of reproduction and development, where the individual stages of ontogenesis are close to each other [4]. Representatives of this family are a convenient object for studying the processes of evolution, development, and aging [5,6]. Members of the genus Nothobranchius are characterized by a wide variety of karyotypes, which may play an important role in the adaptation of fish to unstable environmental conditions [7]. Their short life cycle is due to living in dry bodies of water, which are only filled with water for a certain period of the year. The food supply for such water bodies is not significant and consists of insects and their larvae [8]. Under such conditions, the efficacy of the visual system is a factor in survival and evolution [9]. In shallow ephemeral pools, the amount of solar insolation and the amount of incoming ultraviolet (UV) rays can be destructive if the fish has not formed an effective defense mechanism [10]. N. guentheri have a wide variety of behavioral responses, both in the natural environment and in laboratory conditions. Ethological tests based on the use of different variants of colors as an influencing factor show high learning ability and confirm a high degree of development of the visual system and the corresponding parts of the brain [11]. The intrapopulation hierarchy of N. guentheri males is based on the size and color intensity of individual fish, which is also largely determined by the development of color vision [4].In many vertebrates, long-term exposure to UV rays contributes to the development of lens cataracts [12]. Early stages of lesions affecting lens transparency may be evaluated by the elemental composition, in particular the change in the ratio of potassium (K), sodium (Na), and calcium (Ca) [13,14]. Ultraviolet radiation also leads to a disruption in the structure of lens proteins, e.g., crystallins (most common α-, β- and γ-crystallin) [15]. Preparation of more detailed information about irregularities in the structure and synthesis of proteins may be established through transcriptome analysis, which is currently being studied in the brain [16]. Changes in fish habitat and the pollution of water bodies by technogenic discharges can also be assessed by the chemical composition of organs with slow metabolism, otoliths, and the eye lens [17,18,19,20].The embryonic development of N. guentheri takes a long time (at least 21 days under laboratory conditions) [21], and usually includes three diapauses necessary for the full formation of the larva [22,23]. Such a complex development mechanism is due to the drying out of temporary water bodies and makes it possible for the fish population to survive long-term unfavorable conditions without substantial energy losses [4,9]. The last diapause preceding hatching is the most stable and can last more than six months [18].The visual system of N. guentheri has great importance to its lifestyle, as indicated primarily by the size of the eyes in relation to the body size. According to our observations, the eye size of N. guentheri changes significantly during life. In many species of teleost, there is a correlation between habitat conditions and morphological characteristics of the eye [24]. One of the characteristic features of the visual system of the Nothobranchius is the significant size of the lens.The aim of this work is to identify some aspects of the morphoanatomical parameters of the eye of N. guentheri and its parts throughout life. The short life cycle makes it possible to study the development of the visual system, starting from the early stages of embryogenesis and ending with the death of fish from natural causes, which, under laboratory conditions, occurs at 10–12 months post-hatching. In addition, the study investigated the content of individual elements in the lens of fish in order to identify the patterns of their exchange and accumulation in different age groups.2. Materials and Methods2.1. Object of StudyThe study complied with the guidelines of the Local Ethics Commission of the Institutional Review Board of Moscow State University of Technology and Management (approval number 7, 2 February 2021).The isolate N. guentheri Zanzibar TAN 14-02 was obtained from the collection of the Engelhardt Institute of Molecular Biology of Russian Academy of Sciences at the age of two to eight months. This served as the main base of the experiments in 250 individuals. Fish were divided according to sex and size parameters, and were kept in aquariums (45 × 45 × 15 cm, W × L × H) with a volume of 50 L, with 10 fish each of the same size and gender, and constant aeration at a temperature of 22 ± 2 °C and pH 7.2–7.6. The water was changed according to the following scheme: partial, every two days; full, every seven days. The fish were fed live food (Artemia salina, Daphnia magna, and larvae of the Chironomidae family) twice a day, at 12:00 and 18:00.2.2. EmbryologyEggs for the study of embryogenesis and the development of the visual system were obtained from the group of N. guentheri, consisting of one male and two or three females, 5–7 months old, more than 3 cm in size. Individuals were placed in aquariums with a volume of 50 L. The water layer did not exceed 10–15 cm, at the bottom of which there was a container with a peat substrate crushed to 0.5–0.8 mm. The eggs were taken every seven days and incubated in a moist peat substrate (60–70% water by weight) at a temperature of 24–25 °C.Control of embryotic development was performed every six hours during the first three days (before the beginning of diapause I) and on days 9–14 (from the beginning of diapause II to the beginning of III). The remainder of the time, control was carried out in races once per day.N. guentheri was grown according to the method of Genade et al. [23]. On days 21–24, embryos at stage III of diapause were selected. Then, they were transferred to incubators with distilled water for hatching.2.3. HistologyTo obtain histological sections seven-month-old males without visible damage were selected. The fish were sacrificed in MS-222 solution (250 mg/L), after which they were fixed in 4% neutral formalin solution for 24 h at room temperature. Then, tissue samples were dehydrated in a series of graduated alcohols and embedded in paraffin. To obtain the eye in the horizontal plane [25], serial sections of the frontal plane (anterio-posterior plane) of fish (4 μm) were stained with hematoxylin and eosin (H&E) and examined under a light microscope. Preparation and staining of histological slides were performed according to Suvarna et al. [26].2.4. MicroscopyAn Olympus BX53 light microscope (Olympus Corporation, Japan, Tokyo) with a Carl Zeiss ERc 5s eyepiece attachment (Zeiss, Germany, Oberkochen) and ZEN lite software (Zeiss, Germany) was used for microscopy of eggs and histological preparations. Egg development was examined in plates with 96 wells.In total, 10 sections (n = 10) obtained from 5 fish were examined. The morphometric measurement of the layer thickness in the temporal (caudal) and nasal (cranial) zones of the retina was performed on 3 sections (n = 3) (50 measurements per section).2.5. Fish Photography and Morphometric MeasurementsTo assess the morphoanatomical changes in the fish eye during ontogenesis, individuals of different sizes and age composition were photographed using the techniques described by other authors with some modifications [27,28,29]. The fish were sedated in a solution of MS-222 (0.1 mg/L) [30], after which their physical activity significantly decreased. This made it possible to take photographs without hindrance. The location of the source of artificial lighting was carried out in such a way that the resulting image on the camera was clear and there were no various negative effects (glare, overexposure, etc.). In this work, we used a studio diode ring illuminator Raylab RL-0518 Kit (Russia) with a color temperature of 4200 K (neutral color range) and a Nikon D5000 camera (Japan). For a reliable assessment, the survey was carried out under the same lighting conditions (4200 K) and camera settings (ISO 400–500, F 5.6, 1/60 s) for all the studied fish.A total of 107 fish were photographed during the study, from which 60 fish were randomly sampled.Morphometric measurements were performed using ImageJ software (Wayne Rasband (NIH); https://imagej.nih.gov/ij/; access date: 6 June 2021 ). On all the selected photographs of the fish, the parameters shown in Figure 1 were measured and selected based on the described method [31].2.6. Eye Elemental CompositionThe elemental composition of the lens was measured in 4 males of each age group (2 and 7 month) (n = 4 × 2) by X-ray energy dispersive microanalysis on a two-beam scanning electron microscope (Zeiss CrossBeam 340 with a Schottky cathode) using an X-Max 80 Oxford Instruments detector. To ensure the reliability of the results obtained and to reduce the measurement error, a preliminary calibration of the detector was carried out on standard samples at reduced accelerating voltages of 7.5 and 10 kV. In addition, in the course of measurements, a 25 nm thick carbon film deposited on the surface was taken into account, the signal introduced by the film atoms was subtracted from the total sample signal by specialized software.2.7. StatisticsComparative analysis of various parameters under study was performed using Student’s t-test; the value of p < 0.05 was taken as a significant difference. The Shapiro–Wilk test was used to determine the normal distribution of the data.The correlation between different morphometric parameters and different elements was determined using Pearson’s correlation with Student’s t-distribution to calculate the significance. Paired linear regression was performed for parameters with a significant value of the correlation coefficient. Statistical data processing was performed using GraphPad Prism version 8.0 software (GraphPad, San Diego, CA, USA).3. Results3.1. EmbryogenesisThe formation of the visual system of N. guentheri begins with the formation of the optic (eye) vesicle at 16–18 h post-fertilization (hpf) (not shown). Further differentiation and formation of the optic vesicle ends at 24–26 hpf (Figure 2a,b). At this stage of development, the anterior part of the brain is clearly distinguishable, which indicates the parallel formation of the visual and nervous systems. At this time, the lens is formed by invagination of ectoderm into the optical cup. Further development is suspended due to the onset of diapause I (72–96 hpf), the duration of which depends primarily on the water temperature.After the end of diapause, development is accelerated. In the period from 96–120 hpf, further formation of the nervous system occurs, and the front and middle parts of the brain are already distinguishable. In the lens, the process of forming the lens fibers from the cubic epithelium continues. There is a differentiation of the periocular mesenchyme, which includes melanocytes (Figure 2c,d).Under non-optimal conditions for the course of embryogenesis at this time (five to seven days post-fertilization (dpf)), diapause II occurs. Its duration depends on a large number of factors [32], most of which are insufficiently studied. For these reasons, under laboratory conditions, diapause II can last from several hours to tens of days. In the conditions of our experiment, its completion was noted at 8–12 dpf. After diapause, there is a notable acceleration of organogenesis. By the time of the onset of diapause III (12–14 dpf) in the embryo, the visual system can be considered fully formed, and the main structures of the eye, including the sclera, cornea, and lens, are clearly distinguishable (Figure 2e). Melanocytes, which are part of the choroid, are visibly clearly distributed in connective tissue differentiated from the peculiar mesenchyme. Complete differentiation of all elements of the eye is completed by the end of diapause III (21–26 dpf), which can be determined by the presence of body and head pigmentation in the embryo, which is characteristic of postembryonic fish. In the eye of the embryo, the pupil and the iris are clearly distinguishable (Figure 2f).3.2. Post-Hatching DevelopmentN. guentheri larva, during the transition to a free lifestyle, is characterized by positive phototaxis and the dorsal part of the body is covered with melanocytes to protect it from UV radiation (Figure 3a,b). The eye occupies most of the head section of the fish. In our laboratory, 12–24 h after hatching, N. guentheri larvae switched to active feeding on Artemia nauplii. At 7–10 days post hatching (dph), the larva acquires some of the features characteristic of adult fish: the formation of a scaly cover, a change in the shape of the mouth, and the shifting of the eye to the ventral side (Figure 3c,d).3.3. Histology AssayIn total eyes, the N. guentheri slice shows that the basic structure of the visual system has a standard structure for teleost. Nevertheless, it was possible to identify some features, most of which are associated with the structure of the retina. The lens (Figure 4b,c) has no pronounced features. The lens has clearly distinguishable lens fibers, formed from a single layer of epithelium, which are located at the anterior pole of the eye. The cornea usually consists of three layers: squamous corneal epithelium, corneal stroma, thin corneal. N. guentheri has only one well-defined structure of cornea—the corneal stroma poorly differentiated into separate layers. The ligamentum pectinatum was identified at the junction of the iris and cornea (Figure 4f).The N. guentheri retina has a clear-layered structure, including seven layers: ganglion cell layer, inner plexiform layer, inner nuclear layer, outer plexiform layer, outer nuclear layer, inner segment/outer segment of the photoreceptor cells, retinal pigment epithelium/cells (Figure 4d,e). At the same time, the outer nuclear layer (ONL) presents differentiated photoreceptors with well-developed outer segments (Figure 4g), which makes it possible to assert the high quality of vision of this fish species.Comparison of the temporal and nasal areas of the retina made it possible to reveal a significantly (p < 0.05) greater thickness of retina layers in the temporal area (Figure 4l,m). In addition, the retina of the eye has its own choroid rete, similar to Danio rerio [33]. Choroid in adult fish restricts the choroid rete from the caudal side and consists of connective tissue and melanocytes. The ciliary marginal zone (CMZ) is observed in the most-peripheral retina (Figure 4h,i).3.4. Morphoanatomical MeasurementsMeasurement of the features of the eye size to the size of head and body in different-sized individuals of N. guentheri males made it possible to establish the conservatism of these parameters.There was a significant correlation (p < 0.05) (Figure 5d) between the morphometric parameters of the body (SI, ICA) and eyes (dO, sdO). The highest values of the coefficient of determination were obtained by comparing the parameters dO and sdO of the eye and head length, 0.6995 and 0.8072, respectively (Figure 5b,c). The coefficient of determination between the parameters of the eye and the standard body length were slightly lower (Figure 5e,f).3.5. Elemental Composition of the LensAs a result of measuring the elemental composition of fish lens at the ages of two and seven months, it was revealed that some of the elements retain their concentration throughout the life of N. guentheri. Such elements include silicon (Si) and sulfur (S). The other elements included in the lens during ontogeny are more volatile (Table 1, Tables S1 and S2 and Figure S1).In 2-month old males the detected elements (excluding carbon and oxygen) constitute 2.64% of the weight of the lens composition, while in 7-month-olds, this percentage is 3.2% (Figure 5a,c). Element concentration data can be arranged in the following sequence:For 2-month old: S > Al > Mg > Na > P > Ni > Cl > Ca > Si;For 7-month old: Al > S > Na > Ca > Mg > Ni > P > Cl > Si.The evaluation of the correlation between all detected elements made it possible to establish the presence of positive significant correlations between the content of the following elements: Ca, Na, and Si (p < 0.0001 and p < 0.05, respectively); Cl, Si, and P (p < 0.05 and p < 0.001); S and Si (p < 0.01) (Figure 6b and Figure S2).4. Discussion4.1. Embryology and Post-HatchingThe results of this study suggest that N. guentheri has an effective visual system, adapted to exist in small ephemeral ponds and participate in the formation of the social hierarchy of males of this species of fish.At this time, the embryogenesis of N. guentheri and the factors affecting long-term diapause have been widely described [8,33,34,35,36]. However, these works do not devote much attention to the development of the visual system.The data obtained in the course of this work suggest that the process of the embryonic development of the visual system of N. guentheri forms similarly to that in all other bony fish species [37], but takes much longer due to the presence of the diapause. The obtained terms of development are typical for the conditions of our laboratory and have been tested many times, but it should be recognized that, in other conditions, the stage of development can vary significantly depending on the chemical and temperature parameters of the environment. Due to the shortened life cycle, by 7 dpf the larva acquires some features typical of juvenile fish. Fish in the juvenile stage can occupy all possible ecological niches of shallow ephemeral pools [24].The study of the development of the visual system in other members of the genus Nothobranchius shows that differentiation of retina cells does not stop after hatching and continues throughout the life of the fish [38,39]. The emergence of new cells in the fish retina occurs from the proliferation of multipotent progenitor cells located in the CMZ [40,41] and from dividing Müller glial cells [41]. For N. guentheri, we have shown the possible presence of this zone, but this fact requires additional research.The shortened developmental period and the rapid transition to the reproductive stage make it possible to classify this fish species as precocial. The high development rate is an adaptation to life in drying up water bodies [42], where reproductive success directly depends on growth rate and social hierarchy. Based on this, it can be assumed that the visual system of N. guentheri is functional at the time of hatching, since it plays an important role in survival in the natural environment.The closest analogue of non-annual fish may be Oryzias latipes, whose stages of development of visual system are completely repeated but differs in terms of individual stages and the entire embryogenesis as a whole [43]. From the group of annual killifish, similar development times are observed in Aphyosemion gardneri [44]. The formation of the visual system of N. guentheri proceeds in parallel with the formation of the fish brain and, by the time of the onset of III diapause, it is fully functional (phototaxis) [39]. This is typical for most annual fish [24,44].By the time III diapause begins, the pigment part of the choroid covers most of the eyeball. It has been reported [44] that melanocytes in the choroid prevent light scattering, providing a clearer image. Freshwater ecosystems are characterized by a high degree of UV absorption due to the low proportion of transmitted sunlight caused by large amounts of organic and inorganic particles [45]. However, for the shallow ephemeral pools in which killifish live, the amount of absorbed radiation can be much higher, leading to the formation of a special mechanism that protects the photosensitive cells of the retina primordium. It should be noted that the bottom fish, such as Scophthalmus maximus, have a poorly developed pigment layer of the choroid [46]. It is likely that the visual system is also one of the components of the mechanism of hatching the larvae of annual fish, since even under suitable hydrochemical conditions not all fry hatch simultaneously [47]. In this case, suitable illumination conditions may serve as a stimulus for hatching.4.2. Histological AssayThe study of the eye histology of adult N. guentheri (seven months) showed that the main elements of the visual system are similar to other fish species [48,49,50]. As noted by Seritracul et al. [2], the structure of the retina is very similar in all vertebrates, including humans. In addition, fish retina has the properties of postnatal neurogenesis and regeneration (due to the presence of stem cells). These facts make the fish retina an excellent model in biomedicine. The detected choroid rete restricts the blood vessels that do not penetrate the retina. This indicates the absence of retina vascularization in N. guentheri, as in most other teleost fishes [51,52].In all studied fish, no visible pathological changes in the structure of eye tissues were observed. Despite accelerated aging and the appearance of age-dependent changes in other fish organs [53,54], the visual system continued to function normally. The reason for this is likely the significant role of the visual system for the survival of the organism in the natural environment [55]. While thinning of the retina layers and a decrease in the pigment epithelium layer are often noted in aged individuals of Oryzias latipes and Danio rerio [50,56], these changes will occur in a very short period of time preceding the death of a N. guentheri due to its short life cycle.The thickness of the retina and its layers in N. guentheri was found to be similar to D. rerio (~109 and 113 µm, respectively) [51]. Differences in the thickness of the retina layers in the temporal and nasal areas of the eye are possibly associated with peripheral vision, which plays an important role in various behavioral acts, such as reactions to a predator, and sexual and feeding behavior [8], which is typical for most pelagic fish [57,58]. In particular, the harem sexual behavior characteristic of N. guentheri implies the protection of a certain spawning area from competing males, which determines the success of reproduction with the maximum number of females. Under these conditions, vision plays an important role in maintaining social hierarchy. The increased thickness of the cranial part of the retina is most likely a consequence of the ecological adaptability of fish to life in shallow ephemeral pools [59]. In addition, the amount of light is an important factor in the evolution of the eye of all vertebrates [60]. Thus, the development in the conditions of shallow ephemeral pools contributed to the adaptation of the killifish visual system to the conditions of high light insolation. To obtain more complete data on the thickness of the retina layers, additional studies are needed to compare the dorsal and ventral sides.The similarity of the retina and lens structures between N. guentheri and other fish species such as Oryzias latipes and Danio rerio [43,45] suggests the possibility of using it as a model for studying the structure, development, and occurrence of age-dependent pathologies.4.3. Morphoanatomical CharacteristicRevealing the regularities of the eye size from other morphoanatomical parameters confirms the assumption about the conservatism of the visual system of N. guentheri. The ratio of the eye and head parameters does not change during the entire life span, although the fish of order Cyprinodontiformes are characterized by high population variability in a number of other parameters [61]. The relationship between the diameter and area of the eye and the size of the whole body is not so significant due to the variance in body size. All this speaks in favor of the presence of ontogenetic allometry between the dimensional characteristics of the eye and the body of the fish. Allometry in fish has previously been found between body and lens size [62]. It should also be noted that the size of fish eyes shows the greatest variability among other groups of vertebrates [63]. The location and size of the eyes of N. guentheri are characteristic of fish feeding on plankton and small benthic organisms.4.4. Elemental CompositionThe study of the microelement composition of the lens, as an inert structure of the body, is an effective way to assess age-related changes [19,20]. Since the metabolism of the lens is slow, the substances that enter it persist for a long time, and in many cases, throughout the life of the fish [62]. The elemental composition of inert tissues may reflect the environmental conditions in which fish live, as has been demonstrated for some fish species [19,44]. In addition, pathological changes will occur when the protein composition of the lens is disturbed, which in turn can lead to a violation of transparency [12,64]. Such changes most often occur at the last stages of ontogenesis and can be caused by two main reasons: (i) impaired differentiation of lens fibers from the epithelium in the equatorial zone of the lens (impaired folding and post-translational protein modifications) due to damage, diseases, etc. [65], and (ii) toxic effects associated with the accumulation of pollutants [66,67,68], also leading to disruption of the synthesis of crystallins [69].According to the results of this work, aluminum, which accumulated in significant amounts in the lenses of seven-month old fish, may be one toxic element that can lead to the formation of cataracts in N. guentheri,. The studies carried out made it possible to compile a series of percentage concentrations of elements, which showed that with age, in N. guentheri, the concentration of Na decreases, while the concentrations of Mg and Ca increase. Since these elements have an important biogenic role, it is likely that we can talk about a decrease in or impairment of the metabolism of the lens cortex with the age of the fish. It has been reported that calcium accumulation is a sign of the development of the initial stage of cataractogenesis [70]. The non-biogenic elements of the lens also change with the growth and aging of the fish [19]. The accumulation of these elements is primarily associated with the quality of the aquatic environment, as well as with the high permeability of the lens for ions, water, and other small molecules due to ion pumps and gap contacts [64,69,71]. For example, in fish living in anthropogenically polluted waters, there is an increase in the content of heavy metals and other non-biogenic elements in the lens and otoliths [72].5. ConclusionsAs a result of this work, the development of the visual system of N. guentheri at the main stages of ontogenesis was described for the first time. It was possible to establish that the formation of the visual system occurs similarly to other representatives of this order, but significantly differs in terms of the individual stages due to the presence of diapause. The beginning of the formation of the visual system occurs at 16–18 hpf and the main structures of the eye are formed in parallel with the processes of neurogenesis. The main structures of the eye can be considered formed by the time of the onset of III diapause (12–14 dpf).The histological structure of the eye in adult N. guentheri practically does not differ from most bony fishes, which live in similar environments. The study of the structure of the retina showed the heterogeneity of the thickness of the temporal and nasal areas. None of the studied histological preparations revealed age-dependent changes in the structures of the eye.Morphoanatomical measurements of the body and eye parameters of N. guentheri showed a conservative correlation of these parameters. This indicates the important role of the visual system in the survival of fish in natural conditions.The elemental composition of the lens changes as the fish grows and matures. In individuals in the older age group, a decrease in the content of Na and an increase in Mg and Ca were discovered. Such changes in the lens can be a sign of the initial stage of the impaired metabolism of lens fibers. The accumulation of aluminum with age is most likely an example of the deposition of abiogenic elements in inert structures of the body.The study of the mechanisms of the formation of the visual system of N. guentheri allows us to assert the great role of vision for the survival of fish in their natural habitat of shallow ephemeral pools, in its participation in sexual selection, and in the social structure of the population. In laboratory conditions, it is possible to study many behavioral reactions in N. guentheri in which vision plays an important role. All of this allows us to propose N. guentheri as a model for studying the structure, formation, and stages of the onset of age-related changes in the visual and nervous systems.

animals : an open access journal from mdpi

10.3390/ani13071238

PMC10093262

“Moral personality” is required in order to be entitled to justice in John Rawls’s theory of justice, a famous and influential theory in political philosophy. The concept of moral personality involves the possession of two “moral powers”. One moral power is a capacity for a conception of the good, being a conception of what is regarded as worthwhile in life, while the other is a capacity for a sense of justice. Rawls claims that non-human animals (hereafter, “animals”) do not possess these moral powers, and accordingly he omits them altogether from his theory of justice. In this article, I raise doubts about this omission, outlining how at least some animals may indeed possess the moral powers, albeit to a lesser extent than most humans. In this regard, the distinction between humans and animals can be seen as one of degree rather than kind. A proper acknowledgement of animal abilities suggests that Rawls’s theory requires alteration to accommodate the position of animals.

The relationship between animal rights and contractarian theories of justice such as that of Rawls has long been vexed. In this article, I contribute to the debate over the possibility of inclusion of animals in Rawls’s theory of justice by critiquing the rationale he gives for their omission: that they do not possess moral personality. Contrary to Rawls’s assumptions, it appears that some animals may possess the moral powers that comprise moral personality, albeit to a lesser extent than most humans. Some animals can act in pursuit of preferences and desires (and communicate them non-verbally), which might be taken as implicitly selecting a conception of the good; further, scientific research demonstrating inequity aversion and social play behaviors suggests that some animals can have a sense of justice relating to their own social groups. I conclude that Rawls’s theory needs to acknowledge any animals that can be considered to meet the threshold of moral personality, while the concept of moral personality as a range property may also require reconsideration.

1. IntroductionA limitation of John Rawls’s theory of justice is that the position of non-human animals (hereafter, “animals”) is treated as outside the scope of the theory. Although Rawls considered cruelty towards animals and the destruction of an entire species to be wrong, animals are said not to be entitled to justice [1] (pp. 441–442, 448–449). The reason offered for this exclusion is that animals are not “moral persons”. Moral personality involves two moral powers—the capacity for a conception of the good and the capacity for a sense of justice—that are claimed to be uniquely human attributes. Rawls’s position seems to mean that, as Martha Nussbaum puts it, “[e]ven … the twentieth century’s greatest philosopher of justice … held that it was virtuous to treat animals with compassion, but that they could not be treated justly or unjustly” [2] (pp. 8–9).The relationship between animal rights and contractarian theories of justice such as that of Rawls has long been vexed; a key difficulty is how to include animals in such theories if they are unable to participate effectively in the making of a social contract [3]. Mark Rowlands argues that “there is nothing in contractarianism per se that requires the contract be restricted to rational agents”; even if the framers of the contract have to be rational agents, its recipients do not (p. 236). Meanwhile, Robert Garner claims that Rawls’s approach cannot adequately explain the position of “those humans who are less endowed with rationality or autonomy”, invoking the well-known argument from so-called “marginal cases” [4] (p. 7), though he has also critiqued this argument [5]. The contribution of this article is to consider a different, under-explored issue: the merits of Rawls’s claim that animals do not have moral personality, which underpins their exclusion from his theory of justice.In ascribing moral personality only to humans, Rawls treats humans and animals as qualitatively different. Rawls’s concept of moral personality is widely assumed to “clearly preclude animals” [6] (pp. 2–3). However, as Rowlands says, “the all or nothing manner in which discussions of non-human rationality tend to be discussed is eminently questionable, on both theoretical and methodological grounds” (p. 236). Indeed, I argue that some animals might be better viewed as possessing moral personality, though to a more limited extent than most humans. The difference between humans and animals in this regard may be thought of as one of degree rather than kind, to adopt Charles Darwin’s words [7] (p. 179). Nonetheless, acknowledgment of the lesser degree of moral personality possessed by some animals is likely to require significant alterations to Rawls’s theory. Although I seek to establish some problems with Rawls’s account, modifying the theory to accommodate the possibility of animal moral personality is not attempted here. Further, I do not suggest that possession of some measure of moral personality means that animals (or, for that matter, humans) cannot or do not perform acts that inflict suffering on others.The article proceeds in three parts. First, I address Rawls’s analysis of the position of animals in his theory. As will be shown, Rawls’s theory is already consistent with the view that animals are owed moral concern. Nothing prevents legislators from protecting animal welfare or rights. However, Rawls considers that due to their claimed absence of moral personality, animals are not owed justice, and this may mean that animal protection is of lower priority than the principles of justice. Second, I question Rawls’s approach by analyzing the position of animals in respect of each of the moral powers. I suggest that Rawls’s conclusion that animals do not have moral personality is too simple; there is reason to consider that at least some animals can potentially possess the moral powers, albeit to a lesser degree than humans. Third, I conclude with a short reflection on the possible implications of this analysis for Rawls’s theory.2. Rawls and AnimalsIn A Theory of Justice, Rawls only briefly addresses the position of animals. He describes “the basis of equality” as “the features of human beings in virtue of which they are to be treated in accordance with the principles of justice” and claims that “[o]ur conduct toward animals is not regulated by these principles, or so it is generally believed” (p. 441). That is because it is the “moral persons” who are entitled to equal justice. Such moral persons are “capable of having (and are assumed to have) a conception of their good (as expressed by a rational plan of life)” and “capable of having (and are assumed to acquire) a sense of justice, a normally effective desire to apply and to act upon the principles of justice, at least to a certain minimum degree”. Thus, “equal justice is owed to those who have the capacity to take in and to act in accordance with the public understanding of the initial situation” (p. 442).Moral personality refers to a capacity, not its realization; a being that has the required capacity, whether or not it is developed, receives “full protection” (pp. 445–446). Moreover, it is a “range property”, similar to how “the property of being in the interior of the unit circle is a range property of points in the plane”, because all points inside the circle have the property of being in the interior and they have the property equally (p. 444). In the same way, “provided the minimum for moral personality is satisfied, a person is owed all the guarantees of justice” (p. 443). However, Rawls considers that animals do not meet this minimum, and although Rawls describes moral personality as a sufficient condition for being entitled to equal justice and leaves aside whether it is also a necessary condition (pp. 442–443), he seems to treat it as a necessary condition in the case of animals. He explains his view of the position of animals in a key passage:Last of all, we should recall here the limits of a theory of justice. Not only are many aspects of morality left aside, but no account is given of right conduct in regard to animals and the rest of nature. A conception of justice is but one part of a moral view. While I have not maintained that the capacity for a sense of justice is necessary in order to be owed the duties of justice, it does seem that we are not required to give strict justice anyway to creatures lacking this capacity. But it does not follow that there are no requirements at all in regard to them, nor in our relations with the natural order. Certainly it is wrong to be cruel to animals and the destruction of a whole species can be a great evil. The capacity for feelings of pleasure and pain and for the forms of life of which animals are capable clearly imposes duties of compassion and humanity in their case. I shall not attempt to explain these considered beliefs. They are outside the scope of the theory of justice, and it does not seem possible to extend the contract doctrine so as to include them in a natural way.(p. 448)In short, then, Rawls does not accept that animals are capable of a sense of justice, and therefore we are not required to give “strict” justice to creatures lacking this capacity. However, Rawls’s position that animals are not entitled to “strict” justice raises questions. How do we square this with his view that cruelty towards animals or the destruction of an entire species would be wrong? Presumably, these may be moral duties, but not duties of justice or at least not of “strict” justice. This position is open, of course, to objection that animals might conceivably be owed justice on the basis of their sentience or consciousness [8,9], or experience of pain and pleasure [10], without it being necessary to demonstrate that they themselves have a sense of justice. For Rawls, it seems that such matters might suffice to demonstrate that animals are due moral consideration, but not justice (or “strict” justice).In other words, as Ruth Abbey puts it, Rawls “makes it clear that humans have duties to animals that derive not from the considerations of justice, but from those of morality” (p. 6). Rawls’s position therefore does not seem to preclude animal welfare laws or even animals holding non-basic liberties—legislators can presumably grant animals welfare protections or rights even if this is not compelled by the principles of justice, but merely considerations of morality. Indeed, animal welfare laws have been introduced in many jurisdictions. From one perspective, animals may already have “simple” rights under these welfare laws [11]. Legislators might even choose to confer “fundamental” rights on animals. Such conferrals of rights may take place on the basis of the sentience of animals or other considerations.One might then query the relevance of the analysis in this article. Surely, it might be said, it is human moral personality that matters, as humans make use of their moral reasoning to appreciate their duties to animals. However, Rawls’s position has potentially adverse consequences for animals, including that animal interests might be subordinated to the human interests protected by the principles of justice [12]. A particular difficulty is that the interests of animals could seemingly be overridden by human basic liberties protected by the first principle of justice (the liberty principle). Pursuant to that principle, Rawls outlines a list of basic liberties required for the “adequate development and full exercise of the two moral powers” [13] (p. 45). Since only humans have these basic liberties, any rights of animals appear to be in a lesser position, no matter the relative importance of the interests involved.For example, religious practices are protected by liberty of conscience (often referred to as religious freedom), which is one of the basic liberties within the first principle of justice. As such, if animal sacrifice is a religious practice protected by a basic liberty adopted as part of the basic structure of society, it might be considered to take priority over animal protection laws. According to Rawls’s explanation in A Theory of Justice, basic liberties can be limited when they conflict with other basic liberties, particularly outside what he calls their central range of application (p. 54). He also acknowledges possible limitations on basic liberties in the common interest in public order and security (pp. 186–189). Whether animal protection laws would be subject to the exercise of a basic liberty may therefore depend on whether those laws validly limit a basic liberty. Nonetheless, the failure to acknowledge animal basic liberties creates a risk that a human basic liberty will take precedence over animal interests.This possibility is not merely hypothetical. The case of Lukumi Babalu Aye v City of Hialeah [14], in which the US Supreme Court struck down laws targeting animal sacrifice by Santeria practitioners on the basis that they violated the Free Exercise Clause of the First Amendment to the US Constitution, serves as an illustration of the possibility of prioritization of human rights over even the lives of animals. However, when the interests protected by a human basic liberty do not seem as important as, say, the life of an animal, that may clash with an intuition that “the basic or fundamental interests of animals should not be sacrificed for the sake of the trivial or non-basic interests of humans”, as Garner puts it (p. 11). The question that arises is whether animals can be moral persons whose fundamental interests can be protected by principles of justice at the same level as human interests.In analyzing moral personality, I adopt a different focus from other recent accounts of animals that draw on Rawlsian ideas. For example, Federico Zuolo employs the concept of public justification to develop a theory that builds on the presence of reasonable disagreement about animals [15]. Laura Valentini argues that, from the perspective of an associative account, justice requires that the interests of domesticated dogs “be taken into account in the design of laws and policies within our political communities” [16] (p. 49). That is because, given the history of cooperation between humans and dogs, they can be considered to be fellow cooperators that are objects of moral concern (p. 38). Although these approaches are promising, I do not address public justification or animal cooperation with humans, instead offering a critique of Rawls’s assumptions about animal abilities.3. Animal Moral Personality?Are the moral powers that constitute moral personality exclusive to human beings? If they are not, then we may question Rawls’s conclusion that animals are not entitled to strict justice and his consequent omission of them from his theory of justice. As Raffael Fasel points out, “[i]dentifying the precise features of human nature that made humans (and only humans) special” has long “proved difficult” [17] (p. 4). In this section, I consider each of the moral powers—the capacity for a conception of the good and the capacity for a sense of justice—and conclude that some animals can potentially possess the moral powers, albeit to a more limited extent than humans. In part, the analysis reflects the fact that our scientific understanding of animal behavior and cognition has progressed since Rawls wrote A Theory of Justice. However, since we still have an incomplete understanding of animal abilities, the analysis remains subject to possible future changes in scientific understanding.One difficulty is that moral personality is a range property, such that those who fall short are considered not to have moral personality at all, while those who meet the threshold to any degree have full moral personality [18,19]. There is reason to consider that some animals may well qualify as moral persons even though their extent of moral personality is less than most humans; many other animals are likely to fall short. To adapt Rawls’s theory in a way that makes it truly non-anthropocentric, it might be necessary to rethink the understanding of moral personality as a range property, though whether moral personality could be reconceived as scalar or as having gradations that reflect lesser degrees of moral personality—which would constitute a significant alteration to the theory—is a complex question. In this analysis, I consider moral personality as a range property, though I note the possibility of alternative approaches in the conclusion.A further difficulty that would arise in an attempt to adapt Rawls’s theory to include animals is determining which animals reach the threshold of moral personality. Marc Bekoff and Jessica Pierce, commenting on the possibility of identifying animals that constitute “moral animals” in the sense in which they use the term, observe that “[g]iven the rapidly accumulating data on the social behavior of numerous and diverse species, drawing such a line is surely an exercise in futility, and the best we can offer is that if you choose to draw a line, use a pencil” [20] (p. 8). However, they outline a preliminary list of primates, social carnivores, cetaceans and some rodents for which they consider there is “compelling evidence for moral behavior” (p. 9). The same kind of evidentiary limitations apply to the question of which animals might qualify as Rawlsian moral persons based on our current knowledge; the answer may shift over time.In raising the possibility of animal moral personality, I do not seek to rely on the so-called “marginal cases”—humans who are too severely disabled or unwell to have the capacity for moral personality. Rawls emphasizes that what is required for moral personality is the potential to exercise and develop the moral powers “that is ordinarily realized in due course” (p. 442). Thus, infants and children are included, as are those who have lost their capacity temporarily. However, it could be objected that Rawls does not account for the moral status of those humans whose incapacity may be permanent or lifelong. He observes that the problem of “those more or less permanently deprived of moral personality may present a difficulty” but declines to examine it (p. 446). One possible approach may be to accept that some “scattered” humans, as Rawls puts it (p. 443), do not meet the threshold. In respect of animals, though, the problem is different: it is that Rawls overlooks the actual abilities of many animals, which go a significant way towards human moral personality. Thus, it is to the issue of the abilities of animals that I turn.3.1. A Conception of the GoodIn A Theory of Justice, Rawls explains the moral power of a capacity for a conception of the good as involving a “rational plan of life” (p. 442). In Political Liberalism, the moral power is defined as “the capacity to form, to revise, and rationally to pursue” a conception of the good, being “a conception of what we regard for us as a worthwhile human life”, which normally consists of a “determinate scheme of final ends and aims, and of desires that certain persons and associations, as objects of attachments and loyalties, should flourish” [21] (p. 302). In Justice as Fairness: A Restatement, Rawls describes such a conception as “an ordered family of final ends and aims which specifies a person’s conception of what is of value in human life or, alternatively, of what is regarded as a fully worthwhile life”; he adds that the elements of such a conception are “normally set within, and interpreted by, certain comprehensive religious, philosophical, or moral doctrines” (pp. 18–19).Setting aside the reference to “human” life, which may be viewed as question begging for present purposes, one initial difficulty is that animals usually cannot speak to us to articulate a conception of the good (subject to the possibility of, for example, a primate learning sign language, or other similar developments). Nonetheless, some animals can (and do) make choices as to their preferences. This is similar to what Tom Regan calls “preference autonomy”—that is, “the ability to initiate action because one has those desires or goals one has and believes, rightly or wrongly, that one’s desires or purposes will be satisfied or achieved by acting in a certain way” [22] (pp. 84–85). Moreover, there are animals that can signal their desires to us through other, non-verbal means. The question is whether acting in pursuit of certain stable preferences (and communicating them non-verbally) might, on one view, be taken as implicitly selecting a “conception of what is of value” or “worthwhile” in that animal’s life—in other words, an animal’s conception of the good.Consider, by way of illustration of this, a simple hypothetical example that may be familiar to any owner of a companion animal [23]: a domesticated dog that chooses a favorite chair to sit in in the house when his owner is away. Why he chooses this chair may not be altogether clear: maybe because it is in the sun, or because it is close to the food bowl, or for any other conceivable reason—perhaps because he just likes the chair. Nonetheless, he consistently chooses this chair. On one view, this preference reflects the dog’s “conception of what is of value” or “worthwhile” in his life. That is to say, he is making a choice as to what he prefers in a place to sit and thus what the dog is implicitly doing is to form and pursue a conception of what is of value or worthwhile in his life.Can a dog also “revise” his conception of the good, as required by the other verb Rawls uses? It is hard to see why not. The dog is certainly able to change his mind about his favorite chair. Perhaps, after a long time favoring a particular chair, he simply grows tired of it and decides to sit elsewhere in future. Or perhaps he has come to realize, walking on the way to his favorite chair, that a sofa in a different room gets the afternoon sun in a particularly appealing way, and that becomes his new favorite. In so doing, the dog revises his conception of the good. Moreover, although the example given is of a domesticated dog, the argument may apply equally to wild animals if they have the required abilities.However, there are at least three possible challenges to this view, being ways in which such behavior by animals—making choices as to their preferences—might be considered not to fully match Rawls’s description of this moral power. The first issue that could be raised concerns the immediacy of the aims being pursued. In defining this moral power, Rawls refers to having a “plan of life” or a “determinate scheme of final ends and aims”. In the case of animals, the conception of the good in question may not involve a “plan of life” in a long-term sense, or constitute a “determinate scheme of final ends and aims”—as far as we know, at least, the ends and aims are often shorter term, or in the case of some activities involving planning, medium term, rather than “final”.Nonetheless, it seems clear that animals can plan: consider the case of a bird building a nest for her young, or migrating to some distant place, or a squirrel burying nuts to eat later, or other examples of future planning documented in the scientific literature on “mental time travel” [24,25]. Such behavior appears to reflect ends and aims that, although not “final”, may sometimes lie rather distantly in the future. As Regan says, animals can be “individuals with desires, beliefs, and the ability to act in pursuit of their goals” (p. 116). Those goals are not always immediate and may be medium term or even somewhat long term. Nor is it the case that humans are constantly contemplating their “final” ends (we might also ask whether our ends and aims are ever meant to be truly final, until they turn out to be). Moreover, humans routinely engage in actions that are motivated by immediate or short-term desires; for many of us, long term planning is not a strong point. This does not deny humans moral personality, and nor should it for animals.The second issue is whether animals can have “desires that certain persons and associations, as objects of attachments and loyalties … should flourish”. The concern shown by a domesticated dog for an injured or unwell owner seems to suggest that this is possible [26]. Again, the dog might not have as developed a sense of the human’s life—and would not be thinking, for example, about the human’s job prospects or other aspects of a flourishing life that lie beyond the dog’s comprehension of human civilization—but it seems wrong to suggest that the dog lacks the desire for his owner, as an object of attachment and loyalty, to flourish. Once again, the limitations of the abilities of animals do not seem to deny their possession of moral personality entirely, but merely to constrain its extent.The third issue is whether animals are forming, revising and pursuing their conception of the good “rationally”, in the sense of a “rational” plan of life. Rawls explains that a rational plan of life involves a rational choice, and awareness of the relevant facts and consideration of the consequences (pp. 358–359). This may appear to be a point of difficulty for the argument. On the one hand, there is obviously a thought process behind animal choices, which may include consideration of the circumstances and the possible consequences of actions. Although it can be accepted that an animal is not espousing a comprehensive doctrine on one’s relation to the world, Rawls stipulates that this is “normally” an aspect of a conception of the good, a necessary qualification since many humans might not espouse any comprehensive doctrine; it does not seem required to demonstrate rationality. Rowlands observes that “the claim that animals are not rational is plainly false on at least some conceptions of rationality. Animals can identify causal relations and use their understanding of such relations to solve problems” [27] (pp. 78–79).On the other hand, a concern may be raised that animals are not reflective enough. As part of an argument that most animals do not have an interest in liberty, Alasdair Cochrane observes that “we are on reasonably safe ground to attribute to sentient animals the ability to possess and pursue desires” [28] (p. 667). However, seeking to reject the “autonomy” of particular animals, he claims that “[m]ost animals cannot frame, revise, and pursue their own conceptions of the good. This is not to say that sentient animals do not have different characters, nor is it to deny that they can make choices. It is simply to make the point that most animals cannot forge their own life plans and goals” (p. 668). As to why “most” animals are unable to do this, Cochrane relies on their (presumed) lack of reflective capabilities. He says that what is required is “the capacity to reflect on … desires, and modify them in relation to one’s own conception of the good” (p. 667).According to Cochrane, “such a capacity requires a level of consciousness above mere sentience. Indeed, one might claim that to be an autonomous agent one needs to possess ‘higher-order thought consciousness’, that is, to be able to have thoughts about thoughts” (p. 667). However, there are two difficulties with employing such a requirement as the basis for excluding animals from Rawls’s moral power of a capacity for a conception of the good. First, it seems to add to the list of verbs—what is required, on this approach, is not merely forming, revising and pursuing a conception of the good, but also reflecting upon it, or upon other desires in relation to it. It seems conceivable for animals to effectively pursue a conception of the good, embodying a set of stable preferences, without necessarily reflecting on other desires from the standpoint of that conception—simply by faithfully adhering to those stable preferences until they are revised.Second, although the scientific literature shows evidence of animal metacognition [29], and even deception [30], at this stage we do not know how reflective animals truly can be, including in relation to their desires. Given that some animals have stable preferences, it might be surprising if they were unable to assess other desires by reference to those preferences. A dog that chooses to sit in his favorite chair may well be considering and dismissing other impulses that pop into his head on the way there. However, we cannot say with any confidence how much reflection is taking place. That cuts both ways: a lack of reflectiveness is also not a sound basis on which to exclude animals from moral personality, because it is based on unproven assumptions about their abilities. Until scientific research clarifies the extent of animal metacognition, this will remain shaky ground for possible opposition to animals’ capacity for a conception of the good.In short, although it may be accepted that there are some limitations on animals’ capacity for a conception of the good, those limitations do not detract from the point that they often seem capable of forming, revising, and pursuing a conception of what is of value and worthwhile in their lives, depending, of course, on the abilities of the individual animal or species, but recalling that what is required for moral personality is merely a capacity, not its realization. In the case of animals, this capacity clearly does not rise to the level of sophistication of a human capacity for a conception of the good, and it perhaps may not extend as far into the future, but the difference might be viewed as one of degree, rather than of kind. The behavior of animals constitutes evidence that, rather than not possessing the moral power of a capacity for conception of the good at all, many animals might indeed possess it, albeit in a more limited sense than most humans do.3.2. A Sense of JusticeIn A Theory of Justice, Rawls defines the moral power of a capacity for a sense of justice as “a normally effective desire to apply and to act upon the principles of justice, at least to a certain minimum degree” (p. 442). In Political Liberalism, he refers to it as “the capacity to understand, to apply, and normally to be moved by an effective desire to act from (and not merely in accordance with) the principles of justice as the fair terms of social cooperation” (p. 302). In Justice as Fairness: A Restatement, he describes it in similar terms as “the capacity to understand, to apply, and to act from (and not merely in accordance with) the principles of political justice that specify the fair terms of social cooperation” (pp. 18–19). Any argument around animals’ sense of justice rests on incomplete scientific knowledge of their behavior and cognition, as well as difficult philosophical questions about the meaning of “justice” and a “certain minimum degree”. However, there is reason to doubt Rawls’s assumption that animals necessarily cannot have a sense of justice.It can be acknowledged at the outset that animals may not have a sense of justice in the sense of understanding, applying and acting from Rawls’s own principles of justice, which relate to the basic structure of a human society. These principles, particularly the difference principle, might not be comprehensible to many humans; we cannot suppose that there are any animals capable of understanding them. However, it seems problematic to say that animals therefore cannot have any sense of justice, because even if they cannot understand, apply and act from the principles of justice for a human society that are formulated by Rawls in his theory of justice, there is scientific evidence that suggests that some animals may have a sense of justice relating to their own social groups.Drawing on scientific research, Bekoff and Pierce argue that animals do have a sense of justice in this way (p. 113). They define justice asa set of expectations about what one deserves and how one ought to be treated. Justice has been served when these expectations have been appropriately met. Our justice cluster comprises several behaviors related to fairness, including a desire for equity and a desire … and a capacity to share reciprocally. The cluster also includes various behavioral reactions to injustice, including retribution, indignation, and forgiveness, as well as reactions to justice such as pleasure, gratitude, and trust.Contending that animals can be “moral subjects” that can act for moral reasons, but not “moral agents” that are morally responsible for their actions (issues not addressed in this article), Rowlands objects to the definition of justice offered by Bekoff and Pierce. He writes that “a ‘set of expectations about what one deserves and how one ought to be treated’ does not add up to justice—at least as this is ordinarily understood—for the simple reason that one can have unjust expectations about what one deserves or how one ought to be treated” (p. 30). He goes on to say:If we identify justice with any behavioral norm that happens to be in place in a social group, then who would wish to deny that animals have a sense of justice? That is, who would wish to deny that there are norms of behavior that regulate the interactions between members of a social group, and that an individual’s violation of those norms can often lead to conflict? Without such norms and their enforcement, it is difficult to see how any social group could remain in existence.(p. 32)However, there is a certain imprecision to the way in which Rowlands characterizes Bekoff and Pierce’s definition. According to Bekoff and Pierce, it is not any behavioral norm within a social group that constitutes justice; it is one relating to what one deserves and how one ought to be treated, a more specific subset of norms. Rowlands also objects that one may have “unjust expectations”. However, that seems to amount to merely saying that one’s sense of justice may be incorrect. The issue is perhaps better expressed not as whether one’s expectations are unjust or not—humans’ sense of justice may frequently be wrong—but whether the expectations relate to a concept of fairness (whether correct or not). Although this element of fairness is omitted from Bekoff and Pierce’s definition, they refer to fairness on many occasions elsewhere, so they too seem to be aware of its importance.There is a significant amount of scientific evidence that animals can have such expectations of treatment relating to fairness, and understand, apply and act from them. One famous example is a study in which Sarah Brosnan and Frans de Waal demonstrated inequity aversion in capuchin monkeys through an experiment in which pairs of monkeys exchanged tokens for rewards, and one was given a cucumber and the other a grape (a preferred reward) [31]. Brosnan and de Waal summarized their results as that “a nonhuman primate, the brown capuchin monkey … responds negatively to unequal reward distribution in exchanges with a human experimenter. Monkeys refused to participate if they witnessed a conspecific obtain a more attractive reward for equal effort, an effect amplified if the partner received such a reward without any effort at all” (p. 297). The capuchin monkeys appeared to respond negatively to a situation that they perceived as unfair, even forgoing their own reward due to the perceived unfairness. Similar results have been found in some other animals, such as chimpanzees [32].An additional example is social play behavior, which Bekoff and Pierce consider to offer “the most compelling evidence for a sense of fairness in social mammals”, both wild and domesticated (p. 115). They point out that some animals, such as dogs, coyotes, and wolves, “exhibit fairness during play, and they react negatively to unfair play behavior. In this context, fairness has to do with an individual’s specific social expectations, and not some universally defined standard of right and wrong” (p. 120). Unfair play can include behaving aggressively or attempting to mate. Thus, “during play, when a dog becomes too assertive, too aggressive, or tries to mate … the violation of trust stops play, and play only continues if the playmate ‘apologizes’ by indicating through gestures such as a play bow his intention to keep playing” (p. 121). Play, too, can be seen as an instance of animals understanding, applying and acting from a set of expectations—by enforcing them against other participants—about what they deserve and how they should be treated in a way that relates to fairness.The scientific literature is also replete with examples of animal altruism [33], and there is evidence for animal empathy as well [34,35,36]. Is this relevant to a sense of justice? The presence of altruistic behavior does not necessarily demonstrate a sense of justice, or at least not directly, because one can assist another without considering what that other should be able to expect in terms of what is fair. Nonetheless, altruism helps to illustrate the forms of social cooperation that exist among animals, to which inequity aversion and social play behaviors may be closely related.However, there are at least two possible challenges to the foregoing analysis of an animal sense of justice. The first—a possible response to any argument for animal morality, including one that claims that some animals can have a sense of justice—is that the behavior in question merely reflects instinct. However, to the extent that that is so, it does not seem to be a real point of differentiation between animals and humans, since in both cases moral behavior is the result of a combination of learned practices and instinct resulting from evolution. As Nussbaum explains:No doubt [animal] behavior has its basis in instinct, but here we should make two points. First, our own moral behavior is also based on our instinctual evolutionary endowment: an inherited tendency to assist others has helped humans to survive and flourish. Second, both humans and other animals need teaching to develop their instincts in an appropriate way. We see this cultural element in animal behavior clearly when we live with animals: dogs who are not appropriately trained behave lawlessly, and can even be schooled into dangerous aggression (as when pit bulls, who can be loving and cooperative, are trained, instead, to attack). Dogs who are well-trained internalize norms pertinent to their conduct.(p. 74)Indeed, discussing Rawls’s psychological account of how a sense of justice is acquired by humans (as set out in chapter VIII of A Theory of Justice), Robert Elliot points out that there are overt parallels between the process by which they “begin to understand their place in an association and to appreciate how they and others may benefit from co-operative, regulated behaviour” and cooperation in animal societies, such as a wolf pack [37] (p. 99). He adds:Rawls is quite prepared to support his principles of justice in a way that echoes ethologists’ explanations of animal behaviour. He says that “the capacity for a sense of justice and the moral feelings is an adaption of mankind to its place in nature” and elaborates, saying that, “for members of a species which lives in stable social groups, the ability to comply with fair co-operative arrangements and to develop the sentiments necessary to support them is highly advantageous, especially when individuals have a long life and are dependent on one another”.(p. 100)The second possible objection is that the expectations of treatment relating to fairness that some animals may apply and act upon do not correspond with the principles of justice as the fair terms of social cooperation that Rawls is invoking in his definition of this moral power. That objection has salience if the focus is on human principles of justice, including the principles of justice adopted in respect of human society’s basic structure in Rawls’s theory. As acknowledged above, we have no reason to consider that animals can understand, apply and act from these principles. However, what the evidence seems to suggest is that some animals may understand, apply and act from what might be described as principles that specify fair terms of social cooperation governing the workings of their own social groups.For example, monkeys might not be able to consider the justice of a human political institution, but monkeys can have societies of their own, albeit at a much smaller scale, whose justice they may consider when they take issue with the fairness of others members’ behavior. Further, as the inequity aversion study suggests, they also seem able to take issue with the fairness of human behavior towards them. Thus, although it may be accepted that animals, in applying their own sense of justice, would not be “tak[ing] part in and … act[ing] in accordance with the public understanding of the initial situation” in a Rawlsian sense (p. 442), the fact that animals can have their own societies suggests that their failure to consider the justice of human society may be beside the point in establishing their abilities in respect of this moral power.To the extent that some animals can have a set of expectations about treatment relating to fairness, they seem able to understand, apply and act from their own sense of justice, albeit to a lesser degree and at a smaller social scale than most humans. Nussbaum observes that rather than a “sharp binary division” between the moral rationality of humans and animals, there is really “a continuum”, reflecting a shared evolutionary history between our different species (p. 73). Indeed, humans can be said to be on a different point on a continuum from those animals that possess a sense of justice to a more limited extent. As with animals’ capacity for a conception of the good, Rawls’s claim that animals lack any capacity for a sense of justice appears to be, at best, a misleading characterization of a more nuanced and complex reality.That is not to suggest that recognition of an animal sense of justice would be a simple matter for Rawls’s theory; there is no doubt that it raises difficulties. One particular difficulty is that, even if some animals are accepted as having a sense of justice relating to their own social groups, so long as those animals do not have a sense of justice relating to a human society, they may be unable to engage with humans in a reciprocal way consistent with Rawls’s principles of justice as the fair terms of social cooperation. What this indicates is that, in order to be appropriately inclusive of animals and their abilities, a significant reconsideration of aspects of Rawls’s theory may be required. However, such a reconsideration seems to be unavoidable if the theory cannot convincingly explain the basis for the exclusion of animals from considerations of justice.4. ConclusionsAlthough Rawls does not dispute that animals merit moral consideration, they are omitted from his theory of justice. The basis for this omission is the purported absence of animal moral personality. However, this omission seems doubtful. Some animals may have, to a certain extent, capacities for a conception of the good and a sense of justice. The difference in this regard between humans and animals may be better viewed as one of degree, not kind. Consequently, a deficiency of Rawls’s approach is that it offers no recognition of the abilities of animals. As Garner says, “it seems entirely appropriate to include the interests of animals within the justice equation. To do otherwise implies a qualitative difference between humans and animals that the evidence suggests does not exist” (p. 13).In part, that reflects advancements in our understanding of the behavior and cognition of animals that had not occurred yet when Rawls originally published A Theory of Justice in 1971. It is possible that with the benefit of those advancements in understanding, Rawls would have had a different, more nuanced view about the position of animals within his theory, on the basis that it is too simplistic to say that animals as a class do not possess moral personality. However, it is also possible that he would have maintained the position that it is legitimate to confine the scope of a theory of justice to humans. Nonetheless, the limitation in scope to humans diminishes the persuasiveness of his theory if it is justified in a way that overlooks the actual abilities of animals.There are two possible approaches to addressing this exclusion of animals from Rawls’s theory. The first possibility is that, if it is accepted that some animals meet the threshold of moral personality, then those animals are acknowledged as moral persons entitled to justice. However, many animals might not be thought to meet the threshold. The second possibility is to reconsider moral personality, either to treat it as scalar or to recognize a lesser level of moral personality that encompasses the abilities of a larger group of animals. The problem is that if moral personality is considered scalar, then humans might not possess it equally among themselves. Alternatively, if a lesser level of moral personality is recognized, then it would be necessary to determine the scope and implications of this form of moral personality. These are difficult questions that are not sought to be resolved here.However, whichever approach is adopted to address the issue, if some animals were accepted as entitled to justice, then their interests could not be ignored in determining the basic structure of society, and the principles of justice. The formulation of the basic liberties of animals might become a key issue, with those basic liberties potentially tailored to the particular moral personality of animals. The implication may be that in a clash between human and animal basic liberties, humans would not necessarily win. Despite the difficulties shown in this article about the position of animals, moral personality is a concept that helps to shed valuable light on the normative basis for human basic liberties. There could even be some potential in the concept as a way to understand what the basic liberties of animals should be.Granting basic liberties to animals no doubt raises thorny issues about their position in relation to humans both in society and in nature. Would a human duty to respect animals’ basic liberties be restricted to those animals under our direct control, such as farmed animals and companion animals, or extend to those animals in nature that we affect? Alternatively, might there even be a duty to intervene in nature or natural processes to safeguard the basic liberties of animals? Nussbaum observes that, despite the “horrible suffering” inflicted on animals through predation in nature, “[w]e are very ignorant, and if we tried to interfere with predation on a large scale we would very likely cause disaster on a large scale” (p. 248). That is clearly so, and it is also not necessarily the case that we would be obliged to positively protect—as opposed to avoid violating ourselves—the basic liberties of animals. However, these matters, too, are left for another day.

animals : an open access journal from mdpi

10.3390/ani11061811

PMC8234185

Copper (Cu) is a trace element necessary for biological utility; nevertheless, it can produce significant harmful impacts when existing in abundance. This study examined the efficiency of vitamin C and vitamin E in alleviating the biochemical, genotoxicity, and pathological alterations in the liver induced by copper sulfate (CuSO4) toxicity in chickens. The broilers were fed on five experimental diets; basal diet with no additives or basal diets supplemented with 300 mg CuSO4/kg, CuSO4 + 250 mg Vit. C/kg diet, CuSO4 + 250 mg Vit. E/kg diet, CuSO4 + 250 mg Vit. C/kg diet + 250 mg Vit. E/kg diet for six weeks. The obtained results suggested that addition of vitamin C and E, especially in combination, was beneficial for alleviating the harmful effects of CuSO4 toxicity on growth performance and liver histoarchitecture in broiler chickens.

This experiment was carried out to explore the efficiency of an individual or combined doses of vitamin C (Vit. C) and vitamin E (Vit. E) in alleviating biochemical, genotoxicity, and pathological changes in the liver induced by copper sulfate (CuSO4) toxicity in broiler chickens. Two hundred and fifty-one-day-old broiler chicks were haphazardly allotted into five groups (five replicates/group, ten chicks/replicate). The birds were fed five experimental diets; (1) basal diet with no additives (CON), (2) basal diets supplemented with 300 mg CuSO4/kg diet (CuSO4), (3) basal diets supplemented with 300 mg CuSO4/kg diet + 250 mg Vit. C /kg diet, (4) basal diets supplemented with 300 mg CuSO4/kg diet +250 mg Vit. E /kg diet, (5) basal diets supplemented with 300 mg CuSO4/kg diet + 250 mg Vit. C /kg diet + 250 mg Vit. E /kg diet for six weeks. The results displayed that CuSO4-intoxicated birds had significantly (p < 0.05) decreased bodyweight, weight gain, and feed intake with increased feed conversion ratio from the 2nd week till the 6th week compared with the CON. However, these changes were minimized by single or combined supplementation of vitamin C and E. The FCR was insignificantly different in birds-fed diets complemented with vitamin C and E singly or in combination from the 3rd week of age compared to the CON. Serum aminotransferases (ALT, AST) and alkaline phosphatase (ALP) were elevated in CuSO4-intoxicated birds (p < 0.05). Additionally, they showed a drop in serum total protein (TP), albumin, globulins, triglycerides (TG), total cholesterol (TC), low-density lipoprotein-cholesterol (LDL-C), very low-density lipoprotein-cholesterol (VLDL-C), and high-density lipoprotein-cholesterol (HDL-C) levels compared to the CON (p < 0.05). Concomitantly, histopathological and DNA changes were perceived in the liver of CuSO4-intoxicated birds. Co-supplementation of Vit. C and Vit. E single-handedly or combined with CuSO4-intoxicated chickens enhanced the performance traits and abovementioned changes, especially with those given combinations of vitamins. From the extant inquiry, it could be established that supplementation of vitamin C and E was beneficial for mitigating the harmful effects of CuSO4 toxicity on growth performance and liver histoarchitecture in broiler chickens.

1. IntroductionThe proper activities of iron metabolism-related metalloenzymes are maintained by copper (Cu), a vital microelement involved in poultry diets [1]. Despite the need for unlike enzyme activities and metabolic processes, chronic overexposure to copper caused adverse effects [2]. This leads to cell death because of its ability to stimulate the production of Reactive Oxygen Species (ROS) [3]. Copper has a narrow optimum range between elemental and toxic concentrations [4]. Naturally, some types of soil contain Cu in toxic levels. In contrast, others may comprise high Cu levels through the human release of heavy metals into the environment through melting, mining, farming, industrial, and waste removal practices [5]. In poultry, the extreme type of Cu poisoning is the long-term ingestion of Cu compounds from different sources [6]. Copper metabolism, its release into the circulatory system or excretion through the bile, is controlled primarily by the liver [7]. It accumulates steadily in the liver during chronic Cu toxicity without causing any noticeable signs or symptoms. When the liver storage capacity is surpassed, it can lead to hepatocellular lesions and the Cu releases into the blood circulation causing jaundice, hemolysis, and renal disease [8]. The studies mentioned above have designated that excessive exposure to Cu can cause oxidative stress in the brain tissue of chickens [9], reduce the glutathione peroxidase and copper-zinc superoxide dismutase activities, and increase the contents of hydroxyl radical and malondialdehyde in the liver of ducklings [10]. A discrepancy between production of ROS and the ability of the body to detoxify these intermediate species is indicative of oxidative stress. Wang et al. [4] showed that Cu encourages oxidative damage in skeletal muscles of chickens through autophagy, apoptosis, and mitochondrial dynamics, thus expressing fears about poultry raising zones polluted with Cu. Moreover, Wang et al. [11] reported nephrotoxicity in Cu-intoxicated chickens due to oxidative damage of the kidney. Although Cu supplementation of up to 200 ppm is necessary to promote growth, excessive amounts of dietary Cu reduce growth. It reduces the digestibility and absorption of copper in poultry, which leads to increased excretion in the faeces and environmental pollution. [12].Few reports have displayed that supplementation of antioxidants such as vitamin C vitamin E, polyphenols, alpha-lipoic acid, beta-carotene, and zinc has a protective impact against the toxicity of Cu [13,14]. However, only restricted studies have been achieved on the effects of vitamin C and E single-handedly or incorporating on growth performance, biochemical markers, DNA damage, or pathological findings in broilers fed excess dietary Cu. Consequently, the ambition of the existent inquest was to assess the effects of Cu intoxication on these parameters and then appraise the protecting impact of vitamins C and E against excess dietary supplementation with Cu, individually or in combination. Therefore, this experiment aimed to assess the mitigating effects of single or combined addition of vitamin C and E on the harmful impacts of Cu toxicity in broiler chickens.2. Material and Methods2.1. Experimental Birds, Diet, and ProtocolThis study was conducted in a poultry research unit in the faculty of veterinary medicine, Zagazig University, Egypt. The ethics of the experimental protocol were approved by the Institutional Animal Care and Use Committee of Zagazig University, Egypt (ZU-IACUC/2020). All animal experiments were performed following the recommendations described in “The Guide for the Care and Use of Laboratory Animals in scientific investigations.”Two hundred and fifty-one-day-old commercial broiler chickens (COBB-500) were attained from Al-Kahira Poultry Company, 10th of Ramadan City, Sharkia Governorate, Egypt. The experiment lasted for 42 days with good ventilation. Birds were raised in an open, well-ventilated house with sawdust. Room temperature was controlled and thermostatically regulated by two heaters. Room temperature during the first week was set at 34 °C and gradually reduced by 3 °C every week until it reached 24 °C. The light program for the first week was 24 hours a day and then changed to 16 h of light and 8 hours of dark over 7 to 42 days.Freshwater and feed were offered for ad libitum consumption throughout the experiment. The chicks were given a starter diet from one day until the 10th day of age, a grower diet (11th–22nd day), followed by a finisher diet up to 42-days of age. Ingredients and chemical composition of diets were formulated as designated by the COBB-500 broiler manual guide [15] (Table 1). All birds were vaccinated at 7 and 14 days old against Newcastle disease and 11 and 22 days old for Gumboro disease [16].The chicks were haphazardly allotted into five experimental groups (five replicates/group, 10 chicks/replicates). The birds were fed on five experimental diets; (1) basal diet with no additives (CON), (2) basal diets supplemented with 300 mg CuSO4/kg diet (CuSO4), (3) basal diets supplemented with 300 mg CuSO4/kg diet + 250 mg Vit. C /kg diet, (4) basal diets supplemented with 300 mg CuSO4/kg diet + 250 mg Vit. E /kg diet, (5) basal diets supplemented with 300 mg CuSO4/kg diet + 250 mg Vit. C/kg diet + 250 mg Vit. E/kg diet for six weeks. Copper sulfate (CuSO4·5H2O, El-Gomhoria industry, Zagazig, Egypt), vitamin C (ROVIMIX® STAY-C®35, DSM, Heerlen, The Netherlands), and vitamin E (α tocopherol acetate, Pharco Pharmaceutical Industries, Zagazig, Egypt). The toxic dose of Cu used in this inquest was dogged according to Cinar et al. [17], while the vitamins C and E doses were used after Sahin et al. [18].2.2. Growth PerformanceThe average initial body weight (BW) was recorded at the beginning of the experiment. The BW was then determined every week, and body weight gain (BWG) was determined [19]. The difference between the weight of the provided feed and the feed that remained was used to calculate feed intake (FI) per replicate. Then, the feed conversion ratio (FCR) was calculated. FCR = amount of consumed feed (g)/BWG (g).2.3. SamplingSamples of blood were collected from the wing vein of 10 randomly selected birds in each group at the termination of 3rd and 6th weeks post-supplementations and centrifuged (3000 rpm for 15 min) immediately for separation of serum, which is stored at −20 °C in deep freeze until biochemical analysis [20]. The chicks were euthanized using cervical dislocation, according to the American Veterinary Medical Association (Schaumburg, IL, USA) guidelines [21], and two portions of liver tissues were separated and blotted dry. The first part was put in ice-cold PBS (phosphate buffer saline) for comet assay determination, and the 2nd part was fixed in ten percent formalin for histopathological inspection.2.4. Blood Biochemical StudiesThe serum activities of alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were determined according to the method of Reitman and Frankel [22]. The serum alkaline phosphatase activity was measured according to the modified method of Moss [23].The serum total protein levels were estimated according to Grant [24]. The serum albumin level was evaluated according to Doumas et al. [25]. According to Doumas and Biggs [26], the serum globulins levels were calculated mathematically by subtracting albumin values from total protein values.The total serum lipids, total cholesterol, triglyceride, and high-density lipoprotein (HDL-c) were estimated according to the methods of Zöllner and Kirsch [27], Roeschlau et al. [28], McGowan et al. [29], and Young [30], respectively. Low-density lipoprotein (LDL-c) and very-low-density lipoprotein cholesterol (VLDL-C) were calculated mathematically according to the following relationship described by Friedewald et al. [31].2.5. Detection of DNA DamageThe liver DNA damage was measured using a single-cell gel electrophoresis technique (also known as comet assay) as previously defined by Singh et al. [32]. Comet assay is a quick, accurate, and simple method for detecting DNA damage. In this method, 0.5 g of crushed samples were transferred to 1 mL ice-cold PBS. This suspension was stirred for 5 min and filtered. The cell suspension (100 μL) was mixed with 600 μL of low melting agarose (0.8% in PBS). A 100 µL of this mixture was spread on precoated slides. The coated slides were immersed in lysis buffer (0.045 mol/L Tris/Borate/EDTA (TBE), pH 8.4, containing 2.5% sodium dodecyl sulfate (SDS)) for 15 min. The electrophoresis conditions were 2 V/cm for 2 min and 100 mA. The slides were then washed 3 times, for 5 min each, with neutralization buffer (0.4 Mol/L Tris (pH 7.5)). Finally, the slides were stained with 50 AL of ethidium bromide (2 mg/mL) and covered with a coverslip.The DNA fragment migration patterns of 100 cells at 400 magnifications with the Optika Axioscope fluorescence microscope were calculated for each dose level. The length of DNA migration (tail length) on PX was calculated for each cell from the center of the nucleus to the termination of the tail. By calculating the total intensity (fluorescence) in the cells, which was taken as 100%, the DNA% in the tail was determined, deciding what percentage of this total intensity corresponded to the intensity only measured in the tail. The tail moment was expressed in arbitrary units. Although any image analysis device may be sufficient for SCGE data quantification, Comet 5 image analysis software developed by Kinetic Imaging Ltd. (Liverpool, UK) linked to a CCD camera has been used to determine the degree of quantitative and qualitative DNA damage in the cells by measuring the length of DNA migration and the % of migrated DNA. Finally, the program calculated the tail moment. Generally, 100 randomly selected cells are analyzed per sample.2.6. Histopathological InvestigationsSamples were taken from the liver of euthanized chicks by manual cervical dislocation and fixed in formalin of 10%. The samples preserved with formalin are dehydrated and embedded in paraffin. Five-micron-thick paraffin slices were set and stained with hematoxylin and eosin (H&E) and inspected microscopically [33].2.7. Statistical AnalysisData were analyzed with a one-way analysis of variance (ANOVA) using the GLM procedure in SPSS (SPSS Inc., Chicago, IL, USA) after Shapiro–Wilk’s test was used to verify the normality and Levene’s test was used to verify homogeneity of variance components between experimental treatments. Duncan’s test was used to compare the differences between the means at 5% probability [34]. Variation in the data was expressed as mean ± SD, and the significance level was set at p < 0.05.3. Results3.1. Clinical Signs and Body PerformanceNo clinical signs or mortality were found in all supplemented birds (single or combined Vit. C and Vit. E addition) during the experimental period. The CuSO4-intoxicated group showed mild diarrhea (few cases), decreased appetite, and pale comb.As presented in Table 2, the BW, BWG, and feed intake of broilers were significantly declined in all groups from the 2nd week till the termination of the experiment (6th week) compared to the CON group (p < 0.05). However, these changes were minimized in birds supplemented with vitamins C and E compared with CuSO4-intoxicated group. The FCR was significantly higher in CuSO4-intoxicated broilers at the 2nd to 6th week of age compared to the control group (p < 0.05). Comparatively, with the CON group, the FCR was insignificantly different in birds fed diets supplemented with vitamin C, and E singly or in combination from the 3rd week of age and went back to near average control values at the 6th week (p < 0.05).3.2. Serum Levels of Liver BiomarkersAs shown in Table 3, Table 4, Table 5 and Table 6, CuSO4 induced hepatotoxicity as reflected statistically (p < 0.05) by increased serum activities of ALT, AST, and ALP, whereas serum TP, albumin, globulins, TG, TC, LDL-C, VLDL-C, and HDL-C levels were reduced at 3rd and 6th week compared to CON (p ≤ 0.05). The decrease in serum albumin and HDL-C was only in the 6th week period.On the other hand, co-supplementation of copper with Vit. C and Vit. E significantly lowered the serum AST, ALT, and ALP and increased serum TP, albumin, globulins, TG, TC, LDL-C, VLDL-C, and HDL-C levels in the CuSO4 + Vit. C and CuSO4 + Vit. E groups compared with CuSO4-induced hepatotoxicity group. However, co-administration of CuSO4 + Vit. C + Vit. E restored liver biomarkers’ changes near-normal control values.3.3. DNA DamageThe data in Table 7 and Table 8 and Figure 1 revealed that CuSO4 intoxication significantly elevated the comet %, %DNA in the tail, tail moment, olive tail moment, and tail length at the end of the 3rd and 6th week parallel to CON (p < 0.05). On the contrary, co-administration of CuSO4 + Vit. C, CuSO4 + Vit. E, or their combination (CuSO4 + Vit. C + Vit. E) to birds ensued in an enhancement in the results of comet assay performance, showing a substantial decrease in previous parameters relative to the CuSO4-intoxicated group but did not return to values of standard control. It was evident that the vitamin combination led to a significant decrease in comparison to the CuSO4-intoxicated group.3.4. Histopathological FindingsThe livers of the control chicken showed typical histological arrangement of hepatic lobules at the 3rd and 6th weeks (Figure 2a,b), respectively. In contrast, the livers of chicken from the CuSO4-intoxicated group displayed hyperplastic and necrotic biliary epithelium with various degenerative and necrotic changes at the third week (Figure 3a). Additionally, cholestasis, necrotic bile duct epithelia, besides lymphocytic portal aggregation and fibroblast proliferation, were encountered at the 6th week (Figure 4a). Liver from chicken of the CuSO4 + Vit. C group showed moderate enlargement of hepatic cells and hyperplastic Kupffer cells at the 3rd week (Figure 3b) and partially contracted hepatic cells proliferative Kupffer cells and dilated sinusoids at the 6th week (Figure 4b). Liver from chicken of the CuSO4 +Vit. E supplemented group revealed lymphocytic portal aggregations within apparently normal hepatic parenchyma at the third week (Figure 3c), while intense hyperplasia of Kupffer cells and cloudy swelling of hepatic cells with a few lymphocytic aggregations were seen at the 6th week (Figure 4c). Liver of the CuSO4 + Vit. C + Vit. E group displayed little portal and interstitial lymphocytic aggregations with apparently normal hepatic parenchyma at the third week (Figure 3d), mild portal lymphocytic aggregation, normal hepatic parenchyma, and dilated blood vessels were observed at the 6th week (Figure 4d).4. DiscussionAlthough the level of copper up to 100 to 200 mg/kg as CuSO4 improves growth performance in the broilers [9,35,36,37], incompatible effects such as poor feed intake, decrease in body weight, and hematobiochemical changes at higher doses of copper have been reported [38,39,40]. In this study, no clinical signs or mortality were perceived during the experimental period in Vit. C and E supplemented groups, while CuSO4-intoxicated birds showed mild diarrhea, anorexia, and weight reduction. This was synchronized with Luo et al. [41], who reported zero mortality in male chicks supplemented with 300 and 450 mg/kg CuSO4 for 21 days. Other studies reported no significant differences in mortality % of broiler chickens among the control and other groups given 0, 10, 25, 50, 125, 250, and 500 CuSO4/kg diet for 35 days [42] or 100, 200, and 400 mg CuSO4/kg diet for 42 days [43].The exiting investigation showed that copper sulfate had a toxic effect, as exhibited by a statistical reduction in growth efficiency parameters (BW, BWG, and FI), increasing the FCR over the entire growth-out period (days 1–42) in CuSO4-intoxicated birds. These findings may be attributed to falling in the feed intake and utilization as a result of GIT disturbance caused by a highly toxic dose of Cu or owing to the embarrassment of the satiety center by Cu, leading to loss of interest in feeding [44]. Additionally, the reduction in feed consumption could be due to an anorexic effect of CuSO4 on chickens [41,45]. Our results were consistent with Luo et al. [41], Cinar et al. [17], Abduljaleel [46], Scott et al. [36], and Zhou et al. [47]. Previous studies recorded different outcomes such as unchanged BW and FCR in chicks fed up to 400 mg Cu/kg [48] and improved live weight gain in broiler chick supplemented with 400 mg/kg feed copper sulfate for 6 weeks [43]. The reason for variances between our experimental results and others may be concurrent with the varieties in the CuSO4 particle size, stain, and age of birds, experimental periods, and the dose of copper between earlier and current experiments. Furthermore, the difference between the effects of Cu levels on the growth performance of raising poultry indicates that the bioavailability of Cu can differ. This was supported by other literature stating that copper at sufficient dietary levels has favorable impacts, but severe toxic effects as a result of extra copper are well established [49,50]. The study of Morsy et al. [51] investigated the oral administration of copper oxide nanoparticles (CuO-NPs) at doses 5 mg/kg and 15 mg/kg BW in broiler chickens. They showed a notable decrease in weight gain, FCR in a level-dependent manner. The growth parameters were upgraded meaningfully in CuSO4 + Vit. C and CuSO4 + Vit. E-supplemented groups compared to CuSO4-intoxicated group; however, better results have been found in birds taking both Vit. C and Vit E. Many reports showed better performance by feeding Vit. C or Vit. E to broiler chicks [17,46,52,53,54] or fish [55,56,57,58,59]. The improvements in performance characteristics in vitamin-supplemented groups could be attributable to their antioxidant effects that protect the bird from the oxidative stress caused by Cu exposure [60,61] or the role of vitamins as an immune stimulant [62].Birds supplemented with excess CuSO4 revealed an elevation in serum transaminases (ALT and AST) and ALP activities correlated with diminished serum TP, albumin, and globulins levels equated to the control group, suggesting significant liver damage (hepatotoxicity), which was established by hyperplastic and necrotic biliary epithelium, hepatic degeneration, and necrosis in the histopathological findings. The increased transaminase activities in the CuSO4-intoxicated group may be associated with the influence of Cu on the liver and kidney, so releasing their intracellular enzymes to circulation [63] or due to the cytotoxic effect of Cu, resulting in lipid peroxidation indicating the hepatotoxic effect of Cu [64]. In addition, Kumar et al. [43] determined that a higher level of copper accumulation may have injured the liver to rise these enzymes. Abundant studies indicate that copper can be metabolized in hepatic tissues and converted by glutathione (GSH) into metallothionein; Therefore, copper excess is got, and the immediate results of GSH are depleted of enhanced cytotoxicity [65].The finding of high serum enzymes in Cu-intoxicated broilers agreed with Yigit et al. [49], who stated that copper produced changes in the liver transaminases of broilers. As a result of CuSO4 addition, the upsurge in ALP activity might be recognized for liver, intestine, kidney, and to some degree, bile duct injuries, especially liver cell membrane, which appears to act as a stimulus to increase the synthesis of this enzyme or may be due to hepatic or bile duct cholestasis causing enzyme regurgitation back into the bloodstream [66].The reported hypoproteinemia in this inquiry may be due to impairment protein synthesis or the functional deterioration of the liver and excessive loss of protein caused by nephrosis [67] or could be explained due to the oxidative stress of copper on the liver and kidney tissue [11,39]. Since albumin has unique copper ion binding sites and carries dietary copper to the liver [68], the distinguished drop in the serum albumin levels may be due to the decrease in the synthesis of hepatic albumin in the CuSO4 group. Contrary to our results, Almansour [69] reported intensification in serum protein levels in copper-fed quail. Cinar et al. [17] displayed no alteration in plasma proteins of 300 mg/kg diet copper-fed bird. Dietary addition of Vit. C and Vit. E alone or in combination with CuSO4-intoxicated birds significantly ameliorated the previous changes in liver enzymes, demonstrating that these vitamins have potential protecting effects on cell membranes, thus preventing enzyme leak into the blood [17,70]. Better results were evident with vitamin supplementation than each alone because some liver biomarkers were returned near the average control values. These results were in synchronization with those attained by [71], Abou-Kassem et al. [72], Mashkoor et al. [73], and Hashem et al. [20]. The hepatoprotective influence of vitamin C or E is associated with their antioxidant properties [74]. Additionally, ascorbic acid causes its association with heavy metals, which leads to reduced tissue oxidative stress and restored AST, ALT, ALP and LDH levels [75]. Besides, vitamins C and E may preserve the hepatic cellular membrane and protect hepatocyte from copper’s toxic effects, minimizing the enzyme’s leakage into the bloodstream [70]. In dissimilarity, supplementing cadmium-intoxicated broilers with vitamins C and E did not recover transaminase activities [76].Co-supplementation of copper with Vit. C and Vit. E alone or in combination elevated serum TP, albumin, globulins. The enhancement in the protein profile in Cu-exposed birds fed with Vit. C and Vit. E alone or in combination may be due to the immunostimulant effect [77] or due to impairments of the copper uptake and utilization [17] by dietary supplement of Vit. C or Vit. E. Imik et al. [78] reported a high total protein concentration in blood of quail supplemented vitamin E and C in the diet. Therefore, only the co-administration of copper with both vitamins had a protecting effect to hepatotoxicity caused by CuSO4.Regarding lipogram analysis, CuSO4 administration causes dyslipidemia, evidenced by a numerical reduction in serum TG, TC, LDL-C, and VLDL-C levels; however, serum HDL-C did not display any change in the third week but significantly dropped at 6th week in CuSO4—intoxicated birds compared with the control group. Our findings were consistent with Bakalli et al. [79] and Idowu et al. [71]. Moreover, Wu et al. [80] noticed significantly lessened serum cholesterol and LDL-C levels in broilers fed 3 sources of copper (copper methionate, tribasic copper chloride, and copper sulfate) in the diet. The reduction in plasma cholesterol and triacylglycerol in the blood of Cu-exposed chickens is due to fall cholesterol synthesis, high degradation, or excretion rates [81]. The excess level of Cu supplementation to diet either declines GSH that diminished stimulation of HMG-CoA reductase activity resulted in the reduced synthesis of cholesterol [82], or lead to changes in lipid metabolism, which result in decreasing plasma lipid, 17 beta-estradiol, and hepatic lipogenic enzyme activity [83]. The reduction in LDL level was related to copper toxicity because Cu is an effective catalyst of LDL-C oxidation to an atherogenic form [84] or alkoxyl radicals [85]. The reduction in HDL-C is due to hypocholesterolemia and hypoproteinemia in this inquiry, as more than 40% of HDL-C value represents cholesterol value and the remaining proteins [86]. The same findings were obtained by El-Hady and Mohamed [87] in broiler chickens with dietary supplementation of CuSO4 at levels 50 and 100 ppm for 5 weeks. However, earlier studies exposed different outcomes, for example, rise of total lipid, cholesterol, and LDL-C with no variation HDL-C level in Cu-exposed quail [69] or no alteration in plasma total cholesterol levels in broilers [17]. Additionally, Jegede et al. [35] informed a decrease in plasma triglycerides and cholesterol in Arbor-Acre unsexed broilers fed CuSO4 or copper proteinate at concentrations of 50, 100, or 150 mg/kg diet for 56 days. The breed, dietary components, and the investigational strategy and methodology can explain these differences.Dietary supplementations of CuSO4-intoxicated birds with Vit. C and Vit. E alone or in combination reduce the opposing effects of copper on lipid profile. Treatment of Cu toxicity with Vit. C and Vit. E combination is more effective than using each one separately. This can be due to the ability of the vitamin C to replenish the vitamin E re-mobilizes the free radicals associated with it [88]. The combination of vitamins protects lipid structures against peroxidation [70]. Vitamin C guards against Cu toxicity by preventing excess Cu absorption by minimizing copper absorption from the intestine by reducing soluble Cu levels in the small intestine [89,90]. Besides, Vit. E can prevent cholesterol-related endothelial dysfunction, preventing functional impairment induced by ROS [13].Although the micronucleus assay has been useful and its utility should be considered [91], the comet assay is a more sensitive method for assessing genotoxicity [92]. The true comet assay was used to test for genotoxic agents [93]. Collaboration with others is a sensitive and fast method for distinguishing DNA damage produced by trace metals, such as copper [94]. Data from this research showed that CuSO4 has the genotoxic ability to interact with DNA and induce mammalian cell alterations specified by elevation in comet %, DNA % in the tail, tail length, tail moment, and olive tail moment. Similarly, Banu et al. [95] indicated significant DNA damage with a decrease in mean comet tail-length after adding CuSO4. Copper-induced DNA damage may or may not be constrained at low copper concentrations, as it is closely bound to storage or transport proteins (eg. ceruloplasmin) and thus not available for oxidative reactions [96], but at high concentrations, free Cu can have an enormous genotoxic effect [91]. Free Cu causes ROS and multiple types of DNA damage, such as base alteration and DNA strand breaks, which can cause severe cell death [97]. A copper-induced high ROS production consequence in oxidative destruction to a single DNA base and sugar phosphate and breaks DNA strands [98]. Additionally, copper reduces DNA-binding cell viability, resulting in cell death [99].The current study revealed that supplementing of CuSO4-intoxicated chickens with Vit. C and Vit. E alone or in combination exert a partial genoprotective effect against DNA damage induced by excessive concentrations of copper, which was proved by the decrease of comet%, tail length, and moment. A combination of vitamins is more effective in reducing the genotoxic effects of copper than using each vitamin alone. The genoprotective effect of Vit. C and Vit. E could act as free radical scavengers and antioxidants [100]. Jiraungkoorskul and Sahaphong [101] demonstrated that ascorbic acid reduces genotoxicity in fish induced by copper. Assy et al. [102] exhibited a defending effect of Vit. E (100 mg/kg diet) against DNA damage in rats with CuO nanoparticles toxicity (250 mg/kg diet). Our outcomes are in settlement with a new study reporting potential anticipatory effects of dietary antioxidants, including vitamin E, ascorbic acid, phytosterols, polyphenols, and medicinal plants extracts against vanadium-induced DNA damage [103]. Another study showed that curcumin supplementation reduced genomic and cellular DNA damage in mice exposed to 390 ppm Cu [91]. Morsy et al. [51] reported increased DNA fragmentation percent and microscopic recording in various inspected organs of chickens received CuO-NPs.Histopathological check established the results of CuSO4-intoxicated chickens in the existing inquest and it is in coordination with the earlier report of Oguz et al. [38], Shahzad et al. [40], Wang et al. [11] in chickens and Baruah et al. [104] in ducks with copper toxicity. Similarly, other studies exhibited a significant inflammatory cell infiltration and hepatocyte vacuolar degeneration [105], severe microscopic changes, including vacuolar degeneration, local tissue necrosis, and blurred hepatic lobules in birds-supplemented with 300 mg Cu/kg diet [47]. Morsy et al. [51] observed histopathological changes in chickens getting CuO-NPs with some differences in its severity. Our results differ from other findings that showed that birds fed a high level of copper diet did not alter any histological tissue [106]. These inappropriate associations may be attributed to the variance in the types of test subjects, experimental animals, and test duration. Long-term exposure of birds to high levels of copper, dangerous effects may occur [47]. However, single or combined addition with Vit. C and Vit. E decreased the histopathological alterations with apparently typical liver architecture in some cases. Other findings presented that the addition of vitamins C and E in the feed has successfully offset arsenite’s toxic effects in broiler chicken [107].5. ConclusionsThe obtained findings could conclude that long-term exposure to CuSO4 caused significant alterations in the liver evaluation biomarkers, genotoxicity, and histopathology. Dietary addition of vitamins C and E reduced the harmful effects induced by CuSO4, especially with their combination, which caused an improvement in the growth performance; returned the biochemical parameters in close to average values, with subsidence the histopathological changes and DNA degeneration. Overall, the protective roles of vitamins C and E with their synergistic action against the toxic effects of Cu are seen in our research, but further studies are still needed to understand the full potential of vitamins.

animals : an open access journal from mdpi

10.3390/ani11113066

PMC8614441

Exogenous fibrolytic enzymes can improve nutrient digestibility of feeds high in fibrous content offered to Egyptian lactating buffaloes. The proposed cellulase exclusively produced in-farm using Penicillium Chrysogenum showed higher activity in previous in vitro studies. That is why it was chosen to get tested against a well-known commercial source of cellulase enzyme from the Egyptian markets for its efficiency in increasing milk productivity and composition. Profiles of amino acids and fatty acids were also recorded. The initial results highlighted a superiority of the produced enzyme (FENZ) against the commercial source (CENZ). It was also clear that FENZ can preserve higher proportions of fatty acids in the milk, primarily conjugated linoleic acid. Based on the idea rationale, our conclusion is to promote setting a small cellulase production unit in each farm in Egypt to decrease the cost of feeding by using agricultural and agro-industrial waste during the cellulase production and feeding process.

The experiment was conducted to study the effects of supplementing a cellulase enzymes cocktail to lactating buffaloes’ diet, on the nutrient intake, nutrient digestibility, and milk production performance and composition. Twenty-four lactating Egyptian buffaloes were assigned into one of the following treatments: CON—control consisted of a total mixed ration, CENZ—the total mixed ration supplemented by a commercial source of cellulase enzyme, FENZ—the total mixed ration supplemented with cellulase enzyme cocktail produced in-farm. Supplementing the diet with the in-farm source of cellulase (FENZ) had a significantly higher impact on crude protein, neutral detergent fiber, and acid detergent fiber digestibility. However, FENZ tended to increase the EE digestibility compared to CENZ. FENZ showed significantly higher nutrient digestibility percentages compared to other groups. Supplementing the diet with cellulase enzymes (CON vs. ENZ) significantly increased the daily milk yield and the fat correct milk yield; both yields were significantly higher with FENZ than all groups. Oleic, linoleic, and linolenic acid concentration were significantly higher with cellulase enzymes supplementation (CON vs. ENZ) and the conjugated linoleic acid concentration. Supplementing fungal cellulase enzyme produced on a farm-scale has improved milk productivity, fat yield, and milk fat unsaturated fatty acids profile in lactating buffaloes.

1. IntroductionRuminant-based agriculture relies on several key processes to reach food security worldwide; one of those keys is the ability to provide farm animals products (meat or milk) that cover the increased needs of rising populations, mainly by providing animal feed sources that do not compete with human crops land area [1,2]. Most ruminant feeds consist of forages with high fibrous content, mainly because they can degrade plant cell wall material by the rumen microorganisms and their associated enzymes. However, rumen digestion is not perfect, and usually a high amount of plant fiber bypasses the digestive tract without being used [3,4]. Plant biomass’s significant portion (up to 50% dry weight) consists of cell wall cellulose polymers. It is well known that cellulose, consisting of β-1,4 glucosidic bonds linking D-glucose molecules, is by far the ultimate renewable resource naturally [5]. Recently, there has been a rising interest in using agro-industrial waste and agroforestry products in ruminant nutrition. Regrettably, those materials’ nutritional composition is unbalanced and requires the support of high-quality grains, legumes, or additives to meet the nutritional needs [2].Additionally, the high number of cross-links within plant cell wall carbohydrates and lignin through the growing season decrease digestibility and bounds the utilization of forages by ruminants. Consequently, it will demand several cellulolytic enzymes to degenerate these materials effectively [6,7,8]. As an outcome, many strategies have been exploited to enhance fiber degradation in the rumen. One of those strategies is the application of exogenous fibrolytic enzymes. Enzymes as a feed additive have been used extensively in the last decades to improve the nutritive value of diets by enhancing fiber digestion and, respectively, increase productive ruminant performance [9]. Research has shown that anaerobic fungi categorized extensively in the digestive tracts of ruminants can utilize various carbohydrates and maintain an efficient glycosyl hydrolase system that hydrolyzes plant carbohydrates [6,10]. The mechanisms by which cellulases’ enzyme from anaerobic rumen bacteria digest cellulose are poorly defined. However, the cellulases’ network in fungi consists of three cellulases: (1) endoglucanases, which randomly hydrolyze the 1,4-β-glycosidic bonds within cellulose, producing oligosaccharides with reducing and non-reducing ends; (2) exoglucanases, which separate cellobiose units from their nonreducing ends, and (3) β-glucosidases, which hydrolyze cellobiose and low-molecular-weight cellodextrins, producing glucose [11,12,13]. It is well known that fungi are the main cellulase-producing microorganisms. The Penicillium chrysogenum is a well-known cellulase producer with numerous commercially available products for agricultural and industrial uses. However, there is not enough knowledge on Penicillium chrysogenum produced cellulases on ruminant nutrition and performance [11,12].On the other hand, agricultural and agro-industrial wastes like sugarcane bagasse, rice straw, wheat straw, orange peel, and palm fronds continually expand due to industrialization, and their disposal becomes a problem concerning area and creating environmental pollution. However, these wastes could serve as a cheap alternative source for microbial growth and biomass or enzymes production. This production system emerged as an appropriate technology for managing agro-industrial residues, with numerous advantages, including less pollution, high volumetric productivity, and a relatively greater concentration of products [14,15]. The current study aimed to investigate the effect of supplementing cellulase enzyme cocktail produced in-farm scale from Penicillium Chrysogenum to the diet of lactating buffaloes on the nutrient digestibility, milk production performance, milk amino acids, and fatty acids profile when compared to a commercial cellulase enzyme product from the Egyptian markets. We hypothesized that applying the in-farm enzyme cellulase will better impact the buffaloes’ digestibility, productivity, and milk fatty acid content due to the high activity of the produced fungal source of cellulases.2. Materials and MethodsThe experiment conducted in this study was carried out on Egyptian lactating buffaloes at ALSATAR farm in Khatatba City, Monufia Governorate, Egypt. The buffaloes were cared for following the animal research guidelines and ethics of the Ministry of Higher Education and Scientific Research, Egypt, and authorized by the local ethical committee of the National Research Centre, Egypt.2.1. Experimental Design, Treatments, and Management2.1.1. Enzymes SupplementsCellulolytic enzyme cocktail was produced in a small pilot plant prepared to set and prepared on the farmland for enzyme production using the methodology described earlier in [16]. Briefly, dried (70 °C for 24 h) and ground rice straw, wheat straw, and palm fronds were used as the primary substrate materials for the fungal cultures at 50%, 30%, and 20%, respectively. The fungal strain of Penicillium chrysogenum obtained from the Laboratory of Plant Pathology, National Research Centre, Cairo, Egypt, was used to produce the cellulase enzyme using the mentioned wastes as a source of cellulose. Briefly, 1000-mL conical flasks each containing 100 mL of cellulose powder medium that consisted of NaCl, 6 (g/L); (NH4)2SO4, 1 (g/L); K2HPO4, 1 (g/L); MgSO4.7H2O, 0.05 (g/L); CaCl2, 0.1 (g/L); yeast extract, 0.5 (g/L); Peptone, 0.5 (g/L); primary substrate powder, 4 (g/L). The medium pH was adjusted to 6 and then the growth cultures of fungi were employed as inoculant at rate of 5% (v/v) inoculum size, the fermentation then lasted three days at 30 °C. The fermented substrate for each flask was then mixed with 25 mL of 0.02 M acetate buffer (pH 5.0) to extract the enzyme, by shaking in a rotary shaker (120 rpm) for one hour at room temperature, that to extract the enzyme from the solid-state media. The activity of the extracted enzymes was measured in the obtained filtrate based on carboxymethyl-cellulase activity (CMC) and the reducing sugar liberated which was determined following the method of Dinitrosalicylic acid [16]. One cellulase unit is defined as the amount of enzyme that liberates reducing sugar at the rate of one µmol/mL/min under assay conditions. For comparison, a commercial source of cellulase enzyme (Pan-Zyme; BAYTARA for pharmaceuticals technology, Sadat Industrial City for Guangdong VTR Bio-Tech Co., Ltd., Guangdong, China) was introduced. Generally, the commercial each kilogram of Pan-Zyme had a cellulase activity of 21,081 International Units (IU). In comparison, the produced enzyme cellulase activity was 12,817 IU/kg. Both commercial and produced cellulase enzymes were supplemented with the buffaloes’ diet at 42.16 IU/Kg of dry matter following the recommendation of the commercial source of enzyme.2.1.2. Diet, Treatments, and ManagementTwenty-four lactating Egyptian buffaloes (602 ± 15.2 kg body weight; 3.6 ± 0.55 parity, 65 ± 10 d in milk and 6 ± 1.1 kg/d of milk production; (mean ± SD)) were used in a complete randomized design with a 25-day adaptation period and a 5-day sampling period (total experimental period of 30 days). All buffaloes were initially housed in tie stalls and fed individually according to body weight to meet their requirements for lactation [17]. The buffaloes were then randomly allocated (8 buffaloes per group) to one of the following experimental groups: CON—control consisted of a total mixed ration as described in Table 1, CENZ—the total mixed ration supplemented by a commercial source of cellulase enzyme (42.16 IU/kg DM), FENZ—the total mixed ration supplemented with cellulase enzyme cocktail produced in-farm (42.16 IU/kg DM). Buffaloes were fed the total mixed ration (TMR) twice daily at 7 a.m. and 7 p.m. in equal portions 1 h before each milking (buffaloes were machine milked twice per day). The portions of TMR used for all treatments were prepared once every ten days to check the dry matter concentration and adjust the feed if necessary. The enzyme supplementation was applied to a small portion of the TMR offered individually for each animal twice per day before the morning and evening feeding to ensure that the animal received the specified amount of enzyme. Samples of fresh TMR were collected every ten days (during the TMR preparation) and throughout the sampling period and stored at −20 °C until analyzed chemically. The milk production was tracked daily and was only recorded during the sampling days. Milk samples were collected at each milking from all buffaloes throughout the sampling period. Morning milk samples were stored at 4 °C until the evening samples were collected. The samples were subsequently pooled according to morning and evening milk yield and prepared in two equal parts: one aliquot was immediately analyzed for fat, crude protein, and lactose by infrared analysis (Milkotester LM2, Belovo, Bulgaria), and the rest was stored at 20 °C for amino acid (AA) and fatty acid (FA) analysis. Milk total and non-fat solids content was calculated.2.2. Sample Analysis2.2.1. Nutrient Intake and DigestibilityDuring the last 5 days of the experiment (sampling period), feed intake was recorded daily by weighing each animal’s offered diets and refusals after morning and evening feeding. The apparent nutrient digestibilities were determined during the sampling days according to the method described by Ferret et al. [18], where acid-insoluble ash was used as an internal marker. Fecal grab samples were collected from the rectum of each buffalo twice daily at 6 a.m. and 6 p.m., dried at 55 °C in a forced air oven for 48 h and pooled daily for each buffalo, ground to pass a 1-mm screen using a feed mill (FZ102, Shanghai-Hong Ji instrument Co., Ltd., Shanghai, China) and stored for further chemical analysis.2.2.2. Chemical AnalysisThawed TMR samples were dried at 55 °C for 48 h, milled to pass through a 1-mm screen (FZ102, Shanghai-Hong Ji instrument Co., Ltd., Shanghai, China), and composited by treatment before chemical analysis. Samples were analyzed for analytical DM (method no. 934.01), ash (method no. 942.05), crude protein (CP; method no. 954.01), and ether extract [EE; method no. 920.39); [19]. Neutral detergent fiber (NDF; Van Soest et al. [20] and acid detergent fiber (ADF; AOAC [19]; method 973.18) analyses were conducted using an ANKOM200 Fiber Analyzer unit (ANKOM Technology Corporation, Macedon, NY, USA). For NDF assays, samples were pretreated with an α-amylase and sodium sulfite. Both NDF and ADF are expressed without residual ash, and the organic matter (OM) was calculated.Quantitative amino acid measurements were performed for the milk protein. According to Millipore Corporation, the amino acid composition of experimental samples was determined using the HPLC-Pico-Tag method following the preparation procedure [21]. Phenyl isothiocyanate (PITC, or Edman’s reagent) was used for pre-column derivatization, while reversed-phase gradient elution high-performance liquid chromatography (HPLC) separates the phenylthiocarbamide (PTC) derivatives which were detected by their UV absorbance. The sample corresponding to the protein was weighed into a 25 × 150 mm hydrolyzed tube and was placed in a 110 °C oven for 24 h. The tube contents were quantitatively transferred to a volumetric flask and completed to volume with HPLC grade water. About 1 mL of the solution was filtered through a 0.45-µm sample filter. Together with appropriate standards, aliquots of hydrolysate were placed in disposable glass sample tubes from Waters Associates, soda glass from Fisons), loaded into a reaction vial. Hydrochloric acid was removed from the samples by drying under vacuum in a PICO-Tag workstation, achieved by connecting the vial to the workstation manifold and opening the vacuum control valve. Samples were then redried from redrying reagent (Waters reagents) again using the workstation. Derivatization is initiated by adding freshly prepared reagent mixed using a vortex mixer and allowed to stand at room temperature for 20 min. In this dried state, derivatized samples may be stored at freezer temperatures for several weeks, if required, before analysis. The chromatographic analysis using HPLC was carried out using the following gradient of Pico-Tag solvent (Eluent A and B) at 38 °C, flow rate 1 mL/min, and 20 mL of sample was injected and loaded on amino acids C18 column (100 × 4.6 mm) stainless steel. Detection of the PTC derivatives is by ultraviolet absorption measurements using a fixed wavelength (254 nm) Waters detector. Before injecting the sample, the illustrated was calibrated by two injections of the lysine standards.Samples of fatty acid composition in milk and dried ground feed were analyzed following the method described in [22]. Briefly, 3 mL of 2 M NaOH was added to 500 mg and 100 mg of milk and feed samples, respectively, for hydrolysis of the samples in a closed system using 15-mL screw-cap Teflon-stoppered Pyrex tubes. The hydrolyzed samples were incubated in a block heater at 90 °C for 40 min. Then, samples were extracted and esterified using 0.5 M NaOH in methanol and converted to FA methyl esters (FAME) using boron trifluoride (1.3 M; Fluka-Sigma Aldrich, St. Louis, MO, USA). A gas GC-MS system (7890B, Agilent, Santa Clara, CA, USA) equipped with a 100 m fused silica capillary column (0.25 mm i.d.; coated with 0.25 μm Agilent HP; Chrompack CP7420; Agilent Technologies, Santa Clara, CA, USA) and mass spectrometer detector (5977A). Hydrogen at a flow rate of 1.3 mL/min was used as the carrier gas. The injector and detector temperatures were 200 and 250 °C, respectively. The oven temperature was programmed as follows: initially 120 °C for 7 min, then increased by 7 °C per min to 140 °C, where it was held for 10 min before being increased by 4 °C per min to 240 °C. A 1-μL sample was injected into the GC column. The peaks were identified by comparison with the retention times of appropriate FAME standards (37 FAME Mix, Sigma Aldrich, PA, USA).Moreover, the conjugated linoleic acid peaks were identified by comparison with the retention times of a reference standard (a mixture of cis- and trans- 9,11 and 10,12-octadecadienoic acid methyl esters; Sigma Aldrich, PA, USA) using Galaxie Workstation 10.1 (Varian, CA, USA). Fatty acid compositions were expressed as g/100 g total FA. Both amino acid and fatty acid analyses were conducted at the Central Service Unit, National Research Centre, Egypt.2.3. Statistical AnalysisAll collected data were averaged by buffalo before chemical analysis. All parameter data (intake, digestibility, milk production/composition, milk AA, and FA profile) were analyzed using a model that included the fixed effect of treatment and the random effect of the buffalo within the treatment using PROC MIXED procedure of SAS (SAS® OnDemand for Academics, 2021 SAS Institute Inc., Cary, NC, USA). The sums of squares for treatment effects were further separated into a single degree of freedom comparisons to test for the significance of preplanned contrasts as follows: (1) supporting the total mixed ration CON diet with cellulase enzymes ENZ through supplementation (CON vs. FENZ + CENZ) and (2) the source of enzyme used as a supplement FCENZ (FENZ vs. CENZ). Treatment effects were considered significant or tending towards significance at p ≤ 0.05 and 0.05 < p ≤ 0.10, respectively.3. ResultsThe total mixed ration (TMR) offered to the lactating buffaloes (Table 1) could be characterized by a high forage content. High fibrous constitution and moderate fatty acid profile. The diet components were chosen based on; the season availability (the experiment was conducted in the middle of spring during May and June 2021), the richness in fibrous content, and was set to cover the maintenance and production requirement following the NRC (2001) manual.3.1. Nutrient Intake and DigestibilityThe nutrient intake and nutrient digestibility of buffaloes in the control group (CON), commercial source of cellulase (CENZ) group, and the in-farm source of cellulase (FENZ) group are shown in Table 2. Supplementing buffaloes’ diets with cellulase enzymes (CON vs. ENZ) did not significantly affect nutrient intake; no differences were observed between the cellulase enzymes sources (CFENZ). On the other hand, significant changes were observed in nutrient digestibilities. The dry matter digestibility was higher when supplementing cellulase enzymes (CON vs. ENZ). It was also significantly higher with FENZ when compared with CENZ. The same observation was applicable in the case of the digestibility of crude protein (CP), ether extract (EE), neutral detergent fiber (NDF), and acid detergent fiber (ADF), which were significantly higher with cellulase enzymes addition compared to the control (CON vs. ENZ). Supplementing the diet with an in-farm source of cellulase (FENZ) had a significantly higher impact on CP, NDF, and ADF digestibilities. However, FENZ tended to increase the EE digestibility compared to CENZ (CFENZ). Generally, FENZ showed higher nutrient digestibility percentages compared to other groups.3.2. Milk Production and CompositionThe milk production performance of dairy buffaloes is fully stated in Table 3. Supplementing the diet with cellulase enzymes (CON vs. ENZ) led to a significant increase in the daily milk yield and the fat correct milk yield (FCM); both yields were significantly higher with FENZ compared to all groups. Buffaloes’ diets supplemented with any source of cellulase enzyme had significantly higher fat, total solids, and solids not fat (SNF) percentages compared to control (CON vs. ENZ), it also resulted in a significantly higher fat, crude protein, and lactose yields. Repeatedly, the FENZ group resulted in significantly higher fat percentage and yield, and it is noteworthy that the highest crude protein, lactose, and energy yields were found in the FENZ group.3.3. Milk Amino Acid ProfileThe milk amino acids (AA) composition in lactating buffaloes is shown in Table 4. Amino acids profiles in both the essential amino acids (EAA) group and non-essential amino acids (NEAA) group were not significantly affected by cellulase enzymes supplementations (CON vs. ENZ). There were no significant differences between the cellulase enzyme sources (CFENZ) on the amino acids profile. However, enzyme supplementation showed significantly greater total EAA compared to control (CON vs. ENZ) and tended to be higher with FENZ compared to CENZ.3.4. Milk Fatty Acid ProfileThe milk fatty acids (FA) composition in lactating buffaloes is shown in Table 5. The milk fat concentration of C16:1 cis-9 and C18:0 was significantly lower with enzyme supplementation compared to the control (CON vs. ENZ). As for the C18 unsaturated fatty acids (oleic, linoleic, and linolenic acids), their concentration was significantly higher with cellulase enzymes supplementation (CON vs. ENZ) as well as the profile of the conjugated linoleic acid (cis-9, trans-11 CLA and trans-10, and cis-12 CLA). The response of supplementing lactating buffaloes’ diets with cellulase enzymes (CON vs. ENZ) was significantly higher on the sum of unsaturated fatty acids (UFA), the sum of monounsaturated fatty acids (MUFA), and the sum of polyunsaturated fatty acids (PUFA) which increased compared to CON. As well, the sum of saturated fatty acids (SFA) showed to be significantly lower when adding cellulase enzymes to buffaloes diets (CON vs. ENZ). The milk fat C10:0 tended to increase by FENZ group, while C14:0 tended to decrease by FENZ group (CFENZ), significantly, FENZ showed a lower C16:1 cis-9, linolenic acid, and PUFA profiles, and a higher oleic and MUFA profiles compared to the commercial source of cellulase (CFENZ).4. Discussion4.1. Nutrient Intake and DigestibilityFollowing the obtained results, the nutrient intake was not affected by cellulase supplementations, which, according to several pieces of literature [4,23,24], cellulolytic enzymes could have a marginal effect on feed intake. Consequently, that could suggest that cellulase enzymes positively affect feed intake only under suboptimal feed digestion conditions. Additionally, the amount of enzyme supplementation in this study was lower than those reported in previous studies (Peters et al. [4]; Bhasker et al. [25]), which also could explain the lack of response in nutrient intake due to an insufficient supply of enzyme activity [9]. On the other hand, the significant increase in nutrient digestibility was consistent with results obtained by [26]. That could be explained by: the increase of NDF digestibility, potential variations in gut viscosity, modified ruminal fermentation, enhanced attachment, colonization of the plant cell wall by ruminal microorganisms, and complementary actions with ruminal enzyme are possible causes of higher nutrient digestibility [27]. The increase in nutrient digestibility could also be explained by a possible synergistic effect between exogenous enzymes and endogenous enzymes, which could act as a condition modulator that increases the number of fibrolytic and non-fibrolytic microorganisms in the rumen [28].4.2. Milk Production, Composition, AA and FA ProfileThe increased milk yield and fat correct milk (FCM) were similarly reported by other studies using exogenous fibrolytic enzymes [4,23]. The study by Ortiz-Rodea et al. [29] conducted a large experiment on the effect of fibrolytic enzymes on milk production and composition (29 experiments). They suggested that an increase in milk and FCM yields could be encountered because of improvement in nutrient utilization and digestibility. The increase in fat content and yield due to cellulase enzyme supplementation may be credited to the substantial quantity of fiber digested in the rumen, contributing to more acetate for fatty acid synthesis [24]. The lack of effect on milk crude protein was also reported by Ortiz-Rodea et al. [29] and Peters et al. [23], consistent with a lack of differences in nutrient intake by enzymes supplements.Based on the obtained results, only the total EAA was affected by the cellulase enzyme supplementation, which is in line with results reported by Morsy et al. [24], where high essential amino acid concentration was reached due to the presence of protease enzyme in the supplemented cocktail of enzymes. In our case, the increase of total EAA could only be due to a slight improvement of endogenous protease enzyme due to the better conditions set by the synergistic effect of cellulases.As for the fatty acids’ composition, Rojo et al. [28] stated an increase in monounsaturated fatty acids and a lower profile of saturated fatty acids. Most milk fatty acids originated from plasma or the de novo synthesis in the mammary gland from acetate produced from rumen fermentation, including acetyl CoA carboxylase enzymes and fatty acid synthetase [24]. Ruminants do not synthesize polyunsaturated acids; consequently, their intensity in milk depends on the amount absorbed from the intestines. These results in the present study may be due to the altered contents of acetic and propionic acid production in the rumen due to better fiber digestion. The direct result of shifted VFA proportions could elevate precursor availability for fatty acid synthesis [23,29]. In our study, the conjugated linoleic acid concentration was also increased by cellulase addition, significantly, without increasing the stearic acid concentration, suggesting a slower biohydrogenation process that preserved a higher proportion of CLA without increasing stearic acid or decreasing the sum of unsaturated fatty acids [22,24]. Suggestively, the enzyme produced in the farm (FENZ) based on solid stat fermentation media is harboring multiple cellulases and hemicellulases with different specific functions that are produced in the presence of lignocellulose-based materials, which may explain the significant impact on lactating buffaloes’ performance compared to the commercial enzyme despite the equal supplemented concentration. It is also worth mentioning that the cost of producing 1 kg of farm-based cellulase enzyme cocktail cost 50 Egyptian pounds, which is equal to 0.3 Egyptian pounds for each kg of milk/day, on the other hand, the cost of 1 kg of the commercial enzyme is 300 Egyptian pound, which is equal to 1.16 Egyptian pound for each kg of milk/day.5. ConclusionsThe cellulase enzyme produced from Penicillium chrysogenum had a better impact on nutrient digestibility when supplemented with lactating buffaloes’ diets. Additionally, compared to the commercial enzyme source and the control, the produced cellulase enzyme had a better impact on the milk productivity, fat yield, and fatty acid profile. That is why we recommend establishing a small pilot production unit for cellulase enzymes in Egyptian farms, which will provide entrepreneurs with small farms to benefit from agricultural and agro-industrial wastes located near the farm.

animals : an open access journal from mdpi

10.3390/ani12030261

PMC8833639

Ammonia is the dominant pollutant gas in poultry houses, and it is harmful to broilers, especially in the cold season. Exposure to ammonia leads to damage to the respiratory system of broilers, affects the health of broilers, and reduces production performance. To date, the relationship between lung flora and immune system and brain exposed to ammonia is unclear, and there have been numerous studies on the lung–brain axis in recent years. Therefore, the aim of this study is to explore the effects of ammonia on lung microflora, lung tissue mucosal morphology, inflammatory cytokines, and neurotransmitters. Moreover, exploring these mechanisms can help in the development of strategies that alleviate the negative effects of the performance of ammonia. Our study suggests that the damage caused by ammonia to broiler lungs may be mediated by the lung–brain axis.

Atmospheric ammonia is one of the main environmental stressors affecting the performance of broilers. Previous studies demonstrated that high levels of ammonia altered pulmonary microbiota and induced inflammation. Research into the lung–brain axis has been increasing in recent years. However, the molecular mechanisms in pulmonary microbiota altered by ambient ammonia exposure on broilers and the relationship between microflora, inflammation, and neurotransmitters are still unknown. In this study, a total of 264 Arbor Acres commercial meal broilers (21 days old) were divided into 4 treatment groups (0, 15, 25, and 35 ppm group) with 6 replicates of 11 chickens for 21 days. At 7 and 21 D during the trial period, the lung tissue microflora was evaluated by 16S rDNA sequencing, and the content of cytokines (IL-1β, IL-6, and IL-10) and norepinephrine (NE), 5-hydroxytryptamine (5-HT) in lung tissue were measured. Correlation analysis was established among lung tissue microflora diversity, inflammatory cytokines, and neurotransmitters. Results showed that the broilers were not influenced after exposure to 15 ppm ammonia, while underexposure of 25 and 35 ppm ammonia resulted in significant effects on pulmonary microflora, inflammatory cytokines, and neurotransmitters. After exposure to ammonia for 7 and 21 days, both increased the proportion of Proteobacteria phylum and the contents of IL-1β and decreased the content of 5-HT. After exposure to ammonia for 7 days, the increase in Proteobacteria in lung tissue was accompanied by a decrease in 5-HT and an increase in IL-1β. In conclusion, the microflora disturbance caused by the increase in Proteobacteria in lung tissue may be the main cause of the changes in inflammatory cytokines (IL-1β) and neurotransmitters (5-HT), and the damage caused by ammonia to broiler lungs may be mediated by the lung–brain axis.

1. IntroductionAmmonia is one of the poultry house by-products that are detrimental to broilers and is produced from uric acid and undigested proteins in manure by aerobic or anaerobic bacteria [1]. The ammonia concentration in the poultry houses should not exceed 25 ppm. However, ammonia concentration commonly exceeds 25 ppm during the cold season, and this phenomenon is prevalent all over the world [2,3]. Several studies reported that 25 ppm ammonia can reduce productive performance, amplify immune response [4], and damage the respiratory system [5], leading to low breast muscle and carcass composition in broilers [6]. When ammonia concentration reaches 50 ppm, it can even affect the eyes, causing cornea damage and blurred vision, leading to difficulties in food foraging, which affect the metabolism and performance of broilers [7].In recent years, numerous studies have found that brain diseases such as Alzheimer’s disease are always accompanied by lung infections [8,9]. This has led to the lung–brain axis, a relatively narrow field in recent years. Several studies have shown that pulmonary microflora can regulate pulmonary inflammation and affect the nervous system and immune microenvironment through various mechanisms [10,11,12]. There is evidence suggesting that bacteria may have an impact on the host physiology by regulating neurotransmitters such as dopamine, norepinephrine, serotonin, or gamma-aminobutyric acid (GABA) [13]. Microflora affects neural development and behavior by regulating the hippocampal serotonergic system in the early stages [14]. Although the detailed mechanism of the lung–brain connection remains unclear, it is known that the central nervous system (CNS) can be affected by pulmonary flora through immunity, HPA axis, nervous system, and metabolites. Accumulating evidence suggests that respiratory tract microflora can be disturbed by external environmental factors [15,16]. Pathogenic microorganisms are common in lung diseases, such as asthma and cystic fibrosis. They are related to the increase in inflammatory markers in the stable state of disease. Environmental stress can lead to the disorder of the respiratory tract microflora by increasing the number of pathogenic bacteria, promoting an excessive secretion of toxins that increases the permeability of mucosa, resulting in a large production of inflammatory cytokines such as IL-1β and IL-6 by the host immune system [17]. Therefore, we hypothesized that the damage caused by ammonia to broilers may not only be caused by the lungs but also that the lung–brain axis may be involved in the process of lung injury.Thus, in this study, we investigated the effects of different concentrations of ammonia on lung tissue mucosal morphology, microflora, inflammatory cytokines, and neurotransmitters of broilers by exposure to ammonia for 7 days and 21 days, and we also analyzed the relationship between lung tissue microflora diversity, inflammatory cytokines, and neurotransmitters. Our results aimed to provide reference and recommendations for the study of the lung–brain axis and the therapy of lung diseases such as chronic obstructive pulmonary disease (COPD) and asthma.2. Materials and Methods2.1. Experimental DesignA single factor random design was used in this experiment. A total of 280 1-day-old male broiler chickens (Huadu Co., Ltd., Hebei, China) were housed in cages and a temperature- and humidity-controlled room, with free access to feed and water, then reared to 21-days-old. At the age of 22 days, 264 broilers with similar weight were randomly divided into 4 treatment groups, each comprising 6 replicates single-tier cages with 11 birds (single-tier cage size is 0.80 m length × 0.80 m width × 0.40 m height). The 4 treatment groups were the control group (0 ppm), the 15 ppm, the 25 ppm, and the 35 ppm ammonia groups. From day 22 to day 42, the concentrated NH3 was delivered in the four exposure chambers (4.5 m length × 3.0 m width × 2.5 height). Birds in the control group were housed in a chamber without NH3 addition, whereas birds in ammonia groups were exposed to different ammonia concentrations. The concentrations of NH3 in 4 chambers were monitored with a LumaSense Photoacoustic Field Gas-Monitor INNOVA 1412 (Santa Clara, CA) during the entire experiment. Temperature (21 °C ± 1 °C), relative humidity (60 ± 7%), and airflow were controlled during the exposures to ensure a suitable environment. The manure was removed from the chambers every 3 days to reduce the volatilization of extra NH3.2.2. Experimental Diet and Feeding ManagementThe diet used for age 21 to age 42 was based on a corn-soybean meal diet without antibiotics. The corn-soybean meal basal diet (Table 1) was formulated to meet or exceed the National Research Council requirements for broilers for all nutrients. The broilers were reared with the same diet and free feeding and vaccination against Marek’s disease, Newcastle disease, and bronchitis.2.3. Sample CollectionOn days 7 and 21 of the experiment, three broilers of approximately average weight were selected from each replicate for tissue sample collection. After the broilers were sacrificed, the left lung tissues were isolated and washed with PBS. Three lung tissues from each replicate were mixed into one sample (six samples were harvested from each treatment group) and stored in a −80 °C refrigerator for further analysis.2.4. Morphological Observation of Lung Tissue Mucosa and Determination of Serum Immune Cytokines and NeurotransmittersPieces of lung tissues from birds of four groups were fixed in 4% paraformaldehyde. Fixed tissues were embedded in paraffin, then sectioned to 3 μm thickness and stained with Mayer’s hematoxylin and eosin. The contents of interleukin (IL-1β, IL-6, IL-10) and noradrenaline (NE), 5-hydroxytryptamine (5-HT) in lung tissue were detected by an enzyme-linked immunosorbent assay (ELISA) kit (Jiancheng, Co., Ltd., Nanjing, China).2.5. DNA Extraction and PCR Amplification 16s rDNAMicrobial DNA was extracted from lung tissue samples using the E.Z.N.A.® soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) and according to the manufacturer’s protocols. Then, DNA quality was assessed using a 1% agarose gel electrophoresis. The V3-V4 hypervariable regions of the bacteria 16S rDNA gene were amplified with primers 338F (5′-ACTCCTACGGGAGGCAGCAG-3′) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) by thermocycler PCR system (GeneAmp 9700, ABI, Foster City, CA, USA). The PCR reactions were conducted, and the resulted PCR products were extracted from a 2% agarose gel, further purified, and finally, quantified using the QuantiFluor™-ST (Promega, Madison, WI, USA) according to the manufacturer’s protocol.2.6. Illumina MiSeq SequencingPurified amplicons were pooled in equimolar and paired-end sequenced (2 × 300) on an Illumina MiSeq platform (Illumina, San Diego, CA, USA) and according to standard protocols provided by Majorbio Bio-Pharm Technology Co., Ltd. (Shanghai, China).2.7. Statistical AnalysisAll statistical analyses for factor measurements of the difference between groups were conducted using one-way ANOVA analysis available with the SAS 9.1 software. Differences among means were tested by Duncan’s multiple range test. The data of inflammatory cytokines and neurotransmitters are presented as mean ± SEM. The data on lung tissue microflora were analyzed by a cloud platform (Meiji Co., Ltd., Shanghai, China). The correlation of inflammatory cytokines and neurotransmitters, respectively, with the abundance of lung tissue microbial species was calculated using Spearman’s coefficient. The replicate cage served as the experimental unit, and p < 0.05 was considered statistically significant.3. Results3.1. Effect of Ammonia on Lung Tissue Mucosa of BroilersThe effect of ammonia on lung tissue mucosa is shown in Figure 1. After 7 days of ammonia exposure, compared with the control group, inflammatory cell infiltration occurred in the 15 ppm group, local tissue hemorrhage occurred in the 25 ppm group, and a large number of red blood cells and necrotic cell masses were observed in the bronchus in the 35 ppm group. After 21 days of ammonia exposure, compared with the control group, a large number of red blood cells were found in the bronchus in the 15 ppm group, local tissue hemorrhage occurred in both the 25 ppm group and the 35 ppm group, and focal infiltration of inflammatory cells around the smooth muscle bundle of the bronchus. In addition, connective tissue hyperplasia was also found between lung lobules in the 35 ppm group.3.2. Effect of Ammonia on Pulmonary Microflora of BroilersThe effect of ammonia on the pulmonary microflora of broilers is shown in Figure 2. At the phylum level, Firmicutes, unclassified Bacteria, Proteobacteria and Bacteroidetes are the dominant bacterial phyla. Only the proportion of Proteobacteria was significantly different between four groups after 7 and 21 days long exposure (p < 0.05). At the genus level, only unclassified_f_Ruminococcaceae was significantly different between the four groups after 7 days long exposure (p < 0.05), but no significant difference after 21 days long exposure.3.3. Effect of Ammonia on Inflammatory Cytokines in Lung Tissue of BroilersThe effect of different ammonia concentrations on inflammatory cytokines in lung tissue under exposure to ammonia for 7 days and 21 days is shown in Figure 3. The results show that the level of IL-1β in lung tissue significantly increased under exposure to ammonia for 7 days in the 25 and 35 ppm group compared with the control group (p < 0.05), but there was no significant difference in IL-6 and IL-10 contents between the four groups. After 21 days long exposure, the levels of IL-1β and IL-6 in the 35 ppm group were significantly higher than that in the control group (p < 0.05), but there was no significant difference compared with the 15 ppm and the 25 ppm group. Compared with the control group, the content of IL-10 significantly increased in the three ammonia groups after 21 days long exposure (p < 0.05).3.4. Effect of Ammonia on Neurotransmitters in Lung Tissue of BroilersThe effect of different ammonia concentrations on neurotransmitters in lung tissue after 7 and 21 days long exposure is shown in Figure 4. The results show that there was no significant difference in NE content between the four groups after 7 and 21 days long exposure. The content of 5-HT decreased with the increase in ammonia concentration. Compared with the control group, the content of 5-HT significantly decreased in the 35 ppm group after 7 and 21 days long exposure (p < 0.05).3.5. Correlation Analysis of Pulmonary Microflora with IL-1β, IL-6, IL-10, and NE, 5-HTThe correlations between the top 10 most abundant phylum, inflammatory cytokines, and neurotransmitters in lung tissue were analyzed (Figure 5). The results show that the proportion of Proteobacteria was negatively correlated with 5-HT (p = 0.031) and positively correlated with IL-1β (p = 0.002) after 7 days long exposure. However, after 21 days long exposure, there were no significant correlations between pulmonary microflora, inflammatory cytokines, and neurotransmitters.4. DiscussionIn poultry houses, atmospheric ammonia has been linked to damage to the respiratory tract mucosa and a reduction in resistance to respiratory diseases [18]. Numerous studies have found that ammonia concentration, once exceeding 25 ppm, may have adverse effects on poultry health and production [19,20]. In the current study, we observed that with an increase in ammonia concentration, lung tissue showed different degrees of damage. A previous study reported that broilers exposed to 20 ppm of ammonia over a period of 42 days indicated pulmonary edema, congestion, and hemorrhage [21]. Another study found that exposure to 25 ppm of ammonia resulted in a decrease in performance by impairing the immune response [5]. In this study, we found that the microflora of lung tissue would be disturbed in broiler chickens exposed to 35 ppm of ammonia, the level of inflammatory cytokines increased, and the level of neurotransmitters decreased. These results indicate that exposure to 35 ppm ammonia concentration would have a detrimental impact on broilers’ health.The respiratory tract of animals is connected to the external environment and, therefore, be widely affected by external factors. The airway microbial that colonizes the mucosa may be the first to be affected by external factors [15]. Germ-free mice exhibited a lack of normally developed immune system and mucosal alternations, both of which can be restored by colonizing microbiota [22,23]. Therefore, the lung injury in the ammonia group may be due to the disorder of microflora in lung tissue. In the current study, we found that the dominant flora of lung tissue includes Firmicutes, Proteobacteria, and Bacteroidetes. These results are consistent with previous reports that showed the existence of the phyla Proteobacteria, Firmicutes, Tenericutes, Actinobacteria, Bacteroidetes, and Chlamydia/Verrucomicrobia in the respiratory tract of domestic and wild birds [24]. A recent study found that air pollutants can cause airway microflora disorders in rats, leading to the colonization of pathogenic bacteria, making the body susceptible to infections [25]. In addition, we found that exposure to ammonia causes lung tissue microflora disorders, and the proportion of Proteobacteria was significantly increased under exposure to ammonia. Shin et al. considered that an increased abundance of Proteobacteria is a potential diagnostic signature of dysbiosis and diseases [26]. Specifically, ammonia, as an environmental pollutant, endangers the homeostasis of respiratory microorganisms in broilers, increasing pathogenic bacteria, which is detrimental to respiratory health.Cytokines are known to regulate immune and inflammatory responses. It has been suggested that foreign stress can influence the function of immune cells and cytokines [27]. In the current study, the level of inflammatory cytokines IL-1β, IL-6, and IL-10 in lung tissues were affected by ammonia. A previous study reported that exposure to 25 ppm of ammonia could affect the level of cytokines in serum and tracheal tissue of broilers [28]. Further, the level of IL-1β increased in broilers that were exposed to ammonia for 3 weeks, and these released cytokines led to inflammatory responses and multiple organ damage [29]. Our results show an increase in the levels of IL-1β, IL-6, and IL-10 in the 25 ppm and 35 ppm groups compared with the control group. This demonstrates the occurrence of lung inflammation in broilers following exposure to high concentrations of ammonia. Similarly, our last experiment also demonstrates that this exposure to high concentrations of ammonia causes inflammatory damage to broilers’ lung tissues [30].Neurotransmitters are mainly secreted and released by neurons and are the medium of communication between neurons. They are often used as biomarkers that indirectly reflect the activities and disease status of the central nervous system [31]. One study found that benzo(a)pyrene exposure decreases the content of 5-HT in the urine of 3 to 5-year-old children [32]. At present, there is no study on the effect of ammonia on the content of neurotransmitters in broilers’ respiratory tracts. In the current study, we found that 5-HT in lung tissue significantly decreases following exposure to ammonia. 5-HT can promote the contraction and relaxation of bronchioles and bronchi [33]. Therefore, the decrease in 5-HT content in lung tissue may be due to the decrease in bronchoconstriction and relaxation capacity caused by the continuous inhalation of ammonia. Although due to the presence of the blood-brain barrier, 5-HT in peripheral tissues cannot directly enter the CNS, and studies have shown that peripheral 5-HT can enter the brain through immune cells as a carrier [34]. Furthermore, it has been suggested that the change in 5-HT in the brain was also associated with panic disorder and depression [35]. In addition, previous literature has shown that the concentration of 5-HT in the periphery seemed to have an opposite pattern to that of 5-HT in CNS [34], indicating that the 5-HT concentration of CNS was affected by the periphery 5-HT level. Therefore, when lung inflammation occurs, the changes of 5-HT may be related to CNS through neural and immune pathways.The existence of respiratory microbiota on the respiratory mucosa probably acts as a gatekeeper that provides resistance to the colonization of respiratory pathogens. The respiratory microbes might be involved in the maintenance of the homeostasis of respiratory physiology and immunity [15]. Previous studies demonstrated that Haemophilus influenzae leads to airway inflammation deterioration, induced by cigarette smoke exposure in COPD (chronic obstructive pulmonary disease) mice [36], and the production of the pro-inflammatory mediators interleukin-6 (IL-6) and IL-1β [37]. This indicates that there is a close relationship between respiratory tract microflora and inflammatory cytokines. In this study, Spearman’s correlation analysis showed that the presence of Proteobacteria positively correlates with IL-1β and negatively correlates with 5-HT when the broilers were exposed to ammonia for 7 days. The increase in Proteobacteria has been proved to be positively correlated with intestinal inflammation [38]. Moreover, a comparison of the bacterial composition of patients with or without asthma demonstrates, in different studies, a higher abundance of Proteobacteria in asthmatic patients. It is generally known that an increase in Proteobacteria represents an imbalance of microflora. 5-HT may either be pro- or anti-inflammatory in monocytes. In the current study, lung function was impaired after ammonia exposure, and the decrease in 5-HT level meant that 5-HT might be an anti-inflammatory factor in lung tissue. Therefore, it is proven that the microflora disturbance caused by an increase in the Proteobacteria population under ammonia exposure may be the main cause of IL-1β content’s increase and 5-HT level’s decrease.There is growing evidence that the lung and brain are an interrelated system—if one of them is damaged, the other will also be affected and vice versa. The relationship between them is mediated through inflammatory, neurological, and endocrine signaling pathways [39]. Pulmonary microbiota can regulate pulmonary inflammation, which is correlated with cytokines (IL-6, IL-10, et al.). In addition, the microflora can also be connected to the brain through signal molecules such as neurotransmitters (5-HT, NE, etc.). In 2013, Clarke and colleagues reported that the early-life microbiome regulates the hippocampal serotonergic system in a sex-dependent manner [40]. However, some studies have shown that there is a link between serotonin in the lungs and CNS. In addition, serotonin from the lungs can enter CNS through the carrier. Therefore, 5-HT may transmit signals of lung inflammation to the brain. In general, lung microorganisms are disturbed by ammonia, which leads to lung inflammation and lung tissue and mucosal damage; these abnormal changes may be transmitted to the CNS by 5-HT as a signal molecule. This may be one of the lung–brain axis regulation pathways, and the mechanism needs to be explored by further investigation.5. ConclusionsIn conclusion, 15 ppm air ammonia barely affects broilers, while 25 ppm and 35 ppm disturbs the pulmonary microflora of group broilers, affecting the immune system and 5-HT level of 35 ppm group broilers. Overall, exposure to 35 ppm ammonia had the most severe effect on broilers. Moreover, under ammonia exposure for 7 days, the pulmonary microflora disturbance caused by an increase in Proteobacteria may lead to an increase in IL-1β and a decrease in 5-HT. Furthermore, this change may be mediated by one of the pathways in the lung–brain axis. These findings will provide a reference and new recommendations for the study of the lung–brain axis and the therapy of lung diseases, such as COPD and asthma. Although our understanding of the lung–brain axis is still very limited, research in recent years has yielded greater progress. Nevertheless, this paper suggests that the damage caused by ammonia to broiler lungs may be mediated by the lung–brain axis.

animals : an open access journal from mdpi

10.3390/ani12010116

PMC8749823

The development of cost-effective strategies that can be easily implemented on-farms is pivotal to promote a more judicious use of antimicrobials and its reduction in livestock industry. Indeed, inappropriate use of antimicrobials is linked to the phenomenon of antimicrobial resistance, a global health concern for both humans and animals. Studies on other food-producing species have confirmed the effectiveness of biosecurity measures on the reduction of antimicrobials, of while little is still known in beef production. Thus, this study aimed to investigate the effect of quarantine as a strategy to reduce medications in beef production. This measure resulted to be a viable strategy to reduce antimicrobials in beef cattle without compromising animal health and performance. The reduction was evident especially with regards to treatments administered for respiratory diseases, indeed the bovine respiratory disease is one of the most detrimental health issues affecting beef cattle. Penicillins was the most used class of antimicrobials, highlighting the need for an urgent decrease of such broad-spectrum medications, known for their contribution to the development of resistance. Although implementing new strategies on-farm can be costly for farmers, the reduction of antimicrobials on the long term and the support from EU authority may help to overcome some initial disadvantage.

Judicious antimicrobial stewardship in livestock industry is needed to reduce the use of antimicrobials (AMU) and the associated risk of antimicrobial resistance. Biosecurity measures are acknowledged for their role against the spread of diseases and the importance in reducing AMU in different species. However, their effectiveness in beef production has been scarcely considered. The aim of this study was to investigate the effect of the quarantine strategy on AMU in beef cattle. A total of 1206 Charolaise animals in five farms were included in the trial. Roughly half of the animals followed the standard procedure of the fattening cycle (no-quarantine; NO-QUA group) and half followed a 30-day period of quarantine (QUA group) since their arrival. Performance and antimicrobial data were recorded and a treatment incidence 100 (TI100it) per animal was calculated. Penicillins was the most used class of antimicrobials. Differences between groups were significant for males only, with NO-QUA group having greater TI100it (3.76 vs. 3.24; p < 0.05) and lower body weight at slaughter (713.4 vs. 723.7 kg; p < 0.05) than QUA group. Results suggest that quarantine strategy can reduce AMU in males without compromising their performance, whereas further investigation is needed for females.

1. IntroductionThe importance of antimicrobials in animal production is well-known. Since their discovery in the late 1940s, they have been essential to tackle infectious diseases, especially in intensive farming systems where pathogens are more likely to arise [1,2]. However, the misuse of antimicrobials and the associated risk of antimicrobial resistance (AMR) are global issues that jeopardize both human and animal health alike [3,4]. Animal production is a major contributor to the increase of AMR [3,4,5] thus highlighting the need of addressing antimicrobial use (AMU) in animal food-producing sectors. For example, some studies showed that an overuse of antimicrobials can select for the carriage of resistant bacteria in beef production [6,7,8]. Current research worldwide is mainly focused on the development of cost-effective strategies to be easily implemented in livestock farms and to contribute to the reduction of AMU. Alternatively, management strategies reported in the literature for the livestock species include supplementation of diet with additives (e.g., essential oils and clay minerals), organic acids or probiotics [9,10,11], targeted vaccinations, and changes in husbandry practices [10,12]. These strategies will be pivotal to promote a more judicious AMU. To achieve this goal the collection of accurate data on AMU at farm level is crucial [13]. One of the most recognized indicators to estimate AMU is the treatment incidence 100 [14,15] which is calculated through the defined daily dose animal (DDDA) established at both EU and Italian level.Biosecurity is also important to reduce AMU [12,16] because it accounts for measures which may contribute to prevent the introduction and spread of diseases in the herd [17]. Effective biosecurity measures to monitor and reduce the risk of diseases, such as the bovine respiratory disease (BRD), have been identified in beef production [18,19], yet their implementation is still scarce due to required investments and lack of information on their efficacy [20]. Some examples of biosecurity measures are the separation of younger calves from older cattle, the reduction of the stocking rate, the application of a quarantine period to the imported animals, the testing and culling of clinical suspects only, and the schedule of vaccination programs [10,21]. Unlike for beef production, studies on other livestock species have already confirmed the positive link between biosecurity and animal health, welfare, and productivity as well as the effectiveness of these measures on AMU [22]. For instance, Postma et al. [12] reported the efficacy of improving internal and external biosecurity measures to reduce AMU in pigs without compromising animal performances.The effectiveness of biosecurity in the reduction of AMU would encourage an antimicrobial stewardship in beef production. So far, management indicators recognized as contributing factors significantly associated with AMU in beef cattle have been investigated mainly through observational studies [23,24]. For instance, a study on Swiss veal farms observed a significant association between lack of quarantine upon arrival at the fattening farm and increase of treatment incidence [24]. In addition, the duration of the fattening period, the quarantine, and the space of feeding areas were significantly associated with AMU in veal calves [23]. A recent study carried out on Swiss veal farms evaluated a novel outdoor concept for calf fattening, which resulted in a drastic reduction of AMU without compromising animal health [25]. To the best of our knowledge, little is known about cost-effective strategies that may promote a prudent antimicrobial stewardship in beef production. Therefore, the aim of the present study was to investigate the effect of quarantine of imported beef cattle as a strategy to reduce AMU on-farm.2. Materials and MethodsThis study was approved by the Ethical Committee for the Care and Use of Experimental Animals of the University of Padova, Italy (approval no. 74/2018) and was conducted in accordance with Italian law (Decreto legislativo no. 26/2014) and EU Directive 2010/63/EU on the protection of animals used for scientific purposes.2.1. Specialized Fattening System in Italian Beef CattleApproximately 70% of beef cattle produced in Italy are reared in specialized fattening farms in the north-east of the country and around 90% of young animals farmed in this area are imported from France [26]. In France, animals are reared at pasture until 10–14 months of age, and then transferred to specific collection centers and mixed with animals of other farms located in different French departments to create homogeneous batches according to body weight (BW), breed, and sex. Within 2–4 days, animals are purchased by Italian beef fatteners and transported to Italy. The intensive conditions under which animals are fattened in Italy allow them to reach the slaughter weight after 6–7 months from arrival. The diet supplied to the animals consists of a total mixed ration with high proportion of concentrates, different proportions of feedstuffs according to breed, sex and fattening stage, and mineral and vitamin supplementations. The typical housing system of Italian beef farms, which generally has a turnover of about two fattening cycles per year, consists of closed or open barns with multiple pens. Each pen has fully slatted or concrete floors with straw bedding [27].2.2. Experimental Design and Treatment GroupsA total of 1206 Charolaise cattle (576 males and 630 females) entered five commercial specialized beef fattening farms associated to a cooperative of beef producers (AZoVe) located in Veneto region (Italy) between July 2018 and August 2019. Three farms reared only females and two only males, and all of them purchased animals from France in five different periods between July 2018 and August 2019 with an average of 240 animals per period. For each period, only one truck with animals of the same sex and homogeneous BW arrived to each of the five farms. At their arrival to the farm, animals were weighed and divided in two experimental groups, namely quarantine (QUA) and no-quarantine (NO-QUA), which were allocated in two different buildings of the farm. Two pens per experimental group were available for the trial and all pens were balanced for kg of initial BW per m2. On average, each pen contained 12 animals and the surface available per animal averaged 5.3 m2. Animals allocated to the NO-QUA group followed the standard fattening cycle and thus since the beginning of the trial joined animals that were already present in the farm, i.e., those not included in the trial but fattened in the same building of NO-QUA animals. Instead, animals allocated to the QUA group followed a 30-day period of quarantine since their arrival to the farm before moving to the same building of NO-QUA group. Prior to the allocation of the animals, the building designated for QUA animals was cleaned and sanitized. Both QUA and NO-QUA groups started the fattening cycle on the same day. The animals went through the same health protocol. Specifically, they did not receive any vaccines nor antimicrobial treatments in France. Whereas, at their arrival to the Italian fattening farms, the same vaccination program (i.e., polyvalent vaccine for BRD) and parasitic control program were administered to the animals.In total, 578 animals (48%) from the five periods were assigned to the NO-QUA group (264 males, 314 females) and 628 animals (52%) to the QUA group (316 males, 312 females). The diet provided to the animals was the same for QUA and NO-QUA groups but differed according to sex (Table 1). Since it was necessary to comply with the needs of the farmers and the routine management procedures of their farms, a non-randomized controlled intervention study was considered the most suitable approach. Animals of a truck were a mix of different French farms and this, together with the creation of pens with similar kg of BW per m2, contributed to minimize the potential bias related to the allocation of the animals to the two experimental groups. In fact, for male-rearing farms, animals allocated to the QUA group were purchased from 12 departments and 138 farms of origin while animals of NO-QUA group were purchased from 10 French departments and 108 farms of origin. For female-rearing farms, animals were originally purchased from 18 French departments and 156 farms of origin for QUA group and 21 departments and 162 farms of origin for NO-QUA group.Information on date of birth, date of start and end of the fattening cycle (the mean duration of the fattening cycle was 193 days), BW at arrival to the fattening farm (BW0), BW at 30 days after arrival to the fattening farm (BW30), and BW at the end of the fattening cycle (BWfinal) were collected for each animal. In particular, animals were individually weighed over three time points: (1) once arrived at the farm prior to their allocation to the experimental groups, (2) after 30 days since their arrival to the farm, and (3) prior to transport to the slaughterhouse. Data on BW were used to calculate the average daily gain during the first 30 days from arrival to the fattening farm (ADG30, kg/d) and at the end of the fattening cycle (ADGtot, kg/d). Reasons of culling during the fattening cycle (e.g., injury, death) were also recorded. The afore-mentioned variables were also used to obtain the season of arrival at the fattening farm, the length of the fattening cycle (days), the number of deaths and the mortality rate (%).2.3. Quantification of Antimicrobial UseData on the number of parenteral treatments administered to the animals, the date of treatment, the reason of administration and the amount (mL) of antimicrobial used per each parenteral treatment were recorded throughout the fattening cycle. Antimicrobials were administered by veterinarians employed by the AZoVe cooperative of beef producers. They are all equally trained and stay blind on whether the farms are or are not involved in research trials. Thirteen veterinary medicinal products (VMP) containing antimicrobials were used in the studied farms. A defined daily dose animal for Italy (DDDAit) was assigned to each active ingredient (AI) with antimicrobial activity of those VMP. A DDDAit represents the dose (mg) of the AI administered per kg of BW per day and it was established during the development of the ClassyFarm integrated monitoring system (www.classyfarm.it) of the Italian Ministry of Health. In order to quantify the frequency of treatment, which allows for a better monitoring of AMU [14,15], an index called treatment incidence 100 for Italy (TI100it) was calculated per each VMP [15] at animal level using the following formula, modified from Timmerman et al. (2006) Equation (1):TI100it = amount of AI administered per animal (mg)/[DDDAit (mg/kg/day) × standard body weight (kg) × standard days at risk] × 100(1) where ‘standard body weight’ is the average expected BW of the animal at treatment (400 kg) and ‘days at risk’ the standard number of days of the fattening cycle (230 days). The TI100it of all VMP were summed up to obtain a total TI100it per animal and considering all antimicrobial administrations carried out during the whole fattening cycle for both groups. If a VMP had two AI, both were considered in the calculation of the DDDAit as two different treatments.2.4. Statistical AnalysisData were analyzed using the software SAS 9.4 (SAS Institute Inc., Cary, NC, USA). Animal was the experimental unit. Data were tested for normality. Males and females were analyzed separately because none of the farms had both sexes. Descriptive statistics of BW0, BW30, BWfinal, length of the fattening cycle, ADG30, ADGtot, and TI100it per sex were calculated. In addition, mortality rate, number of animals treated per sex, number and percentage of parenteral treatments per class of antimicrobials, percentage of parenteral treatments according to the reason of administration and number and percentage of parenteral treatments per experimental group per sex (QUA and NO-QUA for males, QUA and NO-QUA for females) were calculated. A Chi-Square test was performed to check for differences regarding the reasons of treatment between groups (QUA and NO-QUA).To investigate the effect of the strategy of quarantine on animal performance, two ANOVA tests (one for male-rearing farms and one for female-rearing farms) were performed using the GLM procedure of SAS. BW0 was transformed into a categorical variable through the creation of three classes of BW (low, medium, high) within sex according to mean ± 0.5 SD. The following linear model was used Equation (2):yijkl = µ + farmi + quarantinej + seasonk + iBWl + (farm × quarantine)ij + (quarantine × season)ik + eijkl(2) where µ is the overall intercept of the model; yijkl is the dependent variable (BW30, BWfinal, ADGtot or ADG30); farmi is the fixed effect of the ith farm (two farms for males and three for females); quarantinej is the fixed effect of the jth experimental group (NO-QUA and QUA); seasonk is the fixed effect of the kth season of arrival of the animal to the fattening farm (spring: March, April, May; summer: June, July, August; autumn: September, October, November; winter: December, January, February); iBWl is the fixed effect of the lth class of BW of the animal at arrival to the fattening farm (low, medium, high); (farm × quarantine)ij is the fixed interaction effect between farm and experimental group; (quarantine × season)jk is the fixed interaction effect between experimental group and season of arrival of the animal to the fattening farm; and eijkl is the random residual. Data are presented as least squares means ± standard error. Multiple comparisons among least squares means of the fixed effects were performed through Bonferroni post-hoc test. The criterion for statistical significance was established at p < 0.05 and for statistical trend at 0.05 < p < 0.10.A third model was built to investigate the effect of the strategy of quarantine on AMU. Only males were included in the statistical analysis as the number of females treated with antimicrobials was very low. Data were not normally distributed, thus the TI100it was analyzed using a generalized linear mixed model with gamma distribution and log link function in GLIMMIX procedure of SAS. Before performing the analysis, a constant of 3 was added to TI100it in order to avoid an over-estimation of the index during the statistical analysis due to the high number of zeros (animals not treated) and the impossibility of modelling the log. The model included farm, quarantine, and season of arrival as categorical fixed effects, BW0 as linear covariate and intercepts of animal ID nested within period of purchasing as random effect. Goodness of fit of the model was evaluated by checking Akaike’s Information Criterion and Bayesian Information Criterion of each step of model building. Results are presented as least squares means ± standard error. Tukey–Kramer post hoc adjustment was used for multiple comparisons of least squares means of the fixed effects. The criterion for statistical significance was set at p < 0.05 and for statistical trend at 0.05 < p < 0.10.3. Results3.1. Descriptive Statistics of Performance TraitsThere was a large variability on performance traits between farms rearing different sexes. In particular, male-rearing farms had heavier animals than female-rearing farms, either at the start (403.4 vs. 320.6 kg, respectively) and end of the fattening cycle (719.6 vs. 559.6 kg; Table 2). The mortality of QUA and NO-QUA animals was 0.96% and 1.04%, respectively.3.2. Effects of Farm, Quarantine, and Season on Performance TraitsThe two male-rearing farms differed significantly for all performance traits, except for ADG30. Quarantine group males had higher BW30, BWfinal, ADG30 and ADGtot than NO-QUA males (p < 0.05; Table 3). There was also an effect of the season of arrival to the fattening farm on all performance traits. In particular, animals that entered the fattening farm in spring had significantly higher BW30, BWfinal, ADG30, and ADGtot compared with animals that entered the farm in other seasons (p < 0.05; Table 3).Across the three female-rearing farms, there were significant differences for BW30, BWfinal and ADG30 (p < 0.05; Table 4). No differences were reported between QUA and NO-QUA groups (p > 0.05), whereas season of arrival affected performance traits. Specifically, females that entered the farm in winter had significantly higher BWfinal and ADGtot than females that entered the farm in other seasons (p < 0.05; Table 4). Least squares means for the interaction between farm x quarantine and quarantine x season were not significant, thus no results were provided.3.3. Descriptive Statistics of Antimicrobial UseOne-hundred-and-fifty-six animals out of 1206 received at least one treatment during the fattening cycle. Specifically, 126 out of 156 were males and 30 females. A total of 675 parenteral treatments were administered during the trial: 57.0% were administered to NO-QUA animals and 43.0% to QUA animals. Three of the 13 VMP used in the farms were composed of two AI belonging to different classes of antimicrobials. Thus, we also calculated the number and percentage of parenteral treatments administered to the animals identifying each treatment with the number of AI of each VMP. If each AI of a VMP is counted as a separate treatment, the number of parenteral treatments increased to 763 (Table 5). In general, penicillins were the most frequently used antimicrobials (29.2%) followed by amphenicols (19.7%), fluoroquinolones (15.7%), and aminopenicillins (13.9%; Table 5). The TI100it averaged 0.76 ± 2.65 for males (range: 0 to 36.02) and 0.13 ± 0.69 for females (range: 0 to 9.17).Overall, the main reasons of administration of antimicrobials were locomotor disorders (58%) and respiratory diseases (37%). The remaining 5% was represented by other reasons such as gastrointestinal diseases, abscess, ear infection and horn fracture. Data on antimicrobial use according to the class and reason of administration are presented in Table 6.NO-QUA and QUA groups did not differ significantly in terms of frequency of animals treated for locomotor disorders (e.g., lameness, interdigital dermatitis and interdigital phlegmon) and other diseases, whereas they differed significantly for respiratory diseases (Table 7).3.4. Effects of Farm, Quarantine, and Season on Antimicrobial Use in MalesAntimicrobial use differed significantly between farms and groups. Indeed, NO-QUA group had higher TI100it than QUA group (3.76 vs. 3.46, respectively). In addition, males that arrived at the fattening farm during the coldest months of the year had higher TI100it compared with animals that arrived in spring and summer (Table 8).4. Discussion4.1. Effects of Farm, Quarantine and Season on Performance TraitsFarm was an important source of variation of performance traits during the fattening cycle in both male-rearing farms and female-rearing farms; in fact, different management strategies applied on-farm are crucial in determining growth performance [28]. Different feeding strategies supplied to the animals also play an essential role in explaining the variability of performance observed in our study. According to the general feeding management applied on specialized Italian beef fattening farms, females receive a diet with lower content of concentrates compared to males (Table 1) [26] to avoid an excessive fat deposition of female carcasses. This helps to explain why farmers are more likely to rear only one sex on their farms, i.e., to easily manage the different diets.The strategy of quarantine showed a positive effect on performance of male-rearing farms. In the present study, beef cattle underwent a long travel from France to Italy (Veneto region) and a process of mixing both before travelling and at arrival to the fattening farms. Currin and Whittie [29] reported that transportation, especially during cross-country travels, is a stressful event for the animals, and Benavides et al. [30] reported higher likelihood of cross-contamination among animals for bovine viral diarrhea virus when farms shared transport vehicles or animals were transferred in contaminated vehicles. The aforesaid stressful events combined to the process of adaptation to a new farm environment can increase animals’ susceptibility to diseases (e.g., BRD and bovine viral diarrhea virus) which in turn may affect their performance [25,31,32]. Indeed, the BRD is one of the most detrimental health issues affecting beef cattle and it is usually associated with a general depression-like status and a decrease of appetite of the animals [33]. Thus, it is likely that the implementation of a 30-day period of quarantine before entering the standard fattening cycle, was an effective measure in reducing diseases in Charolaise males as suggested by an improved animal performance and the lower number of parenteral treatments administered for respiratory diseases. However, a positive effect of quarantine on performance traits was not observed in Charolaise female-rearing farms, likely due to their lower susceptibility to diseases, mainly BRD [34], as supported by the low number of females treated (30 out of 630 females) reported in this study. These findings are also in line with other studies reporting that male beef calves were at higher risk of diseases and had an increased likelihood to die due to BRD than females [35,36]. Another factor that may help to explain the higher susceptibility of males to respiratory diseases, can be the difference in the level of concentrates supplied to their diet [36,37,38]. For instance, Galyean et al. [37] reported a trend for increased BRD morbidity with increasing levels of concentrate, in particular when such level was above 50% of the total diet. The difference in diet composition between male-rearing farms and female-rearing farms is in line with data presented in literature. Indeed, the level of concentrates administered to males was higher than the amount provided to females and with a percentage above 60%. Finally, it is also likely that differences in animal management with regards to care/handling was applied on-farm [39], for instance due to the fact that females are more docile and easier to handle than males. Poor animal handling can lead to high levels of stress which in turn may impair the animals’ immune systems and increase their susceptibility to diseases [40].The season of arrival to the fattening farm is as an important source of variation of animal’s performances [28]. In our study, males had better performances during spring, likely due to warmer temperatures and lower humidity typical of this period of the year in Veneto region, which may contribute to reduce the risk of BRD. However, both heat stress and exposure to diseases may lead to a reduction of animal performances, as observed by Sturaro et al. [28], who reported a reduced ADG in beef cattle with high temperatures experienced during summer.4.2. Effects of Farm, Quarantine, and Season on Antimicrobial UseItalian beef farming is mainly characterized by young animals imported from France and reared under intensive fattening conditions in the north-east of Italy. Although this system is recognized as a positive integration between the two countries in terms of exploitation of resources available [41], the long transport distance, the lack of a preventive vaccination program and stress exposure might increase the need of AMU. Specifically, the long transport distance that animals have to undergo is a common practice in beef industry [42], thus making the findings of our study applicable to other international beef fattening realities. In this study, 156 animals out of 1206 were treated at least once with a VMP. However, we also observed that females were less treated than males (30 vs. 126, respectively), in accordance with our previous studies carried out on a larger sample of beef fattening farms [31,43]. A different level of concentrates between male-rearing and female-rearing farms can also contribute to explain such a difference in the number of animals treated. In fact, according to the study of Fluharty and Loerchthe [38], higher percentage of concentrates led to a greater number of treatments required for sick beef calves. Penicillins was the most used class of antimicrobials, thus highlighting a wide exploitation of broad-spectrum antimicrobials which are known for their contribution to the development of AMR [44,45].The main reason of AMU in the current study was for locomotor disorders followed by respiratory diseases. The Charolaise beef breed is known for its high BW compared to other breeds [26]. This could be one of the reasons that makes these animals more prone to develop lameness [46], a welfare issue that can increase under intensive conditions when there is a lack of an appropriate flooring system and the animals reach a final BW greater than 700 kg [47].The percentage of parenteral treatments and the TI100it were lower in QUA than NO-QUA group suggesting that providing an initial 30-day period of quarantine to the animals arriving to the fattening farm led to a reduction of AMU. Indeed, biosecurity measures, such as the practice of quarantine, are essential to reduce the spread of diseases [16,19,25] and consequently may help to decrease the need of AMU as observed in the current study.Significant differences for AMU among male-rearing farms may be explained by differences in feeding and management practices. Sharma et al. [11] reported the importance of a suitable diet in the prevention of diseases. Indeed, targeted feeding strategies are important to maintain appropriate animal health and welfare conditions and specifically they seem to be associated with the presence of locomotor disorders like lameness [46]. For instance, Compiani et al. [46] reported that a good optimization of the feed ration at the arrival to the farm combined to a gradual transition towards the ration of energy concentrates, helps to manage acidotic events and the associated risk of developing lameness in beef cattle. Therefore, although Charolaise beef breed requires a high level of starch and energy in the diet, a gradual transition to the new diet is essential to avoid locomotor impairments. This may help to clarify why Charolaise males-rearing farms were more likely to have animals more treated with antimicrobials for locomotor disorders.A season effect on AMU was also reported in Charolaise male-rearing farms showing that animals arriving to the fattening farm in winter and autumn had higher TI100it compared to animals arriving in summer and spring. Similar results were reported in our previous study [31,47], where low temperatures and different humidity likely explained the greater likelihood of BRD observed in winter and autumn. Although Becker et al. [25] reported that there was no clear linkage between winter and AMU in young calves, we observed an increase of TI100it during the coldest months of the year.According to the literature, factors such as transportation distance, farmer–veterinarian relationship, and variables associated to the pen such as the m2 can be considered as possible sources of variation of animal health, performance and AMU [48,49,50] whereby justifying their investigation through the statistical model. However, transportation data were not available, and the five farms involved in the study were managed by the same veterinarians equally trained and employed by the cooperative of beef producers (AZoVe). Thus, in this case we did not consider necessary to investigate the latest effect in the model. Instead, we tested the area (m2) of the pens, but we decided to remove this effect from the final model because not significant whereby confirming that differences between pens within and between farms of same sex were not significant. Future intervention studies to further investigate the effect of new strategies on AMU in beef farms are needed and inclusion of other farm characteristics may contribute to provide target and cost-effective information for a more holistic view.5. ConclusionsOverall, this study showed that the implementation of the practice of quarantine was a feasible strategy to reduce AMU in beef production without compromising animal health and performances. Specifically, this strategy was effective in male-rearing farms and in the reduction of AMU administered for respiratory diseases. Building quarantine facilities ex-novo is a cost for beef farmers, but the reduction of AMU on the long term may help to compensate this initial economic investment. Moreover, stricter application of current EU policies that promote high standards of animal welfare may drive farmers towards the implementation of a more welfare-friendly farm helping them not only to apply a more judicious AMU in beef cattle but also to cover such initial costs. Findings of this study can be representative of similar beef fattening farm realities worldwide, specifically those characterized by animals reared at pasture in the first part of their life, followed by a long transportation distance to reach the fattening farm and by intensive fattening conditions. On-farm, a rich energy diet is supplied to the animals to reach the required final BW in relatively short time. Another aspect that can make our results easily exploitable is that the Charolaise breed used for the study is a cosmopolitan breed, thus making our findings applicable to other realities. Nevertheless, further research should investigate AMU in other breeds or multi-breed farms. In addition, it would be interesting to investigate the effect of quarantine on AMU in other types of beef fattening systems.

animals : an open access journal from mdpi

10.3390/ani13061087

PMC10044351

The increase in farms’ size over time while not increasing livestock density, and the extension of the area where animals are kept, has meant that the proximity between human and animals has decreased. More than ever, good management is essential, as it is associated with better performances and overall welfare improvement. The personality traits most commonly studied in farm animals are exploration and sociability, and are usually studied by observing animals’ behavior. The present work aimed to evaluate the behavior of Mertolenga breed young bulls when exiting the chute, the reaction to novelty and human approach tests, as well as to understand if the responses between these tests are related. Data from twenty-nine Mertolenga-bred young bulls was collected and analyzed, and it was observed that older animals showed a tendency to leave the chute more calmly, took longer to touch the novelty ball and touched the ball less often, with a consequent reduction in the likelihood of playing with it. As for the human approach test, animals that came out of the chute more calmly allowed the human to come closer to them. These behavior tests should be further studied on potential sires in order to increase docility and handling in autochthonous beef breeds.

Livestock behavior and welfare are increasingly recognized to be related, not only to the animals’ handling, but also with productivity. The present work was carried out at the Mertolenga Breed Testing Center and its objective was to evaluate the behavior of Mertolenga breed young bulls when exiting the chute, the reaction to novelty and human approach tests, as well as to understand if the responses between these tests are related. Twenty-nine Mertolenga-bred young bulls from 16 different farms, aged between 8 and 13 months, entered the study farm from the end of May to the beginning of June 2021. Data was collected on six different days and analyzed with the SAS® 9.4 software. Older animals showed a tendency to leave the chute more calmly, take longer to touch the novelty ball and to touch the ball less often, with a consequent reduction in the likelihood of playing with it. In the human approach test, animals that came out of the chute more calmly allowed the human to come closer. These behavior tests should be further studied on potential sires, so as to increase docility and manageability of autochthonous beef breeds.

1. IntroductionAlthough the number of beef cattle farms has been decreasing, the stock numbers have been increasing. The increase in population size and the extension of the area where they are kept has reduced the proximity between humans and animals, with the possibility that the brief interactions are limited to unpleasant procedures for the animal, making them afraid of humans and the handling of the animals more dangerous for the operator.The intrinsic characteristics of animals such as genetics, past experiences, and the context they are in, as well as the attitude of humans themselves, will influence their relationship with the production environment [1]. The importance of human-animal relationship has been intensely studied, and its influence on the welfare, health, and performance of cattle is clear. While bad animal handling can lead to production declines due to increased fear and consequent release of stress hormones, good handling can improve performance and animals’ overall welfare [2]. Animals’ reactivity results not only from a set of experiences throughout their lives, but also from genetic factors [3]. Table 1 shows heritability estimates for some behavioral tests in different cattle breeds.There are several methods for temperament assessment that can be used in cattle: from measurements of physiological parameters, such as heart rate or blood cortisol level, to behavioral tests, such as scoring behavior and speed when exiting a chute, the flight distance, and the novelty test, among others.The most studied personality traits in production animals are boldness (defined as the propensity to take risks when exposed to novelty), exploration (behavior by which animals gain knowledge about their environment), and sociability (the propensity to seek contact with others or stay close to conspecifics), and are usually studied by observing animals’ behavior in novelty or isolation tests [8]. Ensuring calmness in animals is also a good strategy to control feeding behaviors, as well as overall activity. Thus, with longer and more frequent meals, feed efficiency and methane emissions are improved [9]. Furthermore, selecting and breeding less temperamental and fearful animals ensures their welfare and makes the production sector more attractive to consumers.The Mertolenga breed is the most numerous out of the 15 autochthonous cattle breeds, currently with about 27,000 breeding females in Portugal. It presents an excellent adaptability and three different phenotypes: red, roan, and red spotted [10]. It is widely known for its nervous temperament that sometimes makes it challenging to work with.Cattle, as preys, are very prone to fear, which causes their responses to stimuli to be unpredictable. Thus, fear is a very important characteristic that often affects welfare in a negative way. According to [11], fear is an emotion induced by a perceived threat that causes animals to quickly move away from its location or to hide. It should be distinguished from emotional anxiety, which typically occurs without a particular or immediate external threat. Some examples of signs of fear are running away, seeking to hide, an exaggerated reaction upon exposure to a sudden novelty, vocalizations, conspecific seeking, and also increased aggression, including attacking the source of fear. The reactivity of cattle to handling and to human contact are consistent over time and may have some genetic influence, thus, temperament traits are already used in some selection programs. However, the diversity of behavioral tests makes it difficult to demonstrate a relationship between them [12].As cattle are gregarious animals, fear expression can also be influenced by social stimuli, and social isolation is probably one of the most important stressors when animals perform fear tests individually. Thus, social isolation may act as a multiplying factor in the ability to feel fear, especially neophobia. Behavior patterns revealing fear are very variable according to the characteristics of the threat, and can be expressive movements, such as head posture or facial expressions. There are also some specific alarm signals, such as odors or pheromones, which allow herbivores to communicate and warn conspecifics [13].Fear tests should be performed when the animals are still young, in order to minimize the influence coming from previous experiences. The results of these tests can not only give us information about the temperament of the animal, but may also give indications about the future productivity of the tested animal [14]. Both the novelty test and the human approach test are often used in animal fear studies.The novelty test, such as the evaluation of the animal’s reaction to a new object, allows us to identify the animal’s confidence, its tendency to explore (as an indicator of its daringness), whether it is ready to take risks, and its adaptability to new environmental conditions, which may affect its well-being and productivity. Animals’ response to a novel object also reflects its degree of fear [15]. This test is usually performed outside the animal’s comfort environment and can be repeated several times in order to study whether the behavior toward the object is maintained or changes over time, reflecting an adaptive capacity.The response to a new object depends on the experimental design, its validity, and its reliability. The repeatability of the test is often questioned, since habituation is to be expected [13]. In case of repeating the test, it should preferably be performed in the same place, with the animal alone. The insertion of a new object in the space will cause abruptness, unfamiliarity, and unpredictability factors to produce (or at least affect) animals’ reactions. The stimuli tested are mostly visual, with a person introducing an object on the floor before the animal’s habituation to the site, or with an object being dropped suddenly near the animal. The habituation period to the test site before confrontation with the new object, can vary from 1 to 15 min, and the animals tested should be calves or young bulls, in order to minimize the impact of previous experiences on their behavior [14]. The most commonly used variables are the latency to approach the object, the minimum distance to the object, the frequency and duration of contact, the type of exploration (e.g., licking, sniffing), body posture, and vocalizations.In a study by [8] almost all of the behaviors studied in the novelty test with an object before and after weaning—such as the latency to touch, the percentage of time focused or not focused, the percentage of time the animal touched the object, and the number of times it played with the ball—tended to be correlated.Other behavior tests are possible when handling animals in a race or chute, including the type and speed of exiting. The exit from the chute has been used as a measure to evaluate temperament and is indicative of the same type of stress that animals suffer when they meet humans. Cortisol levels are positively correlated with altered behavior inside the chute and with speed when exiting it. These behaviors reflect higher stress levels and are usually associated with less time feeding and resting, and with a compromised immune system [16].The exit velocity from the chute can be evaluated using infrared sensors 1.83 m away from the chute according to [17]. This is a simple method to implement and is safe for the operators performing it, while also being simpler for the animals, as they do not have to be further restrained or be manipulated, as in the test where the percentage of eye-white or the behavior of the animal locked in the chute is determined [18]. A study showed that the exit speed from the chute tended to increase with age [18].The human approach test is more easily performed in the animal’s comfort environment and assesses the human-animal relationship (in this case, fear from humans) which impacts the safety of stockpersons, as well as the well-being and productivity of the animals [15].The quality of the human-animal relationship depends on the behavior of both parties. However, animals usually generalize attitudes and behaviors toward humans (although cattle can discriminate between different people). The interactions between the two can be subdivided into five main types: static visual presence, walking among animals without tactile contact (but with the possibility of using their voice), physical contact, feeding, and intrusive [19]. The animal can perceive whether an interaction is positive, neutral, or negative, and this perception is in turn related to the animal’s previous experiences. Humans can unconsciously emit calm or dangerous signals, often ignoring fear, aggression, or calm signals from the animal, therefore human behavior and know-how is crucial. It is humans who mostly determine the number and nature of their contact with animals, with animals only reacting to them (although they may also initiate them). Some studies show that neither the breed nor the age of the animals significantly influences the animals’ withdrawal reactions, but rather the human-animal relationship existing on the farm. The size of the herd also has an effect on the animals’ reactions, with animals coming from larger herds being calmer [2].The objective of the present work was to evaluate the behavior of Mertolenga-breed young bulls when exiting the chute, when exposed to novelty, and in a human approach test, as well as to understand whether the responses between these tests are related. The results may support the inclusion of some of these tests in the classification of bulls to be used as breeders.2. Materials and Methods2.1. Facilities, Animals and ManagementThis study was conducted at the Mertolenga Cattle Breed Testing Center in Évora (Portugal) between May and November 2021.The Mertolenga Testing Center is capable of testing up to 100 males simultaneously. The performance tests carried out aim to contribute to breeding programs, and to identify and select the best bulls to be used as semen donors for use in the breed’s artificial insemination program, and for storage in the Portuguese Bank of Animal Germplasm.The males to be tested are chosen according to their genetic merit for calving interval and maternal ability, their development, and their conformation. The performance tests follow a protocol approved by the Portuguese competent authority (Direção Geral de Alimentação e Veterinária—DGAV) and in accordance with the standards of the International Committee for Animal Recording. Upon entering the center, the animals undergo a 23-day adaptation period to individual feeders, after which the performance test begins. At the end of the testing period, the males that pass the technical evaluation, are morphologically classified and submitted to an andrological examination. Only the approved males whose andrological examination results are considered apt, are validated and registered as future breeding stock [20].In this study, 29 Mertolenga young bulls entered the farm from the end of May to the beginning of June 2021. These animals came from 16 different farms, as described in Table 2.As there was not enough room for all of the animals in the same paddock, they were subsequently divided into two groups according to their weights and were always kept in side-by-side roofed pens, with access to an outside loafing area. Although they were in two adjoining pens, the study was conducted as a single group.Every three weeks, the animals were taken to the chute to be weighed in order to calculate their weight gains and to perform individual novelty tests. The chute was located on the side of the barn where the animals lived, along which there was an access corridor.All tests were performed by the same person, dressed in the same way, in order to reduce possible influences.2.2. Evaluation of the Chute ExitIdeally, as mentioned, this test is performed using an infrared device for better accuracy. In this work, since it was not possible to use this type of equipment, we used the evaluation of the exit of the chute on a scale of 1 to 3, where 1 corresponds to an exit at pace, 2 at trot, and 3 at gallop.This evaluation was always performed by the same operator, in order to minimize human error (Figure 1).2.3. Novelty Test with a BallTaking advantage of the fact that the animals were going to the chute to be weighed, a novelty test was performed with a striking-blue colored ball, of approximately 70 cm in diameter, placed in a fenced area adjacent to the chute (Figure 2). Each animal leaving the chute had to pass through this area, consisting of approximately 50 m2 earthen floor, which then opened to their usual home pen. In a first approach, only the behavior of the animal passing through this route was recorded. After 3 weeks, at the time of the new weighing, each animal had 1 min to become acquainted to the area before the ball was placed as a novelty factor for another 4 min. During this time, all behaviors were continuously recorded (e.g., vocalizations, defecations, time, and minimum distance of approaches to the ball, among others).2.4. Human Approach Test at the Feed CorridorThe third test in this study, the human approach test, was conducted from the feeding corridor and the same animals were included so that the results of the different tests could be correlated. After the novelty test was performed, the animals remained in a pen near the chute, and only after all of the animals had performed the novelty test and the food distributed, were they returned to their holding place where the human approach test was performed.At 3 m away from the feed trough, when an animal was ideally isolated (with no more than two animals within 1 m of it), the operator began the approach by taking one step per second with one arm stretched forward at 45° and the palm facing down toward the animal (Figure 3). When the animal reacted and moved away or turned its face away, the test was terminated and the distance to the trough was recorded.A scale of 0 to 3 was devised to assess the flight distance, with 0 corresponding to touching the animal, 1 to a distance of up to 1 m, 2 to a distance between 1 and 2 m, and 3 to a distance between 2 and 3 m.2.5. Data CollectionAll tests were filmed with a cell phone camera in order to facilitate subsequent analysis, and all field records were made on paper and later entered into an Excel sheet.The records collected in the chute were the identification of the operator and the animal, the weight, and the type of exit from the chute (evaluated as pace, trot, or gallop).In the novelty test, in the first 60 s in which the animal is alone in the pen adjacent to the chute, the following data were recorded: the time (in seconds) in which it was active or stopped, whether it moved toward the camera (located outside the pen) or to the water trough (present in the pen), and whether it vocalized. In the next 4 min the following were recorded: the latency time to touch the ball, the time the animal was attentive to the ball and distracted, the number of touches it gave the ball, whether or not it played with the ball (rated yes—1/no—0), the time it was active or immobile, whether it urinated/defecated, and again, whether it vocalized, went to the camera, or to the water trough.When necessary, comments were added such as whether a person or tractor passed close to the pen during the test, among others.Most of these records were collected directly onto Excel tables during the analysis of the test videos.In the human approach test in the feeding corridor, the final distance from the operator to the feeding trough was calculated in a 0 to 3 scale, recorded on paper, and later confirmed by video analysis.2.6. Statistical AnalysisFor the statistical analysis, data collected on six different days during the performance test were used. In addition to the behavioral characteristics evaluated, the following information about each animal was considered: identification of the animal, date of birth, farm of origin (classified by letters in order to maintain confidentiality), and weight and age at the entrance and exit of the performance test. As mentioned earlier, some nominal variables were converted into numeric ones, namely the exit from the chute (pace—1, trot—2, gallop—3) and the playing behavior (yes—1, no—0). All data were collected in Excel tables.All analyses were conducted with the SAS® 9.4 software [22] using different procedures (Proc’s) with different models and including several variables as factors, but in the final model analysis of each behavior parameter, only the factors that significantly influenced them for p < 0.05 were considered.Initially, several descriptive statistics were estimated with the Proc Means, Proc Freq, and Proc Univariate procedures in order to characterize the available data. A preliminary correlation analysis was also performed among all behavioral parameters to assess their associations, and then each parameter was analyzed individually.Through the novelty test, we evaluated the exit from the chute (classified as pace, trot, or gallop), the activity of the animal, the latency to touch the ball (in seconds), the number of touches to the ball, and whether the animal played with the ball.Chute exit was analyzed using Proc Glimmix, in which the effect of the farm of origin, age, and weight of the animal were included as factors. Farms that had only one animal (total of 5 farms) were removed from this analysis.Activity behavior was initially analyzed using Proc GLM with a model that included the effect of the animal’s farm of origin, and later in another way, with a mixed model using Proc Mixed, in which the effects of the type of exit from the chute, the operator, and the animal were considered (the latter as a random effect).The latency to touch was analyzed using Proc GLM with a model that included the effect of the farm of origin, and subsequently with a mixed model using Proc Mixed with a model that included the type of chute exit as a fixed effect, and the animal as a random effect.The number of touches on the ball was also analyzed with a mixed model and with Proc Mixed, but with a model that included the effects of animal age and latency to touch the ball as covariates, and the animal as a random effect.The parameter play (yes or no) was analyzed with Proc Logistic, modeled for the probability of the animal “play = yes” and with latency to touch as a covariate.As the human approach test depended on the cooperation of the animals, it was not possible to obtain records from all of the animals on every day of the test, therefore only animals with 5, 4, 3, or 2 records were considered. We obtained results from 17 animals that performed the test more than 2 times, 9 animals that performed the test once, and 3 animals that never came to the feeding trough to perform the test.Finally, a chi-square test of independence was performed between the distance to the trough and the type of exit from the chute, in which the distance to the trough was classified into only three classes as no animal was touched (<1 m, 1 to 2 m, and 2 to 3 m), and the type of exit from the chute into two classes (pace and trot, or gallop).3. Results and DiscussionData from 29 Mertolenga young bulls from 16 farms on six different dates were used. These animals started the test with an average of 321.5 ± 30.3 days of age and 320.1 ± 44.6 kg of weight, and finished the test with an average of 426.5 ± 30.3 days and 445.1 ± 55.1 kg.According to Table 3, when evaluating the exit from the chute it can be seen that the animals averaged between pace (1) and trot (2). After leaving the chute, while in the area without the novelty element, they were active on average of 35.5 s. Once the ball was introduced to the park, the latency to touch the ball was 132.8 ± 95.0 s, the number of touches on the ball ranged from 0 to 9 touches, and most of the animals did not play with the ball (88%).The average flight distance was 1.70 (in a scale from 1 to 3) and none of the animals were touched.In Table 4 it can be seen that:the exit from the chute was significantly influenced by the effect of the farm where the animal came from, the age of the animal, and its weight.the activity of the animal during acclimatization to the space was significantly influenced by the animal itself, by the farm of origin, by the operator, and by the type of exit from the chute.the latency to touch the ball was significantly influenced by the animal, its farm of origin, and its type of exit from the chute.the number of touches on the ball was significantly influenced by the age of the animal and the latency to touch.the play behavior was significantly influenced by the latency to touch the ball.the distance from the operator to the feed trough in the human approach test was not significantly associated with the exit from the chute, with an error of 7% (p = 0.071).It would be interesting to find a significant correlation between the exit from the chute and the distance to the trough, in order to be able to correlate the two behavior tests performed. However, the results showed a non-significant positive correlation (0.034; p = 0.81).Table 5 shows significant correlations between the animal’s exit from the chute and its weight and age.As for the exit from the chute analysis, it was possible to observe that the older the animals, the less likely they were to exit the chute at trot or gallop, in other words, the greater the tendency to exit the chute more calmly, meaning that they would leave at pace (Scheme 1). These results are in agreement with those demonstrated by [14].There were significant statistical differences on the exit of the chute between an animal leaving at pace and an animal leaving at gallop, as well as between an animal exiting at trot and an animal leaving at gallop (p < 0.01).According to [18], chute exit speed tends to increase with age. In addition, animals with faster chute exits have been shown to have lower growth rates, worse carcasses, a weaker immune system, and higher cortisol levels when handling them [14]. Not only are they more difficult to handle and dangerous for keepers, they can also influence the behavior of the other animals around them. However, in our study we found that older animals were more likely to leave the chute walking (Scheme 1).Regarding the analysis of the latency to touch, the longer the animal took to touch the ball, the fewer times it touched the ball (Scheme 2). At times, this happened with very frightened animals that would leave the chute very quickly, and when the ball was placed in the park as a novelty factor, they would immediately invest on it. Meanwhile, calmer animals sometimes took a little longer to touch the ball, but then proceeded to play with it. As they became older, the number of touches also decreased (Scheme 3), perhaps also due to the habituation factor of the test. This lack of consistency in behavior between the tests was in agreement with [15], who also stated that the animals’ response to the novelty test may be due to the type of object used; in our study the same colored Pilates-ball was used.According to [13], habituation is common and a possible explanation for these results. In their study, which was conducted three times with a three to four-week interval, a decrease in animal interactions with a new object was observed. Scheme 4 shows that the longer the animal takes to touch the ball, the less likely it is to play with it. That is, the older the animal, the greater the latency to touch the ball and the fewer touches it makes to the ball (Scheme 2 and Scheme 3), with a consequent reduction in the probability of playing with it.As for the human approach test, animals that left the chute more calmly, at pace, allowed themselves to be approached more than the animals that left the chute at trot or gallop (Scheme 5), as also mentioned by [15].According to [3], the flight distance test performed at the feed trough was not affected by the age of the animals, and decreased significantly throughout the repetition of the test. These results were not observed in our study (only in animal 7247), and there was a great variability of results obtained for each animal (Scheme 6). As mentioned previously, not all animals performed the test (3 animals), and curiously, two of these three animals came from the same farm and always left the chute at a gallop pace—these were probably animals with fear/distrust issues. It is also suggested that the results of this human approach test were influenced by the animals surrounding the animal performing the test. Although we always endeavored to apply the test only to animals relatively isolated, this was not always possible.4. ConclusionsThe aim of this work was partially achieved. We managed to evaluate all the behaviors in question and obtain results that were interesting, and showed different responses. There was great variability in the animals’ responses in both tests, likely because the animals came from different farms and therefore had different experiences, but may also be due to genetic predisposition.There was no evidence that the tests’ results were related to each other in any of our animals. It is necessary to continue data collection by performing behavioral tests with as many animals as possible, in order to understand all the factors involved including environmental and genetic factors, and even possible habituation to the tests over time.The importance of animal welfare is being increasingly recognized. The parameters studied are related not only to management and ease of handling, but also to productivity, to products’ quality, and to profitability, as previously mentioned. Hence, there is a need to implement selection programs that take into account criteria such as the temperament of the animals. This could lead to the selection of calmer animals, which in turn tend to be more efficient.We believe that these behavior tests should be further studied on potential sires, so as to increase docility and manageability of autochthonous beef breeds.

animals : an open access journal from mdpi

10.3390/ani12070938

PMC8996914

Unwanted toileting is amongst the most undesirable behaviors in cats. The aim of this study was to test the effect of a chemical cue derived from cat anal glands on the elimination behavior of domestic cats tested individually. A total of 31 cats were tested for 23 h in an enriched test room where they had the choice between two litter trays, one sprayed with the composition tested and the other with the control. We measured the weight of elimination, the weight of urine, we recorded the type of elimination, and counted the urine spots and stool piles. We also looked at the duration of exploration of litter trays, and what trays the cats were choosing to toilet as a first and second choice. We demonstrated, across all parameters, that cats urinated and defecated significantly less in the litter trays where the chemical cue was sprayed. These results confirm and strengthen those of our previous study. Future studies could explore if a chemical cue derived from male cat anal glands might be used to manage unwanted toileting in cats.

Unwanted toileting is amongst the most undesirable behaviors in domestic cats and can lead to conflicts between cats and the communities they are living in. This study aimed to confirm the effect of a semiochemical composition, reconstituted volatile fraction derived from cat anal glands, on the elimination behavior of domestic cats. A total of 31 cats were tested individually, for 23 h, in a blinded randomized choice test, with two litter trays, one sprayed with the treatment and the other with the control. Parameters included elimination weight, urine only weight, the record of the elimination type and counting of urine spots and stools, exploration duration of each litter tray, and first and second choice of litter tray to eliminate. Across all parameters, cats urinated and defecated significantly less in the litter tray where the semiochemical composition was sprayed than in the litter tray where the control was sprayed (for example: elimination weight p < 0.0001; urine only weight p < 0.0001; exploration duration p < 0.0001, and first elimination choice p < 0.0001). These results demonstrate that a semiochemical composition-derived from cat anal glands significantly decreases elimination at the location where it is sprayed. Future research is warranted to explore the possibility to manage unwanted toileting using this semiochemical composition.

1. IntroductionAnimals explore their physical and social environment and gather information to secure the necessary resources for their survival. This environment encompasses other animals, conspecifics, or other species. Gathering information about those animals, whether they are predators, conspecifics or preys, is an essential part of assessing the environment, to avoid threats and conflicts, find and select mating partners. Communication can be defined as the transmission of information from one animal to another [1] using different sensory modalities, such as visual, auditory, and chemical signals [2,3]. All animals emit volatile organic compounds (VOC), some of these compounds have evolved to become chemical signals emitted towards a receiver, either in the same species or other species [4]. Chemical communication allows signals to travel in the air or be deposited on a substrate, releasing the information they carry even when the animals are not present [5]. In a species that can sometimes adopt a solitary lifestyle [6] such as domestic cats, this is essential, as, in natural conditions (without human intervention) and except for cat colonies, conspecific encounters may not be very frequent [5].Research has shown that several secretions such as urine, faeces and secretions from various glands can be vectors of chemical communication [7,8,9,10] and chemical compounds have been identified in urine [11,12], faeces [13,14], and anal glands [9]. In the domestic cat, the anal glands’ structure has been described [15,16] and their chemical composition explored [9,16,17]. Scent communication using secretions from anal glands have been described in several species such as the honey badger (Mellivora capensis) [18], spotted hyena (Crocuta crocuta) [19,20] and red foxes (Vulpes vulpes) [7,21,22,23,24] Anal gland compounds may have a role in individual recognition [13], territorial marking or reproductive advertisement [11]. The research field of linking compounds to a specific role is expanding, such as exploring the interaction of chemical compounds and bacteria present in the glands [17] or the interaction between compounds and their binding proteins [25]. Cat elimination behaviour is a complex process [26,27] that fulfils the role of excreting waste from the body and providing information about species, sex, reproductive status of conspecifics and age for male cats [13,14]. This behaviour is different from marking behaviour where the cat may spray urine, scratch surfaces [28] and rub on objects and individuals [29,30,31] as a form of chemical communication. However, both in elimination behaviour and marking behaviour, compounds are present that seem to influence behaviour. Urinary extracts for example have been shown to influence the choice of location to eliminate in outdoor cats [32]. Regarding anal glands’ secretions, they can be deposited on a substrate or are excreted on faeces. While some of the roles of chemical compounds isolated from cat anal glands have been explored [9], there is a lack of knowledge about the influence of very specific compounds can have on cat behaviour, especially on their choice of location to eliminate. Today, different repellents systems exist to manage cat inappropriate elimination, but elimination behaviour plays a crucial role for the cat and its communication. A new approach is to develop strategies to manage this behaviour, to promote human and cat cohabitation. A previous work performed in a cattery setting showed that a semiochemical composition, a reconstituted partial volatile fraction derived from the secretion of an entire mature male cat (13 years-old) anal glands, deterred cats from defecating at the location where it was sprayed [33]. The aim of our study was to use the same semiochemical composition on cats tested individually in a randomised controlled choice trial, to test its effect on cat elimination behaviour. Our hypothesis was that the semiochemical composition tested would have an influence on cat elimination behaviour and we predicted that the semiochemical composition would deter cats from defecating at the location where it is sprayed.2. Materials and MethodsThe study was approved by the IRSEA ethics committee (National French Ethics Committee C2EA125) under approval number CE_2019_07_CEIS_03.2.1. Animals/SubjectsIt was conducted on 31 cats (10 entire males, 10 entire females, 4 neutered males, and 7 neutered females, see Table 1), in an enriched test room where they were free to move. All cats were examined by a veterinarian before being included in the study and were in good health, free of chronic disease or lower urinary tract disorders.2.2. MaterialThe test room (12 m2) contained shelves, bedding, hiding places, scratching posts, a tunnel, toys, and litter trays (Savic Aseo litter tray, L56 cm × L39 cm × H27.5 cm) (Figure 1). Dry food was provided in a puzzle feeder (Karlie flamingo NORTHMATE puzzle feeder) to match the specific needs of each cat (i.e., depending on the cat: Royal Canin Kibble VCN Adult Cat, VCN Senior Cat Stage 2 and Cat Satiety) and in a quantity sufficient for maintaining an adequate weight. Wet food (Hill’s prescription diet i/d digestive care) was used as enrichment and put in an enrichment feeder (Kong classic and Kong Extreme, size S). Water was provided ad libitum in a water bowl. The room also contained a trap camera (Coolife trap camera model H881, 16 MP 1080P HD) with motion detection sensors, focused on the area where the litter trays were positioned. 2.3. ProcedureBefore testing, the cats underwent a six-week progressive habituation program to ensure that they were comfortable staying alone in the test room, as they usually live in an enriched cattery environment with conspecifics. The habituation procedure consisted of first spending some time in the test room by groups of three to four cats living in the same cattery, a familiar human being present in the room. Then each cat spent some time only with a familiar human present during 20 min to provide positive interaction and opportunities to play. The familiar human then provided a small number of treats and left the room, for 20 min. The time the cat was left alone was gradually increased, from 20 min to five hours. The number of sessions depended on how each individual cat responded to being alone and was a minimum of six sessions and a maximum of ten sessions. Cats were monitored during the habituation process for signs of stress such as excessive agitation, pacing, continuous meowing, pawing/scratching at the door, refusing to eat or play and overgrooming. 32 cats entered the habituation process, and one cat was excluded from the study. All remaining cats (31) were comfortable being left alone in the test room at the end of the habituation process and didn’t show any of the aforementioned signs of stress any more. Their behaviour can be described as follows: they explored the room, played with toys, ate treats, or rested on the resting places. The cats were tested individually and stayed 23 h in the enriched test room, alone, except during short sessions where they positively interacted (positive contact and play) with a familiar human. Four interaction’s sessions (the first one of 30 min, then the following ones of 20 min each) were performed during the test.The chosen experimental design was a Randomized Controlled Trial, where the cats had the choice between two litter trays (named 1 and 2) present in the test room. In each litter tray, one treatment was sprayed on the litter substrate. The treatment used was a reconstituted partial volatile fraction (named CEMS for Cat Elimination Modulation Semiochemical) of the secretion of a male cat’s anal gland, as described in Kasbaoui et al., 2022 [33]. Briefly, the treatment is composed of several volatile compounds present in the anal glands secretions obtained from an entire male cat, following their identification by GC-MS analyses. The actual treatment product is made from these same several pure chemical compounds, industrially purchased (SIGMA, Saint-Quentin-Fallavier, France), at the final concentration of 2% in a mix of ethanol and water (60:40). The other was the control (i.e., a mixture of ethanol and water, (60:40)).The procedure was blinded and randomized, the treatment being applied randomly to litter tray 1 or litter tray 2 for the whole duration of the test of each individual cat. Litter trays were filled with 1.5 kg of non-agglomerant litter substrate (brand “Prop’chat NF) and treatments were sprayed (five sprays) on the litter substrate and mixed with it, according to the randomization list. Litter trays were weighed each day at the start of the experiment (Expondo digital weighing scale, model SBS-PT-40/1, with a precision of 1 g). The litter trays were then installed in the enriched test room and secured at their predetermined location with double-sided tape. The test started at 10 a.m. in the morning. The cat was brought to the test room approximately five minutes after the litter trays were put in place, to allow the ethanol to evaporate. The familiar human stayed with the cat for the interaction session during 30 min, then left the cat alone in the room, having given the Kong toy filled with wet food behind. The familiar human came back at 12:30 p.m., 2 p.m. and 4:30 p.m. for interaction sessions. After the last interaction session, the litter trays were removed from the test room in order to record data, perform the cleaning procedure, renew the litter substrates and re-apply treatments. The procedure was as follows: the litter trays were removed from the test room and weighed. The type of elimination present was recorded (urine only, stools only, or urine plus stools) and the counting of urine spots and stool piles was performed and recorded. Stools were then removed if present and the litter trays were weighed again (giving the parameter of the “urine only” weight). After all of the data were recorded, the litter substrate was discarded in a special bin. For cleaning, each litter tray was sprayed on the inside and the outside with a detergent disinfectant product that destroys liposoluble compounds (DNA02 LeVrai Professionnel, https://www.bernard.fr/, accessed 29 October 2021) and thoroughly wiped until completely clean. Then, it was sprayed a second time with water to rinse the detergent and avoid any contact between the detergent and the cats. After the litter tray was dried, 1.5 kg of fresh litter substrate was weighed and placed in the litter tray. Finally, according to the randomization list, the designated treatment was sprayed on the litter substrate and mixed with it. The litter trays were weighed to record the weight of the unused litter tray with the treatment applied and were put back in place. The whole procedure lasted less than 10 min. Then the cat was left alone for the night, and the test ended at 9 a.m. the next morning, when the cat was brought back to its cattery. Data were recorded again, and the test room was cleaned. The cleaning procedure was to remove all that was washable, then steam clean the floor and walls, spray a detergent that breaks down proteins and liposoluble compounds (VIGOR surpuissant, https://www.bernard.fr/, accessed 29 October 2021), then rinse the detergent with water, mop the floor, and wait 10 min for the room to dry completely before preparing the room for the next test. The parameters studied included total elimination weight (urine plus stools, which was calculated by subtracting the weight of the unused litter tray with the treatment from the weight of the used litter tray), urine weight (which was calculated by subtracting the weight of the unused litter tray from the weight of the used litter tray without stools), type of elimination (0 = no elimination, 1 = urine only, 2 = stools only, 3 = urine plus stools), and frequency of urine spots and stools (counting each urine spot and stool pile for each cat). The videos recorded were exported twice a day and stored. Video analysis allowed to study the exploration duration of each litter tray (i.e., time when the cat sniffed the litter tray, had his head inside the tray, or scratched the litter tray) and the first and second choice of litter tray to eliminate. 2.4. Statistical AnalysisData analysis was performed using SAS 9.4 software Copyright (c) 2002–2012 by SAS Institute Inc., Cary, NC, USA. The significance threshold was fixed at 5%. Continuous parameters (total elimination weight, urine weight, and exploration duration) were analysed according to the treatment applied (Y and Z), the sex of cats (entire male, entire female, neutered male, and neutered female), and the treatment sex interaction. The normality of residues from raw data was verified using the UNIVARIATE procedure. If normality was verified, the different effects listed on these variables were evaluated with a GENERAL LINEAR MIXED MODEL using the MIXED procedure with “block” in the random statement. If normality was not verified, Box-Cox data transformation was performed using the TRANSREG procedure to try to obtain normality. After a Box-Cox transformation, normality was obtained and transformed data were modelled using the MIXED procedure. If there were significant differences, multiple comparisons were analysed with the TUKEY test using the LSMEANS statement.Binary variables (first and second choices of litter tray for elimination) were analysed according to the treatment applied, the sex of cats, and the treatment × sex interaction. These effects were evaluated with a GENERALIZED LINEAR MIXED MODEL (Block being considered as a random effect) using the GLIMMIX procedure, specifying the BINARY distribution in the MODEL statement. If there were significant differences, multiple comparisons were analysed with the TUKEY test using the LSMEANS statement.Polytomous variables (type of elimination, frequency of urine spots and stools) were analysed according to the treatment applied, the sex of cats, and the treatment × sex interaction. These effects were evaluated with the help of the GENERALIZED LINEAR MIXED MODEL (Block being considered as a random effect) through the GLIMMIX procedure, specifying the MULTINOMIAL distribution. 3. Results3.1. Elimination WeightFor the total elimination weight (Table 2), we found a significant effect of: (i)treatment (GLMM; DF = 1; F = 43.44; p < 0.0001) where the elimination weight was significantly lower in the litter tray sprayed with CEMS than in the litter tray with the control;(ii)sex (GLMM; DF = 3; F = 7.48; p = 0.0009) where the weight of elimination in females and the weight of elimination in males were significantly lower than in neutered males; and(iii)interaction between treatment and sex (GLMM; DF = 3; F = 4.60; p = 0.0100).There was a significant difference between females and neutered males (Tukey’s test; DF = 27; t value = −4.28; p = 0.0011) and males and neutered males (Tukey’s test; DF = 27; t value = −4.22, p = 0.0013). Regarding the significant interaction effect, there was a significant difference between CEMS and control within the neutered males (Tukey’s test; DF = 27; t value = 5.37; p = 0.0003) and between neutered males and males within control (Tukey’s test; DF = 27; t value = −5.58; p-value = 0.0002).For the weight of urine only (Table 2), we observed a significant effect of: (i)treatment (GLMM; DF = 1; F = 41.23; p-value < 0.0001) where the weight of urine only was significantly lower in the litter tray sprayed with CEMS than in the litter tray with the control;(ii)sex (GLMM; DF = 3; F = 6.81; p-value = 0.0006) where the weight of urine only in females and the weight of urine only in males was significantly lower than in neutered males;(iii)interaction between treatment and sex (GLMM; DF = 3; F = 4.46; p-value = 0.0072). There was a significant difference between females and neutered males (Tukey’s test; DF = 54; t value = −3.78; p = 0.0022) and male and neutered male (Tukey’s test; DF = 54; t value = −4.06; p = 0.0009).3.2. Type and Number of EliminationsWe observed a significant effect of treatment for the type of elimination (GLMM; DF = 1; F = 15.61; p-value = 0.0005), the number of urine spots and the number of stool piles per litter tray (urine spots: GLMM; DF = 1; F = 14.42; p-value = 0.0008; stool piles: GLMM; DF = 1; F = 6.66; p-value = 0.0152). Cats significantly chose to eliminate less (score of elimination) in the litter tray sprayed with CEMS than in the litter tray sprayed with the control. They also urinated and defecated significantly less in the litter tray sprayed with CEMS than in the litter tray sprayed with the control (Table 3).3.3. Exploration of the Litter Trays and Choice of Litter Tray to EliminateCats explored the litter tray with CEMS significantly less than the litter tray with the control (mean ± SE: CEMS 58.16 ± 8.04 s versus control 132.56 ± 13.60 s; GLMM; DF = 1, F = 27.86; p-value < 0.0001) and chose the litter tray with the control to eliminate significantly more than the litter sprayed with CEMS, for the first choice (GLMM; DF = 1; F = 100.33; p-value < 0.0001) and second choice (GLMM; DF = 1; F = 5.24; p-value = 0.0293) of elimination.For the first choice, a significant difference in the sex effect is observed (GLMM; DF = 3; F = 23.49; p-value = <0.0001). No significant difference in the sex effect is found for the second choice (GLMM; DF = 3, F = 0.02; p-value = 0.9949).These results showed that cats avoided the litter tray sprayed with the CEMS treatment (Table 4). During the study, two cats out of 31 urinated outside the litter tray (6% of the population tested).Taking into account all results, cats significantly chose to eliminate (urinate or defecate) in the litter tray sprayed with the control rather than the litter tray sprayed with CEMS. 4. DiscussionThe aim of this study was to assess the effect of a semiochemical composition, reconstituted volatile fraction derived from male cat anal glands (of which some effect had been shown in a previous study [33]), on the elimination behaviour of cats tested individually, in a randomised controlled choice trial. We showed that cats urinated and defecated significantly less in a litter tray sprayed with the treatment CEMS than in a litter tray sprayed with the control. They also explored the litter tray treated with CEMS significantly less than the one with the control, and their first and second choice for elimination was significantly the litter tray sprayed with the control treatment. There was a significant effect of sex (females versus neutered males and males versus neutered males) for the elimination weights (urine plus stools and urine only) where neutered male produced a higher weight of total elimination and a higher weight of urine only then both females and males.4.1. Effect of TreatmentOur study showed that cats urinated and defecated significantly less in the litter trays sprayed with the CEMS treatment, across all of the elimination behaviour parameters: elimination weight, type of elimination, urines spots, and stools piles. These results confirm the ones of our previous study regarding defecation and give more data about urination (urine weight and urine frequency). The CEMS treatment also had a significant effect on the first and second choice of litter tray for elimination, with the control litter trays being significantly chosen for the first elimination event and for the second elimination event. Some cats even always chose the control litter tray. So, the CEMS treatment seems to have an aversive effect and significantly deter cats from choosing the litter tray where it is sprayed. This is confirmed by the parameter “exploration duration of the litter trays”, with the litter tray sprayed with the CEMS treatment being explored significantly less than the litter tray with the control treatment. The exploration duration parameters and the choice of elimination parameters were co-dependent, as the exploration of the litter tray included the scratching of the litter substrate. Urine, feces, and sebaceous glandular secretions contains chemical compounds that are vectors of chemical communication [7,8]. Anal gland secretions can be excreted in cat urine [7] and also on cat feces [9]. When allowed outside and living in a colony, cats bury their feces in their core home range and tend to bury them less in the peripheral areas of their home range [8]. Entire male cats also sometimes do not cover their feces as well [8] and cat feces contain the chemical basis of species, sex, and individual recognition [13]. The semiochemical composition tested in this study is specific to an entire male cat. Therefore, it is possible that the avoidance behaviour was triggered by this semiochemical composition, considered as a territorial scent marking. 4.2. Effect of SexThere was a significant effect of sex in three out of eight parameters. The first effects were on the elimination weight (i.e., total elimination weight (urine plus stools) and urine only weight). For both parameters, the statistical analysis showed that there was a difference between female and neutered males, and males and neutered males. From a previous study of cats in our cattery [33], we showed that neutered male cats were also significantly heavier than female cats. Neutered male cats are also older than male and female cats. As water intake and urine production can be influenced by cat weight [34], it is possible that the elimination weight of neutered male cats, both depending on their average weight and their condition, could be higher than that of male and female cats that were much younger (approximately three years old) and for the females, significantly lighter. This hypothesis is also supported by the fact that there was no effect of sex on the number of urine spots. Therefore, neutered males did not urinate more frequently, but produced more urine during their urination. The avoidance effect depended on the cat, and in our study, it was present in the four sex categories tested (entire males and females, neutered males, and females), despite significant differences between them. During the study, no cats showed signs of stress making necessary to stop the test. A few cats were more sensitive than others and refused to use either litter tray and sprayed and urinated outside the litter tray. It is possible that the perception of this semiochemical composition was perceived as stressful for the most sensitive cats [35] in our enclosed study setting. However, in the test conditions, the cats could be away from the litter trays, but could not leave the room. In real life conditions, cats could freely leave the location where the CEMS treatment would be applied, rendering less likely the possibility to be stressed by continuous exposure to a signal triggering avoidance. Moreover, chemical signals are a part of the cats’ daily lives [5] so the signal would be, in real life, one signal present in a wealth of other chemical signals, and part of the cats’ natural environment. 5. ConclusionsWe demonstrated that a semiochemical composition, reconstituted volatile fraction derived from cat anal glands, deterred the subjects from defecating and urinating in the litter tray where it was sprayed. This semiochemical composition does not prevent the elimination of the cat but could promote redirection of elimination. Future research is warranted to explore the possibility of using this semiochemical composition to help manage unwanted toileting in real life conditions and if this composition, while triggering avoidance, may also modulate other behaviours of cats and/or have stressful effects on cats.

animals : an open access journal from mdpi

10.3390/ani11092710

PMC8471945

The measurement of stress hormones has become a widely used and effective tool for evaluating adrenocortical activity in animals. However, to correctly interpret stress measurements, the potential sampling bias resulting from an oversampling of individuals in different states of pregnancy has rarely been investigated. We found significant yearly variations in states of pregnancy, which is related to the conditions of the females due to the snow cover duration before and at the start of the reproductive period. These results are important for improving the interpretation of stress hormone concentrations in free-ranging populations during the breeding and reproductive periods.

The measurement of glucocorticoid metabolites (GCMs) in faeces has become a widely used and effective tool for evaluating the amount of stress experienced by animals. However, the potential sampling bias resulting from an oversampling of individuals in different states of pregnancy has rarely been investigated. In this study, we validate a noninvasive method for measuring gestagen metabolites in female mountain hares (Lepus timidus) under controlled conditions. We also measured the concentration of gestagen metabolites of females in a free-ranging population during the early breeding and post-breeding periods from 2014 to 2019. We found significant yearly variations in gestagen metabolites, which were related to the condition of the females due to the snow cover duration before and at the start of the reproduction period. GCMs were significantly influenced by the gestagen metabolite levels. These results are important for improving the interpretation of GCM concentrations in free-ranging populations during the breeding and reproductive periods.

1. IntroductionThe measurement of glucocorticoid metabolites (GCMs) has become a widely used and effective tool for evaluating adrenocortical activity in animals [1,2]. Faecal GCM measurements can be easily obtained without any need to handle the animal, rendering the sampling process almost feedback-free and, therefore appropriate for evaluating stress in free-ranging wild animals [3]. However, GCM excretion may depend on the reproductive state of the animal responsible for the collected sample [4]. To correctly interpret GCM measurements [2], it is important to clarify this relationship.Our model species, the mountain hare (Lepus timidus), is a perfect species for testing the influence of different reproductive states on GCM excretion, because GCM and noninvasive genetic sampling (NIGS) methods have recently been developed specifically for this purpose [5,6,7,8,9]. The mountain hare is an elusive species that is nocturnally active, has no sexual dimorphism, and is sensitive to disturbances [9,10,11]. It is a nonterritorial species, and individual home ranges show considerable overlap [12]. Mountain hare females typically have two to three litters during a single reproductive period in the Alps, which regularly starts in April and ends in August [13]. The reproductive success of mountain hares is threatened by climate change and by stress due to human recreational activities [9,14,15]. To understand how hare reproduction is affected by these stressors, we need to better understand how reproduction affects GCM excretion. To this end, a noninvasive method for endocrine monitoring of pregnancy has to be developed.In this study, we validate a noninvasive method for measuring gestagen metabolites in female mountain hares under controlled and field conditions. First, we evaluated the suitability of enzyme immunoassays as indicators of the state of pregnancy. Secondly, we investigated the seasonal and yearly variations in gestagen metabolites from a collection of faeces in a free-ranging population over six consecutive years, 2014–2019, in the early breeding and the post-breeding periods and how the concentrations of gestagen metabolites influence the GCM levels. Finally, we discuss how this knowledge can be used to improve the interpretation of the results of faecal glucocorticoid metabolite measurements in free-ranging populations.2. Materials and Methods2.1. Noninvasive Method for Measuring States of PregnancyTo establish noninvasive methods for the endocrine monitoring of pregnancy indicated by concentrations of gestagen metabolites in mountain hares, we obtained 81 faecal samples in the “Zoo am Meer” in Germany. We analysed them with a 5ß-pregnane-3α-ol-20-one enzyme immunoassay (EIA) to measure 20-oxopregnanes (20-OP) and a pregnanediol EIA to measure 20α-hydroxypregnanes (20α-OHP). The sampling was conducted from 2017 to 2018 during the early breeding period (March and April), the breeding period (May and June), and the post-breeding period (November–February). Young hares were born during the sampling period (April: n = 1, May: n = 1, and June: n = 4).2.2. FieldworkThe study area comprised 3.5 km2 and was situated along the Ofenpass in the Swiss National Park in South-eastern Switzerland (46°39′N, 10°11′E). The Swiss National Park is designated by the International Union for the Conservation of Nature [16] as a Category Ia nature reserve (strict nature reserve/wilderness area) and is closed to the public in the winter. It was therefore possible to study mountain hares under natural conditions without human disturbance.The study area ranged in elevation from 1693 to 2587 m a.s.l. The habitats within the study area were delimited and classified according to the habitat categories of HABITALP, a standardised habitat classification project for protected areas in the Alps [17]. The study area encompassed seven main habitat types: meadows (29%; with diverse grasses, including Nardus stricta, Festuca sp., Poa sp., Agrostis sp., Luzula sp., and sedges); timber stands (24%); scree slopes (16%); storeyed stands (12%; mixed Larix decidua, Pinus cembra, P. sylvestris, P. mugo spp., and Picea abies); sapling stands (6%; dominated by P. mugo spp.); pole timber (5%); and mature stands (5%). Residual habitats covered 3% of the area. The climate in the Swiss National Park is continental, with mean January and July temperatures of –9 °C and 11 °C, respectively [18]. The mean monthly precipitation measured at 1970 m a.s.l. is 34 mm in January and 108 mm in July [18].We collected fresh pellets over six consecutive years (2014–2019) in the early breeding (end of March until the first half of April) and in the post-breeding periods (October). The samples were collected both systematically and opportunistically, as described in detail by Reference [5]. Systematic sampling was conducted on 91 plots that were preselected on a 200-m-square grid; all hare pellets within each trial plot were collected during each visit. For the opportunistic sampling, we collected pellets as we moved from one systematic plot to the next. Only fresh faecal pellets were collected, because amplification success rates are significantly lower for pellets older than five days [5] and because GCM can be influenced by weather conditions [6]. Samples for the genetic analyses were collected and stored in separate plastic tubes without being touched by hand to minimise DNA contamination [19]. Samples for the GCM analyses were stored in plastic bags. After collection in the field, the samples were frozen and stored until they were analysed in the lab.2.3. Genetic AnalysesWe used nine nuclear microsatellites to identify individuals and assign them to the collected faeces samples: Lsa1, Lsa3 [20], Sat2, Sat5, Sat8, Sat12 [21], Sol30, Sol8 [22], and Sol33 [23]. Seven loci were analysed in R [24], and two (Sat2, Sat12) were scored qualitatively using a description of a phenotypic peak to find consensus genotypes for each replicated sample [25]. The sex of the pellet owner was determined using an assay developed by Reference [26]. The assay is only amplified in male individuals, as it amplifies part of the Y-chromosomal SRY [26]. A genotype is considered female if none of the three replicates are amplified at the SRY locus and male if at least one of the replicates is amplified. For the identification of unique genotypes, the ALLELEMATCH package in R was used, which considers genotyping errors and missing data during the assignment of individuals [27]. A unique individual is identified when the sample differs from all other samples at more than two loci, including the additional loci (Sat2 and Sat12). The DNA samples were genotyped by three independent replicates, and consensus homozygote genotypes were accepted if all three replicates were consistent. Consensus heterozygote genotypes were accepted if at least two replicates were consistent, and no more than two alleles were found across all three replicates [25].2.4. GCM AnalysesFaecal GCMs were measured using an 11-oxoaetiocholanolone EIA, which has proven suitable (based on the results of a validation study including an ACTH challenge test) for evaluating the adrenocortical activity in mountain hares [6]. Every sample was dried and thoroughly homogenised. Afterwards, a portion (0.15 g) was mixed with 5 mL of methanol (80%), shaken (30 min), and centrifuged (2500× g; 15 min). An aliquot of the supernatant (after 1:10 dilution with an assay buffer) was then analysed in the 11-oxoaetiocholanolone EIA. All intra- and inter-assay coefficients of variation were below 12%, and the sensitivity of the method was 2 ng/g of faeces. The details of the extraction procedure and the EIA can be found elsewhere [6,28,29].2.5. Pregnancy State AnalysesThe same extraction procedure as for faecal GCM was applied to measure the pregnancy status of the individuals. Aliquots of the extract were analysed in a 5ß-pregnane-3α-ol-20-one EIA (measuring 20-oxopregnanes: 20-OP) and a pregnanediol EIA (measuring 20α-hydroxypregnanes: 20α-OHP). The details of the EIAs, including any cross-reactions, can be found elsewhere [30,31].2.6. Statistical AnalysesAll statistical tests were conducted using R 3.6.3 [24].To investigate the seasonal differences in gestagen metabolites under controlled conditions, we used 81 samples in a linear model with EIA (20-OP and 20α-OHP) as the response variable and season (early breeding period, reproductive period, and post-breeding period) as the predictor variable.To investigate the variations in the reproduction activity of free-ranging female mountain hares over six years (2014–2019), we used 248 samples from 33 individuals (n = 138 from the early breeding period and n = 110 from the post-breeding period) and measured 20α-OHP based on the higher sensitivity and clearer distinction of the post-breeding period (Figure 1). The compositions of females with different pregnancy states were classified into four categories (≥3000 ng/g, 2000–2999 ng/g, 1000–1999 ng/g, and <1000 ng/g) and used to show the yearly variations in the pregnancy states of the females (Table 1). We used linear models with gestagen metabolites as the response variable, year, and season (early breeding period and post-reproductive period) as the predictor variables and individual ID as a random factor. In the investigation of variations in GCM concentrations, we used linear models with GCM as the response variable, year, season (early breeding period and post-breeding period), gestagen metabolites, and the interaction season gestagen metabolites as the predictor variables and individual ID as a random factor. We used the Shapiro–Wilk normality test to examine the distribution of data. If the variables were not normally distributed, they were transformed to meet the criteria of normal distribution.3. Results3.1. Noninvasive Method for Measuring States of PregnancyThe seasonal differences in gestagen metabolites, with peaks during the breeding period, were significant for 20-OP (F2,78 = 6.32, p = 0.003) and 20α-OHP (F2,78 = 8.77, p < 0.001; Figure 1). 20α-OHP were present at higher concentrations and enabled a clearer distinction between the post-breeding period as compared to the early breeding and breeding periods (Figure 1).3.2. FieldworkThe composition of females with different states of pregnancy during the early breeding period varied among the years (Table 1). In 2014 and 2018, the years in which we found no females with the highest gestagen metabolite levels (≥3000 ng/g) during the early breeding period, we found a higher January–March snow height as compared to the other years. We found no clear pattern for the other variables (number of days with snowfall, average daily temperatures, and number of days with air frost; Table 1). In 2019, we found one female with the highest gestagen metabolites level (≥3000 ng/g) also during the post-breeding period.The GCM concentrations varied significantly from year to year (F5,207 = 20.97, p < 0.001; Figure 2). We found that the GCMs were significantly influenced by gestagen metabolites (F1,207 = 5.42, p = 0.021), where the GCMs increased with the increasing concentrations of gestagen metabolites. For example, in 2016, the highest GCM concentrations were found during the early breeding period, which was when the most pregnant females were found in the study area (Table 1). The GCM concentration was also significantly influenced by the season (F1,207 = 25.64, p < 0.001), with higher GCM excretions occurring during the early breeding period as compared to the post-breeding period (Figure 2). The interaction season gestagen metabolites had no significant influence on the GCM concentrations (F1,207 = 0.68, p = 0.410).4. DiscussionWe validated a noninvasive method for measuring gestagen metabolites as indicators of different states of pregnancy in female mountain hares and showed variations in a population of free-ranging mountain hares over a period of six years. These results are important for understanding how the timing of the mountain hare reproduction period varies from year to year [32]. They also improve the interpretation of faecal glucocorticoid results when collecting “anonymous” faeces in free-ranging populations during reproductive periods.The selected 20-OP and 20α-OHP EIAs [30,31], used here for the first time in the genus Lepus, exhibited the required suitability for measuring the gestagen metabolite concentrations in mountain hare faeces. However, the results of both EIAs differed significantly in the measured amounts and in distinction between the early breeding and the post-breeding periods. The 20α-OHP EIA proved better-suited as an indicator of the state of pregnancy (luteal phase/pregnancy) in mountain hares.We found a high yearly variation in the compositions of females with different pregnancy states in the early breeding season in the Alps, which was in alignment with the observations of mountain hares in North-eastern Scotland [33]. Few females with the highest gestagen metabolites level (≥3000 ng/g) were observed during the early breeding period in our study, but the number of pregnant females could be expected to increase later in the year [33].Year-to-year changes in the distribution of the pregnancy states throughout the year could be attributed to weather conditions. In particular, the snow cover in January–March affects the conditions for reproduction [33,34]. Snow limits access to the nutrient-rich food that is preferred by females in the spring when the reproductive period begins [35]. The temperature seems to play a secondary role in the state of pregnancy, a conclusion supported by observations of other mountain hare populations [36].We would expect the pregnancy state of females during the early breeding period to have the most direct influence on the success of the first litters in April [10], because the pregnancy state influences the onset and length of the reproduction season. This is the case in mammals like snowshoe hares (Lepus americanus) and cottontail rabbits (Sylvilagus floridanus) [37,38,39]. It may be possible that the conditions for reproduction and the survival of the first litters in April may improve as the climate warms and the amount of snow in the Alps decreases [40,41], because more food is available earlier in the year. However, the mountain hare is a cold-adapted species and can therefore only move upward with higher temperatures; it will not find suitable environmental conditions at lower elevations [15]. Mountain hares have to follow the “green wave” up the mountain to find high-quality food until autumn [42].Interestingly, we found one pregnant female with the highest gestagen metabolites level (≥3000 ng/g) in the post-breeding period at the beginning of October. This can be interpreted as another indicator of the flexibility of mountain hares with regards to their reproductive strategy [13].The GCM concentrations in the females were influenced by year, season, and gestagen metabolites. The season affects the GCM concentrations due to the season-dependent energetic costs (which also vary from year to year). The early breeding season has a higher energetic cost due to the activation of catabolic processes associated with hypothalamic–pituitary–adrenal (HPA) axis activation, ovulation, and pregnancy [43,44], all of which lead to higher GCM concentrations in various mammal species [4]. Gestagen metabolites have a positive effect on GCMs, the strength of which depends on the pregnancy states of the females [4].Interestingly, we found differences in the GCMs between the early breeding and post-breeding periods but no differences in the gestagen metabolites. This confirms the influence of other factors on the pregnancy state (like the predation risk) that are tightly correlated with maternal off-early breeding period stress profiles [45,46], individual differences [7], and the population cycle [39,47].5. ConclusionsOur results showed how the evaluated enzyme immunoassays can be used as indicators of the state of pregnancy to improve the interpretation of the results of faecal glucocorticoid metabolites in free-ranging populations of mountain hares. We recommend that other GCM studies take into account the pregnancy state of females during the breeding and reproductive periods to minimise the potential sampling bias resulting from an oversampling of pregnant females when collecting “anonymous” faeces.

animals : an open access journal from mdpi

10.3390/ani11092618

PMC8471473

Wildlife ecological patterns are driven not only by environmental and biological contexts, but also by landscape-management schemes that shape those contexts. The present study aims to determine the effect of different environmental factors (including management schemes) on the occurrence patterns of a southern African small mammal community. Based on a landscape where three land-use contexts that differ in their levels of human presence and/or where activities coexist (private ecotourism reserve, mixed farms and traditional communal areas), and by using a body-size-based approach (i.e., using two size-based rodent groups—medium and small—as models), we found that the mean relative abundance of medium-sized species did not differ across the management contexts, but small species’ mean relative abundance was higher in the game reserve. The overall variation in rodent abundance was negatively affected by ungulate presence (possibly linked to a decrease in food availability) and by human presence (increased disturbance). Rodent abundance seems to be influenced by environmental gradients that are directly linked to varying management priorities across land uses, meaning that these communities might not benefit uniformly by the increased amount of habitat promoted by the commercial wildlife industry.

South Africa’s decentralized approach to conservation entails that wildlife outside formally protected areas inhabit complex multi-use landscapes, where private wildlife business (ecotourism and/or hunting) co-exist in a human-dominated landscape matrix. Under decentralized conservation, wildlife is perceived to benefit from increased amount of available habitat, however it is crucial to understand how distinct management priorities and associated landscape modifications impact noncharismatic taxa, such as small mammals. We conducted extensive ink-tracking-tunnel surveys to estimate heterogeneity in rodent distribution and investigate the effect of different environmental factors on abundance patterns of two size-based rodent groups (small- and medium-sized species), across three adjacent management contexts in NE KwaZulu-Natal, South Africa: a private ecotourism game reserve, mixed farms and traditional communal areas (consisting of small clusters of houses interspersed with grazing areas and seminatural vegetation). Our hypotheses were formulated regarding the (1) area typology, (2) vegetation structure, (3) ungulate pressure and (4) human disturbance. Using a boosted-regression-tree approach, we found considerable differences between rodent groups’ abundance and distribution, and the underlying environmental factors. The mean relative abundance of medium-sized species did not differ across the three management contexts, but small species mean relative abundance was higher in the game reserves, confirming an influence of the area typology on their abundance. Variation in rodent relative abundance was negatively correlated with human disturbance and ungulate presence. Rodent abundance seems to be influenced by environmental gradients that are directly linked to varying management priorities across land uses, meaning that these communities might not benefit uniformly by the increased amount of habitat promoted by the commercial wildlife industry.

1. IntroductionIn South Africa, agricultural intensification, and overgrazing have led to profound land use changes [1]. Historically, most landscapes were converted into livestock farms and farmlands, either as intensive, extensive, or communally managed areas [2], leading to the destruction, degradation and/or fragmentation of natural ecosystems [3]. Consequently, such habitat destruction led to declines in wildlife populations and distribution in much of South African nonprotected areas [4]. However, the establishment of national policies attributing custodial rights over wildlife to landowners, prompted a transition in the governance of natural resources from the state to privates [5]. This political option led to widespread conversion of rangelands, i.e., farmlands and livestock farms, into areas dedicated to commercial wildlife industries, such as game ranching and private game/ecotourism reserves [6]. The positive conservation outcomes of these policies for economically valuable and charismatic species [7] is believed to have an umbrella effect on other taxa, mainly through the increased coverage, representativeness and connectivity of protected/restored habitats [6,8]. However, the effect of such management approaches is unexplored for most overlooked—but functionally important—taxa, such as rodents [9]. Thus, information on the ecological responses of less-charismatic taxa is needed to better gauge the complementary conservation role of South Africa’s private land.In South Africa, game farms and private game reserves often coincide across relatively small scales, rooted in human-dominated landscapes (e.g., communal lands) [10]. These land uses have contrasting management priorities and, consequently, distinct impacts on the landscape structure and wildlife ecological patterns. In game farms, the main objective is to maximize the production of ungulates for meat or hunting, while in private game reserves the goal is to maintain charismatic species, promoting ecotourism-based activities [11]. Often, these wildlife-oriented land uses are surrounded by human-dominated areas with high levels of anthropogenic disturbance. The regional co-existence of all these land uses generates complex multi-tenured landscapes, usually divided by semi-permeable wildlife fences, influencing the biodiversity supported by each of these land uses [12].Management actions directed to charismatic or valuable species may have cascading effects on rodents, usually overlooked and handled like pests [13,14,15]. However, it is crucial to understand the effect of human-induced land-use changes on rodent spatial patterns, as well as the underlying ecological mechanisms thereof, since rodents are fundamental for some ecosystem functions [16]. Rodents are primary consumers [16] and support a large community of predators [17,18], which makes them a vital link in food-chain structuring [19]. Moreover, they are considered useful indicators of ecosystem functioning as they are valuable tools to the description and monitoring of habitat integrity. For these reasons, rodents have been used as model species to understand how land use changes affects wildlife [16]. Several factors have been identified as influential in shaping rodent community and population structures, many of which are often determined by the landscape management options [20]. Some studies have indicated that vegetation type and structure are fundamental drivers of rodent occurrence and abundance [21,22,23]. For example, areas with greater herbaceous coverage favor rodents by providing shelter against predators, food, and adequate microclimatic conditions [24]. Studies have shown negative effects of overgrazing on small mammals’ abundance, by reducing the herbaceous stratum, increasing trampling risk and feeding competition with ungulates [9,24,25,26,27]. Regarding rodent distribution, it tends to be uniform when the habitat is favorable and resources are abundant. However, when disturbances increase the level of habitat heterogeneity, causing landscape fragmentation, their distribution is mostly clumped [28,29]. Rodents are not a homogeneous group, since different species may establish distinct relationships with the environmental and biotic components of the ecosystem. For example, larger rodents’ range over larger spatial scales than smaller rodents [30] and, therefore, are more susceptible to changes at this landscape level [31].Changes in management priorities across South African multi-tenured landscapes will have a direct impact on these environmental drivers and, ultimately, in the distribution and abundance of small mammal species across and within management contexts. For instance, when management measures promote the abundance of ungulates (e.g., as prey for large carnivore populations in ecotourism reserves, or as hunting assets in game farms), grazing pressure will increase, negatively influencing the herbaceous strata [26]. Alongside with long dry and hot seasons [32], these conditions may lead to shrub encroachment, known to reduce food availability (leaves, seeds, and arthropods) for ground dwelling rodents [33]. Nevertheless, some rodent species are usually considered efficient colonizers of human shaped environments [13,34], as they are able to use human-related food resources due to their omnivore character [35].Although the processes that regulate small mammals’ spatial distribution are known for some landscapes (e.g., woodland [29] and mixed forest [36]), there is a lack of information regarding the drivers of rodent-abundance patterns in African savannas (but see [9,37]), as well as how these vary across different management schemes. Here, we evaluated the variation in rodent abundance across three adjacent management contexts, spanning a private ecotourism game reserve, mixed farms and communally owned land, managed by Zulu tribal authorities [12], under the following two main objectives: (1) to estimate heterogeneity in small-mammal-abundance distribution (mean abundance and patchiness) across management contexts (game reserve, mixed farms and communal lands); and (2) to determine the main, fine-scale environmental factors affecting small-mammal-abundance patterns across land-use types. These objectives were tested in two size-based rodent groups, for a more detailed assessment of ecological responses.Linked to these two goals, we tested four hypothetical drivers of rodent communities:(i)An area-typology hypothesis, i.e., cumulative effect of management-induced changes to vegetation, grazing pressure, etc., creates area-specific differences in rodent abundance. Patchiness will also be tested to acknowledge in which area each group is more or less clumped, regarding their abundance values. Although the exact effect of area on rodent abundance is not fully predictable [37] (given the disturbance gradient) we expected the communal lands to have the lowest values of abundance and highest patchiness (i.e., more clumped), followed by mixed farms and the game reserve, with higher abundances and lower patchiness;(ii)A vegetation-structure hypothesis, i.e., areas with higher herbaceous cover will have a positive influence on both rodent size-based groups, since it shapes the ability of the landscape to provide protection against potential predators [21,22,23,25,27,38];(iii)An ungulate-pressure hypothesis, i.e., rodent species abundance is negatively influenced by the abundance of ungulates, since higher grazing pressure tends to decrease herbaceous land cover, increase disturbance due to the trampling effect, and increase landscape fragmentation [9,24];(iv)A human-disturbance hypothesis, i.e., rodent species’ distribution is negatively influenced by human disturbance factors, such as the presence of domestic animals and households that may constrain species’ presence [14,39].2. Materials and Methods2.1. Study AreaThis study was implemented in the Maputaland–Pondoland–Albany Biodiversity Hotspot [40] in northern KwaZulu-Natal, South Africa. Our specific study area is characterized by a spatial gradient of human intervention, ranging from the Mun-ya-wana private game reserve (less subject to human associated activities), to mixed game farms and to communally managed lands, where two distinct Zulu communities are settled (Figure 1b). The Mun-ya-wana private game reserve (27°40′ S–27°55′ S; 31°12′ E–32°26′ E) represents the union of several properties without internal fences, managed by private owners whose goal is to explore eco-touristic products, therefore promoting wildlife and habitat conservation. Those management objectives are commonly related with a more sustainable use of wildlife, typically wildlife-viewing tourism [41]. The reserve is surrounded, to the South, by a mosaic of commercial game ranches for the production of wild ungulate species, occasionally mixed with domestic cattle [42] (hereafter mixed farms) and represents large expanses of natural habitat with low human density. Communal lands to the east are composed of households, interspersed with pasture areas and semi-natural vegetation. The region is characterized by a warm-temperature climate, with a humid and hot summer (October to April), according to the Köppen–Geiger classification. Mean monthly temperatures range from 19 °C in July to 31 °C in January, and the average annual precipitation is 800 mm [43,44]. Elevation ranges from 3 m to 304 m above sea level [45], dominated by a similar mixture of vegetation throughout the area (bushveld, woodland and grassland) [46] (Figure 1b). Nevertheless, the game reserve hosts a higher diversity and abundance of pristine habitats, such as indigenous forests, while mixed farms are mainly composed of pasture areas (low shrubland and grassland–Figure 1). Contrarily, communal lands have the lowest proportion of vegetation and the highest cover of urban–village occupation (Figure 1).2.2. Rodent SamplingRodents were sampled between October and November 2017 (the southern hemisphere’s spring) using ink-tracking tunnels [42], left active in the field for four consecutive nights (open circles in Figure 1c). Ink-tracking tunnels were made of robust corrugated plastic (55 × 10 × 10 cm), open on both ends to allow rodents to enter. Both entrances of the tunnel are equipped with an adhesive paper with the glue side up, and an ink pad (12 × 10 cm) was placed in the floor center [47] (Figure S1B). In the middle of the tunnel, a small PVC-pipe section, hanging from the ceiling, was installed, and contained bait composed of a mixture of peanut butter, oatmeal and sunflower oil [46]. The pipe was used to prevent the consumption of the bait by the animals entering/crossing the tunnel. The ink tunnels were placed on the ground, grouped in clusters of nine, in a Y formation, 10 m apart from each other (Figure 1c). The arms of the Y formation were 120 degrees apart (Figure 1c). This design provided an adequate spatial coverage in relation to the home-ranges of the rodent species, also ensuring some level of independence between sampling units, considering the mean distance between sites (see below). After the four-day sampling period, the plates of each ink tunnel (containing footprints and tracks) were photographed individually, always at the same distance and with a reference scale.The footprint data was used to estimate rodent relative abundance, using the proportion of the tunnels with records (track index; TI–for more details see Supplementary Materials) [48]. To ensure that this approach captured spatial heterogeneity in relative abundance, we conducted a small trial, comparing the abundance indices derived from ink tunnels to those obtained from live-trapping (see Supplementary Materials, PART A). As track identification at the species level is very time consuming and not viable in large-scale studies, and as distinguishing footprints from similar-sized species is very difficult and bias prone, we opted for dividing tracks into groups based on track size (for more details see Supplementary Materials PART A; Figures S1A and S2A, Table S1A). Rodent footprints were grouped into three different size-based groups per body length/weight, assuming a relation between rodent body length/weight and footprint sizes [49,50]: small (body length: 50–100 mm), medium (100–150 mm) and large rodents (150–200 mm) (Figure S2B). Sampling intentionally took place outside the breeding season (which peaks in the wet season, [51]), in order to avoid grouping juveniles in the wrong size-based group. However, considering the low number of detections of large rodents in ink-tracking tunnels, we only analyzed the data from small- and medium-sized rodents (see Results). The most common species captured during live trapping and linked to each group were Mus minutoides and Dendromus melanotis for small rodents, Mastomys natalensis and Saccostomus campestris for medium rodents and Otomys angoniensis and Rattus rattus for large rodents (Table S2A). 2.3. Environmental VariablesVegetation structure variables were collected using two different approaches: field measures and remote-sensed products [52]. All variables collected have been previously detected as influential to rodent presence elsewhere (e.g., vegetation cover) [21,23]. Shrub-and-grass cover were visually estimated and assigned the corresponding Edwards classification category [53] (see Table 1 for details), within a 30 m radius buffer, centered on the ink tunnel’s Y formation. Regarding the land use, the predominant categories were selected (thicket, grassland, sand forest and urban villages) and, for each buffer, was assigned the category with the highest cover. According to the type of crops present in the study area, the harvesting season occurs mostly between April and June [54], not coinciding with the study period. Therefore, we assumed that there would be no influence of crop productivity on the distribution/abundance of rodents in our study. The percentage of tree cover was assessed based on the Global Forest Watch database (Table 1). We also selected the Normalized Difference Vegetation Index (NDVI), widely used as a vegetation productivity proxy, collected from Landsat 8 Images [55].Variables of ungulate pressure and human disturbance were collected from Curveira-Santos et al. [12] camera-trap surveys. Cameras, located in the center of the Y formation, were active for 60–90 days, and attached to a tree or metal stake, 30 cm above the ground, without any bait and set to photograph at minimum delay (1 s for daytime and 30 s for night-time) (see [12] for details). Each of the defined ink-tunnel clusters (i.e., one cluster includes nine ink tunnels and one camera-trap; Figure 1c) were spaced approximately 1.4 km apart (Figure 1b). In total, were sampled 196 points: 100 points in Mun-ya-wana eco-tourism/game reserve, 50 points in mixed farms and 46 points in communal lands. Capture rates, expressed as the number of independent camera records (>1 h interval between photographs of the same species, per 100 trap-days) for livestock (cows and goats), wild ungulates and human disturbance, were used as surrogates of disturbance in the modeling procedure (Table 1). Wild ungulates were grouped according to two criteria: weight, since trampling is one of the main negative impacts of ungulates on rodents [26], and/or the fact that they are actively managed in all studied areas (Table S1B). Only ungulates weighing between 45–200 kg and actively managed were used in the analysis, since they are more abundant than other ungulates, as they are present throughout the areas under study, and because they have a greater impact on rodents, due to their weight (Table S1B). Livestock were also separated in two weight classes: i.e., goats and cows. 2.4. Data Analyses/Modelling2.4.1. Spatial Patterns of Rodent Relative Abundance Across Areas and Size-Based GroupsDifferences in mean abundance values of size-based groups (small and medium) between study areas (Mun-ya-wana game reserve, mixed farms and communal lands) were tested using GLM with 3-level area covariate and binomial error distribution. The magnitude of patchiness in each area was ascertained by spatial-point pattern analysis of count data using Lloyd’s index of patchiness [62]. A Lloyd’s index of 1 indicates a random distribution, whilst one <1 suggests uniformity and >1 patchiness.2.4.2. Influence of Environmental Variables on Rodent Relative AbundanceDue to the high number of candidate variables and to avoid multicollinearity bias, we first estimated the nonparametric Spearman’s correlation (rs) using the “psych” R package [63]. When a high correlation between two covariates was detected (rs ≥ 0.7; [62]), the variable that was less correlated with the dependent variable was excluded from the analysis [64]. The influence of all candidate variables on rodent relative abundance was tested using a boosted-regression-tree (BRT) approach, implemented with the “gbm” package [65] in R [66,67]. This modelling technique encompasses the advantages of regression trees (e.g., predictor variables can be of any type, analysis is insensitive to outliers and can accommodate missing data [68]), overcoming their low predictive capacity through the boosting algorithm [69]. The final model is a linear addition of several regression models in which the simplest term is a tree [68,70].Boosted-regression-tree models are resilient to model overfitting but, to have a better predictive performance, we defined, a priori, the model’s input parameters based on Carslaw and Taylor’s suggestions [70]. In BRT, learning rate (lr) is the shrinkage parameter that controls the contribution of each tree to the model, and tree complexity (tc) determines the number of nodes in a tree and, consequently, its size. These two parameters control the number of trees in the model, while the bag fraction (0.5) selects the proportion of data being used at each step [61,70,71]. All models were fitted to allow interactions using a ten-fold cross validation to determine the optimal number of trees for each model. The largest learning rate and the smallest tree complexity were selected to allow a minimum of 1000 trees in the BRT fitting process (see [68]). Non-informative variables were removed during the fitting process, allowing the simplification of the set of variables [68]. This simplification consisted of defining how many variables the function can test to remove, based on relative influence and total number of variables. Then, a graph was produced showing differences in the predicted deviance according to several scenarios, each one with a different number of variables removed. Next, the number of variables to eliminate was decided, and they were removed in order of minor relative influence. We defined a threshold value and only reported the interactions with relative influence values >10%. The final relative influence of each variable was calculated by averaging the number of times a covariate is used for splitting, weighted by the squared improvement to the model as the result of each split. It is then scaled, such that the values sum to 100 [72]. Fitted values were plotted in relation to the most important predictors, revealing their effects on rodent abundance. Explained deviance was calculated using the following formula from Abeare (2009) [73] D2=1−(residual deviancetotal deviance)The 95% confidence intervals of each variable were estimated for the fitted function by taking 500 bootstrap samples of the input data, with the same size as the original data. A BRT was fitted to each sample, and the 5th and 95th percentiles were calculated for the points of each function. Models were built separately for small- and medium-sized rodents. For each model performed, interactions between typology and the other influential independent variables (i.e., relative importance above >10%) were estimated, to evaluate context-dependency in the influence in the effect environmental variable associated with the management context. All analyses were implemented in R via R Studio Version 1.1.463 [66,67].3. Results3.1. Spatial Patterns of Rodent Abundance Across Areas and Size-Based GroupsFrom the 192 sampling points monitored, 85% presented small rodent tracks, while 76% detected the occurrence of medium rodents, with an overlap in 35% of sites and inter-area variation in detection (i.e., number of tunnels with signs/total number of tunnels, Table S2B). Mean abundance in Mun-ya-wana game reserve was 0.52 ± 0.26 (mean ± SD) for small rodents and 0.43 ± 0.34 for medium rodents; in mixed farms, 0.31 ± 0.21 for small rodents and 0.52 ± 0.32 for medium rodents; and in communal lands was 0.26 ± 0.23 for small rodents and 0.36 ± 0.24 for medium rodents (Figure 2). Regarding the GLM result for size-based groups, it revealed significant differences in relative abundances only for small rodents, between Mun-ya-wana game reserve and the remaining areas (Table S3B, Supplementary Materials). No significant differences were detected in relative abundances of medium rodents between areas (Figure 3). Between groups, significant differences were only found in mixed farms (Table S3B, Supplementary Materials), with medium rodents being more abundant (0.52 ± 0.37) than small-size rodents (0.31 ± 0.26) (Figure 3). Based on these results, the effect of environmental drivers on rodent abundance was evaluated separately for each of the size-based groups. Rodent PatchinessLloyd’s Index of Patchiness revealed that for every area and size-based group, all abundance values were aggregated (γ > 1; Table 2). Both medium and small rodents are heterogeneously distributed within the three study areas (Figure 3), demonstrating a heterogeneity gradient. According to Table 2, we can observe that the highest values for small rodents are in communal lands, followed by mixed farms and finally, the game reserve. For medium rodents, there is a greater clustering pattern in the game reserve, followed by communal lands and mixed farms. With these results, it is possible to state that the abundance patterns differ between the size-based groups, and within each area.3.2. Drivers of AbundanceCapture rate of goats and cows were both correlated with human presence (p = 0.75; p = 0.76, respectively), and intercorrelated (p = 0.79). Therefore, both former variables were removed from the analysis. 3.2.1. Small-Size RodentsThe predictive deviance for the BRT model produced for small rodents was 38.8%. After the simplification of the model, and consequent removal of two variables, predictive deviance increased to 50%, indicating that the final model explained an important part of the total variability [68]. Distance to houses, wild ungulates, human presence, NDVI, grass cover and area were identified as the most influential drivers of small rodent abundance (Figure 4). Small rodents were more abundant in areas far from human settlements, with lower abundances of wild ungulates and low presence of humans. Regarding the NDVI, values between 0.29 and 0.35 affect positively the abundance of small rodents. Semi-open grass cover had the most positive effect on small rodent abundance, as well as the Mun-ya-wana ecotourism/game reserve. Interactions with area typology within this model were found for wild ungulates (0.20, interaction size) and NDVI (0.34). As it is possible to see, in the Figure 5, that the most evident and distinct responses for both variables occur in Mun-ya-wana game reserve, revealing a clear influence of this area on wild ungulates and NDVI.3.2.2. Medium-Size RodentsFor this rodent group, the initial predictive deviance of the model was 40.6%, but, after the removal of one variable during the model simplification, the predictive deviance increased to 50%. The set of variables identified as important for this group was very similar to that described for the previous rodent groups (Figure 4). Medium-size rodents’ abundance was also higher in areas with low abundance of wild ungulates, human presence and which were far from human settlements. However, this group seems to thrive in more continuous grass cover and it is positively affected by low values of NDVI (0–0.18). Contrarily to the small-rodents group, this model did not include the area variable, which may indicate a lower relevance of area typology in shaping the abundance patterns of these rodents. 4. DiscussionRodent abundance, although often an unheeded aspect of conservation management, is crucial to understand ecosystem functioning, since rodents are primary consumers [16] and support a large community of predators [17,18], making them a vital link in food-chain structuring [19]. In our study area, spatial heterogeneity in rodent-abundance patterns appears to be influenced by environmental gradients that are directly linked to varying management priorities across land uses (e.g., ungulate pressure associated with wild game), which means that these rodent communities, and groups within these communities, might not benefit uniformly from the increased amount of habitat promoted by the commercial wildlife industry.4.1. Context-Specific Responses and Variation Across Management SchemesArea typology was an important abundance driver for small rodents (thus, just partially supporting our first hypothesis-H1), with higher abundances being estimated for Mun-ya-wana game reserve than for the remaining areas. Medium-size rodents did not show any significant differences in their abundance between areas (Figure 3).The difference in small rodent abundance between areas (Figure 3) is supported by the interactions of the NDVI and wild ungulates abundance with the area typology (Figure 5). Overall, small rodent abundance decreased with an increase in wild ungulate abundance, irrespective of the management scheme, as predicted (H3: ungulate pressure hypothesis). Similarly, small rodent abundance increased with an increase in NDVI. However, the game reserve displayed a higher small rodent abundance, relative to the other land uses, and there is a differential effect of wild ungulates and NDVI on abundance between areas. Within the game reserve, these variables have a greater influence on this group probably due to the applied management practices. The greater variation in small rodent abundance in response to variation in wild ungulate abundance in Mun-wa-wana game reserve may be driven by the greater vegetation spatial heterogeneity of this area. The game reserve has a greater habitat heterogeneity compared to the other study areas due to better conservation derived from its protection status. This habitat heterogeneity results in a heterogeneous distribution of wild ungulates, owing to differences in habitat preference or selection (e.g., [74]). Thus, this wider variation of ungulate abundance across the reserve induces a more pronounced response in rodents, leading to the detected typology effect. Regarding the NDVI, the response may be influenced by the same factor (better conservation status of native forests-sand forests), which assure a lower disturbance regime, and thus create conditions to support a more abundant rodent community. However, the conservation character of some environments may induce the opposite trend in other taxa. Studies that analyzed the influence of protected areas in the conservation of small mammals found that these areas exhibit lower abundances compared to neighboring areas, since their conservation aims is mostly focused on wild ungulates and predators [37]. This induces small mammals’ movement to nearby areas, such as farms and agriculture lands, where they can find more resources (e.g., food) [9], and sometimes lower predation pressure. A study conducted in the same studied game reserve, based on live trapping measures, revealed a higher abundance of small mammals in adjacent farms and former cattle farms [9]. This pattern seems to be corroborated by our study data, but only for medium rodents that are less abundant in the more protected area (i.e., Mun-ya-wana game reserve). Small rodents respond differently, and the pattern may be associated with the environmental conditions provided by the game reserve, that seem to promote this group abundance. As mentioned above, the game reserve has a greater habitat heterogeneity derived from its protection status. This allows the conservation of certain vegetation patches that do not thrive in the other two areas. In this case, the NDVI values that promote a higher abundance of small rodents (between 0.28–0.35, Figure 4) correspond to native forest that exist in greater coverage in the game reserve (i.e., sand forests, Figure 1). Despite a greater abundance of wild ungulates and possible predators, the presence of these native habitats establishes more favorable conditions for small rodents. Considering that these rodents use the landscape on a smaller scale due to their size [30], these minor patches of vegetation create a significant difference in the abundance of this group.Rodent abundances vary not only between areas (linked to areas specificities, and small mammals’ requirements), but also show an inter-group variation within areas. The spatial variation of abundances within-areas seems to be linked to the type of management implemented in each area that affects the vegetation structure and thus may have important implication in species conservation [9]. Lloyd’s Index supports that aggregation levels differ between size-based groups, since rodents preferentially aggregate in different areas (medium-size rodents in mixed farms and small rodents in Mun-ya-wana game reserve), which supports an allopatric distribution of both rodent groups. Furthermore, the highest abundances of each size-based group occurred in distinct areas (small in Mun-ya-wana, medium in mixed farms). Places where rodents occur in a more regular pattern, usually have better conditions (i.e., higher, and more regularly distributed resources), while sites where rodent distribution is more aggregated/clustered, indicate a more heterogeneous distribution of resources [75]. Our results show that the lowest values of Lloyd’s Index i.e., less patchy distribution, match the highest abundance values for both groups. This pattern is verified for small rodents in the game reserve and medium rodents in mixed farms (Table 2). Area typology influences the patchiness, since conditions will be more or less suitable for rodents according to the type of management applied (e.g., reserve and communal lands; [76]). A greater patchiness may lead to isolated populations, causing more sensitive species to disappear [75]. Thus, it is crucial to determine which type of management best promotes rodent abundance.4.2. Fine-Scale Environmental Drivers of Rodent Abundance Across the LandscapeOur data also reveals that the abundance of both rodents groups is overall promoted by grass cover, which supports our second hypothesis (H2). However, the type of grass cover that enhances rodent abundance varies between groups. While medium-size rodents reached higher densities in continuous grass cover, small rodents are more abundant in semi-open grass cover. Grass cover, especially continuous layers, can provide protection against potential predators [25,27], reducing predation risk, and therefore allowing medium-size rodents to reach higher abundances. The different results might be associated to habitat preferences. Small rodents occurred predominantly in forested savanna areas (ex. Mun-ya-wana game reserve center area), while medium rodents occurred predominantly in open savanna areas (ex. north and south areas of the game reserve–see Figure 1 and Figure 2). The continuous grass cover patches may be more important in these open areas, since they provide an efficient protection against predators [24]. In forested regions (where small-size rodents seem to be more abundant), grass cover may be less important compared to its potential cover under better conservation of native forests, which guarantees a greater diversity of microhabitats and assures a lower disturbance regime, thus creating conditions to support a more abundant rodent community. The presence of ungulates (wild or domestic) has been associated with a reduction of habitat quality for rodents, by decreasing the availability of food and shelter for these small mammals [26,37]. This general pattern is reflected in our results, corroborating our third hypothesis (H3), i.e., species abundance is negatively influenced by the abundance of ungulates. This negative impact of ungulates may be linked to their impact on vegetation [26], since higher grazing pressure tend to decrease herbaceous land cover [9,24]. A study conducted in central Kenya showed an increase in small mammals’ abundance in the absence of ungulates, revealing the existence of food competition between ungulates and African rodents [77]. Although being omnivores, rodents feed mainly on seeds and grasses [78], which are highly depleted when ungulates are present. Furthermore, the ungulates trampling impacts on small mammals are also a possible explanation for this negative influence, since the soil compaction due to ungulates movements hampers burrows maintenance [26,79]. Other studies highlighted the impact of a reduction of the herbaceous layer, as it decreases refuge availability and increases predation risk by improving small mammals’ detection by predators [31,80,81,82]. Therefore, these two-fold effects (decrease in food and shelter availability), acting in isolation or in synergy, may be the underlying processes that constrain rodent abundance in the presence of ungulates. The distance to human settlements and human presence are also two factors that we identified as having a negative effect on both rodent groups’ abundance, which corroborates our fourth hypothesis (H4). Rodents revealed lower abundances in areas closer to houses, especially in communal lands, the area with the highest density of settlements (while houses are almost absent from the other two areas). Thus, the effect of this variable cannot be linearly interpreted as a distance to the nearest house, but probably as a distance to the communal lands themselves, as both groups’ abundances are low in this area (see Figure 2). The average abundance values confirm that the least preferred zone for both groups of rodents are the communal lands, as it is the place where the lowest values of abundance were estimated (Figure 3). However, these negative effects of anthropic disturbance may also be linked to the presence of domestic animals (livestock and goats), that occur concomitantly with settlements, and that also negatively affect rodent abundance, due to the same processes described above for wild ungulates [26].This different patterns between rodent groups, as well as the variation of the drivers and their importance on the abundance variation of both species, supports the division of our dataset into size-based groups. This means that not only rodents should be taken into consideration, but also heterogeneity within rodent communities, which is important given their different functional roles (e.g., as prey, as consumers–granivory and insectivory–and seed dispersers). Although we acknowledge some limitations of this approach, based on footprint size, we have tried to minimize this by sampling only in seasons where the misclassification effect of juveniles’ presence is negligible. Nevertheless, this time-limited sampling hampers the validity of extrapolating results. Interpretation of the overall (annual) pattern of abundances’ spatial distribution must be done with care. Rodents numerically respond to variations in rainfall and food availability, which vary throughout the year. Thus, by sampling in only one season, we may have gotten a partial image of the processes shaping rodent abundance. However, in terms of wildlife management and conservation, it is always better to have a partial understanding of the ecological patterns and processes than having none.5. ConclusionsOur study contributes to the current view that landscape-management options shape the ecological patterns of species, by modifying the composition and structure of habitats. Moreover, responses to land composition are species/group-specific. These results highlight the need to expand conservation actions beyond protected areas. For biodiversity conservation to succeed in these habitat mosaics, landscape-level policies and management are required to integrate both protected and managed areas, as the later also host a large number of species, acting as a metapopulation source-sink. We encourage future work that evaluates the transferability of our findings to other southern African multi-use landscapes.

animals : an open access journal from mdpi

10.3390/ani12050574

PMC8909077

During the weaning period, the change of feed and separation from ewe induce weaning stress and may affect the growth and health of kids. The application of antibiotics could relieve weaning stress; however, their prophylactic application coerces researchers to find antibiotic alternatives to relieve weaning stress. Fucoidan is a natural plant extract widely used in animal production with antioxidant and immune-modulatory properties resulting in beneficial effects on the intestinal tract. In the present study, fucoidan dietary supplementation boosted antioxidant and immune functions, improved the morphology of the intestinal tract and promoted the growth performance of kids. These results indicated that fucoidan could be used to alleviate weaning stress in kids.

The purpose of this study was to evaluate the effects of fucoidan dietary supplementation on growth performance, organs’ relative weight, serum anti-oxidation markers, immune function indices and intestinal morphology in weaned kids. A total of 60 2-month-old weaned castrated male kids (Chuanzhong black goat) were used for this 30-day experiment and randomly allocated to four groups. The control group (CON) fed a basal diet, while the other three groups were provided with the same diet further supplemented with fucoidan at 0.1%, 0.3% or 0.5%, namely, F1, F2 and F3 groups, respectively. The results indicated that dietary fucoidan supplementation significantly increased (p < 0.05) the activity of catalase (CAT) when compared to the CON group on day 15. Moreover, the addition of fucoidan at 0.3% and 0.5% significantly increased (p < 0.05) the activities of glutathione peroxidase (GSH-Px) and total superoxide dismutase (T-SOD). On day 30, dietary fucoidan supplementation significantly reduced (p < 0.05) the feed conversion rate (FCR), contents of tumor necrosis-α (TNF-α), interleukin-1β (IL-1β) and interleukin-6 (IL-6), while it significantly increased (p < 0.05) the activity of total superoxide dismutase (T-SOD), the content of immunoglobulin G (IgG) and the villus height (VH) of the duodenum. Moreover, dietary 0.3% and 0.5% fucoidan supplementation significantly increased (p < 0.05) the villus height (VH) of the jejunum and ileum and significantly reduced (p < 0.05) the crypt depth (CD) of ileum. In conclusion, dietary fucoidan had positive effects on growth performance, serum anti-oxidation, immune function and intestinal morphology of weaned kids.

1. IntroductionIn modern small ruminants’ production, weaning stress is an inevitable problem observed in small animals. Weaning stress had negative effects on antioxidant capacity, immunity, intestinal morphology and growth performance [1,2,3]. Formerly, antibiotics were widely used to alleviate weaning stress [4]. However, the ban of antibiotic use in feed worldwide forces researchers to find antibiotic alternatives. Fucoidan, a kind of macromolecular polysaccharide rich in sulfate, occurs in the cell walls of brown algae, mucous matrix and some marine invertebrates [5,6]. Fucoidan has been proven to possess many biological properties, including immunomodulatory, antioxidant and antibacterial properties, resulting in the promotion of animal growth [7,8,9,10]. Thus, fucoidan was widely used in functional foods and animal production. Dietary fucoidan administration increased feed intake, daily gain and feed efficiency and also had beneficial effects on intestinal morphology, antioxidant capacity and immune function in weaning pigs [11,12,13]. Similar results are also reported in chickens and fish [14,15,16]. In summary, fucoidan could be used as an environmentally friendly substitute for antibiotics in diets to improve growth in pigs, chicken and fish. However, data on the application of fucoidan in small ruminants are scarce. Therefore, the purpose of this study was to evaluate the effects of fucoidan on growth performance, organs’ relative weight, antioxidant markers, immunity indices and intestine morphology in weaned kids.2. Materials and Methods2.1. Animals, Diet and Experimental DesignThe experimental protocol applied in this study followed the guidelines of the Animal Care and Use Committee of Guangdong Ocean University.The fucoidan used in this study was provided by a company (Mingyue Hailin Fucoidan Biotechnology Co., Ltd., Qingdao, China). Fucoidan had the form of a powder with a yellow color and a smell of seaweed. The purity of fucoidan was 98%, and the sulfate ion content was 28.9%.A total of 60 two-month-old weaned castrated male kids (Chuanzhong black goat) with an average initial body weight of 12.5 ± 0.5 kg were used in this 30-day experiment. Kids were weaned at 60 days and randomly allocated to 4 treatments with 15 replications. The control group (CON) was fed with a basal diet, while the other three treatment groups were fed with the same diet further supplemented with fucoidan at 0.1% (F1 group), 0.3% (F2 group) and 0.5% (F3 group). Three kids were placed in a pen (each pen = 3.1 m × 2.5 m × 1 m). The basal diet (Table 1) was formulated to meet or exceed the nutrient requirement of the Feeding standard of Goat, China (NY/T 861-2004).Before the trial started, the sheep house was cleaned and sterilized. A 7-day pre-trial was conducted first, during which vaccination, deworming and numbering were performed. The roughage consisted of silage and Aneurolepidium Chinese hay. The kids were fed twice a day at 8:30 am and 17:30 pm, with access to clean drinking water available ad libitum. Fucoidan was manually mixed into the concentrate. The kids were fed concentrate first and then roughage.2.2. Growth PerformanceBody weight was determined on day 1, 15 and 30, the feed intake was recorded daily. The average daily gain (ADG), average daily feed intake (ADFI) and feed conversion rate (FCR) were also calculated.2.3. Sample Collection and Organs’ Relative WeightOn days 15 and 30, blood samples were collected from the jugular vein, and then they were centrifuged at 3500 g for 10 min (4 °C). The serum was collected and stored at −20 °C for later analysis.Kids were fasted for 12 h prior to slaughter at the end of the trial. Six kids of similar weight per group were selected for slaughter. About 2.5 cm segment of the duodenum, jejunum and ileum were trimmed and used for morphological indices. Samples of tissue were washed with PBS, then fixed in paraformaldehyde for histological evaluations. Finally, organs were weighed.The organ relative weight was calculated by the following formula: organ index (%) = organ weight/body weight × 100%.2.4. Serum AntioxidantThe activities of total antioxidant capacity (T-AOC), total superoxide dismutase (T-SOD), catalase (CAT), glutathione peroxidase (GSH-Px) and malondialdehyde (MDA) content were measured using commercial kits according to manufacturer’s guidelines (Nanjing Jiancheng Bioengineering Institute, Jiangsu, China).2.5. Serum ImmunityThe contents of immunoglobulin G (IgG), interleukin-1β (IL-1β), interleukin-6 (IL-6), interleukin-2 (IL-2), interleukin-10 (IL-10) and tumor necrosis factor-α (TNF-α) were estimated by enzyme-linked immunosorbent assay kits according to manufacturer’s guidelines (Jiangsu Meimian industrial Co., Ltd., Jiangsu, China).2.6. Intestinal HistomorphologyThe samples from the duodenum, jejunum and ileum were fixed in paraformaldehyde for 24 h at room temperature and subsequently dehydrated through a graded ethanol series, cleared with xylene and embedded in paraffin. Then, tissues were cut into 5 μm-thick continuous sections. Finally, the sections were stained with hematoxylin for 2 min and eosin for 40 s, and then dehydrated and mounted on slides. The morphological parameters were measured by Image Pro Plus 6.0 software (Media Cybernetics, Silver Spring, MD). The morphological parameters of the intestinal tract included villus height (VH), crypt depth (CD) and radio of villus height to crypt depth (VCR).2.7. Statistical AnalysisAll statistical analyses were performed using SPSS 26.0 via one-way ANOVA, and differences were detected by Ducan’s multi-range test. The results are expressed as mean ± standard error of the mean (SEM), and differences are considered significant at p < 0.05.3. Results3.1. Grouth PerformanceAs shown in Table 2, kids fed F2 and F3 diets had a higher final body weight than those fed the CON and F1 diets. Kids fed diets with fucoidan significantly reduced (p < 0.05) FCR compared to those fed the CON diet during days 16 to 30 and the overall period. Kids fed F2 and F3 diets had significantly higher (p < 0.05) values for ADG and ADFI than those fed the CON diet during days 16 to 30 and the overall period. No significant differences were observed for growth performance among treatments during days 1 to 15.3.2. Organs’ Relative WeightAs shown in Table 3, no significant differences were observed for organs’ relative weight among the experimental groups.3.3. Serum Antioxidant CapacityAs shown in Figure 1, on day 15, kids fed diets with fucoidan had a significantly increased CAT activity (p < 0.05). Lambs fed F2 and F3 diets showed significantly increased (p < 0.05) activity of GSH-Px and T-SOD than those kids fed CON and F1 diets. No significant differences were observed for MDA and T-AOC among treatments.On day 30, kids fed diets with fucoidan showed significantly increased (p < 0.05) activity of T-SOD. Kids fed F2 and F3 diets had a significantly reduced (p < 0.05) MDA content than those fed CON and F1 diets. Kids fed F2 diet had a significantly increased (p < 0.05) activity of CAT than those kids fed CON, F1 and F2 diets. No significant differences were observed for GSH-Px and T-AOC content among treatments.3.4. Serum Immuntiy IndicesAs shown in Figure 2, kids fed diets with fucoidan had significantly higher IgG content than that fed with CON diet on day 30 (p < 0.05). Kids fed diets with fucoidan had significantly lower TNF-α, IL-1β and IL-6 contents than those fed CON diet (p < 0.05). On the other hand, F2 and F3 animals had significantly higher IL-2 and IL-10 contents than those fed CON and F1 diets (p < 0.05).3.5. Intestinal MorphologyThe effects of dietary fucoidan administration on intestinal morphology are illustrated in Figure 3. From the HE staining, we could see that the intestinal villi were denser and longer than the CON in the duodenum, jejunum and ileum. As shown in Table 4, kids fed with fucoidan-supplemented diets had significantly higher VH in the duodenum on day 30 (p < 0.05). Moreover, kids fed with F2 and F3 diets significantly increased (p < 0.05) the VCR in the duodenum and ileum than those fed with CON and F1 diets. Kids fed with F2 and F3 diets had significantly lower (p < 0.05) CD values in the ileum than those fed with CON and F1 diets. Kids fed with F2 and F3 diets significantly increased (p < 0.05) the VH in the ileum than those fed with CON and F1 diets. However, no significant differences were observed for CD in the duodenum and CD or VCR in the jejunum among treatments.4. Discussion4.1. Growth PerformanceIn general, weaning stress adversely affects the growth performance of kids by reducing feed intake, feed efficiency, immune suppression and increasing intestinal damage [17,18]. In our study, dietary fucoidan administration had no significant effects on growth performance among treatments during days 1 to 15. In support of our results, Rattigan et al. [19] reported that dietary fucoidan supplementation had no significant effects on feed intake and daily gain during days 1 to 14; however, supplemented weaner pigs showed higher feed intake and ADG than those in the CON group during days 16 to 30 and 1 to 30. Similarly, Draper et al. [20] indicated that dietary fucoidan administration increased feed intake and decreased FCR in weaning pigs. We speculated that the improved effects of fucoidan on the growth performance in this study might be related to the improvement of feed intake, antioxidant capacity, immune function and intestinal structure.4.2. Serum AntioxidantWeaning can lead to the excessive production of oxygen-free radicals, which results in oxidative stress and is linked with reduced growth performance [21,22]. SOD, GSH-Px and CAT activities are the first lines against oxidative injury, and MDA was the final product of lipid peroxidation [23,24]. In this study, dietary fucoidan administration increased the activities of GSH-Px, SOD and CAT on day 15 and decreased the content of MDA on day 30. Similarly, Yang et al. [25] reported that dietary supplementation with 0.1% fucoidan increased the activities of CAT and SOD and decreased the content of MDA in Pelteobagrus fulvidraco. Zhang et al. [26] indicated that supplementation with fucoidan improved the activities of SOD, CAT and GSH and reduced the content of MDA in drosophila geriatric. In summary, the improvement of antioxidant capacity with dietary fucoidan supplementation might be related to the increased activities of antioxidant enzymes and the reduced lipid oxidation values.4.3. Serum ImmunityPrevious research confirmed that weaning reduced immunity function, making invasion by external pathogenic microorganisms easier [27]. Immunoglobulin and cytokines are an important part of the immune system [28]. The immunoglobulins protect the body by removing pathogenic microorganisms and harmful molecules [29]. Cytokines play a key role in inflammation and anti-inflammatory processes [30]. In this study, dietary fucoidan administration increased the content of IgG, IL-2 and IL-10 in serum while reducing the content of IL-1β, TNF-α and IL-6. In agreement with our study, Lean et al. [31] reported that dietary fucoidan administration reduced the content of IL-1β in colons in mice with colitis. Tomori et al. [32] indicated that fucoidan increased the contents of IgG and IL-2 in serum by promoting the proliferation of immune cells and reducing the contents of IL-4 and IL-5 in spleens of mice. Moreover, Delma et al. [33] indicated that fucoidan exerted anti-inflammatory activity by regulating the NF-κB signaling pathway and reducing the expression of p53. Taken together, the improvement of immunological functions with dietary fucoidan supplementation may be related to the regulation of inflammatory-related factors release.4.4. Intestinal MorphologyAt weaning, the reduced villus height and increased crypt depth are related to stress [34,35]. In this study, dietary fucoidan administration increased the villus height of the duodenum, jejunum and ileum and decreased the crypt depth of the ileum. Our results are in line with that of Leonard et al. [36], who reported that dietary fucoidan administration in weaning pigs increases villus height and VCR of jejunum. Similarly, Walsh et al. [12] indicated that dietary fucoidan supplementation increased the villus height of small intestinal in post-weaning pigs. According to current studies, it indicated that dietary fucoidan administration could improve the intestinal morphology of the small intestine.5. ConclusionsIn summary, this study showed that fucoidan dietary supplementation improved the feed intake, daily gain, antioxidant capacity, immune status and intestinal morphology in weaned kids.

animals : an open access journal from mdpi

10.3390/ani12030324

PMC8833476

Approximately 20% of older horses develop pituitary dysfunction (PPID), which is associated with haircoat changes, muscle loss, and a higher risk of developing an infection or laminitis. Elevated plasma adrenocorticotropic hormone (ACTH) is used to diagnose PPID; however, ACTH is not stable in blood samples. Therefore, samples should be kept at 4 °C until analysis. In ambulatory veterinary practice, blood samples can be left at room temperature (20 or 30 °C) or inadvertently left in a vehicle without refrigeration where they might be exposed to temperatures of up to 70 °C in hot climates. To evaluate the effects of temperature on ACTH concentrations, we experimentally subjected blood samples from horses with and without pituitary dysfunction to temperatures of 4 (reference), 20, 30, and 70 °C for 1 h prior to laboratory measurement. The stability of ACTH was affected by short-term exposure to high temperatures in horses with and without pituitary dysfunction with both higher and lower ACTH concentrations measured unpredictably. Our results suggest that samples should be kept at 4 °C to reflect the true ACTH concentration. Exposure to temperatures of up to 40 °C for 1 h can still provide an appropriate assessment of pituitary function in most cases, but the ACTH concentration changed by 12% in healthy horses and 5% in horses with PPID.

Pituitary pars intermedia dysfunction (PPID) is diagnosed by increased basal or post thyrotropin-releasing hormone (TRH) stimulation ACTH concentrations. ACTH is known to be unstable; however, the effect of different temperatures and TRH stimulation on equine ACTH stability is poorly described. In total, 15 horses, including 8 PPID positive (ACTH > 35 pg/mL at baseline or >65 pg/mL 30 min after TRH stimulation), were divided into 2 groups: 9, including 5 PPID positive, with basal ACTH concentrations and 6, including 3 PPID positive, with post TRH stimulation ACTH concentrations. Whole blood was stored for 1 h at 4, 20, 30, 40, or 70 °C. After centrifugation, immunoreactive ACTH concentrations were determined using a chemiluminescent assay. Linear mixed effect models were used to detect the effects of temperature, PPID status, and TRH stimulation on the immunoreactive ACTH concentration. Temperature had a significant effect (p = 0.003) on immunoreactive ACTH concentrations, and this effect was greater in PPID-negative horses (p = 0.01), with the changes in immunoreactive ACTH concentrations being slightly unpredictably higher or lower than samples stored at 4 °C. Even at 20 °C, mean immunoreactive ACTH concentrations minimally changed by 5% in PPID horses and 12% in non-PPID horses after 1 h. No significant effect of TRH stimulation was identified. Although ACTH concentrations should ideally be determined from samples kept at 4 °C, samples inadvertently left at temperatures of up to 40 °C can provide valid results if analyzed within 1 h; however, this increases the risks of altered ACTH concentrations, occasionally influencing the diagnosis of PPID.

1. IntroductionPituitary pars intermedia dysfunction (PPID) is a common disease of older horses and ponies, affecting 21% of horses over 15 years of age [1,2]. Up to 73% of PPID-positive horses are euthanized due to complications associated with PPID and only 50% of horses are alive 4.5 years after diagnosis [3].Histologic examination of the pars intermedia is considered the gold standard to diagnose PPID; however, there is only fair agreement between pathologists, and histopathology can only be performed postmortem [4,5]. Recommended antemortem diagnostic tests include baseline ACTH concentration and ACTH response to thyrotropin-releasing hormone (TRH) stimulation, which increases diagnostic test sensitivity to diagnose early or subclinical cases of PPID [6,7,8,9,10].Studies found that human ACTH is highly unstable due to proteolytic degradation [11,12]. Preanalytical stability of human ACTH depends on both the time to centrifugation and temperature, with the measured ACTH concentration being altered if blood samples are left at room temperature for ≥4 h prior to centrifugation. In order to obtain accurate values of plasma ACTH concentrations in humans, it has been recommended that if samples cannot be transported to a laboratory for analysis within 2 h (at room temperature), they should be kept at 4 °C and analyzed within 8 h [13]. The effects on the measured ACTH concentration of delaying centrifugation of horse blood by even 1 h resulted in a mean reduction in the measured ACTH concentration by 11.6 pg/mL in one study [14]. The stability of equine ACTH has been investigated at room temperature (20 or 22 °C) and with archived samples (−20 or −80 °C), but as far as the investigators are aware, there have been no studies investigating the stability of ACTH when exposed to high environmental temperatures [15,16]. Some countries or regions are susceptible to high environmental temperatures and samples obtained by equine ambulatory veterinarians might be subjected to these conditions if not kept in adequate cooling devices. In some areas, environmental temperatures might occasionally exceed 40 °C and temperatures inside vehicles might exceed 70 °C [17,18]. It is also uncertain if ACTH post-TRH stimulation has the same stability as basal ACTH. Previous research could not detect an effect of TRH stimulation on the stability of equine immunoreactive ACTH after up to five freeze/thaw cycles, suggesting that ACTH post-TRH stimulation would have a similar stability to basal ACTH; however, further investigation is warranted to determine if there is an effect of TRH stimulation on the stability of ACTH after exposure to high temperatures [18].If inappropriate sample handling conditions affect the measured ACTH concentration, then false positive and negative diagnoses of PPID might occur. This could have animal welfare implications if the ACTH concentration is reduced, and a false negative result is obtained [6]. In contrast, a false positive result could lead to unnecessary medication and might even prevent animals from competing in high-level competitions where pergolide is a banned substance (FEI Equine Prohibited Substances List, January 2020).Therefore, the aim of this study was to investigate the short-term effect of a range of temperatures, and the effect of TRH stimulation, on ACTH stability in horses with and without PPID, and the effect on the subsequent diagnosis of PPID.2. Materials and Methods2.1. HorsesIn total, 15 institution-owned mature horses (median 16 years of age, range 11 to 27 years) were enrolled. There were 6 geldings and 9 mares of various breeds: Australian Stock Horse (n = 6), Standardbred (n = 5), Warmblood (n = 2), Arabian (n = 1), and Quarter Horse (n = 1). Horses were kept on the same pasture. In total, 8 horses were diagnosed with PPID in mid-summer (December and January) by either a baseline ACTH concentration over 35 pg/mL or ACTH concentration over 65 pg/mL 30 min post-TRH stimulation [8,9,10,19]. All horses assigned to the PPID-positive group either had clinical signs or developed them during the following 12 months. Beyond clinical signs of PPID for some horses, all were healthy based on physical examination [20]. No horses were receiving any medication including pergolide.Horses were randomly divided (coin toss) into two groups. Baseline ACTH was determined in 6 horses, of which 3 were PPID positive. The second group included 9 horses, of which 5 were PPID positive, and ACTH concentrations were determined 30 min post-TRH stimulation.2.2. Sample ProcessingFor horses in which the basal ACTH concentration was determined, 50 mL of blood were collected by jugular venipuncture into 5 ethylenediaminetetraacetic acid (EDTA) plastic blood collection tubes from each horse. For horses that underwent TRH stimulation, blood was collected 30 min after receiving 1 mg of TRH intravenously (Sigma-Aldrich Pty Ltd. (subsidiary of Merck), North Ryde BC, New South Wales, Australia). Blood was sampled between 7 and 8 am. All procedures were approved by the University of Queensland Animal Ethics Committee, approval number SVS/474/17, 22nd December 2017.Samples were taken to the laboratory within 1 h of collection. They were then stored as whole blood either in a temperature-monitored refrigerator (4 °C), in a temperature-controlled room (20 °C), or in incubators (30, 40, or 70 °C) for 1 h. Samples were then centrifuged at 3500× g at 4 °C for 10 min. The plasma was separated and analyzed within 2 h of collection from the horse. The plasma immunoreactive ACTH concentration was measured by a chemiluminescent assay (Immulite 1000 Chemiluminescent Assay, Siemens. Bayswater, VIC 3153, Australia). The intra-assay coefficient of variation used for this assay was 4.8% [18]. Samples were analyzed using kits from the same ACTH lot numbers that had been purchased and transported to our laboratory as a single acquisition. Due to the necessity of analyzing all samples within a 2-h window from collection, and limitations on the speed of the analyzer, the 15 horses were sampled over a 4-day period. The samples from each horse were run using the same kit.2.3. Data AnalysisTo satisfy distribution normality as tested with the Kolmogorov–Smirnov test, the changes in the immunoreactive ACTH concentrations in each sample were calculated utilizing both the percentage of the reference sample (ACTH concentration in sample/ACTH concentration in reference sample × 100) and the absolute percentage change from the reference sample (absolute value of (1-(ACTH concentration in sample/ACTH concentration in reference sample)) × 100). The reference sample was defined as the sample stored at 4 °C. The percentage of the reference sample (presented as mean ± standard deviation) was the percent deviation above or below the ACTH concentration of the reference sample (recorded as 100%) while the absolute percentage change from the reference sample (presented as mean ± standard deviation) was used to show an absolute value that deviated away from the ACTH concentration of the reference sample (recorded as 0%).The effects of temperature, TRH stimulation, and PPID status were then analyzed with a linear mixed effect model using “temperature”, “TRH stimulation”, and “PPID status” as fixed effects and “horse” as a random effect to account for repeated measures. Then, concentrations of immunoreactive ACTH were compared to baseline (expressed as either 100% or 0%) using a repeated measures ANOVA and Tukey’s multiple comparison’s test. The statistical analysis was performed by commercially available software (IBM® SPSS® Statistics Version 25) and p < 0.05 was considered to be significant.To determine the effects of short-term high-temperature exposure on a diagnosis of PPID, a qualitative analysis was performed for each sample, with a cut-off value of >35 pg/mL in the horses in which the basal ACTH concentration was determined, or >65 pg/mL in the horses in which the post-TRH stimulation test ACTH concentration was determined. A false positive and false negative rate was then determined for each temperature.Bland–Altman plots were used to determine the bias (95% limits of agreement (LOA)) and visualize the difference between ACTH concentrations in samples kept at 4 °C versus 20, 30, 40, and 70 °C for 1 h.3. ResultsThere was a difference in the effects of short-term storage temperature (p < 0.0001) on immunoreactive ACTH concentrations, with a difference between storage at 4 °C and storage at 70 °C (p = 0.0003; Figure 1). Bland–Altman plots (Figure 2A–D) demonstrated mean bias of −2.8 pg/mL [−17.04–11.46 pg/mL] for 4 versus 20 °C, −1.38 pg/mL (−10.0 to 7.0 pg/mL) for 4 versus 30 °C, −0.16 pg/mL (−9.3 to 9.0 pg/mL) for 4 versus 40 °C, and 10.1 pg/mL (−1.5 to 21.7 pg/mL) for 4 versus 70 °C for 1 h.Immunoreactive ACTH concentrations analyzed as a percentage of the reference 4 °C sample showed that only temperature had a statistically significant effect on the measured immunoreactive ACTH concentration (p = 0.003) and there was no significant effect of PPID status or TRH stimulation on the measured ACTH concentration. Samples kept at 20, 30, and 40 °C all showed deviations above and below the ACTH concentration of the reference sample, ranging from 84.3 to 122.1% in the samples kept at 20 °C, from 89.4 to 115.6% in the samples kept at 30 °C, and from 86.5 to 116.3% in the samples kept at 40 °C; however, none of these changes were statistically significant. Significantly lower immunoreactive ACTH concentrations were observed in samples that were kept at 70 °C, with all ACTH concentrations below the ACTH concentration of the reference sample, ranging from 72 to 94.4% (p = 0.001, Figure 3). However, out of the 15 samples that were kept at 70 °C for 1 h, only 9 could be analyzed due to solidification of the sample.Immunoreactive ACTH concentrations analyzed as the absolute percentage change from the reference sample showed that both temperature and PPID status had a significant effect on the immunoreactive ACTH concentration (p = 0.001 and p = 0.01, respectively); however, there was no significant effect of TRH stimulation on the ACTH concentration (p = 0.8). For the 7 non-PPID horses, when compared to the 4 °C reference sample, samples kept at 20, 30, 40, and 70 °C had significantly different ACTH concentrations (p = 0.02, 0.01, 0.04, and p = 0.003, respectively) ranging from 1.3 to 22.1%, 2.1 to 15.6%, 0.4 to 16.3%, and 5.6 to 28%, respectively, above or below the reference sample (Figure 4).For the 8 PPID horses, when compared to the 4 °C reference sample, samples kept at 30, 40, and 70 °C had significantly different ACTH concentrations (p = 0.01, 0.02, and p = 0.001, respectively) ranging from 8 to 9.0%, 0 to 9.6%, and 7.2 to 23.1%, respectively, above or below the reference sample (Figure 5). No significant difference in the immunoreactive ACTH concentration was detected between samples kept at 20 °C compared to samples kept at 4 °C (p = 0.07), with differences ranging from 0.6 to 16.5% above or below the reference sample.Although the immunoreactive ACTH concentrations significantly changed with temperature and PPID status, the change in the ACTH concentration was only clinically relevant in one horse, a non-PPID horse that underwent TRH stimulation. This horse had a reference 4 °C ACTH concentration of 60.9 pg/mL, but after storage of the whole blood sample at 20 and 40 °C for 1 h, the measured ACTH concentrations were 69.1 and 70.8 pg/mL, respectively. This would have led to a false positive diagnosis of PPID if a diagnostic cut-off value of 65 pg/mL was used.4. DiscussionThe main results of this study are that (1) statistically different changes in the measured immunoreactive ACTH concentration occurred when samples were stored as whole blood at 70 °C for as little as 1 h, (2) clinically significant changes were observed for the measured immunoreactive ACTH concentration in 1 of 15 horses when samples were stored at 20 and 40 °C, (3) the stability of immunoreactive ACTH from PPID and non-PPID horses was different, and (4) no significant effect of TRH stimulation on immunoreactive ACTH stability was detected.Previous research has investigated the effects of ACTH stability at room temperature (21 °C) on ACTH concentration, finding no difference between samples kept at room temperature verses those kept at 4 °C when analyzed prior to 8 h [15]. After that time, the ACTH concentration of samples stored at room temperature began to decline, compared to refrigerated samples; however, it should also be noted that those results were mostly from PPID-negative horses [15]. These findings suggested that the storage of samples at room temperature could provide valid results. In contrast, our study found that samples kept at 20 °C for only 1 h led to deviations in the measured ACTH concentration of up to 22% above and 18% below the ACTH concentration measured in the same sample stored at 4 °C, suggesting that samples kept at 20 °C could potentially lead to occasional misdiagnoses. The Bland–Altman plots suggest that the bias is small for a given ACTH value at 20, 30, and 40 °C, with reasonably small increases in the 95% LOA as the temperatures increased to 40 °C. Therefore, accidently keeping a sample at room temperature, or in a pocket or car up to 40 °C for 1 h is unlikely to be clinically relevant in most cases. Leaving a sample in a car in the sun on an extremely hot day where vehicle temperatures can reach 70 °C is not recommended, with reductions in the ACTH concentration in samples that did not solidify.We also found that individual horses with and without PPID had small random and unpredictable rises or declines in the measured immunoreactive ACTH concentration in response to short-term exposure to high temperatures (>20 °C). Horses without PPID showed more variability in the absolute percentage change in the ACTH concentration compared to horses with PPID, but their reference values at 4 °C were lower so the percentage change was higher. A previous study investigating the effect of processing delays on ACTH concentrations also found that the stability of ACTH was affected by the PPID status of the horse [16].The differences in stability observed between PPID-positive and PPID-negative horses could be due to the site of ACTH production. PPID-negative horses release the majority of their stress-induced ACTH from the corticotropes of the pars distalis within the pituitary gland, and only a small amount from the pars intermedia [21]. In contrast, PPID-positive horses release more proopiomelanocortin (POMC)-derived peptides, such as α-melanocyte-stimulating hormone (α-MSH), corticotrophin-like intermediate peptide (CLIP), β-endorphins, and ACTH, from the melanotropes within the pars intermedia of the pituitary gland. These peptides likely have different structures than pars distalis-derived ACTH [22]. In both PPID-negative and -positive horses, prohormone convertase 2 (PC2) cleaves ACTH into α-MSH and CLIP [18,22]. It is likely the Immulite™ assay also cross reacts some of these POMC peptides. It is possible that degradation of ACTH to CLIP occurs in vitro. Horses with PPID have been shown to have an upregulated expression of the messenger ribonucleic acid of PC2 [22]. This mutation reduces the conversion of ACTH to α-MSH, resulting in high plasma concentrations of an ACTH peptide with poor bioactivity [18,22]. These ACTH peptides could be increasingly more susceptible to proteolytic degradation or have a structure that interacts differently with the antibodies used in chemiluminescent assay, resulting in variable immunoreactive ACTH concentrations [18].It has been shown that the peptides measured as basal ACTH are different to the ACTH fragments measured after TRH stimulation, with marked differences in peptide detection using different analyzers [23]. The Immulite™ 1000 is a chemiluminescent immunometric assay that uses two-site sequential mouse monoclonal and rabbit polyclonal anti-human ACTH antibodies. The Immulite™ 1000 assay consistently measures higher values of the ACTH concentration when compared to an immunofluorescence assay in samples concurrently measured from ponies or horses for basal and TRH-stimulated ACTH [23,24,25]. Poor agreement between the Immulite 1000 and a commercial ACTH radioimmunoassay has also been documented [26]. The season also affects basal and post TRH ACTH concentrations in horses with and without PPID, and seasonally specific thresholds for the diagnosis of PPID specific for each type of analyzer should always be utilized [8,27,28]. It was unknown if post-TRH stimulation ACTH would have the same stability as basal ACTH. We were unable to detect a difference in the effects of TRH stimulation on ACTH stability, suggesting that both basal ACTH and ACTH post TRH stimulation have equal stability.A limitation of our study was the small sample size. However, by separating the eight PPID-positive and the seven PPID-negative horses, the variability in the ACTH concentration was reduced compared with other studies [16]. Significant effects of temperature on ACTH stability were detected even when using small sample sizes, indicating adequate power of the study when absolute values were utilized. Although storage at 70 °C caused a uniform reduction in the measured ACTH concentrations, storage at temperatures between 20 and 40 °C caused some ACTH measurements to increase and some to decrease in an unpredictable manner, which was more than the intra-assay coefficient of variation of the analyzer in 43 of the 60 samples (72%). Utilizing the absolute values of the percentage change in the ACTH concentration compared to the reference samples in the analysis prevented statistical cancelation of the results that either increased or decreased. The horses that were included in the study were of ages relevant to the diagnosis of PPID, with horses ranging in ages from 11 to 27 years. Various breeds and sexes meant the group was an adequate representation of a general practice population of full-sized horses. If a larger number of horses were utilized, potentially more clinically significant effects resulting in misdiagnosis of PPID might have occurred. Although these changes in the immunoreactive ACTH concentration only caused 1 misdiagnosis within our study (7%), there is potential that the effects of stability could have a significant effect when monitoring ACTH concentrations in individual PPID horses over time, which is recommended for individual pergolide dosage adjustment. An erroneous result could be enough to alter treatment decisions.Another limitation was that samples in general ambulatory practice might be exposed to a variety of temperatures over a short period of time. A sample could be exposed to summer environmental temperatures (20 to 45 °C) for a period until being placed in a polystyrene cooler box with ice blocks or a portable fridge (4 °C). If no cooling device is available, and the sample is left in the vehicle, it could then be exposed to vehicle temperatures up to 70 °C in regions that experience high environmental summer temperatures [17]. These variations in temperatures could potentially cause ACTH concentrations to be even more unpredictable then those found in this environmentally controlled study.5. ConclusionsIn conclusion, the stability of equine ACTH is affected by short-term exposure to a temperature of 70 °C. When ACTH variation is reduced by grouping horses with and without PPID, our results using absolute values suggest that samples should be kept at 4 °C to reflect the true ACTH concentration. Exposure to temperatures of up to 40 °C for 1 h can still provide an appropriate diagnosis of PPID in most cases, but ACTH concentration might change by up to 28%, and 40% of samples exposed to 70 °C will be unmeasurable. TRH stimulation does not appear to affect immunoreactive ACTH stability.

animals : an open access journal from mdpi

10.3390/ani11102911

PMC8532645

Late pregnant dairy cows housed outdoor can be exposed to hot weather conditions for several weeks prior to calving affecting their physiology and behavior. We aimed to determine whether access to an artificial shade for outdoor-housed dairy cows during the three weeks prior to calving had a positive effect on lying, rumination, feeding, and drinking behaviors. Also, the relationship between access to shade and health status was investigated. Shaded cows increased rumination time, but the daily lying time was similar to unshaded cows. Also, shaded cows spent half of the time drinking during the warmest hours of the day and spent more time feeding during the morning feed than unshaded cows. The prepartum and postpartum body fat mobilization and presentations of clinical diseases after calving were similar between both treatments. This study presents evidence that shade is an important resource for cows during temperate summers, observing effects mainly on behavioral variables.

Cows are affected by environmental factors associated with warm weather conditions; however, little is known about the effect of shade access especially during the prepartum period of dairy cows in temperate regions. This study assessed the effect of shade on the behavior (lying, rumination, feeding, and drinking), body fat mobilization, and health status of outdoor-housed dairy cows during the prepartum period under temperate summer conditions. During the 3 weeks prior to calving, 24 multiparous Holstein cows were grouped (4 cows/group) and assigned to either an open corral without shade or with access to shade until calving. We daily measured shade use, lying, rumination, feeding, and drinking behavior. Weekly, prepartum non-esterified fatty acids (NEFA) and postpartum b-hydroxybutyrate (BHB) concentrations were measured. Clinical examination was periodically performed individually until 21 postpartum days. Shade use averaged 45.6, 46.0, and 19.8% during the hottest hours of the day (11–18 h) in weeks 3, 2, and 1 prior to calving, respectively. Shaded cows had higher values for rumination time and feeding time during the morning but spent less time drinking during the warmest hours than unshaded cows. NEFA and BHB concentrations and clinical diseases were similar between both treatments. These findings suggest that under temperate summer conditions the access to an artificial shade is an important resource, observing beneficial effects mainly on behavioral variables.

1. IntroductionSeasonal calving is often used in grazing dairy farms in temperate regions to maximize pasture nutrient utilization to support lactation, minimize costs associated with purchased feeds and maintain the supply of milk through the year [1,2,3]. Cows that calve at the beginning of the autumn calving season might be especially affected by environmental factors associated with warm summer weather during their prepartum periods such as high air temperature, relative humidity, and high solar radiation levels [4]. Considerable research has shown that exposure to warm weather imposes challenges to the welfare and productivity of animals in pasture-based systems, but most of the work to date has focused on lactating dairy cows [5,6,7]. A better understanding of the responses of dry pregnant cows to summer weather even in temperate zones would be an important first step for designing better management practices during this important period.Although behavioral changes have been documented in both indoor and pasture lactating dairy cows exposed to warm weather, little is known about how these climatic conditions may affect cow behavior during the weeks before calving. Karimi et al. [8] showed that prepartum housed cows adapt to heat stress (defined as the temperature–humidity index, THI > 72, equivalent to 25 °C at 50% relative humidity) through increasing meal size and reducing meal duration, as well as increasing standing times. Similarly, Paudyal et al. [9] reported that the average rumination time during the days prior to calving in indoor cows was lower in the hot season (monthly average THI ≥ 76) compared to the cool season (monthly average THI < 76). More recently, Edwards et al. [10] found that cows kept on pasture and exposed to moderate heat stress conditions (THI > 68 and ≤ 79) during the week before calving decreased their lying time compared to those not experiencing heat stress conditions (THI ≤ 68). To date, no studies have quantified the effect of warm environmental conditions on the feeding and ruminating behavior of prepartum dairy cows managed outdoors. Previous research on late gestating dairy cows housed indoors has also determined that hot conditions reduce their feed intake, causing body fat mobilization and weight loss [11]. Excessive lipid mobilization, reflected by an increase of non-esterified fatty acids (NEFA) and β-hydroxybutyrate (BHB) levels, contributes to oxidative stress and inflammatory responses [12,13] and increase the risk for common postpartum disorders [14,15]. Thus, exposure of prepartum cows to summer warm weather may influence their health status after calving. In pasture-based systems, one of the most economical ways of modifying the environment to mitigate the effect of summer weather is the provision of shade [16]. Previous studies have shown that seeking shade is an important behavioral coping strategy for pasture lactating cows when exposed to hot summer conditions in temperate climates [17,18,19,20]. For instance, shade access on sunny days with temperatures higher than 25 °C is beneficial to reduce the body temperature in lactating dairy cows [17,21]. Aside from changing the body temperature, access to shade has also been shown to increase feed intake [22] and, consequently, milk production of dairy cows on pasture [17,21]. Although the beneficial effects of providing shade to heat-stressed cattle in hot climates have been well documented in both indoor and outdoor-kept dairy cows ([23,24]), little is known about how to shade access may influence the behavior and body fat mobilization of dairy cows throughout the dry period in temperate regions. Therefore, our study aimed to evaluate the effects of providing access to artificial shade on lying, rumination, feeding, and drinking behavior of prepartum dairy cows housed outdoor during a temperate summer. Additionally, the effects of shade access on energy analytes in the blood (prepartum NEFA and postpartum BHB) and clinical disease occurrence after calving were also evaluated. We hypothesized that prepartum cows with access to a shade would spend more time lying, ruminating, and feeding, and less time drinking in comparison to prepartum cows without shade access. We also predicted that energy analytes in the blood and disease occurrence would be reduced in dairy cows with prepartum shade provision.2. Materials and MethodsThis study was carried out between January and March 2019 during the summer season in southern Chile, at the Austral Agricultural Experimental Station of the Austral University of Chile, located in Valdivia (39°46′42′′ S, 73°13′38′′ W). Valdivia has a Temperate Oceanic Climate (Cfb: without a dry season and warm summer) according to Köppen-Geiger Classification [25]. The experimental farm managed a total of 160 Holstein-Friesian cows with an average of 305-d milk production of approximately 7000 kg (4.56% fat and 3.51% protein). 2.1. Animals, Management and Experimental DesignThe study included 24 multiparous Holstein-Friesian dairy cows that were selected from a group of 40 pregnant dry cows, based on their expected calving date and with no signs of clinical health problems (i.e., lameness, injuries). There was a difference of 28 days between cows with the earliest and latest calving dates. The animal’s parity number was [mean ± standard deviation (SD)] 3.3 ± 1.3, and the bodyweight (BW) and body condition score (BCS) at the start of the experiment was 620 ± 80 kg and 3.2 ± 0.4 on a 5-point scale, respectively. From the dry-off (60 days before their expected calving date) to approximately 30 days before the expected parturition date (far-off group), cows were housed together in paddocks with pasture (mixture of grasses and legumes) and managed daily in a rotational grazing method. Cows had free access to water bins and no access to shelter. The stocking rate in these paddocks was maintained at approximately 102 cow/ha. Approximately 30 ± 5 days before the expecting calving date, cows were allocated into six groups (4 cows per group) based on parity and BW, and each group was randomly assigned to one of two treatments: corrals without shade access (3 groups/n = 12 cows) or corrals with shade access (3 groups/n = 12 cows). All cows were moved at the same time from the far-off group to the experimental prepartum groups and were evaluated during the last 3 weeks of gestation. Data obtained during the first 4 days were excluded from the final analysis because that time was used to acclimatize the groups of cows.The corrals in both treatments measured 17 m × 12 m (51 m2/cow), had a bare soil surface with no grass cover, and were delimited by an electric fence allowing visual and auditory contact but limiting tactile interaction between cows. A freestanding shade structure (10 m long, 6 m wide, and 2.5 m above the ground) was constructed in 3 out of the 6 corrals. The structure’s tops were covered with a shade cloth that blocked 80% of solar radiation, providing a shaded area of 15 m2/cow (measured between 10:00 h and 18:00 h). Shade structures were orientated in the same direction with the 6 m side facing north. Corrals with shade access were located interspersed among the corrals without shade access to prevent any overlap in shadow casting (Figure 1). All corrals were daily cleaned of manure in the morning during the study period.Cows were fed twice a day at 08:30 and 18:00 h and the ration was provided in two feed bins placed outside the shaded area. The ration was formulated following the NRC, [25] guidelines and consisted of grass silage (40.5 ± 5.1% DM, 10.5 ± 1.7% CP, 54.8 ± 7.3% NDF and 10.9 ± 0.67 MJ/kg DM [mean ± SD]) and ~3 kg of commercial concentrates/d on an as-fed basis (87% DM, 22.5% CP, 12.0% NDF, and 11.56 MJ/kg DM [mean]) with anionic mineral mix (Mg 4.0%, Cl 33.0%, S 2.6%, Ca 0.7%, K 0.3%, 2 mg Na kg−1, 1050 mg Cu kg−1, 2100 mg Mn kg−1, 3500 mg Zn kg−1, 140 mg I kg−1, 13 mg Co kg−1, 10 mg Se kg−1) at a rate of 0.20 kg cow per day. Water was provided ad libitum in a water trough (one in each corral, 600 L).After calving, cows remained in the experimental corrals for 6 h with their newborn calf. Then the calf was moved to the calf barn and the dam to the lactation group. The newly calved cow was not replaced by another cow to prevent changes in the social structure of the group. The experimental cows in the postpartum lactation group continued to be monitored until 21 days in milk (DIM). These cows were fed under a daily rotational grazing system conformed by perennial ryegrass and white clover (21.1% DM, 19.3% CP, 47.1% NDF, and 11.7 MJ/kg DM [mean]), providing 6 kg of commercial concentrates/d on an as-fed basis (86 to 88% DM, 11.5% CP, 32.3% NDF, and 13.19 MJ/kg DM [mean]) during each milking. A mineral mix (Ca 14.0%, P 10.0%, Mg 6.0%, Na 4.0%, S 0.2%, 5000 mg Zn kg−1, 1500 mg Cu kg−1, 200 mg I kg−1, 20 mg Co kg−1, 14 mg Se kg−1) was offered with the concentrate at a rate of 0.25 kg cow per day. All the animals were milked twice a day (at 06:00 and 15:00 h).2.2. Data Collection2.2.1. Environmental MeasurementsClimatic variables including air temperature (°C), relative humidity (%), wind speed (km/h), solar radiation (W/m2), and precipitations (mm) were recorded continuously using a weather station (A720, ADCON Telemetry GMBH) located 1 km from the research location. Data were recorded every 60 min over a 24-h period during the prepartum experimental period. These data were used to calculate THI and Heat Load Index (HLI) for each prepartum day, according to Ravagnolo and Misztal, [26] and Gaughan et al. [27], respectively. For the calculation of black globe temperature, we used the formula reported by Hahn et al. [28]. We then categorized each hour based on THI levels (<68 = no heat stress, and ≥68 = heat stress) [29,30], and HLI levels (<70 = thermoneutral conditions, and ≥70 = warm/hot conditions) [27].2.2.2. Behavioral MeasurementsTo determine the use of shade as well as the behavior under the shade (lying/standing), two infrared video cameras (Ezviz CS-CV310-A0-1B2WFR, Ezviz, City of Industry, CA, USA) were mounted on two corners of each corral at a height of 2.2 m, and both directed down the center of the shade structure. Video recordings were analyzed over the prepartum period using scan sampling at 3-min intervals during the hottest part of the day (time with the highest solar radiation) between 11:00 and 17:59 h. Prior to the video recording, each cow was marked with a unique number on both sides using hair dye to facilitate individual recognition. A cow was considered to be using the shade when at least 2 hooves were in the shade cast by the structure [18]. Lying underneath the shade was defined as a cow having its flank in contact with shaded ground, with no weight supported by any of the legs. Standing underneath the shade was defined as a cow in an upright posture without the forward motion in the shaded area. One trained observer transcribed the behavioral data from the videos, and the intra-observer percentage agreement was above 98% for all behaviors (shade use and lying/standing under shade). These data were used to calculate a percentage of daily use of the shade. Daily lying time, number of lying bouts per day (i.e., frequency of transitions from lying to standing positions), and duration of lying bouts per day (i.e., calculated as the ratio of minutes lying to the number of lying bouts per day) were measured using electronic data loggers (HOBO Pendant G Acceleration Data Logger, Onset Computer Corporation, Bourne, MA, USA, validated by Ledgerwood et al. [31]. The devices were configured to record at 1-min intervals and then each data logger was attached to the lateral side of one of the hind legs of the cow, along with a self-adhesive Coban™ self-adherent wrap, and remained in position up to three weeks after calving. Every 10 days, the data loggers were changed and the information from those days of change was not considered for the analysis. The data were downloaded weekly, isolated standing or lying events were removed from the base following the recommendation of Ledgerwood et al. [31].Rumination time was recorded using an individual rumination tag on a neck collar. This automatic system (Hr-Tag, SCR Engineers Ltd., Netanya, Israel) records rumination time continuously over a 24-h period as validated for dairy cows by Schirmann et al. [32]. Data were stored in the memory of the logger and summarized as min/2-h intervals. The tag information was transferred to and stored in the control unit through radio frequency. This information was downloaded onto the database weekly. If one 2-h interval was not recorded by the system, the entire day was discarded.Feeding and drinking behavior was recorded using one infrared trail camera (Q-See QT5682-411, Anaheim, CA, USA) situated 2 m above each feeder and water bin and fixed to stands. Both behaviors were evaluated on four days of each prepartum week (12 days/cow), using scan sampling every 3 min for each cow. Before the start of the experiment, we perform a preliminary study with other animals to measure the time that cows spend eating the morning and afternoon ration in order to define our observational period for assessing this behavior. Cows consumed the total delivered feed within 2.5 h in the morning and the afternoon. For this reason, to assess feeding time, video recordings were analyzed for 3.5 h after morning feeding (from 08:30 to 12:00 h) and for 3.5 h after afternoon feeding (from 18:00 to 21:30 h). A cow was considered as feeding when it was standing and having the head above/or in the feed bin, and silage was visible in her mouth as described by Schütz et al. [33]. To assess drinking time, video recordings were analyzed between 11:00 and 18:00 h. A cow was considered as drinking when she was standing and facing the water throw and swallowing water as described by Huzzey et al. [34]. Videos were analyzed by a one trained observer (intra-observer reliability was 96% for feeding behavior and 97% for drinking behavior). 2.2.3. Blood Sampling and Health RecordsBlood samples from all cows were taken weekly during the same day from the time of enrollment until 3 weeks after parturition. During the prepartum period, between days −21 to −15 to represent week −3, days −14 to −8 to represent week −2, and days −7 to −2 to represent week −1 relative to calving. Postpartum blood samples were collected between days 2 to 7 to represent week 1, days 8 to 14 to represent week 2, and days 15 to 21 ± 1 to represent week 3. Samples were collected at approximately 09:00 h during the prepartum period and after morning milking during the postpartum period. Blood was collected from the coccygeal vein through vacutainer® system into 5-mL EDTA sterile tubes and then transported to the laboratory within 5 h of collection. Serum was separated immediately upon arrival at the laboratory and stored at −20 °C until analysis. NEFA concentration (Randox, Crumlin, UK) in weeks −3, −2, and −1 before calving and BHB concentrations (Ranbut, Randox, Crumlin, United Kingdom) in weeks 1, 2, and 3 after calving were measured by enzymatic analysis using an auto-analyzer (Metrolab 169 2300, Wiener Lab, Rosario, Argentina). The intra-assay coefficient of variation for the BHB and NEFA assay were 3.8% and 2.8%, respectively. Clinical disease diagnosis was performed on each cow by a trained veterinarian on days 0, 3, 6, 9, 12, and 21 after calving, and after morning milking. Clinical examination included direct observation of the animal, measurement of the rectal temperature, ruminal auscultation, and evaluation of the vaginal discharge. The retained placenta was defined as failure to expel the placenta within 24 h after parturition. Clinical hypocalcemia was defined as any recumbent cow within 72 h after parturition exhibiting anorexia, nervous symptoms, staggering, varying degrees of unconsciousness, and good response to intravenously administered calcium. Puerperal metritis was characterized by a fetid watery red-brown vaginal discharge and rectal temperature ≥ 39.5 °C. Clinical mastitis was detected by the milker using inspection of the foremilk and palpation of the udder at milking and was characterized by the presence of abnormal milk or by signs of inflammation in 1 or more quarters. The presence of any other clinical health disorders such as digestive or respiratory problems was also recorded. Additionally, the gait score of the cows was assessed by the trained veterinarian when they exited the milking parlor using a 5-point numerical rating system (NRS), where 1 = sound and 5 = severely lame [35]. Lameness was categorized as clinical lameness (prevalence of cows scored as NRS = 3) and severe lameness (prevalence of cows scored as NRS = 4 or 5).2.3. Statistical AnalysisOne cow in the group without shade access was excluded from the experiment during the acclimatization period due to the display of frequent agonistic behavior (butting, pushing, fighting, and threatening) against other cows and its data was excluded from the analyses. Statistical analysis was performed using R language [36] with cow as the experimental unit. Environmental (temperature, relative humidity, wind speed, solar radiation, THI, and HLI) and behavioral measures (shade use, behavior under the shade, and rumination, feeding and drinking behavior) were summarized for each day and then summarized by observational periods based on days relative to parturition (week −3: days −21 to −15; week −2: days −14 to −8; week −1: days −7 to −2). Blood analytes values were stratified according to the prepartum (NEFA) and post-partum week (BHB) in which they were evaluated.Mixed models, using the lme4 package [37] and lmerTest [38] was used to determine differences in the use of shade between the weeks prior to calving. The model included the fixed effects of the period relative to calving and the random effect of cow nested within the group.To determine the effect of climate conditions on time spent under the shade, data of daily weather conditions were combined with shade use data for each cow according to the calendar date. Data were analyzed using mixed models (lme4 package and lmerTest), considering as fixed effects the period relative to calving and the daily weather variables (solar radiation, wind speed, THI, HLI, and their interactions), and as the random effect, the cow nested within the group. Temperature and relative humidity were strongly related (r > 0.8) to THI and, therefore, excluded from the model. The calculated AIC of the overall model determined whether additional climate factors improved the model fit. To detect differences in behavior and blood metabolite concentrations between cows with and without shade access during the prepartum period, data were analyzed using mixed models (lme4 package and lmerTest). The models considered treatment group (with and without shade access), period relative to calving and their interactions as fixed effects, and cow nested within the group as a random effect. The association between treatments (with and without shade) and health outcomes (absence or presence of health events) were analyzed using 2 × 2 contingency tables and Fisher’s exact test statistic.Residuals were examined in all models to verify normality and homogeneity of variances and to detect possible outliers and influential points. No observations were removed from the analyses. The significance level was defined as p < 0.05, p < 0.01, p < 0.001, and tendencies were considered p < 0.1.3. Results3.1. Environmental ConditionsA summary of the weather conditions during the prepartum period is shown in Table 1. No precipitation was recorded on any of the experimental days. The percentage of hours with THI values ≥ 68 between 11:00 to 17:59 h was 86 ± 10%, 63 ± 8.6%, and 43 ± 11% during week −3, −2, and −1 relative to calving, respectively. During the same period of the day, the percentage of hours with HLI values >70 was 91 ± 7.1%, 72 ± 6.3%, and 50 ± 9.2% during week −3, −2, and −1 relative to calving, respectively. Hourly means of solar radiation during the study period are shown graphically in Figure 2.3.2. Shade Use During 11:00 to 17:59 h, we found that cows spent approximately half of their time beneath the shade during week 3 (45.6 ± 2.7%; range: 10–96%) and 2 (46.0 ± 3.0%; range: 11–100%) prior to calving. However, shade use was lower on the week prior to calving than on previous weeks (19.8 ± 2.1%, range 0–67%; p < 0.05). On average, cows spent a low proportion of their time lying down beneath the shade during week 3 (10.6 ± 2.1%, range 0–25%), and 2 (8.9 ± 1.3, range 0–19%) before calving, and during the week prior to calving was even lower than on previous weeks (6.4 ± 1.2%, range 0–21%; p < 0.1). Most of the time, prepartum cows used the shade individually (30.1 ± 1.7%) or in pairs (36.7 ± 2.0%). Only 21.5 ± 2.3% and 11.7 ± 1.7% of the time we observed three or four cows using the shade simultaneously, respectively. Prepartum cows used the shaded area at different hours of the day, reaching the highest values between 13:00 and 15:00 h during weeks 3 and 2 prior to calving (Figure 3). The use of the shade by cows was influenced by some environmental conditions during the entire prepartum period; the increase in HLI and solar radiation increased the use of shade (HLI Estimate: 0.7%; SE: 0.02; p < 0.01; solar radiation Estimate: 0.1%; SE: 0.02; p < 0.05), and the wind speed decreased the time that cows spent lying under the shade (Estimate: −6.4%; SE: 2.0%; p < 0.01). THI had no effect on the cow location and the posture under the shade (lying/standing).3.3. Lying, Ruminating, Feeding, and Drinking BehaviorTreatments had no effect on total daily lying time, lying bout, and lying bouts duration pre-calving (p > 0.05; Table 2). Overall, we observed a week effect in which cows spent less time lying and had more lying bouts of shortening duration during the week prior to calving than in other weeks, regardless of the treatment (p < 0.01). There was a tendency for cows with shade access to have a greater mean daily rumination time during week 3 prior to calving compared to cows without shade access (p < 0.09; Table 2).Between 11:00 to 17:59 h, cows with access to shade spent less time drinking than cows without shade access (Table 3). Although cows in both treatments consumed all the feed within the morning and evening observation periods, cows with access to shade spent more time feeding than cows without shade access only during the morning hours (Table 3).3.4. NEFA, BHB, and Health StatusWe found no treatment effect on NEFA or BHB concentrations pre-calving (p > 0.05; Table 4). The proportion of cows that were considered as clinically ill or healthy after calving was similar between treatments, where 33% (n = 4) of cows with shade access (n = 12) and 27% (n = 3) of cows without shade access (n = 11) were diagnosed with at least one clinical disease postpartum (Fisher´s exact test, p = 0.99). 4. DiscussionAlthough the importance of shade access for cattle exposed to summer weather conditions in temperate climates has been the focus of several studies (reviewed by Van Laer et al., [39] and Webster et al., [40]), to our knowledge no studies have described the effectiveness of providing shade for prepartum dairy cows managed outdoor. We found that when prepartum cows had free access to an artificial shade during the warmest hours of the day, they spent approximately half of this time beneath the structure (~3.2 h; from 11:00 to 18:00 h), especially in the third and second week prior to calving. This result is in agreement with previous findings that mid-lactating cows on pasture used the shade approximately 3 h during the same period of the day and under similar weather conditions in New Zealand [18,20]. However, our results showed a decrease in shade use of approximately 20% (1.4 h) during the week prior to calving. This finding was most likely because during this week the percentage of hours under hot weather conditions was lower (THI 43% and HLI 50%) compared to the other weeks (week −3, THI 86% and HLI 72%; week −2, THI 63% and HLI 70%). When prepartum cows were beneath the shade between 11:00 and 18.00 h, they spent less than 10% of the time lying down. It is possible that the motivation for lying during the hottest period of the day decreases as an attempt to dissipate the heat gained from the environment and the fetus [41], since a standing position increases the available body surface for heat dissipation [17,42,43]. The use of shade and the behavior beneath the shade (i.e., standing/lying) were influenced by environmental conditions, specifically solar radiation and wind speed. Regardless of the week relative to calving, cows spent more time under the shade as the average solar radiation increased, showing that prepartum cows try to reduce body temperature by seeking shade, in the same way, it has been described for lactating cows [17,18,20]. Providing prepartum cows with shade is likely beneficial in terms of reducing heat load due to the protective effect of these structures on solar radiation (reviewed by Lees et al., [44]). We also observed that the wind speed reduced the time that cows remained lying under the shade. This finding disagrees with Tullo et al., [45] who observed longer lying times in housed lactating dairy cows as the wind speed increase under a shade roof from 1 m/s to 1.8 m/s. These authors suggested that air movement underneath the shade favors the dissipation of heat to the environment, generating a cooler environment and promoting lying behavior. It is possible that in our study the artificial shade did not create a cool environment under it, favoring the cows to remain standing position exposing their body to the wind. Unfortunately, we did not measure the floor temperature under the shade to test this hypothesis. A better understanding of the effects of environmental conditions on shade use and behavioral responses in prepartum dairy cows is needed. In our study, all cows used the shade but most of the time they were alone (30%) or in pairs (37%). One possible explanation might be that the shaded area was a limited resource for the group, leading to agonistic interactions [20,46]. However, the space granted in our experiment (15 m2/cow) was higher than the 3.5 to 5.6 m2/cow recommended for mature dairy cows [47]. Schütz et al., [48] observed a similar behavior, where lactating dairy cows rarely used simultaneously the shade in pasture paddocks, even at low densities (8.1 and 8.8 m2/cow). In our study, factors such as hierarchy or social isolation, which is important in cows close to calving [49], may have influenced this behavior and are necessary to consider in future research.The daily lying time was similar for shaded and unshaded cows (10.7 and 11.2 h, respectively) over the prepartum weeks which agrees with previous studies carried out in lactating grazing dairy cows with and without shade access under temperate summer weather conditions [7,18,19]. Previous studies have found that temperatures below 21 °C overnight for 3 to 6 h allow the opportunity for the dairy cattle to dissipate the heat gained during the day [50]; thus, there is evidence that during the warm day weather conditions dairy cattle are motivated to lie down during the cooler night hours [7]. We did not investigate lying time during the day and night hours separately in this study, but we argue that both groups of cows were motivated to lying during night hours, independently of the shade access during the day. We also found that access to shade did not affect the daily lying bouts or bouts duration during the prepartum weeks. However, we observed that during the week previous to calving the daily lying bouts increased and lying bouts duration decreased significantly in both groups. These findings are in line with recent studies that found an increase in lying bouts and a reduction in the duration of lying bouts of healthy grazing dairy cows in the days close to calving [51,52]. These changes in lying patterns might be an indicator of discomfort and restlessness associated with parturition [34,52].Shaded cows tended to have a higher rumination time compared to non-shaded cows (~59 min) during the prepartum period. Shorter rumination times have been described in housed lactating cows under warm (THI > 76; Soriani et al., [53]) and temperate (THI > 52; [54]) environment conditions. However, others have shown no relationship between rumination time and THI conditions [55]. Paudyal et al., [9] observed that rumination time in prepartum cows was shorter in the hot season (monthly average THI ≥ 76) than in the cool season (monthly average THI < 76; 432 vs. 487 min/d) under subtropical weather. Further research is encouraged to determine the effect of shade access on the rumination behavior of cows kept in outdoor conditions during the prepartum period in temperate regions.Feeding time tended to be higher in shaded cows during morning feeding compared to unshaded cows, and the reason for this remains unclear. Moallen et al., [56] and Soriani et al., [53] observed reductions in the dry matter intake of lactating dairy cows in the next days when exposed to THI values > 76. In the present study, although the dry matter intake was not evaluated, shaded and unshaded cows ate the whole ration both in the morning and evening, affecting only the feeding time. In this sense, small but more frequent meals are described in lactating dairy cows exposed to heat stress [57], which could have reduced the feeding time in cows without access to shade. Also, we hypothesized that cows without access to shade generated frustration in the animals, as has been previously reviewed by Polsky and von Keyserlingk [58]. This could increase competition in the feeder and, therefore, a reduction in feeding time as reported by Crossley et al., [59]. However, this was not evaluated in the present study.Drinking time during hot hours of the day (11:00–17:59) tended to be twofold higher in cows that had no access to shade when compared to shaded cows throughout the prepartum period. This result is in line with previous studies carried out under grazing conditions, where lactating cows without access to shade increased the time spent near a water trough [7,20]. The increase in the amount of water in unshaded cows could be related to increased water loss via skin and respiratory evaporation to dissipate the body heat [60], so shade access could be beneficial in reducing water demands. During the period around calving, the benefits of shade access may be greater because during this period cows are more thirsty because of exhaustion, stress, and colostrum production, leading to long drinking times such as the case of early lactation [61].We did not observe any effect of access to shade on serum NEFA concentrations in cows during the weeks before calving. This result might be explained by the relationship between NEFA concentration and food consumption [62], and both shaded and non-shaded cows consumed the whole ration of food, varying only the time spent eating the meal. It is also possible that due to the few heat stress events that the cows faced within this experiment, these animals had fewer energy costs associated with thermoregulation [63].Similarly, we found no differences in BHB concentrations between cows with access and cows without access to shade during the prepartum period. Taken together, these results suggest that access to shade under moderate warm weather conditions did not affect fat mobilization in prepartum cows. These findings could have also influenced the health results because in both groups NEFA and BHB concentrations were similar or lower than the limits proposed by Ospina et al., [14] as a risk factor for the development of postpartum diseases (i.e., clinical ketosis, abomasal displacement, placental retention, and metritis).5. ConclusionsThe results from this study showed that access to an artificial shelter during the last 3 weeks of the dry period under temperate summer conditions facilitated the daily rumination time, increased the feeding time after morning food delivery, and decreased drinking time, but it did not have a significant effect on NEFA or BHB blood concentrations or health status after calving. These results provide evidence that when prepartum cows are kept outdoors during the autumn calving season, cows should be given access to shade to reduce any negative effects of the warm weather on their behavior.

animals : an open access journal from mdpi

10.3390/ani11092679

PMC8469846

Early weaning is used to improve efficiency in pig production. However, early weaning may trigger liver oxidative stress in piglets. In this study, we evaluated the effects of early weaning on the development and antioxidant function of the liver in piglets. Our findings show that early weaning significantly decreases piglet body weight and suppresses liver development. We find that early weaning also suppresses the activities of superoxide dismutase (SOD) and catalase (CAT) (p < 0.05). It could be concluded that weaning may reduce the growth performance and liver antioxidant function of piglets.

This study examined the impact of early weaning on antioxidant function in piglets. A total of 40 Duroc × Landrace × Large White, 21-day-old piglets (half male and half female) were divided into suckling groups (SG) and weaning groups (WG). Piglets in WG were weaned at the 21st day, while the piglets in SG continued to get breastfed. Eight piglets from each group were randomly selected and slaughtered at 24th-day (SG3, WG3) and 28th-day old (SG7, WG7). The body weight, liver index, hepatocyte morphology, antioxidant enzymes activity, gene expression of antioxidant enzymes, and Nrf2 signaling in the liver of piglets were measured. The results showed that weaning caused decreased body weight (p < 0.01), lower liver weight (p < 0.01), and decreased the liver organ index (p < 0.05) of piglets. The area and size of hepatocytes in the WG group was smaller than that in the SG group (p < 0.05). We also observed that weaning reduced the activity of superoxide dismutase (SOD) and catalase (CAT) (p < 0.05) in the liver of piglets. Relative to the SG3 group, the gene expression of GSH-Px in liver of WG3 was significantly reduced (p < 0.05). The gene expression of Nrf2 in the SG3 group was higher than that in the WG3 group (p < 0.01). The gene expression of NQO1 in the SG7 group was higher than that in the WG7 group (p < 0.05). In conclusion, weaning resulted in lower weight, slowed liver development, and reduced antioxidant enzymes activity, thereby impairing liver antioxidant function and suppressing piglet growth.

1. IntroductionWeaning is one of the most important periods of piglet growth and may cause weaning stress. Early weaning causes physiological stress to piglets, leading to abnormal physiological and immune functions, as well as more severe oxidative stress [1,2]. The liver is the main site of piglet metabolism and contains a large number of mitochondria, where reactive oxygen species are produced [3,4]. A previous study has reported that weaning may cause liver oxidative stress in piglets, affecting liver cell apoptosis via MAPK signaling [5]. Antioxidant enzymes are important modulators of liver metabolism, and liver metabolic disorders are the main pathological factors leading to liver disease [6].Oxidative stress is caused by reactive oxygen species (ROS), which may damage proteins, nucleic acids, and cell membranes, and affect immune responses and barrier functions in weaned piglets [7,8]. Oxidative stress affects the development and physiological functions of tissues and organs in piglets [8]. It has been reported that early weaning could reduce the activity of digestive enzymes, impair tight junctions, and damage intestinal barrier function of weaned piglets [9,10]. Weaning also disrupts the redox balance, triggering oxidative stress and apoptosis [11,12]. Some other studies have also showed that the antioxidant enzymes activity is the main indicator of the liver function in weaned piglets [13,14]. The Nrf2 signaling pathway plays an important role in the process of oxidative stress and autophagy [15,16,17]. Under homeostatic conditions, Nrf2 binds to Kelch-like ECH-associated protein 1 (Keap1) in the cytoplasm. At the same time, Keap1 binds cullin 3 and forms an E3 ubiquitin ligase complex, which can promote the occurrence of Nrf2 ubiquitination, thus causing the rapid degradation of Nrf2 [18]. In response to stress, Nrf2 will dissociate from Keap1 and then enter the nucleus and combine with the antioxidant response element (ARE), and then activate the expression of antioxidant-related genes such as NAD(P)H quinone dehydrogenase 1 (NQO1) and heme oxygenase 1 (HO-1) [19]. In this study, we examined the effects of early weaning on the expression of antioxidant-related genes and antioxidant enzymes activity in weaned piglets, which are associated with Nrf2 signaling. This study was to study the effects of early weaning on the development and antioxidant function of the liver in piglets.2. Materials and Methods2.1. Ethics ApprovalAll procedures were approved by the Institutional Animal Care and Use Committee of Yangzhou University (PR China).2.2. Experimental DesignTwenty pairs of 21-day-old piglets ((Duroc × Landrace) × Large White) were chosen and randomly divided into SG and WG group. One pair includes two piglets with similar body weight from the same mother. The two piglets from the same mother were divided into two groups randomly, respectively. The sows were of similar parity (3rd or 4th) and fed with the same gestating diet that met nutrient requirements (formula were made according to the company standard). The piglets in WG group were weaned on the 21st-day old, while piglets in SG were breastfed by their own mothers until the end of the whole experiment. Eight piglets from SG and WG were randomly selected and slaughtered on the 24th-day old (SG3, WG3) and 28th-day old (SG7, WG7), respectively. The diets of the weaning piglets were designed according to the NRC (2012) (Table 1). During the experiment, the suckling group was allowed to breastfeed freely while the weaning group was given ad libitum access to the feed and water. Daily management, disinfection, and epidemic prevention were carried out according to the routine procedures of the pig farm.2.3. Sample CollectionAll piglets were fasted for 12 h before weighing and necropsy. The liver tissue samples were collected in 2 mL cryotubes and stored at −80 °C for later use. The formula for the calculation of organ index was: organ index (g/kg) = organ weight (g)/live weight (kg).2.4. Histomorphometry DeterminationLiver tissue was fixed in 4% polyoxymethylene for 48 h at 4 ℃. The tissues were then immersed in wax thrice, 40 minutes each, and embedded. They were then sectioned at 2 µm. The sections were then placed in a 37 °C water bath, dried, and stored for hematoxylin-eosin (H&E) staining. Two slices for each tissue sample were made, each slice was photographed under 10, 20, and 40× objective. A representative image of the stained part was taken on a microscope, and the size of liver cells quantified on Image-Pro Plus (6.0) (Media Cybernetics, Inc., Rockville, MD, USA). The software evaluated cell area and particle sizes of 30 cells per field of view, and finally calculated their average.2.5. Determination of Activity of Antioxidant EnzymesThe study was carried out 3 and 7 days after weaning (24 and 28 days old). All piglet liver samples were collected in 2 mL cryotubes and stored at −80 °C. Liver samples (0.9–2.1 g) were homogenized in 9 volumes of normal saline on an ice water bath. They were then centrifuged at 3000 rpm for 15 min, supernatant transferred into a new centrifuge tube, and stored at 4 °C. Total oxidizing capacity (T-AOC), superoxide dismutase (SOD), catalase (CAT), glutathione peroxidase (GSH-Px), and malondialdehyde (MDA) were measured using kits that were purchased from Nanjing Jiancheng Co., Ltd. (Nanjing, China).2.6. RT-qPCR AnalysisLiver samples were taken from −80 °C and total RNA extracted using TRNZOL. RNA purity (Tiangen Biotechnology, Beijing, China) was then measured using Applied Biosystems 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) and RNA quality examined by agarose gel electrophoresis. RNA was then reverse transcribed into cDNA and stored it at −80 °C for later use. Extraction of total RNA and its reverse transcription were performed according to the instructions of the TAKARA kit (TAKARA RR047A, Dalian, China). Primers were designed with Primer 5.0 according to the gene sequence of the pig (http://www.ncbi.nlm.nih.gov/pubmed/ accessed on 20 December 2019). Primer sequences are shown in Table 2. The primers were synthesized by Yingwei Jieji (Shanghai, China) Trading Co., Ltd. The data was calculated using the method of 2−ΔΔCT according to the reference [20].2.7. Statistical AnalysisThe statistical package for the Social Sciences (SPSS) 23.0 was used for statistical analysis. The body weight and liver weight of piglets were analyzed by independent t test. The other data was first analyzed by double factor variance analysis with group (G) and time (T) as fixed effects. The interaction effect of G × T was also considered. Independent t test was further used to analyze the difference between SG and WG on the different time point. Statistical significance was set at p < 0.05.3. Results3.1. Effects of Weaning on Piglet Liver IndexThe results showed that the body weight, liver weight, and liver index of the WG piglets was significantly lower than that in the SG piglets (p < 0.05), and the liver index decreased especially on the 7th day after weaning (p < 0.01) (Table 3).3.2. Effects of Weaning on Piglet Liver MorphologyThe results showed that the particle size and area of hepatocytes in the piglets of WG group was smaller than that of the SG group (p < 0.05), and the area of hepatocytes decreased, especially on the 3rd day after weaning (p < 0.05) (Table 4).3.3. Effect of Weaning on Antioxidant Enzymes Activity in Piglet LiverThe results show that the activity of SOD and CAT in the liver of WG was lower than that of SG (p < 0.05), and the activity of CAT decreased especially on the 7th day after weaning (p < 0.05) (Table 5).3.4. Effect of Weaning on the Expression of Antioxidant Enzymes Genes in Piglet LiverThe results illustrate that there were no significant differences in gene expression of PRDX, SOD, and CAT between SG and WG groups (p > 0.05) (Table 6). However, the gene expression of GSH-Px in WG3 was significantly lower than that in SG3 (p < 0.05).3.5. Effect of Weaning on the Expression of Nrf2 Related Genes in the Liver of PigletsEffects of weaning on gene expression of Nrf2 pathways in the liver of piglet are shown in Table 7. The gene expression of Nrf2 in livers of WG3 was significantly lower than that in SG3 (p < 0.01). Compared with SG7, the expression of NQO1 in livers of WG7 was lower (p < 0.05).4. DiscussionWeaning is an important stage in pig production. Weaning stress alters physiological and immune function of piglets, which affects piglet immune responses and intestinal barrier function [21,22]. Oxidative stress can damage mitochondrial function and increase the release of mitochondrial (ROS) [23]. The increase of ROS will increase the pro-inflammatory factors, and damage the immune system of piglets [24,25]. Previous studies have shown that weaning upsets redox balance and triggers oxidative stress in piglets [23,25]. The liver is a crucial digestive organ in piglets and contains a large number of mitochondria, which are susceptible to oxidative stress [26]. The effects of weaning on the antioxidant ability of liver are rarely searched at present. Here, we evaluated how weaning affects the growth and the antioxidant ability of liver in piglets.The liver is an important metabolic organ that oxidizes triglycerides to produce energy and plays an important role in lipid metabolism and immune regulation [27,28]. Liver development closely correlates with the growth of pigs [29]. Pigs weaned at 28-days old exhibit better growth performance than that of pigs weaned at 21-days old [30]. The earlier weaning time will slow down the growth of piglets, which is consistent with our findings. The results in this study suggested that weaning caused the loss of body weight and liver weight on the 3rd and 7th day after weaning. The results of hematoxylin-eosin staining showed that the area and particle size of hepatocytes in WG groups was smaller than that in SG groups, indicating that weaning stress significantly affects growth of weaned piglets, slowing liver development.Oxidative stress and reactive oxygen species ROS induce cell and tissue damage through apoptosis [31,32,33]. The liver is the main site of ROS generation. Oxidative stress influences inflammation, metabolism, proliferative liver disease, and chronic liver disease [34,35]. Enzymatic and non-enzymatic antioxidant systems are essential for cellular responses and regulate oxidative stress under physiological conditions [14,36]. Antioxidant enzymes like catalase, superoxide dismutase, and glutathione peroxidase are indicators of oxidative stress [37]. Yin et al. found that early weaning inhibited the activity of plasma SOD and CAT of piglets [38], which is consistent with our findings that the activity of SOD and CAT in liver of WG piglets is significantly lower than that of SG piglets, suggesting that weaning affects antioxidant enzymes activity in piglets.The decrease in the expression of antioxidant mRNA in piglets is an important reason for the decreased antioxidant capacity of piglets [39]. Feng et al has reported that low-birth-weight caused decreased mRNA expression of GSH and SOD, thus caused decreased antioxidant capacity in the liver of piglets [40]. In this study, weaning down-regulated gene expression of GSH in the liver of piglets on the 3rd day after weaning. However, the enzyme activity of GSH did not change significantly after weaning, the reason for which might be that the modification and activation of antioxidant proteins are affected when piglets are under weaning stress [41,42].Nrf2 signaling is a research hotspot in oxidative stress [43,44] and is the main regulator of cellular redox levels. Nrf2 is reported to be involved in transcriptional activation of ARE response genes, including glutathione S-transferases (GSTs), γ-glutamylcysteine synthetase (γ-GCS), heme oxygenase-1 (HO-1), antioxidants, proteasomes, and drug transporters [45,46]. Nrf2 and Keap1 form a conjugate, and then Nrf2 dissociates from the conjugate and stably translocates into the nucleus, driving expression of antioxidant enzymes genes [47,48]. In the oxidative stress model induced by diquat, Nrf2 signaling is essential for regulating the redox state of cells [49,50]. The results in this study showed that weaning caused the down-regulated expression of Nrf2 and NQO1 in the liver of piglets on the 3rd and 7th day after weaning, respectively.5. ConclusionsIn conclusion, early weaning reduced piglet growth, liver index, and liver antioxidant enzymes activity. Our findings indicate that weaning reduces the antioxidant capacity of liver in piglets.

animals : an open access journal from mdpi

10.3390/ani13091538

PMC10177403

Mycotoxin-contaminated feed is a current, worldwide problem. Mold-contaminated cereal is rich in heat-resistant and harmful metabolites such as fumonisins. The amount of fumonisins consumed as part of animal feed, including livestock feed, is unknown, and thus their influence on animal health, the costs associated with veterinary health care and the associated economic losses in breeding are impossible to calculate. Prenatal exposure of fetuses to fumonisins, through exposure of the mothers to fumonisins during pregnancy, is an important problem during breeding, often resulting in disturbed development of offspring in the postnatal period as a result. Thus, studies focused on the effects of fumonisin-exposure during the prenatal period on postnatal development are required.

Animal feed is very frequently contaminated with different types of mold, the metabolites of which are toxic to living organisms. Mold-contaminated cereal is rich in heat-resistant and harmful metabolites such as fumonisins (FBs). The amount of FBs consumed as part of animal feed, including livestock feed, is unknown. Therefore, this study aimed to evaluate the effects of maternal oral FB intoxication on basal duodenum morphology and the immunolocalization of gut hormones responsible for food intake (leptin and ghrelin), as well as their receptors, in newborn rat offspring. Pregnant Wistar rats were randomly allocated to one of three groups: a control group or one of two FB-intoxicated groups (60 or 90 mg FB/kg b.w., respectively). Basal morphological duodenal parameters changed in a dose- and sex-dependent manner. The intensity of the ghrelin immunoreaction was unchanged in females, while in males it increased after FB exposure (60 mg/kg b.w.), with a simultaneous decrease in expression of the ghrelin receptor. Leptin and its receptor immunoreaction intensity was decreased in both sexes following FB exposure. The current study highlighted the potential involvement of intestinal ghrelin and leptin in the metabolic disturbances observed later in life in offspring that were prenatally exposed to fumonisins.

1. IntroductionAnimal feed is very frequently contaminated with different types of mold, the metabolites of which are toxic to living organisms [1]. Fusarium is a common genus of imperfect fungi which produces heat-resistant fumonisins [2,3]. Although Fusarium produces the A, B, C, and P types of fumonisins, type B (FB) is the most common environmental threat, which includes two types: FB1 and its cytotoxic analog, FB2. Both types of FB inhibit sphingosine N-acyltransferase [4], and in addition, FB2 specifically inhibits the protein serine/threonine phosphatase. [5]. They disturb sphingolipid metabolism [6], and besides their carcinogenicity, they are also nephrotoxic and cause hepatocarcinogenesis, immunosuppression and neurotoxicity [7,8,9]. Signs of FB-induced intoxication differ depending on the animal species, their age and sex, as well as on the FB dose, duration, and route of FB exposure. There are observed non-species-specific symptoms in sheep, pigs, poultry, horses, and rodents; and organ-specific symptoms in the lungs and esophagus of pigs or in the brains of horses [10,11,12]. Some animals, such as ruminants, are resistant to the effects of FB intoxication, since the rumen microbiota degrade the mycotoxins [13]. Assessment of the dietary intake of FBs is difficult since (1) their concentration in feed components is unknown, (2) infrastructure is contaminated, and (3) FBs are also produced by Aspergillus niger, a mold which is commonly found in soil, water and fecal matter [14]. Although the acceptable level of FBs in feedstuff is regulated by the EU Commission legislation within the European Union [15,16] and by the FDA in the USA [17], complete control of the amount of FBs in animal feed is impossible and the effects of FB exposure are still not fully known. Economic costs of veterinary care and livestock losses due to exposure to mycotoxins are impossible to calculate, because animals are sometimes already exposed to FBs prenatally and the subsequent FB-induced developmental disturbances result in many negative long-term effects later in life. Prenatal FB exposure results not only in the altered development, structure, and function of bone or the gastrointestinal tract (liver and intestines, including the enteric nervous system), impaired hematopoiesis, and skeletal muscle inflammation, it also leads to endocrine alterations, resulting in disproportional development [18,19,20,21,22,23].Considering the indirect impact of maternal FB exposure during pregnancy on fetal development and the resulting health problems in offspring during postnatal life, it is crucial to investigate the effects of FB exposure during pregnancy not just on intestinal development but also on the gut hormones that regulate food intake in newborn offspring. It is in line with the hypothesis of the prenatal origin of health and disease according to fetal programming [24,25]. New evidence suggests that prenatal exposure to FBs could lead to the development of overweight and diabetes by altering metabolism [19], involving changes in hormonal gastrointestinal regulation of food intake.We hypothesize that prenatal FB exposure not only affects the intestinal morphology but also can alter the immunolocalization of basal gut hormones involved in food intake, resulting in postnatal developmental impairment.Therefore, the objective of the present study was to evaluate the impact of maternal oral FB intoxication on the small intestine of newborn rat offspring. Specifically, the study analyzed changes in the basal duodenum morphology and immunolocalization of gut hormones responsible for food intake (leptin and ghrelin) and their receptors. These analyses are essential for gaining a better understanding of the consequences of FB exposure during pregnancy on the disproportional postnatal development of rat offspring.2. Materials and MethodsThe animal study protocol (#132 676-Adm/08/2020) was approved by the Institutional Review Board of the State Scientific Research Control Institute of Veterinary Medicinal Products and Feed Additives (SCIVP) in Lviv, Ukraine. The experiment was performed in accordance with the EU Directive 2010/63/EU.2.1. Preparation of Fumonisins Stock SolutionFBs were produced in vitro using F. moniliforme on a maize grain medium, as previously described [26]. Briefly, the process included inoculating autoclaved, coarsely cracked grains with F. moniliforme cultures and allowing them to cultivate for 4 weeks at 24 °C. The resulting contaminated maize was then autoclaved, dried, ground, and subjected to liquid chromatography to analyze FB1 and FB2 levels, which showed a ratio of FB1 to FB2 of 3:1 (73% to 27%). FBs were subsequently extracted from the ground grains using ethanol solution, quantified with an ELISA kit (R-Biopharm, Darmstadt, Germany), and concentrated to a 100 mg/mL FBs extract stock solution. During the experiment, to achieve the required concentrations in 0.5 mL, the FB extract stock was diluted in a 0.9% saline solution based on daily measurements of the body weight of individual pregnant dams.2.2. Animals and Experimental DesignPregnant Wistar rats, at the age of five weeks (n = 18), were housed individually in polypropylene cages (380 × 200 × 590 mm). The rats were kept at a temperature of 21 ± 3 °C and a humidity of 55 ± 5% under a 12 h light/dark cycle. A standard diet for laboratory rat was provided ad libitum, with unlimited access to water. After an acclimatization period, the pregnant dams were randomly assigned to one of three groups (with six rats in each group): a control group (the FB 0 group) or one of two FB-intoxicated groups, which received FBs at a dose of either 60 mg FB/kg b.w. (the FB 60 group) or 90 mg FB/kg b.w. (the FB 90 group), respectively. FBs at a dose of 90 mg FB/kg b.w., which is equivalent to 1/10 of the established LD50 value, is sufficient to cause sub-clinical intoxication in adolescent rats, while the 60 mg FB/kg b.w. dose, which is equivalent to 1/15 of the established LD50 dose, does not elicit clinical or subclinical signs in adolescent rats [18,20,27]; however, both doses resulted in higher body weight at weaning with probable insulin resistance noted mainly following the 60 mg/kg b.w. dose [19]. The two doses are routinely used in developmental studies (omitting the first 6 days of pregnancy) carried out at SCIVP, Lviv, Ukraine [27,28].The FB mixture, diluted from FB extract stock in 0.5 mL of 0.9% saline, was given to the dams daily from the 7th day of pregnancy to parturition, intragastrically via oral gavage. Pregnant control animals were administered saline solution at the same volume. All pregnant females were under veterinary care. After natural parturition, all newborn offspring were allocated to the following groups, in line with their mothers: the FB 0, FB 60 or FB 90, and two newborn offspring from each mother (one male and one female, n = 6 males and n = 6 females in total) were weighed and euthanized (CO2 inhalation). Immediately after euthanasia, the duodenum was carefully dissected out and fixed in 10% buffered neutral formalin solution for morphometrical assessment and immunohistochemical evaluations.2.3. Tissue Sampling and Morphometrical AnalysisAfter being fixed in 10% formalin for 24 h, intestine samples were washed in water, dehydrated in a series of graded ethanol solutions (30, 50, 70, 80, 90, 96, and 99.8%). They were then paraffin-embedded, cut into 5 µm sections, mounted on a microscope slide and stained with Goldner’s trichrome staining, a histological staining technique used in pathology to visualize and differentiate various tissue components. It typically stains collagen fibers green, muscle fibers red, and nuclei black or brown, allowing for the identification and assessment of connective tissue, muscle tissue, and cell nuclei in histological sections [29,30]. The specimens were observed under standard bright illumination with an optical microscope (BX63 and CX43, Olympus, Tokyo, Japan) using the 20× objective. The length and thickness of intestinal villi, as well as the length and thickness of the crypts, thickness of the villus epithelium, mucosa, submucosa, and both myenterons (longitudinal and circular muscle layer), and the number of villi and crypts per mm of mucosa, were measured as described in detail previously [21,28] (Figure 1). The villus height/crypt depth ratio was calculated for individual villi–crypt pairs. The number of enterocytes and goblet cells per 100 µm of villus was also determined [31,32]. Special attention was paid to the correct orientation, and only sections containing longitudinally cut villi and crypts were included in the analyses. For each parameter, 10 replicate measurements were made per animal. The analyses were performed using ImageJ software (version 1.51k) [33].2.4. Leptin, Ghrelin and Their Receptors ImmunostainingThe immunohistochemical (IHC) stain was performed with rabbit anti-leptin, anti-leptin receptor, anti-ghrelin, and anti-ghrelin receptor as primary antibodies (#DF8583, #DF7139, #DF6389, #DF2794, respectively, all Affinity Biosciences, Jiangsu, China), diluted (1:300) in Diamond antibody diluent (Cell Marque Corp., Rocklin, CA, USA). The sections were then processed with a ready-to-use, two-step detection system of poly-HRP-anti-mouse/rabbit IgG (#IM-DPVB110HRP, BrightVision, Immunologic WellMed B.V., Duiven, Netherlands), developed in DAB (#D5905, Sigma-Aldrich, St. Louis, MO, USA) and counterstained with Mayer’s hematoxylin (Patho, Mar-Four, Konstantynów Łódzki, Poland) [34]. The IHC images, acquired using a bright-field microscope (BX-61, Olympus, Tokyo, Japan) and the 20× objective (200× total magnification), were analyzed using an ImageJ-compatible plugin for analyzing cytoplasmic staining patterns by assigning a histogram profiling and scoring the deconvoluted DAB images using IHC Profiler [35], an open-source plugin for semiquantitative and quantitative evaluation and automated scoring of immunohistochemistry images. An optical density (OD) quantitative score was calculated according to algebraic formula previously described by Jafari and Hunger [36].The measurements were carried out separately for villi and crypts in five randomly selected areas of the positive signal by an associate who was blinded to the treatment.2.5. Statistical AnalysisThe data collected in the study were analyzed using the statistical software system STATISTICA (version 13, StatSoft, Inc., Tulsa, OK, USA) and GraphPad Prism (version 9.5.1.733, GraphPad Software, San Diego, CA, USA). An experimental unit consisted of a single rat, so measurements were averaged per animal before statistical analysis. The data for males and females were analyzed separately, and all data were confirmed to be normally distributed through the Shapiro–Wilk normality test. The analysis was conducted using a one-way analysis of variance (ANOVA) based on the model: Yij = m + Di + eij, where Y represents a single observation, m represents the general mean, D represents the effect of the FB exposure (i.e., 0, 60, and 90 mg FB/kg b.w.), and e represents the error. Post hoc comparisons were conducted using the Tukey HSD test. Additionally, to determine the linear and quadratic effects among the means, an orthogonal contrast analysis was conducted.3. Results3.1. Duodenal Basal MorphologyBasal morphological changes after prenatal/maternal fumonisin exposure in female newborn rat are presented in Table 1. Maternal FB exposure, irrespective of the dose, resulted in a decrease in the thickness of the longitudinal lamina, compared to that observed in the control group, with both linear and quadratic effects being statistically significant (linear and quadratic, p < 0.001 and p = 0.002, respectively). The thicknesses of circular muscle layer, villus and villus epithelium decreased after prenatal FB exposure, irrespective of FB dose, compared to the control group (linear, p = 0.008, p = 0.001, and p < 0.001, respectively). Total crypt number was significantly lower in newborn offspring exposed to fumonisins at a dose of 60 mg/kg b.w., compared to that of the other groups (linear, p < 0.001). No other changes in serum parameters were observed.In the case of male newborn offspring, prenatal exposure to FBs resulted in a trend of decreased thickness of the longitudinal muscular lamina (p < 0.001). Moreover, crypt depth (p = 0.025), goblet cell number per 100 µm of villus (p < 0.001), villi length/crypt depth ratio, and villus epithelium thickness (p < 0.001) decreased linearly, while mucosa thickness, crypt thickness, and villus length decreased linearly and quadratically (p = 0.028 and p = 0.024; p < 0.001 and p = 0.008, respectively) following FB exposure (Table 2).3.2. Ghrelin, Leptin, and Their ReceptorsAs shown in Figure 2, the optical density (OD) of the ghrelin immunoreactivity (IR) in crypts and villi of female newborn rat offspring significantly decreased following prenatal exposure to 90 mg/kg b.w. FB (linear contrast in crypts and villi, p < 0.001 for both), with a greater decrease in crypts when compared to the control group. The OD of the ghrelin receptor IR in crypts showed a significant decrease regardless of the FB dose, with a linear and quadratic trend in crypts (p < 0.001 for both) and a linear trend in villi (p < 0.001). The OD of the leptin IR decreased following FB exposure in crypts (linearly and quadratically, p < 0.001 for both), while in villi the OD of the leptin IR decreased linearly (p < 0.001), showing a significant difference only for the FB 90 group when compared to the control. The OD of the leptin receptor IR decreased linearly in crypts (p < 0.001) with the lowest value observed in the FB 90 group, while in the villi the decrease was linear and quadratic (p < 0.001 for both), but without a difference between the FB-intoxicated groups.Figure 3 shows representative images and results of quantitative analysis of the OD of IR observed in male offspring. The OD of the ghrelin IR in crypts showed a quadratic trend (p < 0.001), with the highest OD observed in the FB 60group. The OD of the ghrelin IR in villi showed the highest value in the FB 60 group as well, but both a linear (p = 0.033) and quadratic trend (p < 0.001) were observed. The OD of the ghrelin receptor IR in crypts and villi significantly decreased following prenatal exposure regardless of the FB dose. The decrease showed a linear trend (p < 0.001, for both) and a quadratic trend (p < 0.001 and p < 0.05 for crypts and villi, respectively). In the crypts of the duodenum of newborn male rat offspring, no effect of prenatal exposure to FBs was observed on the OD of the leptin IR. In the villi, a quadratic trend was noted (p < 0.001), with the lowest OD observed in the FB 60 group. The OD of the leptin receptor IR significantly decreased in the crypts and villi of the FB 90 group when compared to that observed in the FB 0 and FB 60 groups (p < 0.05).4. DiscussionFumonisins have been shown to inhibit lipid biosynthesis-related proteins (ceramide synthase is inhibited by all fumonisins, whereas the serine/threonine phosphatase is only inhibited by FB2), resulting in disturbances in sphingolipid synthesis, the biological effects of which include altered cellular differentiation, cell proliferation, programmed cell death, apoptosis, and inflammation [18,19,27]. Maternal nutrition has been shown to be a very strong determinant of the prenatal formation of all fetal structures, as well as their function, which in turn influences all aspects of postnatal life [24,25]. Postnatal functioning of living organisms is dependent not only on the postnatal development of all the structures and systems of the organism (skeletal system determining movement or gastrointestinal system determining the whole metabolism) which in turn influences physical activity, but also on their mental health, which determines their position in the environment and in their peer and social groups. Thus, prenatal nutrition could be a factor that will determine which traits will be preserved and passed on epigenetically to the next generation. Many harmful or toxic substances, including fumonisins, are often present in the diet of animals, including that of pregnant animals [12,37]. To address this concern, the European Commission has released recommendations on the acceptable levels of fumonisins in maize and animal feed [16]. Prenatal development is characterized by a period of developmental plasticity during which the formation and maturation of many internal organs and systems occurs, which are crucial for postnatal health [25,38,39]. Fumonisins cross the placental barrier and exert many negative effects, dependent on the dose, duration of exposure, and time at which the fetuses are exposed [40,41]. During the current study, fumonisins were administered from the 6th day of the rats’ pregnancy, when the small intestine, including the duodenum, is not yet formed [42]. Smooth muscle and epithelial cells of the small intestine are not yet differentiated on day 15.5; however, the intestinal tight junctions are formed. Circular and longitudinal muscular layers develop on day 16.5, and on day 17.5 the first duodenal villi appear, epithelial cells differentiate to enterocytes, and the first endocrine cells develop. The villi epithelium is complete on day 18.5, and at the same time goblet cells appear. The last 3 days before birth are crucial for completion of intestinal development, and the intestinal mucosa becomes comparable to that found postnatally, with a significant increase in intestinal villi and microvilli in the brush border. On day 21, villi shape and size are comparable to that characteristic of postnatal structures [42]. Newborn rats have a short intestine (20 cm), weighing about 0.066 g, and during the first 21 days of postnatal life there is a large increase in intestinal weight (about eightfold) and a twofold increase in length [43]. The duodenum is the shortest and most proximal segment of the small intestine and the initial site of contact of food with the gastric secretions, bile and digestive enzymes. The duodenum plays an important role in the alkalinization of gastric chyme, which is important in the prevention of mucosal damage, because duodenal luminal pH fluctuates between 2 and 7 due to the presence of gastric acid. The acid is neutralized not only by bicarbonate, but also by the mucus produced by the goblet cells. The duodenum is also involved in the control and regulation of digestion, absorption, and motility, since it produces hormones which are critical for the coordination of all these processes [44].The current study was designed to take place during the period of development when the most important alterations occur in the intestine, and the results showed that prenatal fumonisins exposure reduced mucosa thickness in a sex-dependent manner. This FB-induced change could not only influence secretion, digestion, and absorption, but also the protective functions of the intestine, involving non-specific (defense against injury) and specific immunity (involving mucosa-associated lymphoid tissue) [45]. The negative effects of prenatal fumonisin exposure are still not fully known and there are very few previous studies which present such effects, making it difficult for us to discuss them.The only available study on rats prenatally exposed to fumonisins at the same doses as those used in the present study (60 and 90 mg/kg b.w./day) found that the reduced mucosal thickness observed in the duodenum of offspring at birth (such as that observed in the rats in the current study) seems to be temporary, since it disappears at weaning [21]. However, this study did not differentiate between males and females upon weaning. Another previous study observed no effect of FB exposure on mucosal thickness; however, it was performed on adolescent male rats [28].The smooth muscle fibers of the muscular lamina are arranged in two different layers (the inner circular and outer longitudinal) and determine the spiral passage of chyme to the periphery. Physiologically, the muscular lamina is the thickest in the proximal small intestine and it thins as it moves away from the pylorus. The present study only focused on the duodenum, where a thinning of both muscular layers was noted. Although we are not sure how prenatal FB exposure affects the other segments of the small intestine in rats, we can assume that it could lead to disturbances in the mixing of the chyme with digestive juices and in the movement of the chyme forwards, ultimately leading to incomplete emptying of the intestine. Moreover, it is unknown how prenatal exposure to fumonisins and the consequent thinning of the muscular lamina noted in female newborn rats could influence the basic electrical rhythm (BER), triggered in the interstitial pacemaker cells of Cajal, as well as the other physiological properties of the duodenum, such as excitability and contractility, in which the muscular layer plays a significant role [46].Crypts play a role in the renewal and regeneration of the villi epithelium, because immature enterocytes undergo proliferation and then rise from crypts and move along the villi and further undergo exfoliation on the villus’s tip. When proliferation and exfoliation are in equilibrium, proper villi length and shape are maintained [47]. Changes in villi morphology at weaning, when shortening is observed, have been shown to be temporary and are associated with the change in feeding from liquids to solids [48].The current study showed that prenatal fumonisin exposure resulted in a reduction in the length and thickness of the villi. These results are in agreement with a previous study with a similar study design, where narrower villi were observed in weaned rats following FB exposure [21]. Similar effects were observed in previous studies, with shorter villi observed in adolescent male rats [28] and villi atrophy in laying hens [26] following FB exposure. Other studies also confirmed the negative effects of fumonisins on intestinal villi, irrespective of the species or fumonisin dose [49,50,51,52,53]. Additionally, a linear decrease in villi/crypt ratio with an increase in fumonisin dose could indicate villi atrophy [54].The current study not only showed a reduction in thickness and narrowing of the newborn rat villi, but also narrower and more shallow crypts. Taken together, our results are in agreement with previous studies [26,28] and indicate that fumonisins negatively impact cell proliferation and possibly intensify cell death, leading to extensive cell exfoliation on the villi tips and shortening [53,55].Crypts are important for non-specific immunity since Paneth cells are located inside them, which produce many different bacteriocidial secretions, as well as specific endocrine amine precursor uptake and decarboxylation cells, which produce various biogenic amines or hormones which act in a paracrine or endocrine manner [56]. The secretory role of the crypts could be disturbed by prenatal fumonisin exposure, since a decrease in the number of crypts was observed in the newborn rat offspring following FB exposure. The same effect was also previously noted in weaned rats following FB exposure [21]. Taking into account the results observed in weaned rats from a previous study and the results of the current study, it seems that the reduction of the crypt number is permanent. However, this effect was not observed in adolescent rats directly intoxicated with fumonisins [28].A decrease in the number of goblet cells and villi epithelium thickness was also observed in the current study. The villi epithelium, together with the microbiota, immune cells, goblet cells, and tight junctions, constitute the intestinal barrier [57]. Although a reduction in villi epithelium thickness has not been observed in previous studies, a reduction in the number of goblet cells has been noted and is in agreement with our results [28,51,52].The proper development of intestinal structures is necessary for further general growth and development of organisms, and any changes in normal intestinal development can be associated with disease and impaired growth later in life [58]. As previously reported, newborn offspring from dams exposed to fumonisins, at a dose of 60 or 90 mg/kg b.w., weighed significantly less compared to control offspring, and the effects were sex-dependent. Prenatal FB exposure reduced term body weight in males, irrespective of fumonisin dose, while in females this effect was only observed after exposure to the higher FB dose (90 mg/kg b.w.) [18]. This fumonisin-induced effect on body weight and the weight of all vital organs was still observed at weaning in a sex-dependent manner [19]. Additionally, the disproportionate neonatal development observed in offspring at weaning, following maternal fumonisin exposure, has been linked with hormonal dysregulation, including that of growth hormone and insulin. Taking into account the disproportional growth, final body weight, glucose, and insulin concentration, the risk of diabetes mellitus and obesity should be considered [19].Gut hormones, leptin and ghrelin, play a role in the regulation of feed intake. Ghrelin is an orexigenic hormone, synthesized in the stomach and other segments of the small intestine, and released directly into the bloodstream, although during the fetal period it is also expressed in the pancreas [59,60]. Ghrelin has been shown to stimulate appetite and worsen glucose tolerance in rats through a reduction in insulin sensitivity, leading to an increase in plasma glucose [61]. Food intake rapidly decreases plasma ghrelin concentrations and a decrease in ghrelin concentrations are accompanied by a simultaneous increase in leptin concentrations during obesity. However, lowered plasma ghrelin concentrations are sufficient to maintain ghrelin’s blood glucose-enhancing effect and it can even be paradoxically enhanced in obesity. This effect is considered a physiological adaptation to this state involving these two hormones [62]. Leptin, which is also produced in the stomach, has various biological effects [63]. It not only initiates puberty, participates in inflammatory and immune responses, angiogenesis, hematopoiesis and bone formation, but also plays a regulatory role in energy homeostasis. Leptin is important in the control of body weight through the stimulation of metabolic rate and the suppression of food intake, which has also been demonstrated in rodents [64].Since ghrelin and leptin play a regulatory role in food intake through the gut–brain axis [65], it is important to know what effects prenatal fumonisin exposure has on the immunolocalization of leptin and ghrelin and their receptors in the duodenum in newborn offspring, which was previously unknown.The current study showed that the OD of the ghrelin IR was unchanged in females and increased in males following FB exposure at a dose of 60 mg/kg b.w. with a simultaneous decrease in the ghrelin receptor. On the other hand, the OD of the leptin IR and that of its receptor, decreased in both sexes. Ideally, the mRNA gene expression and Western blot analyses of protein levels should also be performed, as well as the measurement of the concentration of these hormones in peripheral blood, to comprehensively examine all potential mechanisms of prenatal FB exposure and its effects on postnatal animal development [19], which could be in line with the Developmental Origin of Health and Disease theory which suggests that disturbances in the intrauterine environment can increase the risk of developing various diseases (such as diabetes and obesity) later in life [66].To support these findings, further studies should be conducted to investigate whether prenatal exposure to fumonisins can trigger obesity and diabetes. Given the size of the newborn rats and the amount of samples we were able to collect, future studies should likely involve larger animals.5. ConclusionsThe present study is the first to suggest that disturbances in the levels of intestinal ghrelin and leptin and their receptors could contribute to the development of metabolic abnormalities in offspring, at a later stage of life, as a result of prenatal exposure to fumonisins. Due to the common presence of fumonisin in feed and the widespread risk of prenatal intoxication, research should be continued, including studies of other hormonal regulators of food intake.

animals : an open access journal from mdpi

10.3390/ani11092609

PMC8472471

The effect of partial or complete substitution of soybean oil (SO) by poultry fat (PF) on growth, nutrient digestibility, plasma lipids, and the pectoral muscle content of fatty acids (FAs) was examined in this study. Dietary PF supplementation improved breast muscle FA profile but did not affect muscle vitamin E content and liver thiobarbituric acid reactive substances (TBARS). By adding PF to the diet, economic efficiency was greatly improved in a dose-dependent manner. Therefore, the results of this study revealed that PF could be used as a partial or total replacement of SO in broiler nutrition without affecting their performance or physiological response with a tendency to improve their meat products.

Continuous genetic improvements of commercial broiler strains has led to the necessity of using fats in their rations to fulfill a large portion of the energetic requirements. Several fat sources have been introduced in poultry nutrition, such as rendering poultry fat (PF) an available and cheap lipid source compared to conventional sources such as soybean oil (SO). The present study investigated the effect of partial or full replacement of SO by PF on performance, nutrient digestibility, blood lipids, and fatty acids (FAs) content of pectoral muscle. Four hundred and eighty one-day-old male Ross-308 chicks were distributed into four experimental groups (12 replicates each): the first group (control) was fed a diet formulated with soybean oil as a fat source while the second to fourth groups (PF25, PF50, and PF100) were fed diets formulated with 25, 50 and 100% of PF as a fat source instead of SO. Results revealed no synergistic effect between SO and PF in any of the studied parameters. Replacing SO by PF did not alter birds’ growth, carcass characteristics, and plasma indices of birds. Abdominal fat% was increased (p < 0.01) in PF50 and PF100. Dry matter digestibility was improved (p < 0.05) in PF50 and PF100, while crude fat and protein digestibility was not affected. Contents of palmitic and docosahexaenoic acids in the pectoral muscle of PF50 and PF100 were reduced (p < 0.01) while concentrations of oleic and linolenic acids, total unsaturated FAs, and polyunsaturated FAs/Saturated FAs ratio were elevated (p < 0.05) in the same groups. Liver thiobarbituric acid reactive substances (TBARS) and muscle vitamin E contents were not altered. The dietary addition of PF greatly improved economic parameters. In conclusion, PF can be used as a lipid source in broiler diets to produce inexpensive meat while maintaining its growth performance.

1. IntroductionThe metabolizable energy requirements of modern commercial broiler strains are increasing with continuous improvements in their performance and the swift evolution of the intensive poultry industry. Additionally, due to the limited amount and incessant rising in prices of conventional energy sources feedstuffs, it has necessitated the search for alternative materials to ensure the future profitability of poultry production [1]. Lipid sources (oils and fats) are commonly included in broiler feeds to fulfill their high energy requirements. Fats added to broiler diets can improve the absorption and digestion of fat-soluble vitamins and other nutrients, enhancing their growth performance under armor neutral and heat stress conditions [2,3]. However, the prices of different lipid sources vary, and their utilization also varies depending on their physical and chemical properties [4]. The chain length of fatty acids (FAs) and their saturation degree affect the digestibility and metabolizable energy of the fat source [5]. Although animal fats are cheaper than vegetable oils, it is generally believed that the nutritional value of the former is less than that of the latter. Furthermore, among animal fats, poultry fat (PF) can be utilized by poultry species at a higher rate than tallow and lard oil [6].In the Middle East, soybean oil (SO) is the most common fat source used in poultry diet formulation, particularly after several hydrations, filtration, and degumming [7]. However, limited supply and consistently higher prices are anticipated for it. Rendered poultry fats, obtained from processed wastes of poultry slaughterhouses, can be sustainable alternatives to SO due to their wide availability and relatively low price. Moreover, dietary addition of PF has other benefits, including reducing dustiness, improving feed texture, increasing palatability, and enhancing nutrient absorption by reducing digesta rate of passage through the gut. However, rancidity and low utilization, particularly in young chicks, are disadvantages of using animal fats in poultry diets [8]. No differences in performance parameters were noticed in the literature when PF was used instead of SO [9,10]. However, it’s worth noting that a synergistic effect was observed when vegetable oils were mixed with animal fats such as PF and tallow [6,11].PF is commonly used in feed mills; however, the prohibitions and specific rules of the European Union regarding this practice have caused problems for these facilities with the EU countries [12,13,14]. Nevertheless, these by-products must be reintroduced into production and economy because of their high nutritional content, economic value, large production cost, and the high cost of alternative implementation. Therefore, the legal regulations and prohibitions put into action must be reconsidered to keep pace with the continuous improvement in production, use, and sales following current scientific developments.The present study aimed to investigate the potential synergistic impact of mixing SO with PF in broiler diets as well as the effect of full replacement of SO by PF on performance parameters, nutrient digestibility, plasma lipids, and contents of fatty acids and α-tocopherol in pectoral muscle and TBARS in hepatic tissues.2. Materials and Methods2.1. Ethical StatementThe study was approved by the Ethics Committee of Local Experimental Animals Care Committee and conducted following the guidelines of Kafrelsheikh University, Egypt (Number 4/2016 EC). All precautions were followed to minimize suffering during the entire experimental period.2.2. Chemical Analysis of Fat SourcesPoultry fat by-product and soybean oil were supplied by the Al-Sabeel Al-Gadidah Company (Tanta, Al-Gharbia, Egypt). Poultry fat was produced from the processed waste of poultry slaughterhouses (feathers, non-edible viscera, feet, head, blood, etc.) in a rendering unit with a batch-cooker and fat presser at the Fat Hanz Company (Tanta city, Egypt). Poultry fat was assessed for Escherichia coli and Salmonella spp., and it was incorporated into the diets only after confirming the absence of these pathogens. Metabolizable energy, peroxide value, and fatty acids profile of SO and PF were determined according to the procedures described by AOAC [15], Abd El-Moneim and Sabic [1], and Abd El-Moneim, et al. [16]. Values of the thrombogenic and atherogenic indexes were estimated following the equations of Ulbricht and Southgate [17]:(1)Atherogenic index=C12:0+4∗C14:0+C16:0Sum of unsaturated FAs (2)Thrombogenic index=C14:0+C16:0+C18:00.5∗OA+0.5∗(MUFA−OA)+0.5∗n6PUFA+3∗n3PUFA+(n3PUFAn6PUFA) where OA = oleic acid (C18:1), MUFA = monounsaturated FAs, PUFA = polyunsaturated FAs.2.3. Experimental DesignA total of 480 one-day-old male Ross-308 broiler chicks (43 g) were allocated into 48 ground pens and distributed equally into four experimental groups (12 replicates each). Pens, the stocking density of which was ten birds/m², were equipped with an automatic nipple cup drinker and a chain feeder system. The starter (0–10 d), grower (11–24 d), and finisher (25–35 d) experimental diets (Table 1) were formulated to meet the recommendation of Aviagen [18] for male broilers. The first diet (basal diet; control) was formulated using soybean oil as a fat source. In the second to fourth experimental diets, 25, 50, and 100% of the soybean oil were replaced by poultry fat. Diets and freshwater were offered ad libitum to the birds. The feed trial took place in a temperature-controlled chamber, started from 33 ± 1 °C and decreased by one degree per 3 days until reaching 24 ± 1 °C and kept till 35 days of age, with a proportional humidity between 50% and 70% and a 22:2 h light: dark cycle. Mortalities were recorded throughout the experimental phases.2.4. Growth Performance and Organ WeightsInitial and final body weight and feed consumption were measured individually on a pen basis. Feed conversion ratio (FCR) was calculated as a g feed:g gain. The European production efficiency factor (EPEF) was calculated as liveability (%) × body weight (g)/FCR × age (d). At the end of the trial (35 days), 48 birds (one bird/replicate; 12 birds/treatment) were individually weighed, slaughtered by cutting the carotid artery, and dissected to evaluate the relative weights of the thigh and breast muscles, abdominal fat, and liver [19]. Blood samples were collected with heparinized test tubes for blood biochemical analyses.2.5. Nutrient DigestibilityAt the end of the experiment (35 d), the twelve birds per group were weighed and individually caged in metabolic pens for collecting their excreta for four days. Before the commencement of this period, an adaptation period of 24 h had elapsed. Fresh water and diets were offered ad libitum to all birds during the manure collection stage. The approximate analysis of dry matter (#930.15), crude protein (#954.01), and crude fat (#920.29) of dried excreta and diets were performed according to AOAC [15]. The trichloroacetic acid procedure was used to estimate fecal nitrogen [20].2.6. Plasma Biochemical AnalysisAt 35 days, to separate plasma, collected blood samples were centrifuged (2500× g for 15 min at 4 °C). Plasma samples were kept at −20 °C pending analysis. Plasma high-density lipoprotein (HDL), total cholesterol, aspartate aminotransferase (AST), total protein, and albumin were calorimetrically evaluated using commercial kits and following the manufacturer’s instructions (Diamond Diagnostics, Cairo city, Egypt), using spectrophotometric analysis (Spectronic 1201; Milton Roy, Ivyland, PA, USA).2.7. Muscle and Liver Biochemical AnalysisThe analysis of superficial pectoral muscle FAs was conducted on 48 birds (1 bird per replicate; 12 birds per treatment) using gas-liquid chromatography (GLC) as described by [21,22]. The concentrations of muscle vitamin E [23] and liver thiobarbituric acid-reactive substances (TBARS) [24] were also determined.2.8. Economic EfficiencyAverage feed cost per bird was calculated as feed consumption per bird × cost of one kg diet (0.53, 0.52, 0.51 and 0.50 US $ for control, PF25, PF50, and PF100, respectively, considering the price of one kg of SO (1.47 US $) and PF (0.32 US $). Feed cost per kg gain was estimated by multiplying the cost of a one kg diet by the FCR. Total costs were measured by summing the feed cost/bird and all fixed costs, including housing labor, vaccines, drugs, day-old chick, disinfectant, veterinary supervision, etc. Subtracting total costs from the total return, considering the average price of the bird (1.72 US $ per one kg live body weight), was considered the net return. Benefit/cost ratio (B/C ratio) was estimated by the following equation: net return/total costs × 100 [25].2.9. Statistical AnalysisDifferences between the experimental groups were analyzed using one-way ANOVA was applied to determine the effects of replacing SO with PF, in which pens were the statistical units for performance parameters, birds for the carcass, organ weights, and samples for biochemical and other parameters, the General Linear Model package of SPSS (Version 19.0, Chicago, IL, USA). Tukey’s multiple range test was used to identify the significant (p < 0.05) differences among means of experimental groups.3. Results3.1. Chemical Analysis of Fat SourcesThe main difference between the FAs profile of SO and PF (Table 2) can be found in the ratio of unsaturated FAs over-saturated FAs (U/S ratio). The U/S ratio of SO (3.84) was 1.7 times that of PF (2.28). Concentrations of polyunsaturated (PU) FAs in SO were higher than that of PF, while the latter was richer in monounsaturated (MU) FAs. The PU/S ratio of SO was 2.4 times that of PF, while the MU/PU ratio of PF was 2.2 times that of SO.3.2. Growth Performance and Organ WeightsThe partial or full replacement of SO by PF did not alter the final body weight, feed consumption, and FCR of broiler chickens at marketing age (Table 3). The EPEF of PF groups was slightly higher than that of the control. The highest value of EPEF was recorded in PF100. Additionally, the dietary inclusion of different lipid sources had no significant impact on carcass percentage and the relative weight of thigh and breast muscles and liver of 35-day-old broiler chicks. The abdominal fat percentage was increased (p < 0.01) in PF100 and PF50 compared to control and PF25.3.3. Nutrient DigestibilityAs presented in Table 4, digestibility coefficients of dry matter were improved (p < 0.01) in PF50 and PF100 compared to the control. Digestibility coefficients of crude protein and fat were not influenced by dietary replacement of PF instead of SO.3.4. Plasma Biochemical AnalysisThe data presented in Table 5 shows the impact of partial or total replacement of SO by PF on plasma biochemical parameters of broilers at marketing age. Total protein, albumin, AST, HDL-cholesterol, and total cholesterol were not significantly affected by dietary inclusion of PF compared to the control.3.5. Muscle and Liver Biochemical AnalysisThe fatty acid profile of the pectoral muscle was influenced by the type of dietary fat source (Table 6). The main differences in the FAs profile of breast muscle can be found in the concentrations of certain FAs and the PU/S ratio. Concentrations of palmitic acid and docosahexaenoic acids were reduced (p < 0.01) in PF50 and PF100. However, levels of oleic acid, linolenic acid, total unsaturated FAs, and PU/S ratio were elevated (p < 0.05) in the pectoral muscle of birds fed 50% and 100% PF. Pectoral muscle concentration of α-tocopherol and content of TBARS in the liver were not significantly affected by replacing SO with PF (Table 6). However, numerical reduction in α-tocopherol and elevation in TBARS levels in PF treated groups were observed.3.6. Economic EfficiencyAs presented in Table 7, economic parameters were greatly influenced by partial and full replacement of SO by PF in broiler diets. Feed cost/bird, feed cost/kg gain, and total cost/bird were decreased (p < 0.001) in PF25, PF50, and PF100 compared to the control. Net return and benefit/cost ratio were increased (p < 0.001) in PF25, PF50, and PF100, and the highest values were recorded in group PF100.4. DiscussionThe analyzed composition of the lipids sources in the present study revealed that PUFAs in SO were higher than that of PF while the MUFAs in PF were higher. The U/S and PU/S ratios of SO were 1.7 and 2.4 times that of PF, while the MU/PU ratio of PF was 2.2 times that of SO. These findings are almost similar to NRC [26] and previous findings [8,27]. The predominant FA in PF was oleic acid followed by linoleic acid, while in SO, it was linoleic acid followed by oleic acid. This revealed that the two fat sources were rich in the unsaturated FAs with the superiority of SO by 12.5%, which may explain the effects of these lipids on studied parameters.Fats are a high-energy feedstuff commonly incorporated in the formulation of commercial poultry diets. Results of earlier studies investigating the impact of dietary addition of different lipid sources on poultry performance were equivocal. Several investigations have reported that supplementation of vegetable oils to poultry feed can improve their performance, carcass traits, and production efficiency by elevating the diet’s energy level better than animal fat [28,29,30]. Others reported a synergistic effect between vegetable oils and animal fat [6,11]. Nevertheless, some studies revealed non-significant differences between animal fat and vegetable oils [9,10,27]. In the present study, no synergy effect between PF and SO was noticed. We also found that partial and total replacement of SO by PF in broiler diets had no significant impact on final body weight, FCR, EFEF, and carcass traits except abdominal fat, which increased PF50 and PF100. A similar trend was reported by Okur [31] and Sanz [32], who noticed an elevation in abdominal fat weight when animal fats were used in broiler diets. The growth performance of broilers is greatly influenced by dietary fat sources and their FA profiles, particularly essential FAs, such as α-linolenic acid and linoleic acid, as their deficiency may retard broilers growth [33]. As our results revealed, the differences between SO and PF in these FAs were insufficient to induce significant differences in birds’ growth performance. The lack of differences in growth performance of birds fed diets with SO or PF could also be attributed to the equilibrium ratios of energy-to-protein and energy-to-amino acid in these diets [27,31]. Additionally, Pesti et al. [27] reported that feeding on fat sources with high metabolizable energy resulted in a high amount of fat being deposited. This might explain the increase of abdominal fat in birds fed diets with high levels of PF.In our study, replacing SO by PF improved dry matter digestibility while digestibility of crude protein and crude fat was not affected. These results are in line with previous findings [34,35]. Fatty acids chain length and U/S ratio greatly affect nutrient digestibility. Fat sources such as tallow and palm oil characterized by a low U/S ratio showed drastic negative impacts on nutrients digestibility [36,37]. Tancharoenrat et al. [35] noticed a reduction in crude fat digestibility for broilers fed diets with fat sources with low U/S ratio such as palm oil (U/S 0.93) and tallow (U/S 0.80) compared to SO (U/S 5.07).In the present study, the difference between the U/S ratio of PF (2.28) and SO (3.84) was lower than mentioned in the study of Tancharoenrat et al. [35], which might explain the insignificant changes in fat and protein digestibility. Nevertheless, they did not observe a significant difference for crude fat digestibility between PF- (U/S 2.07) and SO-based diets. The authors suggested that the changes in FAs composition of PF were not enough to exert a drastic effect on crude fat digestibility. Furthermore, dietary addition of PF could enhance nutrients’ digestion and absorption by reducing digesta rate of passage through the gut, which might explain the improvement in dry matter digestibility [8].To our knowledge, limited investigations have studied the impact of dietary addition of fat sources on the blood biochemistry of poultry species. The majority of these studies focused on the impact of fat types on the quality of animal products for human uses without studying their effect on birds’ health status during production [38]. In the present study, we evaluated the effect of replacing SO by PF on broiler chickens’ hepatic function and blood lipids. No significant alterations were observed in all studied parameters among experimental groups. Hu [9] noticed similar results who reported insignificant impact of dietary SO and PF on HDL levels- and LDL-cholesterol and total cholesterol in the serum of Cherry Valley ducks. Donaldson et al. [38] also noticed non-significant changes in serum levels of AST, total protein, albumin, and cholesterol of Japanese quail-fed diets with SO and lard. However, results of serum triglycerides as affected by various dietary fat sources and levels were somewhat contradictory. Some studies reported significant elevated serum triglycerides of humans and birds fed high dietary fat [39,40]. Others reported non-significant changes [9,41], while some observed a significant reduction in its level [38,42]. Donaldson et a. [38] attributed the decrease in serum triglycerides to a possible reduction in de novo synthesis of FAs in the liver as large amounts of FAs were being supplied to the birds via dietary fat sources. The lack of consistency between findings of these studies suggests some differences in lipid handling pathways and multiple potential mechanisms contribute to regulating serum concentrations of cholesterol and triglycerides between different avian species, including postabsorptive lipid metabolism and/or hepatic uptake of HDL-cholesterol.Consumers have become more concerned about the nutritional aspects, including the lipid profile and FA contents. Chicken meat with its low-fat and high-protein contents has been characterized as the main source of PUFAs [43]. Functional and beneficial foods that contribute to preventing chronic diseases, such as coronary heart disease and metabolic disorders, are characterized by higher concentrations of PUFAs [44,45]. It has been documented that FAs and lipid profiles of chicken meat can be modified by changing broilers’ feed composition [46,47]. Reducing SFAs and elevation of PUFAs contents in chicken meat would improve its nutritional value and quality [48]. Our results showed that pectoral muscle levels of palmitic acid and docosahexaenoic acids decreased while concentrations of oleic acid, linolenic acid, total UFAs, and PU/S ratio were elevated PF50 and PF100. These findings are considered positive as the reduction in PU/S ratio, thrombogenic index, and the atherogenic index, and the elevation of linolenic acid (n-3) are favorable in healthy and functional food for human consumption. The added value of n-3 PUFAs to human foods and their favorable impacts of on human health were investigated. Bostami et al. [47] reported the health benefits of long-chain n-3 FAs to animals and humans, such as reducing the risk of heart diseases and lowering the concentration of circulating cholesterol.Moreover, Pinchasov and Nir [49] documented that PUFAs can inhibit the activity of the 9-desaturase enzyme complex, which responsible for converting SFAs to MUFAs, thereby downregulating the synthesis of MUFAs. Furthermore, the reduction in the atherogenic and thrombogenic indexes in PF50 and PF100 is considered favorable, as Ulbricht and Southgate [17] recommended the low values of these indexes in healthy human diets. Generally, supplementation of PF in broiler diets instead of SO tends to improve the lipid profile of breast meat.Lipid oxidation, one of the major factors responsible for the deterioration of the quality of meat products, is primarily initiated in the UFAs of membrane phospholipids [17]. Vitamin E (α-tocopherol) plays a fundamental role in protecting these susceptible cellular structures against oxygen-containing free radicals and reduces their content of TBARS [50]. The primary location of vitamin E is within the biological membranes, such as mitochondria and microsomes, which allow its effective function compared to other antioxidants [17]. Therefore, increasing the muscle membrane content of α-tocopherol by dietary manipulation is required. Dietary fat sources generally contain high fat-soluble vitamins, including α-tocopherol, but they vary among themselves. Vegetable oils and most plant-origin feedstuffs are rich in vitamin E, while their content is lower than most animal products [51,52,53,54]. In the present study, levels of α-tocopherol in pectoral muscle and TBARS in the liver did not differ by replacing SO with PF. Numerical reduction in α-tocopherol and elevation in TBARS levels in PF treated groups were observed. These findings agree with those of Polycarpo et al. [55], who reported significant changes in hepatic contents of vitamins E and A (fat-soluble vitamins) of broilers fed on corn-based diets with SO or beef tallow as lipid sources. Lauridsen, et al. [56] also noticed that fat sources did not influence the concentration of vitamin E in muscle membranes. Contrarily, Dänicke, et al. [56] and Gatellier, et al. [57] observed higher hepatic vitamin A and muscular TBARS concentrations in birds fed diets with SO compared with tallow. The lack of significance in α-tocopherol concentrations in the present study might be attributed to the low incorporation levels of SO and PF in broilers diet, eliminating the added value of vitamin E to the feed.Furthermore, the insignificant differences in fat digestibility observed in this study may be considered another explanation as absorption of fat-soluble vitamins depends on fat digestibility and the emulsification process. Knarreborg, et al. [58] documented that good conditions of micelle formation and emulsion increase the bioavailability of α-tocopherol. The numerical reduction in α-tocopherol and increase in TBARS concentrations in the breast and liver tissues of birds fed on PF may be due to the relatively high susceptibility of broiler meat to lipid oxidation when fed diets incorporated with PF [59,60].As expected, the economic efficiency and benefit-to-cost ratio were significantly improved by dietary replacement of SO by PF in a dose-dependent manner. This effect is due to the large difference between the prices of SO and PF; under our study’s condition, the price of SO was 4.6 times that of PF. The differences in the prices of SO and PF are reasonable; since SO is one of the most commonly used fat sources in poultry rations and its various industrial uses. While PF is a cheap by-product of poultry slaughterhouses, its utilization is affordable and reduces its adverse impacts on the environment. Several studies have reported the economic benefits of using PF instead of SO [4,9,34,38].5. ConclusionsThe present study investigated the effect of partial or full replacement of SO by PF on growth performance, nutrient digestibility, plasma lipids, and vitamin E and FAs contents of the pectoral muscle of broilers. Neither a synergistic effect between SO and PF nor effects of the dietary changes on broilers’ growth, carcass parts, and blood biochemistry were noticed in the present study. Dietary supplementation of PF improved the FA profile of breast muscle but did not affect muscle content of vitamin E and liver TBARS levels. Economic efficiency was greatly improved in a dose-dependent manner by dietary addition of PF. Therefore, this study revealed that PF can be used as a partial or total replacement of SO in broilers’ nutrition without affecting their performance or physiological response to improve their meat products.

animals : an open access journal from mdpi

10.3390/ani11092516

PMC8466291

The climatic environment within calf housing can have an effect on calf health, but also on growth and performance. Calves have a lower threshold environmental temperature (lower critical temperature, LCT), below which can impact on the calf’s ability to maintain its core body temperature. This can cause the calf to partition more of its available energy into heat production and less into growth. The LCT decreases as the calf gets older. This year-long study followed 299 dairy-bred calves on one farm in Scotland from birth until approximately 28 days of age, and looked at the proportion of time for which the temperature was below the LCT for the individual calf, as well as the daily liveweight gain (DLWG; kg/d) of the calves during this time. For their first 6–14 days of life the calves were individually housed, and then subsequently group housed. Air temperature (°C), relative humidity (%), and wind speed (m/s) were recorded every hour of every day throughout the study, and calves were weighed regularly so that DLWG could be calculated. The study demonstrated that calves that spent a high proportion of their time below their LCT had a lower DLWG compared to calves that spent a low proportion of their time below their LCT.

Calf housing is naturally thermodynamic, with interactions between various elements such as wind speed, air temperature, and humidity. This study investigated the effect of the proportion of time for which calves were exposed to effective environmental temperatures below their lower critical temperature (LCT) on their daily liveweight gain (DLWG) within their first month of life. This study used the naturally occurring climatic environment, whereas other such studies have been conducted under climatically controlled conditions. Air temperature (°C), relative humidity (%), and wind speed (m/s) were recorded within the calf housing from birth until approximately 28 days of age, with calves being health-scored and weighed at regular intervals. Calves were housed from birth until 6–14 days old in individual hutches, and then moved into group housing igloo pens. Whilst individually housed, calves that spent less than 32% of their time below their LCT had a DLWG of 0.06 ± 0.34 kg/d (mean ± SE) compared to calves that spent more than 97% of their time below their LCT, which had a DLWG of −0.19 ± 0.045 kg/d. When group housed, calves that spent less than 1% of their time below their LCT had a DLWG of 0.59 ± 0.18 kg/d, whereas calves that spent more than 28% of their time below their LCT had a DLWG of 0.53 ± 0.23 kg/d. The proportion of time for which calves were exposed to effective environmental temperatures below their LCT had a significant effect on DLWG when calves were individually housed. Therefore, exposure to effective environmental temperatures below the LCT can be detrimental to the growth of the calf in the early stages of its life.

1. IntroductionFor the farmer, the main objective of rearing calves is to produce a healthy calf that is able to achieve target growth rates as economically as possible. It has been well documented that poor performance in an individual calf in the pre-weaning phase can affect its future productivity [1,2,3,4,5]. Therefore, calf growth—specifically, daily liveweight gain (DLWG)—is a key performance indicator for monitoring success in calf rearing. DLWG is influenced by colostrum provision [6,7], ongoing nutrition, incidence of disease, and the quality of the calf’s environment [8,9].Following the birthing process, calves in the UK are exposed to an environmental temperature that is typically below their thermal comfort zone (TCZ), with the most recent summary data available from the Met Office (from 1981 to 2010) showing that there is a mean annual maximum temperature of 12.4 °C and a mean annual minimum temperature of 5.3 °C [10]. The TCZ is defined as the environmental temperature at which the calf is not motivated to perform any thermoregulatory behaviour, and lies within the thermal neutral zone (TNZ) for the calf. The TNZ in the first week of life is estimated to be in the range of 15–25 °C [11]. An environmental temperature below the lower boundary of the TNZ is regarded as the lower critical temperature (LCT). Below the LCT is the point when the calf increases metabolic heat production to maintain thermal balance, which can be done by contraction of the skeletal muscles (shivering) or through non-thermogenic processes, such as increasing energy intake. The calculation of the LCT can be expressed as follows (Equation (1)):Tlc = (Tr + He’min·Ie) − Hmin (It + Ie)(1) where Tlc represents the lower critical temperature, Tr is the rectal temperature (°C), Ie represents external insulation (the resistance to heat loss via the skin provided by the hair coat) (°C·m2/watts), He’min is the minimum evaporative heat loss (watts/m2), Hmin is the thermoneutral heat production (watts/m2), and It represents the insulation provided by tissue (°C·m2/watts) [12]. Studies have investigated the effect of temperature on growth in calves. Cockram and Rowan [13] carried out a study using calves less than four weeks of age in controlled environmental chambers of 10 °C and 25 °C. They found that calves housed within the chambers at 10 °C had lower liveweight gains at 17–22 days of age than the calves housed in the chambers at 25 °C. Broucek et al. [14] reported that there was only a slight decrease in DLWG in calves that were exposed to high temperatures (average of 26.5 °C) compared to those exposed to temperatures regarded by the authors as moderate (averages of 19.5 °C and 15.7 °C). These high temperatures were recorded between June and September, which are regarded as summer and early autumn months in the Northern Hemisphere. Additionally, various studies have indicated that the season of birth can influence DLWG in calves. Place et al. [8] found from their study conducted in Pennsylvania, US, that there was a tendency for calves born in winter to have higher average DLWG than calves born in the other seasons of the year. This result was also found from the study carried out in Minnesota, US, by Chester-Jones et al. [5], where calves born in the autumn and winter gained 0.66 kg/d compared to 0.62 kg/d in calves born in summer. Therefore, season of birth and temperature are confounded, as lower temperatures would be expected during the winter months and higher temperatures during the summer months. As well as temperature and season of birth, Kelly et al. [15] highlighted the effect of relative humidity (RH) levels on DLWG; they examined the DLWG of calves housed at two temperatures (7 °C and 15 °C) and two levels of RH (75% and 95%), and found that there was no difference in the DLWG between the RH levels when housed at 15 °C; however, there was a significant increase in DLWG when housed at 7 °C with 75% RH compared to 7 °C with 95% relative humidity levels. This demonstrates that there is more than one aspect of the thermal environment that can influence DLWG.The consumption of milk provides energy and protein to support growth, but the energy can be diverted to other metabolic processes if required. A study by Rosenberger et al. [16] showed that giving calves a higher milk volume allowance resulted in a higher weight gain; however, a compensatory mechanism for calves that are exposed to temperatures below the LCT is to increase feed consumption if additional feed is available [17]. There is also evidence to show that nutrition and the environmental temperature can affect daily liveweight gain [13] and, therefore, it is generally advised that calves are provided with more energy when the environmental temperature falls below the LCT [18,19] in order to allow the calf to partition some of this energy towards growth as well as the heat production required for maintenance. This increase in energy can be achieved by methods such as increasing the volume of milk offered or the concentration calf milk replacer (CMR) fed to the calf. The practice of feeding dairy calves milk replacer is common in the UK [20], as it is a consistent product in terms of fat, protein, and energy content, in contrast to whole milk. As would be expected, various studies have shown that increasing the quantity of CMR fed each day increases calf DLWG [21,22,23].The objective of this longitudinal observational study was to investigate the effects of environmental conditions on calf growth using naturally occurring environmental conditions with the same management regime throughout. This is in contrast to studies such as that by Cockram and Rowan [13], conducted under controlled environmental conditions, or where the management of the calves was adjusted. Additionally, as an alternative to using air temperature as the sole environmental parameter, this study aimed to examine the effect of the proportion of hours for which calves were exposed to effective temperatures below their LCT on their DLWG for two management phases: the first days of life when the calf is in an individual hutch (phase 1), and then when the calves are moved into group pens (phase 2). Effective temperature was chosen, as although air temperature is widely used to assess the thermal conditions, other parameters such as relative humidity and wind speed can amplify the perception of high and low temperatures [24].2. Materials and Methods2.1. Calf Housing and ManagementThe calves used in this study were sourced from the dairy herds at SRUC Dairy Research and Innovation Centre, Crichton Royal Farm, Dumfries, Scotland, and their management followed normal farm management practices. All calves born within the study period were eligible for recruitment into the study, and consisted of dairy (Holstein) and dairy–beef cross (Holstein–British Blue cross, Holstein–Limousin cross, and Holstein–Aberdeen Angus cross) calves. Both male and female calves were eligible for recruitment. In total, 299 calves were eligible for recruitment into the study. A full descriptive summary of the calves can be found in Section 3.1 (Calves—Descriptive).All calves were removed from their dam within 24 h of birth and given an ear tag for identification, had their navel dipped with an iodine solution, and then were oesophageal tubed with four litres of thawed, quality-tested pasteurised colostrum, and weighed (birthweight). The calves were then taken to the main calf-rearing unit and placed within a straw-bedded individual calf hutch (Calf-Tel Compact, Calf-Tel, Hammel Corporation, Germantown, WI, USA) (Figure 1a), where they had access to fresh water and ad libitum starter pellets (VitaStart + Deccox, ForFarmers, Suffolk, England; crude protein 18%, crude fats and oils 4%, crude fibre 11.5%, crude ash 7%, 3 mm diameter) daily via buckets. From this point on, the calves received three litres of reconstituted milk replacer (Omega Gold, ForFarmers, Suffolk, UK) twice per day (7:30 a.m. and 4:00 p.m., approximately) via individual buckets with teats whilst in their individual hutches. The milk replacer consisted of 23% crude protein, 18% crude fat, 0.1% crude fibre, and 8.5% crude ash, and was fed at a concentration of 15%. All of the buckets for each calf remained with that calf for the duration of time for which it remained in the individual calf hutch. As a matter of routine, all calves received 4 mL Halocur® (Intervet International, Boxmeer, The Netherlands) for their first six days of life due to the history of cryptosporidiosis on the farm. Once calves were able to suckle from the teat confidently, and were assessed to be strong, healthy, and above 6 days of age, the calves were eligible to move into the group housing igloo pen (Figure 1b).The group housing igloo pen consisted of a hand-laminated fibreglass constructed dome (igloo) (Holm and Laue, Westerfield, Germany) (3.9 m, 4.4 m, 2.2 m: length, width, and height, respectively; volume: 20 m3). In front of the igloo there was a roofed pen (5.1 m × 5.1 m, length and width). The flooring of both areas was covered in straw. New straw was added weekly, and all bedding was removed and replaced every 2 weeks. The main calf-rearing unit consisted of 2 rows of 4 group igloo pens (Figure 2). Every pen contained an automatic milk feeder, from which every calf was allowed up to 7.2 L (7 L plus 0.2 L carry-over allowance) of reconstituted milk replacer every day; this was fed at a concentration of 15%.Upon leaving the individual hutch and entering the group housing igloo system, every calf received 2 mL Rispoval® RS + PI3 IntraNasal (Zoetis, Brussels, Belgium). It was normal practice for the farm to move calves into the group housing igloo pen between 6 and 14 days of age.2.2. Climate DataAir temperature (°C), relative humidity (%), and wind speed (m/s) were automatically recorded hourly throughout the study period using a Ventus W831 Weather Station (NSH NORDIC A/S, field 4, DK-8740 Brædstrup, Denmark). A sensor to measure air temperature and relative humidity and an anemometer to measure wind speed were located in the central passage of the calf shed at 0.8 m and 1.5 m, respectively. The data from both sensors were downloaded twice per week.2.3. MeasurementFor this study, measurements were taken for two management phases: Phase 1 covered the period from when the calves went into the individual hutch after birth until they left it and went into the group housing igloo pen; this phase is hereafter referred to as “B2G” (birth-to-group). Phase 2 covered the period from when the calves left the individual hutch and entered the group housing igloo pen until the end of the study; this phase is hereafter referred to as “G2E” (group-to-end).Data on date of birth, birthweight (kg), calving ease, and parity of dam were collected from farm records. The calf’s date of birth was used to define its season of birth (winter (December, January, February), spring (March, April, May), summer (June, July, August), autumn (September, October, November)). Data on air temperature (°C), wind speed (m/s), and relative humidity (%) for the period from when the calf went into the individual hutch until it left the individual hutch were also collected retrospectively from the weather station in the calf shed.2.3.1. Individual Hutch (B2G)Calves were weighed when they were removed from their dam using a manually operated calibrated weigh crate with Tru-Test digital load cells and an EziWeigh5 weigh head attached (Ritchie Agriculture, Forfar, Angus, Scotland); this weight was referred to as “birth weight”. It was not noted whether or not the calf was weighed after or prior to colostrum consumption, as this information was unable to be retrieved. The calves were weighed again, using the same weigh crate, at the point of leaving the individual hutch and entering the group housing igloo pen (LH weight). Daily liveweight gain (DLWG) (kg/d) for B2G was calculated by dividing the weight gain by the number of days between birth weight and group housing igloo pen entry weight. Calf-level health treatments were collected from farm records, and calves were classified as either having received treatment or not during this period (No, Yes).2.3.2. Group Housing Igloo Pen (G2E)Whilst in the group housing igloo pen, there were two recording days per week—typically Mondays and Thursdays (12:30 p.m.–2:00 p.m.), and occasionally on other days where circumstances intervened. Various measurements were taken from each of the calves on these days, as described below.LiveweightA record of liveweight (kg) was taken for each calf on the recording days. This weight was assigned a weighing number to represent whether this was the calf’s first, second, third, etc. recording of liveweight whilst in the group housing igloo pen for the duration of the study (WGT1, WGT2, WGT3, WGT4, WGT5, WGT6, WGT7). Daily liveweight gain (DLWG) (kg/d) for G2E was calculated but, to take into account the possible changing rate of DLWG over time between entering the group housing igloo pen and the end of study period, a linear regression was applied for G2E, with the value of the slope used as DLWG [9,25].Health AssessmentA health assessment of each calf was made on the day the calf left the individual hutch, and on each recording day, using the Wisconsin method [26,27]. This was carried out by the same trained operator (D.J.B.). This method involved taking a rectal temperature, visually assessing ocular and nasal discharge, head/ear positioning, and the presence or absence of a cough. Each aspect was given a score on a scale of 0–3, with 0 being described as “normal” and 3 as “severe”. A sum of these scores represented the overall health score for the calf, with the lowest possible score being 0 and the maximum score being 15. For this study, rectal temperature was taken using a digital thermometer (Genia Digiflash, Saint-Hilaire-de-Chaléons, France). A score for faecal consistency was not able to be carried out, as the calves were group housed and, thus, faeces from individuals could not be identified, so this was not included in the analyses of health status.From the scoring, calves were defined as either “healthy”, “diseased”, or “intermediate” based on the criteria in Table 1.Health status for the calf was then categorised for the G2E period of the study into “ever showed clinical or mild signs of disease” (Intermediate and Diseased—Signs of disease, Yes) or “never showed any signs of disease” (Healthy—Signs of disease, No).Farm staff involved in the care of the calves were made aware of the results of the health assessments, and treatment was administered at the discretion of the farm. Calf-level treatments were collected from farm records, and calves were classified as either having received treatment or not during this period of the study (No, Yes).Milk Intake DataMilk consumption data (quantity of calf milk replacer (CMR) consumed each day) were collected for each calf for the period during which it was in the study (from entering the group igloo pen to the last time it was weighed). The milk-feeding equipment was changed within the study period. Initially, calves were fed from an H and L100 (Holm and Laue, Westerfield, Germany—183 calves), and latterly from a BioControl milk feeder (BioControl AS, Grimstad Gård, N-1890 Rakkestad, Norway—116 calves). No calf was fed using both milk-feeding systems.The number of days from the calf entering the group pen until the last day of measurements being taken was calculated and used to determine the average daily intake of milk (l/d) and CMR (g/d) for each calf. Only average daily CMR intake was used in the analysis.2.4. Data Analysis2.4.1. Inclusion and Exclusion CriteriaFor calves to be included in the analysis of the B2G stage, they must have had a birthweight recorded and been 14 days of age or less when they left the individual hutch and entered the group housing igloo pen. As well as these criteria, for calves to be included in the analysis of the G2E stage, they also must not have been sold or died before reaching the end of the study period, must have remained in the group housing igloo pen for the full duration of the study period, and must have had a full set of data for milk intake from the automatic milk feeder.2.4.2. Proportion of Hours Below LCTBased on the hourly measurements taken from the weather station, the effective temperature (°C) for every hour was calculated using the etv function from the “ThermIndex” package in R [28]. This function incorporates air temperature, relative humidity, and wind speed, and is based on the equation by Suping et al. [29]. The lower critical temperature (LCT) was calculated per day per calf from birth (day 0) to the day they left the study. LCT at birth (day 0) was defined as 15 °C [11], and decreased with the age of the calf [18] by 0.5 °C per day.The proportion of total hours for which each calf experienced an effective temperature below this LCT was then calculated; this was done by calculating the number of hours for which the effective temperature was below that of the associated age-related LCT for each calf, and then dividing that by the total number of hours for which the calf was in each phase of the study (individual hutch (B2G) and group igloo pen (G2E)). The proportion of hours below the LCT was then categorised into quartiles based on the examination of the distribution plots: B2G (≤0.32, 0.33–0.58, 0.59–0.96, ≥0.97), G2E (≤0.01, 0.02–0.06, 0.07–0.27, ≥0.28).2.4.3. Statistical AnalysisAll statistical analyses were conducted using R software [29]. Datasets were compiled for the two time periods of interest for the study: birth to group pen entry (B2G), and group pen entry until the end of the study period (G2E). Each individual calf was regarded as the experimental unit within both datasets.Table 2 shows which variables were considered in the analysis of each dataset.Any calves missing a birth weight were excluded, as their DLWG could not be calculated. As it was normal practice for the farm to move calves into the group housing igloo pen between 6 and 14 days, calves older than 14 days were excluded. Moreover, calves older than 14 days were likely to have experienced a health-related issue in early life. Other reasons for exclusion included the sale or death of the calf before the end of the study period, the return of the calf to an individual hutch after initially entering the group housing igloo pen, milk consumption data being unavailable, or only having one weight in the group housing igloo pen. The dependent variable—DLWG—was investigated for association with the independent variables mentioned in Table 2.To analyse the dataset for the time period of B2G, multiple univariable linear models using the lm function in R [30] were used to screen the independent variables for level of significance.To analyse the G2E dataset, multiple univariable linear mixed-effects models with random effects for the group in which the calf was placed, the group igloo pen in which they were placed, and the milk-feeding system used were constructed. The lmer function from the “lme4” package [31] was used. For both sets of analyses, independent variables that had a p-value of less than 0.20 were carried forward and included in a multivariable model.A maximal model was constructed and then optimised using backward step selection until only variables significant at p < 0.05 remained. Tukey’s post hoc test was used to obtain significance between factor levels of significant variables only.3. Results3.1. Calves—DescriptiveIn total, 299 calves were enrolled onto the study. Of these, 226 were dairy calves (137 female, 89 male) and 73 were dairy–beef cross calves (34 female, 39 male). Of these 299 calves, 109 were born from primiparous dams and 190 from multiparous dams. In terms of calving ease, 263 had an unassisted birth and 36 had an assisted birth. Out of the 299 calves, 80 were born in winter (26 dairy, 54 dairy–beef cross), 85 in spring (76 dairy, 9 dairy–beef cross), 54 in summer (45 dairy, 9 dairy–beef cross), and 80 in autumn (51 dairy, 29 dairy–beef cross). One calf was excluded from the data due to a missing birth weight, and 27 were excluded as they were older than 14 days when moved into the group housing igloo pen. Data from the remaining 271 calves were used in the analysis of the B2G dataset.In terms of the age at which these 271 calves left the individual hutch and entered the group housing igloo pen, 19 calves were 6 d, 52 calves were 7 d, 45 calves were 8 d, 45 calves were 9 d, 29 calves were 10 d, 28 calves were 11 d, 27 calves were 12 d, 16 calves were 13 d, and 10 calves were 14 d.Of the 271 calves during the B2G stage, 37 received treatment whilst in the individual hutches (19 diarrhoea (6.1 d ± 2.05), 13 respiratory disease (6.1 d ± 2.36), and 5 other disease (3.6 d ± 1.63) (mean age at treatment ± SD). During the G2E stage, of the 221 calves, 95 calves were treated for 110 incidents (36 scour (14.5 d ± 4.58), 62 respiratory disease (22.0 d ± 6.46), and 12 for other health related reasons (18.8 d ± 5.97) (mean age at treatment ± SD). Of the 95 calves treated during this period (G2E), 12 were treated for more than one disease. A total of 16 calves were treated in both the B2G and G2E stages.Five calves died or were euthanized during the study (one experienced a seizure, two suffered traumatic leg injuries, one had an umbilical abscess, and one had an injury to an eye that was refractory to treatment) and one calf was sold before completing the study. Three calves were also excluded as they returned to the individual hutches from their group igloo pen. There were two separate occasions throughout the study where milk-feeding data could not be recovered as a result of power failure; this affected 41 calves, which were excluded from the study. Following exclusions, the dataset used for the analysis of the G2E time period contained 221 calves.The calves, on average, lost weight (mean DLWG −0.07kg/d) during the B2G time period, although there was a large variation in DLWG (Table 3). This variation was slightly less for G2E, where the calves had a mean DLWG of 0.60 kg/d. For the G2E period, 98 calves showed no signs of disease (all scores in Healthy category) and 123 showed signs of disease (one or more scores in the Intermediate or Disease categories).3.2. Climate—DescriptiveOver the course of the study, there was a range of climatic conditions experienced by the calves (Table 4). For the time period from birth until leaving the individual hutch (B2G), some calves experienced an effective temperature that was always below their age-related LCT. On average, calves in the individual hutches spent 60% of the time below their LCT.3.3. Daily Liveweight Gain—Birth to Group Pen (B2G)The final multivariable linear model for the daily liveweight gain of calves from birth until they left the individual hutch and entered the group housing igloo pen (B2G) contained the following variables: the proportion of hours for which the calf was exposed to effective temperatures below its age-related LCT (ProphrsLCT), the weight of the calf at birth (Birth weight), and the age at which the calf left the individual hutch (Age leaving individual hutch) (Table 5). No other variables were significant in the model (Season of birth: p = 0.454; Parity of dam at birth: p = 0.755; Sex of calf: p = 0.929; Breed classification: p = 0.391; Calving ease: p = 0.142; (Farm) Treatment administered: Individual hutch: p = 0.273). A potential reason for these variables being non-significant could be the management of the calves throughout the study.There was a significant effect of ProphrsLCT on DLWG (B2G) (F (3, 265) = 6.098, p < 0.001). There was a significant difference found between the categories ≤ 0.32 and 0.59–0.96 (p = 0.015), ≤0.32 and ≥0.97 (p < 0.001), and 0.33–0.58 and ≥0.97 (p = 0.041). The results suggest that the greater the proportion of time that a calf spent below the LCT, the greater the reduction in DLWG (Figure 3).The birth weight of the calf had a significant effect on DLWG for the period between birth and leaving the individual hutch (F (1,265) = 35.154, p < 0.001). When accounting for all other variables in the model, for every kg increase in birth weight, DLWG (B2G) reduced by 0.02 kg/d, indicating that heavier calves grew more slowly during this period.The age at which the calf left the individual hutch also had a significant effect on the daily liveweight gain (DLWG, kg/d) (F (1,265) = 11.196, p < 0.001); DLWG (B2G) increased by 0.03kg/d for every day older the calf was when leaving the individual hutch.3.4. Daily Liveweight Gain—Entering Group Pen until End of Study Period (G2E)The final model examining factors affecting the daily liveweight gain of calves for the G2E period of the study contained the proportion of hours for which the calf was exposed to effective temperatures below its age-related LCT (ProphrsLCT), the age of the calf on entry to the group pen (Entry age), and its average daily intake of CMR (CMR intake) (Table 6).There was no significant effect of ProphrsLCT on daily liveweight gain (G2E) (chisq = 2.747, 3 df, p = 0.432).The age at which calves entered the group housing igloo pen also had a significant effect on DLWG for G2E (chisq = 6.343, 1 df, p = 0.012); the final model indicated that for every day older the calf was on entry to the group pens, their DLWG (G2E) increased by 0.01 kg/d.There was a significant effect of the average CMR intake on the daily liveweight gain for G2E (chisq = 348.686, 1df, p < 0.001). Furthermore, there was a significant positive correlation between average CMR intake and DLWG (G2E) (r = 0.77, p < 0.001) (Figure 4); the more CMR consumed by the calf, the higher the growth rate achieved.4. DiscussionThe DLWG values measured for the calves in this study were similar to those reported in other UK studies. Bazeley et al. [32] reported that there was no weight gain in calves in their first eight days of life, which was similar to this study, which showed a mean DLWG for B2G of −0.07 kg/d, although the range was −1.33 to +1.00 kg/d. The range of DLWG shows that it is possible for calves to have a positive DLWG in their early stages of life. For the G2E time period, the mean DLWG was 0.60 kg/d for calves 29 days of age (mean age at end of study). The DLWG values for the present study were slightly higher than those reported for the top performing herds in the study of Bazeley et al. [32] (0.52 kg/d) and the non-jacketed calves in the study of Scoley et al. [33]. Differences in the housing and management systems on each of the farms in these studies will likely account for these minor differences in reported growth rates.4.1. Proportion of Hours Below LCTIn the present study, there was a significant effect of the proportion of hours for which the calves were exposed to effective temperatures below their LCT on their daily liveweight gain (DLWG, kg/d) when in the individual hutch from birth until leaving the individual hutch and entering the group housing igloo pen. However, this had no significant effect on DLWG for the older age group in the time period from entering the group housing igloo pen until the end of the study (~28 days of age). An important point worth noting is how rapidly the vulnerability of calves changes with age, which is often a message that the industry often fails to appreciate. For this study, the proportion of hours for which the calves were exposed to effective temperatures below their LCT during the B2G period was split into quartiles (≤0.32, 0.33–0.58, 0.59–0.96, >0.96), as it was for the G2E period (<0.01, 0.02–0.06, 0.07–0.27, >0.28). The B2G period covered the period from birth until leaving the individual hutch (between 6 and 14 days of age) and the period G2E covered the period from entering the group housing igloo pen (between 6 and 14 days of age) until the end of the study (approximately 28 days of age). It can be seen from the quartiles from both time periods that the older calves (G2E period) were not exposed to effective temperatures below their LCT to anywhere near the same extent as they were during the B2G period.4.2. Climatic Environment ParametersThe proportion of hours below age-related LCT and/or effective temperatures are not the conventional climatic environmental parameters used to assess thermal conditions. In some studies, the temperature–humidity index (THI) has been used. Shivley et al. [34] found that calves exposed to a THI of less than 50 during the pre-weaning period had a higher DLWG compared to calves exposed to a THI between 50 and 59 or greater than 70. However, it is suggested that the THI is more of an indicator of heat stress rather than of general conditions in cooler climates. Moreover, the THI is based solely on a combination of air temperature and relative humidity; it does not take wind speed into consideration, which effective temperature does. The housing for the calves in the present study was within an open-sided umbrella-like structure and, therefore, the calves had the potential to be exposed to wind; thus, the use of the effective temperature was thought to be most appropriate. As Hahn et al. [35] state, there are limitations to the use of air temperature as a representation of the thermal environment, and a combination of parameters should be used. There is a limitation in the work presented as well, as the calculation of effective temperature did not take the radiant heat from sunlight into consideration and, therefore, was not completely representative of the heat exchange with the environment. Some work has been carried out to examine the appropriateness of human comfort indexes for use in livestock—particularly heat stress [36]—but further work should be carried out to develop a more general index for livestock, in particular for calves.4.3. B2G PeriodFor the B2G period, it is likely that the young calves struggled to acclimatise to the environment in which they were kept. According to Nienaber and Hahn [37], this process can take days or even weeks to be achieved. Rowan [38] reported that acclimatisation develops with the age of the animal. Another possible explanation for this result could lie within the calculation used for the proportion of hours below the LCT. It is acknowledged that the calculation used for the proportion of hours below the LCT is imperfect, but this calculation is a significant improvement on previous studies that merely measured air temperature, as it at least takes into account the influence of the age of the calf on its LCT. There would be very few calves who were exposed to effective temperatures that were above their age-related LCT on an hourly basis for the individual hutch period. Most calves in Scotland, including those in the present study, would be exposed to effective temperatures at some stage during the day—but mostly at night—that were below this LCT in winter months. As this study has demonstrated, calves that were exposed to a high proportion of hours below their LCT had significantly lower DLWG. The calves were offered a high-volume, energy-dense milk diet during this management stage (B2G) (6 L/d at 15% concentration which equated to 900 g CMR/d), regardless of season of birth. Therefore, despite this, the climatic environment still had an impact on the performance of these calves. However, milk intake was not recorded during this management stage, and it was assumed that calves drank all of the milk that was offered to them at both feedings. Therefore, this result suggests that the period following birth is when the calf is most vulnerable to the climatic environment, and management procedures such as the application of calf jackets could potentially be beneficial by acting as a barrier to reduce heat loss to the environment [39].4.4. G2E PeriodThe non-significant effect of proportion of hours below LCT on DLWG for the time period from entering the group housing igloo pen until the end of the study period (G2E) could be as a consequence of the calves’ development. By this stage, their LCT was below the average climatic conditions for the region. The calves in the present study entered the group housing igloo pen when their age-related LCT was 10 °C (mean entry age was 10 days). Every day after this, their LCT declined by 0.5 °C and, therefore, there would be a higher chance that the calf would not be below its LCT unless there had been a sudden dramatic change in environmental conditions—especially in southern Scotland, where the mean daily air temperature has been estimated to be around 9 °C [40]. Another reason is related to the behavioural response of the calves. For the G2E time period, the calves were group housed, whereas for B2G they were housed individually. One behavioural response to low environmental temperatures is huddling [17]. Once in the group housing igloo pen, calves had the opportunity to keep warm by huddling with the other calves in the group, whereas in the individual hutch, each calf was reliant on other processes, such as the ability to nest in the bedding material to maintain warmth. Although not recorded in the present study, it would have been of interest to see whether this behavioural response (huddling) was occurring, and in particular whether it occurred during specific times of the day (e.g., around dawn and overnight) and its relationship with the climatic environment.4.5. Birth WeightIt was found that birth weight had a significant effect on DLWG for the period from birth until leaving the individual hutch (B2G). Donovan et al. [41] and Yaylak et al. [42] also both found that birth weight had a significant influence on subsequent daily liveweight gain. However, daily liveweight gain for both of these studies was recorded over longer periods of time than in the present study (birth to 6 months and birth to weaning, respectively). This effect of birth weight on subsequent growth rate may be a result of the volume of milk given to the calves during this early phase of life. When they were housed in the individual hutch (B2G), calves in the present study were given 6 L/d of milk. In the case of the lightest calf birth weight (31 kg), this equated to nearly 20% of bodyweight, whereas for the heaviest calf (67 kg), the 6 L/d only equated to 8.9% of bodyweight. Therefore lighter born calves potentially had more available energy, over and above that was required for maintenance, which could be used for growth. This explanation is supported by the results of Khan [43], who showed that calves from 3 d of age can safely consume 20% of their bodyweight per day, and that an increase in consumption supports an increase in DLWG.4.6. Age Leaving Individual Hutch and Entering Group Houisng Igloo PenThe age at which the calf left the individual hutch and entered the group housing igloo pen had a significant effect on DLWG in both the B2G and G2E periods. For the B2G time period, the older the calf was upon leaving the individual hutch and entering the group pen, the higher the DLWG. The same trend was evident for the G2E time period, with the greater the age at entry to the group pen, the higher the DLWG. A factor that could have influenced this result is the use of the automated feeders, and the possibility that older calves take less time to learn to use them. Fujiwara et al. [44] reported that calves that took a considerable time to learn and adapt to the automated feeders had reduced milk intake and poor growth rate in the initial weeks of entering the group pen. They also reported that calves introduced around 6 days of age took longer to voluntarily consume milk from the automated feeders than calves at 9 days of age. This result was also found in the study by Jensen [45].4.7. CMR IntakeThe amount of CMR consumed per day had a significant effect on DLWG for the G2E time period. This result is consistent with those found in other studies [2,22,23,46]. Morrison et al. [21] showed that an increasing level of milk replacer produced a higher DLWG between 0 and 28 days of age, while Johnson et al. [46] found that there was a positive correlation of CMR intake with daily liveweight gain. The study by Khan et al. [43] demonstrated that calves can safely be fed 20% of their bodyweight per day, which can result in an increase in DLWG. These results could be due to the extra energy that the increase in CMR intake provides above maintenance energy. Further work could consider examining the proportion of maintenance energy provided by the daily CMR intake, in terms of being above and below that which is required by the calf, to validate this theory.4.8. Strategies for Achieving Thermal ComfortThere are a few farm-level practices that could assist with achieving thermal comfort other than feeding more energy [18]; one such practice would be the application of calf jackets. Scoley et al. [33] reported a 6.37 °C increase in the skin temperature of calves that received calf jackets compared to calves that did not receive a calf jacket when the ambient air temperature was 7.7 °C (mean of their study). Rawson et al. [47] concluded that there was a 52% increase in whole-animal insulation when calves wore a calf jacket at extremely low ambient temperatures. However, as Robertson [39] conveys, there is still a gap between the scientific evidence and the popularity as well as the benefit of use of calf jackets.Another practice could be to apply more bedding material to the housing in order to enable the calf to “nest”. Nesting behaviour is when the calf tries to bury itself within the bedding material, hence creating a nest-like feature. Nesting scoring can be carried out as a quantification of thermal comfort, with an example of its use being captured by Lago et al. [48].5. ConclusionsThe present study has shown that the key performance indicator of daily liveweight gain is affected by the climatic housing environment in the very early stages of the rearing phase of the calf. Therefore, emphasis should be placed upon the management of the calf at this stage, and providing an appropriate plane of nutrition will assist with achieving target daily liveweight gain. The application of amelioration strategies may also be beneficial, especially in times when temperatures are below the LCT. Further investigation should be carried out to examine the various comfort indices and assess which is the most suitable for use with pre-weaned calves and different types of housing. It would be hoped that other studies would consider using hours below LCT and reject the use of average environmental values as a variable for analysis.

animals : an open access journal from mdpi

10.3390/ani11082162

PMC8388387

Rodents are the most abundant and diversified group of mammals. These animals show genetic and physical diversity in different ecosystems of the world, including the desert ecosystem. The current study was undertaken to check the morphometric pattern of three commensal rodent species, viz, Mus musculus, Rattus norvegicus, and Rattus rattus, in Qatar. One hundred forty-eight rodents were captured and studied for body and cranio-mandibular measurements. The study found R. norvregicus as the most prevalent rodent in Qatar. Most of the rodents were collected from Al Rayan municipality, were adults, and were from livestock farms. The rodents’ average body weights were 18.8 ± 2.2 gm, 264.3 ± 87.5 gm, and 130 ± 71.3 gm for M. musculus, R. norvegicus, and R. rattus, respectively. The average morphometric measurements of the external body and skull were normally distributed and can be used as a reference of R. norvegicus and R. rattus for Qatar.

The current study was undertaken to estimate the morphometric pattern of three commensal rodents, i.e., Mus musculus, Rattus norvegicus, and Rattus rattus in Qatar. One hundred forty-eight rodents were captured from different facilities throughout Qatar. The captured rodents were used to identify the external body and cranio-mandibular morphometry. The study found that R. norvregicus was the most prevalent (n = 120, 81%, 95% CI: 73.83–87.05). Most of the rodents were collected from Al Rayan municipality (n = 92, 62%), were adults (n = 138, 93.2%, 95% CI: 87.92–96.71), and were from livestock farms (n = 79, 49%, 95% CI: 41.02–57.65). The rodents’ average body weights were 18.8 ± 2.2 gm, 264.3 ± 87.5 gm, and 130 ± 71.3 gm for M. musculus, R. norvegicus, and R. rattus, respectively. The research found that the studied rodents are smaller than those of other countries such as Turkey, Tunisia, and Iran. The study of morphometry is a useful tool for the traditional identification of small mammal species, including rodents. The average morphometric measurements of the external body and skull were normally distributed and can be used as a reference of R. norvegicus and R. rattus for Qatar. A further comprehensive study is required to investigate the rodent population index, eco-friendly control program, and public health importance in Qatar.

1. IntroductionRodents are the largest group of mammals, distributed on every continent of the world except Antarctica [1]. Globally, there are 2552 rodent species available, of which three species, i.e., house mice (Mus musculus), brown rat (Rattus novegicus), and black rat (Rattus rattus), occupy different habitats with higher density than other species of rodents [2,3]. These human commensals live in diverse ecosystems throughout the world, showing high morphological and genetic variation. For instance, the brown rat showed at least 13 evolutionary clusters globally [4]. Several evolutionary factors, such as climate and geography, predators, urbanization, and agricultural settlement, are behind these evolutionary changes [5,6,7]. The desert environment is also a factor for the phenotypic and genotypic evolutionary change of mammals. For example, fur coloration and its covariation with habitat have been reported for desert gerbils [8]. Genetic analysis and phenotypic and morphometric assessments provide unique ways of identifying different mammalian species and evaluating animal diversity evaluation [7,9]. The external and cranio-mandibular morphologies are valuable tools in the classification of rodent species. The bones of a skull have some variation between and within a mammalian species that lead their species or subspecies to a distinguished morphological identity [9].The state of Qatar is a small country in the Arabian Peninsula, whose terrain comprises sand dunes and salt flats across a low barren plain [10,11]. The country has a dry, subtropical climate, with very low annual rainfall (33.1 mm in 2010 and 114.1 mm in 2015), intensely hot (42.7–48.1 °C) and humid (32–72% relative humidity) summer, and warm (10.7 °C) winter. Due to the climate and geography, agricultural practices are limited in Qatar [10,12]. Rodents have importance for animal and public health in this country [13]. Rodent-borne pathogens, such as Coxiella and Toxoplasma, are common causes of livestock abortion in Qatar [14]. Taenia taeniaeformis, Toxoplasma godii, and Toxascaris leonina were reported among pet animals [15,16]. Zoonoses that can be associated with rodents, such as Escherichia coli, Giardia duodenali, and Hymenolepis nana, were reported among human populations in this country [17,18]. Moreover, the zoonotic cestode, Hymenolepis diminuta, was identified among R. norvegicus in Doha city of Qatar [19,20]. The country has governmental [21] and non-governmental rodent control programs. Minimal research, however, has been done on rodents in this country [13,19,20]. There is no documented report of rodent identification guidelines, such as morphometry of rodents in Qatar. Therefore, the present research aimed to study three commensal rodents, such as Mus musculus, Rattus norvegicus, and Rattus rattus, to identify the specific species of the rodents and to understand their physical and behavioral characteristics that are potentially found in the Qatar.2. Materials and Methods2.1. Study Season, Area, and Rodent CollectionA cross-sectional study was done from November 2019 to February 2020 as a part of routine pest control program in Qatar. A total of 250 traps were used, which include 150 single rodent traps (SRT) and 100 multi rodent traps (MRT). We used different types of baits such as bread (Arabian khubj), biscuits, potato chip, and cheese for capturing the rodents [22]. An SRT or MRT was used randomly, without targeting any specific rodent species or the species behavior. A water bottle containing 5% glucose was affixed to each trap to reduce dehydration and stress of the captured animals in the harsh Qatari environment. The trappings covered six facilities: family residents, bachelor residents, agricultural farms, livestock farms, industrial areas, and commercial areas throughout Qatar (Figure 1). The traps were set for a single night. Successful traps were collected in the morning and transferred at the earliest convenience to the veterinary laboratory, Doha, Qatar. A comfortable temperature was maintained (20–25 °C) in the transportation car and veterinary laboratory rodent room. The traps were washed with soap and pressurized water and air-dried to avoid any residual contamination and transmission from the previous rodent to the next.2.2. Rodent Identification and Morphometric AssessmentThe captured rodents were euthanized using 5% isoflurane inhalation for five minutes in a desiccator. After weighing with an electronic balance (Serial No. 057700082, Kern EG420-3NM, Kern & Sohn GmBH, Balingen, Germany), morphological appearance and external measurements were recorded as per species, age, sex, and pregnancy [22,23,24,25]. Rodent species were identified based on morphologic characteristics and measurements. The animals were assessed for sex (female or male) using external and internal aspects of reproductive organs such as testicles, penis, seminal vesicles, vagina, mammary teats, and possible pregnancy signs. For age detection, we only identified the adult rodents. Developed genital organs and pregnancy were the sign of an adult rodent. Additionally, we considered prominent temporal ridges and postorbital processes of the skull to determine a rodent as mature. The presence of a gravid uterus served as the indicator of pregnancy.Five standard external measurements were made for the animals using a ruler (Figure 2). Following the morphological characterization, the rodents were dissected, skulls were collected, cleaned, and dried according to the standard procedure [26]. The cranium and mandible morphometric variables were recorded using a digital caliper (TESA TWIN-CAL IP67, Hexagon, Switzerland) described previously [9,27,28,29] and illustrated in Figure 3, Figure 4, Figure 5 and Figure 6.2.3. Statistical AnalysisThe data were analyzed using statistical software StatSoft (2011) to study the descriptive analysis of the number of captured rodents and their morphometric variables that included mean, percentage (%), 95% confidence interval (CI), standard deviation (SD), skewness, standard error of skewness, kurtosis, and standard error of kurtosis. The data were tested with the Kolmogorov–Smirnov test, skewness, and kurtosis to validate the normality. If the skewness and kurtosis were outside −2 and +2, the measurement was considered significantly skewed or kurt [30,31]. The student t-test was performed to examine the variability of the morphometric traits among sex (female vs. male) and pregnancy (pregnant vs. non-pregnant). The chi-square (χ2) test was performed to examine the level of significance (p < 0.05) among the area (municipality) and trapping location types.3. Results3.1. Demographic InformationThe study captured 148 rodents from all seven municipalities of Qatar (Table 1, Figure 1). A total of 79 rodents were captured by SRT and 69 rodents by MRT. The thirty-two MRT captured more than one rodent (2–5) at a time. Based on the morphologic and morphometric characters of the body and skull, three species of rodents were identified, i.e., M. musculus, R. rattus, and R. norvegicus. R. norvegicus comprised 81.1% (n = 120) of the total captured rodents, whereas R. rattus (n = 24) and M. musculus (n = 4) showed low density. Most of the collected rodents (n = 138, 93.2%) were adults. A major portion of the captured rodents was collected from Al Rayan municipality (n = 92, 62%). This municipality harbors all the three commensal species (M. musculus and R. rattus, and R. norvegicus), showing (χ2 = 21.02, p < 0.05) the highest density for R. norvegicus (n = 64). The majority of the rodents (n = 79, 49%) (χ2 = 35.29, p < 0.05) were collected from the livestock farms.3.2. Morphometric Assessments of RodentsThe overall means of body weight, external morphometry, and cranio-mandibular variables per species are presented in Table 2, Table 3 and Table 4. Out of the 148 rodents, 108 rodents were dissected, comprised of 86 R. norvegicus, 18 R. rattus, and 4 M. musculus. The average body weight was variable among three rodent species (18.8 ± 2.2 gm, 264.3 ± 87.5 gm, and 130 ± 71.3 gm for M. musculus, R. norvegicus, and R. rattus, respectively). The skewness and kurtosis statistics of all the studied external body measurements of R. norvegicus and R. rattus were within −2 and +2. This indicated that the observed values were normally distributed. In general, the tail is longer than the length of the body and head of M. musculus and R. rattus, which is the opposite in R. norvegicus. Compared to the general length of a rodent, the ears and legs of R. rattus are longer than that of R. norvegicus. As the captured number of M. musculus was small, no further statistical comparative analysis could be considered on their body or cranio-mandibular measurements.The t-test showed that there is no sexual or pregnancy-related dimorphism (p > 0.05) in any of the presented characteristics in the case of R. norvegicus (Table 5, Table 6, Table 7, Table 8, Table 9 and Table 10). However, the right ear length measurements showed that females have longer ears than males in R. rattus. Moreover, the mandibular characters, such as the length of lower incisors and the distance between lower incisor to coronoid process, lower incisor to condyloid process, lower incisor to angular process, ramus to molar tooth 1, and lower incisor to molar tooth 1 of R. rattus, were significantly higher in females than males (p < 0.05). In addition, the value of lower molar tooth 1 to molar tooth 3 was higher in the case of males than females in R. rattus (Table 7). Furthermore, the right hind leg was longer (p > 0.05) in non-pregnant than pregnant R. rattus (Table 8).4. DiscussionThe study of rodent demography is essential from ecological and public health perspective [32]. The present study identified three commensal rodent species in Qatar captured during routine pest control activities. These rodents have a cosmopolitan distribution and are mainly facilitated by anthropic activities [2]. Four species of rodents were reported previously in Qatar, viz., Arabian Jerboa (Jaculus loftusi, previously included in Jaculus jaculus), house mouse (M. musculus), brown rat (R. norvegicus), and black rat (R. rattus) [13,19,20,33]. Jaculus loftusi is a wild dipodid rodent that lives in the desert ecosystem, like the sandy and rocky places [34], so this species is not in the scope of the present study. However, the current study found that a significant component of commensal rodents in Qatar is R. norvegicus. This is supported by the previous reports [19,20], which captured only R. norvegicus during their studies in Qatar.Our study revealed that most of the rodents were from livestock farms. The livestock farms are mostly made up of mixed livestock species with poor management and biosecurity [35], making an ideal place for rodents to colonize and why we captured a major part of rodents from these places. A previous study reported that over 75% of the livestock farms were infested with rodents, mainly by R. norvegicus, and the incidence of house mouse M. musculus was detected less in Qatar [13], which is congruent with the present study. Out of the 148 captured rodents, only four were M. musculus.Traditional morphometry is a valuable tool for species identification in small mammals, including rodents [28,36]. The present study found the body weight and general body length of R. norvegicus as 264.3 gm and 398.5 mm, respectively, which were 259 ± 85.2 gm and 405 ± 54.7 mm, respectively, for the same species in Turkey [37]. In the case of cranial morphometry, the condylobasal length and the zygomatic breadth of R. norvegicus in the current study were 45.2 mm and 22.4 mm, which were 45.52 mm and 23.75 mm in the case of Turkey [37] and 46.84 mm and 21.64 mm in the case of Iran [38], respectively, for the same species and measurements. The overall body length of R. rattus in Turkey was 378.43, which was 324.4 mm for the same species of Qatar. The cranial length and zygomatic width of R. rattus in the current study were 37.2 mm and 18.2 mm, which were 39.15 mm and 19.86 mm, respectively, for Turkey [37] and 39.08 mm and 19.97 mm, respectively, for Tunisia, respectively for the same species and measurements [39].Similarly, the body length of M. musculus in Qatar was 78.5 mm, which was 85.41 mm [24] and 88.0 mm [40] for the same species from different parts of Iran. Due to the small sample size (n = 4), we do not have strong support in the results of M. musculus morphometry. However, the overall body and cranial size indicate that the three studied rodent species in Qatar are comparatively smaller than the same species from the countries like Turkey, Tunisia, and Iran. This variation may be due to Qatar harsh environmental effects [6,7,8], which is supported by Bergmann’s rule [41]. Rodents of the colder environment are bigger in body size than the wormer environment [42,43]. This further highlights the necessity of performing traditional morphometry on the geographic population of rodents, specifically cosmopolitan species.Based on the average general body and skull morphometric measurements, males were slightly larger than females, although there is no significant sexual dimorphism. This finding is supported by a previous study by Ventura and Lopez-Fuster [7]. However, the present study showed that the body and cranio-mandibular linear measurements of commensal rodents in Qatar were normally distributed for the two species, R. norvegicus, and R. rattus. Bodyweight and body and skull linear measurements distribution shape were approximately symmetric since the statistic of skewness measures were between −0.2 and 0.2 [30,31]. Normality analysis of the biometric traits can be considered typical characteristics of the two rodent species, R. norvegicus and R. rattus, in this country. To the best of the authors’ knowledge, such work is the first time in Qatar. Therefore, the current study can be used as a reference for morphometric measurements of the commensal rodents in this country, especially for R. norvegicus and R. rattus.5. ConclusionsThe current study estimated, identified, and characterized the morphometric variables of three commensal rodents in Qatar. The research identified that the commensal rodents of Qatar are comparatively smaller than the same species of some other countries, such as Iran, Tunisia, and Turkey. The is the first study on rodent morphometry in Qatar and even in the Arabian Peninsula. Due to geo-ecological similarities, the present study can be a reference study to rodent or small mammal identification in Qatar and other countries of the Arabian Peninsula.

animals : an open access journal from mdpi

10.3390/ani11061573

PMC8229765

In recent decades, the ban of antibiotic growth promoters together with the reduction in the feed of trace minerals with antimicrobial properties such as zinc and copper, has increased the demand to identify organic alternatives with antimicrobial properties that may improve the production efficiency and sustainability in an antibiotic-free system. The sustainability of pig production can be enhanced, by increasing the feed efficiency, modulating the microbiota, and reducing the impact of different diseases through the use of natural substances. Considerable research has focused on the gut environment and microbiota modulation as key to boosting pig health. Sustainable dietary interventions that positively modulate the gut environment and microbiota in pigs are required to enhance pig health and welfare. In the field of natural substances, seaweed and their bioactive compounds have assumed particular importance as feed ingredients for pigs. In fact, seaweeds include numerous bioactive substances with prebiotic, anti-microbial, antioxidant, anti-inflammatory, and immunomodulatory activities. The present paper reviews the prebiotic effects of seaweed polysaccharides in pigs.

To ensure environmental sustainability, according to the European Green Deal and to boost the One Health concept, it is essential to improve animals’ health and adopt sustainable and natural feed ingredients. Over the past decade, prebiotics have been used as an alternative approach in order to reduce the use of antimicrobials, by positively affecting the gut microbiota and decreasing the onset of several enteric diseases in pig. However, dietary supplementation with seaweed polysaccharides as prebiotics has gained attention in recent years. Seaweeds or marine macroalgae contain several polysaccharides: laminarin, fucoidan, and alginates are found in brown seaweeds, carrageenan in red seaweeds, and ulvan in green seaweeds. The present review focuses on studies evaluating dietary seaweed polysaccharide supplementation in pig used as prebiotics to positively modulate gut health and microbiota composition.

1. IntroductionGut health, which is described as a generalized condition of homeostasis in the gastrointestinal tract [1], has been recognized as playing a key role in maintaining pig health. In fact, the gut plays an important role in efficient feed digestion and absorption, for the protection of the gut barrier, the microbiota composition, and the improvement in the immune status [2]. In fact, commensal bacteria such as Lactobacilli and Bifidobacteria are necessary to sustain the host immune system, protecting against the colonization of opportunistic pathogens [3].Since the ban on in-feed antibiotics, reliable dietary interventions are needed that are capable of sustaining pig performance and improving gut health, by minimizing the use of antimicrobials. A large amount of evidence has reported the beneficial effects of some feed ingredients or additives in modulating gut health and microbiota in pig. The review by Xiong et al. [2] focused on the effects of several feed ingredients or additives such as functional amino acids, natural extracts, and short-chain fatty acids and prebiotics on gut health in weaned pigs. Over the past few decades, prebiotics have been used as an alternative approach aimed at reducing the use of antimicrobials, by positively affecting the gut microbiota and decreasing the onset of several enteric diseases in pig [4]. However, dietary supplementation with seaweed polysaccharides as prebiotics, has also gained attention in recent years. In fact, natural bioactive compounds have been considered as attractive dietary interventions in pig in order to ensure environmental sustainability, in line with the European Green Deal plan and to improve animal health according to the One Health approach.Marine macroalgae, or seaweeds, are classified as brown algae (Phaeophyceae), red algae (Rhodophyta), and green algae (Chlorophyta) and include thousands of species. The chemical composition and the bioactive metabolite content of several species have been extensively studied, along with the variations related to species and genera, harvesting season, environmental conditions, and geographical location [5,6]. Seaweeds also contain large amounts of carboxylated and sulfated polysaccharides, with important functions for the macroalgal cells including structural and energy storage [7]. Seaweed polysaccharides are safe, environmental-friendly, and economical natural polymers. Seaweed polysaccharides, such as fucoidan, laminarin, ulvan, carrageenan, and alginates, show several biological activities in vitro and in vivo studies [8,9]. In fact, polysaccharides and oligosaccharides originating from seaweeds have been shown to regulate intestinal metabolism and fermentation and reduce the adhesion of pathogenic bacteria [10]. Several seaweed polysaccharides have also shown anti-inflammatory, antiviral, and antioxidant activities [11]. Considering the above mentioned properties, the present paper reviews the prebiotic effects of seaweed polysaccharides in pig nutrition.2. Seaweed PolysaccharidesThe polysaccharides contained in brown, red, and green seaweeds present different bioactive molecules such as fucoidan, laminarin, alginate, ulvan, and carrageenan, which are reported in Table 1.The yield of seaweed polysaccharides varies in relation to the species-growing conditions, extraction method, such as solvent concentration and extraction time [14]. The polysaccharide content of brown, red, and green seaweeds is reported in Table 2. The total polysaccharide content in seaweeds is highly variable, fluctuating from 4 to 80% of dry matter (DM), according to the data of Lafarga et al. [12].In green seaweeds, the content ranges from 15 to 65% of DM with the highest value for Ulva spp., in red seaweeds from 53 to 66% of DM with the highest value in Chondrus crispus, and in brown seaweeds from 10 to 66% DM with the highest amount in Ascophyllum nodosum and Saccharina spp. [15]. Carrageenans and agars are the two main polysaccharides in red seaweeds, but porphyran and xylan have also been observed [16]. Carrageenans are sulfated polysaccharides, composed of d-galactose units, with a structural role, similar to cellulose in plants, and are present in some red algae, such as Chondrus, Gigartina, and Hypnea [17], with the highest amount in Chondrus and Kappaphycus spp. [18]. Agar is largely observed in the Gelidium and Gracilaria spp. and is composed of agarose and agaropectin [19]. Fucoidans, alginates, and laminarin are the main polysaccharides in brown seaweeds. Alginates are the main cell wall polysaccharides in brown algae, such as Laminaria spp., Fucus spp., Ascophyllum nodosum, and Macrocystis pyrifera [20]. Besides alginates, fucoidans are cell wall water-soluble polysaccharides in brown seaweeds, containing L-fucose and sulfate groups, in addition to monosaccharides such as mannose, glucose, xylose, and glucuronic acid [21].animals-11-01573-t002_Table 2Table 2Polysaccharides composition of brown, red, and green seaweed (g kg−1 DM) 1.SeaweedPolysaccharides, %AlginatesCarragenanFucoidanLaminarinUlvanReferences Brown Ascophillum nodosus 62(42–70)285(240–330)-75(11–120)118(12–120)-[8,18,22,23,24,25] Laminaria hyperborea 39.9(14.4–65.5)215(22–408)-30(20–40)125(0–320)-[8,18,26,27,28] Laminaria digitata 57.3(44–70.7)435(350–520)-49.5(22–112)120(0–350)-[8,18,26,27,29,30,31]Laminaria sp. *45(13–77)309(225–343)-147.5(22–550)153(62.4–340)-[8,26,29,32,33,34,35]Fucus sp. #57(34.5–66)162-105(11–200)2.3(0.4–3.8)-[18,23,24,27,34,36,37,38]Sargassum sp. ⁑36 (4–68)296(93–499)-38(31–45)3(0–6)-[18,34,36]Saccharina sp. **69(58–80)242.5(200–285)-33(13–80)97.5(0–330)-[5,8,18,23,27,29,39] Undaria pinnafitica 40(35–45)425(340–510)-219(30–690)30-[8,18,29,40] Red ----- Chondrus crispus 60.5(55–66)-439.5(338–510)---[18,34,41] Kappaphycus alvarezii 58(53.5–64)-448.5(187–756)---[32,41,42,43] Green - Ulva sp. §42(15–65)----176(11–400)[18,34,44,45,46,47,48]1 Data are reported as mean values and range (minimum-maximum). * Values from Laminaria claustonii and japonica. # values from Fucus vesciculosus, serratus, spiralys. ⁑ Values from Sargassum patens, hemifhyllum, henslowianum. ** Values from Saccharina longicruris, latissima, cichorioides, japonica, longissimi. § Values from Ulva armoricana, lactuca, intestinalis, meridionalis, pertusa. - Polysaccharides not present in the considered seaweed.Laminarin, also called laminaran, is a storage polysaccharide in brown seaweeds which is composed of (1–3)-β-d-glucan. The laminarin structure differs in the degree of branching and polymerization. The highest laminarin content is found in Laminaria spp. and Saccharina spp. (32% DM), however it is also present in small amounts in Ascophyllum, Fucus, and Undaria spp. [18]. Ulvan is the constituent of the cell wall of green seaweeds and is constituted by β-(1–4)-xyloglucan, glucuronan, and cellulose in a linear arrangement [49]. The ulvan content varies from 2.7% DM in Ulva flexuosa to 40% DM in Ulva Armoricana [48].3. Seaweed Polysaccharides as PrebioticCarbohydrates, which are indigestible to hydrolytic enzymes and are fermentable, are considered as prebiotics. They must not be digested or adsorbed in the first tract of the gut, however they should be fermented in the colon by Lactobacillus and Bifidobacterium, enhancing their growth and decreasing the concentration of other invading pathogens in the large intestine [50]. Digestion can affect the seaweed polysaccharide activity as prebiotics. The first step is to verify the resistance to hydrolysis by acids and enzymes in in vitro conditions.Laminarins from different seaweeds vary in terms of the structural characteristics such as the degree of polymerization and the presence of inter-chain hydrogen bonds. These complex structures are resistant to hydrolysis in the first tract of the gut and are studied as dietary fibers [51]. In brown seaweed, polysaccharide laminarins were indigestible in an in vitro model with hydrochloric acid and enzymes [52]. In addition, laminarin from Laminaria saccharina and digitata were fermented, producing short-chain fatty acids (SCFA) [53]. Another study reported that SCFA that are produced from the fermentation of Laminaria digitata and Undaria pinnatifida are not metabolized well compared to the sugar beet fibers [54]. In a simulated digestion model, Ascophyllum nodusum was fermentable, thus producing SCFA and reducing the concentration of total anaerobe bacteria such as Escherichia coli and Streptococcus in batch systems inoculated with porcine cecal suspensions [55]. The same result was obtained in the Porphyra spp., a red seaweed with galactoside, which was not digested by salivary, gastric, pancreatic, or intestinal enzymes and was fermentable by Bifidobacterium [56]. Alginate, agarose, and carrageenan, from brown and red seaweeds are not digested but fermented by the gut bacteria [57]. Oligosaccharides enzymatically hydrolyzed from alginate, agarose, and carrageenan have shown prebiotic activities, stimulating Bifidobacteria and Lactobacilli growth and producing SCFA [58,59]. A recent in vitro study evaluated the fermentability of the three aforementioned oligosaccharides using pig fecal microbiota. The data reported that all the oligosaccharides considered were able to enhance SCFA concentration, in particular butyric acid. A positive shift in gut microbiota was also observed for alginate and agarose oligosaccharides with a decrease in pathogenic bacteria [60].Ulvan, the main sulphated polysaccharide in green seaweeds, is a water-soluble dietary fiber resistant to digestive tract enzymes, however it is poorly fermented by colonic bacteria and thus in its form, it is not studied as a prebiotic, but could be hydrolyzed to bioactive oligosaccharides [61]. Low molecular weight Ulva armoricana aqueous extract, whose main component is ulvan, has shown antibacterial activity against several Gram-positive and Gram-negative bacterial strains [62].Several in vivo studies in rats, highlighted the prebiotic activity of seaweed polysaccharides, as reviewed by O’Sullivan et al. [63]. In fact, it is reported that in rats, dietary supplementation with alginate oligosaccharides or agarose hydrolysate increased the cecum Bifidobacterium and Lactobacillus count. Another study on rats reported that dietary supplementation with Chondrus crispus increased the SCFA concentration and reduced the Clostridium and Streptococcus concentration [64]. In mice fed Laminaria japonica, a higher production of SCFA and butyric acid was observed, together with a decrease in Clostridium, Escherichia coli, and Enterobacter [65]. Considering the data on in vitro studies and laboratory animals, several studies on dietary supplementation with seaweed polysaccharides have been conducted in pigs to improve microbiota composition, and reduce the onset of several diseases and the subsequent antibiotic treatment. 4. Seaweed Polysaccharides as Prebiotics in SowsThe effects of algae polysaccharides as prebiotics in sows have been evaluated by several authors as presented in Table 3. The effects of polysaccharides in the gut are usually assessed by evaluating the SCFA content and the intestinal microbiota composition and/or the presence of beneficial bacteria [72]. The effects of dietary supplementation with seaweeds in sows can modulate the productive performances and health of lactating piglets, making them more resistant to pathogens.Dietary brown seaweeds were evaluated as a prebiotic in field conditions. In the study by Leonard et al. [66] an antimicrobial effect of seaweeds was observed. In fact, the establishment of neonatal gut microbiota was mediated by the fecal microflora of sows or by the colostrum and milk composition. This early gut colonization is fundamental for the development of healthy microbiota and for modulation of the immune system [73].In piglets after 9 days of weaning, a decrease in Escherichia Coli was observed, however the same trend was also detected for Bifidobacteria and Lactobacillus. Moreover, a positive effect on the intestinal morphology was observed in the treated piglets. Similar result on the modulation of the microbiota was observed in a subsequent study from the same authors [67]. A reduction in fecal Escherichia coli in sows may positively affect piglet’s microbiota with a lower Enterobacteriaceae count which influences the onset of enteric pathologies [74]. In fact, although the Enterobacteriaceae count was higher in sows receiving dietary laminarin and fucoidan, the diarrhea score was lower in piglets from treated dams [68]. A positive impact on the ileum morphology of piglets from treated sows has also been observed, with an increase in villi height at weaning. In the study by Bouwhuis et al. [70] a lower expression was observed of heat-labile enterotoxin and enteroaggregative heat-stable enterotoxin, which are responsible for the intestinal colonization of enterotoxigenic Escherichia coli and represent the most frequent cause of diarrhea. It has been reported that laminarin, which is a source of β-glucans, lowered intestinal Enterobacteriaceae and reduced Escherichia coli numbers in feces. This could be ascribed to the potential agglutination activity, as earlier observed for yeast β-glucans which prevent bacterial adhesion and the colonization of the epithelial mucosa [75,76].An antibacterial activity of fucoidan has also been reported [77]. An increase in Lactobacilli count was also observed which was also shown to be helpful for gut health with a positive modulation of the immune system [78,79].In the considered studies the effects of laminarin and fucoidan from Laminaria spp. on Lactobacilli population were inconsistent.5. Seaweed Polysaccharides as Prebiotics in Post Weaning PigletsWeaning is a critical phase in pig production, often characterized by high antibiotic and microelement use. In fact, at weaning the gastrointestinal tract and immune system of piglets are not yet fully developed and the social, environmental, and physiological challenges, predispose the piglets to dysbiosis [80]. These challenges lead to a lower feed intake and growth rate and a high incidence of post-weaning diarrhea (PWD) due to the presence of enteric pathogenic bacteria [81].In fact, at weaning, a lower Lactobacilli count has been observed, with a high growth of facultative anaerobes bacteria such as Enterobacteriaceae, Proteobacteriaceae, Clostridiaceae, and Prevotellaceae [80]. After weaning, structural and functional alterations of the small intestine have also been observed with negative effects on the absorptive capacity [82].Feeding strategies in the post-weaning phase can reestablish the gut eubiosis that was lost at weaning, aimed at restoring the Lactobacillus count, promoting the growth of beneficial bacteria that boost the mucosal immune system and lowering the pathogenic bacteria proliferation [83].The role of diet in the post weaning health status is widely recognized, in fact feed ingredients and additives can exert selective pressure on the gut microbiota. It has also been reported that dietary fermentable carbohydrates play a key role in positively affecting the intestinal microbiota of post-weaning piglets [84].Several studies have evaluated the effects of seaweed polysaccharides as prebiotics in post weaning piglets, as reported in Table 4.The dietary inclusion of Ascophillum nodosum in the piglets’ diet can reduce the Escherichia Coli content in the small intestine of weaned piglets [85]. The Lactobacillus/Escherichia coli ratio in the small intestine was shown to increase in the piglets receiving dietary seaweeds suggesting a helpful microbial modification. A reduction in the Enterobacteriaceae count was also observed. These are opportunistic pathogens such as Salmonella Typhimurium and enterotoxigenic Escherichia coli (ETEC) that induce PWD in piglets [85]. No effects of dietary Ascophillum nodosum on the small intestinal morphology were observed but an increase in total SCFA and butyric acid content were reported [85,86]. Similar data on gut health improvement have been observed with dietary supplementation with Laminaria spp.Laminarin and fucoidan, as sources of seaweed polysaccharides with prebiotic effects, are able to decrease fecal Escherichia coli counts in the feces, thus improving post-weaning piglet health with a positive effect on growth performance and gain to feed ratio [88,90]. An improvement in Lactobacillus count has also been detected [88,89,91,92,93].It has been also reported that laminarin, modifying the resident microbiota, may indirectly enhance mucin synthesis and secretion, as adherence of beneficial bacteria to mucosal epithelia up-regulates the mucin production. An enhancement of cytokine gene expression was also observed after a lipopolysaccharide (LPS) challenge [101]. Fucoidan also supports Lactobacillus growth with a positive effect on feed digestibility [87,101].The increase in butyric acid reported in several studies, is usually related to carbohydrate fermentation which has a positive effect on gut health [87,91,92,97]. However, in the study by Sweety et al. [91], after a Salmonella typhimurium challenge, proliferation of Salmonella shedding was observed after dietary supplementation with laminarin and fucoidan. This is probably related to the decreased content of propionic acid and increased content of acetic acid which can modulate the gene activation in Salmonella pathogenicity Island 1 (SPI−1) [102]. A recent study showed that the alginic acid oligosaccharide from seaweed, which contains 96% α-L-guluronic acid and 4% β-D-mannuronic acid, has antibacterial and anti-biofilm activities. These oligosaccharides can modulate mucosal cytokine expression and antibody production and promote the growth of Lactobacillus [103]. As recently reported, dietary seaweed alginic acid oligosaccharides in weaned piglets can boost the intestinal barrier integrity, upregulating occludin mRNA expression in the caecal mucosa and claudin—1 mRNA expression in the caecal and colonic mucosa [98].A recent study in piglets also showed that supplementing a seaweed product in drinking water (Algo-Bio®) decreased the Enterobacteria resistance rate to tetracycline [104]. In addition, in piglets treated with a seaweed supplement, the antibiotic susceptibility test showed a high sensitivity of Escherichia coli to imipenem, amikacin, and netilmicin with low chloramphenicol resistance. In fact, in agreement with the data of Berri et al. [105], the use of seaweed in pig nutrition to overcome antibiotic resistance is promising.6. Seaweed Polysaccharides as Prebiotics in Growing-Finishing PigsTable 5 reports the effects of seaweed carbohydrates as prebiotics in growing-finishing pigs. The antimicrobial properties of seaweed polysaccharides in growing-finishing pigs could be effective in reducing pre-slaughter enteric pathogens. In fact, several studies have reported a lower enteric Escherichia coli count in growing-finishing pigs fed Ascophillum nodosus or Laminaria spp. [101,106,107]. Dietary seaweed polysaccharides reduced intestinal and fecal Salmonella Typhimurium and lowered the colonic gene expression of pro-inflammatory cytokines in growing-finishing pigs after an experimental Salmonella Typhimurium challenge [108,109]. The effects of Laminaria digitata extract in wet, spray-dried, and freeze-dried forms have also been evaluated [110,111]. The laminarin and fucoidan from L. digitata in wet form showed a prebiotic effect in pig, enhancing the Bifidobacterium sp. count in the ileum. In fact, several technological processes, such as spray drying techniques, have been shown to reduce the functionality of some bioactive compounds [112]. An increase in the colon and cecum SCFA was also detected following dietary supplementation with laminarin [101] and fucoidan [113]. In the growing finishing phase, the decline in enteric Escherichia coli and Salmonella Typhimurium count and the improvement in beneficial bacteria, such as Bifidobacterium and Lactobacillus, suggest that brown seaweed polysaccharides are a sustainable dietary strategy to improve gut health, by modulating microbiota and reducing pre-slaughter pathogens.animals-11-01573-t005_Table 5Table 5Prebiotic effects of seaweed in growing finishing swine.Seaweed SupplementDoseAnimalPrebiotic EffectEffect vs. Control %Reference Ascophyllum nodosum Dried intact x (3–6–9 g kg−1)Pig,48.7 kg LWIleum:Coliform Lattobacillus Adherent Lactob.Biphydobacteria−19%NSNSNS[106] L. digitata Wet (W) orspray dried (SD)seaweedPig14.5 kg LWIleum:BifidobacteriumLactobacillusEnterobacteriaceae+18% wetNSNS[110] L. digitata Wet (W) orspray dried (SD)seaweedLAM + FUC(0.50–0.42 g kg−1)Pig,14.5 kg LWIleum: BifidobacteriumLactobacilli spp.+45% SD+66% W+31% SD+47% W[111] L. digitata Seaweed extract yLAM (0.3–0.6 g kg−1)Pigs, 18 kg LWColon:EnterobacteriaceaeCaecum:SCFA−29% LAM 0.3−22% LAM 0.6+9% LAM 0.3+12% LAM 0.6[101] L. hyperborea Seaweed extractLAM (0.30 g g kg−1)FUC (0.36 g kg−1)LAM + FUC(0.3 + 0.36 g kg−1)Boars,65 kg LWProximal colon:Enterobacterium spp.Lactobacilli spp.SCFADistal colon:Enterobacterium spp.Lactobacilli spp.SCFA+14% LAM + FUC+16% FUC+16% FUC+30% LAM + FUC+43% FUC+36% FUC[113] L. hyperborea Seaweed extract z(0.7–1.4–2.8–5.6 g kg−1)LAM = 0.08–0.16–0.32–0.64 g kg−1FUC = 0.06–0.12–0.24–0.48 g kg−1Boars,76 kg LWCaecum:Bifidobacterium spp.Enterobacterium spp.Colon:EntorobacteriaBifidobacterium spp.Lactobacilli spp.−3, +5, +2, +3%+3,− 4, −7, −4%−3, −9, −7, −2%+3, +5, +2, −3%0, −1, −2, −3%[107]Laminaria spp.Seaweed extractLAM + FUC(0.18 + 0.34 g kg−1)Female pigs, 30.9 kg LWChallenge Salmonella Typhimurium 11 dCaecum:S. TyphimuriumLactobacillus spp.ColonS. TyphimuriumLactobacillus spp.−16%+3%−24%+1%[108]Laminaria spp.Seaweed extractin dietLAM + FUC (0.18 + 0.34 g kg−1)Female pigs,30.9 kgChallenge Salmonella Typhimurium at 11 dSalmonella TyphimuriumCaecumColon−16%−24%[109]LAM: laminarin; FUC: fucoidan; SCFA: short-chain fatty acids. x The Ascophillum Nodosum seaweed extract (Kerry Enhancer, Kerry Algae, Curraheen, Tralee Co., Kerry, Ireland). y Purified laminarin (990 g/kg) laminarin was sourced from Bioatlantis Limited, Tralee, County Kerry, Republic of Ireland. z The seaweed extract contained laminarin (112 g kg−1), fucoidan (89 g kg−1).7. ConclusionsOverall, the present data demonstrate that the prebiotic activity of seaweed polysaccharides could be used to improve pig health in several production phases, thus reducing the use of antimicrobials. The gut health enhancement and reduction in pathogenic bacteria in lactating and post weaning piglets are fundamental in modulating productive performance and health, by reducing gastrointestinal diseases and enhancing feed digestibility. In the growing-fattening phase, a modulation of microbiota and a reduction in pre-slaughter enteric pathogens have also been observed.In conclusion, dietary supplementation with brown seaweed polysaccharides seems to be a valid strategy to modulate microbiota, making pigs more resistant to pathogens and thus reducing antimicrobial use. Considering that prebiotic and antibacterial activities have also been observed for red and green seaweed polysaccharides, further studies are needed to evaluate their effects on pig health.

animals : an open access journal from mdpi

10.3390/ani11092549

PMC8472153

New technologies allow researchers to improve the methods for immediate, accurate data collection, cleaning and analysis, with minimal geographical limitations. Although much has improved in the field of equine anaesthesia in recent years, we are still far from reducing anaesthetic-related mortality in this species in comparison with small animal anaesthesia. The aim of this multicentre study was to probe the usefulness of an internet-based method that utilised an electronic questionnaire and statistical software to show the data and report outcomes from horses undergoing general anaesthesia and certain procedures using standing sedation. Within six months, 8656 cases from 69 centres were collected: 6701 procedures under general anaesthesia and 1955 under standing sedation. The results demonstrated (i) the utility of the method and (ii) that some horses died unexpectedly when undergoing not only general anaesthesia, but also standing sedation. Finally, (iii) we present some descriptive data that outline the current anaesthesia practice compared with the previous CEPEF2. We concluded that our internet-based method is suitable for this type of study. New techniques may reduce the mortality rate. However, the results presented here should be interpreted cautiously as these are only preliminary data with lower numbers than CEPEF2.

It is almost 20 years since the largest observational, multicentre study evaluating the risks of mortality associated with general anaesthesia in horses. We proposed an internet-based method to collect data (cleaned and analysed with R) in a multicentre, cohort, observational, analytical, longitudinal and prospective study to evaluate peri-operative equine mortality. The objective was to report the usefulness of the method, illustrated with the preliminary data, including outcomes for horses seven days after undergoing general anaesthesia and certain procedures using standing sedation. Within six months, data from 6701 procedures under general anaesthesia and 1955 standing sedations from 69 centres were collected. The results showed (i) the utility of the method; also, that (ii) the overall mortality rate for general anaesthesia within the seven-day outcome period was 1.0%. In horses undergoing procedures other than exploratory laparotomy for colic (“noncolics”), the rate was lower, 0.6%, and in “colics” it was higher, at 3.4%. For standing sedations, the overall mortality rate was 0.2%. Finally, (iii) we present some descriptive data that demonstrate new developments since the previous CEPEF2. In conclusion, horses clearly still die unexpectedly when undergoing procedures under general anaesthesia or standing sedation. Our method is suitable for case collection for future studies.

1. IntroductionInternet-based research has become commonplace during the last 10–20 years and makes online data collection possible from a large pool of participants with few geographical limitations. Online data collection is quick, cheap and increases the accuracy and efficacy of data entry [1]. Moreover, data can be analysed interactively with the ability to follow-up with participants [1,2,3]. Apart from understanding how to use new technology and the expertise required to match the study design with data collection, adequate statistical software for data analysis and cleaning is mandatory [4,5]. Immediate technical support and fluent communication between the administrators and participants is also crucial. The high risk of mortality associated with general anaesthesia in horses remains one of the biggest concerns for equine practitioners and veterinary anaesthetists. Many studies report anaesthesia-related mortality risks. Most of these studies are retrospective [6,7,8,9,10,11,12,13,14] but a few prospective single-centre [15] and some prospective, multi-centre [16,17] investigations have been carried out. To date, the Confidential Enquiry into Perioperative Equine Fatalities 2 (CEPEF2) published in 2002 remains the largest observational, multicentre study with 41,824 cases collected from 62 clinics over a period of 6 years [18]. The overall death rate up to seven days was 1.9%, 0.9% in noncolics and 7.8% in colics [18]. Although much has changed since then, we are still far from reducing these numbers [19], and the need for an update on the CEPEF data was proclaimed eight years ago [20]. Avoiding general anaesthesia by undertaking some procedures in standing horses may reduce the mortality, but there are no data as yet to support this hypothesis. The first objective of this report was to describe the usefulness of an internet-based method for data collection and a strategy for data cleaning for the CEPEF4 study, whose final aim is to identify the risk factors associated with equine anaesthesia and standing sedation. Second, to report the outcomes/fatalities within a period of seven days of equidae undergoing (i) general anaesthesia and (ii) certain procedures performed under standing sedation during the first six months of CEPEF4 and to describe the key preliminary findings that will be analysed in detail in the future. Our first hypothesis was that the proposed method would be a fast, reliable tool to collect, clean and analyse data. Second, that mortality was lower than reported by Johnston et al. almost 20 years ago [18] and that standing sedation procedures are not exempt from the risk of death, although it is reduced. Finally, we hypothesised that there would be trends towards new practices in anaesthesia and analgesia compared with the previous CEPEF2.2. Materials and MethodsThe study design was multicentre, cohort, observational, analytical, longitudinal and prospective. For this phase of preliminary results, the six-month data collection period was from 1 November 2020 until 30 April 2021.Inclusion criteria comprised cases of horses, donkeys and mules of all ages from clinics all over the world, recruited specifically for this project. The equidae involved were all cases from each clinic undergoing (i) general anaesthesia whatever the reason and (ii) standing sedation for surgery or advanced diagnostic imaging (magnetic resonance imaging (MRI), computed tomography (CT) or scintigraphy) requiring a continuous rate infusion (CRI) or at least one extra top-up apart from the initial sedation bolus. Once recording started in any one clinic, all cases had to be included. However, to allow for holidays and staff absence, all cases within a specified period could be omitted; recording and sending all cases started again after the specified period.Exclusion criteria were cases outside the agreed recording periods, cases from centres that did not follow the communication process and cases from centres that did not send all the cases from a recording period. General anaesthetics for terminal procedures were excluded, as well as standing sedations without top-ups or CRIs, and standing sedations for reasons other than surgery or advanced diagnostic imaging, such as cast changes or sinuscopies.The steps followed in this study are detailed below:(1)Preparation of the questionnaireA user-friendly, online questionnaire used in small animals [21] was adapted for this equine study using the feedback received from a group of researchers and clinicians with special interest in equine anaesthesia and analgesia [22]. It was designed to be used to compile information for both general anaesthesia and standing sedation. Briefly, the questionnaire collected data about the centre, the level of training of the responsible anaesthetist, the patient, the procedure, the anaesthetic and analgesic protocols and other details about management of the anaesthesia or sedation. Information about the anaesthetic induction and recovery phases was included, with the potential to indicate any intraoperative complications and details of the postoperative period for up to seven days. The questionnaire and the instructions to complete it are available on a website created for support and promotion at https://cepef4.wordpress.com (accessed on 6 June 2021), and are provided here as Supplementary Materials S1.(2)Definition of perioperative equine fatalitiesThe same classification of fatalities was followed as used in previous CEPEF studies [16,18]. After induction of general anaesthesia or the first sedation bolus for standing sedation, the outcome was recorded at day seven for each case as (i) alive (or discharged), (ii) put to sleep (PTS)/euthanised or (iii) dead. The time of PTS or dead was recorded as premedication, induction, maintenance, recovery or the day up to day seven. Each centre was encouraged to use an outcome logbook to communicate further which colic or noncolic cases were alive, PTS or dead during this seven-day period.When a horse was euthanized due to an inoperable lesion found at surgery, due to pre-existing disease or financial constraints, the outcome was classified as PTS. When a horse died unexpectedly or was euthanised due to a perioperative complication, such as a fracture in recovery, the outcome was classified as death. Other examples of death include intraoperative cardiac arrest, spinal cord malacia requiring euthanasia during recovery or post-operative myopathy requiring euthanasia on humane grounds up to seven days after surgery. Those classified as noncolic deaths were required to complete a more detailed online survey at https://edinburgh.onlinesurveys.ac.uk/equine-ga-mortality-form (accessed on 6 June 2021).Classification of outcomes was performed by M.G.-M. and J.I.R., and later confirmed by R.B.-W., M.J. and P.M.T. Ultimately, before final data analysis, the outcomes were confirmed with the contact person from each centre, referred to as the ambassador (see (10) in Methods).(3)Ethics statementThe study was approved by the international Ethical Review Committee of the Association of Veterinary Anaesthetists (AVA), under protocol 2020-009 on 4 September 2020. Full details of the application form were provided to any centre that also required Local Ethical Committee Approval from their institution.(4)Recruitment of collaborating centresRecruitment included an abstract presented at the AVA Spring meeting in Dublin 2020 [23], and Correspondence to the Editor published in peer reviewed journals targeting equine practitioners [24] and veterinary anaesthetists [22], respectively. Finally, we also used our professional network and contacted several centres personally. (5)The Ambassador figure and the Agreement FormAmbassadors for each centre volunteered or were recruited to be the contact person who also took responsibility for the centre’s data collection.Before sending the data, the ambassador signed an agreement form (https://cepef4.wordpress.com/forms-instructions-and-help/ (accessed on 6 June 2021)) ensuring to cooperate/supervise selflessly, providing good quality data including all the cases within a period of time to avoid bias. The CEPEF team certified the anonymous handling of the data.(6)Anonymity and confidentiality of each patient, owner and centreAs in previous CEPEF studies, the anonymity and confidentiality of the patients, owners and centres was ensured. Identification of each patient was provided by a number given by the centre for further communication with the CEPEF team. These numbers were used only for communication with each centre. Each case entered in the database received a unique CEPEF number for data handling and analysis. Moreover, each centre had a code known only by M.G.-M. and J.I.R. The remaining authors, R.B.-W., J.I.R. and P.M.T., were not aware of centre coding for further evaluation of the data. (7)The communication process (recruitment and follow-up meetings)We instituted a method of interactive communication with the ambassador of each centre always performed by the same investigator (M.G.-M.) in order to standardize the information and, therefore, reduce data inconsistencies. For each centre, the project was first presented to the ambassador, and ideally, also to the whole team involved in data collection, via an online meeting (Teams® (Microsoft Corporation, Redmond, WA, USA), Zoom® (Zoom Video Communications Inc., San Jose, CA, USA), Skype® (Microsoft Corporation, Redmond, WA, USA), Meet® (Google Inc., Mountain View, CA, USA)) of about one hour’s duration. This included a standard presentation of the project, introduction to the study website and explained the questionnaire and its technicalities. At least a second, and ideally a third, follow-up meeting was subsequently organised to ensure complete familiarity with the system. (8)Data collection and storageThe electronic questionnaire was used to collect the data (Supplementary Materials S1). A .pdf file was programmed using Adobe Acrobat Pro® (Adobe Inc., San Jose, CA, USA). This .pdf file can be completed using the free app Adobe Acrobat Reader DC (Adobe Inc., San José, CA, USA), available for the different operating systems (Android® (Google Inc., Mountain View, CA, USA), IOS® (Apple, Cupertino, CA, USA), Windows® (Microsoft Corporation, Redmond, WA, USA) and macOS® (Apple, Cupertino, CA, USA)). The questionnaire can be completed on any internet-connected device, including mobile phones, tablets, laptops or computers. Once completed, the questionnaire is sent by e-mail to a specific e-mail account in which all the questionnaires are stored. The collected metadata is converted to a .csv file that is added to the main database as a new case. (9)Data cleaning and statisticsData processing and statistical analyses were performed using R 4.1.0 [25]. Data quality was improved using an initial data cleaning phase performed with specific scripts searching for inconsistencies. R scripts were introduced to detect duplicated cases or blank fields as non-available data (-NA-) in the studied variables. Other R scripts detected more sophisticated inconsistencies using Boolean algebra, for instance, in general anaesthetics coded as inhalation anaesthesia only (INH) but where an alpha2-agonist CRI was ticked, or when an inhalant agent was ticked for a standing sedation procedure. In addition, further R scripts were used to detect uncommon anaesthetic practices, such as induction with an inhalant agent or when premedication was not administered. The full list of the R scripts used in this study can be found in the Supplementary Materials S2.The ambassador of any centre where data inconsistencies were detected was contacted and sent an Excel® (Microsoft Corporation, Redmond, WA, USA) file by e-mail to confirm whether the inconsistencies were real and to request revision if so. These detected inconsistencies were corrected manually within the database, only with the strict agreement of the ambassador.Once data were cleaned, a descriptive analysis was performed. Tables and figures were created using the following R packages: table1 (v1.4.1) [26], ggplot2 (v3.3.3) [27] and rnaturalearth (v0.1.0) [28]. Variables that followed a normal distribution are shown as mean ± standard deviation, whereas those non-normally distributed are shown as median [range].(10)Final meeting with each ambassador to double-check data and outcomesThe individual database from each centre was sent to the ambassador of the centre before the final meeting. This included all the cases sent up until 30 April 2021. A personalized report for the centre was included for its approval. Data were always double-checked by the principal investigator (M.G.-M.) and the ambassador for each centre during a scheduled online meeting of about 30 to 90 min, depending on the volume of the data and potential inconsistencies. The ambassadors were asked for feedback and to grade the project and the communication process between 0 (worse) to 10 (excellent). These meetings occurred between Monday 17 May and Thursday 3 June 2021.Figure 1 shows the flow diagram of the process.(11)Reporting the resultsThe STROBE-Vet guidelines, an extension of the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) statement [29], as recommended for reporting of observational studies in veterinary medicine (https://strobevet-statement.org, accessed on 6 June 2021), (Supplementary Materials S3) were followed in order to maximise reporting quality.3. ResultsDuring the specified first six months, 69 centres from 20 countries in four continents collaborated to collect data (Figure 2, Table 1). As a result of continuous recruitment, each centre had a different starting date for data collection.Figure 3 shows the flow diagram describing the cases included or excluded from the current analysis. The data cleaning process detected 193 empty data fields and 1310 inconsistencies. After a discussion with the ambassador of each centre, 50 inconsistencies were redeemed, but were noted as unusual practices. Sixty two empty data fields and 1260 inconsistencies were corrected. For standing sedations, 131 durations were not found in the records of the centres and were coded as missing data. After data cleaning, 8752 cases (8656 horses, 92 donkeys and 4 mules) were confirmed. The median and range of cases received per day were 48 [1–146]. These preliminary results contain horse information only as the sample size for donkeys and mules is too small at this stage.Of the 8656 horse cases, 6701 were general anaesthetics and 1955 were standing sedations. Of these, only 39 cases were procedures in the field, 31 total intravenous anaesthetics and 8 standing sedations. The demographic data for general anaesthetics and standing sedation are shown in Table 2.3.1. Perioperative Equine FatalitiessSixty six of the 6701 horses that underwent general anaesthesia were classified as having died (1.0%) (confidence interval 95% (CI 95%): 0.76–1.25%). Of these, 35 (out of 5784) were classified as noncolic deaths (0.6%) (CI 95%: 0.42–0.84%) and 31 (out of 917) as colic deaths (3.4%) (CI 95%: 2.3–4.8%). Four of the 1955 horses undergoing standing sedation died (0.2%); all were noncolic surgeries (0.2%) (CI 95%: 0.06–0.52%). Of the 6701 horses that underwent general anaesthesia, 329 were classified as PTS (4.9%) (CI 95%: 4.4–5.5%): 76 of 5784 as noncolic PTS (1.3%) (CI 95%: 0.88–1.45%) and 253 of 917 as colic PTS (27.6%) (CI 95%: 25.0–31.0%). From the 1955 horses undergoing standing sedation, 13 were PTS (0.7%) (CI 95%: 0.35–1.13%): 11 noncolic surgeries out of 1949 (0.6%) (CI 95%: 0.28–1.01%) and 2 out of 6 colic surgeries (33.3%) (CI 95%: 4.3–77.7%).The outcomes at seven days post-anaesthesia or standing sedation are shown in Table 3. Details of the deaths classified as noncolics are presented in Table 4. Table 5 indicates the time of death or PTS of all the cases whether general anaesthesia or standing sedation was used. 3.2. General AnaesthesiaThe centres sent a median of 100% of cases (100 [95–100]%). Of the 69 centres, 63 confirmed that 100% of the cases were sent, three sent more than 98% and three sent more than 95%.Table 6 gives the details of the individual drugs used at each phase of general anaesthesia. The protocols for the general anaesthetics are described in Table 7, Table 8, Table 9, Table 10 and Table 11. Information about the method of induction and recovery from general anaesthesia is shown in Table 12 and Table 13, respectively. Figure 4 describes the monitoring of the horses under general anaesthesia. 3.3. Standing SedationFrom the 69 collaborating centres, 57 sent standing sedation cases that fulfilled our inclusion criteria (median 99 [10–100]%). Forty-three centres sent more than 90% of their cases, four sent between 80 and 90%, five sent between 50 and 80% and five sent between 1 and 49%. The drugs and protocols for standing sedations are shown in Table 14 and Table 15, respectively. Table 16 reports the different CRIs used to maintain standing sedation. Figure 5 describes the monitoring of the horses under sedation.3.4. Results of Survey for Feedback in Final MeetingThe ambassadors of the collaborating centres gave a final mark for the project and the communication process of median 10 [7–10] (0—worse to 10—excellent).4. DiscussionThe results presented here demonstrate the utility of the electronic questionnaire and the internet-based method for data collection, with interactive data handling and cleaning for the performance of the CEPEF4 study. The data collected from the 1 November 2020 until the 30 April 2021 show that overall, 1% of horses undergoing general anaesthesia still die within the seven-day outcome period. Even with this preliminary evaluation of a small number of horses, it appears that a fatal outcome in both noncolics (0.6%) and colics (3.4%) was less frequent than that reported by Johnston et al. almost 20 years ago [18]. As anticipated, and despite the higher risk of bias collecting standing sedation data, the risk of death with standing sedation (0.2%) appears lower than with general anaesthesia. However, even with a lower risk, some noncolic horses died unexpectedly when undergoing standing sedation within the seven-day outcome period. The data also demonstrate certain changes in the routine anaesthetic practice and protocols that have developed since CEPEF2.In light of our findings, and focussing on our main objectives, the discussion is structured in the following four parts: (1) the proposed method for data collection with interactive data handling and cleaning, (2) the outcomes of horses that underwent general anaesthesia and standing sedation, (3) the preliminary data from horses undergoing general anaesthesia and (4) the preliminary results of standing sedation in horses. (1)The proposed method for data collection with interactive data handling and cleaningOur first hypothesis was confirmed as the proposed internet-based method proved to be a reliable, easy, quick and cheap means of collecting data, with minimal geographical limitations. This preliminary phase of CEPEF4 was initiated and executed during the COVID-19 pandemic but our methodology allowed us to communicate with the collaborating centres and to collect and handle data arising from many locations worldwide. The feedback from the ambassadors of each centre indicated that the communication process (recruitment, follow-up and final meetings) allowed fluent communication that undoubtedly contributed to the collection of good quality data [2]. Once collected, our R scripts detected many inconsistences that were reduced almost to zero prior to final submission. The three steps followed for this data cleaning proved effective: first, data screening to detect missing or excess data, outliers, inconsistencies or strange patterns. Second, diagnosis to detect missing data, true extremes or true errors. Finally, data editing or treatment to correct, delete or leave unchanged the detected inconsistencies [5] as agreed with each centre’s ambassadors. Even with this large dataset, we achieved a reliable method with a clear strategy for communication and data collection/cleaning. This approach will help to minimize inconsistencies that might lead to data misinterpretation for the current CEPEF4 study. For the first six months, we compiled 8656 horse cases: 6701 general anaesthetics and 1955 standing sedations. CEPEF1 collected 6255 general anaesthetics in a period of two years (February 1991–March 1993), using paper copies and communication by post and telephone [16]. This reflects the importance of new technologies that allow interactive communication, data collection and analysis. Furthermore, the rapid availability of a community of researchers and equine clinicians familiar with the CEPEF studies made this work possible in such a short period of time. The impact of the CEPEF2 study, cited by 448 other publications (according to Google Scholar on the 29 August 2021), reflects the importance of an update on these data [20]. With the first hypothesis confirmed, our final goal is to use this methodology to collect approximately 45,000 cases of general anaesthetics for CEPEF4, to increase the statistical power and to compare the results with those of 20 years ago [18]. In order to ensure a robust comparison, we followed the same approach as in CEPEF2. Both M.J. and P.M.T., authors of the previous series, carefully supervised the study design and the subsequent implementation of data collection and cleaning. On the basis of the cumulative cases collected per week to date, the proposed cases numbers should be reached in approximately two years.(2)Outcome of horses under general anaesthesia and standing sedationAs reported in CEPEF2 [18], the overall equine anaesthetic mortality rate is still higher than in other veterinary species such as dogs and cats [30]. However, these preliminary data from a small population do suggest that the current rate is lower than 20 years ago. As stated in 2016 by Dugdale and Taylor [19], “we still lose horses after anaesthesia to a range of catastrophes that would not occur if the horse were not anaesthetized”. Our preliminary results confirm this statement. We recorded 39 horses that were classified as noncolic deaths: 34 general anaesthetics, four standing sedations and one that started as standing sedation and later changed to general anaesthesia. Focusing on the noncolics undergoing general anaesthesia, with the exception of 17 cases classified as moderate- or high-risk patients (ASA III, IV or V), the remaining 18 were healthy (four ASA I, 14 ASA II). This indicates that noncolic horses undergoing general anaesthesia still die unexpectedly; 0.6% at this stage, versus the 0.9% of CEPEF2. Further recruitment of cases to achieve our final CEPEF4 goal should help to clarify why this occurs and hopefully suggest how these numbers could be reduced. With regard to colics undergoing general anaesthesia, our preliminary data recorded 3.4% that died, compared to the 7.8% (457 colic deaths out of 5846) in CEPEF2. This relative reduction may indicate an improvement in the peri-operative management of horses with colic. However, the number of colic PTSs remain extremely high, 27.6% in this data versus the 25.2% of CEPEF2. It is generally assumed that standing procedures are safer than general anaesthesia in horses. However, this is the first study to report mortality rates associated with standing sedation in a multicentre, prospective, cohort study. The four standing sedations (one ASA I, one ASA II and two ASA III) classified as noncolic deaths (0.2%) provide evidence that standing sedation procedures still carry a risk, albeit less than general anaesthesia.(3)Preliminary results for general anaesthetics in horsesOne third of the horses were premedicated with a combination of an alpha2-agonist and a partial/agonist-antagonist opioid, mainly butorphanol. When acepromazine was also given, the percentage increased to 48%. Alpha2-agonists with a pure opioid, with or without acepromazine, was used in about 26% of the cases. Only 13% of the cases received an alpha2-agonist alone. This is in contrast with the CEPEF2 data, in which 43% of the horses received an alpha2-agonist alone, 29% combined acepromazine with an alpha2-agonist and no more than 8% included opioids such us methadone or butorphanol [18]. Our data show the tendency for premedication with drug combinations rather than a single agent. Drug combinations calm the horse, enhance sedation and provide analgesia by employing low doses of each drug, which reduces their potential side effects [31,32,33]. Compared with around 16% in CEPEF2 [18], premedication with acepromazine alone has not been reported in the current study.Ketamine was the most commonly used intravenous agent for the induction of general anaesthesia, usually combined with a benzodiazepine as a central muscle relaxant (88%). Diazepam was frequently used, and increasingly midazolam, which was not reported at all in CEPEF2. Midazolam has recently gained Market Authorisation for use in horses in Europe, which would support its increased use. Thiopental and guaiphenesin, an intravenous anaesthetic and a central muscle relaxant, respectively, were rarely used. This combination was historically the evolution from chloral hydrate and pentobarbital [34]; however, the often prolonged and violent recovery from barbiturate anaesthesia provoked the transition towards ketamine [35]. A few techniques were popular in some individual clinics, although the overall numbers were not high. For example, the induction and brief maintenance of general anaesthesia in foals with propofol, sometimes combined with ketamine [36]. Isoflurane was the most common inhalation agent used for the maintenance of general anaesthesia (88%), followed by sevoflurane (9%) and desflurane (3%). None of the anaesthetics reported in these preliminary results used halothane in contrast with the previous CEPEF studies [16,18]. In 2004, halothane was considered an acceptable anaesthetic for the maintenance of anaesthesia in horses [37]; however, it is no longer manufactured and is now rarely used. Our data also showed the current tendency to use an inhalant agent in combination with an intravenous CRI, the so-called partial intravenous anaesthesia (PIVA) [38,39,40]. Out of all the 6000 inhalant-based general anaesthetics, 3718 were PIVA (62%) versus 2282 (38%) that were pure inhalation anaesthesia. Since the publication of CEPEF2 nearly 20 years ago, there have been numerous reports of PIVA techniques using alpha2-adrenergic agonists [41,42], lidocaine [43], ketamine [44] and opioids [45,46]. Only 701 of the total 6701 were general anaesthetics where total intravenous anaesthesia (TIVA) was used, 31 under field conditions. Very short procedures were carried out in the anaesthetic recovery box or in the field using repeated boluses of a range of drugs and combinations. Ketamine, thiopental, alpha2-agonists and benzodiazepines were all used. Less than 300 of the 701 TIVA cases were maintained with various combinations of the so-called “triple drip”, usually combining ketamine with an alpha2-agonist and guaiphenesin [40]. The maintenance of anaesthesia for short procedures was also carried out with ketamine alone or occasionally with thiopental.Our data show that it is now common practice (63%) to administer a small dose of an alpha2-agonist after the end of general anaesthesia before the recovery phase, either when the patient is still on the surgical table and before transport to the recovery box, or once in the recovery box. Santos et al. (2003) [47] first demonstrated its benefits, later confirmed by others [48,49]. This study collected relevant information about the monitoring undertaken during general anaesthesia. Although it has been recently suggested that improvements in monitoring have reduced the risk of anaesthetic mortality [19], no data from a prospective, multicentre study support this statement. Our data indicated that around 90% of the horses undergoing general anaesthesia (INH, PIVA or TIVA) had an electrocardiogram and a pulse oximeter. Arterial blood pressure (invasively) and the end-tidal carbon dioxide (EtCO2) concentration were measured in more than 75%. Inspired oxygen and an end tidal volatile agent were measured in fewer cases, but still more than 50%. Arterial blood gas analyses (70%) and lactate (43%) were performed most often in horses with colic. In general, body temperature was not frequently monitored (33%), and the use of spirometry in horses was minimal. Non-invasive blood pressure was sometimes used in foals, when an arterial line had not been placed, or even with invasive blood pressure. These data are similar to those reported via the online questionnaire in 2015 by Wohlfender et al. [50].The data show the current protocols for the induction of anaesthesia were usually assisted, with either personnel (61%) or using a gate (25%). For recovery, 51% were free, 41% were assisted with ropes and 8% were manually assisted, mostly foals. These data reflect the ongoing controversy about which is the best method of recovery from general anaesthesia. Head and tail rope systems do not completely prevent fractures during recovery [51,52,53], but some cases may benefit from this technique. Slings were used in some cases but again did not ensure a good recovery: one young horse became excited with sling assistance and recovered badly after an elbow fracture repair; this was one of the unexpected noncolic deaths. (4)Preliminary results of standing sedations in horsesThe data presented here are the first reported from a prospective, multicentre, cohort study of standing sedation, but the numbers are small and cautious interpretation is still in order.Whereas the data collected from general anaesthetics can be considered as strong and reliable, only 43 of the 57 centres sent more than 90% of their standing sedation cases, which could lead to biased data. The main challenge for centres with a high case load was to collect every single case. Alpha2-agonists were used in all except three of the premedication combinations. Acepromazine was not given alone but, combined with at least an alpha2-agonist, it was given to 47% of the cases. Opioids, mostly butorphanol, morphine and methadone were used in 85% of the cases. Boluses of detomidine, butorphanol and romifidine are the most commonly used for the maintenance of standing sedations. For CRIs, detomidine was the most commonly used, followed by far, by butorphanol. Monitoring in standing sedation procedures was minimal. Many of the procedures previously performed only in anaesthetised horses are now carried out using standing sedation. Although the inclusion criteria were more restrictive and not all the centres were able to supply these procedures, 23% of the cases in this study were standing sedations. Further study of these data is required. Creating a subgroup of collaborating centres may ensure more reliable data.(5)LimitationsOur study is not free of limitations. First, the electronic questionnaire. Its use required a learning curve and sometimes raised technological difficulties, all solved with online support. The potential for creating a specific app is under consideration; however, the adaptation to different software can be challenging. Second, the current version of the questionnaire could be improved. For instance, the field “castrated” and “pregnant” is not mandatory in the current version and could lead to misinterpretation. Drugs commonly used, such as mepivacaine, were not included (but could be added manually). An updated version will be created in response to the feedback from the collaborating centres. Third, there is an inherent bias as some areas of the world and types of practice are underrepresented. Further strategies are to be implemented in this respect, although language barriers should not be underestimated. Finally, the amount of information collected by the questionnaire is enormous and cannot be covered in a single scientific paper. However, and as stated in the agreement form, “sub-studies can be proposed, as long as this does not involve duplicate use of the CEPEF4 data”. This may allow our community to benefit from our method for collecting multicentric data for purposes other than mortality up to seven days.5. ConclusionsWe have designed a reliable method, with a clear strategy of communication and data collection/cleaning that can be used to collect cases for CEPEF4. This approach will help to minimize inconsistencies that may lead to data misinterpretation in future CEPEF studies. This preliminary report shows that horses still die unexpectedly during and within the seven-day postoperative period of general anaesthesia and standing sedation. Our results also show that current practice in anaesthesia has changed over the last 20 years.

animals : an open access journal from mdpi

10.3390/ani11071956

PMC8300385

South American camelids (SAC) are increasingly popular in Switzerland for keeping as pets. For many farmers, SACs are a supporting leg when offered for trekking. Products made from alpaca fibers are in demand, and even the market for meat is increasing as they are extensively hold. Low reproductive rates and abortions/stillbirths are one of the main complaints among alpaca and llama breeders. The purpose of this study was to investigate the reasons for pregnancy loss and perinatal death in SAC herds in Switzerland. Aborted and stillborn fetuses were collected, necropsied and analyzed for infectious causes. This was complemented with a nationwide survey among breeders. During a 1.5-year period, eight cases of aborted or stillborn alpaca and llama have been analyzed. For the first time, the bacterium Coxiella burnetii, a known cause of abortions and stillbirths in small ruminants and cattle, could be detected in aborted alpacas and llamas. As this pathogen has the possibility to infect humans (=zoonotic potential), it is important to gain more knowledge about its significance among SAC. This study has set the ground stone for further investigations about abortifacients in Swiss SAC herds.

Over the last decade, South American camelids (SAC) have gained increasing popularity in Switzerland. They are used for several purposes such as fiber and meat production, as companion or guard animals and for trekking activities. The purpose of this study was to investigate the frequency and reasons for pregnancy loss and perinatal death in SAC herds. Within the scope of this study, early embryonic losses could not be identified, as pregnancy examinations by ultrasonography are not performed routinely. Aborted and stillborn fetuses were collected, necropsied and analyzed for infectious abortifacients. A nationwide survey among breeders was carried out. During a 1.5-year period, only eight cases of aborted or stillborn alpaca and llama (out of a population of 6550 animals) were reported by the breeders, and their causes were subsequently analyzed. In half of the cases, Coxiella burnetii was identified in the fetoplacental material. Abortions and stillbirths were reported to be rare in Swiss herds. As a conclusion, recording of embryonic losses through ultrasound training of veterinarians should be impaired and breeders motivated to have abortions and perinatal mortality examined. Special focus should be laid on C. burnetii due to its zoonotic risk.

1. IntroductionRaising healthy offspring plays a vital role in South American camelid husbandry. Thus, fetal and perinatal loss are amongst the most important factors contributing to lowered reproductive efficiency and one of the major complaints amongst camelid North American breeders [1,2]. Birth rates of 50% and 45.9%, respectively, in alpaca (Vicugna pacos) and llamas (Lama glama) have been described, pinpointing early pregnancy loss (50–57.8%) as one of the main factors contributing to the low fertility of SAC [1]. Most embryos (50%) seem to be lost in the first month [3]. Camelid reproductive physiology exhibits some differences compared to other ruminants. Ovulation in camelids only takes place upon induction [4]. Gestational length in SAC ranges from 342 to 350 days [5]. Normal cria weight at birth ranges from 6.7 to 9.1 kg in alpacas and 10.2 to 13.6 kg in llamas [6].Previous studies in New Zealand detected an abortion rate of 5–17% [7,8]. Other studies have shown that finding the causative agent for the abortion can be difficult, leading to a rather big proportion of cases where the reason stays unknown [9,10]. Abortion is defined as the expulsion of a fetus after organogenesis is completed and before fetal independent viability [11]. Stillbirth describes the death of the mature fetus before or under parturition [11]. Known reasons for abortion and stillbirth can be categorized into two main categories: infectious and non-infectious causes. Infectious abortifacients can be of viral, bacterial or parasitic nature. Bovine viral diarrhea virus (BVD) is one of the most common viral abortifacients found in SAC [10,12]. Bacteria are another important group of infectious organisms provoking abortions and stillbirths. So far, Brucella, Campylobacter fetus subsp. fetus spp., Chlamydophila spp., Leptospira spp., Listeria monocytogenes, Salmonella spp. and Trueperella pyogenes have been isolated from aborted SAC fetuses [1,10,13,14,15,16]. Additionally, parasites may cause abortions and stillbirths in SAC; namely, Neospora caninum and Toxoplasma gondii are known parasitic abortifacients [1,17].Early embryonic death is the most common form of reproductive loss in SAC, estimated to affect 10–15% of all pregnancies in the first 60 days of pregnancy [18]. In extreme conditions, up to 60–80% in the first 90 days of gestation might occur [18].Twin pregnancies, stress, umbilical torsion and placental insufficiency account for the non-infectious causes of abortion and perinatal mortality [4,19]. The definition of perinatal mortality varies among studies and is in some cases used as a synonym of stillbirth [20]. Usually, perinatal mortality refers to death at full term up to 48 h after parturition [21]. Perinatal mortality accounted for 22.4% alpaca and 8.6% llama deaths in the UK SAC population [22]. The most common causes of perinatal mortality were a lack of colostrum and hypothermia [6,22]. The age of the dam also has an impact on perinatal survival of the cria. Younger dams were found to have higher perinatal loss (11.3%) of their crias, than more mature dams (5.9%) [6]. Bravo et al., 2009, hypothesized that younger dams give birth to smaller and lighter crias [6]. The smaller size of the crias means increased difficulty in reaching the udder, which leads to higher risk of perinatal death due to starvation and lack of colostrum. South American camelids are popular in Switzerland [23]. The number of alpacas increased by 4.3% in the year 2020, whereas the llama population decreased by 1%, according to the Swiss Federal Office for Statistics [24]. In 2019, the SAC population counted 2882 llamas and 3668 alpacas in Switzerland [25]. The animals are used in a variety of purposes such as fiber and meat production, hobby, breeding or as guard animals [23,26]. At the level of the Swiss legislation, SAC have production animal status [27]. Camelid husbandry in Switzerland differs to that in South America, and, hitherto, the frequency and reasons for abortions and perinatal mortality in Switzerland might also differ.The objective of this study was to investigate the causes and to gain knowledge about the impact of abortion, stillbirths and perinatal mortality in Swiss SACs. In a first part of the study, carcasses of aborted, stillborn and perinatally diseased crias were collected over a 1.5-year period (2018–2020). Subsequently, necropsies were performed, and samples taken and analyzed for known abortifacients. Jugular vein blood was taken from the dams for determination of antibodies and trace element status. Currently, there are no data on trace element values in SAC in Switzerland. In the second part of the study, breeders were sent a questionnaire to gain information about Swiss SAC husbandry and how often breeders have to deal with pregnancy loss and perinatal mortality.2. Materials and MethodsThis investigative model involved: (1) recruitment of SAC herds with abortions, (2) collection of samples (fetus/cria, placenta, serum of the dam) from these herds and distribution of questionnaires to the owners, (3) necropsy and laboratory examinations of the samples collected and (4) distribution of a nationwide questionnaire about reproductive losses to SAC owners.2.1. Field Samples from South American Camelids and Farm DataAborted crias were collected throughout Switzerland during a 1.5-year (July 2018–February 2020) sampling period. Alpaca and llama breeders were informed about the project by means of a short article in a magazine for breeders distributed by the Swiss Counseling and Health Service for Small Ruminants. Furthermore, a short description of the project was presented in a nationwide continuing education program for food animal practitioners in Bern in June 2018. Additionally, veterinarians and breeders were informed about the project at a congress about SAC reproduction in April 2019. The fetoplacental unit was retrieved within a day of the notification by the owners and then kept in cold storage (4 °C) until immediate necropsy. Simultaneously, blood samples [S-Monovette® (no anticoagulant), 10 mL Luer tubes, Sarstedt, Nümbrecht, Germany] were taken from the jugular vein of the dam. The blood samples were centrifuged within 4 h of collection at 3500 revolutions per minute for 10 min (Hettich® EBA 20centrifuge, Hettich, Switzerland) and the sera stored in Eppendorf tubes. One sample was immediately sent to the Institute of Veterinary Bacteriology (Vetsuisse-Faculty, University of Bern, Switzerland) for serologic analysis of brucellosis, a second tube was sent to the Institute of Parasitology (Vetsuisse-Faculty, University of Bern, Switzerland) for serology of neosporosis and toxoplasmosis, a third sample was sent to the Institute of Virology and Immunology (Vetsuisse-Faculty, University of Bern, Switzerland) for Bovine Herpes Virus 1 (BHV-1) serology, and a forth tube was stored at −20 °C for trace element analysis. The owners were asked to fill in a questionnaire with 50 questions. The questionnaire consisted of a general part, asking questions about the farm structure and management, and a specific part, inquiring about the abortion, health status of dam and sire and recent management changes. Between July 2018 and February 2020, a total of eight aborted alpacas and lamas and their corresponding placenta and serum of the dam were retrieved. All animal experimentation was performed with permission and in accordance with Swiss law. The following approval number was allocated by the animal experimentation commission (elected by the cantonal executive counsel) BE 30/18. All participating breeders gave written consent for all examinations.2.2. NecropsyThe fetuses were necropsied at the Institute of Animal Pathology (Vetsuisse-Faculty, University of Bern, Switzerland) at the latest within 48 h of parturition. First the weight of fetus and placenta were determined, as well as the crown–rump length (CRL, distance between the external occipital protuberance and the end of the sacral bone using a measuring tape) of the fetus. Fetal gender was assessed, and the body was searched for malformations. The carcass was dissected according to standard operation protocols [28,29], and laryngeal muscles, lung, heart and liver tissue were collected and fixed in 10% formalin for histologic examination. These tissue samples were paraffin-embedded and processed to hematoxylin and eosin-stained tissue sections. For microbiological analysis, an ear notch sample, tissue samples of the myocardium, liver, lung, as well as a brain hemisphere, the ligated compartment three and body fluid of the fetus were collected. The placenta was controlled for completeness and macroscopic lesions. Tissue samples were retrieved for microbiologic examination, and a piece was fixed in 10% formalin for histologic analysis, taking care not to collect a part from the physiologically avillous area. The histopathological tissue slices were evaluated by board-certified pathologists for the presence of inflammatory and degenerative lesions.2.3. Laboratory Testing2.3.1. Parasitological AnalysisTotal DNA was extracted from fetal brain, myocardium and placenta samples at the Institute of Parasitology (Vetsuisse-Faculty, University of Bern, Switzerland) as described [30], and real-time PCRs to detect N. caninum [31] and T. gondii DNA [32] were performed. In addition, serum samples of the dam and fetal body fluids were analyzed by commercial enzyme-linked immunosorbent assays (ELISA) to detect antibodies against N. caninum and T. gondii [33].2.3.2. Virological AnalysisBovine viral diarrhea virus was detected by ELISA, performed on fetal ear notch samples sent to the IDEXX Diavet laboratory (Bäch, Switzerland). The commercial IDEXX BVDV Ag/Serum Plus kit (IDEXX Switzerland, Liebefeld-Bern, Switzerland) was used [34]. In order to detect antibodies against BHV-1, serum samples of the dam were analyzed with antibody ELISA IDEXX IBR gB Ab Test (IDEXX Switzerland, Liebefeld-Bern, Switzerland) according to international standards [35]. Placenta, fetal lung and liver tissues were analyzed at the Institute of Virology and Immunology (Vetsuisse-Faculty, University of Bern, Switzerland) by direct immunofluorescent detection of equine herpes virus 1 (EHV-1) antigens according to OIE standards [35].2.3.3. Bacteriological AnalysisSerum samples of the dam as well as placenta, fetal lung, liver and the ligated compartment three were all analyzed at the Institute of Veterinary Bacteriology (Vetsuisse-Faculty, University of Bern, Switzerland).Broad-Spectrum CultureBroad-spectrum bacterial and fungal culture was performed for all samples of placenta, fetal lung, liver and compartment three content.Staining and MicroscopyAll tissue samples were evaluated with the Stamp’s modification of the Ziehl–Neelsen staining [36].Molecular Detection of BacteriaIn order to detect Coxiella burnetii and Chlamydia abortus, total DNA was extracted as described [30]. Subsequently a qPCR was performed [37,38].SerologyBrucella abortus and Brucella melitensis antibodies were detected using the IDEXX Brucellosis Serum X2 test kit (IDEXX Switzerland, Liebefeld-Bern, Switzerland).2.3.4. Trace Elements AnalysisFrozen serum samples (stored at −20 °C) of the dam were used to determine their trace element levels at the time of abortion or stillbirth. The samples were collected, frozen and then sent in bulk to be analyzed at an external laboratory (Labor Zentral, Geuensee, Switzerland). Zinc and copper values were detected by means of photometry. Inductively coupled plasma mass spectrometry was used to identify serum selenium and iodine concentration. In order to detect manganese, an atom absorption spectroscopy was performed.2.4. QuestionnaireIn order to assess putative risk factors and gain a general overview of how often owners experienced abortions in their herd, a standardized web-based questionnaire was sent to all the SAC members of the Swiss Counseling and Health Service for Small Ruminants (n = 377) by email. The questionnaire contained 12 questions including general questions about the farm (i.e., SAC species, herd size, other animals on the farm), management (who takes care of the animals, purchase) and occurrence of abortions and stillbirths (Supplementary Materials Questionnaire S1 and Questionnaire S2). Answering of the questionnaire was voluntary, and all collected data were anonymous.2.5. Statistical AnalysisThe continuous variables (gestation length, body weight of fetuses, placental weight, CRL, maternal blood trace element concentration, age of the dam) were described using median and quartiles. The categorical variables (breed, purchase, other animals present, defined necropsy findings, presence of pathogen infection) were described using frequencies and proportions. Due to the limited number of cases, the statistical analysis was limited to descriptive statistic only. Data were analyzed using the statistical software NCSS 11 (2016) (NCSS, LLC., Kaysville, UT, USA).3. Results3.1. Field Samples of Abortion Cases3.1.1. Farm CharacteristicsThe eight cases of aborted/stillborn crias originated from seven Swiss farms. One farm submitted two abortions of the same dam in two consecutive gestations. The seven participating farms were all located in the German-speaking part of Switzerland in the cantons of Bern (four cases on three farms), Lucerne (two cases), Uri (one case) and Zurich (one case). All herds were taken care of by members of the family. The mean herd size was 25.5 animals with a minimum of five animals and a maximum of 65. All farms kept their animals in a combination of pasture and barn. More than half of the participating farms (5/8) kept other animals such as cattle (n = 3), cats (n = 3), dogs (n = 2) or chicken (n = 2) on their premises.3.1.2. Data about the Dam, Cria and the Abortion/StillbirthOf the eight cases included in this study, six of the dams were alpacas, while two were llamas. The mean age at the time of abortion or stillbirth was eight years, with the youngest dam being four years and the oldest 13 years old. The majority of the dams (5/8) were healthy at the time of abortion. One dam was diagnosed with Mycoplasma haemolamae and endoparasitosis prior to giving birth. A second one showed a swelling at the level of the mandible ante partum, whereas a third dam was ill of unknown cause at the time of parturition. Three of eight dams had a history of pregnancy loss in a previous gestation.The fetuses were expelled at a mean of 306 days of gestation. In two cases, the gestational age was unknown, because the owner had not noted the exact day of mating. The majority of fetuses were female (n = 5). Most of the offspring were aborted without any assistance. In two cases, the owner helped the dam expel the fetus. Three crias were stillborn. Only one parturition took place during the night, while the majority of dams (7/8) gave birth during daylight hours.3.1.3. Necropsy and Histopathological FindingsThe mean placental weight, fetal weight and CRL were 1.3 kg, 5.6 kg and 59.5 cm, respectively. More details concerning weight, CRL and placental weight are shown in Table 1. None of the aborted fetuses showed signs of malformation. The following observations were made during necropsies: fetal pulmonary atelectasis (8/8), bronchopneumonia (1/8), macroscopically visible (and/or histologically) visible meconium (3/8), placental mineralizations (4/8), placental edema (1/8), placental necrosis (1/8) and placentitis (3/8). In one case of placentitis, bacterial colonies were seen histologically on the surface of the chorion, as well as infiltrating into the chorion.3.1.4. Serological and Microbiological FindingsPathogen detection and serology results are shown in Table 2.Antigens/DNA of BVDV, EHV, Neospora caninum, Toxoplasma gondii, Brucella abortus, Brucella melitensis, Campylobacter foetus subsp. Foetus, Campylobacter foetus, subsp. Venerealis, Chlamydia abortus, Salmonella sp. and pathogenic Leptospira spp. have not been detected. Antibodies against BHV-1, Neospora caninum, Brucella abortus and Brucella melitensis have not been detected. In all dams, antibodies against Toxoplasma gondii were found.Direct evidence of infection (Ag-ELISA, culture, PCR in combination with histopathology results) was found in 5/8 animals. There was one co-infection found in a dam with C. burnetii and Strep. pluranimalium. The dam had a history of already having aborted in a previous gestation. Another dam aborted twice in a row, and both fetuses were submitted to this study. In both cases, C. burnetii could be detected by qPCR. Despite detailed case histories and extensive necropsy and laboratory examinations, the ultimate cause of death (UCOD) could not be assigned in 3/8 cases.Indirect evidence of exposure to an infectious agent (Ab-ELISA) was found in all dams. All of the tested females had developed antibodies against T. gondii. Although T. gondii DNA could not be detected in any of the fetuses or corresponding placentae by PCR, antibodies against T. gondii were detected in one aborted fetus (Table 1 and Table 2: No. 8) and one stillborn cria (Table 1 and Table 2: No. 7) indicating infection.3.1.5. Trace Mineral AnalysisIn two animals (dams 1 and 2), the serum selenium level was low, in four animals (dams 2, 3, 4 and 6) the zinc level was low, respectively. Trace mineral status of all the dams are shown in Table 3.3.2. QuestionnaireThe questionnaire with 12 questions about herd management and experiences with abortions and stillbirth was sent to 377 members of the Swiss Counseling and Health Service for Small Ruminants. The questionnaire was returned by 43 members (11.40%), fully completed by 32 members (8.48%), while 11 questionnaires were only partially completed (2.91%), four questionnaires being submitted empty (1.06%).The majority of the participating herds (93%) are taken care of by family members. About half of the farms (53.1%) stated to have 1–5 dams, whereas 25% have 6–10 dams, 9.4% 10–20 dams and 12.5% more than 20 dams in their herd. Keeping other animals beside SAC seems to be quite common, with 61% of the participants answering to keep another species. Companion animals being the most popular. Cats were mentioned as pets in 48.7% of households and dogs in 38.5% of homes. Cattle were kept alongside 23% of SAC herds, whereas horses were only present in 12.8% of participating farms. Other animals mentioned to live with alpacas and/or llamas were small ruminants (n = 8), poultry (n = 8), donkeys (n = 2) and rabbits (n = 2). Most of the purchased SAC originate from Swiss farms (71%). None of the farms purchased animals from North or South America. One in 39 herds each made acquisitions in Europe or Australia and New Zealand. In 11 of 26 herds, the new arrivals were isolated prior to introduction into the existing herd. The majority of breeders (19/27) stated that no abortion occurred in the previous two years. In six (22%) herds, one abortion had happened in the previous two years, and in two herds (7%), even two abortions occurred. Only 36% answered the question if they experienced a stillbirth in the previous two years. One stillbirth had occurred in nine (64%) herds, two stillbirths in two (14%) herds, whereas three (21%) breeders answered that they had never lost a cria due to stillbirth in the previous two years. In summary, this results in a total of 19 (49%) reproductive losses during the previous two years, either in the form of abortion or stillbirth, amongst the participating owners. Only 3.6% (1/28) breeders separate the dams for parturition. Parturitions were observed by the majority of breeders (21/27).No evidence of significant association was found with herd size, purchase, other animals present and abortion or stillbirth, respectively.4. DiscussionIn the 1.5-year field sample collection period, only eight cases of abortion or stillbirth could be collected and necropsied from a total population of 6550 animals. Considering the reported 10% of abortions in SAC in a recent survey in Switzerland, Austria and Germany of [40], one could expect 655 cases per breeding season. Other authors described difficulties in gaining access to cases due to remote farms and great distances to travel [41], which is less of a problem in Switzerland. The number of abortions reported in the seven Swiss farms was similar to the number of cases included in our study. With the data at hand, it is not clear whether abortions really are a minor concern in Swiss SAC herds or if the herds with more losses did not participate in our study. By legislation, SAC abortions do not have to be examined. This is probably why we have such low numbers. Despite strong advertising for this pilot study at veterinary conferences, in journals, newsletters and the information that we come to collect the abortions and that everything is free of charge, we were not able to acquire more abortions for pathological examination over the defined period. In Swiss SAC herds, ultrasonographic pregnancy diagnoses are not routinely undertaken. Therefore, no statement on early embryonic losses is possible. Statistics on abortions in Switzerland only concern those detected visually by owners. Further investigations are necessary to elucidate the prevalence and cause of abortions in Switzerland.All dams that reported abortions or stillbirths in the current study were seropositive for T. gondii. This is in accordance with a previous seroprevalence study in Switzerland, showing 83.2% of Swiss SAC to be seropositive for the parasite [33]. In the questionnaire of our study, cats were reported as the most popular pet to be kept alongside alpacas and llamas. This fact leads to a higher probability to have contact with T. gondii and becoming seropositive. In the study of Basso et al., 2020, “absence of cats in the last two years” was found to be a protective factor [33]. From none of the aborted or stillborn fetuses of the seropositive dams, T. gondii DNA could be detected by qPCR. However, one aborted fetus and one stillborn cria showed antibodies against T. gondii in their body fluid. As maternal antibodies do not cross the placental barrier in SAC, this indicates that those fetuses got infected with T. gondii during pregnancy. If the infection was also the cause of the pregnancy, loss remains unclear. We could assume that the dam became primo-infected during pregnancy and transmitted the parasite vertically to the fetus. Stillbirth caused by T. gondii has so far been described once, in which case T. gondii tissue cysts could be found in both kidneys, along with mild non-suppurative lesions in the placenta, liver and heart [42].In five of eight of the aborted or stillborn fetoplacental units, C. burnetii could be isolated from fetal organs, the placenta or both. Only in three cases, corresponding histological lesions were found as placentitis and another case revealed edema of the lamina propria of the placenta. Two of the corresponding dams had a history of a previous pregnancy loss. To our knowledge, it is the first report of an abortion caused by C. burnetii in a SAC. C. burnetii is a well-known cause of abortion in small ruminants [43,44]. In a Swiss prevalence study among small ruminants, C. burnetii could be isolated in 44.4% of aborted sheep and 44.2% of aborted goats [45]. The numbers are in line with the publication of [30] with submitted abortion material during the years 2012–2016 in sheep with 46.3%, respectively, in goats with 52.8%. In a Swiss study on PM in cattle, C. burnetii was detected in 32% of calves by qPCR, but only 15% of placentae and 19% of calves had histological signs of inflammation (placentitis and bronchopneumonia), respectively [20]. The authors suggested that acute infections with C. burnetii might occur without histological lesions [20]. Regarding the zoonotic potential of C. burnetii and the high numbers of bacteria shed after an abortion [45,46], it is essential to investigate abortion cases in SAC in order to implement protective measures for the owners and the remaining animals in the herd.One of the aborted fetuses showed histological signs of a severe bronchopneumonia. Microbiological analysis isolated large amounts of E. coli in the fetal liver, lung and compartment three, as well as in the placenta, additionally to C. burnetii detected in the placenta by PCR. These findings might point to an opportunistic infection or a co-infection leading to the stillbirth of the animal. Normally, as also in this case, the mother shows no clinical signs. Opportunistic infections with E. coli leading to stillbirth are also described in cattle [47] and humans [48].In two cases where no infectious agent could be detected in the fetus, the corresponding dams showed signs of systemic illness, one dam aborted after transport. In these cases, we regard the systemic illness or stress to be the reason for the pregnancy loss in these dams. Additionally, others have made the same observation that illness of the dam led to abortions with no microbiological or pathological findings [49,50].Regarding the existing knowledge about abortions and perinatal mortality in SAC, the available literature is scarce. The majority of available publications originates in North America, and the question remains what the situation in Europe is like. There are still large gaps of knowledge in terms of which infectious agents are of significance in SAC in Switzerland and Europe. BVD proofed to be the most frequently studied infectious agent. Additionally, results of the studies show that infections with BVD can lead to abortions and stillborn fetuses, but the highest risk lies in the development of persistent infected (PI) animals, which can spread the virus more effectively [12,51,52,53]. Especially in SAC, the risk of spreading BVD with PI crias is serious, due to the often seen habit of taking dams and their suckling cria to other farms to be bred [53,54]. When Switzerland started its BVD eradication program, a study about the impact of SAC in spreading the virus was performed [55]. A prevalence rate of 5.75% for BVD in the year 2008 was found. Which led to the conclusion that SAC play no major role in the spread of BVD in Switzerland, and thus, SAC were not included in the BVD eradication program of Swiss cattle [55]. Apart from the studies about BVD, there are only isolated case reports about other infectious abortion pathogens. Infectious causes described in case reports include: EAV [56], T. gondii [42], N. caninum [17,41], Sarcocystis [57], L. monocytogenes [14], C. fetus fetus [16], Strep. sanguinis [58] and Coccidioides posodasii [59]. Regarding the zoonotic risk of some of the abortifacients and the popularity of SAC and the close contact to humans, it is of great importance to keep track of occurring abortions in SAC herds. Furthermore, it is vital to sensitize SAC owners to notify their veterinarian about abortions and send the appropriate material to laboratories for further investigation of the etiology of the abortion and taking appropriate safety measures.As prepartum maternal dietary trace element imbalances (selenium, zinc, copper, iodine) are associated with PM in calves [60], trace minerals were analyzed in dams, especially as no data in SACs are available in Switzerland. In two dams, selenium values were below indicated reference value [14], but no corresponding histologic pathologies could be seen. The selenium findings may be explained by the low selenium status of Swiss soils [61]. Low zinc values were described in mummification and abortion in ewes [62], but the low values found here may be incidental.Concerning perinatal mortality, the review revealed a wide variety of definition, due to which not many publications met our inclusion criteria. There is a correlation between the age of the dam and the chance of survival of the cria [6,63]. This suggests that it is useful to monitor young dams around parturition to ensure that the offspring receive sufficient colostrum and, hence, have a higher chance of survival.5. ConclusionsWith the data at hand, it is not clear whether abortions really are a minor concern in Swiss SAC herds or if the herds with more losses did not participate in our study. Gaining sufficient data by voluntary participation proved to be a major challenge. Further information campaigns, incentives and prevalence studies are necessary to obtain a more precise picture of reproductive losses in Swiss herds. Recording of embryonic losses through ultrasound training of veterinarians should be impaired and breeders motivated to have abortions and perinatal mortality examined. For the first time C. burnetii could be identified as an abortifacient in SAC, and our results indicate that it should be considered as a differential in cases of abortion.

animals : an open access journal from mdpi

10.3390/ani12070803

PMC8996834

We present chromosome and DNA analysis of a normal Thoroughbred mare and her abnormal foal born with neurologic defects. We show that the foal has an abnormal karyotype with three copies of chromosome 26 (trisomy chr26), instead of the normal two. However, two of the three chr26 have fused, forming an unusual derivative chromosome. Chromosomes of the dam are normal, suggesting that the chromosome abnormality found in the foal happened during egg or sperm formation or after fertilization. Analysis of the foal and the dam with chr26 DNA markers indicates that the extra chr26 in the foal is likely of maternal origin and that the unusual derivative chromosome resulted from the fusion of two parental chr26. We demonstrate that although conventional karyotype analysis can accurately identify chromosome abnormalities, determining the mechanism and parental origin of these abnormalities requires DNA analysis. Most curiously, this is the second case of trisomy chr26 with unusual derivative chromosome in the horse, whereas all other equine trisomies have three separate copies of the chromosome involved. Because horse chr26 shares genetic similarity with human chr21, which trisomy causes Down syndrome, common features between trisomies of horse chr26 and human chr21 are discussed.

We present cytogenetic and genotyping analysis of a Thoroughbred foal with congenital neurologic disorders and its phenotypically normal dam. We show that the foal has non-mosaic trisomy for chromosome 26 (ECA26) but normal 2n = 64 diploid number because two copies of ECA26 form a metacentric derivative chromosome der(26q;26q). The dam has normal 64,XX karyotype indicating that der(26q;26q) in the foal originates from errors in parental meiosis or post-fertilization events. Genotyping ECA26 microsatellites in the foal and its dam suggests that trisomy ECA26 is likely of maternal origin and that der(26q;26q) resulted from Robertsonian fusion. We demonstrate that conventional and molecular cytogenetic approaches can accurately identify aneuploidy with a derivative chromosome but determining the mechanism and parental origin of the rearrangement requires genotyping with chromosome-specific polymorphic markers. Most curiously, this is the second case of trisomy ECA26 with der(26q;26q) in the horse, whereas all other equine autosomal trisomies are ‘traditional’ with three separate chromosomes. We discuss possible ECA26 instability as a contributing factor for the aberration and likely ECA26-specific genetic effects on the clinical phenotype. Finally, because ECA26 shares evolutionary homology with human chromosome 21, which trisomy causes Down syndrome, cytogenetic, molecular, and phenotypic similarities between trisomies ECA26 and HSA21 are discussed.

1. IntroductionMultiple forms of chromosome rearrangements have been reported in the domestic horse, Equus caballus (ECA) and most are associated with decreased fertility, embryonic or fetal loss, congenital and developmental disorders, causing significant economic loss to breeders and the equine industry [1,2]. The most commonly found chromosomal abnormalities in horses are X-monosomy and XY male-to-female sex reversal (also known as XY disorder of sex development or XY DSD) [1,2,3], which owe to the specific features of equine sex chromosome organization [2,4,5]. Rearrangements involving autosomes, however, are rare in horses and include mainly a few translocations and autosomal aneuploidies [2].Aneuploidies cause genetic imbalance, due to which most of them are lethal [6], and the 14 reported live-born cases of autosomal trisomies involve only the six smallest equine autosomes—ECA23, 26, 27, 28, 30 and 31 [1,2,7]. Autosomal aneuploidies are equally rare in other domestic species. There are 16 reported cases of autosomal trisomies in cattle involving the 10 smallest autosomes, typically resulting in fetal death or postnatal culling by breeders due to congenital defects [8,9]. In the domestic pig, there are no reports of live-born animals with whole autosome aneuploidies [10], and all autosomal trisomies in dogs have exclusively been found in tumor cells [11]. Likewise, although aneuploidies occur in at least 5% of clinically recognized human (Homo sapiens, HSA) pregnancies and account for over 25% of spontaneous abortions, only trisomies of HSA13, 18 and 21 have been found in live born, of which only trisomy HSA21 survives to adulthood [12,13].Extensive studies of human autosomal aneuploidies show that the majority are caused by errors in maternal meiosis I (MI) with advanced maternal age being a critical contributing factor, whereas only 5–10% of trisomies are caused by paternal errors [13]. At the same time, human data also show remarkable variation among trisomies regarding the parent and meiotic stage (MI or MII) of origin of the extra chromosome. For example, paternal errors account for nearly 50% of trisomy HSA2 but almost never for trisomy HSA16. Likewise, errors in maternal MI account for almost all cases of trisomy HSA16, whereas trisomy HSA18 is predominantly caused by errors in maternal MII, suggesting that the patterns of non-disjunction may have chromosome specific effects [13,14]. In rare occasions, trisomies of acrocentric autosomes are combined with Robertsonian fusion or isochromosome formation [15,16,17], so that despite of aneuploidy, the diploid chromosome number remains normal. For example, about 5–6% of cases with Down syndrome carry unbalanced heterologous or homologous fusions involving HSA21 [15,17]. The mechanism for heterologous fusions is Robertsonian translocation, of which the most common (82%) in Down syndrome patients is rob (14q;21q), with the remaining 8% represented by rob(13q;21q), rob(15q;21q) and rob(21q;22q) [17]. On the other hand, trisomy due to homologous fusion of (21q;21q) can result from different mechanisms—by isochromosome i(21q) formation or due to Robertsonian translocation rob(21q;21q). Since isochromosomes result from the duplication of a single chromosome arm [18], the duplicated parts are genetically identical and can be distinguished from homologous translocation by genotyping for allelic variation using chromosome specific polymorphic short tandem repeat (STR) markers [15,16,18,19]. In domestic animals, the only case of autosomal trisomy combined with centric fusion or isochromosome formation has been reported in horses for trisomy ECA26 [20,21]. The karyotype formula of the affected Thoroughbred mare was presented as 64,XX, −26,+t(26q;26q), but because polymorphic STR markers were not available for horses at that time, the researchers could not determine whether the abnormal chromosome (26q;26q) was an isochromosome or the result of a Robertsonian fusion. In the present study, we report and characterize the second equine case of trisomy 26 involving homologous fusion 26q;26q. We will characterize the case using classical and molecular cytogenetic approaches and genotype the affected individual and its dam with ECA26 STR markers to determine the mechanism and likely parental origin of the aberration. 2. Material and Methods2.1. Ethics StatementProcurement of blood samples followed the United States Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training. These protocols were approved as AUP and CRRC #2018-0342 CA at Texas A&M University.2.2. Case Description and SamplingA Thoroughbred foal (ID: H1063) was euthanized at the age of 5 months and 3 weeks due to stupors that gradually developed into ataxia, due to failure to thrive despite nursing well and being initially treated for possible neonatal mal-adjustment syndrome, and due to being inappropriate mentally. Although cervical radiographs did not provide an explanation for progressing ataxia, necropsy revealed axonal degeneration in brainstem and spinal cord suggestive of equine degenerative myeloencephalopathy. This was the first foal of a 5-year-old maiden Thoroughbred mare boarded on a large, well-managed farm. The sire had had several normal foals before. Peripheral blood samples in EDTA- and sodium heparin-containing vacutainers (VACUTAINERTM, Becton Dickinson) were obtained from the affected foal and its dam (ID: H1066) for cytogenetic and DNA analysis.2.3. Cell Cultures and Chromosome PreparationsMetaphase chromosome spreads were prepared from peripheral blood lymphocytes following standard protocols [22]. Briefly, 1 mL of sodium heparin stabilized peripheral blood was grown for 72 h in 9 mL of culture medium RPMI-1640 supplemented with HEPES and Glutamax (Gibco), 30% fetal bovine serum (FBS; R&D Systems Inc., Minneapolis, MN, USA), 1X antibiotic-antimycotic (100×; Invitrogen, Waltham, MA, USA), and 15 µg/mL pokeweed mitogen (Sigma Aldrich, St. Louis, MO, USA). Lymphocyte cultures were harvested with demecolcine solution (10 µg/mL; Sigma Aldrich), treated with Optimal Hypotonic Solution (Rainbow Scientific, Windsor, CT, USA), and fixed in 3:1 methanol/acetic acid. The cells were dropped on clean, wet glass slides and checked under phase contrast microscope (×200) for quality. 2.4. Karyotyping and Cytogenetic AnalysisChromosomes were stained by GTG-banding [23] for karyotyping. Karyotyping and chromosome analysis were performed with a motorized fluorescence microscope Axio Imager M2p (Zeiss) equipped with a high-resolution progressive scan CCD camera CoolCube 1 and Ikaros v5.3.18 software (MetaSystems GmbH, Altlußheim, Germany). Images of a minimum of 30 cells were captured and analyzed per individual. Horse chromosomes were identified and arranged into karyotypes according to the International System of Cytogenetic Nomenclature of the Domestic Horse [24] and chromosome aberrations were described following Human Cytogenomic, Nomenclature [25].2.5. Fluorescence In Situ Hybridization (FISH)The rearrangements identified by conventional cytogenetic analysis were validated by two-color FISH with ECA26-specific Bacterial Artificial Chromosome (BAC) clones (Table 1) from horse genomic BAC library CHORI-241 (https://bacpacresources.org/, last accessed 1 December 2021). The probes were labeled with biotin or digoxigenin by nick translation using Biotin or DIG Nick Translation Mix (Roche Diagnostics, Basel, Switzerland), following the manufacturer’s protocol. Hybridization and signal detection followed standard protocols described elsewhere [22]. Biotin-labeled probes were detected with Alexa Fluor® 488 streptavidin conjugate (Molecular Probes, Life Technologies, Carlsbad, CA, USA) and digoxigenin-labeled probes with DyLight® 594 anti-digoxigenin conjugate (Vector Laboratories, Burlingame, CA, USA). Chromosomes were counterstained with 4′,6-diamidino-2-phenylindole (DAPI). At least 10 cells were captured and analyzed for each experiment using Isis v5.3.18 software (MetaSystems GmbH, Altlußheim, Germany).2.6. DNA Isolation, PCR Analysis and STR GenotypingGenomic DNA was isolated from EDTA-stabilized blood with QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany). Both horses were tested by PCR for the Y-linked SRY gene and X-linked androgen receptor (AR) gene as described earlier [27], followed by genotyping for the 15 autosomal STRs of the standard equine parentage panel [28], and an additional 24 STRs specific for ECA26 (Table 2). Genotyping was performed either with directly fluorescently labeled primers [29] or with three-primer nested PCR where the forward primer in each primer-pair had an M13-tail which was targeted by a fluorescently labeled universal M13 primer during PCR reactions [30]. Annealing temperature for all PCR reactions was 58 °C. The PCR products were resolved with an ABI PRISM 377 (Applied Biosystems, Foster City, CA, USA) and allele sizes were determined using GeneScan-500 LIZ Size Standard and GeneMapper® v4.1 (Applied Biosystems, Waltham, MA, USA). 3. Results3.1. Chromosome AnalysisCytogenetic analysis showed that the affected foal (H1063) had normal 2n = 64 diploid number, XY sex chromosomes, one copy of normal ECA26, and the karyotype contained a morphologically abnormal metacentric derivative chromosome (Figure 1A,B). Analysis of GTG-banding suggested that the derivative chromosome was composed of two copies of ECA26 likely fused at the centromeres. Molecular cytogenetic analysis by FISH with two ECA26 BAC clones, one corresponding to the proximal (BAC 9N4) and the other, to the distal (BAC 91H11) portion of the chromosome, confirmed that the derivative chromosome was the result of homologous centric fusion 26q;26q (Figure 1C). Thus, despite the normal diploid number, the foal carried trisomy ECA26 in all cells analyzed. However, by cytogenetic analysis alone, it was not possible to determine whether the derivative chromosome resulted from Robertsonian fusion rob(26q;26q) or from isochromosome formation i(26q).Karyotype analysis of the dam (H1066) of the abnormal foal showed normal 64,XX female karyotype (Figure 1D,E) indicating that chromosomal abnormality of the foal must have originated from a parental meiotic error or a post-fertilization zygotic event. As a standard part of cytogenetic analysis, both horses were tested by PCR for the SRY and AR genes and the results agreed with karyotype analysis and the phenotypic sex of the two horses: the XY foal H1063 was SRY-positive, the XX dam H1066 was SRY-negative, and both horses were positive for the X-linked control marker AR.3.2. STR Genotyping: Parentage and the Origin of ECA26 TrisomyGenotyping for 15 genome-wide autosomal STRs [28] qualified the cytogenetically normal Thoroughbred mare H1066 as the dam of the affected foal H1063. The two horses were also genotyped for 24 STR markers which were evenly distributed over ECA26, starting with UMNe588 as the most proximal marker and ending with TKY523 as the most distal one (Table 3). As expected, the STR markers showed the presence of one or two alleles in the cytogenetically normal dam H1066. However, five STRs had three alleles in the abnormal foal H1063 (Figure 2, Table 3), indicating that the metacentric derivative chromosome was the result of Robertsonian fusion rob(26q;26q) and not an isochromosome. The karyotype of the foal was designated as 64,XY,der(26),rob(26q;26q) [25].Further comparison of the genotyping patterns between the foal and the dam showed that in all 5 cases where the foal had 3 alleles, two of the alleles were identical with those of the dam (Figure 2). Additionally, of the 10 markers that were heterozygous both in the foal and the dam, the two horses shared the same alleles (Table 3). Based on these observations, and despite having no genotype information for the sire, it is very likely that the extra ECA26 in the foal was of maternal origin.4. DiscussionHere, we characterized by chromosome analysis and STR genotyping an equine case of trisomy for chromosome 26 with homologous fusion 26q;26q (Figure 1 and Figure 2, Table 2). Genotyping ECA26 STRs in the affected horse and its dam showed that the abnormal chromosome was the result of Robertsonian translocation and most likely of maternal origin. Since the dam of the affected foal had normal 64,XX karyotype (Figure 1D), the aneuploidy must have originated from maternal meiotic nondisjunction, though the following fusion could have taken place either in meiosis or post-fertilization.It is certainly curious that this is the second case of trisomy ECA26 with a derivative chromosome 26q;26q in horses. The first case was described more than three decades ago [20,21], but because of uninformative blood typing, the mechanism (Robertsonian fusion or isochromosome) or parental origin of the aneuploidy remained unknown [21]. In our case, the presence of three alleles for 5 ECA26 STRs in the affected foal (Figure 2, Table 2) was a compelling piece of evidence that the derivative metacentric chromosome resulted from Robertsonian fusion. Furthermore, since all heterozygous STRs of the dam had the same two alleles also present in the affected foal (Table 2), we concluded that the extra chromosome ECA26 was likely of maternal origin. Though, complete evidence for the parental origin requires STR genotyping of the sire, whose samples were not available. Nevertheless, the findings underscore the importance of combining STR genotyping with cytogenetic analysis of possible isochromosomes or Robertsonian fusions. Isochromosome is formed by centromere mis-division of sister chromatids resulting in a bi-armed chromosome with identical genetic material in each arm [18,32]. Homologous Robertsonian fusions, on the other hand, result in genetically distinct arms preserving the heterozygosity from the parent from which the extra chromosome came from [16,18,33].Another intriguing aspect of the present and the previous case [21] was that there have been no reports about ECA26 trisomy with three separate copies of the chromosome. This contrasts with other recurrently reported equine trisomies: all cases of cytogenetically studied trisomies of ECA27 (4 cases), ECA30 (5 cases), and ECA31 (2 cases) (reviewed by [2]) involve three separate chromosomes without homologous fusions. Furthermore, the trisomy ECA26 described in this study, is so far the only confirmed Robertsonian fusion in equine clinical cytogenetics [2], even though Robertsonian type rearrangements have been a normal part of equid and Perissodactyl karyotype evolution [34].Can it be that ECA26 is more prone for centric fusion than other equine small acrocentric chromosomes? Chromosome-specific effects have been observed in humans where a small percentage of cases of Down and Patau syndrome with trisomy HSA21 and HSA13, respectively, have the extra chromosome in the form of Robertsonian fusion or an isochromosome [15,16,17,35,36]. In Down syndrome, there are even rare mosaic cases where one cell line carries HSA21 isochromosome and another, a Robertsonian fusion [17]. It is thought that some human chromosomes, such as HSA21, are inherently unstable and more prone to rearrangements [17] due to certain features of their sequence architecture (e.g., region-specific low copy number repeats) [18]. Based on our current knowledge of the horse genome [37], ECA26 does not stand out with any sequence peculiarities. Additionally, unlike HSA21 and other human acrocentric autosomes, ECA26 does not carry the satellite with multicopy rRNA genes that may contribute to instability [18]. On the other hand, and based on comparative chromosome painting [38] and gene mapping [26], ECA26 is more similar to HSA21 than to any other human chromosome because about 30 Mb (70%) of ECA26 shares evolutionary homology with the entire HSA21. However, the remaining 13 Mb (30%) of ECA26 is homologous to a part of HSA3 and this happens to be the pericentromeric/proximal portion of ECA26 which is involved in homologous fusion 26q;26q. Therefore, it is perhaps not relevant to expand the known instability of HSA21 [17] to ECA26 and it remains unclear whether the two cases of ECA26 trisomy with 26q;26q fusion were merely a coincidence or true reflections of presently unknown sequence properties of this horse autosome.On the other hand, it is also possible that ECA26 instability and rearrangements are due to sequence variants segregating in certain horse breeds or families and not due to the genomic architecture of ECA26 per se. Indeed, the case described in this study and the one reported earlier [20,21], both occurred in Thoroughbreds. However, then again, two cases are too few for any conclusions.Besides cytogenetics, there are several other shared features of interest between the two cases of trisomy ECA26 (this study; [20,21]). In both, the dams of the affected foals were young—5 years-old in this case and 3 years-old in the one described by Bowling et al. [21], thus excluding advanced maternal age as a contributing factor and rather supporting chromosome-specific effects. Additionally, both affected horses had gait deficits (ataxia), were not thriving, and had behavioral and mental issues. However, because the case presented in this study resulted in euthanasia at a young age but the horse described by Bowling et al. [21] lived many years, the basis of comparison is rather limited. It is, though, noteworthy that necropsy of the present case showed axonal degeneration in brainstem and spinal cord as seen in equine degenerative myeloencephalopathy (EDM) [39]. Although genetic basis for EDM is suspected but currently unknown [39], the present findings suggest that possible contribution of chromosome abnormalities/genetic imbalance should be considered. The fact that both cases were described as “inappropriate mentally” (this study) or “mentally dull” [21], and because of the homology between ECA26 and HSA21, there is a temptation to compare equine trisomy 26 with human Down syndrome. Indeed, there are some similarities: the horse described by Bowling et al. [21] lived many years and it is well-known that trisomy HSA21 is the only human autosomal trisomy surviving to adulthood [12,13]. Furthermore, at the age of 4, the mare with trisomy ECA26 gave birth to a chromosomally normal colt [21], and there are many cases of fertile women with Down syndrome in humans [40]. Despite this, drawing parallels between the two cases of ECA26 trisomy in horses with human Down syndrome should be taken with great caution. Firstly, genetic homology between ECA26 and HSA21 is not one-to-one since ECA26 is homologous also to part of HSA3 [26,38]. Secondly, stupors and ataxia which were the prevailing features of the two equine cases, are not the predominant characteristics of Down syndrome [41]. Most importantly, however, it is extremely narrow to compare the few phenotypic characteristics of two equine cases with the extensive research and clinical material available for Down syndrome since 1866 [41]. Furthermore, phenotypic features of the two equine cases share similarities with the phenotypes of other reported equine autosomal aneuploidies. For example, gait deficiencies, behavioral abnormalities and poor thriving have also been found in cases of trisomy ECA27 and ECA30 (reviewed by [2]), thus not being unique to trisomy ECA26. All in all, it is hard to tell which phenotypic features of trisomy ECA26 are the specific consequences of ECA26 overdose and which ones are due to general genomic imbalance.5. ConclusionsWe demonstrated that proper characterization of an autosomal (ECA26) trisomy with homologous fusion (26q;26q) and determining the mechanism and parental origin of the rearrangement, require the use of complementary approaches—cytogenetics and genotyping. To date, equine trisomy with homologous fusion has been unique to ECA26. However, to determine whether this is an ECA26-specific effect or just a coincidence, requires more cytogenetic cases and improved knowledge about the genomic architecture and functional annotation of ECA26. The latter is also needed to shed more light on the possible homology between trisomy ECA26 in the horse and the Down syndrome with trisomy HSA21 in humans.

animals : an open access journal from mdpi

10.3390/ani11041150

PMC8072609

The aim of this study was to analyze the whole blood transcriptome of lactating goats fed a dietary supplementation with 10% olive leaves, one of the main by-products deriving from the olive oil chain supply. This evaluation was effective in identifying the differential regulation of the gene coding for apolipoprotein B mRNA editing enzyme catalytic subunit 2 (APOBEC2), which showed downregulated in goats that received the dietary supplementation. Taking into account the strong association between plasma apoB and low-density lipoprotein, an evaluation was performed of both blood and milk cholesterol. The obtained data demonstrated a significant lower concentration of circulating cholesterol and cholesterol released into the milk through the mammary gland, demonstrating positive effects of olive leaves feeding on animal welfare and potential health benefits for consumers.

Agro-industrial by-products represent an important source of compounds credited with high biotechnological potential. In the last decade, considerable interest has developed toward the use of these matrices as dietary supplements in the zootechnical field, paying particular attention to the qualitative aspects associated with animal products. However, less is known about the effect of these matrices on gene expression and thus on animal metabolism. Therefore, the aim of this study was to analyze the whole blood transcriptome of lactating goats fed a dietary supplementation with 10% olive leaves (OL), one of the main by-products deriving from the olive oil chain supply. By applying a false discovery rate (FDR) < 0.05 and a Log2 Fold change (Log2Fc) lower than −0.5 or higher than +0.5, it was possible to identify the differential regulation of gene coding for the apolipoprotein B (apoB) mRNA editing enzyme catalytic subunit 2 (APOBEC2), which showed downregulation in goats that received the dietary supplementation. An evaluation of both blood and milk cholesterol was performed, taking into account the strong association between plasma apoB and low-density lipoprotein (LDL). Results showed significantly lower concentrations of circulating cholesterol and cholesterol released into the milk through the mammary gland, demonstrating positive effects of OL feeding on animal welfare and potential health benefits for consumers.

1. IntroductionThe olive oil chain supply is responsible for the production of waste residues whose disposal represents a pivotal issue both from an environmental and economic point of view. This problem is obviously strongly felt in Mediterranean countries, where we observe a marked production of olive oil and non-negligible amounts of related by-products are therefore accumulated, specifically represented by olive pomace and wastewaters, but also leaves and other plant pruning residues [1,2]. As for other agro-industrial sectors, the strategy of valorizing these by-products is now well established, trying to identify alternative uses justified by the high richness of these plant matrices in bioactive compounds [3,4]. From this point of view, much has been done in the zootechnical field, through the development of feeding strategies based on the use of these by-products, both for ruminants and monogastrics [5,6,7]. Specifically for olive oil by-products, over time, considerable information has been collected on the effects induced on both the qualitative and quantitative aspects of animal products. Most studies in both meat and dairy products have specifically shown an increase in concentration of polyunsaturated fatty acids and greater oxidative stability, with significant implications in the improvement of the health functionality and products’ shelf-life [8,9,10].With specific regard to olive leaves (OL), reference is made to a matrix well characterized for the high content of phenolic compounds, mainly oleuropeosides, flavonoids, and phenolic acids, which confer a high anti-inflammatory and antioxidant potential [11]. For that reason, in the last decade, OL have found widespread application in animal nutrition, especially for dairy ruminants. In the review presented by Molina-Alcaide and Yáñez-Ruiz [12], this by-product was presented as fibrous with a low digestibility, especially of crude protein, as well as capable of negatively influencing microbial protein synthesis and fermentation in rumen. A significant improvement in the milk fatty acid profile was however evidenced in lactating animals compared with animals fed conventional diets. More recently, the efficacy of an OL-supplemented diet in inducing an increase in concentration of linolenic acid in goat milk was confirmed; furthermore, an improvement in the oxidative stability of derived dairy products was observed, probably as a consequence of the reduction of lipolytic events during storage and ripening [13,14,15].All these findings represent a stimulus to the use of OL as an ingredient in the diet of lactating ruminants in order to obtain milk and cheese improvements from a nutritional point of view. However, little or nothing is known about the effect of this feeding strategy on the animal transcriptome. The use of experimental diets based on the integration of plant matrices rich in bioactive compounds, as in the case of OL, should positively contribute to the regulation of gene expression in dairy ruminants. This approach has allowed other studies to highlight the effect of diet on the molecular mechanisms responsible for variations observed not only on productive traits, but also in aspects related to animal health [16,17,18]. Therefore, the specific objective of this study was to analyze the whole blood transcriptome of goats by using an RNA-sequencing approach, considering the hypothesis that a dietary supplementation with OL could be effective in modifying the host’s metabolic pathways. It should not be underestimated that this can also represent a useful point of view for optimizing nutrition protocols for farm animals.2. Materials and MethodsThe study was conducted in a commercial company that, during the spring months, habitually integrates the diet of dairy goats with OL and other residues coming from pruning. For this reason, no breeding practices other than those normally adopted have been introduced. Additionally, the trial has been planned and executed by taking into account the Directive 2010/63/EU of the European Parliament (European Union, 2010) and Directive 86/609/EEC (European Economic Community, 1986), which deals with the protection of animals used for scientific purposes [19,20]. Blood sampling was performed by authorized veterinarians concurrently with planned blood withdrawal for the brucellosis prophylaxis; therefore, no ethical declaration is necessary.2.1. Experimental Plan, Animals and DietsThis study is part of a wider experimentation concerning the effect of an OL-supplemented diet on lactating goats, with the purpose of focusing the attention on both the qualitative characteristics of milk and cheese and the biochemical and molecular aspects influenced by the diet. For this reason, the experimental design has already been described [15]. Briefly, the trial lasted 30 days and involved 30 Saanen goats homogeneous for age (46 ± 2 months), weight (52.7 ± 4.3 kg), lactation days (86 ± 7 days in milk), milk yield (2296 ± 281 g/day), and a body condition score (BCS) equal to 2.78 ± 0.18.At the beginning of the experimentation, the animals were evenly separated into two groups that were indoor housed in two adjacent but separate areas characterized by a space shared by goats belonging to the same group, a drinking trough, and provisional single bunks on straw, useful for the individual administration of the diets. Each goat received polyphite daily hay ad libitum, while twice daily (in the morning and in the evening) was administered a custom-formulated concentrate (a total of 1 kg/head) whose ingredients and chemical composition have been previously reported [15]. The control group (CTR) was fed a standard diet that was formulated by taking into account the nutritional needs of lactating goats, while the experimental group (OL+) received the same basic diet but supplemented with 10% OL on a dry matter (DM) basis.2.2. Milk and Blood Samples CollectionAt the end of the experimental period, individual milk samples were collected from all the animals involved. Milk samples were then immediately aliquoted and stored at −20 °C until the time of analysis.Additionally, in the case of whole blood (WB), sampling was performed at the end of the 30 days of dietary supplementation on a total of 20 animals randomly selected, 10 from the CTR and 10 from OL+. In order to evaluate the blood biochemical parameters, WB was collected in Venoject glass tubes containing ethylenediaminetetraacetic acid (EDTA) or sodium heparin (Terumo Italia, Rome, Italy). Samples were immediately cooled, and plasma separation took place within 30 min of collection. Following a centrifugation step at 500 RCF for 15 min, the obtained samples were stored at −20 °C until analysis.Although a larger number of animals was used in the study, for the RNA-Seq analysis, sampling was done by drawing 2.5 mL of blood from the jugular vein on a total of 10 goats randomly selected among those already subjected to WB sampling for biochemical evaluations (5 from the CTR and 5 from OL+). Duplicate WB samples from each of the selected animals were collected in PAXgene™ tubes (Qiagen SpA, Milan, Italy), stored overnight at room temperature and then at −20 °C until RNA extraction, as prescribed by the manufacturer.2.3. Blood AnalysisBlood samples obtained from CTR (n = 10) and OL+ (n = 10) animals were analyzed for blood cell count with leukocyte formula (total white blood cells, lymphocyte, monocyte, neutrophils, eosinophils, and basophils). Analysis was performed by using a laser-based hematology analyzer with software applications for animal species (ADVIA 120 hematology system, Siemens, Munich, Germany). Serum samples were instead evaluated by using an automatic biochemistry analyzer (ILAB 650, Instrumentation Laboratory-Werfen, Milan, Italy), in order to quantify different target compounds (triglycerides, glucose, cholesterol, calcium, urea, aspartate aminotransferase (ASAT), alanine aminotransaminase (ALAT)).2.4. Cholesterol Evaluation in MilkCholesterol quantification was performed in raw milk samples obtained from the same animals involved in blood sampling (n = 10 from CTR and n = 10 from OL+). The analysis was carried out by applying the method previously described by Oh et al. [21] with slight modifications. In total, 1 mL of milk was transferred into a 15 mL tube and subjected to saponification by adding 1 mL of 10% KOH in ethanol (w/v) for 30 min at 70 °C. The addition of 5 mL of diethyl ether and 2 mL of H2O therefore allowed us to proceed with the extraction of the unsaponifiable fraction. The extraction was repeated 3 times, and the diethyl ether extract was transferred into a 50 mL round-bottomed flask and dried with a Strike-Rotating Evaporator (Steroglass s.r.l., Perugia, Italy) with the bath temperature set at 50 °C. The sample was then recovered with 1 mL of methanol, and a 20 µL aliquot was directly injected into the HPLC. Cholesterol identification was achieved through a HPLC chromatographic system (Varian, Harbor City, CA, USA) equipped with a Supelcosil LC-18 HPLC column (25 cm × 4.6 mm, 5 μm; Sigma-Aldrich, Milan, Italy). Isocratic conditions with a mobile phase containing 75% acetonitrile and 25% methanol were used. The flow rate of the mobile phase was 1.5 mL/min, and the column temperature was set at 38 ± 0.1 °C. The cholesterol peak was detected at 205 nm. Cholesterol (Sigma-Aldrich, Milan, Italy) was used as standard in order to obtain a calibration curve that was linear in the range of concentration from 0.01 to 0.50 mg/mL (R2 = 0.983). The obtained regression lines were exploited to calculate the amount of cholesterol that was expressed in mg/mL.2.5. Library Preparation and RNA-Seq AnalysisThe RNA-Seq experiments, inclusive of the RNA extraction and bioinformatics evaluations, were performed by an external company (Genomix4life SRL, Baronissi, Salerno, Italy). Total RNA extraction, library preparation, and sequencing were performed by following the procedure already described by Iannaccone et al. [18].The quality control of the obtained raw sequences (fastq files) was performed by exploiting the FastQC tool (Version 0.11.8) for high throughput sequence data, available on http://www.bioinformatics.babraham.ac.uk/projects/fastqc (accessed on 12 October 2020) [22]. The bioinformatic tool cutadapt (version 2.5) [23] was then used in order to remove the adapter sequences and the very short reads (reads length < 20). The sample was mapped on reference Capra hircus genome (ARS1) using the bioinformatics tool STAR (version 2.7.5c) [24], by referring to the standard parameters for paired reads. The reference track was the assembly Capra hircus (ARS1, INSDC Assembly GCA_001704415.1) obtained from Ensembl (release 101—August 2020—EMBL-EBI) [25].The featureCount algorithm (version 2.0) [24] was then used to quantify the expressed transcripts for each replicate of sequenced samples. A false discovery rate (FDR) less or equal to 0.05 (FDR ≤ 0.05) was considered in order to identify differentially expressed genes (DEGs). The Bioconductor package DESeq2 [26] was used to normalize the data, using the median of ratio, to perform the differential expression analysis.2.6. Statistical AnalysisStatistical analysis of the obtained data was performed by using the software SigmaPlot 12.0 (Systat software Inc., San Jose, CA, USA). The ANOVA model was used, and the Student’s t-test was applied for means comparison. p-values lower than 0.05 have been associated with the presence of significant differences.3. Results3.1. Hematochemical CharacterizationThe complete blood cell count and main biochemical parameters in serum samples were evaluated in order to characterize the general health status of animals fed the dietary OL supplementation. No significant variations (p > 0.05) were evidenced in the cell count of the different considered families (Figure 1).The only significant difference was instead observed in serum (Figure 2), in which a lower concentration of cholesterol was recorded in the OL+ samples compared to those obtained from animals fed the standard diet (128.5 ± 25.7 mg/dL vs. 102.9 ± 17.2 mg/dL in CTR and OL+, respectively; p < 0.05). No differences were highlighted for triglycerides, glucose, calcium, urea, ASAT, and ALAT (data not shown).3.2. Cholesterol Amount in Goat MilkThe administration to goats of a diet supplemented with 10% OL (on a DM basis) did not induce changes in the cholesterol content of milk (Figure 3), although it should be reported that the p-value associated with this variation showed a trend toward significance as it was lower than 0.1 (0.119 ± 0.009 mg/mL vs. 0.108 ± 0.008 mg/mL in CTR and OL+ samples, respectively; p = 0.066).3.3. Influence of Dietary OL Supplementation on Whole Blood TranscriptomeThe sequencing of RNA extracted from whole blood samples (CTR, n = 5; OL+, n = 5), allowed us to analyze the transcriptomic signature in dairy goats that received the OL dietary supplementation. By applying a false discovery rate (FDR) < 0.05 combined with a log2 fold change (log2FC) higher than 0.5 or lower than −0.5, only a differentially expressed gene (DEG) was identified. Specifically, in OL+ animals, there was evidence of a downregulation in the expression of Apolipoprotein B mRNA editing enzyme catalytic subunit 2 (APOBEC2; NCBI Gene ID: 102186351; FDR = 0.008; Fold Change = −47.636906).A matrix of all genes expressed in all samples with the corresponding read-counts was created, and data were then normalized by using the median of ratio, to perform the differential expression analysis. In particular, the counts were divided by sample-specific size factors determined by median ratio of gene counts relative to geometric mean per gene.4. DiscussionIn the last decade, experimentations concerning the use of agri-food by-products as dietary ingredients for farm animals had a significant boost mainly due to the need to valorize these plant matrices, the disposal of which represents both an environmental and economic issue. In the specific case of OL as a feeding ingredient for dairy ruminants, several studies have been conducted, which, even recently, have characterized the aspect concerning the potential effects on the quality of milk but also fresh and ripened cheeses [13,14,15]. To our knowledge, there is, however, a lack of information regarding the effect of this feeding strategy on animal metabolism, an aspect that can contribute to obtain clarifications about variations observed in the nutritional qualities of obtained dairy products, as well as on the animals’ health.In this study, we decided to investigate the peripheral blood of goats fed a dietary OL supplementation, since in previous nutrigenomic studies on ruminants, it was possible to identify differentially expressed genes with extreme sensitivity and accuracy [16,18,27]. In addition to this, the characterization of gene expression carried out on blood tissue allows us to collect information capable of reflecting the molecular and biochemical mechanisms occurring in other tissues or organs [28]. Specifically, the main finding of this study concerns the fact that dietary OL supplementation was shown to be effective in modifying the gene expression of lactating Saanen goats. This datum is therefore in a condition of affinity with what has already been shown in recent studies in which the enrichment of ruminants’ diet with plant matrices led to changes in the transcription process of specific genes [16,18,27]. The aspect that in any case must be immediately highlighted concerns the fact that generally, the RNA-seq approach on whole blood is capable of highlighting variations of expression that involve entire gene clusters and not only a few transcripts or even one, just like in our case. This aspect therefore deserves to be further investigated to verify whether this finding can be confirmed or enriched by identifying other molecular targets influenceable by the administered feeding strategy. In any case, to our knowledge, this study represents the first whole blood transcriptome profiling of goats fed a dietary OL supplementation.Analyses concerning blood cell count and the evaluation of specific parameters in serum samples were performed in order to obtain a preliminary assessment of the health status of the animals involved in the trial, and whether the diet had somehow induced significant effects from this point of view. In the counts of the various cell populations, no significant differences were found, while OL administration seems to have been effective in inducing a reduction in blood cholesterol. Although in this case, it is not possible to perform a comparison with similar trials, it may however be useful to refer to the study conducted by Olmez et al. [29], who evaluated the effect of an olive leaf extract in rats fed a cholesterol-enriched diet. Administration of the extract was shown to be effective in reducing cholesterol and LDL-cholesterol levels in blood, a datum therefore associated with a lower predisposition to the onset and progression of atherosclerosis. Similarly, a phenolic-rich extract obtained from olive mill wastewaters induced hypocholesterolemic effects in rats fed a cholesterol-rich diet. In this case, the finding was also correlated with broader spectrum assessments aimed at determining lipid peroxidation and antioxidant enzymatic activities in heart, liver, kidney, and aorta. Specifically, an improved lipid oxidative stability was found in the analyzed tissues, together with an increase in the catalytic activities mediated by catalase and superoxide dismutase. The study performed by Iannaccone et al. [30], in which laying hens received a dietary supplementation with olive pomace with consequent modulation of biochemical pathways related to cholesterol biosynthesis (a finding that was consistent with the reduction in egg yolk cholesterol), also fits into this area. As regards ruminants, there is a lack of specific references to the effect of the dietary administration of olive by-products in the modulation of blood parameters, and in cases where this was done, no consideration was given to cholesterol. For instance, in the study presented by Obeidat [31], the effects induced on Awassi lambs as a consequence of a feeding strategy based on the use of an olive cake supplementation were characterized. In that case, only an increase in the amount of serum glucose was highlighted, while no variations were evidenced for urea, aspartate aminotransferase, alanine aminotransferase, and alkaline phosphatase.The results obtained from the blood analysis led to also considering the data concerning the cholesterol content in milk. Very interesting is the fact that such a parameter was found to be lower in OL+ milk samples, with a trend toward statistical significance (p = 0.066). The presence of limited concentrations of cholesterol in foods is of great interest from a nutritional point of view because high levels of this compound in human plasma correlate with an increasing risk of cardiovascular disease. Moreover, in the presence of specific conditions, especially following heat treatments [32], cholesterol in food can undergo oxidation generating products, namely oxysterols, which are in turn very harmful to human health [33]. In this regard, a comparison is useful with what has been previously reported by Gómez-Cortés et al. [34], who explored alternative strategies to decrease cholesterol content in sheep milk cheeses. Specifically, authors highlighted that feeding animals a diet rich in specific polyunsaturated fatty acids was effective in reducing the cholesterol content in milk. In particular, dietary supplementation was performed by including extruded linseed in the ration. As is known, linseed is rich in polyunsaturated fatty acids, with relevant concentrations of linolenic acid (C18:3 cis-9, cis-12, cis-15), which also represent the major fatty acid detectable in the OL used in this trial [15]. Contrary to what has just been reported, no changes in milk cholesterol were observed in the study conducted by Symeou et al. [35], in which lactating Chios ewes were fed diets supplemented with ensiled olive cake. In this case, the only noteworthy changes concerned the reduction of saturated fatty acids and an increase in concentration of fatty acids of interest for consumers’ health, such as conjugated linoleic acids and linoleic acid.The sequencing of the transcripts isolated in whole blood revealed a downregulation in the expression of APOBEC2 in goats fed the dietary OL supplementation. This protein belongs to the wide family of apolipoprotein B messenger RNA-editing enzyme catalytic polypeptides, which are able to deaminate mRNA and, in specific conditions, single-stranded DNA [36]. Specifically, this protein family is easily identifiable through the amino acid similarity found at the level of the catalytic zinc-dependent domain, responsible for cytidine or deoxycytidine deamination [37].The best known and characterized member of this group of enzymes is certainly APOBEC-1, which is involved in apolipoprotein B (apoB) RNA editing. Specifically, apoB is reported to be present in mammals in two distinct forms, apoB100 and apoB48; apoB100 is secreted in the liver, while apoB48 is obtained in the small intestine following the deamination of a cytidine base in the nuclear apoB transcript, with the consequent introduction of a translational stop codon. Overall, apoB represents a fundamental element in the assembly and secretion process of lipid structures, including triglycerides and cholesterol of both dietary and endogenous origin. Furthermore, apoB glycoproteins also mediate the intravascular transport and cellular uptake of different lipoproteins; therefore, the importance of apoB is simultaneously associated with processing, absorption, and regulation of circulating lipoproteins [38]. With specific regard to this last function, recent findings allowed us to hypothesize that apoB concentration could be useful in providing a direct measure of the number of circulating atherogenic lipoproteins. This position was advanced following the analysis of epidemiological data which attributed a greater importance to apoB than cholesterol alone as a risk index for the onset of vascular diseases [39].The specific function of APOBEC2 in mammals, and especially in ruminants, has not been fully characterized; however, a discrete sequence homology with APOBEC1 has been highlighted [40], and therefore, it is not excluded that this enzyme may contribute to the apoB secretion. A significant reduction in the expression of this enzyme could therefore have positive effects on lipid metabolism, at least partially justifying the encouraging results obtained in this study in the assay of blood and goat milk cholesterol.5. ConclusionsThis is the first study to report whole-transcriptome profiling of lactating Saanen goats fed a dietary OL supplementation. The obtained findings suggest the downregulation of a gene encoding for a member of the APOBEC family. This condition seems to correlate with a reduction in blood cholesterol and a tendency for the compound to accumulate in lower concentrations in the produced milk, thus generating a condition consistent with an improvement of both animal welfare and the potential health benefits for consumers.

animals : an open access journal from mdpi

10.3390/ani11092625

PMC8469024

Animal-based weeding in vineyards is an ecological approach that cannot be implemented throughout the year, since animals are a threat to the fruits and lower branches of the vines. The SheepIT project addressed the challenge of monitoring and conditioning sheep posture by an autonomous collar. By modifying sheep behaviours, SheepIT collars allows them to be used as a vineyard weeding method. Pilot-test results showed that most animals can be conditioned using a proper combination of stimuli. As such, they interrupt their posture after audio cues. Additionally, some sheep could not be conditioned. The progression of the stimuli counters over the test days showed that the number of audio cues was higher than the number of electrostatic stimuli, proving the principle of the conditioning process, although oscillations associated with animal activity were found. The animal-conditioning analysis, and the results of the blood samples, showed that sheep bearing a collar did not face any additional stress. Additionally, the leaf-count process and the analysis of phenological evolution show that the animal’s presence did not spoil the vine’s development.

Weed control in vineyards demands regular interventions that currently consist of the use of machinery, such as plows and brush-cutters, and the application of herbicides. These methods have several drawbacks, including cost, chemical pollution, and the emission of greenhouse gases. The use of animals to weed vineyards, usually ovines, is an ancestral, environmentally friendly, and sustainable practice that was abandoned because of the scarcity and cost of shepherds, which were essential for preventing animals from damaging the vines and grapes. The SheepIT project was developed to automate the role of human shepherds, by monitoring and conditioning the behaviour of grazing animals. Additionally, the data collected in real-time can be used for improving the efficiency of the whole process, e.g., by detecting abnormal situations such as health conditions or attacks and manage the weeding areas. This paper presents a comprehensive set of field-test results, obtained with the SheepIT infrastructure, addressing several dimensions, from the animals’ well-being and their impact on the cultures, to technical aspects, such as system autonomy. The results show that the core objectives of the project have been attained and that it is feasible to use this system, at an industrial scale, in vineyards.

1. IntroductionWeed removal often consists of a combination of mechanical and chemical techniques. Between rows, simple mechanical methods, such as mowing or shredding, may be used without harming the vines. Additionally, for the space between vines, special machinery for tillage and mowing, often featuring automatic vine-skipping mechanism, is usually employed, to minimize the risk of damaging the vines. Brush-cutters can also be used and, they remove weeds more efficiently. However, this method raises the risk of damaging the vines and is more labour-intensive [1]. The problems associated with the use of mechanical methods eventually led to the widespread adoption of chemical herbicides. These are cheaper, easier and faster to apply, and can help in the suckering process (removal of unnecessary buds from the vines) [1]. However, even when correctly applied, grapes can be negatively affected, e.g., due to drift effects, eventually groundwater, fruits and soils may become contaminated [2]. Additionally, eventual long-term effects of the use of herbicides in ecosystems and human health are still under evaluation. For these reasons, wine producers, particularly those ones of high-quality wines, are either abandoning, or at least reducing, the use of these methods. Mechanical systems also have issues in terms of sustainability and environmental impact, with respect to their corresponding emissions of greenhouse gases.The use of ovine animals, usually sheep, for weeding vineyards, is an ancient, sustainable, and environmentally friendly practice used around the world [3,4]. However, this method can only be applied during part of the production cycle, namely until the formation of the fruits, as animals tend to feed from them. Therefore, the need for mechanical and chemical processes is reduced but not eliminated. Additionally, animals must be permanently monitored by human shepherds, thus incurring a significant cost, a situation aggravated by the scarcity of this kind of manpower.The SheepIT project [5] aims at developing a technological solution that ultimately eliminates the need for human supervision during grazing, by automating the functions of human shepherds. The project leverages the concepts of Cyber–physical systems (CPS), Internet of Things (IoT), cloud computing and machine learning, to monitor and condition behaviour and collect data about sheep grazing in vineyards. The abundance of real-time data brings additional benefits, such as the capability of detecting abnormal situations, e.g., animals’ illness or predator attacks, and keeps precise records of grazed areas, for management purposes. Such a system enables an effective, sustainable and environmentally friendly weed-removal process. Additionally, this solution can have a significant positive economic impact due to soil fertilization and by-product generation (e.g., milk and meat), turning a procedure that incurred a significant cost into one that generates profit.A set of pilot installations in which animals were allowed to graze within vineyard parcels, being monitored both by the SheepIT platform and by project staff, were included in the project. During these field tests, animals’ wellbeing was monitored, namely by regularly collecting blood samples to identify possible stress conditions. Additionally, the vineyard parcels where sheep grazed were also monitored, to determine the effectiveness of the automated posture-control mechanisms in preventing sheep from damaging vines and grapes. Several technical aspects were also extensively evaluated, including the effectiveness of the posture-control and localization mechanisms, as well as the system’s scalability and autonomy.This paper presents the main results gathered during field experiments carried out to assess the fulfilment of the project’s objectives and to identify aspects in which further research and development (RandD) is required. The paper is organized as follows: Section 2 reviews the state-of-the-art, both in terms of commercial solutions and academic works related to animal monitoring and posture control. Section 3 briefly introduces the SheepIT project, and Section 4 describes the tests and conditions under which they were exposed. Section 5 presents and discusses the results of the field trials and, finally, Section 6 concludes the paper by presenting the most relevant conclusions and identifying future work.2. Related Work2.1. Animal LocalizationMonitoring animals’ location is one of the more important requirements of the livestock industry, and one of the activities in which technological solutions were first adopted and are more widely disseminated. Over time, a significant RandD effort developed in this scope, resulting in the availability of several commercial solutions, for both wild and domestic animals.The scientific literature reports the use of devices incorporating the global positioning system (GPS) to locate cattle [6,7], white-tailed deer [8], griffons [9], crocodiles [10] and sheep [11]. In the last case, a GPS is combined with a jaw and a lying/standing sensor to monitor the grazing areas of domestic sheep. The device has an estimated autonomy of just a few days, but a larger battery was infeasible as it weighs almost 2 kg and needs to be transported on the back of sheep; one of the common limitations of GPS-based localization devices is the short autonomy that results from the relatively high-energy consumption of GPS modules. Additionally, its high cost and frequent loss of satellite connection [12] makes GPS unsuitable for animal localization, particularly for small–medium sized animals. Several solutions have been developed to mitigate these GPS disadvantages, often based on duty-cycling techniques [13], monitoring activity of just a few animals, taking advantage of animal’s gregarious behaviour [14] or a combination of both [15]. These processes, while, albeit, allowing the reduction of devices’ cost and energy consumption, do not allow for the monitoring of all animals of a group continuously, with the required accuracy.Relative localization solutions, in which the animals’ positions are estimated with respect to fixed landmarks with known localizations, opens the way to more compact and energy-efficient solutions, such as range-based localization mechanisms based on received signal strength indicators (RSSI). Such an approach is particularly interesting, as the application already includes a wireless network, as, today, RSSI information is available in most of radio transceivers. This method is adopted in [16] for example, where a ZigBee communication stack based on the IEEE 802.15.4 standard was used, with errors ranging from 5.2 m to 43 m. Thorstensen et al. [17] proposed a system to monitor the position of a flock of sheep in mountainous terrain, based on UHF radio modules carried by the sheep that were capable of communicating with access nodes (gateways), consisting of an UHF module, an Internet connected IEEE 802.11 and GPS.Commercial products typically offer easy-to-deploy and remotely accessible localization services, usually made available to the user via a web browser or a mobile application. The available solutions are mainly distinguished by the type of animals they address, the autonomy of the devices, the localization and communication technologies used to acquire and report the coordinates and the availability of a virtual-fence mechanism. Some of the products that target the livestock industry [18,19,20,21] share the use of GPS for determining animals’ absolute location, although the way the information is uploaded to backend services differs.2.2. Animal MonitoringThe recent and unprecedented evolution of IoT technologies, big data and machine learning (ML) technologies has created opportunity for retrieving additional and valuable information from agricultural systems. ML techniques are increasingly used in different domains of the agricultural sector, such as crop, water, soil and livestock management [22]. In the last case, most of the applications are related to animal monitoring. For instance, welfare monitoring [23], reproductive cycles optimization [24,25,26] and pasture management [27] are just some examples where ML techniques are being applied to enable the development of more efficient, accurate, cheaper and less labour-intensive animal monitoring applications.The commercial solutions presented above aim at large-scale high-value cattle enterprises. However, sheep are medium-size low-value animals, for which few commercial solutions can be found in the market, one exception being [20], whose main goal is tracking animal’s location, although it also has some behaviour-monitoring features.The scientific literature reports several relevant studies, differentiated by their methods used to attach a monitoring device to sheep (ear tag, collar, leg tag or halter), their number and type of classification states, the features used in their models and the sampling rates of their incorporated sensors. Giovanetti et al. [28] present a system that identifies sheeps’ jaw movements to build a supervision-based decision system capable of distinguishing three different feeding behaviours, grazing, ruminating and resting. The overall accuracy of the model was 93%. Decandia et al. [29] follow a similar approach, aiming at discriminating three main sheep behaviours, namely grazing, ruminating and other behaviours, while studying the effects of technical aspects (e.g., observation-window duration) in the performance of the prediction system. Compared with previous studies, the authors added one additional sensor to their device to measure the force applied by the jaw during its movements (a force sensor). In these conditions, the global accuracy attained was 89.7%.With the aim of developing a device capable of real-time monitoring of sheep for allowing the precocious detection of lameness situations, two complementary strategies were followed. In a study by Walton et al. [30], the authors discuss the most suitable choices for the sensor position, sampling frequency and observation window (OW) size. Based on the results provided by this study, a subsequent study [31] analysed the importance of different features for use in this kind of case and evaluated the behaviours of several ML algorithms for the classification of three sheep behaviours, lying, standing and walking. Both studies used a device composed of a microcontroller, a radio, an integrated accelerometer and gyroscope and a battery. This device was either attached to the ear (ear tag) or to the neck (collar)of the animal. The best overall accuracy was 95%, but in practice this value can be lower, as the study demonstrated that there was a trade-off between the number of features (and consequently, higher computing cost) and overall accuracy. Thus, a higher accuracy comes at the expense of higher energy consumption and, therefore, reduced autonomy.Barwick et al. [32] aimed at evaluating the effects of three different mounting locations of an accelerometer (ear tag, collar and leg) and at the classifying of a more extensive set of sheep states, namely: grazing, walking, standing and lying. The results showed that the mounting location had a significant impact on the detection of some behaviours while penalizing the detection of others. Accuracies above 90%, were obtained for some behaviours, while for others the performance was significantly worse. Marais et al. [32], [33] extended sheep-behaviour analysis to five states: lying, standing, walking, running and grazing. These studies also used an accelerometer for monitoring purposes and considered various technical aspects, such as sampling rates, OW sizes and ML algorithms, reporting accuracies between 80% to 90%.From the above analysis, it is possible to conclude that the accuracy of ML algorithms is closely related with the corresponding complexity and to the number of states. Therefore, obtaining high accuracies for a rich set of states in low processing-power microcontrollers is challenging. Moreover, most of the studies found in the literature employ off-line data collection and analysis, which is unsuitable for real-time behaviour control, as is proposed in the scope of the SheepIT project.2.3. Behaviour ConditioningThe most common and ancient animal-conditioning systems aimed at confining animals within predefined grazing parcels. To mitigate traditional fence issues [34], electric fences started to be used. Initially, they were used together with physical fences to avoid damages caused by the animal’s aggressive behaviours. Then, electrical fences started to be used as a standalone method for bounding grazing areas. Despite still requiring individual placement by humans, the labour involved is drastically less than that required for installing conventional fences. Furthermore, animals tend to recognize the presence of electrified wires, which enables them to avoid such structures. However, electric fences require an electrical source, typically a battery, that needs to be recharged and protected against weather conditions and other external threats. Additionally, they are not suitable for use during wet days or in places with dense vegetation, since these conditions may lead to accidents from unwanted electric conduction.The concept of virtual fences emerged to improve the flexibility offered by traditional solutions, as an electronic system that allows the definition of fence boundaries [35], in a programmatic way, based on animal-localization systems. Virtual fences require a combination of cues to compel animals to stay inside designated areas. These cues can include various sources, such as visual, auditory, olfactory, vibration or electrostatic stimulus [36]. A combination of different stimuli, for instance, sound, vibration and electrostatic discharge, to confine cattle in a virtual fence, is the most common method reported in the literature [37], where a collar producing a 6 kV electrostatic stimulus after audio and vibration cues proved that the combination of several stimuli improved the animals’ learning.The effect of sound and electrostatic stimuli was also evaluated in sheep [38], where the boundaries of the virtual-fence mechanism were based in an electromagnetic emitter. The study incorporated food and social challenges, including food with higher nutritional value and free animal grazing within the forbidden areas. Additionally, a training-process evaluation was conducted, and the results from trained and untrained animals were compared. The study found that trained animals demonstrated greater success in the conditioning process and that social challenges had a bigger impact than food challenges. The study also reported several situations where untrained animals ignored the electrostatic stimulus and joined the group of free sheep, proving that social challenges can supercede the learning from electrostatic stimuli.A number of studies report findings that may have impacts for animal well-being (e.g., [37,39,40,41]). The main conclusions are that properly designed, systems that have predictable operation, apply adequate stimuli combinations and include suitable training processes can be safely used on animals. That is, without causing any relevant discomfort to them. The Electronic Collar Manufacturer’s Association maintains a document defining good practices in the implementation of electronic collars [42] to be followed by its associates to preserve animals’ integrity and well-being.3. SheepIT ProjectThe SheepIT RandD project [43] aims at developing an autonomous system able to control sheep posture and monitor their location in real-time. The main objective is to allow the use of these animals in vineyards throughout the entire production cycle, and avoid the use of mechanical and chemical processes to remove weeds. As additional features, the system provides complementary real-time information, such as alarms (e.g., panic resulting from potential predator attacks, animals who have become lost) and health status (e.g., illnesses) relative to all animals of the herd. Additionally, a detailed record of grazing activities in each parcel (e.g., time and duration of grazing periods, number of animals) is maintained to allow the effective management of animals and parcels.3.1. Posture ControlAllowing a flock to move freely within a vineyard and feed from ground weeds but not from grape and vine leaves, requires the posture of each animal to be continuously monitored and, when necessary, interrupted by means of the application of corrective stimulus. To accomplish this objective, each animal carries a collar that features a set of sensors (inertial and ultrasound) and a microcontroller and actuators (i.e., stimulation devices, namely sound and electrostatic). The collar also includes a radio operating in the license-free ISM 868 MHz band that provides the communication link with the remaining system, and a battery.The microcontroller is at the core of the logic operation of the collar, implementing the state machine represented in Figure 1. The microcontroller periodically collects sensor data and applies decision-tree conditions, designed using ML algorithms, to estimate animal state behaviours in one of the following categories: eating, standing, moving, running and infracting (INF) [44].Figure 1 depicts the posture-control-mechanism state machine. When an infracting behaviour (INF) is detected, the system initiates a stimulation sequence. First an audio cue (cue state) is applied. If the animal persists in the infracting state, an electrostatic stimulus starts to be applied (penal state). If, during the stimulation sequence the animal ceases its infracting behaviour, the system resumes the normal monitoring operation (Idle state). The duration of the stimulus depends on a set of configurable thresholds (t_CUE, t_PEN). These thresholds, together with a stimuli configuration sequence, are used to decide when to move from the CUE state to the PEN state, as well as when to stop the stimulation (STOP state) when an animal is deemed as unresponsive. This stop conditioning is fundamental for guaranteeing animal well-being, since it allows the posture-control state machine to remain blocked until a receiving a specific signal (!BLK).3.2. Wireless Sensor NetworkThe posture control mechanism has an eminently local functionality that depends on the collar’s resources (sensors, actuators and processing). However, other system functionalities depend on aggregating, storing and post-processing data from all animals. To enable collection of this data, the SheepIT project encompasses a networking infrastructure based on wireless sensor network (WSN) and IoT principles.As illustrated in Figure 2, the SheepIT WSN is composed of three types of nodes: collars, beacons and a gateway. Collars are the mobile nodes attached to animal’s necks that send data to a gateway, over a set of fixed nodes named beacons. Beacons are planted in the vineyard area to guarantee radio coverage. To ensure an energy-efficient operation, communications predominantly follow a time-domain multiplexing scheme [45] that assigns predefined and exclusive time slots to each node for all periodic traffic. As such, collars may stay in low-power mode for extended time periods.The gateway aggregates all WSN data and sends it to a cloud application to be persisted, post-processed and made accessible to the users. Since the gateway is the WSN device empowered with the biggest computational capacity, in addition to connecting the WSN with the Internet, it also acts as the network coordinating element (e.g., managing functions such as collar registration) and supports the animal location and alarm generation mechanisms.3.3. Animal Localization MonitoringTo minimize collar energy consumption, a hybrid localization mechanism was employed, using RSSI data collected during periodic communications between beacons and collars and beacon GPS absolute location. Implementing such a localization mechanism was relatively complex. The reason for the complexity is because the implementation is carried out at the gateway, which is the device with more relaxed energy and processing constraints.At system boot, the gateway obtains the beacons’ GPS location, which remains fixed during normal system operation as the beacons are fixed devices. Then, RSSI values, measured during the communications between beacons and collars, are continuously sent to the gateway. These RSSI values are used to obtain a distance estimate between collars and beacons, using a calibrated path loss signal propagation model. Then, the relative distance between each collar and the beacons that are within communication range (or a subset of them, if more than three are in range) is used to estimate the relative location between each collar and the corresponding beacons, through a trilateration method. This relative location is then combined with the absolute location of the beacons to obtain collars’ absolute location. RSSI-based distance estimates are subject to several impairments that result from the appearance of obstacles, signal reflections and hardware variability, that impact the correctness of the propagation model. For this reason, a set of filtering processes are employed to improve the estimation of the collars’ position [46].3.4. Cloud Computational PlatformThe Computational Platform (CP) enables the storage and processing of data collected via the WSN, in a secure and scalable environment. Additionally, it allows using data -mining and machine -learning tools to create additional valuable services to the user, such as the detection of illness and lameness conditions, the monitoring of animal reproductive cycles, the evaluation of preferred grazing areas and timings, or the identification of social relationships. The CP follows the architecture illustrated in Figure 3, being composed of five different interconnected modules. The message-oriented middleware (MOM) was implemented as an AMQP broker (RabbitMQ [47]) and it is the data-interface point between the CP and the WSN. It routes incoming messages sent from the gateway to the CP components that process and store the data.RabbitMQ receives data from the M2M network through the gateways and serves two components: a processing framework implemented, based on an Apache Spark server [48], and a rule manager. The Apache Spark sever is a processing framework that subscribes and consumes data from the RabbitMQ queues and orchestrates all the operations, such as alarm generation, data processing and data persistence. It combines the processing of real-time stream traffic with a batch process that handles non-periodic traffic coming from processing tasks performed within the platform. Both processes persist relevant data into the data storage, which is updated regularly with new data. As a complement to the Spark processing framework, the rule management module [49] was included for the definition of complex event processes and event-stream processes as the generation of predictions, detection of patterns or triggering alarms and real-time notifications.Finally, a representational state transfer (REST) application programming interface (API) was included to ease the access to data, both for web-based user access, and for automatic integration of legal animal information systems.4. Materials and Methods4.1. Vineyard Parcel and FlockTo assess the impact and effectiveness of animal weeding on the vines, as well as to evaluate the animal well-being when using the SheepIT collar, a pilot test was conducted during 2018 growing season. The vineyard was located at the Agrarian Superior School of the Polytechnic Institute of Viseu, Portugal. This vineyard is in the Dão wine region, latitude 40°38′17.33″ N, longitude 7°54′57.07″ W and altitude of 452 m. It is a non-irrigated vineyard, installed in the year 2008, with “Touriga National” cultivar vines. Its approximate area is 0.57 ha and includes about 4000 plants. The vines were spaced 1.0 m between plants within a row and 2.30 m between rows, trained on a vertical shoot, positioned with a pair of movable wires and pruned on a unilateral Royat pruning machine with six fruiting units.The sheep were of the “Serra da Estrela” breed and were used in all experiments. The number of sheep used differed according to the experiment, as detailed in the corresponding experimental description.4.2. Installed PlatformAll sheep selected for each experiment carried a SheepIT collar for monitoring and conditioning sheep posture. To collect data, it was installed an infrastructure composed of a set of seven beacons, spread out evenly to cover all vineyard plots and to enable real-time data collection and localization. Beacons were placed on the top of poles, around 80 cm from the upper vine line (Figure 4) to improve radio coverage. The animals grazed freely throughout the test lot, and communications from collars were received by one or more of the beacons. Beacons were configured to relay collar data to the gateway (Figure 5), which was then sent by this device, through a cellular interface, to the CP’s broker.4.3. Grapevines Phenological Development and Leaf CountTo evaluate the impact of sheep on the vines, two methods were used, (i) the analysis of vines’ phenological development and (ii) a count of the number of vine leaves. The former is concerned with the periodic phenomenon of the vine growing cycle (e.g., bud burst, flowering and veraison), that influences the timing of the numerous operations in the vineyard (e.g., phytosanitary protection, defoliation, crop thinning) and depends on environmental factors, such as soil management [50,51,52,53]. The latter aims at counting the number on leaves on a sample of vines, before and after grazing.To implement both methods, the vineyard was divided into six experimental plots (Figure 6) with about 300 m2 each. These plots were split into two equal-size groups, one where sheep were allowed to graze (G) and the other for control purposes (non-grazing—NG), following a cross distribution.For the phenological development analysis, we chose five representative vines in each one of the six plots. On each of these five vines a fruiting unit with two buds was labelled. The thirty marked vines were monitored every week, by direct observation of the phenological state of each organ (shoot/bunch), using the numerical scale of the modified Eichhorn–Lorenz system [54].The analysis of leaf area consisted of the use of an empirical non-destructive methodology proposed by Lopes and Pinto [55] that serves to determine leaf area indirectly. In this experiment, the monitoring was performed on two shoots of the same plant, for every five plants in the experimental unit. This monitoring was performed before and after grazing.4.4. Animal Behaviour, Animal Location, Posture Control and Animal Well-BeingDuring behaviour-conditioning tests, twelve sheep were used to graze in the vineyard. Sheep were allowed to feed within the vineyard during two different periods, with two different purposes. First, to collect data to optimize the operation of the behaviour-detection mechanism and second, to evaluate the effect of the posture-conditioning mechanism on sheep behaviour and welfare.In the first period, a postural behaviour data collection was conducted to tune the posture-detection logic. During this phase, a 3-h experiment was conducted, where a single animal at a time was released onto a single plot, its activity recorded by video and its collar’s sensor data gathered. Both video recording and sensor data acquisition were time-synchronized to allow the correlation of their data.The collected data was then processed, in a data-preparation phase. This processing included removing redundant and duplicated data, verifying its integrity and removing null values. The resulting dataset was subsequently split randomly into a ratio of 75–25%: a training set and a test set.In the second period, the flock was brought together within the vineyard and allowed to graze freely for four days. During this period, animals were conditioned solely by the collars, without human intervention. The communication platform was kept in the sensor data gathering state, allowing the project staff to remotely monitor animal behaviour.During these experiments, sheep were supervised by zootechnicians and veterinarians, who observed the animals, checked their behaviour, collected blood samples for determining serum cortisol levels and recorded heart and respiratory rates. Blood sampling and heart and respiratory rates were also measured before and after the tests to assess any stress induced by the collars.5. Results and Discussion5.1. Posture DetectionAccording to the observed behaviours recorded by video, each entry of the dataset created during the gathering phrase was labelled by project team members with one of the following set of behaviour states: infracting, resting, eating, moving and running. Since animals tend to be grazing most of the time, most of the dataset observations were of the type eating (70%). The less representative states were infracting (3%), running (2%) and standing (2%). This distribution was expected, as sheep were free to graze on the area.After applying feature-transformation and feature-selection techniques, the following features were selected to train the model: distance measured by the ultrasound transducer; pitch, yaw and roll angles; mean, standard deviation, maximum, minimum, number of zero crossings and dominant frequency of the dynamic acceleration vector; static acceleration on the y and z axes; and nEqualStates (number of samples on which no changes in the state are registered).Decision trees (DTs) were used to model the system, since they allow an easy interpretation of modelling results, as well as provide a set of conditions that can be implemented directly in the collar. This set of features not only allowed us to obtain a global accuracy like those obtained in the mentioned state-of-the-art studies, but also allowed a reduction of the number of false positives and false negatives associated with the infracting state that were registered in preliminary tests.5.2. Animal LocalizationRSSI values measured through communication messages were transmitted and stored for collar localization and tracking purposes. Additionally, a GPS logger was attached to one of the sheep to enable comparison between GPS-tracking- and SheepIT-generated tracks post-experiment.The SheepIT localization mechanism version 1, presented in [56], allowed us to get an average error of 4 m at 10-m distances, and an average error of 10.9 m at 30-m distances. These results are considered acceptable, considering the expected application of animal tracking for activity-monitoring purposes.Figure 7 illustrates the projection of beacons (yellow pins) and tracks on the ortho-photo map of the vineyard plot from Google Earth. The blue line is the track registered by the GPS device, and the red line is the track generated by the SheepIT localization algorithm.During the experiment, several accidents were detected with the collars’ antennae. Particularly, as they were fixed to the collars’ supporting straps, they were often damaged with use. During the prior SheepIT industrialization process, it was decided to place the antenna inside the collar’s case, to reduce damage and thus reduce maintenance and cost. However, this increased robustness created a loss of precision in the localization process, particularly because of the high impact of radio-frequency propagation issues, such as signal reflections and signal fading.The initial results, after including the antenna inside the box (Figure 8), confirmed the expected loss of precision of the localization process, which required the implementation of filtering techniques [46] (Figure 9). We included Kalman and speed-limitation filters, allowing the device to achieve a localization accuracy like that obtained when the antenna was outside the collar’s case. It was also confirmed that the localization error strongly depended on a proper beacon density and placement. To be effective, the location system may require additional beacons that would not be strictly necessary for ensuring radio coverage, thus negatively impacting on the installation and operation costs. Even though, as beacons are the simplest devices and have a relatively low cost, this is not considered an issue.5.3. Posture ConditioningEvaluating the data collected during the first day of tests had, as its main purpose, understanding the effectiveness of the training process, which is considered fundamental to the success of the system. Figure 10 illustrates the evolution of the number of audio cues (posture buzzers) and electrostatic stimuli (posture shocks) on a sheep with a favourable behaviour (collar D). It can be observed that there was a sharp growth in the number of detected infractions at the beginning of the day (in the first hours of the morning), together with a small increase in the number of penalizations. In contrast, in the afternoon period it was observed that there was a stabilization of both indicators, registered as smaller and less frequent number of infractions, as well as an almost null increase in the number of penalizations. Such behaviour is in line with the behaviour observed in several animals, as shown in Figure 11. It can be seen that the ram-born collar M remains more idle in the early afternoon. This behaviour seems to be evidence of a successful training process among all the sheep.In contrast to previous examples, Figure 12 typifies a clear example of a refractory sheep. As can be seen, the infraction and penalization indicators rise continuously and are of an order of magnitude larger than those presented previously. Despite a certain level of softening early in the morning, the indicators started to increase continuously, suggesting that this animal di not react to any of the stimuli. Another interesting observation is that, besides the high number of stimuli, it was not possible to observe any kind of discomfort in this particular sheep. The notion of some malfunctioning of the collar was raised, but such could not be confirmed. Therefore, this sheep was identified as refractory and removed from the test.Generally, animals repeated the first day behaviour in the remaining days of the experiment, and most of animals presented a low number of penalizations. Figure 13 depicts an example of a sheep with favourable behaviour over the whole experiment (Collar H). At the beginning of the day, it was noticed that a high number of infractions were detected, but it stabilized and, at the end of day one, only two penalizations were registered. It can be observed that the sheep behaviour during the second day is quite like that observed during day one, even if the number of infractions increased.On the third day, a significant increase in the number of infractions and penalizations was observed in all animals. Even if the sheep continued reacting to the posture-control mechanism, a clear deterioration in the behaviour was detected and for which we have no explanation. On the last day, the evolution of the counters followed the trend of the previous days, but the tests were shorted, as they were stopped just before the lunch break.The ram, being of a larger size (Collar M), revealed its infracting behaviour on the first day, as shown in Figure 14. The animal insistently fed from vines, despite being observed having a reaction to the conditioning mechanism. On day two, a similar behaviour was observed in the morning, with a high increase in the number of infractions and penalizations.Nevertheless, the values appeared to stabilize over the rest of day. This did not happen on day three, however, i.e., the animal seemed to ignore the stimuli. To understand the behaviour of this animal, and how interpret it from the infraction counters, it is important to note that its state distribution followed the common pattern, identified in Figure 11. Sensors present in the collars continuously read the values and classified the animal’s behaviour. Despite the high number of situations in which the animal’s posture was considered inadequate, the number of samples collected indicated that the posture was evaluated adequate.5.4. Power Consumption and System AutonomyCollar autonomy is a function of battery capacity (the collars used in the tests had one lithium battery 18,650 3.6 V 2650 mAh), the energy consumption of the various modules of the system and their duty-cycles, and number of stimuli applied. To minimize energy consumption, the collar’s software is designed to enter low power modes whenever there are no active functions. To this end, it employs a time-triggered architecture, in which local activities are synchronized with the communication infrastructure so that, in turn, it uses a communication mechanism based on TDMA [45], wherein an exclusive communication timeslot is assigned to each device.The animal conditioning system has a critical impact in collar’s power consumption because of the significant amount of energy consumed by the electrostatic stimulation. Even if the collar’s sensors are activated at specific periodic instants, in coordination with the communications state machine, the triggering of stimuli depends on animal behaviour. Therefore, a collar coupled to a well-behaved animal would present a high autonomy, while a collar coupled to a misbehaving animal is expected to drain the battery much more quickly.Taking advantage of the capability of collars to measure and report their battery levels, we used the values gathered during the posture conditioning tests (Section 5.3) to assess the evolution of the battery charge in real operating conditions. Figure 15 illustrates the results of this assessment, where it can be seen that collar autonomy exceeded the week of its operation (since, after three days, the lowest value was around 85%). Even so, there is a set of possible optimizations in collar operation for increasing its autonomy, from increasing its battery capacity, to redesigning the state machine so that the collar goes to sleep while the animal is lying down, and turns collars off at night.5.5. Animal Well-BeingAll tests involving animals were fully monitored by zootechnicians and veterinarians, to ensure their welfare and safety. To assess the stress caused by using collars, each animal was individually checked by the veterinarian, before and after the tests. During each of these procedures, cardiac and respiratory rhythms were measured and blood samples were taken for analysis.The animals’ heart rates and respiratory rates showed normal values for the species (104.7 ± 13.7 and 30.0 ± 5.1, respectively). Additionally, there was no significant variations in cortisol values measured before (12.21 ± 1.070 μg/L) or after the application of stimuli (14.75 ± 1.617 μg/L). Since these values did not exceed the reference values [57,58,59], it is possible to conclude that the use of collars does not cause panic nor apparent pain. Nevertheless, the gathered data and manual observations taken during the project allowed us to detect a few refractory animals (one during animal behaviour tests), and those animals were excluded from the tests to assure their welfare.5.6. Impact on Vine Leaves and on PhenologyWithin each plot, the vine leaves were counted twice, once before starting the experiment and again at the end of each test day. The mostly significant loss of leaves occurred on the first day, when 32 leaves were lost (out of a total of 4555 or 0.7%). On the remaining days, the number of leaves lost per day decreased significantly (Table 1).The effect of grazing on leaf area was insignificant (p > 0.05), with a maximum of 0.7% of leaves lost, a minimum of 0.2% and an average of 0.3%. This result is in line with the posture-control results, particularly those shown in Figure 11. It should be noted that grazing-effect tests over the vineyard were not carried out at the same time, nor in the same lot as the animal-behaviour tests, although both took place over four days.Thus, according to Lopes and Pinto [55], who also used leaf area to estimate vigor [60], the results suggest that neither vine productivity nor vigor were altered during grazing. There were no significant differences between the grazing and non-grazing plots, which leads us to conclude that the presence of the sheep did not impact the development of the vines. The main phenological stages of the grapevines, budbreak or flowering and veraison, occurred in almost all the vines at the same time.6. ConclusionsThe SheepIT pilot tests, despite being short, verified the animals’ conditioning by the collars. Most of sheep were successfully conditioned by SheepIT’s collars, although a few refractory sheep were identified. Unwanted behaviours were successfully detected and reverted, most of them using exclusively audio cues as stimuli. This was essentially indicated by the results of the leaf count, which, together with the conducted phenological analysis, confirmed that the presence of animals equipped with collars does not pose a threat to the vines.The tests were carried out to ensure no ill effect of wearing the collars for animal well-being. These tests confirmed what has been reported in the literature—the effect of appropriate stimuli does not cause stress to sheep.The duration of the field tests was short and there are several questions regarding animal behaviour to be answered, such as how effective the device is in long term use, and whether a predictive and static conditioning mechanism is sufficient to ensure an effective posture-control mechanism. More information is required on whether animals will try to circumvent the postural control mechanism, and if any potential long-term impact in animals’ health, milk production or calving exist.Though the technical implementation of the posture-control mechanism could be validated in a real-world scenario use-case, there is still more research to be conducted, particularly regarding the field of animal learning. For instance, we suggest the developing of a deeper understanding on how the stimuli configurations and sequences could be dynamically adjusted to become suitable and effective to all sheep.

animals : an open access journal from mdpi

10.3390/ani11092634

PMC8466846

This study examined the quantitative discrimination abilities (using two-dimensional objects) in a shark and a stingray species. Both species underwent a training procedure, followed by a series of transfer tests designed to investigate whether they could extrapolate and apply learned knowledge to a set of new quantity discrimination tasks. Sharks and rays successfully mastered the training tasks as well as most of the transfer tests. This included the discrimination of 4:1, 5:2 and 7:5. The present study is the first to describe numerical abilities in elasmobranchs, in which any potentially confounding, non-numerical factors (e.g., size and area) were controlled for. This study adds to the growing number of studies on fish that are key to understand the evolution and development of cognition in vertebrates.

Over the last decade, studies examining the cognitive abilities of fish have increased, using a broad range of approaches. One of the foci has been to test the ability of fish to discriminate quantities of items and to determine whether fish can solve tasks solely on the basis of numerical information. This study is the first to investigate this ability in two elasmobranch species. All animals were trained in two-alternative forced-choice visual experiments and then examined in transfer tests, to determine if previously gained knowledge could be applied to new tasks. Results show that the grey bamboo shark (Chiloscyllium griseum) and the ocellate river stingray (Potamotrygon motoro) can discriminate quantities based on numerical information alone, while continuous variables were controlled for. Furthermore, the data indicates that similar magnitudes and limits for quantity discrimination exist as in other animals. However, the high degree of intraspecific variation that was observed as well as the low rate of animals proving to be successful suggest that the ability to discriminate quantities may not be as important to these species as to some other vertebrate and invertebrate species tested so far.

1. IntroductionThe ability to discriminate quantities has shown to be a widespread cognitive skill in a variety of animals including mammals [1], insects [2], birds [3], fish [4], amphibians [5] and reptiles [6]. This is not surprising, as quantity discrimination is advantageous for many behaviours including foraging [7,8], mating [9,10] and predator avoidance [11,12]. There is an ongoing debate in the literature on how quantities are perceived and processed [13]. On the one hand studies suggest that two core systems exist, which play a prominent role in the ability to quantify larger and smaller numerical ratios—the object file system (OFS) and the approximate number system (ANS) [14,15,16,17]. The OFS is faster, more accurate and enables identification of individual objects with a one item difference being enough for distinction. However, it is limited to a small number (≤5) of presented objects [15,18,19,20]. The ANS is associated with estimating larger numerical magnitudes (>5) based on Weber’s Law ‘of the just noticeable difference’ [15,21,22]. On the other hand it has been suggested that these two are in fact not individual systems but that the ANS is in charge of processing both small and large quantities [23,24]. Studies in mammals, fish and other animals have provided evidence to support either of the two hypothesized systems [4,25,26,27]. Agrillo et al.’s review argues that the results of most studies on fish are supporting the two-system hypothesis [13]. Therefore, the present study is focusing its analysis based on the hypothesized two core system and will not engage in debating whether a two core or single core system exists.In the past, the research focus concerning numerical abilities was mainly on primates and birds; however, an increasing number of studies on a few fish species emerged over the last decade [28]. The majority of these studies involved guppies (Poecilia reticulata), angelfish (Pterophyllum scalare) and mosquitofish (Gambusia holbrooki) [28]. Guppies have shown to successfully discriminate 5 vs. 4 items while three-spined sticklebacks even distinguished 7 vs. 6 [12,29]. For larger quantities (>5) fish, like other vertebrates, seem to rely on Weber’s law and are only able to distinguish between the size of two sets of given objects when the ratios are lower. Mosquitofish and guppies successfully discriminated quantities up to 12 vs. 8 (mosquitofish) and 12 vs. 6 (guppies), while both species failed at 12 vs. 9 [30,31]. Ratios tend to have an important influence on the species’ ability to discriminate quantities, especially when the numbers fall within the ANS. Mosquitofish showed a decrease in performance when the ratio increased and the numerical distance between two sets of stimuli decreased [30]. Animals have been shown to use both numerical (discrete) and continuous variables (e.g., surface area, size, density etc.) to solve quantitative tasks [32]. Discrete quantities are countable and represented by a certain value while continuous variables are a measurement of a specific quantity that correlates with the number of items that are presented. This means for example that the surface area of a fixed object increases with its numerosity and therefore needs to be controlled when investigating the ability to discriminate discrete quantities. This control prevents the influence of actively avoiding or choosing the larger area occupied by the larger amount of stimuli presented [32]. Agrillo et al., showed that the availability of both continuous and numerical information combined leads to a more accurate outcome than one of them alone [33]. Accordingly, when testing numerical discrimination, the use of continuous variables needs to be controlled for, and this “represents one of the most critical issues in this research field” [34]. In one of the first studies on fish, where continuous quantities were controlled for, mosquitofish demonstrated the ability to discriminate 2 vs. 3 [35]. Since then, discriminating abilities relying only on numerical information have been demonstrated in guppies (Poecilia reticulata) [36,37], angelfish (Pterophyllum scalare) [38], Siamese fighting fish (Betta splendens), zebrafish (Danio rerio), redtail splitfins (Xenotoca eiseni) [39], goldfish (Carassius auratus) [40] and a blind cavefish (Phreatichthys andruzzii) [41].Several elasmobranch species, including the grey bamboo shark (Chiloscyllium griseum) and the ocellate river stingray (Potamotrygon motoro), have been part of a range of cognition experiments and have shown various visual discrimination as well as orientation abilities (see reviews [42,43,44,45,46,47,48,49,50,51,52,53]). Adding to that, freshwater stingrays (Potamotrygon castexi) are able to use water as a tool to extract food from a tube [54]. Even though the class of Chondrichthyans is at least 450 million years old and represents the oldest lineage of vertebrates [55], only one study demonstrating the ability to discriminate quantities in Port Jackson sharks (Heterodontus portusjacksoni) has been published [56]. The overall aim of the study was to investigate potential influences of elevated ocean temperature on cognitive abilities. During the research, continuous variables were not controlled for and therefore only the low cognitive load to distinguish quantities was shown.For the current study, C. griseum and P. motoro were trained to visually discriminate quantities, as both species have shown to be well suited for visual cognition experiments. By training animals, opposed to testing for spontaneous choice, one can control all variables potentially influencing the animal’s choice [34]. In previous studies, grey bamboo sharks were able to distinguish between symmetrical and non-symmetrical shapes, various stationary objects (including geometric forms) as well as moving stimuli and categorized two-dimensional objects [49,52,57]. They also performed well in a range of optical illusion experiments [50,58]. In recent studies a wide range of visual discrimination abilities were shown in P. motoro, including first evidence for memory retention [45], serial reversal learning [46] and colour discrimination [45]. In addition, C. griseum and P. motoro both seem to be able to use spatial maps to orientate, can use different orientation strategies and memorize spatial tasks over a period of several weeks or months [51,59].Individuals of both species underwent similar training procedures for this work and were subsequently tested on whether they could transfer numerical abilities they have previously learned to unknown ratios. This study is the first to examine the ability of elasmobranchs to use numerical information in which continuous variables were controlled for. It adds valuable insights to the field of shark and ray cognition and to understanding the evolution and development of cognition in vertebrates.2. Materials and Methods2.1. Housing and Experimental FacilitiesThe sharks were kept in five interconnected aquaria (each 300 L) and two 450 L aquaria, either individually or as pairs. All tanks were connected to an additional tank equipped with a filter unit, a protein skimmer and a UV lamp. To ensure constant water parameters throughout the setup (ca. 1.0217 kg salt/dm3; conductance: about 50 μS/m; 25 ± 1 °C; pH 8, KH 9–12°, NO2− < 0.2 mg/L), parameters were checked twice a week, adjusted using Aqua Medic AB Reef Salt and Reef Life System Coral B buffer (AB Aqua Medic GmbH, Bissendorf, Germany). Weekly water changes (about 20% of total volume) were conducted. Experiments took place in a separate tank (50 cm ∗ 50 cm ∗ 100 cm) (Figure 1a), filled with water from the housing-system every morning and emptied and rinsed at the end of each day. The experimental tank was divided into a Starting Compartment (SC) and a Decision Compartment (DC). Both were separated by a grey PVC panel containing a hand-operated guillotine door (Figure 1a). All walls and the bottom side were covered with blue non-transparent adhesive foil on the outside of the tank to prevent the sharks from any visual disturbance. The lower half of the wall at the end of the DC was left uncovered to allow light from a projector (Optoma ES521, Optoma Corp., Taipei, Taiwan) to hit a milky colored plastic pane (50 cm ∗ 21.5 cm) on which the two-dimensional stimuli were displayed. The stimuli were prepared and presented as a PowerPoint presentation (ver. 16.24) and operated from a laptop next to the experimental setup. Both stimuli were arranged on the same slide during each trial. To ensure clear lateral decision-making, the milky pane featured a separating 10 cm broad and 9 cm high divider in its center. The divider’s breadth also indicated the beginning of the Decision Area (DA) at which a choice was counted if the animal passed over it. To distinguish the DA from the rest, the tank’s underside in the DA was covered in white foil. This created a line to ensure unambiguous data collection of the decision-making process. A webcam (Logitech c170, Logitech, Lausanne, Switzerland) was mounted on the ceiling above the setup to allow for visual observations of the shark during the experiment. It was only used to observe the shark’s choice, without the shark being able to see the experimenter during the trial.Each stingray group (5 and 9 individuals) was kept in a large wooden tank (233 cm ∗ 233 cm ∗ 45 cm) lined with black pond foil covered in sand to enable rays burying themselves (Figure 1b). EHEIM Jäger Aquarium Heaters, two airstones, a pump and filter system and a weekly water change ensured steady water quality (conductance: 380–420 μS/m; 28 ± 1 °C; pH 6–8, KH 2–7°, NO2− < 0.3 mg/L). Parameters were checked twice a week and adjusted using Aqua Medic AB Reef Salt and Reef Life System Coral B buffer (Aqua Medic, Germany). Experiments took place in a separate compartment (71 cm ∗ 125 cm ∗ 28 cm) within the pool (Figure 1b). Its partitioning was the same as the one used for the sharks, with a starting box for the rays added to the SC. The set-up only really differed in the way the stimuli were presented. Instead of using a projector to display stimuli, the experimenter changed a set of printed and laminated cards by hand (due to the wooden nature of the tank).2.2. Experimental ProcedureAll animals were trained in two-alternative forced-choice visual experiments (for an overview of all experiments see Table 1 in the Section 3). Each experiment consisted of up to 30 sessions with ten trials per session, in which a defined learning criterion (LC) needed to be met. To reach the LC, the animals had to make at least seven out of ten correct choices (70%) in three consecutive sessions before reaching 30 sessions (c 2 (1) ≤ 0.05). This predefined criterion ensured statistical significance in choice when reaching the LC. For all animals that completed an experiment, a two-tailed binomial test (95% confidence interval) was run in R (ver. 1.4.1103) to assess whether the sharks had a significant choice towards either the negative (lower quantity) or positive (higher quantitiy) stimulus.An animal was placed in the SC of the respective experimental tank and given a few minutes to acclimatize. A trial commenced once the shark or ray crossed the opening of the guillotine door with the tip of its head, entering the experimental area. A choice was made by crossing the DA line with the tip of the head on the respective side (Figure 1). The time was stopped manually with a stopwatch (CG-501, Genutek Electronics Co. Ltd., Guangdong, China). A maximum time of 120 s was allowed for the decision-making process. When choosing correctly, the animal was fed and given a couple of seconds in front of the stimulus. A wrong choice was not rewarded with food and the animal was immediately guided back to the SC using a rubber kitchen scraper. To ensure olfactory cues were not influencing the experiment, the water was stirred after each trial. An inter-trial time of 30 s, which started when the shark or ray had returned to the SC and the door was shut, separated the trials. After the experiment, the animals were transferred or guided back to their resident tank. Sharks and rays underwent up to two experimental sessions per day, six times a week. The time between the two daily sessions for one individual was at least five hours.2.3. Experimental AnimalsFor this study, two juvenile grey bamboo shark (Chiloscyllium griseum) groups and two ocellate river stingray (Potamotrygon motoro) groups were trained in an experimental procedure. Animals were divided into separate groups due to housing limitations and the experiments taking place over the period of two years on different individuals. Except for one group of rays (Group 3), all individuals were experimentally naïve and distinguishable by phenotypic characteristics (for details on size and age see Section 2.3.1 and Section 2.3.2). Food consisted of pieces of shrimp, fish, squid and earthworms and was provided during experiments only. Once a week, ten drops of a vitamin solution (Atvitol, JBL GmbH and Co. KG, Neuhofen, Germany) was added to the food. By using automatic time switches, all animals were kept in a 12 h light and 12 h dark cycle.2.3.1. Grey Bamboo Shark (Chiloscyllium griseum) (Group 1): 8 individuals (7 females, 1 male), head-tail length 35–40 cm, 1 session per day The first group of sharks was trained to discriminate 4 vs. 1 geometric symbols (square, circle, triangle) where all continuous variables (color, surface area, density (space occupied on slide) and geometry) were controlled for (except size of symbols) and choosing the higher quantity was rewarded with food. The first transfer test was therefore controlled for size and all stimuli had the same individual size (Figure 2e). Twenty transfer test trials for each shark were implemented. For the second part of the experiment the sharks were trained to discriminate 5 vs. 2 geometric symbols where all continuous variables (size, color, surface area, density (space occupied on slide) and geometry) were controlled for and choosing the higher quantity was rewarded with food. Transfer tests for this experiment consisted of unifying the stimuli color, showing non-geometric stimuli, reducing the surface area of the positive stimulus and examining ratios of 0.57 (7 vs. 4), 0.6 (5 vs. 3) and 0.8 (5 vs. 4) (Figure 2). (Group 2): 7 individuals (5 females, 2 males), head-tail length 25–35 cm, 2 sessions per day The second group of sharks underwent training to discriminate 4 vs. 1 and 5 vs. 2 geometric symbols with the color not being controlled for. As none of the sharks was able to reach the LC the experiments were further simplified following Agrillo et al.’s and Bisazza et al.’s approaches [33,36]. A second 5 vs. 2 and a 2 vs. 1 training experiment was implemented with the surface area and the color not being controlled for. A repetition of the Pretraining was conducted after finishing all regular experiments to exclude any visual or cognitive impairments throughout the study.2.3.2. Ocellate River Stingray (Potamotrygon Motoro) (Group 3): 5 individuals (1 female, 4 males), disk diameter 17–22 cm, 2 sessions per day The first group of stingrays consisted of animals that had previously been trained on a different non-numerical experiment, including the Pretraining procedure. For this study they underwent training to discriminate 5 vs. 2, 3 vs. 6, 4 vs. 7, 5 vs. 7, 6 vs. 7 and 7 vs. 2 while only the surface area was not controlled for. In the subsequent transfer tests the surface area was controlled for and the rays were tested whether they could still discriminate the quantities they were trained on. (Group 4) 9 individuals (5 females, 4 males), disk diameter 11—14 cm, 2 sessions per day The second group of stingrays were only trained on 4 vs. 1, 5 vs. 2 while continuous variables (size, colour, surface area, density (space occupied on slide) and geometry) were controlled for. During the transfer tests, a variety of different ratios were tested to examine the ability of the rays to discriminate quantities based solely on numerical information. In addition to the tested ratios, a transfer test was designed to quantify whether the animals used a relative or absolute approach to discriminate quantities.2.4. Experiments2.4.1. PretrainingAll animals were tested in a “Pretraining” experiment to ensure their cognitive and visual wellbeing and their habituation of the experimental procedure. Animals needed to differentiate between a black circle and a blank, by choosing the side on which the black circle was displayed (Figure 2a). The correct choice was rewarded with food. All animals were examined during the Pretraining and evaluated whether they successfully managed to reach the LC before the 30th session. Those individuals that were experimentally naïve and did not reach the LC before the end of the 30th session, were given an additional 15 sessions after which their performance was analyzed. If a statistically significant choice (p < 0.05) for the correct stimulus occurred, they were permitted to the subsequent training sessions, else they were excluded from all experimental procedures (Figure 3).2.4.2. Numerical Discrimination TrainingIn the numerical discrimination training, two different stimuli that differed in the number of objects displayed were presented to the animals. Four different randomized object/side orders were developed and circulated to prevent animals from simply memorizing the order of the presented stimuli. During a session, each stimulus was presented five times on the right and five times on the left side, but never more than twice consecutively on the same side. For each group, a different set of symbols (but all consisting of geometric forms) was used. Once the LC was met and animals performed consistently above chance level, an 80% rewarding scheme was implemented, i.e., animals were maximally rewarded in 8 out of 10 correct trials. Two trials were randomly assigned to not be rewarded before each session, irrespective of actual choice. This procedure prepared animals for transfer testing, in which trials were generally unrewarded, by preventing a strong association of ‘no food reward’ with an ‘incorrect’ choice (which could have kept animals over time from participating in these trials).2.4.3. Transfer TestsOnce the LC was met, and the 80% rewarding scheme implemented, transfer testing began. Transfer tests were designed to examine the animal’s ability to transfer and extrapolate what it learned during training onto a new task. The regular training trials continued, but one or two transfer tests were added to the regular ten-trial session and were not rewarded. Transfer tests were carried out to examine (i) the ability to discriminate low and high ratios, (ii) a possible influence by shading and size, (iii) the ability to discriminate quantities of non-geometrical stimuli and (iv) whether a relative or absolute strategy was used.3. Results3.1. PretrainingFor group 1 six out of eight sharks reached the LC within 27 ± 7 sessions and six out of seven sharks from group 2 within 12 ± 9 sessions. All five rays of group 3 obtained the LC in 7 ± 10 sessions and eight out of nine rays from group 4 in 28 ± 17 sessions. It needs to be considered that the rays of group 3 were not experimentally naïve and have been previously trained in this setting. An exemplary graph of a Pretraining learning curve in a grey bamboo shark is shown in (Figure 4), where the LC was reached in the 13th session.3.2. Sharks3.2.1. Group 1Seven out of eight sharks were successfully trained to discriminate 4 vs. 1 within 30 sessions. On average the sharks needed 14.57 ± 7.56 sessions to reach the LC. A graph showing the performance of an individual shark during training and subsequent training while transfer tests were added can be found in Figure 5. The seven sharks were permitted to take part in twenty transfer tests each, where the area of the single negative stimulus was adjusted to the same size as one of the positive ones. Pooled as a group, the sharks chose the higher numerosity significantly more often (n = 7, p < 0.0001) (Figure 6). Three out of eight sharks were successfully trained to discriminate 5 vs. 2 within 30 sessions. On average, the sharks needed 17.33 ± 10.01 sessions to reach the LC. Subsequently, all three individuals each took part in twenty trials for each of the following transfer tests. Pooled as a group, the sharks chose the higher numerosity significantly more often while the stimulus color was the same (n = 3, p < 0.0001), non-geometric stimuli were shown (n = 3, p < 0.0001), the positive stimulus was reduced in size (n = 3, p < 0.0001) and the ratio was increased to 0.6 (5 vs. 3) (n = 3, p < 0.0001) and 0.57 (7 vs. 4) (n = 3, p < 0.001). No significant choice for the higher numerosity was shown for 5 vs. 4 (n = 3, p = 0.155) (Figure 6).3.2.2. Group 2Six out of seven sharks passed the Pretraining test but were not able to be successfully trained on specific ratios, even after they were provided with additional continuous variables to ease the ability to discriminate quantities (following [33]). A repetition of the Pretraining was conducted after finishing all regular experiments to exclude any health-related, visual or cognitive impairments throughout the study. Compared to the first time the sharks were examined in the Pretraining, the average number of sessions to reach the LC slightly decreased from 12 ± 9 to 12 ± 6 sessions, while the number of sharks reaching the LC (six out of seven) remained the same.3.3. Rays3.3.1. Group 3As described before, this group of rays was always trained on a new ratio while the surface area was not controlled for. Individuals that reached the LC in the specific training were additionally tested in transfer tests, where the surface area was controlled for. This procedure examined whether the rays were still able to discriminate the previously learned quantity. In the 5 vs. 2 training, four out of five rays reached the LC in 15 ± 11 sessions. During the 3 vs. 6 training four out of five rays reached the LC in 8 ± 3 sessions. In the 4 vs. 7 training four out of five rays reached the LC in 7 ± 3 sessions and during the 7 vs. 2 training, one ray reached the LC in 13 sessions. While three out of four rays were able to the reach the LC in the 5 vs. 7 training in 7 ± 4 sessions, none of the rays was able to reach the LC to discriminate 7 vs. 6 and therefore no transfer tests were implemented. Pooled as a group within the transfer tests the rays chose the higher numerosity significantly more often when tasked to discriminate 6 vs. 3 (n = 4, p < 0.01), 7 vs. 4 (n = 3, p < 0.0001) and 7 vs. 2 (n = 1, p < 0.02) (Figure 7a). No significant choice for the higher quantity was recorded for 5 vs. 2 (n = 3, p = 0.052) and 7 vs. 5 (n = 1, p = 0.263) (Figure 7a).3.3.2. Group 4Two rays were successfully trained to discriminate 4 vs. 1 and one was successfully trained to discriminate 5 vs. 2 (Table 1). An exemplary graph showing the performance of one ray during the training and while the transfer tests were added to the training scheme can be found in Figure 8. As a group, the rays chose the higher numerosity significantly more often when tasked to discriminate 3 vs. 1 (n = 3, p < 0.0001), 3 vs. 2 (n = 3, p < 0.00002), 5 vs. 2 (n = 2, p < 0.0002), 4 vs. 3 (n = 3, p < 0.0002), 5 vs. 3 (n = 3, p < 0.00001), 7 vs. 4 (n = 3, p < 0.00001), 7 vs. 5 (n = 3, p < 0.003), 9 vs. 5 (n = 1, p < 0.05), 12 vs. 6 (n = 2, p < 0.0001), 15 vs. 5 (n = 2, p < 0.001) and 12 vs. 8 (n = 3, p < 0.04) (Figure 7b). No significant choice towards the higher numerosity was recorded for 5 vs. 4 (n = 2, p = 1), 7 vs. 6 (n = 1, p = 1) and 12 vs. 9 (n = 2, p = 0.6358) (Figure 7b). The 7 vs. 4 and 9 vs. 5 transfer test were additionally used to see whether the rays used a “relative” (higher or lower quantity) or “absolute” (“4” or “5” depending on what they were trained on) approach to discriminate quantities. In both tests, the rays choose the higher numerosity (“relative strategy”) significantly more often than the absolute number of items they were trained on. Furthermore, they were tested whether the size of the stimuli influenced their choice. Therefore, the negative stimuli were downscaled, with individual symbol size being smaller than the individual symbol size of the positive stimuli. The rays still showed a significant choice towards the higher quantity on which they have been trained on (4 vs. 1 or 5 vs. 2) (n = 3, p < 0.000001).4. DiscussionThe present study examined the ability of C. griseum and P. motoro to discriminate two different quantities of two-dimensional objects. The results of the experiments show that individuals of both species successfully solved the training tasks based on numerical information while continuous variables were controlled for. The results corroborate findings from several other publications examining this ability in fish over the last decade [35,60]. Previously, the ability to discriminate quantities had only been examined in Port Jackson sharks (Heterodontus portusjacksoni) [56]. Unfortunately, continuous variables were not controlled for. Nonetheless, similar to the present study it was shown that there was a high degree of individual variability in regard to learning performance and the development of side preferences in those sharks, that did not successfully perform during training. In the current study, this resulted in only a small number of individuals (7 out of 15 sharks, 7 out of 14 rays) reaching the stage where transfer tests were implemented. Especially shark group 2 experienced great difficulty to learn the task. Within all groups, side preferences (either side) were noticeable. In those cases where side preferences persisted and prevented animals from reaching the learning criterion, individuals were assumed to not have comprehended the task. However, having a side preference can generally be considered a reliable replacement strategy, as it enables an animal to secure 50% of the rewards. The low number of individuals advancing to the transfer test stage may not only reflect individual learning variation but could also indicated that numerical discrimination abilities may not be as fundamental to these species as to others [7,13]. Information on the ecology and the natural behavior of C. griseum and P. motoro is scarce and whether there is a need to discriminate quantities for mating, foraging or other behaviors is not known [61,62].The following discussion is divided into sharks and rays. As shark group 2 was not examined in transfer tests at all, the variety of transfer tests for C. griseum differed compared to the ones used for P. motoro.4.1. Grey Bamboo Shark (Chiloscyllium griseum)As expected, the two shark groups performed successfully in the Pretraining test, having shown consistenly high performances in previous cognition studies and in various setups [52,59]. However, no individual of shark group 2 reached the LC within their discrimination training sessions, and therefore no transfer tests were conducted. The following discussion is therefore based on the performance of group 1. Seven out of eight individuals were successfully trained on a 4 vs. 1 task. In the following transfer test, the effect of the only potentially confounding and uncontrolled variable during training was tested; i.e., thesize of the single negative stimulus was adjusted to the size of a single positive stimulus. All sharks still managed to choose the higher numerosity significantly more often. This shows that the sharks actively chose the higher numerosity during training instead of basing their choice on stimulus size. However, the low numerosity and “easier” ratio of 0.25 (4 vs. 1) might have diminished the possible effect of evading a larger (size-wise) negative stimulus. After animals were trained in 4:1, they proceeded to a 5:2 training. Here it showed that more than half of the sharks had problems reaching the LC. This result matches the observations by Vila Pouca et al. (2019) and underlines high individual differences [56]. Only three sharks managed to successfully reach the LC in the 5:2 task and maintained a high performance, so that transfer tests were conducted. First, a set of transfer tests was applied to evaluate the possible influence of shading and individual stimulus size, and the ability to transfer training skills to new shapes. Removing the coloration of the stimuli excluded the sharks’ option to make a choice based on the higher numerosity of shading gradients within the stimulus set. Additionally, the cumulative surface area of the positive stimulus was reduced to half of the negative one, and in the last transfer test, unfamiliar non-geometrical stimuli were presented. All sharks successfully chose the higher quantity in all transfer tests. This could be a seen as first evidence that C. griseum can rely on numerical information and is able to transfer and apply the previously learned information onto an alteredtask.The subsequent transfer tests only included a change in numerosity and ratio, the three sharks were tested on their ability to discriminate 5 vs. 3, 5 vs. 4, and 7 vs. 4. All ratios fall within the range of the hypothesized border area between the OFS and ANS system [15,20]. Sharks were able to successfully discriminate 5 vs. 3 and 7 vs. 4 but not 5 vs. 4. The results match the findings of other studies in fish, mammals, and amphibians [35,63,64]. OFS limits have shown to be at approximately 3 for the angelfish, goldbelly topminnow and redtail splitfin [65,66,67], 4 in primates (Macaca mulatta) and possibly 5 in guppies [1,68,69]. The present data could indicate that the OFS in C. griseum is limited to numbers of 4 or below. The 7 vs. 4 task is likely within the range of the ANS. Future studies could build upon the data from this study and investigate the limit of the two systems in this species further. The combined results show that C. griseum is able to use numerical information when solving quantity discrimination tasks, and its performance is comparable to those of other species. Whether numerical abilities are advantageous for C. griseum in the wild cannot be answered from artificial laboratory experiments and too little information is available about its natural behavior to speculate on this. In light of the astounding results of other cognition studies on the grey bamboo shark (e.g., [41,47,51,53]) and considering how few animals eventually proceeded to the transfer phase, it can be argued that, this ability is unlikely to be crucially important to the species.4.2. Ocellate River Stingray (Potamotrygon motoro)Just like the sharks, both ray groups successfully underwent Pretraining. Group 3 was subsequently trained in a series of consecutive discrimination tasks and then, following each training, tested in matching transfer tests. In contrast, the individuals of group 4 were only trained once (4 vs. 1 or 5 vs. 2) and then examined in many transfer tests using a wide range of different ratios. The transfer tests implemented for both groups varied to maximize the number of different ratios and quantities tested for this species. Individuals of group 3 were successfully trained to discriminate 6 vs. 3, 7 vs. 4, 7 vs. 2, 5 vs. 2, and 7 vs. 5 during the consecutive training procedures. None of the rays was able to discriminate 7 vs. 6. In the unrewarded transfer tests, rays discriminated 6 vs. 3, 7 vs. 4, and 7 vs. 2. This shows that this group of P. motoro was able to discriminate quantities up to a ratio of 0.57. Surprisingly, the rays failed to discriminate 5 vs. 2 (0.4) and expectedly failed to discriminate 7 vs. 5 (0.71). Nevertheless, while not significant, the three individuals tested for 5 vs. 2 showed a tendency towards choosing the correct quantity (p = 0.052). The findings correspond to those found for shark group 1 and the study by Villa Pouca et al. [56]. The results of group 4 are based on three individuals that were continuously trained in a 4 vs. 1 (n = 2) and 5 vs. 2 task (n = 1). Transfer tests were implemented during overtraining, i.e., after the LC was reached. Due to their high and reliable performance (Figure 7), 16 different transfer tests were conducted. While all continuous variables were controlled for, the rays managed to discriminate quantities up to a ratio of 0.75 in a lower numerosity (4 vs. 3) and 0.66 in a higher numerosity (12 vs. 8). Comparable studies in fish showed a similar ability when distinguishing higher and lower numerosities [39,40]. P. motoro has outperformed the results reported in infants [70] and angelfish [39] while showing similar performance compared to primates [1]. Only guppies were able to discriminate 4 vs. 5 after extensive training, a task the stingrays failed to achieve in the transfer test [69]. Another question, that transfer test results of this group answered, was whether rays used an ‘absolute’ or a ‘relative’ strategy [37]. It was examined whether the rays chose the higher numerosity based on choosing the “absolute” number ‘4’ (4 vs. 1 training) or ‘5’ (5 vs. 2 training) or if they simply choose the larger one of two quantities (relative). Transfer tests were conducted were the two strategies were placed into conflict, i.e., the higher training numbers were now the lower numbers of the two transfer test alternatives (7 vs. 4 and 9 vs. 5). All three stingrays used a “relative strategy” to distinguish between the different ratios provided.When analyzing the results of both stingray groups, individual differences seem to be playing an important role, just like in the shark groups. While individuals of group 3 could not discriminate 5 vs. 2 and 7 vs. 5, the three rays of group 4 could. Individual differences are present in many species and need to be accounted for when analyzing and comparing data [71,72]. Even the design of the experimental setup has shown to have an influence on numerical abilities in fish like goldbelly topminnows (Girardinus falcatus) and guppies (Poecilia reticulata) [66,73]. Within group 4, P. motoro was able to discriminate 3 vs. 1 and 3 vs. 2, that fall within the range of the OFS [15,20]. While the stingrays as well as the sharks, were able to discriminate 5 vs. 3 and 7 vs. 4, the rays failed to do so for 5 vs. 4 and 7 vs. 6. Again, this seems to indicate the transition from the OFS to the ANS, in which a difference of one item is not sufficient for distinction anymore. The range of the OFS found for the stingrays therefore appears to be one to four, which is similar to other animals tested [35,63,64]. Within the ANS range, the rays managed successfully to discriminate 9 vs. 5, 12 vs. 6 and 12 vs. 5 in addition to the previously mentioned 12 vs. 8 task. Increasing the ratio to 0.75 in a 12 vs. 9 task, resulted in none of the individuals performing successfully. This matches previous findings by Agrillo et al., where guppies, zebrafish, redtail splitfin, and Siamese fighting fish could discriminate 12 vs. 8 but not 12 vs. 9 [39]. The limits of the ANS and the range of the OFS for sharks should be investigated further, and additional studies on other shark and ray species may show how numercal abilities are generally distributed within the Chondrichthyes.5. ConclusionsThe data gathered within this study shows that the ability to discriminate quantities of two-dimensional objects based on numerical information alone is present not only in teleosts but also in elasmobranchs. Intraspecific differences and a considerably high number of unsuccessful individuals indicate though, that despite being able to ‘count’ (by pick the larger of two quantities within a particular range) numerical abilities may not be of particular importance to both species. However, individuals that passed the training stages maintained very high-performance levels and successfully showed impressive abilities to extrapolate learned knowledge to new tasks. The OFS is likely to range between 1 and 4 and both species preferred using a relative (choosing the larger quantity) over an absolute (e.g., always choosing ‘5’) strategy. Results confirm previous findings that elasmobranchs possess many of the same cognitive abilities and to a similar extent as other vertebrates.

animals : an open access journal from mdpi

10.3390/ani13081311

PMC10135347

Homocysteine is an organic compound that can be measured in the blood of humans and animals. High levels of homocysteine in human blood are associated with an increased risk of heart disease, diseases of blood vessels, formation of blood clots and brain damage. However, the role of homocysteine in the health and disease of domestic animals is poorly understood. This review critically appraises the literature concerning homocysteine in animals, focusing on horses. It aims to clearly define the existing knowledge gap to path an avenue for future research into homocysteine as a potential diagnostic marker of health and disease in this species.

Homocysteine is an endogenous, non-protein sulfuric amino acid, an intermediate metabolite formed by the methionine transmethylation reaction. Its elevated serum concentration in humans, hyperhomocysteinemia, is a sensitive indicator and a risk factor for coagulation disorders, cardiovascular diseases and dementia. However, the role of homocysteine in veterinary species has not been unequivocally established. Although some research has been conducted in dogs, cats, cattle and pigs, relatively few studies on homocysteine have been conducted in horses. So far, it has been established in this species that homocysteine has an atherogenic effect, plays a role in early embryo mortality and is responsible for the induction of oxidative stress. These preliminary findings support establishing a reference range in a normal population of horses, including horses in training and merit further investigations into the role of this amino acid in health and disease in this species.

1. IntroductionHomocysteine (Hcy) is an endogenous, non-proteinogenic, sulfur-containing amino acid formed as an intermediate metabolite in the process of intracellular transmethylation of an essential amino acid methionine to semi-essential cysteine (Figure 1) [1]. According to the latest hypothesis, homocysteine and its thiolactone could have been involved in protein synthesis at the origins of life on Earth [2]. Hcy is normally catabolized via the transsulfuration pathway to cysteine, but it can also be recycled back to methionine via (re)methylation [3,4]. The (re)methylation process takes place in the presence of a biologically active form of folic acid (vitamin B9), N5-methyl-tetrahydrofolate (N5-methyl-TH4) [5]. N5-methyl-TH4 is responsible for lowering Hcy concentrations, and its “sweeping” properties highlight the health benefit of dietary supplementation of folic acid [6,7,8,9,10]. A similar relationship exists with other nutrients. Methionine synthesis requires riboflavin (vitamin B2) and cobalamin (vitamin B12) as cofactors [11,12], whilst the transsulfuration of Hcy to cysteine requires pyridoxine (vitamin B6) [13]. Deficiencies of vitamins B2, B6 and B12 contribute to harmful hyperhomocysteinemia, and their supplementation reduces its effects [10,14,15,16,17,18,19,20,21,22,23].In humans, increased serum Hcy concentration, hyperhomocysteinemia, is a sensitive marker and a risk factor for coagulation disorders, cardiovascular diseases, especially of thrombotic aetiology, neurodegenerative diseases and dementia [24,25,26,27,28,29,30,31,32]. Due to the coexistence of those medical conditions and hyperhomocysteinemia, particularly in older people, Hcy has been named by the medical profession as the cholesterol of the 21st century [33]. Recently, homocysteine has been investigated during COVID-19, chronic kidney disease, oncogenesis and infertility, emphasizing an interest in this biomarker within the medical and scientific community [34,35,36,37,38,39].However, relatively little is known about the role of homocysteine in the health and disease of veterinary species. This review aims to critically appraise existing literature and clearly define the knowledge gap to pave the way for future research into serum homocysteine as a potential biomarker of health and disease in horses.2. MethodsThis paper is a narrative review. Research studies were ascertained by searching PubMed, The Web of Science, Google Scholar, and citation searching. Search terms included: homocysteine, AND horses OR equine OR pets OR dogs OR farm animals. The arbitrary start date for studies included in this review was not set. However, searches were completed in January 2023.3. Homocysteine in Pets and Farm AnimalsThe role of hyperhomocysteinemia has not been clearly understood in veterinary medicine. However, a statistically significant increase in serum concentrations had been demonstrated in dogs with heart disease, dogs with chronic enteropathy, dogs and cats with kidney disease, and dogs with hypothyroidism [40,41,42,43].During canine hypothyroidism, as in humans, the concentration of Hcy was negatively correlated with the concentration of total thyroxine (TT4) [42,44,45]. It is partially explained by the positive correlation between thyroid hormone levels and the activity of methylenetetrahydrofolate reductase (MTHFR), which is involved in the (re)methylation of Hcy to methionine [46,47,48,49]. This relationship indicates a likely role of thyroid gland disorders in developing homocysteine-dependent pathologies [44,50]. Therefore, any investigation of Hcy disturbances should ideally involve concurrent measurement of the thyroid hormone concentrations to assess their involvement.Hcy has been studied in farm animals, but the evaluation of its serum concentration has not gained clinical application. The observations in pregnant and lactating sows provide interesting but somewhat ambiguous conclusions [51]. In a study assessing the effect of folic acid and vitamin B12 supplementation on the parameters of growth and the immune status of their offspring, there was a positive correlation between the concentration of Hcy and the growth performance of piglets, but on the other hand, a negative correlation with some indicators of humoral and cellular immune responses [51]. Furthermore, it has been demonstrated that a methionine-rich diet was correlated with serum hyperhomocysteinemia and predisposed pigs to atherogenesis [52]. Hcy has also been investigated in cattle [53,54,55,56]. Its serum concentration was significantly higher than in healthy animals during bovine theileriosis [53], and in moderate, long-term cobalt deficiency [54].4. Homocysteine in HorsesThe role of Hcy in disease pathogenesis has also been studied in horses, but to a limited extent and only in some areas in this species (Table 1).4.1. Cardiovascular DiseaseThe role of Hcy in the etiology of cardiovascular disease in humans is well understood [64]. Hyperhomocysteinemia in humans has been associated with vascular inflammation and atherosclerosis [65,66,67,68,69,70,71,72]. Therefore, inhibition of vascular thromboresistance in hyperhomocysteinemia is of particular interest. Hcy increases the synthesis of thromboxane A2 (TxA2), activates coagulation factor V, inhibits the synthesis of anticoagulants at the DNA level, and suppresses the maturation of the endothelial matrix cells [73,74,75,76,77,78,79].Furthermore, Hcy negatively affects the regeneration of already damaged vascular endothelial cells because it strongly inhibits DNA and protein methylation [80,81,82,83]. For example, Hcy inhibits the methylation of p21 Ras protein (p21ras) and decreases the expression of the gene encoding this protein [84]. This leads to a reduction in cellular DNA synthesis and impaired tissue repair [81]. Hcy also has a negative effect on myocytes and myocardial contractility [85].The role of Hcy in horses in the course of laminitis is debatable. Although Hcy interacted with vascular endothelial cells in vitro, there was no association between hyperhomocysteinemia and the risk of laminitis [58].However, a relationship between hyperhomocysteinemia, impaired chorionic angiogenesis and early embryo mortality was demonstrated in mares [57]. Therefore, determining serum Hcy concentration may be useful in assessing the risk of the above-mentioned fertility impairment in this species.Because of its involvement in the proliferation of endothelial cells and vascular smooth muscle, Hcy affects the structure and function of the cardiovascular system. However, no relationship between hyperhomocysteinemia and cardiac dysfunction has yet been demonstrated in Equidae. In a study assessing horses with atrial fibrillation (AF), there were no differences in serum Hcy concentration between horses with AF (n = 55) and healthy animals (n = 27) [61]. Furthermore, there was no relationship between serum Hcy concentrations and the likelihood of AF recurrence after successful cardioversion [61].In another study assessing serum Hcy, cardiac (troponin I, creatinine kinase and d-dimer) and renal biomarkers (urea, creatinine and cystatin-C) in horses naturally infected with Theileria equi, all examined parameters, including cardiac ones significantly positively correlated with the magnitude of parasitemia [62]. This may indicate the possible involvement of Hcy in myocardial damage, although other causes, including oxidative stress and the systemic response to parasitemia, are also possible.Although based on small sample size, these observations emphasize a significant variation in the disease pathogenesis between horses and other species and highlight the differences in the potential utility of serum Hcy concentrations as a marker of cardiac disease in Equidae.As aforementioned, more research is needed to further assess Hcy as a risk factor for vascular disease in horses.4.2. Neurodegenerative DiseaseHyperhomocysteinemia in people leads to cognitive impairment and depression [86,87,88,89,90,91,92]. It is also a significant risk factor for Parkinson’s disease (PD) [93]. In a rat model of the disease, hyperhomocysteinemia reduces the number of dopaminergic neurons, likely by increasing their sensitivity to endogenous toxins [94,95]. In addition, Hcy contributes to neuronal degeneration by inducing oxidative stress, enhancing mitochondria dysfunction, DNA damage and apoptosis [96,97,98,99].Amongst other neurodegenerative diseases, Alzheimer’s disease (AD), the most common cause of dementia in older adults, merits special consideration [100,101]. AD is primarily characterized by the deposition of the β-amyloid peptide (Aβ) in the brain parenchyma and cerebral blood vessels [102,103]. Although not entirely clear, hyperhomocysteinemia has been implicated in the pathogenesis of AD and other types of dementia [30,104,105,106,107,108,109]. It has also been demonstrated in a large meta-analysis that every increase of serum Hcy concentration by 5 µmol/L increases the risk of Alzheimer’s disease by as much as 12% [105]. It is fascinating that apart from being involved in vascular changes, Hcy may also play a role in the development of AD by antagonizing gamma-aminobutyric acid (GABA) receptors and acting as a neurotransmitter competing with GABA [110,111,112]. In addition, Hcy has been implicated in the pathogenesis of other neurological and psychiatric diseases in humans, including autism, epilepsy, depression, bipolar disorder and schizophrenia [31,113,114,115,116].There is no data on the role of Hcy in equine neurological diseases. However, the involvement of Hcy in Parkinson’s disease (PD), which is pathogenetically similar to equine pituitary pars intermedia dysfunction (PPID) [117], provides an exciting area for future research into the role of Hcy in this disease entity in horses. Other potential research avenues include the investigation of the role of Hcy in equine degenerative myeloencephalopathy (EDM) and equine motor neuron disease (EMND) [118,119].4.3. Physical ActivityHorses are exceptionally well-adapted athletes [120]. This adaptative behaviour has been crucial in the species’ survival in the face of the threat from predators [121]. During exercise, a horse’s heart rate increases by more than eight times concerning resting conditions [122]. Additional adaptative manifestations to increased physical load include an increased ratio of lung capacity to body weight, the ability to double the number of peripheral blood erythrocytes by splenic contraction in the initial phase of physical effort, and even adaptation to hypercapnia under extreme exercise load [122,123,124].Nevertheless, despite the adaptation mechanisms, sports horses are subjected to adverse effects of physical exertion. Apart from injuries, myopathies, post-exercise pulmonary haemorrhage and other clinical conditions, strenuous exercise and increased oxygen consumption generate oxidative stress, and free radical formation is responsible for numerous metabolic disorders [125,126,127,128]. Furthermore, in humans, strenuous exercise has been linked to cardiac injury by the induction of oxidative stress and systemic inflammatory response [129].It has been demonstrated that acute exercise increases Hcy levels in humans, and the magnitude of those elevations depends on prior training preparation [130,131]. Although the basal Hcy serum concentration is lower in people with regular physical activity, up to 47% of athletes develop hyperhomocysteinemia after exercise [132,133]. It raises concerns for amateur runners participating in triathlons, marathons or ultramarathons. Prolonged post-exercise oxidative stress and hyperhomocysteinemia may lead to vascular endothelial damage in these individuals [131].In one of the few studies evaluating the relationship between Hcy and strenuous exercise in horses, animals were subjected to jumping over ten obstacles with a maximum height of 140 cm over a distance of 350 m [60]. The serum Hcy concentrations were obtained before, 30 and 60 min after the training, with a tendency for increased Hcy concentration documented immediately after training [60].The influence of strenuous exercise on the levels of antioxidants and vitamins is significant as an increased breakdown of glycogen and an increased demand for B6 post-training is well described [134]. Interestingly, post-exercise increase in serum Hcy levels has been detected in humans with appropriate concentrations of folic acid and vitamin B6 [131]. Hyperhomocysteinemia in those subjects has been explained by natural post-exercise changes in glomerular filtration and plasma clearance of Hcy. None of those relationships has been studied in horses.Physical exercise can cause various biochemical changes that may affect horses’ oxidative stress-dependent Hcy metabolic pathways. It has therefore been suggested to study the physiological range of serum Hcy and its changes depending on the frequency, intensity and duration of training [60].4.4. Oxidative StressHcy is an important inducer of prooxidative-antioxidative imbalance associated with increased intracellular calcium ion concentrations and DNA damage [98,135]. Hcy contributes to oxidative stress by reducing glutathione peroxidase activity and lowering the level of vitamins A, E and C [136]. By reducing the de novo synthesis of glutathione, Hcy also leads to the impairment of redox mechanisms [137]. Stimulation of the synthesis of reactive oxygen species is one of the most important links between Hcy, particularly its oxidized sulfhydryl groups, and the development of atherosclerosis [138]. Atherosclerosis is also favoured by Hcy-induced hyperuricemia, which further promotes the formation of reactive oxygen species and subsequent vascular endothelium dysfunction [85].Oxidative stress contributes to increased uptake and use of methionine, which is subsequently converted to Hcy. It has been demonstrated that the maximal physical activity of racehorses over 1200, 1600 and 2000 m causes an increase in reactive oxygen metabolites (dROMs) and serum Hcy concentration. The mentioned indices remained positively correlated with each other also after strenuous physical activity [63].In another study examining the relationship between Hcy and oxidative stress in racing horses, there was a high positive correlation between the serum concentration of Hcy and dROMs. The Hcy concentration increased immediately after exercise, returned to resting values 30 min later, and remained stable until the end of observation (i.e., 180 min after the race). The concentration of dROMs initially increased immediately after exercise, then decreased. However, it remained higher concerning the resting value for the remainder of the observation period. Post-exercise increases and decreases in serum Hcy concentration were also positively correlated with the antioxidative barrier (Oxy-Adsorbent), a compensatory response to oxidative stress [59].The effects of oxidative stress on serum biomarkers have already been identified in racehorses, where an elevation in gamma-glutamyl transferase (GGT) associated with training has been documented [139]. A relationship between poor performance and elevated gamma-glutamyl transferase (GGT) beyond this anticipated increase has also been identified. Although the exact mechanism responsible for the increase in GGT is not fully understood, a poor response to oxidative stress (overtraining) has been proposed [139]. A recent investigation to determine the potential role of infectious agents with elevated GGT in a population of racehorses also supports the role of oxidative stress and, in addition, identified a decrease in vitamin B6 in the horses with exercise-associated elevation in GGT [140]. These data further support the investigation of Hcy in racehorses, particularly in a subset of horses with poor performance and elevated GGT.Those findings suggest that sports horses may be at risk of certain Hcy-related diseases and, at the same time, provide the rationale for serum Hcy as a potential marker of redox disorders in this species.4.5. Reference Values and HyperhomocysteinemiaHyperhomocysteinemia in humans is defined as serum Hcy concentration exceeding 15 µmol/L, while an increase of 2.5 µmol/L above this value increases the risk of cardiovascular disease by 10% [141,142].Currently, it is difficult to clearly define hyperhomocysteinemia in horses due to a lack of established and standardized reference ranges for serum Hcy and a limited number of studies on the role of this amino acid in disease pathogenesis. However, under physiological conditions in horses, elevated post-training serum Hcy levels are usually short-lived, lasting for around 60 min post-exercise [60]. Furthermore, some evidence suggests that the length of post-exercise serum Hcy elevation in horses is directly proportional to the length of prior strenuous exercise [131].To the best of our knowledge, only three studies assessed normal serum Hcy concentration in horses. A study in ponies reported normal serum Hcy concentration in the range of 1.3–14.7 µmol/L [58]. Another small research estimated average resting serum Hcy at 6.16 µmol/L (SD 0.36), with interindividual variations likely caused by differences in husbandry factors [60]. In a research of a more diverse group of 27 healthy horses, the serum Hcy reference range was established at 1.5–7.8 µmol/L with an average value of 4.65 µmol/L [61]. More research is needed to reliably show the physiological range of serum Hcy and its normal variation in horses.4.6. Determination MethodsSerum Hcy concentration has been successfully and reproducibly determined in dogs by chemiluminescent immunoassay incorporated in the ADVIA Centaur XP automated system (Siemens, Munich, Germany) [42]. The correlation between the results obtained by this method and the results obtained by high-performance liquid chromatography (HPLC) is almost complete (r = 0.96, p = 0.0001) [143]. Similarly, an automated enzymatic method (Homocysteine Cobas C INTEGRA 800 assay, Roche Diagnostics International Ltd., Rotkreuz, Switzerland) has been reliably used to determine serum Hcy in horses [61]. Unlike previously utilised HPLC, both methods are rapid and easily available.5. ConclusionsThe involvement of Hcy in numerous pathogenetic processes in humans and animals provides a solid ground for further research in horses. Investigations into the role of Hcy in various diseases and physiological conditions, including physical exertion of varying intensity, are warranted. Most importantly, however, developing a standardized, unified and reliable reference range of serum Hcy in healthy horses is crucial. After meeting these criteria, serum Hcy will likely serve as a clinically useful surrogate marker for evaluating redox disorders in this species.

animals : an open access journal from mdpi

10.3390/ani12030318

PMC8833503

It is known that the bioactive compounds (N-benzyl-palmitamide, benzyl isothiocyanate, glucosinolates and phenolics) in the maca plant are appetizing, as well as having antioxidant effects and improving reproductive ability. The aim of this study was to determine the effects of adding maca plant powder to the ration at different levels on growth, slaughter carcass, partial egg production and some reproductive characteristics of Japanese quail (Coturnix coturnix japonica). Addition of maca powder to the diet increased the feed consumption but did not affect growth, slaughter carcass, partial egg production or fertility. In addition, maca powder reduced embryonic deaths and improved chick quality. According to these results, the positive effects of short-term application of maca powder are not observed during the fattening period in Japanese quails.

Maca plant contains rich nutrients and in addition, it has various bioactive substances (N-benzyl-palmitamide, benzyl isothiocyanate, glucosinolates and phenolics). It is used to improve reproductive properties and has antioxidant effects for both humans and animals. The aim of this study was to determine the effects of adding maca plant powder to the ration at different levels on growth, slaughter carcass, partial egg production and some reproductive characteristics in Japanese quail (Coturnix coturnix japonica). The experimental groups were formed by adding 0% (control), 0.05% and 0.1% maca powder to the diet, and a total of 300 birds were used. Growth (weekly body weights, parameters of Gompertz growth function), feed efficiency and carcass characteristics of quails in the 42-day fattening trial were determined. Reproductive characteristics were measured up to 22 weeks of age. Addition of maca powder to the diet increased the feed consumption (p < 0.05) but did not affect body weights at 35 and 42 days of age, β0 and β1 parameters or point of inflection weight of the Gompertz model, carcass traits, partial egg production or fertility. It may be advisable to add 0.1% maca powder to the diets of breeders. Besides, maca powder reduced embryonic deaths and improved chick quality (both p < 0.05). It is thought that different results for reproductive traits can be obtained if maca powder is used for a longer period in the diets of breeder quail flocks.

1. IntroductionAntibiotic (tetracyclines, avoparcin, virginiamycin, tylosine, spiramycin, etc.) use, which aims to increase the digestion level by controlling the intestinal microbiota in chicks, is a common practice. Thus, feed efficiency is improved, and more live weight gain can be achieved. However, this practice, which is considered risky for public health, has been banned in European Union countries since 2006. For this reason, new natural or biotechnological feed additives are being researched in order to create the effect of antibiotics in poultry. Plants have been used for therapeutic purposes since prehistoric times. In the structure of each plant, there are many important bioactive compounds such as alkaloids, flavenoids, glycosides, mucilages, saponins, tannins, and phenol, phenolic acids, coumarin, terpenes, essential oils, lectins and polypeptides [1]. These chemical compounds in plants show antiviral and antibacterial effects in other living things that consume plants and cause many effects that strengthen the immune system. Bioactive compounds such as N-benzyl-palmitamide, benzylisothiocyanate, glucosinolates and phenolics in maca have important effects on lipid, mineral and antioxidant metabolisms [2,3,4,5]. Many plants and their extracts as feed additives in animal nutrition have been used for many years for growth, reproduction, improvement of product quality and health protection.Maca (Lepidium meyenii) is a medicinal plant with high nutritional value (12.8% protein in dry matter content, 23.5% fibrous substance) from the Brassicaceae family and is a plant from the Andes of Peru. Maca plant contains important fatty acids, macacids, macaridin, alkaloids and glucosinolates [6]. In addition, maca plant has an aphrodisiac compound called p-methoxybenzyl isothiocyanate [7]. Researchers have focused on the effects of using a powder or extract of maca plant as a food additive for humans on some reproductive traits in men and women [8,9]. In addition, it was reported that bone mineralization against osteoporosis and antioxidant activity against stress were improved in rats consuming extracts obtained from the maca plant [2,10]. There are limited studies on the use of maca as a feed additive in livestock diets. In the study conducted by Clément [11], it was determined that maca powder given to breeding cows as a feed additive increased their sperm quality and number. A similar study was carried out in sheep, and it was determined that maca powder given to breeders as a feed additive increased the number of copulation and ejaculation [12]. Bilal et al. [13] determined that the addition of 50 and 75 g/day of maca powder to racehorse rations had no effect on body weight and average daily feed consumption characteristics. In a study by El-Sheikh et al. [14], the addition of maca to the diet did not have a significant effect on the live weight, daily feed consumption, litter size, litter weight or milk production of rabbits. To our knowledge, the only study examining the effects of adding maca to poultry diets was conducted by Korkmaz et al. [15]. The researchers investigated the effects of maca powder supplementation on the performance, egg quality, serum parameters, hormones and antioxidant enzyme levels of laying hens in the post-peak period. In their study, maca powder had neither positive nor negative effects on performance, egg quality, egg yolk cholesterol content, serum parameters (excluding magnesium) or hormones. The aim of this study was to determine the effects of adding maca plant powder to the ration at different levels on growth, slaughter carcass, egg production and some reproductive characteristics of Japanese quail (Coturnix coturnix japonica). Among avian species, Japanese quail has been used in many studies related to early sexual development and maturation [16,17]), as well as for the high production of the eggs and their regular deposition [18].2. Materials and Methods2.1. Animal MaterialThe study was carried out in poultry facilities of the Faculty of Veterinary Medicine and the Faculty of Agriculture of Namık Kemal University, based on the decision of Namık Kemal University Animal Experiments Local Ethics Committee, dated 18 October 2017 and numbered 78. The animal material of the study consisted of a total of 300 quail chicks, which were randomly mated and obtained simultaneously from a parent flock that had not been genetically selected before.2.2. Rearing PeriodOne-day-old chicks were randomly assigned to each experimental group (100 birds per group), and wing numbers were assigned immediately after hatching, and individual weights were recorded. Weekly live-weight measurements were repeated individually thanks to the wing numbers. The chicks, which were randomly allocated to each experimental group, were fed in brooder cages (battery type with 5 floors, 90 cm2/quail) until sex determination on the 21st day after hatching. The chicks were housed at 32 °C for the first three days, and this was lowered by 1 °C every three days, and the temperature was adjusted to 27 °C at the end of the second week. After the 21st day, the quails were housed as a group in the rearing cages (battery type with 5 floors, 160 cm2/quail). A constant light intensity (60 lux) was used throughout the study. A lighting program of 23 h of light and 1 h of darkness was applied to all quails during the first six weeks.Experimental groups were formed by adding 0% (control; C), 0.05% (M1) and 0.1% (M2) maca powder to the diet. Three different rations were used throughout the experiment, and these are presented in Table 1 according to the periods in which they were used. In the study, feed consumption was determined as a group in brooder cages (from hatching to 3 weeks of age). Then, it was determined individually in the rearing period (between 3 and 6 weeks of age). The Gompertz equation was fitted to the growth data of Japanese quails to model the relationship between body weight and age [19]. The model expression of the Gompertz function and their coordinates of the point of inflection are presented in Table 2.In the equation, “t” denotes time, “Y” weight, “β0” the maximum body weight the animal is assumed to be able to reach, “β1” the biological constant of the shape of the curve, “β2” the biological of about the growth rate and “β3” the shape parameter [20]. Model parameters were analyzed using with SAS 9.3 software NLIN procedure Levenberg–Marquardt iteration method [21,22]. Fifty randomly selected quails from each experimental group and a total of 150 quails (71 females and 79 males) were sent to slaughter at the age of 42 days. Feed was removed for 4 h before slaughter, and slaughter weights of quails were determined. All weight measurements during cutting were carried out with a digital scale with 0.01 kg precision. Following slaughter, wet plucking and evisceration, hot-carcass weights were determined, including neck and belly fat, excluding edible internal organs. At this stage, edible visceral weights consisting of abdominal fat, heart, liver and empty gizzard were determined [23]. After the carcasses were kept at +4 °C for one day, the cold carcass weight was measured, and the carcasses were shredded, and the breast weight, chest muscle weight, thigh and wing weights were measured. By dividing the cold carcass, edible internal organs, abdominal fat, breast, thigh and wing weights to the slaughter weight, phenotypic values were obtained for cold carcass ratio, edible internal organs ratio, abdominal fat ratio, breast ratio, breast muscle ratio, thigh ratio and wing ratio, respectively [5,24].2.3. Reproductive PeriodOf the quails that were not selected for slaughter, 27 females and 9 males (1:3 mating ratio) from each group were transferred to individual breeding cages to determine egg production and some reproductive characteristics (fertility, embryonic mortalities, chick quality). We aimed to ensure high fertility by placing the male quail of each family in the cage of only one of the three females every day [25]. Hen-housed day egg yields were individually kept up to 22 weeks of age. When the quails were 10, 14, 18 and 22 weeks old, the eggs collected for 5 days were put into the incubator daily, and the hatching results were evaluated. For this purpose, fertility, total embryonic death, early embryonic death and late embryonic death were determined by macroscopic examination. All hatched chicks were examined by experienced operators to determine the Tona chick quality score of chicks as previously described by Tona et al. [26]. The Tona scoring method is a qualitative scoring system that assesses a total score index of 100, based on a wide variety of visual parameters, such as activity, appearance, retracted yolk, eye condition, leg and feet condition, navel deformities, and status, remaining egg membrane, beak condition and remaining yolk [27].2.4. StatisticsTo determine the differences between the means of the groups in terms of all the characteristics obtained, analysis of variance was applied. Duncan multiple range test was applied when the H1 hypothesis was accepted. The significance level was accepted as 0.05 in all statistical analyses. All statistical analyses were performed using SAS 9.3 statistical software.3. ResultsThe average values of some performance characteristics of female and male quails at 35 and 42 days of age and the results of the variance analysis are given in Table 3. As can be seen from Table 3, the absence of a statistical difference between hatching weights of quails indicates that the subjects were randomly assigned to the experimental groups. It was determined that the quails in the M1 and M2 groups did not have different averages from birds in the control group in terms of body weight and feed conversion ratio at 5 and 6 weeks of age (p > 0.05 for both traits and weeks). However, it was determined that adding maca powder to the diet increased feed consumption linearly (p < 0.05). In the study, statistical differences were determined between the sexes in terms of mean values of body weight and the feed conversion ratio. While female quails had higher body weights than males, they also had a better feed conversion ratio. The treatment–sex interaction effect was not found to be significant in terms of body weight, feed consumption or feed conversion ratio (p > 0.05 for all).The results of the Gompertz growth curve analyses performed using the weekly live-weight data of female and male quails in the experimental groups are given in Table 4. In addition, the growth curves drawn according to the experimental groups and genders are also presented in Figure 1. The coefficients of determination in all nonlinear regression analyses were found to be between 0.996 and 0.999 (not included in any table). In this case, it was determined that the Gompertz growth curve model was quite sufficient to explain the quail data. The mean values of the mature (asymptotic)-weight parameter (β0) of the Gompertz growth model were estimated as 241.57, 258.63 and 260.23 g for the C, M1 and M2 groups, respectively (Table 4). The mean β0 value of the control group was not statistically different from those in the M1 and M2 groups (p > 0.05). At the same time, the mean value of the β0 parameter of female quails was higher than that of males (p < 0.05). There was no difference between the experimental groups (p > 0.05) and between the sexes (p < 0.05) for the β1 parameter, which is defined as the integration constant or maturation rate. The highest mean value (3.90) for the β1 parameter was found in females in the M2 group, while the lowest mean value (3.56) was determined in males in the control group. There was a significant difference between the experimental groups in terms of growth rate parameter (β2), and the β2 values of quails given maca powder were lower than those of group C (p < 0.05). It can be seen that quails in the control group reached the point of inflection of the Gompertz growth model at an earlier age (p < 0.05), but there was no difference between the experimental groups in terms of the weight of the point of inflection. Contrary to this situation, there was no difference between the time of the point of inflection of female and male quails, while females reaching the point of inflection were found to be heavier (p < 0.05). The inflection point age of the Gompertz model was between 16.61 and 19.79 days in all groups. The highest point of inflection weight average was found in the female quails of the M2 group (101.68 g), while the lowest mean value of point of inflection weight was found in the control group male quails (83.59 g).The mean values determined for carcass yield and breast, breast meat, leg, wing and abdominal fat ratios (% body weight) of quails slaughtered at the age of six weeks in the control, M1 and M2 experimental groups, as well as variance analysis results, are presented in Table 5. The addition of maca powder to the ration did not affect any carcass characteristics. Similarly, no difference in carcass characteristics was observed between males and females. Carcass yields of quails varied between 67.46 and 71.74%.The mean values of egg yield, fertility, early, late and total embryonic mortality and the analysis of variance are presented in Table 6. There was no statistical difference between the experimental groups in terms of egg yield and fertility. However, the mean values of the groups fed with maca powder were found to be lower in terms of embryonic deaths. While the mortality rate of quails in the control group was 11.26%, embryonic mortality in the M1 and M2 groups was 8.98% and 4.89%, respectively (p < 0.05). A similar situation occurred in terms of chick quality, the Tona score of quails fed with 0.1% maca powder added to the ration had the highest mean value (p < 0.05). The mean values of Tona chick quality scores in the C, M1 and M2 experimental groups were determined as 90.67, 89.44 and 95.42.4. DiscussionIn the study, it was determined that the addition of 0.05% and 0.1% maca powder to the diet had no effect on body weight and feed efficiency at 35 and 42 days of age but negatively affected the amount of feed consumption (p < 0.05). It was determined that the quails in the M2 group consumed higher amounts of feed (594.78 g and 781.93 g) at 35 and 42 days of age compared to the other groups. It has been determined that the effect of adding maca powder to the feed consumption increases linearly with age. Maca powder is thought to have appetizing properties, and it can be seen that the addition of 0.05% and 0.1% maca powder, especially at the age of six weeks, significantly increased the amount of feed consumption compared to the control group. There is only one study on the addition of maca powder to the diets of poultry. In the mentioned study (Korkmaz et al., 2016) it was determined that the addition of 5 and 10 g/kg maca powder to the rations of laying hens at 56 and 72 weeks of age did not affect the live weight and feed efficiency. The results of this study are consistent with the results reported by Korkmaz et al. [15]. In another study conducted in ruminants, it was reported that the addition of maca powder to the ration did not affect the feed intake, growth rate or carcass performance of bulls [28]. In the study performed by Uchiyama et al. [29], it was determined that the growth rates and feed intakes of rats fed with diets with three different doses of maca powder for seven weeks were not affected by maca supplementation.It was determined that the gender effect influenced the body weight and feed efficiency characteristics of quails (p < 0.05). According to the body weights at 35 and 42 days of age, higher values were found for females (187.99 g and 211.40 g, respectively) than males (177.00 g and 199.51 g). Although female and male quails consumed similar amounts of feed, it was determined that females with a higher body weight also had better feed conversion ratios (Table 2). There is reverse dimorphism between males and females in Japanese quail compared to other poultry species. In many studies, it has been reported that weekly live-weight values and feed consumption characteristics of female quails are higher than males [30].According to our knowledge, there is no study on the effect of adding maca powder to poultry rations on growth curve parameters. There are few studies on the effects of feeding poultry with different rations on growth curves [20]. Lilburn et al. [31], who raised two different turkey genotypes (Nicholas and British United Turkeys) using two lighting programs and two feeding programs, compared the growth patterns of the experimental groups with the Gompertz model. There was no significant dietary effect on any growth curve parameter in their study [31]. Taroco et al. [32] investigated the genotype–environment interaction effects of adding five different levels of threonine:lysine to Japanese quail diets in terms of Gompertz model parameters. Researchers reported that this application had significant effects on the heritability and phenotypic values of the asymptotic weight parameter and inflection point age characteristics of the Gompertz model. Hashiguchi and Yamamoto [33], who examined the growth of Japanese quails fed with diets containing different ratios of protein with the Gompertz model, reported that this practice affected the growth curve parameters. The mature-weight parameter values (140.3–156.4) estimated by the researchers were found to be considerably lower than the averages determined in this study. It is thought that the reason for this situation is due to the low weekly live weight values of the quails used in the study by Hashiguchi and Yamamoto [33]. There are quite a few differences between the live weight values of Japanese quails due to the domestication, adaptation to cage conditions and genetic selection studies. While body weight values at the age of six weeks have been reported as 100–130 g in some studies [34,35], these averages have been reported to be in the range of 250–300 g in other studies [36,37]. In studies in which the growth of Japanese quails was examined with the Gompertz model, the mature-weight parameter was found in the range of 224–295 g [38,39]. In the study, β0 parameter averages (241.57–260.23 g) obtained from all three experimental groups were found to be compatible with mature-weight parameter values reported in the literature. In the study, the integration coefficient parameter (β1) of the Gompertz growth curve model for quail growth samples was estimated in the range of 3.60–3.83. The results obtained were found to be compatible with the estimated values (3.40–3.89) for randomly mating herds that were not selected by many researchers [39,40,41]. In the study, it was determined that adding maca to the diet decreased the instantaneous growth rate parameter; therefore, the birds reached the inflection point of the sigmoid growth curve in the later period. It is thought that this situation is caused by the effect of increasing feed consumption. Small values for β2 indicate late maturation and a high adult weight. On the other hand, high β2 values represent early maturation and a lower adult weight [40]. It is thought that if the maca dose in the ration is increased, the mature-weight parameter will also be positively affected by this situation. In all experimental groups in the study, the inflection point age of the Gompertz model was between 17.08 and 19.01 days, and the mean weight of the inflection point was between 88.87 and 95.73 g. According to the results of many studies in which the growth samples of Japanese quails were analyzed with the Gompertz function, it was reported that the values obtained for the inflection point age of the curve were between 14.76 and 24.62 days of age, and the weight of the growth curve inflection point was between 76.22 and 124.56 g [19,38,39,42,43,44]. The values of inflection point age and weight determined for the Japanese quails included in this study were found to be consistent with the averages reported in these studies.It was determined that the addition of maca powder to quail rations did not cause any difference in terms of yield of carcass, breast, breast meat, thigh, wing or abdominal fat (Table 5). The maca powder used in the study did not affect the weekly live weight or slaughter weight and did not affect the slaughter or carcass characteristics as expected. Ginseng, ginger and licorice root are medicinal plants that contain phenolic and bioactive components similar to maca. These plants have also been used as additives in poultry feeds for years. Reda et al. [45] investigated the effects of 250, 500, 750 and 1000 mg/kg of licorice root supplementation on the performance of Japanese quails. In parallel with our results, researchers reported that the doses applied did not have any effect on the carcass. On the other hand, it was stated that abdominal fat decreased in broiler chickens when 0.3 g/L of licorice root was added to drinking water and 2 g/kg of licorice root was included in feed [46]. Azazi et al. [47] reported that adding different levels of ginseng to rations of a layer breeder flock improved semen quality, fertility and hatchability. Researchers reported that adding ginseng to the diet did not affect percentages of edible inner organs, but surprisingly increased carcass yield. It is thought that the reasons for this difference are that the animal material used is old and that ginseng has been applied for a long time. In accordance with the results of this study, many researchers [48,49,50,51,52] reported that the carcass yields of Japanese quail were between 68.33 and 73.00%. Having obtained similar results, Walite et al. [53] reported that the mean values of the breast, leg and wing were 29.1%, 15.9% and 11.8%, respectively. Akbarnejad et al. [54] found lower averages than the mean values determined in this study, and the yield of cold carcass, breast and leg were found to be 65.0%, 24.7% and 15.1%, respectively. Egg production, fertility, early, late and total embryonic death rates and mean analysis of variance are given in Table 6. While there was no statistical difference between the experimental groups in terms of egg production and fertility, the average values of the groups fed with maca powder were found to be lower in terms of embryonic deaths. The alkaloids present in the maca plant work as the main stimulant of the ovarian follicles. It also has the effect of eliminating free radicals and has an antioxidant function; antioxidants protect the body against their harmful effects by neutralizing free radicals that are released as by-products during cellular metabolism and energy production in mitochondria. Antioxidants may be a mechanism of increasing fertility by enhancing LH levels [55]. It is a known fact that the use of plant-derived antioxidants improves reproductive traits such as quality of the sperm, semen, oocyte and embryo. It was determined that the addition of 5 and 10 g/kg maca to the laying hens’ rations did not have any positive or negative effect on the egg production performance [15]. The results of the study performed by Korkmaz et al. [15] are consistent with our findings. However, Osfor [56] reported that adding 2 and 4 mg/kg of ginseng plant, which has similar antioxidant effects to maca, to quail rations increases egg production and weight and improves feed efficiency. Similarly, Jang et al. [57] reported that the addition of fermented ginseng by-product to laying hens’ rations increased egg weight and yield. Al-Kassie [58] stated that cumin used at different levels in broiler rations causes significant differences in mortality rates; the lowest mortality rate was determined at the level of 0.5% and 1 cumin supplementation (4.1–3.4%) to the ration. However, Ebrahimi et al. [59] reported that adding medicinal and aromatic plants to broiler rations did not affect mortality rates. It was concluded that the addition of some plant seeds and root extracts with antioxidant effects to the rations of breeder turkey hens improves fertility, hatchability and embryonic viability [60,61].5. ConclusionsMaca plant had no effect on body weight or carcass characteristics of quails. The most important positive effects in the study are related to total embryonic death and chick quality characteristics. It may be advisable to add 0.1% maca powder to the diets of breeder flocks. It is claimed that the use of plant-based antioxidants in livestock has a positive effect on reproductive characteristics, but compatible results were not obtained in our study. This study was carried out during a short interval of the laying period. For this reason, it is thought that different results for reproductive traits can be obtained if maca powder is used for a longer period in the diets of breeder quail flocks. These results show that the effects of maca plant on reproductive traits in Japanese quail should be investigated in future studies.

animals : an open access journal from mdpi

10.3390/ani11030716

PMC8001850

The unique digestive properties of rabbits consist of highly specialised communities of intestinal microbes that, unfortunately, make them susceptible to metabolic diseases. This is why breeders, to improve the functions of the digestive tract, often use special feed additives, i.e., probiotics, prebiotics or synbiotics. The need to become independent from soybean meal (SBM), which is currently the basic source of protein in animal nutrition, and the need to stimulate the gastrointestinal tract (GIT), has increased interest in fermented components that have a positive effect on the intestinal microbiota and are a source of valuable protein. In this study, the impact of the diversified proportion of fermented rapeseed meal (FRSM) in the diet of rabbits on the immune parameters and the microbiota of the digestive tract was assessed. The reducing effect of the tested feed component against coliform bacteria and Escherichia coli within the small intestine and colon of animals and the anaerobic biota of Clostridium perfringens in the duodenum and cecum of animals was observed while in the duodenum—an increase in the beneficial biota of lactic acid bacteria. The conducted analysis also showed many complex correlations between the number of intestinal microbiota groups and the level of immunoglobulins. The results of the conducted research indicate that FRSM, in addition to valuable nutritional values, may play an important probiotic role in the GIT of rabbits. Research of this type is especially important in terms of reducing the use of antibiotics for therapeutic purposes through nutritional prevention of animals.

The present study was conducted to determine the effect of the use of varying amounts of fermented rapeseed meal in diets for rabbits on the immune status and microbiota of segments of the GIT. Forty 35 day old rabbits used in the experiment were assigned to four groups: the control group (group C) were fed a standard diet and the experimental received 4%, 8% or 12% fermented rapeseed meal (included in place of standard soybean meal). Class A, G and M immunoglobulins were determined in the blood plasma. In the food content collected after slaughter, microbiological parameters were determined for individual sections of the digestive tract. Rabbits from the groups receiving a diet with an increased proportion of fermented rapeseed meal (8% or 12%) had lower concentrations of anaerobic bacteria and Escherichia coli in the intestinal contents. Research has shown that the increase in intake of fermented rapeseed meal was correlated with an increase in the correlations between the immunoglobulin level and the size of the microbial population in the GIT. In light of the presented results fermented rapeseed meal, by supplying valuable bioactive substances, appears to be a good component in the diet of rabbits, enhancing immune system development and helping to prevent disturbances of the gut microbiota.

1. IntroductionThe complex microbiota of the gastrointestinal tract (GIT) of animals plays an important role in the digestion of nutrients and protection against infections resulting from the presence of pathogens and environmental bacteria. It also functions as a barrier against harmful exogenous substances and ensures normal metabolic, immune and neurological functions in the host. The GIT is involved in numerous physiological processes, from nutrition to behavioural and stress responses [1]. The unique digestive properties of rabbits and specialised microbial communities can help them to adapt to fibre-rich foods but often make them susceptible to metabolic diseases. For this reason, control and modulation of the microbiota of rabbits is an important aspect of breeding practice. Digestive problems can be controlled through the immunostimulatory activity of the microbiota and competitive exclusion, which is particularly important in young animals after weaning.As in the case of all mammals, the introduction of new species to the GIT of rabbits is determined by ecological succession. Microorganisms, competing for nutrients, colonise an ecological niche and consume all food resources. These bacteria are able to inhibit the growth of competing bacteria by producing antimicrobial substances [2]. A community of intestinal microbes is formed by random colonisation from the surrounding meta-community (the mother, bed, cage, air, etc.) and is highly variable between individuals up to the age of 49 days [3]. By the age of 70 days, the composition of the caecal microbiota is highly homogeneous and shows a certain degree of stabilisation. In rabbits, a disturbance of this normal microbiota, known as dysbiosis, is widely considered to be the cause of enteritis with symptoms of diarrhoea, followed by dehydration and potentially death. The aetiopathogenesis of intestinal inflammation in rabbits is complex, and there are generally multiple factors. Enteric pathogens, such as Escherichia coli, Clostridium spiroforme, Lawsonia intracellularis, Clostridium piliforme, Salmonella spp., rotaviruses, coronaviruses, parvoviruses and astroviruses, are most often found in individuals with diarrhoea. However, there is still little information on the factors that initiate it, as infection with these microbes is not synonymous with disease. Most cases of enteritis in rabbits are caused by a combination of multiple environmental factors and infectious agents, including a low-fibre diet, an overall weakened state of health, stress associated with management and inadequate welfare, and age, as well as the presence of one or more potentially pathogenic microbes [2,4].During microbiological fermentation in favourable intestinal conditions, rabbits receive products that stimulate colonization by symbiotic microbes. In rabbits, this is referred to as a combined model of competition and cooperation of the gut microbiota. However, the balance of this ecosystem is fragile and can be destroyed during digestive disorders [5]. An appropriate diet therefore plays a key role in prevention, which is why breeders often use special feed additives, such as probiotics, prebiotics or synbiotics, to improve gastrointestinal function. These additives, due to the fact of their antagonistic effect on pathogenic and opportunistic microbes, are the subject of growing interest among livestock breeders.Improvement in the efficiency of digestion through optimisation of the composition of the microbiota directly improves nutrient digestibility and stimulates immune processes, increasing the profitability of production. This is one of the main reasons that breeders are searching for alternative solutions involving administration of probiotic microbes with feed in the form of monocultures or a mixture of different strains. The most commonly used probiotics in breeding practice include species of the genera Lactobacillus, Enterococcus, Pediococcus, Bifidobacterium, Saccharomyces and Bacillus. Work is also being conducted on recombined probiotics, which are among the most innovative biomedical applications of genetically modified organisms [6].Rabbits, like other farmed animals, are fed complete pelleted feed containing soybean meal (SBM) as the main source of protein. In some parts of the world, especially where soybean is not cultivated, recent years have seen a trend towards elimination of SBM from feed for most livestock animals. This is due to the desire to become independent of imported feed or to concerns about products containing genetically modified organisms [7]. In the diet of rabbits, the possibility of replacing SBM with other high-protein plant-based feeds has been studied. The use of dried kernels of barley, wheat and maize, white lupin seeds [8,9], peas [10,11,12] or other plants of the Fabaceae family [13,14] has been shown to have a beneficial effect on production efficiency. However, a large share of these ingredients in the diet results in health problems associated with the presence of anti-nutrients (e.g., tannins, antitrypsin factor, haemagglutinin, α-galactosides and alkaloids) or problems involving the balancing of the diet [15,16,17].Fermented protein components are currently a subject of great interest. Owing to their synergistic effect involving stabilisation of the gut microbiota and their valuable nutritional properties, they are becoming a sought-after bioproduct in the feed market. At the same time, efforts to limit the use of animal feed containing genetically modified organisms (GMOs) have prompted the search for an alternative source of easily digestible protein. A new material that meets these expectations may be fermented rapeseed meal (FRSM). Due to the microbial fermentation process, rapeseed meal (RSM), on the one hand, loses its anti-nutritional substances and, on the other hand, acquires probiotic properties. Moreover, it becomes a source of sulphur-containing amino acids, more easily digestible protein, digestive enzymes, and antioxidant compounds. In research in pigs, its inclusion in the diet has been shown to improve nutrient digestibility, resulting in improved growth performance. At the same time, by reducing unfavourable gut microbes and stimulating immune processes, FRSM has a prophylactic function and positively affects animal health [18].Bacteria of the genus Bacillus, used in fermentation of plants, can perform a probiotic function, and the products of their metabolism can beneficially modulate immune system activity in animals [18,19]. De-Yu Hung et al. [19] demonstrated that Bacillus bacteria have a positive effect by alleviating diarrhoea and reducing the number of gut pathogens. The studies cited show that FRSM can partially replace SBM in diets for monogastric animals and can be a valuable additive stimulating immune processes in the body and, thus, improve the condition of animals.The available literature lacks studies on the possibility of using FRSM in the diet of rabbits. Moreover, there are few studies characterising the gut microbiota of rabbits, and these often focus only on individual segments of the GIT. Therefore, the present study was conducted to determine the effect of the use of varied amounts of FRSM in diets for rabbits on the immune status and microbiota of segments of the GIT, i.e., the duodenum, small intestine, caecum and colon.2. Materials and Methods2.1. Preparation of FRSMRSM was fermented using the strain Bacillus subtilis 87Y from the strain collection of InventionBio Ltd. (Bydgoszcz, Poland). The bacteria were multiplied on Lysogeny Broth (LB) medium (10 g/L NaCl, 10 g/L peptone, 5 g/L yeast extract), from which a bacterial suspension of OD 600 = 0.1 was obtained in MIM1 medium (8.4 g/L Na2HPO4, 3.9 g/L NaH2PO4, 2.3 g/L urea, 0.5 g/L MgSO4, 60 g/L saccharose, 1.2 mg/L FeSO4, 1.6 mg/L CuSO4, 5 mg/L MnSO4), [20]. Prior to fermentation the RSM was sterilised for 15 min and then inoculated with the previously prepared bacterial suspension in a 1:1 ratio while maintaining 50% moisture by adding sterile water and aerating at 50 L/min. Fermentation was carried out for 24 h at 37 °C with continuous mixing (20 rpm), after which the RSM was dried to 10–11% moisture with a fluid bed dryer.2.2. Experimental Animals and DietForty 35 day old rabbits (Oryctolagus cuniculus) were used in the experiment. The rabbits were crosses of two breeds: New Zealand White, a typical meat breed, and Popielno White.The animals were assigned to four groups of 10 each, with similar body weights in each group. Each group comprised 5 females and 5 males. The animals were kept in cages of two, which gave five repetitions within each group. The animals in the control group (group C) were fed a standard diet. The experimental groups received 4% (group FR4), 8% (group FR8) or 12% (group FR12) FRSM, (Table 1). FRSM was included in place of standard SBM, and the chemical composition of the feed ration in all groups was balanced according to standards for feeding fur-bearing animals (Table 1), [21]. During the experiment, the animals were under the constant supervision of the farm veterinarian, were fed ad libitum and had free access to drinking water.In each group, 5 animals of each sex were kept in two-storey wire mesh cages equipped with a feeder and a nipple drinker, in a closed, climate-controlled building. The rabbits received feed containing FRSM from weaning at 35 days of age until slaughter at 120 days. All tests on animals were performed with the approval of the Local Ethics Committee for experiments on animals (approval no. 65/2020).2.3. Experimental Procedures and Sample Collection2.3.1. Chemical Analysis of FRSM and Feed MixturesThe chemical composition of the FRSM and of the diets for animals in groups C, FR4, FR8 and FR12 (i.e., the content of total protein, dry matter and crude ash) was analysed according to Latimer [23]. We also determined the quantitative composition of amino acids (Lys, Met and Cys) by ion-exchange chromatography with spectrophotometric detection (IEC-Vis). The content of Ca2+ and Na+ was determined by atomic absorption spectrometry [23]. Total phosphorus content was determined by spectrometry according to Fiske and Subbarow [24].In addition, also determined in the diet and fermented components were the content of phytate phosphorus according to Oberleas [25], lactic acid content according to Taylor [26], glucosinolates according to ISO 10633-1 [27] and tannins according to Canbaş et al. [28].The analyses of FRSM, RSM and feeds were carried out in three batches and in duplicate.2.3.2. Experimental Procedures and Sample CollectionCollection of Biological Material for AnalysisAt the end of the experiment, when the animals were 120 days old, 6 rabbits (3 male and 3 female) with average body weight/group were selected for slaughter.Before slaughter, at 119 days of age, 6 rabbits (3 male and 3 female) with average body weight/group were selected for blood collection. Blood was drawn following local anaesthesia from the marginal ear vein into Monovette tubes with a clot activator (Sarstedt). The blood was centrifuged in a laboratory to obtain plasma for determination of immunoglobulins.At the end of the experiment, at 120 days of age, the animals from which blood had been taken were slaughtered. The animals were decapitated after being stunned by a blow to the head, owing to which the body of the animal was relaxed and sensitivity to pain was completely and immediately eliminated. The animals were slaughtered in accordance with Council Regulation (EC) no. 1099/2009 of 24 September 2009 on the protection of animals at the time of killing [29].Immediately after an incision was made in the abdominal wall, the entire GIT tract was removed from each carcass, and then the intestinal contents from each segment of the GIT were sampled in sterile conditions using an MSC Advantage class II laminar safety cabinet with HEPA filters (Thermo Scientific, Waltham, MA, USA). From each part of the GIT about 10 g of intestinal contents were sampled. The samples were homogenised with a CAT Unidrive X1000 handheld homogeniser (Deerfield, IL, USA) for 60 s at a blade speed of 4000 rpm.Blood AnalysisClass A, G and M immunoglobulins were determined in the blood plasma using ELISA assays (Genorise Scientific Inc., Glen Mills, PA, USA).Microbiological AnalysisFrom the homogenised intestinal contents, 20 g of material was weighed out and placed in sterile bottles containing 180 mL of Ringer solution. The solution was shaken for 5 min and left to settle for 15 min. Then ten-fold serial dilutions were prepared in Ringer solution and plated on previously prepared Petri dishes with an appropriate microbiological medium. The following were determined in the material:The total number of mesophilic aerobic bacteria, on enriched agar medium for 48 h at 37 °C (BTL Ltd., Łódź, Poland);The total number of fungi, on Sabouraud agar, for 5–7 d at 25 °C (BTL Ltd., Łódź, Poland);The total number of coliform bacteria on Endo LES agar, for 24 h at 37 °C (BTL Ltd., Łódź, Poland);The total number of Escherichia coli, on mFC agar for 18–24 h at 44 °C (BTL Ltd., Łódź, Poland);The total number of Clostridium perfringens—sulphate-reducing bacteria growing in anaerobic conditions, on iron sulphite agar (TSC) for 48 h at 37 °C (BioMerieux, Marcy l’Etoile, France);The total number of lactic acid bacteria of the genus Lactobacillus, on MRS agar for 3–5 d at 30 °C (BTL Ltd., Łódź, Poland);The presence of Salmonella, on SS agar (Salmonella–Shigella and XLD) after prior multiplication of samples in buffered peptone water and Rappaport–Vassiliadis broth (BTL Ltd., Łódź, Poland) for 24 h at 37 °C. Final identification was carried out using API tests (BioMerieux, Marcy l’Etoile, France) and polyvalent sera (Biomed Inc., Kraków, Poland).Each sample was plated in triplicate. Following incubation, the colonies were counted with a Scan 300 automatic counter (Interscience, Saint Nom la Brétèche, France) and enumerated in accordance with ISO 4832 [30] and ISO 7218 [31]. Next, the numbers of each morphological type were determined and expressed as colony forming units per g of intestinal contents [CFU/g].2.4. Statistical AnalysisThe results obtained in each group were presented as arithmetic means (M) and standard deviations (SDs) as well as the standard error of the mean (SEM). The normality of the distribution was tested by the Shapiro–Wilk test. If the distribution was normal, one-way analysis of variance was performed, and the differences between groups were determined by the Tukey test. The analysis was performed using Statistica software version 13.1 (StatSoft Inc., Tulsa, OK, USA).Values were considered to differ significantly at p ≤ 0.05. Values designated with the same superscript letters a, b between groups differ significantly (p ≤ 0.05). Correlations marked with an asterisk (*) are statistically significant (p ≤ 0.05).3. Results3.1. Chemical Composition of RSM before and after FermentationThe effect of fermentation on the chemical composition of the RSM is presented in Table 2. The fermentation process caused a minor increase in the content of protein and ash and a decrease in the content of fat, fibre and dry matter. The increase in the content of protein and lactic acid can be explained by the bacteria accompanying fermentation and by protein synthesis from the biomass of microbes (Table 2).3.2. Microbial Population in the GIT of RabbitsMicrobiological diagnostics of the bacterial microbiota of the GIT is a valuable parameter in assessment of the bacterial balance of the intestines. The microbial activity of the intestinal contents was determined as the total numbers of microbes of each group, including the number of aerobic and anaerobic bacteria, as well as their concentrations in each part of the GIT (Table 3). The number of mesophilic aerobic bacteria in the duodenum, caecum and colon was similar in all groups. A statistically significant increase in the number of mesophilic aerobic bacteria was observed only in the contents of the small intestine of group FR4 relative to the control (group C). The concentration of microscopic fungi was similar in all experimental groups and in the control group (Table 3).Analysis of the numbers of coliforms did not reveal any significant effect of FRSM on their concentration in the duodenum or caecum (Table 4). In the small intestine, on the other hand, there was a gradual decrease in the number of coliforms that was significantly lower in the group of rabbits receiving a diet with 12% FRSM (FR12) than in the control (p < 0.05). In the colon, the increase in the number of coliforms in the control group was statistically significant in comparison to all experimental groups (FR4, FR8 and FR12). A similar relationship was noted for Escherichia coli bacteria (Table 4). A statistically significant increase in the number of E. coli bacteria was noted in the small intestine of the animals in group C compared to groups FR8 and FR12 (p < 0.05) and in the colon of animals in group C relative to the other experimental groups.In the duodenum of the rabbits receiving a diet with FRSM (FR4, FR8 and FR12), there was a statistically significant increase in the number of lactic acid bacteria (LAB) relative to group C. In the small intestine and colon, a significant increase in LAB was noted in the groups receiving 8% and 12% FRSM (Table 4).The number of anaerobic bacteria of the species Clostridium perfringens in the duodenum and caecum of the animals in the control group was statistically significantly higher than in the animals whose feed contained FRSM (Table 5). The diet with 12% FRSM (group FR12) caused a statistically significant (p < 0.05) decrease in the number of Clostridium perfringens bacteria in the contents of the small intestine and colon relative to the control group and other experimental groups (Table 5).Pearson correlation analysis of the numbers of microorganisms of various groups in different sections of the GIT showed numerous significant effects of lactic acid bacteria on the other microbial populations. In the duodenum, a strong negative correlation was observed between the number of lactic acid bacteria and the number of bacteria of the species Clostridium perfringens (−0.748, FR12), and in the control group between the number of LAB and the population of mesophilic aerobic bacteria (−0.986, C). Correlation analysis showed a very strong relationship (0.999) between the concentration of lactic acid bacteria in the duodenum and the number of fungi in this section of the GIT in group C (Table 6). In the caecum of rabbits in group FR8, a strong positive correlation (0.797) was noted between the number of lactic acid bacteria and the total number of mesophilic aerobic bacteria. Statistically significant relationships between LAB in the caecum contents and the number of Escherichia coli bacteria were confirmed by a strong negative correlation (−0.975). Similar relationships were noted in the small intestine of groups FR8 (−0.998) and FR12 (0.606). The number of lactic bacteria was lowest in the colon of rabbits in group C and was statistically significant. A very strong positive correlation was noted between the numbers of LAB and Clostridium perfringens: 0.877 for group C and 0.925 for FR4 (Table 6).3.3. Titres of IgA, Ig and IgM in Rabbit PlasmaThe level of class G immunoglobulins in the plasma of rabbits from group FR12 was statistically lower than in the control group C. The level of IgG decreased as the level of FRSM in the diet increased (Table 7). The statistical analysis showed no differences in the titres of class A and M immunoglobulins between groups (Table 7). However, it is worth noting the lower IgA and IgM values in the plasma of the rabbits in group FR4 compared to the other groups (p > 0.05).Correlations were also sought between the level of class G immunoglobulins in the plasma and the numbers of microorganisms in the GIT of rabbits depending on the amount of FRSM in their diet. The Pearson analysis showed numerous complex correlations (Table 8). The table shows that the increase in intake of FRSM was correlated with an increase in the correlations between the immunoglobulin level and the size of the microbial population in the GIT. The concentration of class G immunoglobulins in the control group was strongly negatively correlated with the number of anaerobic C. perfringens bacteria in the duodenum (Table 8). In this part of the intestine, the IgG level was also strongly correlated (0.980) with the number of E. coli (FR8). In group FR12, a strong positive correlation (0.897) was noted between the concentration of IgG and the number of fungi. These antibodies in the colon of group C were strongly correlated with the number of mesophilic aerobic bacteria (0.970), the total number of fungi (0.989) and the number of E. coli (0.988).4. DiscussionAn essential factor for maintaining a balance of the populations of the microbiota is a diet with an appropriate composition. Disturbances of their activity may result from the presence of anti-nutritional substances in feed components such as glucosinolates or tannins in RSM. During fermentation, anti-nutritional substances undergo reduction, leading to an increase in the level of lactic acid that exerts a beneficial effect by stimulating the immune system. FRSM, owing to its nutritional value and health-promoting properties, can beneficially affect production parameters, nutrient digestibility, the gut microbiota and metabolic processes [18].During digestion in rabbits, changes take place in physicochemical parameters, such as pH, redox potential, and metabolite concentrations, that directly affect the balance of microbial communities in the GIT [32]. Michelland et al. [33] demonstrated that the bacterial community of the caecum of rabbits can change and adapt to varying qualities and quantities of nutrients in the diet and can achieve a balance quickly following intake of new feed. This was confirmed in the present study in which the inclusion of FRSM in the diet of rabbits was not shown to affect most key groups of microorganisms taking part in feed digestion.The microbial species variation in the GIT of rabbits is affected by numerous factors, from intestinal peristalsis and mixing of the intestinal contents determined by physiological structure, to periodic gastrointestinal strictures (as in the duodenum), to the presence of a number of gastrointestinal hormones and peptides, such as cholecystokinin, somatostatin, vasoactive intestinal peptide and “substance P” [34]. Bowel transit time is highly variable and shorter than in other herbivores. Peristaltic contractions occur slowly, every 10–15 min, and do not change with the stages of the caecotrophic cycle. Food retention times are estimated at 10–20 min in the jejunum and 30–60 min in the ileum [35]. The type and number of microorganisms in the GIT depends on the section of the GIT. In its first sections, i.e., the stomach and superior part of the duodenum, due to their acidic pH, the presence of bile, shorter transit time and limited mucus production, the number of microbes is smaller. These sections of the GIT are dominated by lactobacilli and enterococci. From the superior part of the duodenum to the jejunum and ileum, the number of bacteria steadily increases, until it reaches a value from 1011 to 1012 CFU/g of faeces in the colon and caecum. The availability of oxygen and changes in pH in different sections of the GIT are conducive to the development of Bacteroides spp., which carry out fermentation processes leading to the generation of volatile fatty acids. These, in turn, regulate intestinal pH, serve as an energy source for intestinal epithelial cells and promote gastrointestinal motility [36].A high level of high-quality fibre (NDSF) and protein in the diet of rabbits is important as well. Intake of rapidly fermenting fibre has a beneficial effect on the activity and concentration of volatile fatty acids in the caecum and improves the functioning of the intestinal mucosa and the structure of the microbiota. A diet with a well-balanced content of protein and essential amino acids, rich in arginine, reduces the population of Clostridium spp. and Helicobacter spp. bacteria in the intestines, thereby reducing mortality in the herd [37]. The diet proposed in the present study, containing FRSM (8–12%), can be a valuable source of these substances that benefit rabbit health and can play a role in preventing imbalances in the gut microbiota. The probiotic and prebiotic substances contained in it are believed to stimulate the growth of Bifidobacterium and Lactobacillus bacteria, which are beneficial for the host. Probiotics have been shown to have a positive effect on the gut microbiota of rabbits in research by Bónai et al. [38]. The authors observed a decrease in the number of anaerobic Clostridium spp. and E. coli bacteria following administration of inulin. These results are consistent with those obtained in the present study. Rabbits from the groups receiving a diet with an increased proportion of FRSM (8% or 12%) had lower concentrations of anaerobic bacteria and E. coli in the intestinal contents. Bacteria of the genera Bacillus and Lactobacillus, used in fermentation of rapeseed, may perform an important probiotic role in the GIT of rabbits. The biological effects of their activity depend on the strains of microorganisms and their ability to maintain metabolic activity in the gut environment. Studies by Bónai et al. [39] and Pascual et al. [40] indicate that the addition of a probiotic in the form of Bacillus cereus bacteria or Saccharomyces cerevisiae yeast to the diet of rabbits has a beneficial effect on their health. The authors observed a decrease in the number of Clostridium and E. coli bacteria, which is in agreement with the results of our experiment. Research by Kimse et al. [41], on the other hand, found that probiotic bacteria did not affect the structure or diversity of the bacterial population in the intestine of rabbits. In the present study as well, the probiotic microorganisms contained in the FRSM had no significant effect on the size of the major groups of microorganisms or their total number. Although the occurrence of Lactobacillus strains in the rabbit microbiota is low, supplementation with them increases the number of cellulolytic bacteria and increases the number of ureolytic bacteria [42]. In the present study, a significant increase in the number of lactic acid bacteria in the duodenum and colon was observed as the proportion of FRSM in the diet increased.The gut microbiota is closely linked to the immune system (gut-associated lymphoid tissue—GALT) and in most mammals is primarily located in the small intestine and colon. Rabbits have special additional structures—the sacculus rotundus, located at the ileocaecal junction, and the vermiform appendix, at the end of the caecum. Together with the gut microbiota, lymphoid aggregates, such as Peyer’s patches and isolated cells dispersed in the lamina propria and the epithelium of the villi, play an important role in the acquisition of immunity and differentiation of antibodies [43]. FRSM, rich in bioactive peptides, can stimulate the activity of the intestinal immune system (GALT). By stimulating the microbiota of the GIT and reducing the proportion of potentially pathogenic microbes, such as Clostridium or E. coli bacteria, it can counteract excessive immunisation in the intestinal mucosa and prevent inflammation. This is confirmed by Stappenbeck et al. [44], who observed a decrease in GALT activity in mice that were not exposed to pathogens. The authors also showed weak immunisation of the immune system and a low level of antibodies in the blood serum. In rabbits, levels of antibodies are dependent on the gut microbiota, which influences the level and diversification of the repertoire of antibodies. A study by Severson et al. [45] showed that the GALT system, despite continual immunisation by the commensal bacteria of the digestive tract, is able to acquire tolerance to Bacillus antigens with no increase in the immune response. This is confirmed by the present study using RSM fermented with Bacillus strains. This suggests that tolerance to the antigens of these microbes can be acquired, and the strength and direction of the effect of FRSM was established based on correlation analysis. The formation of mechanisms of tolerance also depends on the presence of probiotic microorganisms and the substances they produce.FRSM, by supplying valuable bioactive substances, appears to be a good component in the diet of rabbits, enhancing immune system development and helping to prevent disturbances of the gut microbiota. The results of the present study indicate that the use of new biotechnological research solutions can lead to changes in agricultural practices. The use of fermented feed products in the diet can diversify protein sources and stimulate the beneficial microbiota of the GIT of animals. These measures are particularly important because they can help to reduce the use of antibiotics for therapeutic purposes through nutritional prevention and optimal development of the immune system.

animals : an open access journal from mdpi

10.3390/ani11051190

PMC8143165

Pork is the most consumed meat source for humans, and the utilization of nutritional approaches to produce pork with an appropriate content of intramuscular fat (IMF) and a balanced ratio of different kinds of fatty acid is an important objective pursuit of swine production. We speculated that dietary supplementation of beta-hydroxy beta-methyl butyrate (HMB) may provide benefits in lipid metabolism of skeletal muscle. In this study, we try to investigate the effects of dietary HMB supplementation on muscular lipid metabolism in Bama Xiang mini-pigs. We found that HMB supplementation could decrease the IMF content and increase n3 polyunsaturated fatty acids as well as regulate the related metabolites (N-Methyl-l-glutamate and nummularine A) in the serum of Bama Xiang mini-pigs, thus improving their meat quality.

This study aimed to investigate the effects of dietary beta-hydroxy beta-methyl butyrate (HMB) supplementation on muscular lipid metabolism in Bama Xiang mini-pigs. Thirty-two piglets (8.58 ± 0.40 kg, barrow) were selected and fed a basal diet supplemented either with 0 (control), 0.13%, 0.64%, or 1.28% HMB for 60 days. Throughout the experiments, they had free access to clean drinking water and diets. Data of this study were analyzed by one-way ANOVA using the SAS 8.2 software package, followed by a Tukey’s studentized range test to explore treatment effects. The results showed that compared to the control, 0.13% HMB decreased the intramuscular fat (IMF) content and increased polyunsaturated fatty acids (PUFAs) in Longissimus thoracis muscle (LTM), and increased the n3 PUFAs in soleus muscles (SM, p < 0.05). Moreover, HMB supplementation led to alterations in the mRNA expression of genes related to lipid metabolism. Serum metabolome profiling showed that in both LTM and SM of Bama Xiang mini-pigs, N-Methyl-l-glutamate was positively correlated with SFA and nummularine A was negatively correlated with C18:3n3 PUFA (p < 0.05). Therefore, N-Methyl-l-glutamate and nummularine A might be potential biomarkers of the HMB-supplemented group. These results suggested that dietary HMB supplementation could decrease the IMF content and increase n3 PUFAs as well as regulate the related metabolites (N-Methyl-l-glutamate and nummularine A) in the serum of pigs.

1. IntroductionPork is the most consumed meat source for humans, and hence there has been a growing focus on pork quality [1]. An increasing body of literature has identified intramuscular fat (IMF) and the fatty acid composition of muscle tissues as important contributing factors to pork quality [2,3,4,5]. There is a close relationship between IMF and meat edible qualities such as juiciness, flavor, drip loss, and tenderness [6]. The recommended range of IMF content is 2.5–3% [7]. Pork with low IMF content has poor palatability, whereas pork with high IMF content will greatly decrease the purchasing decisions of customers due to the existence of too rich visible marbling [8]. Meanwhile, the ratios of different kinds of fatty acids in muscle tissues are also central to the nutritional value of meat and contribute importantly to human health [9]. Specifically, excessive content of saturated fatty acid (SFA) and n6 polyunsaturated fatty acids (PUFA) together with increased ratios of SFA/PUFA and n6/n3 PUFA are associated with the development of cardiovascular and metabolic disorders in humans [10,11]. Nutritional modulation is an efficient strategy to alter the fatty acid profile and concentration of pigs’ muscles [3,12]. On this basis, the utilization of nutritional approaches to produce pork with an appropriate content of IMF and a balanced ratio of different kinds of fatty acid is an important objective pursuit of swine production.Beta-hydroxy-beta-methyl butyrate (HMB), a metabolite of leucine, has gained popularity as a nutritional supplement in humans to augment the mass and strength of skeletal muscles [13] and in animals to modulate muscle fiber types as well as to improve carcass quality [14,15]. Apart from these roles, HMB also exerts important roles in the modulation of energy homeostasis and lipid metabolism in adipose tissues [16,17]. For example, in human studies, there is evidence that HMB led to a significant reduction in body fat (−1.1% vs. −0.5%) relative to a placebo, accompanied by a greater elevation in lean body mass (+1.4 kg vs. +0.9 kg) [18]. Similar results were obtained in high fat diet-fed mice, in which HMB supplementation (1% wt/vol) reduced high fat diet-induced body weight gain and the weight of white adipose tissues [17]. Further evidence comes from a swine model showing that HMB supplementation has the ability to modulate adipose tissue function such as lipolysis and fatty acid oxidation [16]. Of note, excessive leucine in tissues and blood may result in mitochondrial dysfunction and insulin resistance, thus increasing the future risk of developing diabetes and its related metabolic disorders. However, HMB cannot be converted reversibly to leucine and thus would not share similar functions to leucine in this regard [19,20]. Despite the importance of HMB in regulating lipid metabolism in adipose tissue, few studies have sought to investigate its effects on lipid metabolism in skeletal muscles of pigs. Therefore, it is posited that HMB may provide benefits in lipid metabolism of skeletal muscle similar to adipose tissue.Metabolomics is an approach for analyzing quantitatively and qualitatively endogenous metabolites in a biological system; these metabolites are intermediates or the end products of cellular processes and have the ability to directly affect the phenotype [4,21]. How fatty acid composition of skeletal muscle is changed by dietary manipulations can be determined by a systematic study of metabolites and metabolic pathways.Bama Xiang mini-pigs, an indigenous minipig breed from Bama Country of China, grow slowly and are lighter and fatter than the modern breeds such as Large White at slaughter, which makes them an ideal experimental animal model for biomedical research [22,23]. Notably, too-high fat content in skeletal muscle will greatly decrease the purchasing decisions of customers, thus decreasing the economic value of Bama Xiang mini-pigs [8]. Based on the roles of HMB in regulating lipid metabolism of adipose tissue, we hypothesized that HMB could improve lipid metabolism in skeletal muscle as well. Therefore, this study aimed to investigate the roles of dietary HMB supplementation in IMF, muscle fatty acid composition, the mRNA expression of lipid metabolism-related genes in skeletal muscle, and serum metabolomics in Bama Xiang mini-pigs.2. Materials and Methods2.1. Animals and DietsAll animal procedures were performed under the guidelines of the Committee on Animal Care of the Institute of Subtropical Agriculture, Chinese Academy of Sciences under ethic approval number ISA-2017-023.Thirty-two Bama Xiang mini-pigs with similar initial body weight (barrow, 8.58 ± 0.40 kg; 60 ± 2 days) were chosen and randomly divided into four groups (with eight piglets per treatment), that is, basal diet (control group, CON); basal diet +0.13% HMB-Ca (low-dose, HMB1 group); basal diet +0.64% HMB-Ca (moderate-dose, HMB2 group); basal diet +1.28% HMB-Ca (high-dose, HMB3 group), in a completely randomized design according to the body weight. All diets were formulated to be isoenergetic and isonitrogenous and to meet the nutrient requirements and physiological needs of growing mini-pigs (Table 1) [24]. Pigs were housed individually in cages (0.8 × 1.8 m) throughout the experiment and had free access to clean drinking water and diets. The experiment lasted for 60 days.2.2. Sample CollectionAll the pigs were fasted overnight at the end of the feeding trial and slaughtered by electrically stunning (250 V, 0.5 A, for 5–6 s) and exsanguinating. Notably, before slaughter, serum samples were collected and then stored as previously described [2]. After slaughter, samples of the Longissimus thoracis muscle (LTM) and soleus muscle (SM) dissected from the left side of the carcasses were stored at −20 °C or placed in liquid nitrogen (about 10 g of each tissue) and then stored at −80 °C until further analysis.2.3. Measurements of Serum Amino Acids and Lipid-Related SubstancesAmino acid concentrations in serum were determined as previously described [25]. The concentrations of glucose-6-phosphate dehydrogenase (G6PD), adipose tissue triglyceride lipase (ATGL), lipoprotein lipase (LPL), and 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMGR) were measured with the corresponding commercial ELISA kits (Jiangsu enzymebiao Biotechnology Co., Ltd., Jiangsu, China) according to their instructions by using an enzyme-linked immunosorbent assay plate reader (Bio-Tek, Winooski, VT, USA).2.4. Intramuscular Fat and Fatty Acid CompositionThe IMF content of LTM and SM was analyzed using the Soxhlet extraction method as previously described [2]. The fatty acid composition of selected muscle tissues was analyzed via gas–liquid chromatography of methyl esters using an Agilent 7890A GC as previously described [26]. The concentration of individual fatty acids was expressed as a percentage of total fatty acids. Based on the fatty acid composition, we calculated the following parameters: the sum of SFAs and PUFAs; the PUFAs/SFAs ratio; the n6/n3 PUFA ratio; hypocholesterolemic/hypercholesterolemic ratio (h/H); atherogenicity index (AI); and thrombogenicity index (TI).2.5. Quantitative Real-Time PCR AnalysisThe quantitative real-time PCR analysis was conducted in LTM and SM samples as previously described [27]. Briefly, the TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used to isolate the total RNA of LTM and SM. The RNA quality and quantity were checked by 1% agarose gel electrophoresis and by ultraviolet spectroscopy using a NanoDrop® ND-1000 spectrophotometer (Thermo Fisher Scientific, Inc., Waltham, MA, USA). Thereafter, approximately 1.0 μg of total RNA was used to be reverse-transcribed to complementary DNA by using First-Strand cDNA Synthesis Kit (Fermentas Inc., Hanover, MD, USA). The primers for the target genes were purchased from Sangon Biotech Co., Ltd., (Shanghai China) and their sequences are shown in Table S1. The relative expression levels of the selected genes were analyzed using SYBR Green Detection Kit (TaKaRa). The reaction system used for PCR detection includes the following substances: 5 μL SYBR Green mix, an aliquot (2 μL) of a complementary DNA template (corresponding to 25 ng of total RNA) solution, 0.2 μL ROX Reference Dye (50×), and 0.2 μL each of forward and reverse primers. PCR conditions were as follows: incubation for 10 min at 95 °C, followed by 40 cycles of denaturation for 15 s at 95 °C, annealing and extension for 60 s at 56–64 °C. The fluorescent signal was detected by the ABI 7900HT RT-PCR system (Applied Biosystems, Branchburg, NJ, USA). A melting curve was generated for each sample at the end of each run to ensure the purity of the amplified products. The values of cycle threshold (Ct) represent the means of triplicate measurements. The mRNA expression levels of target genes were normalized by the reference gene β-actin (an internal control) and determined by the 2−ΔΔCt method [27].2.6. Liquid Chromatography—Tandem Mass Spectrometry (LC-MS/MS) AnalysisSerum samples (100 μL) of the CON and 0.13% HMB groups (n = 6) were spiked with 20 μL of internal standards (0.3 mg/mL L-2-chlorophenylalanine in acetonitrile) and 400 μL of acetonitrile:methanol (1:1, v/v) solution, and then extracted according to the manufacturer’s instructions (Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China). LC-MS/MS analysis for the extracted samples was performed using a Vanquish UHPLC system (Thermo Fisher) coupled with an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific, Inc.) operating in the data-dependent acquisition mode. The quality control (QC) sample was a pooled sample in which aliquots of all the extracted samples were mixed, and then analyzed using the same method as used for the analytic samples. It helped to represent the whole sample set, which would be injected at regular intervals (every 4 samples) in order to monitor the stability of the analysis. As presented in Figure S1A,B, the total ion chromatogram of the QC samples in positive and negative ion modes showed a good overlap in both ion modes, suggesting good stability of the proposed method.The parameters of chromatography were as follows: column: BEH C18 (100 mm × 2.1 mm i.d., 1.7 μm; Waters, Milford, MA, USA); gradient mobile phase: water with 0.1% formic acid as solvent A and acetonitrile and isopropanol (1:1, v/v) (containing 0.1% formic acid) as solvent B; flow rate: 0.4 mL/min; sample injection volume: 2 μL; column temperature: 40 °C. The gradient mobile phase program applied as follows: t = 0 min, 95% A; t = 3 min, 80% A; t = 9 min, 5% A; t = 13.1 min, 95% A; t = 16 min, 95% A. The parameters of the mass spectrum were as follows: sheath gas flow rate 40 psi; aus gas flow rate 10 psi; aus gas heater temperature, 400 °C; ion-spray voltage floating (ISVF), −2800 V in negative mode and 3500 V in positive mode, respectively; normalized collision energy, 20–40–60 V rolling for MS/MS. The detection was carried out over a mass range of 70–1050 m/z.2.7. Statistical AnalysisFor the metabonomics data, the raw data were imported into the Progenesis QI software (Waters Corporation, Milford, MA, USA) for data preprocessing after UHPLC-QE-HFX/MS analyses. In order to detect metabolites, we carried on the principal component analysis (PCA) and orthogonal partial least-squares (OPLS-DA) analysis. To reveal the differences in the metabolic composition among groups, T-test (Student’s test) and the multivariate analysis of OPLS-DA were used to explore the potential marker or differential metabolites (variable importance in the projection (VIP) > 1, p-value < 0.05). For analysis software, the online platform of Majorbio ISanger Cloud platform (https://cloud.majorbio.com/) was used (assessed on 6 February 2020).Data of this study were analyzed by one-way ANOVA using the SAS 8.2 software package, followed by a Tukey’s studentized range test to explore treatment effects. Results are presented as means ± standard errors (SD). Differences between significant means were viewed to be statistically different at p < 0.05.3. Results3.1. Growth Performance and IMFAll groups of pigs exhibited similar average daily feed intake and feed conversion ratios (p > 0.05, Table S2, https://kns.cnki.net/kcms/detail/11.5461.S.20210316.0920.012.html, assessed on 30 October 2020). However, compared to the CON, 0.13% HMB significantly increased the average daily gain (ADG) (p < 0.05). When fed to pigs at dosages of 0.64% and 1.28%, HMB failed to elevate ADG of pigs (p > 0.05, Table S2). The IMF content of the LTM in the 0.13% HMB group was significantly lower than that in the other three groups (p < 0.05), and no differences were observed among the three groups (p > 0.05, Table S2).3.2. Serum Amino Acid Profile and Lipid-Related SubstancesAs shown in Table 2, relative to the CON, the 0.13% HMB augmented the serum concentrations of alanine, O-phosphoserine, and ornithine (p < 0.05), decreased the concentrations of aspartic acid and β-alanine (p < 0.05), and tended to increase the concentrations of threonine (p = 0.0841) and total NEAA (p = 0.0867). Compared to the CON, HMB at the level of 0.64% significantly increased the concentration of histidine (p < 0.05). Compared to the CON, HMB at the levels of 1.28% significantly increased the concentrations of alanine and arginine (p < 0.05). Dietary treatments exerted no significant effects on the concentrations of other amino acids (p > 0.05).As shown in Table 3, compared to the CON, the 0.13% and 1.28% HMB reduced the concentrations of G6PD and LPL (p < 0.05), while the 0.64% HMB increased the LPL concentration (p < 0.05). The ATGL concentration was the lowest in the 0.64% HMB group and highest in the 0.13% HMB group, with intermediate values in the CON and 1.28% HMB groups (p < 0.05), and no differences were detected between CON and 0.13% HMB groups. The HMGR concentration in the CON group was similar to that in the 0.13% HMB group (p > 0.05), and was significantly lower than that in 0.64% and 1.28% HMB groups (p < 0.05).3.3. Fatty Acid Composition of Skeletal Muscle TissueAs depicted in Table 4, dietary HMB supplementation greatly affected the fatty acid composition in the LTM. The concentrations of n6 PUFAs, such as C18:2 n6c, C20:3 n6, and C20:4 n6, were increased in all samples within the 0.13% HMB group compared with the CON (p < 0.01). Compared with the CON, C18:3 n3 and the sum of SFA were significantly increased and decreased in the 0.13% group (p < 0.01), respectively. The sum of PUFA, the ratio of PUFA to SFA, the sum of n3 PUFA and n6 PUFA contents, and the n6:n3 PUFA ratio in the 0.13% group were higher than those in the CON group (p < 0.01). Compared with the CON, the 0.13% HMB group significantly increased the h/H and decreased the AI and TI (p < 0.05).The results regarding the fatty acid composition of the SM of pigs fed diets with different dietary HMB levels are presented in Table 5. No difference was observed in the contents of C18:2 n6c, C20:3 n6 and C20:4 n6 as well as the sum of n6 PUFA among the groups (p > 0.05). The C18:3 n3 content and the sum of n3 PUFA were the highest in the 0.13% HMB group and lowest in the 1.28% HMB group, with intermediate values in the CON and 0.64% HMB groups. The sum of SFA in the 0.13% HMB group was significantly lower than that in the other three groups (p < 0.01). The n6:n3 PUFA ratio was the highest in the CON and lowest in the 0.13% HMB group, with intermediate values in the other two groups (p < 0.05). Compared with the CON, the 0.13% HMB group significantly decreased the TI (p < 0.05). Unlike in the LTM, no significant difference was observed in PUFA contents, the PUFA:SFA ratio, the h/H, and AI in the SM among the groups (p > 0.05).3.4. mRNA Expression of Lipid Metabolism-Related Genes in Skeletal MusclesFigure 1 shows the mRNA expression levels of genes associated with lipogenesis (acetyl-CoA carboxylase, ACC; fatty acid synthase, FAS), lipolysis (hormone-sensitive lipase, HSL; carnitine palmitoyl transferase 1B, CPT-1B), lipid uptake (fatty acid transport protein, FATP-1; fatty acid-binding protein 4, FABP4), and transcription factor (peroxisome proliferator-activated receptor γ, PPARγ; CCAAT/enhancer-binding protein α, C/EBPα) in the LTM (Figure 1A) and SM (Figure 1B), respectively.In the LTM, no difference was obtained in mRNA levels of C/EBPα among the groups (p > 0.05). Compared to the other three groups, the 0.13% HMB group increased the mRNA expression levels of FAS and ACC, and decreased those of HSL and CPT-1B (p < 0.05), and no difference was observed among the three groups. The FABP4 mRNA expression level was the highest in the 0.13% and 0.64% HMB groups and lowest in the 1.28% HMB group, with an intermediate value in the CON group (p < 0.05). The PPARγ mRNA expression level was the highest in the 1.28% HMB group, followed by the 0.64% HMB group, and lowest in the other two groups (p < 0.05).In the SM, the FABP4 mRNA expression was increased in all the treatment groups relative to the CON (p < 0.05). No observable differences in the mRNA expression of ACC, FAS, HSL, CPT1B, FATP1, C/EBPα, and PPARγ were observed upon HMB supplementation (p > 0.05).3.5. Metabolomics AnalysisTo further determine the slight differences in metabolic profiles among all groups, PCA and OPLS-DA were carried out. As shown in Figure 2A, the scatter plots of serum were cross-distributed among the two groups. Even so, we cannot draw conclusions that no differences were observed in the metabolic profiles of serum among the groups. Supervised OPLS-DA can reveal the specific variables that caused differences among groups relative to unsupervised PCA [28]. Therefore, the multivariate statistical analysis using OPLS-DA was further performed to detect serum-specific metabolites that differ between the CON and HMB treatment, and the scatter plots are presented in Figure 2B. The scatter plots of the CON and HMB groups were distributed in different quadrants, suggesting a significant difference in serum metabolic patterns between the CON and HMB groups.Then, a heat map was plotted for the CON vs. HMB (0.13%) chemical compositions to show the alterations in the metabolite concentrations. As shown in Figure 3, the HMB group exhibited higher content of L-histidine and LysoPE(16:0/0:0), and a lower content of N-Methyl-l-glutamate and nummularine A (VIP > 1, p < 0.05).3.6. The Relationship between Serum Metabolites and Fatty Acid Composition of MusclesAs shown in Figure 4 and Figure 5, N-Methyl-l-glutamate was positively associated with the total SFA content of both LTM and SM (p < 0.01), as well as the n6:n3 ratio in the SM (p < 0.05). Nummularine A was negatively associated with the C18:3n3 content in the LTM and SM (p < 0.05). Of note, the reasons why the number of serum metabolites in Figure 4 and Figure 5 are different from that in Figure 3 are that metabolites in Figure 4 are selected under the condition of VIP ≥ 1.4. DiscussionIn response to increasing consumer demands, a major objective of Bama Xiang mini-pig production is to produce minimal visible fat in skeletal muscle without affecting meat quality. In this study, IMF content in LTM was decreased in the pigs fed diets supplemented with 0.13% HMB relative to the CON (3.02% vs. 3.73%), with higher levels of HMB being ineffective. Consistently, our and other’s previous studies have shown that higher HMB concentrations fail to exert beneficial effects on lipid metabolism in adipose tissues of high-fat diet-fed mice [17] and on protein metabolism in skeletal muscles of neonatal pigs [29] and LPS-challenged pigs [30]. Therefore, the results shown in previous studies and in this study suggest that lower levels of HMB are superior to higher levels of HMB in regulating lipid and protein metabolism. However, up to now, the reasons why that happens are still elusive so that it needs to be further explored. In addition, our results indicated that no observable difference in IMF content in SM was noted following HMB treatment. These observations suggest that HMB regulated muscle lipid metabolism in a tissue-specific manner, and the success of increasing fat loss occurred mainly in LTM (muscles mainly composed of fast-twitch glycolytic fibers) rather than in SM (muscles containing predominately slow-twitch oxidative fibers). Muscles with different fiber types possess different morphological, biochemical, and physiological properties. In particular, slow-twitch fibers have more mitochondria than fast-twitch fibers, and fast-twitch fibers grow faster than slow-twitch fibers [31]. Previous studies using growing pigs (Large White × Landrace) have demonstrated that HMB supplementation led to increased mitochondrial biogenesis and fatty acid oxidation in LTM [14]. Therefore, HMB may promote the fat loss of LTM to ensure that there is enough energy to promote its rapid growth. In addition, IMF content is closely related to the characteristics of meat quality, such as juiciness and tenderness. In general, pork with IMF content 2.5–3% exhibits good meat quality and high economic benefits [6,8]. Therefore, HMB could positively affect the meat quality of muscle mainly containing fast-twitch glycolytic fibers, and the appropriate dosage was 0.13%.Due to its roles in the oxidative stability and nutritional value of muscles, the muscular fatty acid composition is also closely related to meat quality [6,32]. In particular, SFA and PUFA are positively and negatively correlated with meat quality, respectively [5]. However, excessive SFA can affect the degree of fat firmness, subsequently influencing the quality and acceptability of meat products [33]. Human studies have also shown that decreasing dietary SFA and increasing the PUFA:SFA ratio led to a lower risk of cardiovascular diseases [10]. The target for all PUFA:SFA ratio is at least 0.40 [6]. In the current study, although the ratio of all PUFAs to SFAs in selected muscle tissues did not achieve the goal of 0.40 or above, dietary HMB supplementation significantly increased the ratio in LTM relative to the CON, and the maximum elevation of the PUFA:SFA ratio occurred at the HMB level of 0.13% (0.29 vs. 0.10). Similar to IMF content, the PUFA:SFA ratio in SM upon the HMB diets did not achieve the statistical significance in this study. In agreement with these findings, previous studies using growing pigs (Large White × Landrace) also show that HMB supplementation improved fatty acid composition in LTM rather than in SM, as evidenced by increased PUFAs and decreased SFAs as well as SFA/PUFA [34]. However, there is hardly any information available on HMB-affected alterations in muscular fatty acid composition in other animal models, and we do not have a plausible explanation for these observations. Therefore, further investigations are certainly warranted. Apart from the PUFA:SFA ratio, the indices h/H ratio, AI, and TI also can evaluate the effect of fatty acid composition on cardiovascular diseases. Of note, the h/H ratio is more accurate than the PUFA:SFA ratio in reflecting the effect of fatty acid composition on cardiovascular diseases. A relatively high h/H ratio and low content of AI and TI result in a lower incidence of cardiovascular diseases [35]. Consistent with our data of PUFA:SFA ratio, diets supplemented with 0.13% HMB significantly increased the h/H ratio, and decreased AI and TI in the LTM of Bama Xiang mini-pigs, suggesting a beneficial role of HMB in the nutritional value of LTM. Altogether, our results suggested that the synthetic capacity of PUFA in muscles mainly containing fast-twitch glycolytic fibers is improved when Bama Xiang mini-pigs were fed diets supplemented with HMB.In muscle tissues, the ratio of n6 to n3 PUFAs is mainly regulated by diets since it is difficult to alter this ratio genetically in the short term [36]. Increasing n3 PUFAs and decreasing n6 PUFAs are important to ensure superior meat quality and to increase protection against chronic diseases [37,38]. Therefore, to supply pork rich in valuable n3 PUFAs for humans and improve the health of humans, it is of great importance to increase the proportion of n3 PUFAs and to maintain an appropriate ratio of n6 to n3 PUFAs [2,3,26,39]. The recommended dietary ratio of n6/n3 PUFAs for health benefits is 1:1–2:1 [40], yet the typical Western diet often contains 10 or more times the amount of n6 relative to n3 PUFA [41]. Previous studies using growing pigs (Large White × Landrace) have demonstrated that HMB supplementation could augment n3 PUFAs concentrations in both glycolytic and oxidative skeletal muscles [34]. Similar to the observations made by Zhong et al. [34], we also report here that both LTM and SM of Bama Xiang mini-pigs fed diets supplemented with 0.13% HMB exhibited the greatest increase in the concentrations of n3 PUFAs. However, in parallel, we observed that the n6 PUFAs concentrations and the ratio of n6:n3 PUFAs were also greatly increased in LTM of Bama Xiang mini-pigs. Therefore, our results confirm and extend the observations of Zhong et al. by demonstrating that although not influencing the total PUFA amount, HMB supplementation improved the nutritional value of muscles via enhancing n3 PUFA amount. However, the reasons why HMB differently regulated concentrations of n6 PUFAs in muscle with different fiber types are still elusive, and future studies can pay more attention to the relationship between HMB supplementation and muscular n6 PUFAs metabolism.The sources of muscular fatty acids are predominately from diets (such as PUFA) and/or lipogenesis (such as SFA), and the destination of muscular fatty acids are mainly mitochondrial oxidation and/or esterification into triglyceride [42]. In this regard, alterations in the muscular fatty acid profile can, to some extent, reflect the balance among fatty acid uptake from food, lipogenesis, and lipolysis. ACC and FAS are two key enzymes for the biosynthesis of long-chain fatty acids (LCFAs) [43]. In the process of lipogenesis, fatty acid transporters such as FATP1 and FABP4 facilitate and regulate the entry of exogenous fatty acids into the skeletal muscles [44]. When lipolysis occurs, HSL is the key enzyme involved in hydrolyzing stored triglycerides in tissues [45], and CPT1B is the key enzyme responsible for transporting LCFAs into the mitochondria for oxidation [43]. Moreover, the expression of lipogenic- and lipolysis-related genes is under the control of nuclear transcription factors such as C/EBPα and PPARγ [46]. In the current study, 0.13% HMB supplementation affected markers of lipogenesis and lipolysis in LTM of Bama Xiang mini-pigs, as manifested by downregulated mRNA expression of ACC and FAS and upregulated mRNA expression of HSL and CPT1B, thus decreasing lipid stores. However, similar observations were not noted in SM. These findings, to some extent, support our hypothesis mentioned above, that is, increased fatty acid oxidation might be a contributing mechanism for HMB-induced fat loss in LTM of Bama Xiang mini-pigs. Taken together, we summarize that an appropriate HMB level could improve lipid metabolism, especially in muscles mainly containing fast-twitch glycolytic fibers, thus regulating IMF content close to the recommended range (2.5–3%).To further reveal the metabolites and pathways behind the HMB diet regulating the IMF content and fatty acid composition of selected muscles in Bama Xiang mini-pigs, serum samples from the CON and 0.13% HMB diets were subjected to UHPLC-QE-HFX/MS analyses. The total ion chromatogram of the QC samples in positive and negative ion modes supported the stability and reliability of the metabolomics profiles obtained. Moreover, we found that in both LTM and SM of Bama Xiang mini-pigs, N-Methyl-l-glutamate was positively correlated with SFA and nummularine A was negatively correlated with C18:3n3 PUFA. In response to HMB supplementation, the two serum metabolites N-Methyl-l-glutamate and nummularine A were significantly decreased. Therefore, based on these data and the abovementioned data concerning fatty acid composition, it is postulated that N-Methyl-l-glutamate and nummularine A might be potential biomarkers of the HMB-supplemented group and might be implicated in various pathways for lipid metabolism upon HMB supplementation.In conclusion, the current study indicated that dietary HMB supplementation at the level of 0.13% could improve the IMF content in LTM and fatty acid composition in SM of Bama Xiang mini-pigs. This might be due to the alterations in mRNA expression levels of lipogenic- and lipolysis-related genes in muscle tissues. Additionally, dietary inclusion of HMB regulated the related metabolites (including N-Methyl-l-glutamate, L-histidine, and nummularine A) in the serum of pigs. This finding opens a new avenue to understand the mechanisms of action of HMB in improving the meat quality of local pig breeds in China.

animals : an open access journal from mdpi

10.3390/ani11051253

PMC8145495

In vitro and in vivo studies on the supplementation of rumen-protected microencapsulated fatty acid from linseed oil (MO) on rumen digestibility, physiological profile, growth performance, meat quality, and meat fatty acid profile in Korean native steers were conducted. The in vitro study showed that 3% MO is an optimal dose, as there were decreases in the neutral detergent fiber and acid detergent fiber digestibility at 48 h. Supplementation with 3% MO not only promotes growth performance but also enhances the omega-3 fatty acid concentration of meat in Korean native steers.

We evaluated the effects of a rumen-protected microencapsulated supplement from linseed oil (MO) on ruminal fluid, growth performance, meat quality, and fatty acid composition in Korean native steers. In an in vitro experiment, ruminal fluid was taken from two fistulated Holstein dairy cows. Different levels of MO (0%, 1%, 2%, 3%, and 4%) were added to the diet. In an in vivo experiment, eight steers (average body weight = 597.1 ± 50.26 kg; average age = 23.8 ± 0.12 months) were assigned to two dietary groups, no MO (control) and MO (3% MO supplementation on a DM basis), for 186 days. The in vitro study revealed that 3% MO is an optimal dose, as there were decreases in the neutral detergent fiber and acid detergent fiber digestibility at 48 h (p < 0.05). The in vivo study showed increases in the feed efficiency and average daily gain in the 3% MO group compared to the control group on days 1 to 90 (p < 0.05). Regarding meat quality, the shear force produced by the longissimus thoracis muscle in steers from the 3% MO group was lower than that produced by the control group (p < 0.05). Interestingly, in terms of the fatty acid profile, higher concentrations of C22:6n3 were demonstrated in the subcutaneous fat and higher concentrations of C18:3n3, C20:3n3, and C20:5n3 were found in the intramuscular fat from steers fed with 3% MO (p < 0.05). Our results indicate that supplementation with 3% MO supplements improves the growth performance and meat quality modulated by the omega-3 fatty acid content of meat in Korean native steers.

1. IntroductionThe purpose of improving meat quality is to promote the nutritional and commercial value of beef. The intramuscular fat content of beef is an important factor in quality evaluation, as it greatly affects the taste, juiciness, flavor, and tenderness of beef [1]. In addition, to increase the content of intramuscular fat in beef, the use of grain-based feed is encouraged. However, recently, negative perceptions about fat in beef have gradually been spreading. It has been reported that the use of grain-based feed for ruminants induces increases in the content of omega-6 fatty acids in beef, causing an imbalance in the omega-6/omega-3 ratio [2]. It has been previously documented that the consumption of foods with an unbalanced omega-6/omega-3 ratio is associated with various diseases such as obesity, diabetes, cancer, and high blood pressure [3,4]. Accordingly, research on enrichment with omega-3 to improve the balance of omega-6/omega-3 in beef is actively being conducted. Therefore, to maintain an adequate amount of intramuscular fat, it is necessary to increase the content of omega-3 fatty acids, which are beneficial for human health, and to reduce the omega-6/omega-3 ratio.The strategies for increasing the content of omega-3 in beef include the use of forage feeding [5,6], rumen-unprotected supplementation [7,8], rumen-protected supplementation [9,10], and oil seed feeding [11]. Among these, supplementation with linseed oil (enriched omega-3 fatty acid), which is not protected in the rumen, has been shown to increase the omega-3 fatty acid concentration in plasma and fat tissues [12]. Meanwhile, a loss of omega-3 fatty acids was found to occur through extensive biohydrogenation within the rumen. Moreover, soybean oil was shown to adversely affect rumen fermentation, causing negative effects, such as a reduction in productivity [13]. To prevent these problems, rumen-protected supplements have been developed in various ways.Although many studies have investigated rumen-protected supplementation [9,10], the effects on performance, meat quality, and fatty acid composition in Korean native steers require further research. Therefore, the objective of this study was to conduct an in vitro experiment to determine the optimal dosage of rumen bypass microencapsulated fatty acids from linseed oil that does not negatively affect ruminal fermentation. Based on the results of the in vitro experiment, we investigated the effect of supplementation with rumen bypass microencapsulated fatty acids from linseed oil on the growth performance, physiological indicators, meat quality, and fatty acid composition in Korean native steers.2. Materials and MethodsAll experimental procedures involving animals were performed according to the Animal Experimental Guidelines provided by the Animal Care and Use Committee of Konkuk University, Republic of Korea (Approval number, KU19146).2.1. In Vitro Batch Culture2.1.1. Experimental Materials and MethodsRuminal fluid was taken from two fistulated Holstein dairy cows 2 h before feeding. The cows were fed tall fescue and concentrate (a ratio of 6:4) once a day at 0.900 h. Water was supplied ad libitum. Ruminal fluid was filtered with two layers of cheese cloth and mixed in the same ratio. The basal diet fed to the cows was milled through a 1 mm screen and used as a substrate in a ratio of 6:4. Samples of the basal diet (0.5 g) were put in ANKOM bags (filter bag 58, ANKOM Tech, Macedon, NY, USA) to analyze the digestibility. Menke’s buffer solution [14] adjusted to pH 6.8 was prepared under continuous flushing with CO2 at 39 °C. Then, buffer solution and ruminal fluid were mixed in a ratio of 3:1. A quantity of 250 mL of buffered ruminal fluid was filled into the Erlenmeyer flask and added to five ANKOM bags. CO2 gas was flushed into the headspace of the flask. The experiment was conducted on three replicates of five treatments, and samples were incubated in a shaking incubator (JSSI-300C, JSP Corp, Gongju, Korea) for 48 h at 39 °C.The preparation of rumen bypass microencapsulated fatty acids from linseed oil (MO; Microtinic® Omega) was obtained from Vetagro S.p.A (Reggio Emilia, Italy). The material was prepared via a patented spray-cooling microencapsulation process. Briefly, after heating the hydrogenated palm oil, linseed oil was mixed with hydrogenated palm oil and an active principle. A mixture of materials was moved to a chilling chamber and then cut to a size of 1000 to 1500 microns. Finally, the MO was sealed and stored at 4 °C. The ingredients of the rumen-protected microencapsulated fatty acid from linseed oil (MO) were linseed oil (35%), vitamin E (0.5%), rosemary extract (0.3%), and hydrogenated palm oil (64.2%). The main components of the MO were crude fat (95%) and ash (5%) [15].2.1.2. AnalysisThe pH, ammonia-N, volatile fatty acid (VFA) content, and digestibility of the ruminal fluid were analyzed. Measurement of the ruminal pH was conducted with a pH meter (S20 SevenEasy pH, Mettler Toledo, Greifensee, Switzerland), and 5 mL residual ruminal fluid samples were mixed with 1 mL of 2% HgCl2 (wt/vol) and 25% meta-phosphoric acid (wt/vol) solution for ammonia-N and VFA analysis. Four milliliters of culture solution was collected for the ammonia (1 mL) and VFA (3 mL) analyses. All samples were kept frozen at –20 °C until analysis. Ammonia-N was determined via the method presented by Chaney and Marbach [16] using spectrophotometry (Synergy2; Biotek Instruments, Winooski, VT, USA). To determine the VFA content, 3 mL of the culture solution was centrifuged at 20,000× g for 20 min at 4 °C, and 1 mL of supernatant was collected. The 1 mL supernatant sample was mixed with 50 μL of 2% pivalic acid (wt/vol) and used as an internal standard. The VFA composition was analyzed using a gas chromatograph (HP 6890, Agilent Technologies, Santa Clara, CA, USA) equipped with a flame ionization detector and a capillary column (DB-FFAP; Agilent Technologies). The chemical analyses were conducted using the standard method for the association of official analytical chemists (AOAC) [17]. The dry matter (DM) content was determined using the ANKOM bag, and the ANKOM bag was weighed after drying for 48 h at 60 °C in a dry oven. The digestibility of DM was calculated as follows:DM digestibility (%) = dried residue sample/sample weight × 100.(1)The neutral detergent fiber (NDF) and acid detergent fiber (ADF) contents were determined using an ANKOM bag technique. Feed samples (0.5 g) were put into ANKOM bags. The bags were immersed in acetone for 10 min, dried for 10 min, and then put into an ANKOM 200 fiber analyzer (ANKOM Tech, Macedon). The NDF or ADF solution, 4 mL of α-amylase (not included in the ADF analysis), and 20 g of anhydrous sodium sulfate were added. The mixture was heated at 100 °C for 70 min. After heating, the sample was washed with 2 L of distilled water for 12 min at 70 °C and immersed in acetone for 10 min. The samples were dried for 10 min and then further dried in a dry oven for 48 h at 60 °C before weighing. The levels of NDF and ADF digestibility were calculated based on the following equations [15,18]:NDF, ADF (%) = [dried residue − (filter bag weight × C)]/sample weight × 100.(2) Correction (C) = dried blank bag weight/original blank bag weight.(3) NDF, ADF (g) = sample weight (g) × DM (%) × NDF or ADF (%)/10,000.(4) NDF, ADF digestibility (%) = NDF or ADF (g) of zero time/NDF or ADF (g) of each time.(5)The nitrogen content (N%) was analyzed using an analyzer (KjeltecTM 8400, Foss, Denmark) with sulfuric acid and a catalyst. The crude protein (CP) content was calculated by multiplying N% by 6.25. CP digestibility (%) = 100 × [CP (g) of each time/CP (g) of zero time] × 100.(6)2.2. In Vivo Experiment2.2.1. Animals and DietsEight Korean native steers (average BW = 597.1 ± 50.26 kg; average age = 23.8 ± 0.12 months) were randomly and evenly assigned to two groups, control (without MO) and MO (3% MO supplementation on a dry matter basis), for 186 days. The animals were placed four steers per pen and fed concentrate and straw. The chemical and fatty acid compositions of the experimental diets are shown in Table 1. Water was provided ad libitum. Feed was provided once a day at 0.900 h. As mentioned previously, 3% MO was added to the diets by top dressing. The body weight of each steer was measured every 3 months prior to feeding using a digital weighing indicator (CAS Co. Ltd., Seoul, Korea). The feed intake (FI) was calculated after measuring the amount of feed remaining from the day before. Then, the following equation was used to calculate from pen data to individual data: FI/BW% = 1.2425 + 1.9218 × net energy for maintenance (NEm)−0.7259 × (NEm)2 [19].2.2.2. Blood Collection and AnalysisBlood samples were taken blood cell counting [20], and metabolite analysis [21] was performed from each cow via the jugular vein before feeding on the first day at 0, 90, and 180 days. In brief, blood was collected into EDTA-containing tubes (Becton Dickinson, Franklin Lakes, NJ, USA) and then subjected to a complete blood cell count test using an HM2 (VetScan HM2 Hematology System, Abaxis, Union City, CA, USA). For the serum samples, blood was collected in heparin-containing tubes (Becton Dickinson) and centrifuged at 3000 rpm for 15 min to extract the serum. The samples were then used to determine the blood metabolite concentrations using a chemical analyzer (Furuno CA-270, Nishinomiya, Japan).2.2.3. Carcass Traits and Sample PreparationAfter slaughtering, we further measured the effect of 3% MO supplementation on the carcass properties using the loin (longissimus thoracis) samples. After approximately 12 h of fasting, the steers were slaughtered following the commercial slaughtering procedure at a commercial slaughterhouse (Cheongju, Korea). Following 24 h of chilling, the carcass weight, yield grade (carcass weight, back fat thickness, and ribeye area) and quality grade (marbling score, meat color, fat color, firmness, and maturity) were evaluated by an official quality grader via the Korean Carcass Grading System of Korea Institute of Animal Products Quality Evaluation (KAPE, 2019). The quality grade evaluation was carried out on the surfaces of longissimus thoracis muscles in the 13th rib section.2.2.4. Meat QualityEach longissimus thoracis sample was cut and vacuum-packed into a polyethylene bag at 24 h postmortem and subsequently transported to the laboratory. Samples were stored in a chilled room at 4 °C overnight. The longissimus thoracis samples were sliced to observe the color, pH, cooking loss, shear force, and fatty acid content. The samples were allowed to bloom for 30 min.Color was determined using a colorimeter (CR-210, Minolta, Tokyo, Japan; illuminate C, calibrated with white plate, L* = +97.83, a* = –0.43, b* = +1.98). The L* (lightness), a* (redness), and b* (yellowness) values were measured on the surface of meat. All cthe onditions were assessed in triplicate.The pH of each meat sample was measured using a portable pH meter (HI98163, Hanna Instrument, Woonsocket, RI, USA), which was calibrated in buffers (pH 4.0 and 7.0) at room temperature. The pH probe was 20 mm deep into the samples. Measurements were made at three locations in each meat sample.The cooking loss of each sample was measured using a previously described procedure [22]. Each sample was initially cut and weighed and then packed in closed polyethylene bags and immersed in a temperature-controlled water bath (80 °C) until the core temperature reached 75 °C. Cooked samples were transferred to a cooled water bath until reaching equilibrium and then reweighed. The cooking loss was calculated as the difference in weight before and after cooking as follows:Cooking loss (/100 g) = [weight of raw sample (g) − weight of cooked sample (g)]/weight of raw sample (g) * 100.(7)The sample preparation to determine the shear force was similar to the cooking loss procedure. Meat samples 1 cm in diameter were cut and tested at room temperature with a texture analyzer (TA-XT2i, Stable Micro Systems, Surrey, UK). Each sample of meat was cut perpendicular to the direction of the muscle fibers. The conditions of the texture analysis were as follows: pre-test speed 2.0 mm/s, post-test speed 8.0 mm/s, load cell 25 kg, maximum load 2 kg, head speed 2.0 mm/s, distance 20.0 mm, and force 5 g. The maximum force at the calculated peak was determined.Digital color images of beef obtained from 8 steers were used for the image analysis of marbling characteristics. A mirror-type digital camera (D3200, Nikon Co., Tokyo, Japan) was used to photograph the marbling of each sample surface (at the 13th thoracic vertebrae). A strobe (Nikon Co.) was used to prevent irregular reflection on the surface of each sample. An image analysis was performed using the Beef analyzer II (Hayasaka Ricoh Co. Ltd., Sapporo, Hokkaido, Japan), following a previously reported method [23]. Digital color images were converted to binarized images to analyze the marbling characteristics. After binarization, marbling flecks were classified into two forms: big (>0.5 cm2) and small (0.01 to 0.5 cm2) flecks. However, the smallest particles (<0.01 cm2) were not used for the image analysis of marbling characteristics. After calculating the number of marbling flecks and the marbling area, these factors were used to measure the marbling area ratio, coarseness (C), fineness (F), and F/C ratio. The marbling area ratio, the C index, and the F index were calculated following a previously reported method [24].2.2.5. Fatty Acid Profiles of MeatThe lipid content of each sample was determined using a solvent mixture of chloroform:methanol (2:1, v/v) as described in a previous study [25]. FAME Mix standard (Sigma, St. Louis, MO, USA) was the standard used to determine the fatty acid content and fatty acid ratio. The measurement equipment used was a gas chromatograph/FID (7890B GC System, Agilent Technologies), and the analysis conditions followed those used in a previous study [26]: a Sp-2560 capillary column (dimensions: 100 m × 0.25 mm × 0.2 μm, film thickness); injection: split 30:1, heater 255 °C, pressure 32.64, total flow 39.5 mL/min, split flow 36.0 mL/min, injection volume 1.0 μL, helium 1.2 mL/min as the carrier gas; oven program: 70 to 100 °C for 5/min (Hold 2 min), 100 °C to 175 °C for 10/min (hold 40 min), and then 175 to 225 °C for 5/min (hold 40 min); detector: FID System, heater 260 °C (H2 flow: 40 mL/min, air flow: 400 mL/min). Conjugated linoleic acid, omega-3 (C18:3n3, C20:3n3, C20:5n3, and C22:6n3), and omega-6 (C18:2n6t, C18:2n6c, C18:3n6, C20:3n6, and C20:4n6) fatty acids were used as standards (Sigma).2.3. Statistical AnalysisThe results of the in vitro experiment were analyzed as a one-way ANOVA using the JMP pro 14 (SAS Institute, Cary, NC, USA). In the in vivo experiment, the JMP 7.0 (SAS Institute) program was used for all statistical analyses. Comparisons between two groups were evaluated using Student’s t-test. The significance level was declared at p < 0.05 and a tendency was declared at 0.05 ≤ p < 0.1.3. Results3.1. In Vitro ExperimentWe investigated the pH, ammonia-N and VFA contents and digestibility of ruminal fluid by adding various levels of MO (Table 2). Supplementation with 1% to 4% MO did not affect the ruminal pH. Moreover, there were no differences in the ammonia-N and VFA contents among the treatment groups. The digestibility of DM and CP were not different between the control group and all treatment groups (Table 3).Compared to the control, there was no difference in the digestibility of NDF and ADF with 1% to 3% MO, but the group supplemented with 4% MO showed decreased digestibility of NDF and ADF (p < 0.05). Considering these results, we determined 3% MO as the optimal dosage for the in vivo experiment.3.2. In Vivo Experiment3.2.1. Growth PerformanceThe growth performance is shown in Table 4. The average daily gain (ADG) was increased from days 1 to 90 (p = 0.007) in the 3% MO group. MO supplementation also showed a tendency to increase the ADG throughout the entire experiment (p < 0.1). There were no differences in the BW or FI between the two groups throughout the entire experiment period (total average; day 1 to 180; Table 4). As a consequence, the feed efficiency (FE) increased from days 1 to 90 and for the entire experiment period (p < 0.05) in the 3% MO group.3.2.2. Blood Parameter AnalysisBlood collected from veins of steers was analyzed to determine the hematological (Supplementary Table S1) and biochemical parameters (Table 5). For the hematological parameters, there were no significant differences between the two groups throughout the entire experiment (Supplementary Table S1). The biochemical analysis revealed increases in the levels of glucose (GLU; p = 0.041) and high-density lipoprotein-cholesterol (HDL-C; p = 0.033) in the 3% MO group on day 90 compared to the control. In addition, the total cholesterol (TCHO) and non-esterified fatty acid (NEFA) contents on day 90 tended to be higher (p < 0.1) in the 3% MO group compared with the control group. However, there was no difference in the triglyceride (TG) content between the two groups throughout the entire experiment period.3.2.3. Carcass TraitsCarcass traits were evaluated over 180 days (Table 6). There were no differences in meat yield or quality traits between the two groups.3.2.4. Meat QualityThe meat quality results are shown in Table 6. Compared with the control, the 3% MO group showed a tendency for the pH of the meat to increase (p < 0.1). The meat from the 3% MO group had a higher (p < 0.05) lightness (L*) and tended to be higher (p < 0.1) in yellowness (b *) compared with the control. There was no difference in redness (a *). On the other hand, the shear force was lower (p < 0.05) for steers fed 3% MO than for the controls. Supplementation with 3% MO did not affect the cooking loss or marbling characteristics.3.2.5. Meat Fatty Acid ProfileThe fatty acid composition in the subcutaneous and intramuscular fat of the two groups was investigated (Table 7 and Table 8). Compared to the control group, the concentrations of C15:0, C18:1n9t, and C20:1 fatty acid increased (p < 0.05) in the subcutaneous fat of the 3% MO treated group. Moreover, the omega-3 fatty acid (C22:6n3) content increased (p < 0.001), and the omega-6 fatty acid (C20:4n6) content decreased (p = 0.013). The 3% MO group showed a tendency for C20:2 to decrease (p < 0.1).As shown in Table 8, the concentrations of C17:1 and C18:3n6 fatty acids and omega-3 fatty acids (C18:3n3 and C20:3n3) increased (p < 0.05), and the concentrations of C15:0, C16:1, C18:1n9t, and C20:5n3 fatty acids tended to increase (p < 0.1) in the intramuscular fat of the 3% MO group. Also, a decrease in C22:0 fatty acid was observed (p < 0.05). As a result, the level of omega-3 fatty acid increased (p < 0.5), and the omega-6 / omega-3 ratio tended to decrease (p < 0.1).4. Discussion4.1. In Vitro ExperimentRuminal fluid pH and the ammonia-N content were not affected in any of the treatment groups (Table 2). Benchaar et al. [7] reported similar results when they added 2%, 3%, and 4% linseed oil to dairy cows. Benchaar et al. [7] suggested that the level of LO supplementation was not high enough to have an effect, and for that reason, there were no changes in ruminal metabolism. Fat supplementation has toxic effects on cellulolytic microbials and digestion, which leads to microbial changes [27]. These microbial changes shift the fermentation pattern to propionate, while the concentrations of acetate and butyrate decrease [11]. In this study, the total VFA and acetate contents tended to decrease in 4% MO, but there was no significant difference in the amount of VFA between the control and MO treatment groups (Table 2). Palmquist et al. [28] reported that the rumen-protected fat maintains normal rumen fermentation. In agreement with a previous study, we suggest that there was no difference in the VFA content, as there was no change in the fermentation pattern [28]. On the contrary, the digestibility of NDF and ADF decreased following supplementation with 4% MO (Table 3; p < 0.05). According to previous studies [27], fat supplements coat the rumen microbials, resulting in reduced digestion. In addition, microbials attach to the feed particles to digest fiber, but when fat supplements are added, the fat attaches to the feed particles and interferes with the activity of microbial enzymes [27]. Considering these findings, we considered that with 4% MO, the oil from the MO was eluted, negatively affecting ruminal microbials and resulting in reduced NDF and ADF digestibility (p < 0.05). Therefore, we suggest that supplementation with 4% MO led to an overdose of MO that negatively affected fiber digestion and that 3% MO is the optimal dosage for rumen fermentation and digestibility.4.2. In Vivo ExperimentFat supplements increase the energy density of feed and improve the feed efficiency, but the increased energy density causes a decrease in the FI [12]. Moreover, the decreased FI leads to decreases in animal performance parameters, such as the ADG and the FE. Previous studies noted decreases in the FI and the ADG when steers were fed fat supplements [13]. As mentioned earlier, one of the reasons for the decrease in feed intake is the toxic effect caused by fat supplements coating the rumen microbials and feed. However, in the present experiment, ADG and FE increased (Table 4; p < 0.05). Doreau and Chilliard [29] noted that the absence of a negative effect on rumen digestion may be due to the presence of hydrogenated Ca salts. Thus, we suggest that because the MO is coated with hydrogenated palm oil, it did not cause decreases in FI, and ADG, whereas the FE was increased by the increased energy density from the MO. This experiment was performed from March (spring) to September (early fall), so there were seasonal differences. The average daily gain only increased from days 1 to 90 (March to June; THI 58) and not days 91 to 180 (June to September; THI 73). Animal performance depends on a variety of factors such as the FI, environmental conditions, and management system. Therefore, the reasons for observing inconsistent results might be the influence of seasons between days 1 to 90 and days 91 to 180. If there is no adverse effect on rumen fermentation, animals can receive energy from fat supplementation, and the increased energy intake has positive effects on performance.Hematological parameters were measured for the purpose of investigating the stability of experimental supplementation in this study (Supplementary Table S1). There were no differences in the white blood cells, red blood cells, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, or mean corpuscular hemoglobin concentration between the two groups. These results indicate that feed with 3% MO had no negative effect on the immunological response in steers. The results of the biochemical analysis are shown in Table 5. Arave et al. [30] reported that the GLU content, which changes in response to energy intake, increased in serum when animals were fed high-energy diets. Consistent with this experiment, the GLU content increased (p < 0.05) on day 90 in the 3% MO group. In addition, Arave et al. [30] reported that the blood cholesterol concentration reflects the overall nutritional intake condition, similar to the GLU content. As expected, on day 90, the TCHO tended to be higher (p < 0.1) in the 3% MO group. The addition of fat is known to increase the blood cholesterol content. In addition, the results of this experiment show that the increased HDL-C content contributed to TCHO. The content of non-esterified fatty acids also tended to increase (p < 0.1) on day 90. Dryden and Marchello [31] reported that an increase in dietary fat was associated with an increase in serum NEFA. Relling and Reynolds [32] also suggested that supplementation with fat increased the NEFA due to the increase in fat absorption. Consistent with previous studies [12], the HDL-C increased (p < 0.05) in the 3% MO group. However, these changes were only observed on day 90. Whether these are the results of a reaction due to other factors needs further research.There were no differences in carcass yield traits or quality traits (Table 6). In agreement with this experiment, Suksombat et al. [8] showed that dietary palm oil and linseed oil, regardless of the type and level of supplementation, do not affect the carcass traits. Also, these results are similar to those of Castro et al. [33], who reported no difference in carcass traits following supplementation with palm oil, olive oil, or soybean oil in Blonde D’ Aquitaine steers. However, Song et al. [34] found that supplementation with soybean plus fish oil tended to increase (p = 0.08) the meat color score, and supplementation with soybean oil plus monensin tended to increase (p = 0.07) the texture score. These inconsistent results might have occurred due to the difference between the types of supplement and species used in this experiment and previous studies.The pH of meat is the basis for meat quality evaluation, because it affects the quality of meat, as assessed by factors such as the tenderness, water holding capacity, and meat color. The pH values of the control and 3% MO groups were 5.4 and 5.5. Samples from the group supplementation with 3% MO tended to have higher pH values compared with those from the control group (Table 6). The optimal pH of meat is considered to be 5.4 to 5.8 [35]; thus, in this study, meat from the control and MO supplementation groups had normal pH values. Meat color is one of the factors influencing consumer purchase, and the color values for L* (Lightness), a* (redness), and b* (yellowness) are used to measure meat color. The variation ranges are 3 3 to 41, 11.1 to 23.6, and 6.1 to 11.3 (L*, a*, b*, respectively) [36]. In this experiment, the L* of meat decreased (p = 0.042), and the b showed a tendency to increase in the 3% MO group (Table 6). There was no difference in a * values between the two groups. Considering the variation range of each value, the values for L*, a*, and b* in the MO group were within the normal ranges. Meat color depends on age, weight, and nutritional status. Fat supplementation is associated with problems of reducing elements of the meat quality such as taste, meat color, and oxidation resistance, because an increase in the degree of unsaturation promotes oxidation. Fat oxidation changes the color of meat due to red oxymyoglobin being converted to brown metmyoglobin, and this reaction is usually accompanied by the production of a foul odor. In this experiment, however, we proved that 3% MO did not negatively affect the meat color, because the meat color factors were within the accepted ranges. Meanwhile, there was no difference in cooking loss between the two groups (Table 6). Consistent with this study, Wistuba et al. [37] reported that there was no effect on cooking loss in Angus crossbred steers fed 3% fish oil. Pukrop et al. [38] also reported no difference in cooking loss in Angus-Simmental crossbred steers fed 1 g/steer essential oils. The tenderness of meat is considered to be related to the shear force. Kook et al. [39] reported that shear force decreased in early finishing Korean cattle bulls and steers fed 5% fish oil (p < 0.001). They found that fish oil affected the fatty acid composition of meat, and the change in fatty acid composition altered the physical properties of meat, which affected the shear force [39]. It has been reported that changes in the fatty acid composition through supplementation or feed affect the tenderness of meat. In addition, the melting point of fatty acids affects the hardness of adipose tissue [40]. Therefore, the changes in fatty acids composition following MO supplementation may have decreased the shear force (p < 0.05) in the 3% MO group (Table 6). The marbling characteristics are related to the palatability. It has been suggested that the shear force with F of marbling is lower than the C of marbling [41]. Nakahashi et al. [42] investigated the concentration of monounsaturated fatty acids (MUFA) according to the size of marbling particles in meat (small < 0.4 cm2, medium 0.4 to 2.0 cm2, large > 2.0 cm2). They showed that larger marbling particles were associated with a greater MUFA concentration. In this experiment, there was no difference in marbling characteristics or the level of MUFA in meat in the 3% MO group. However, there have not been many studies on the relationship between fat supplementation and marbling characteristics, and therefore, further research is needed.We further determined the fatty acid compositions of subcutaneous- and intramuscular fat in steers (Table 7 and Table 8). In terms of the fatty acid composition in subcutaneous fat (Table 7), no differences in saturated fatty acids (SFA), MUFA, and polyunsaturated fatty acids (PUFA) or the omega-6/omega-3 ratio were found. However, the most unexpected result of this experiment was an increase (p < 0.05) in docosahexaenoic acid (DHA; C22:6n3) in the 3% MO group. This result is supported by a previous finding reported by Kim et al. [43]. They found that the DHA level was increased (p < 0.05) with dietary whole flaxseed (WFS), regardless of the level. However, there is no exact explanation for this finding. This might be due to an increase in α-linolenic acid (ALA; C18:3n3) following 3% MO supplementation, and the promotion of an increase in DHA in the subcutaneous fat by ALA. The MO supplement contains a high concentration of ALA, a precursor to DHA. The metabolic pathway of omega-3 and omega-6 fatty acids proceeds step by step by adding a double bond through desaturases and increasing the length of carbon through elongases. Long-chain omega-3 fatty acids are converted from ALA to eicosapentaenoic acid (EPA; C20:5n3) through a process of elongation by elongases and desaturation by desaturases, such as Δ5-desaturase and Δ6-desaturase. Then, eicosapentaenoic acid is converted to DHA through elongation, unsaturation, and β-oxidation process [44]. However, except for these findings, no specific effect of MO on subcutaneous fat was shown. Generally, the effect of supplementation on subcutaneous fat is comparatively small in steers compared with bulls and cows [43].The composition of fatty acids in the intramuscular fat of steers was investigated (Table 8). The intramuscular fat content of steers fed 3% MO had a higher C18:1n9t concentration than the control group (p < 0.05). This result might be related to the presence of stearic acid (C18:0; SA) in the MO supplement. There is a desaturase that acts on SFAs and converts SFAs into MUFAs, and this desaturase converted SA into C18:1n9t [44]. Supplementation with 3% MO increased (p < 0.05) the contents of C18:3n6 and ALA in the intramuscular fat. Similar results have been reported by Kim et al. [43] in steers fed 10% or 15% WFS. Additionally, many previous studies [45] have reported an increase in ALA when linseed was added. Linoleic acid (LA; 18: 2n6) and ALA are essential fatty acids that cannot be synthesized and must be consumed as feed. When fat supplementation was conducted, as mentioned earlier, many shifts in the fatty acid composition occurred through enzyme systems, such as desaturation and elongation [44]. As a result of desaturation and elongation, the fed LA was unsaturated with ∆6-desaturase, resulting in an increase in C18:3n6. In addition, ALA increased the EPA content through ∆6-desaturase, elongase, and ∆5-desaturase in intramuscular fat. Supplements containing ALA always elevated EPA [46]. However, DHA, which existed in the subcutaneous fat, was not observed in the intramuscular fat (Table 8). In general, there is a metabolic difference between subcutaneous fat and intramuscular fat in ruminants. Smith and Crouse [47] suggested that the main precursor of subcutaneous fat is acetate and the main precursor of intramuscular fat is likely to be glucose. Although ALA is a precursor of long-chain omega-3 fatty acids, there is limited conversion. Noci et al. [48] reported that conversion is affected for a variety of factors. In short, it is due to the activity of low desaturase and competition with other long-chain fatty acids. When ALA is converted to DHA, it is controlled by a complex enzyme system that also acts on other long-chain fatty acids. Thus, due to the low conversion efficiency, the conversion from ALA to DHA was not observed consistently. Considering all of these results, the level of omega-3 fatty acid was elevated (p < 0.05) and the omega-6/omega-3 ratio tended to decrease (p = 0.05) in the 3% MO group. Overall, the supplementation of steers with 3% MO caused many shifts in fatty acid composition. In general, intramuscular fat is affected by several internal and external factors such as age, gender, species, castration, temperature, and nutrition. Therefore, further studies are needed to clarify the effects of these factors on intramuscular fat in steers.5. ConclusionsFeeding with 3% MO, a rumen protected fatty acid enriched with omega-3 fatty acid, improved the growth performance of Korean native steers and decreased the shear force produced by the muscles. By increasing the level of omega-3 fatty acids in intramuscular fat, MO supplementation tended to decrease the omega-6/omega-3 ratio and improve the meat quality. In conclusion, it was demonstrated that supplementation with 3% MO increased the productivity of Korean native steers by producing healthful beef enriched with omega-3 fatty acids.

animals : an open access journal from mdpi

10.3390/ani11040951

PMC8066058

Methane is a potent greenhouse gas. It is 80-times more effective at heating the earth than carbon dioxide over the first 20 years following release into the atmosphere. Ruminant animals have diverse microbial populations in their stomachs that employ anaerobic fermentation to digest feed. Methane is belched into the atmosphere as a by-product of the digestive process. This gut, or enteric methane, primarily from cattle, but also sheep and goats, contributes 30% of the methane released into the earth’s atmosphere each day, and is more than any other single methane source. A major reduction in methane emissions from ruminants is crucial to preserve ecosystems on the planet. Various strategies to reduce enteric methane emissions in farm operations are reviewed to quantify their mitigation potential, determine their impact on animal productivity and likelihood of adoption. Two feed supplements, a commercial product, 3-NOP (Bovaer®), and the seaweed, Asparagopsis, can reduce methane emissions by 40+% and 90%, respectively, with associated increases in animal productivity and no adverse effects on animal health or product quality. The rumen microbial population can also be changed to provide long-term intergenerational reduction in methane emissions, if treated herds remain isolated from non-treated animals.

Anthropomorphic greenhouse gases are raising the temperature of the earth and threatening ecosystems. Since 1950 atmospheric carbon dioxide has increased 28%, while methane has increased 70%. Methane, over the first 20 years after release, has 80-times more warming potential as a greenhouse gas than carbon dioxide. Enteric methane from microbial fermentation of plant material by ruminants contributes 30% of methane released into the atmosphere, which is more than any other single source. Numerous strategies were reviewed to quantify their methane mitigation potential, their impact on animal productivity and their likelihood of adoption. The supplements, 3-nitrooxypropanol and the seaweed, Asparagopsis, reduced methane emissions by 40+% and 90%, respectively, with increases in animal productivity and small effects on animal health or product quality. Manipulation of the rumen microbial population can potentially provide intergenerational reduction in methane emissions, if treated animals remain isolated. Genetic selection, vaccination, grape marc, nitrate or biochar reduced methane emissions by 10% or less. Best management practices and cattle browsing legumes, Desmanthus or Leucaena species, result in small levels of methane mitigation and improved animal productivity. Feeding large amounts daily of ground wheat reduced methane emissions by around 35% in dairy cows but was not sustained over time.

1. IntroductionAnthropomorphic production of greenhouse gases is raising the temperature of the earth to levels that are threatening the sustainability of ecosystems on the planet [1]. Temperatures at regions on earth have risen since the 1960s by more than 1.2 °C and annual soil moisture is estimated to have declined by 20 to over 40% in the cropping regions of southern Australia [2]. Atmospheric concentrations of the major greenhouse gases, carbon dioxide and methane, have risen since 1950 from 350 to 410 ppm (28%) and 1100 to 1875 ppb (70%), respectively [3]. Methane is 28 times more potent than carbon dioxide as a greenhouse gas over 100 years and 80 times more potent over 10–20 years from release [4]. Enteric methane from the microbial fermentation of plant material by ruminant animals, primarily cattle, contributes 30% of methane released into the atmosphere, which is more than any other single source [5]. Enteric methane is the largest contributor (40%) to global greenhouse gas emissions from livestock supply chains [6], contributing 6% of total anthropogenic greenhouse gas emissions [7,8]. Ruminants also produce a substantial amount of carbon dioxide (CO2), with a methane:CO2 ratio of approximately 4:1 [9], making a total contribution of ruminants to anthropogenic greenhouse emissions of 8%. Consequently, there are demands from sections of the community for major reductions in the consumption of red meat to reduce the number of ruminants and amount of methane they release [10,11]. However, the Food and Agriculture Organization [12] projects that demand for red meat will continue to increase at the rate of around 1.5% per year to meet the growing population and rising living standards in developing countries. An alternative mitigation approach, rather than reducing red meat consumption or ruminant numbers, is to reduce methane emissions from ruminants.Australia introduced the Emissions Reduction Fund (ERF) in 2016, where primary producers are paid for reducing greenhouse gas emissions provided an approved Method for methane reduction (or any other greenhouse gas) has government acceptance [13]. Methane emissions from ruminants reduce the energy from feed available for production, with losses to the animal of 3–12% of digested energy [14,15]. Reduction in ruminant methane emissions would decrease national greenhouse gas emissions and increase energy available to livestock for productivity.As a forerunner to the Australian red meat industry establishing a goal to be carbon neutral by 2030 [15], researchers examined a wide range of potential methane mitigation strategies for ruminants in the National Livestock Methane Program (NLMP) coordinated by Meat and Livestock Australia [16]. The research incorporated numerous methane reduction strategies with some identified from previous reviews [17,18]. Results from this Australian research, complemented by published results from other sources, are used in this review to identify those strategies that are considered to have the highest potential for reducing methane emissions from ruminants.2. Materials and MethodsMitigation strategies examined in NLMP included: (i) the role of genetic selection; (ii) use of various feed supplements; (iii) potential role of anti-methanogenic forages; (iv) potential for methane emission reductions from a detailed understanding of rumen function; and (v) best pasture management practices. The potential role of vaccination against rumen Archaea, including peptide sequences, the commercially developed 3-nitrooxypropanol (3-NOP) [19] and biochar were not included in NLMP research, but were evaluated.The evaluation involved assessment, for each potential mitigation strategy, of: (i) status of current knowledge; (ii) likely mitigation potential for individual animals and across Australia; (iii) impact on animal productivity; (iv) cost of and barriers to implementation; (v) chances of research success and its cost; (vi) likely delay before being implemented on production enterprises; and (vii) opportunity for ruminant enterprises to sell ‘carbon credits’. Steps (iv) to (vi) were based on subjective assessments made by 17 people comprising the NLMP investigators, members of the Investor Advisory Group, with a representative from each organization contributing funds to the program, and the NLMP management team [20].These assessments have been used recently by Davison et al. 2020 [20] to determine the financial outcome for each mitigation strategy when applied to the Australian ruminant industry based on livestock types and numbers provided by Mayberry et al. [15]. Financial benefits from 2020 to 2030 were determined as the discounted additional income from changes in productivity and carbon credits based on the ERF price of AUD $16.14 per t of carbon dioxide equivalents, assuming methane to be 28 times more potent as a greenhouse gas than carbon dioxide. The discounted costs for purchasing and applying each strategy were then subtracted from the additional income to provide the net financial change, which was multiplied by the assumed adoption rate to provide an assessment of national monetary benefit to the Australian ruminant industries. Consequently, the methane mitigation strategies were ranked in declining order for likely financial benefit to the Australian cattle and sheep industries (Table 1).3. Methane Mitigation Strategies3.1. Red Macro Algae as a Feed SupplementThe red marine alga, Asparagopsis taxiformis, when collected in the filamentous tetrasporophyte phase, air-dried, ground and added at 2% of substrate organic matter, reduced methane emissions by up to 99% without depressing substrate digestibility or volatile fatty acid production in laboratory rumen fermentation cultures [21]. When dried red alga was fed to sheep for 75 days at five concentrations from 0 to 3% of organic matter, methane emissions declined linearly to over 80% at the highest dose [21]. Methane emissions did not increase over time, suggesting the rumen methanogen population did not adapt to the alga. A recent experiment with penned cattle showed that the inclusion of only 0.2% of feed organic matter as dried Asparagopsis reduced methane emissions by 98% and increased weight gain by 42%, without negative effects on feed intake or rumen function [22]. There were no changes in meat quality or detectable residues in the meat. Similar declines in methane production from 55% to 80% have been found in dairy cows fed 0.5% dry matter of bioactive Asparagopsis taxiformis [23]. Similarly, a decline in methane of 67% was observed when Asparagopsis armata was fed a 1% inclusion in a diet for lactating dairy cows with no residues observed in milk [24].Stefenoni, et al. [23] observed a 5-fold increase in iodine and an 8-fold increase in the iodine and bromide concentration in milk from dairy cows fed total mixed ration diets containing 0.5% Asparagopsis taxiformis collected from Portugal compared with the milk from cows fed the control diet. Iodine concentrations are high in most seaweeds fed to animals. Although these higher-than-normal concentrations in milk could be detrimental to the health of young children, Stefenoni, et al. [23] believe this is not of major concern due to the normal mixing of milk from different farms. Stefenoni, et al. [23] did not detect significantly higher concentrations of bromoform in milk with 0.5% Asparagopsis, but it was detected in milk for the first 10 days after introduction at rates of 6.5% or more in a supplement to a dairy cow diet by Muizelaar, et al. [25].Li, et al. [21] and Stefenoni, et al. [23] suggest that the palatability of diets containing Asparagopsis may be reduced, particularly at higher concentrations in the diet. Muizelaar, et al. [25] found dairy cows offered diet supplements with 6.5% or more Asparagopsis regularly refused feed and selected against these feeds. Previous studies have shown bromochloromethane can reduce methane emissions in cattle and goats by more than 90% without affecting feed intake [26,27,28]. The observed reduction in feed intake in several experiments with diets including Asparagopsis have been suggested to be due to the extremely high mineral content of the seaweed [24].There are two reports of some animals showing changes in rumen papillae morphology and rumen ulceration with either prolonged (75 days) feeding [21] or feeding for shorter periods at 6.5% or greater of Asparagopsis in the diet supplement [25]. The diets used by Muizelaar, et al. [25] also contained large amounts of wheat that can cause rumen ulceration [29].The Asparagopsis alga contains approximately 0.22 mg/g dry matter (DM) of 22 different halogenated metabolites but varies between strains and locations [30,31]. The concentrations of the bioactive compounds depend greatly on the drying procedures, with freeze drying being most effective [23]. The methanogenic activity of Asparagopsis declines over four months when incorporated into total mixed rations [23].Halogenated methane analogues, such as bromochloromethane, inhibit methane production by reacting with reduced vitamin B12, which inhibits the cobamide-dependent enzyme methyl-coenzyme (CoM) reductase step in methanogenesis in a similar process as 3-nitrooxypropanol (3-NOP, see Section 3.2). The experiments reported have been conducted under pen conditions with diets containing grains or being high fiber, but pelleted [21]. The longest experiment was 90 days [22] and the potential impact of changes in vitamin B12 within the animal should be investigated over years if Asparagopsis is to be used in breeding cows.In summary, the research conducted to date on Asparagopsis suggests that low concentrations of 0.5% or less of highly bioactive bromoform compounds in the diet can reduce methane emissions by up to 90% without detrimental effects on feed intake or product quality. Further research is needed to develop methods for decreasing the decline in bromoform activity when incorporated into diets. Research is also needed to show that the supplement can be fed to rangeland animals through lick-blocks or other methods to prove applicability for reducing methane emissions from rangeland and grazing breeding herds and flocks. If such applications are successful, Asparagopsis supplementation could be used across the whole of the Australian ruminant population.Asparagopsis used in experiments has been collected manually in the wild. This process is expensive and not practical for commercial application. Research at James Cook University, Townsville, Australia, has measured variation in the concentrations and types of halogenated metabolites produced by algae due to genetic strains, sea water temperature and nutrient availability [31]. Asparagopsis can be cultured on ropes similar to the commercial culture of mussels and other macro-alga species. The alga can be grown in association with other aquaculture enterprises and could be valuable for reducing pollution from these industries. Due to the enormous potential of Asparagopsis to virtually eliminate methane emissions from ruminants and increase productivity, a company called Sea Forest has commenced commercial production of the alga in ponds in Tasmania [32]. A price for Asparagopsis being sufficiently low for practical methane mitigation in ruminant industries appears highly achievable. An ERF method needs to be developed if ruminant producers are to be paid for reductions in methane emissions. First practical application of Asparagopsis should occur within a year for feedlot and dairy animals, but will be longer for grazing stock.3.2. 3-Nitrooxypropanol (3-NOP) as a Feed Supplement3-nitrooxypropanol (3-NOP) and the ethyl variant, ethyl-3-nitrooxypropanol, were developed by DSM Nutrition Products Ltd. (Kaiseraugst, Switzerland) [33]. The compounds, bind to the active site of the enzyme methyl-coenzyme reductase that catalyses the last step in the reduction of CO2 to CH4 by the hydrogenotrophic methanogenic Archaea [34]. A potential advantage of 3-NOP over Asparagopsis is that it does not alter vitamin B12 metabolism. Laboratory and animal experiments show a significant reduction in methane when 3-NOP is added to the diet [19,35,36,37,38]. Methane emissions are reduced linearly in sheep, cattle and dairy cows as the proportion of 3-NOP is increased up to 2.8% of the diet [39].Most animal experiments show a reduction in methane emissions of 8–30% except for the experiment of McGinn, et al. [40], where the reduction was 70%. The magnitude of the reduction in animal experiments is less than in fermentation assays where it frequently reaches 95%. The discrepancy between the types of experiments could result from the high volatility of the compound or a high rate of degradation within the rumen, which may be associated with the method of feeding in the animal experiments. With effective administration of the 3-NOP compound, methane emission reduction may exceed 40%.There is no evidence 3-NOP reduces feed intake or digestibility of the diet and can be associated with an increase in productivity [39]. Hristov, et al. [19] reported that feeding 3-NOP increased body weight gain in lactating dairy cows, with no effect on milk production, but milk protein content increased. An increase in energy available to the animal would be expected [41] because hydrogen concentrations in the rumen are known to rise [39].3-NOP may be available for all production situations because of the low dose rate required. An effective dose appears to be around 2% of the diet. The compound must be continually available in the rumen to be effective, which means it is best provided in total mixed rations to feedlot cattle and dairy cows. 3-NOP could be provided in lick-blocks or other supplements for grazing animals, but its effectiveness would depend on the frequency animals consumed these products.DSM Nutrition Products is continuing research into feeding protocols for 3-NOP and is currently seeking registration for its use under the name Bovaer® in ruminant diets in various countries around the world [33]. Bovaer is likely to be an important product for reducing enteric methane around the world, with its uptake depending on its price and whether there are payments for carbon reduction.3.3. Manipulating the Rumen Microbial PopulationGlucose from cellulose and starch consumed by ruminants is the main substrate fermented by rumen microorganisms in the rumen. Glucose is degraded by five competing biochemical pathways in the rumen, which produce different amounts of methane and convert different quantities of glucose energy to volatile fatty acids [42]. The most efficient pathway converts 93% of glucose energy to volatile fatty acids, with the production of no methane and high concentrations of propionate, whereas the least efficient pathway converts only 62% of glucose energy to volatile fatty acids and produces one mol of methane for each mol of glucose fermented with high concentrations of acetate. The other three pathways have an efficiency of conversion of glucose to volatile fatty acids from 72% to 86% and produce from 0.66 to 0.25, respectively, mol of methane per mol of glucose fermented.Each of these fermentation pathways is predominantly performed by specific microbial species with only trace amounts of products from other pathways [20]. Methane mitigation strategies that substantially reduce methane emissions such as bromochloromethane, Asparagopsis or 3-NOP alter microbial populations to favour those that produce lower amounts of methane [43].Abecia, et al. [43] and Meale, et al. [44] have demonstrated that if dams and their progeny are treated with a strong anti-methanogenic agent for up to two months after birth and the progeny remain isolated from other animals that have not been treated, the changes in microbial population that favours increased propionic acid production and reduced methane emissions can persist for up to a year after the treatment ceased. Abecia, et al. [43] showed methane emissions were 33% lower and live weight gain 18% higher four months after treatment for goats with the altered microbial population.Meale, et al. [45] fed 3-NOP to one group of calves for 3 weeks post weaning and then ceased the supplement. Methane emissions were 10.4% lower in the treated group than the control at the end of the 3-week treatment. The difference in methane emissions between the previously treated group of animals and the control group was maintained at between 11.9% and 17.5% for up to 47 weeks after the treatment ceased. Distinct microbial populations remained evident between the treatment and control groups for the entire experiment.The possibility generated from these experiments is for whole herds with desired rumen populations to be created and maintained through generations provided they are isolated from animals with different rumen populations [20]. The magnitude of the depression in methane emissions following cessation of the treatment is likely to be related to the size of the initial reduction. Asparagopsis supplements can reduce methane emissions by over 90% and, even allowing for some reversion in the microbial populations, a continuing reduction for long periods after cessation of the supplement of 30% would appear reasonable to assume. The strategy of manipulating rumen microbial populations should have practicality once the Asparagopsis or 3-NOP supplements are commercially available. However, further research is required to demonstrate the feasibility of maintaining isolated herds with modified rumen microbial populations and to demonstrate the magnitude of the reduction in methane emissions.3.4. Tropical Legumes as Plantation Forage or Feed SupplementsLeucaena is a legume shrub grown in many tropical and subtropical regions of the world including the northern coastal environment of Australia, central Africa, Colombia, Mexico and Indonesia. Leucaena can be browsed and has the potential to be harvested, dried and included in rations for feedlot cattle or other ruminants. Leucaena contains high molecular weight condensed and hydrolysable tannins which bind to rumen microbes, reduce the population of Archaea and reduce methane formation [46]. However, Leucaena also contains the toxic amino acid mimosine and its breakdown product, 3-hydroxy-4-(1H)-piridon, which reduces growth rate, causes hair loss and ulcerations in the mouth and stomach [47]. These toxic products limit the acceptable intake of Leucaena to no more than 40% of the diet.Cattle grazing Leucaena plantations with Rhodes grass or naturalized pasture showed a substantial increase in growth rate and a reduction in methane compared with cattle grazing pasture alone [48,49]. Similar improvements in cattle performance with reductions in methane emissions have been observed in Colombia [50] and Mexico [51,52].Piñeiro-Vázquez, et al. [51] replaced a low-quality grass with up to 80% Leucaena in a pen experiment with cattle and found a linear fall to 60% of pretreatment enteric methane emissions. Browsing cattle on naturalized pasture appear to consume around 20% to 40% of their diet as Leucaena [53], which results in a reduction in methane emissions of 25% to 35%.Leucaena may provide an alternative to silage or cotton seed in feedlot rations, but the effect of drying Leucaena on its ability to reduce methane emissions from ruminants is unknown. Based on feeding fresh Leucaena, a 20% inclusion in a feedlot diet would be expected to reduce methane emissions by 7% at the same feed intake [54].Procedures for establishing Leucaena plantations are well known. The plant is generally not grazed until 18–24 months after establishment. The agronomy and costs for Leucaena planting and establishment are known. Developing an ERF method, so that producers can claim carbon credits when adopting a Leucaena feeding system, requires research to provide an accurate algorithm for predicting methane reduction and performance of cattle. The experiments could also provide near infrared (NIR) calibrations to estimate the proportion of Leucaena in the diet of an animal based on a scan of faeces [55]. Research is also needed to assess the effectiveness of dried Leucaena, followed by work with animals in feedlots.Another tropical legume Desmanthus also contains condensed tannins that have been shown to reduce methane emissions in laboratory studies by 26% compared with a Rhodes grass control [56]. One experiment with cattle showed that including 31% Desmanthus cultivars in the diet replacing Rhodes grass resulted in a 10% decrease in methane emissions per kg of dry matter intake [57]. Desmanthus has an advantage over Leucaena because it can grow in semi-arid regions and has a potentially wider range of environments than Leucaena [58] where it can be incorporated into cattle production systems. However, further research is required to determine its potential for methane mitigation under grazing. As for Leucaena, an ERF method needs to be developed for Desmanthus if livestock producers are to obtain payment for methane mitigation.3.5. Grape Marc as a Feed SupplementGrape marc consists of skins, seeds, stalks and stems remaining after grapes have been pressed for wine. Grape marc contains condensed tannins, high concentrations of oils and tartaric acid, all with potential to reduce methane emissions in ruminants [59]. Tannins, particularly extractable tannin with shorter polymer chains and lower cis/trans ratio, were the most effective for reducing methane production by up to 50% and with little effect on total digestion in continuous 10-day laboratory rumen fermentation assays [60,61].Grape marc has high fiber and low metabolisable energy (ME) content because of its high stalk and stem content [59]. Effectiveness of grape marc for reducing methane, without a negative impact on productivity, depends on the ME content of the marc relative to the diet ingredient it is replacing. Substituting 36% of a lucerne hay diet with marc pellets of similar ME content in feed for dairy cows in late lactation reduced methane emissions by approximately 20%, with little effect on milk yield [59]. However, when either white or red grape marc replaced fresh pasture with high ME content in diets for dairy cows in early lactation, milk yield was reduced by approximately 10%, while methane emissions were depressed by 15% with no differences between marc type [62]. The authors of the latter study conclude that the reduction on milk yield was related to the lower ME of grape marc, but the reduction in methane emissions was most likely due to the higher crude fat and lignin content reducing rumen fermentation than a direct effect of tannins. However, there is strong evidence from in vitro rumen fermentation studies that smaller sized extractable tannins reduce methane emissions without affecting the efficiency of fermentation by rumen microbes [62].Replacing 30% of an oaten hay diet with grape marc of similar ME content reduced methane emissions from sheep by 10% without affecting animal productivity [63]. An Angus cattle feedlot experiment, where 20% of maize silage was replaced with marc, reduced emissions by approximately 10%, but growth rate was reduced by 25% due to lower energy intake [64].These results suggest there is limited application for feeding grape marc to reduce methane emissions, without detrimental impacts on productivity. Grape marc sourced directly from the winery has no cost except for loading and transport. Costs increase with processing including ensiling, steam distilling, roller milling and drying for longer storage. Ensiling grape marc is the most practical method for storage on farms. Effective use of grape marc would be restricted to animal enterprises close to source because of the cost of transport. The best option for exploitation appears to be for sheep enterprises in southern Australia in proximity to vineyards during the summer-autumn feed-gap when pasture quality is poor or to sheep and cattle when feed supply is low.3.6. Genetic SelectionGenetic variation in methane emission is inherent between individual ruminant animals [65]. Lower methane emissions may be manifest through a lower methane production for the same feed intake (i.e., residual methane production, RMP) and/or a lower methane associated with a lower feed intake at the same growth rate (i.e., low residual feed intake, RFI). The reduction in methane emissions in selected animals appears to be due to smaller rumen volumes, increased outflow rate of digesta and reduced fermentation in the rumen [66]. Heritability for methane emission, expressed as methane per unit of feed intake, is moderate (0.13–0.35) in cattle, sheep and dairy cattle [65,66,67,68,69]. Although heritability for methane emissions is moderate, the genetic variation between animals is small relative to other traits such as growth rate or milk production [70].The impact of genetic selection on reducing methane emissions depends on the relative selection pressure placed on lowering methane emissions compared with other traits such as growth and reproduction. Changes to national methane emissions depends also on the speed the genes are passed through national herds. The potential for genetic improvement through direct selection for methane traits is limited at 0.2–0.4% per year, cumulating in a reduction of only 4–8% over 20 years for beef cattle and sheep [70]. The rate of genetic gain for methane mitigation is higher for dairy cattle because the widespread use of artificial insemination results in fewer sires used across the industry. Greater selection pressure can be applied in dairy sire selection than in the more extensive beef and sheep industries. Selection models for dairy cattle suggest methane reductions as large as 20–26% over 10 years are theoretically possible, but only at the expense of a 6 to 18% decrease in genetic gains for production traits [71]. Other models suggest the reduction in methane emissions from dairy cows would be only 3% over 10 years if there is genetic improvement in milk production traits [72].Implementation of genetic selection for beef cattle would be relatively straight forward in Australia through the beef cattle genetic evaluation tool, BREEDPLAN. This procedure is used by cattle breed societies in Australia and has been modified to include RMP and RFI, which allows individual farm enterprises to include methane mitigation traits when predicting profit maximizing selection indices for breeding stock. The price for carbon credits, provided a payment method is available, influences selection pressure applied to methane mitigating traits [73], particularly when a price for carbon is higher, the selection pressure for lower methane emissions is increased relative to other traits influencing productivity [70].Measuring methane from individual animal on-farm requires expensive equipment and intensive animal handling. Selection of low methane producing seed-stock for the beef cattle and sheep industries appears feasible only through genomic selection [74]. Although methane emissions have been measured on a limited number of Angus cattle and sheep in Australia, an estimated several thousand animals, representative of each industry, would be needed to develop a reliable dataset to utilize genomic selection for an individual breed. An exception is in the dairy industry, where genomic selection is established, few bulls dominate genetic improvement, and their genes are spread widely through artificial insemination. Superovulation and embryo transplant are also well established for dairy cows for rapid gene transfer. Methane emissions can be readily and relatively cheaply measured on a subset of their progeny.Nevertheless, selection of low methane emitting sires by seed-stock breeders to produce commercial bulls or rams will be expensive and unlikely to be implemented. Similarly, the apparent conflict between reducing methane emissions in dairy cows and maintaining genetic gain in production traits, suggests genetic selection is not the most appropriate strategy.3.7. Nitrate as a Feed SupplementNon-protein nitrogen (NPN), typically urea, is fed to ruminants to increase microbial growth, feed digestibility, feed intake and productivity when crude protein concentration in the diet is less than about 60 g/kg DM [75]. However, when NPN is provided from nitrates, hydrogen is used in the conversion of nitrate to nitrite and then to ammonia. These nitrate reduction reactions have a lower free energy change than reactions utilizing hydrogen for methane production and therefore have a competitive advantage. Consequently, adding nitrate to ruminant diets reduces methane emissions, while providing NPN for microbial growth [76,77]. However, if the concentration of nitrite in the rumen rises and nitrite is absorbed into the blood, nitrite poisoning can occur through excess production of methaemoglobin in the blood [78]. Methaemoglobin reduces the oxygen carrying capacity of the blood and can result in death.Numerous experiments have shown feeding nitrate can reduce methane emissions to a maximum of approximately 50% [78,79,80]. Theoretically, one gram of nitrate reduces methane production by 258 mg, but the average efficiency of hydrogen uptake is around 90% [80,81,82,83,84]. A rounded estimate is that 10 g nitrate/kg DM intake can reduce methane emissions by up to 10% [79]. The estimated maximum intake of nitrate without causing nitrite poisoning is approximately 20 g/day for beef cattle grazing low quality pasture [81], resulting in a methane emissions reduction of only 6.5%.From evidence across many experiments, feed intake and growth rate of cattle are positively affected by nitrate feeding when rumen microbes respond to non-protein nitrogen [79]. However, research with feedlot cattle receiving 10 g nitrate/kg feed DM, which reduced methane emissions by 10%, also depressed feed intake by 10% [85,86]. In contrast, lactating dairy cows fed 20 g nitrate/kg of diet DM, reduced methane by about 15%, without an effect on feed intake or milk yield [77,80].Experiments with sheep indicate similar responses to cattle when nitrates are included in either total mixed rations or supplements to lower quality forage diets [83,84,87]. However, there appears to be a consistent increase in wool growth up to 40% [82,86]. This increase is thought to be caused by nitric oxide formed from nitrite dilating blood vessels and increasing blood flow to the skin.Nitrate is readily included in diets or lick-blocks as either calcium nitrate or ammonium nitrate as a full or partial replacement for urea. Integrity of lick-blocks is sometimes reduced because a greater proportion of the block is from the nitrate compound than from urea. The cost of nitrate compounds is higher than an equivalent amount of nitrogen from urea. The major concern with feeding nitrate for reducing methane emissions is the risk of nitrite poisoning, which is particularly dangerous when animals are subjected to exercise [81].Encapsulation of calcium ammonium nitrate with a mixture of sesame gum and sesame oil cake enhanced the concentration of ammonia in an in vitro rumen fermentation system compared with the free nitrate, suggesting lower concentrations of nitrite [88]. Grazing steers supplemented with encapsulated nitrate at the rate of 47 g/100 kg live weight/day produced an average 10.6% less methane over 13 months compared with an equivalent amount of urea and grew slightly faster (p = 0.055), with no signs of nitrite poisoning [89].Despite the potential benefits from encapsulation of nitrate sources slowing the rate of nitrite production, there is little practical value in using nitrates to reduce methane emissions from ruminants in grazing systems. The risk of nitrite poisoning remains a major concern and means the amount of nitrate offered would be conservative, with a small impact on methane mitigation. Although an ERF method for using nitrate to reduce enteric methane emissions has been approved by the Australian government, to date no enterprise had applied for a project using this practice.3.8. Australian Shrubs or Plant Compounds as Feed SupplementsSeveral plant species, including the Australian Tar Bush shrub, Eremophila glabra, and the legume pasture plant, Biserrula, reduce methane emissions in laboratory rumen fermentation cultures and in sheep when compared with control diets [90,91,92]. Increasing amounts of E. glabra in a continuous rumen fermentation system reduced methane emissions linearly by up to 45% [92].Introducing Australian native shrubs with anti-methanogenic properties into pastures is an effective method for reducing methane emissions and increasing productivity from sheep in south-west Australia [93]. The autumn feed-gap, with poor quality, senescent pasture is a major limitation to sheep productivity in the region. Traditionally, sheep are offered expensive supplementary grain during this period. Many Australian native shrubs grow well in the region and provide relatively high quality feed, with high protein content, particularly when consumed with senesced pasture [94].Sheep offered native shrubs, with preserved inter-row pasture species, for 6–8 weeks in autumn, reduced methane emissions by 26%, increased growth rate from 69 to 142 g/day and eliminated the need for supplementary feeding [93]. Simulation modelling predicted the introduction of shrubs with inter-row pasture species increases whole farm profitability by an average of 24% when occupying an optimal 10% of the farm area [95,96].Other bioactive compounds, extracted from native Australian Melaleuca and Leptospermum plants, reduce methane emissions by up to 97% in fermentation assays, but have not been tested in animals [97]. An estimate of the scale of reduction in methane emissions when bioactive compounds are provided as a supplement was obtained by comparing the reduction in methane emissions in the laboratory assay with reduction when the plants Eremophila and Biserrula were fed to sheep [97]. These comparisons suggest methane emissions may be reduced by approximately 25% if these bioactive compounds were fed to ruminants.Laboratory studies suggest that the bioactive compounds could be included in diets at concentrations of 25–50 g/kg feed and reduce methane emissions, when offered as supplements or in lick-blocks. Although the cost of the bioactives is difficult to estimate, it is likely that they could be manufactured commercially. However, for the usefulness of these compounds to be assessed dose response experiments are required for the bioactive compounds to quantify their effects on methane over the longer term in ruminants. Initial experiments could be conducted with sheep in respiration chambers to determine responses over three months. If these experiments show significant and persistent reductions in methane, the effects of the compounds on feed intake and productivity will be needed as well as experiments with cattle. However, further research into these compounds appears unnecessary as the extraction procedures would be expensive and other supplements such as Asparagopsis species and 3-NOP (Bovaer®) are already proven enteric methane mitigants and are in commercial production.3.9. Vaccination against ArchaeaVaccination against rumen Archaea has potential as a low-cost option to reduce methane emissions in sheep and cattle. The strategy is particularly attractive because it would require only one or two treatments in young animals for a lifetime effect and the practice is applicable to all ruminant production systems. Methane emissions were reduced by 8% in sheep vaccinated against a mixture of rumen methanogens [98]. A 20% reduction in methane emissions was regarded as highly probable when the ‘entire genetic repertoire’ of Archaea is examined to identify motifs common to all Archaea, but not to rumen bacteria [99]. Research in New Zealand is proceeding to identify possible antigens and develop a vaccine [100].Attempts to reduce methane emissions through vaccination have returned varying results from 20% methane increase to 69% methane reduction with half the experiments being unsuccessful [101]. For a vaccine to be successful, high concentrations of anti-methanogenic Archaea IgG and IgA antibodies must be transferred from blood to saliva, which has proved difficult to achieve. Williams, et al. [102] found no effect of vaccination on methane emissions or on the number of methanogenic organisms within the rumen. These mixed results suggest that the likely success of a vaccination strategy for substantially reducing methane emissions from grazing ruminants is lower compared to other strategies.3.10. Feeding Wheat to Grazing Dairy CowsFeeding crushed wheat to dairy cows at a rate of approximately 9 kg across two feeds daily, reduced methane production by 30% to 40% with either freshly cut ryegrass pasture or chopped lucerne hay compared with crushed maize grain or pasture alone [103,104]. Milk yield was more than 20% higher on fresh pasture plus wheat. Moate, et al. [105] compared feeding daily with lucerne hay 10 kg of single rolled maize, or with single rolled wheat, or single rolled or double rolled barley replacing the maize. Methane emissions for the cows receiving the rolled wheat was 49% lower than for the maize diet, 73% lower than for the single rolled barley and 78% lower than for the double rolled barley. There was no effect of grain source on feed intake, but the energy and fat content of the milk from cows fed wheat was less than for the other diets.However, when crushed wheat of small size with low starch content was fed at 9 kg/day with a long-cut lucerne hay compared with 9 kg/day of crushed maize fed, methane emissions were similar [63]. Milk yield was lower for the cows consuming crushed maize than for those consuming wheat, but methane emissions were similar between treatments. Although the composition or energy content of the wheat samples used in these experiments were not determined, the results suggest that normal, high starch wheat with rapid fermentation in the rumen is needed to substantially lower methane emissions.Moate, et al. [105] showed a strong relationship between rumen pH and methane yield. The longer period of time rumen pH was below 6 the lower the methane yield. However, the positive effect on methane emissions of feeding high amounts of crushed wheat to dairy cows appears to dissipate over longer periods of feeding and the reduction in methane is at the cost of reduced milk fat [106].Feeding 9 kg DM daily of wheat to dairy cows is not widely practiced for pasture-based systems in Australia, although some producers feed up to 12 kg DM wheat daily. The relative financial return from additional milk, but lower milk fat needs to be compared with the cost of feeding wheat. Providing wheat, with the correct starch composition and morphology, to cows consuming total mixed rations would be a simpler practice because it would replace other cereal grains and ingredients. Dose response curves are required to define relationships between wheat quality, daily wheat intake, methane emissions and milk yield for pasture based and total mixed ration feeding systems. This information would be required to develop an Emissions Reduction Fund method for dairy producers to claim carbon credits. Feeding high amounts of highly digestible wheat could lead to rumen acidosis unless managed carefully. Feeding large quantities of crushed wheat may only be practical in dairies using total mixed rations.3.11. Biochar as a Feed SupplementBiochar or biocarbon is produced from the partial pyrolysis of organic matter at high temperature to generate an extremely porous material with high surface area that is bioactive and binds organic compounds. The properties of biochar vary widely depending on the nature of the organic material and the conditions of partial pyrolysis [107].Laboratory fermentation assays [108,109,110,111,112] and an early experiment with cattle [113] suggested biochar, as a feed supplement, may reduce methane emissions from ruminants. The experiment with young cattle offered a low-quality diet of dried cassava root chips and cassava foliage with 0.6% biochar increased growth rate by 25% to 140 g/day and reduced methane emissions by 22% without affecting feed intake.Leng, et al. [110] showed the reduction in methane emissions in laboratory assays varied with the physical characteristics of biochar. Leng, et al. [113] postulated that the porous structure of biochar stimulates microbial colonization and biofilm formation, which enhances microbial growth and increases volatile fatty acids and protein supply to the animal. Methanogens are found on the outer surface of biofilms and are thought to remove hydrogen, which stimulates the digestion of cellulose and other feed compounds by maintaining a low hydrogen tension. The additional microbial growth and incorporation of hydrogen into microbes may be one reason for the decrease in methane production and increase in animal growth rate.Despite the positive indications from in vitro experiments and theory, recent experiments with cattle fed from 0.5% to 3% of the diet as biochar made from pinewood [114] or whole pine trees [115] have failed to show significant declines in methane emissions. These latter results suggest that biochar is unlikely to have a significant impact on methane mitigation from ruminants.3.12. Best Grazing Management PracticesA great deal of research has been conducted that has allowed development of on-farm practices that improve reproductive performance and efficiency of feed utilization. The major source of enteric methane emissions in Australia is the breeding female for cattle and sheep. A focus of management practice to reduce methane emissions is a reduction in the proportion of feed used to maintain this class of animal [18]. Improving reproductive performance and increasing growth rate of animals for sale reduces the total feed eaten by the herd used for animal maintenance and improves methane intensity (methane per unit of product sold). Reductions of approximately 20% have been indicated from improved grazing management systems for beef cattle [116]. The key red-meat advisory organization, Meat & Livestock Australia, sets out four methods for reducing methane emissions on cattle enterprises: (i) increase the ratio of live weight to age of animal in the herd; (ii) reduce the average age of the herd; (iii) reduce the proportion of unproductive animals in the herd; (iv) change the relative numbers of each livestock class. These changes can be implemented by improving grazing management to increase growth rate of progeny, weaning at younger ages and removing non-pregnant and non-producing cows. Adoption of best management practices across grazing enterprises in Australia was considered to improve productivity by approximately 20%, while reducing methane emissions by around 5% [16].Factors limiting uptake of management practices known to increase the efficiency of feed use for productive functions are complex. Issues for consideration include relative advantages in productivity and profit, added complexity in management, compatibility with current practices and the risk associated with adopting the practice. Some best management practices may be simple to implement and others difficult. The cost of change within an individual enterprise can be significant in producer time and finances. Those changes that do not involve additional time or financial inputs, other than closer management of stock and resources, are most likely to be adopted. More use of the research into drivers and inhibitors of sustainable adoption of best practices may increase the speed and level of uptake of best practices on individual farms, regionally and nationally [117,118]. The strategy ‘Best Management Practices’ was not included in Table 1 because the wide range of practices potentially available t producers made allocation of costs impracticable.4. DiscussionAll herbivorous mammalian species, whether ruminants, foregut or hindgut fermenters emit methane and, except for equids, macropods and rabbits, produce on average 84 L/d per body mass raised to the power of 0.84 [9]. However, since the advent of farming around 8000 years ago, livestock numbers including ruminants, have increased in close association with the human population. In 2019, an estimated 43 Gt of carbon dioxide equivalents were produced annually from the world cattle population assuming 1.5 billion cattle [119], each producing 1500 kg of methane annually [20]. Thus, to contribute to anthropogenic reduction of greenhouse gases on global warming and help preserve ecosystems on the planet, it is crucial for ruminant enterprises to reduce enteric methane emissions. Most available strategies for reducing enteric methane emissions were reviewed to identify their mitigation potential, likely impact on animal productivity and practicality of adoption. Only four of the strategies: supplementation of diets with small amounts of the red seaweed Asparagopsis or the specifically designed chemical 3-NOP; long-term manipulation of the rumen microbial population; or feeding anti-methanogenic tropical legume shrubs appear to substantially reduce methane emissions and have potential practical application.Asparagopsis and 3-NOP as feed supplements appear to have the greatest potential for reducing methane emissions by greater than 90% and more than 40%, respectively. Care must be taken to ensure the Asparagopsis has high bioavailability and is not fed at concentrations above 0.5% of the diet dry matter. These two products are theoretically available for ruminants in all livestock systems. They are undergoing commercialization and should be available for application, at first in intensive feeding systems, provided the costs are appropriate in relation to implementation and returns from any price on carbon dioxide equivalents. Further research is needed to develop appropriate processes for providing these supplements to grazing livestock through lick-blocks or other means.There is also potential for intergenerational mitigation of methane emissions from ruminants through permanent manipulation of the rumen microbial populations. There is strong evidence [43,45] that rumen microbial populations can be changed for periods at least as long as one year to substantially reduce methane emissions. Such a strategy would be easy to apply, but careful management of herds to keep them isolated from other non-treated animals may incur practical difficulties. More research is needed to determine whether keeping herds isolated from other animals that have normal microbial populations is feasible and whether substantial levels of methane mitigation can be maintained over the long-term. Clauss et al. [9] speculate that the lower than herbivore average methane emissions from equids, macropods and rabbits is most likely caused by variations in the microbial populations fermenting forage.Cattle browsing Desmanthus and Leucaena species resulted in methane emissions reductions of 10–20% and improved productivity of animals grazing naturalized pastures. However, their overall contribution to mitigating methane emissions will be relatively small because of restrictions imposed by the limited geographic regions over which these plants grow.Many of the other strategies reviewed have a small impact on methane emissions from individual animals, apply to only a small proportion of the world’s herds, are a risk to animal health, are difficult to apply to production systems and demonstrate a negative commercial production cost/benefit analysis [20].For example, genetic selection over 10 years, vaccination, grape marc, nitrate or biochar supplementation result in reductions in methane emissions of 10% or less. Best management practices resulted in small methane mitigation, but improvements in animal productivity, which makes them worthwhile for ruminant enterprises, though it can be difficult to sustain their application. Although, feeding large amounts daily of ground, high quality wheat reduced methane emissions in lactating dairy cows by around 35%, the persistence of the effect appears to diminish with time and there are risks to animal health through rumen acidosis.The suggested priorities for application of methane mitigation strategies are similar to those listed by Beauchemin, et al. [8], except we have quantified impacts and placed a higher priority on programmed manipulation of the rumen microbial population and potential intergenerational persistence of lower methane emissions. However, most studies have investigated the impacts of individual methane mitigation strategies, with little assessment of at the effects of combined strategies.The review did not consider variations in feed formulation through differing mixtures of forage and concentrates, addition of lipids, rumen modifiers such as ionophores, defaunation and additional phytocompounds including essential oils, saponins or flavonoids. Many of these methane mitigation strategies have been considered in recent publications [8,120] and are all regarded as having low methane mitigation potential. The value of combining mitigation strategies with different biochemical actions in the rumen, as suggested by Beauchemin, et al. [8], requires more than the few experiments that have been conducted to date. There is an opportunity to investigate such combinations, but if the supplementation of Asparagopsis and 3-NOP can be made practical and economic, persistence with strategies that have low methane mitigation potential appears unnecessary.Uptake of mitigation strategies within Australia is encouraged through the ERF, which is a reverse auction process where projects bid for Australian Carbon Credit Units (ACCU) [121]. The current price is around AUD $16/ACCU, but only applies to ‘Methods’ which have been approved by Government. The Method must have specific parameters that can be measured to estimate the number of carbon credits achieved through methane mitigation from livestock. Currently, for ruminants in Australia, there are only Government endorsed methods for nitrate supplementation and beef herd management. The Beef Herd Management Method applies to projects that demonstrate reduced methane emissions as a result of new forages, new supplement use or new management methods that increase the growth rate of cattle to market weight at an earlier age or improve reproductive performance. Generic Methods are proposed in Australia for anti-methanogenic supplements in feedlot, dairy and grazing systems along with anti-methanogenic legumes, but are still dependent on further research.Davison, et al. [20] estimated that by 2030 Asparagopsis supplements could be applied to 20% of Australian ruminants. With Australian greenhouse gas emissions in 2019 being 539 Mt carbon dioxide equivalents and ruminant enteric methane contributing 12%, a 20% reduction in enteric methane would result in 13 Mt less annually of carbon dioxide equivalents being released into the atmosphere by Australian farmers. That annual 13 Mt (in 2030) decrease in greenhouse gas emissions from adoption of Asparagopsis compares well with a reduction of 4 Mt in 2019 due to all renewable electricity production in the country and the 1.57 Mt reduction in enteric methane in the last 10 years (2009 to 2019) [122].With over 1.5 billion cattle in the world [119] and each animal emitting around an average of 1500 kg carbon dioxide equivalents each year [20], adoption of Asparagopsis supplements at the same 20% rate by 2030, as assumed for Australia, would reduce global methane carbon dioxide equivalent emissions by half a gigaton annually. Application of methane reduction strategies have large potential for reducing global greenhouse gas emissions, while allowing the continued consumption of red meat.Although this review is based predominantly on an analysis of the Australian ruminant industries, the assessments of the practical value of each methane mitigation strategy should be applicable to other countries and production systems. The feed supplements, Asparagopsis and 3-NOP, can be readily applied to any mixed ration formulation for feedlot and dairy animals, but methods need to be developed to ensure bioactivity of Asparagopsis can be maintained once it is incorporated into a diet. There is a greater challenge in utilizing these supplements in rangeland environments, where in Australia approximately 80% of enteric methane is produced from the breeding herd. The research being undertaken in Australia to investigate practical methods for supplementing rangeland animals would be applicable to many similar environments in other countries. Similarly, manipulation of rumen of the rumen microbial population should be applicable to production systems around the world where treated animals can remain isolated from untreated stock.5. ConclusionsThere is an urgent need for managers of ruminant enterprises to adopt strategies that reduce enteric methane emissions to avoid further increases in anthropomorphic associated greenhouse gas emissions. In Australia, this will most likely happen when alternate income sources from carbon projects can be realized and in turn this requires a substantive increase in research investment that results in new Methods that can be employed in carbon projects under the ERF. Potential methane mitigation strategies studied within an Australian research program and others were reviewed to identify their methane reduction potential, their impact on animal productivity and ease of application. The review concluded that only supplements of the red seaweed, Asparagopsis, and 3-NOP have potential for cost-effective practical application and can reduce methane emissions by 90% and 40+%, respectively, with likely increases in animal productivity and without negative effects on animal health or product quality, provided Asparagopsis does not exceed 0.5% of the diet. Both products are undergoing commercialization. There is also potential for intergenerational mitigation of methane emissions through permanent manipulation of the rumen microbial populations and keeping these animals isolated from other non-treated animals. However, further research is needed to determine the long-term feasibility of the strategy and magnitude of methane mitigation. An assumed annual reduction of enteric methane emissions of 20% by 2030 would reduce greenhouse gas emissions by half a gigaton each year.

animals : an open access journal from mdpi

10.3390/ani11071981

PMC8300213

Animal welfare assessment is an essential tool for maintaining positive animal wellbeing. Validated welfare assessment protocols have been developed for farm, laboratory, zoo, and companion animals, including horses in managed care. However, wild and free-roaming equines have received relatively little attention, despite populations being found worldwide. In the UK, free-roaming ponies inhabit areas of Exmoor, Dartmoor, and New Forest, England, and Snowdonia National Park in Wales, amongst others. Visitors and local members of the public who encounter free-roaming ponies occasionally raise concerns about their welfare, as they are not provided with additional food, water, or shelter. In this study, we evaluated the feasibility, reliability, and repeatability of welfare indicators that can be applied to a population of free-roaming Carneddau Mountain ponies to address such concerns. Our findings indicate that many of the trialed indicators were successfully repeated and had good levels of inter-assessor reliability. Reliable and repeatable welfare indicators for free-roaming and semi free-roaming ponies will enable population managers and conservation grazing schemes to manage the welfare of free-roaming horses and ponies.

Validated assessment protocols have been developed to quantify welfare states for intensively managed sport, pleasure, and working horses. There are few protocols for extensively managed or free-roaming populations. Here, we trialed welfare indicators to ascertain their feasibility, reliability, and repeatability using free-roaming Carneddau Mountain ponies as an example population. The project involved (1) the identification of animal and resource-based measures of welfare from both the literature and discussion with an expert group; (2) testing the feasibility and repeatability of a modified body condition score and mobility score on 34 free-roaming and conservation grazing Carneddau Mountain ponies; and (3) testing a prototype welfare assessment template comprising 12 animal-based and 6 resource-based welfare indicators, with a total of 20 questions, on 35 free-roaming Carneddau Mountain ponies to quantify inter-assessor reliability and repeatability. This pilot study revealed that many of the indicators were successfully repeatable and had good levels of inter-assessor reliability. Some of the indicators could not be verified for reliability due to low/absent occurrence. The results indicated that many animal and resource-based indicators commonly used in intensively managed equine settings could be measured in-range with minor modifications. This study is an initial step toward validating a much-needed tool for the welfare assessment of free-roaming and conservation grazing ponies.

1. IntroductionKnowledge of the welfare of animals under human care is integral to their successful management; equally important is an understanding of the welfare of free-living animals to guide how we interact with wildlife and their habitats [1]. To gather knowledge to improve animals’ welfare, a validated, reliable, and repeatable method of assessment is required [2,3]. Recently, welfare assessment has moved from resource-based or simple indicators of environmental parameters to include indicators that monitor the behavioral responses and physiological conditions of individual animals over time [4]. Animal-based indicators are particularly relevant in the welfare assessment of wild or free-ranging animals. Indicators related only to environmental parameters do not allow for the assessment of the behavioral or physical responses to the prescribed condition and are not representative of the animal’s welfare state [4]. An animal’s ability to adjust to both predictable and unpredictable change in its environment is vital to maintaining welfare [5]. For example, seasonal changes in forage and grass availability for grazing animals may elicit periods of fasting (hunger) or seasonal weight gain (fat reserves) to cope with available resources [4]. While periods of hunger may be considered a welfare issue, this may not be a factor for the animal itself if its adaptive capacity (physical and mental abilities) has not been exceeded [4]. Using a multifactorial approach including animal-based indicators (AB) (physical/physiological outcomes) and resource-based indicators (RB) (what is available in the environment) for assessment enables the evaluator to quantify levels of individual welfare [6]. There are a range of welfare audit protocols that have recently been developed to determine welfare by evaluating RB and AB indicators for farm [7,8], companion [9], laboratory [10], and zoo animals [11,12,13,14]. In contrast, there are few protocols for extensively managed animals [3,15,16] or, indeed, free-living wild populations [1].Populations of free-living and free-roaming horses are found throughout the world. Significant numbers occur in Australia, with an estimate of over 300,000 free-roaming brumbies [17]. In the US, smaller populations of free-roaming horses occur. There are just over 79,000 mustangs and 15,546 burros managed by the United States Department of the Interior’s Bureau of Land Management (BLM), inhabiting approximately 31.5 million acres of land across 10 states in the Western United States [18]. In the UK, free-roaming ponies inhabit areas of Exmoor, Dartmoor, New Forest in England, and the Carneddau mountains in Wales, amongst others.Public concern regarding the welfare of free-roaming equids has become more prevalent in recent years. Whilst increased public awareness and demand for the improvement of equine welfare are evident across various equine disciplines, e.g., sport, working horses, racehorses, and those kept for pleasure [19,20], there is also concern regarding the welfare of feral populations. Equine stakeholders participating in a study carried out by Horseman et al. [20] identified overbreeding, a lack of food in winter, and gatherings (rounding-up for health checks) as areas of welfare concern for free-roaming ponies specifically. The public’s concern for the welfare of mustangs in the US has been particularly well-documented, with many groups urging the Bureau of Land Management to cease all gatherings, removals and contraceptive strategies [21]. Visitors and local members of the public who encounter free-ranging ponies in the Carneddau mountains also occasionally raise concerns about their welfare because they are not provided with additional food, water, or shelter (Carneddau Pony Society, personal communication, 3 November 2019). There is therefore a need for objective indicators of welfare in these populations at both the individual and group levels. Despite several validated equine welfare assessment protocols in existence for sport, pleasure, and working horses, e.g., [22,23,24,25,26], there is currently only one audit available for free-roaming horses. This describes a 10-step protocol using the Five Domains model that can be used to form a template for welfare assessment in free-living terrestrial species and uses Australia’s brumby horses as an example [1]. Here, we therefore trialed specific welfare indicators to ascertain their feasibility, reliability, and repeatability for the welfare assessment of free-living horses using free-roaming Carneddau Mountain ponies as an example population.2. Materials and Methods2.1. Ethics This study was granted ethical approval by the University of Chester’s Faculty of Medicine and Life Sciences Research Ethics Committee on 6 February 2020, reference number 1609/19/JH/BS. The Carneddau Mountain Pony Society provided written permission to access the study population on 3 November 2019.2.2. Study PopulationThe Carneddau Mountain ponies are classified as semi-wild (surviving and breeding without human intervention) and believed to be free-roaming since the Bronze Age [27]. Approximately 220 free-roaming ponies comprising mares, their offspring, and 12–15 dominant stallions inhabit about 5377 hectares of habitat [28]. Genetic studies have revealed that the Carneddau Mountain ponies are genetically distinct and isolated from other Welsh pony populations, and they could therefore represent a valuable future genetic resource [29]. The ponies graze on common land; families whose homesteads border the Carneddau mountains are each entitled to grazing rights under the Common Rules 1966 [30], and members of nine families with rights to graze ponies on the Carneddau mountains are represented in the Carneddau Mountain Pony Society, a non-profit organization responsible for managing the Carneddau ponies. This population is therefore semi-feral; it is not heavily managed, yet each pony is privately owned.2.3. Choosing the Welfare Indicators A systematic review of existing equine assessment literature identified several AB and RB indicators of welfare feasible for free-roaming ponies [1,19,23,24,25,31,32,33,34,35]. Management indicators were excluded from consideration because the subjects are free-roaming, and management intervention in-range is minimal. Each identified AB and RB indicator was reviewed and discussed by the lead investigator with a small team of experts in the UK. Experts included a first opinion equine veterinarian (BVetMed and MRCVS); a specialist equine surgeon (MVB, MSc, DipECVS, and MRCVS); a behavior, equine science, and farm animal welfare scientist (PhD); a welfare scientist with wild animal experience (PhD); a conservation grazing scheme manager with extensive knowledge of the Carneddau landscape and ponies; and a behavior and welfare scientist with wild animal and equine experience and a working knowledge of the Carneddau ponies (PhD). Indicators were discussed based on criteria of significance to free-roaming equine welfare, feasibility, and practical application of the indicator in-range for non-specialist’s stock persons and grazing scheme managers, as well as their validity in terms of evidence from the published literature [19,31,34,35,36].In total, 12 AB and six RB welfare indicators were chosen for further evaluation. Most welfare indicators were either categorized with a zero or one to represent a negative state of different degrees of severity, followed by two = neutral, and sometimes three = positive. This type of multi-level numerical scoring system has been used in several of the published welfare assessment templates, e.g., the WQ® system [8] and the Standardised Equine Based Welfare Assessment (SEBWAT) [26]; however, in WQ®, zero = positive, 1 = neutral, and 2 = negative, and in SEBWAT, they use a mixed approach for scoring criteria. In this study, based on the intended end-users, we determined that it was more intuitive to have a higher number equate to a positive category. If a score of zero was applied, a significant risk of welfare compromise was indicated and escalation was required (i.e., reporting the pony to the population managers). A few indicators were categorized using letter scores alone. Reproductive status was classified as lactating, not lactating, or not applicable (e.g., sub-adult female and stallion). Hoof shape was classified as (A) for long overgrown hooves, (B) for hoof cracks (vertical/horizontal), and (N) for normal. Fecal consistency was classified using a mixture of numerical and letter categories; (N) was assigned for normal feces, and feces that were not well-formed, cow-dung-like, or mostly comprised large fibers received an (A) for abnormal. However, feces that had a water-like consistency (diarrhea) received a category of zero. This mixed approach was used for feces due to the rapid progression of the disease state (dehydration, electrolyte derangement, and endotoxemia) that can accompany acute diarrhea resulting from typhlocolitis in horses. Acute diarrhea resulting from typhlocolitis is a major clinical sign observed in horses with salmonellosis, intestinal clostridiosis, cyathostomiasis, and (to a lesser extent) strongylosis [37].Though some indicators could only be assigned a score of 1 or above, others had the option of a zero score because these were deemed to indicate a more severe level of welfare compromise. Those that include an option of zero included BCS, mobility, ocular discharge/swelling, and wounds/swelling. If a pony received a zero in any of these categories, the “second level” application of the Horse Grimace Scale (HGS) was indicated using the HGS mobile application, in addition to reporting the issue to the Carneddau Pony Society. This process of classification ensured that the correct weighting was assigned to a welfare indicator that posed an immediate danger to the health and welfare of the pony, e.g., this assured that alopecia of mane/tail was not weighted at the same level as an open wound > 7 cm involving deep tissue and muscle. The HGS also had a zero score option. All remaining indicators are outlined in Table 1.2.4. Feasibility of Assessing Welfare Indicators In-RangeThe majority of welfare indicators were derived from literature focusing on the welfare assessment of intensively managed horses (stabled); therefore, it was essential to determine whether the selected indicators could be carried out in-range and whether the minor modifications that would be required for in-range setting were appropriate. All 12 AB and six RB indicators were explored and tested in-range at conservation grazing sites and in the Carneddau Mountain on free-roaming ponies prior to the main testing phase to confirm their feasibility. This enabled us to develop a decision tree-based question for each of these welfare indicators. The pre-testing also enabled the identification of auxiliary equipment necessary to facilitate observations of ponies, e.g., binoculars. Finally, any significantly modified indicator, e.g., body condition score (BCS) or mobility, was individually tested for observer reliability during a pre-trial of indicators.In an early stage of the study, it was determined that forage availability, quality and accessibility were difficult to evaluate. For an assessor to achieve an objective appraisal of resource provision, it would be time-consuming and require specialist knowledge of the species of plants in-range and those which are consumed by the ponies, so it was not practical for a rapid welfare assessment in a free-roaming habitat. Therefore, it was not included in the list of RB indicators for nutrition, and the focus was placed on BCS as the primary AB indicator for nutrition. Hoof shape was feasible in-range, but it was also a time-consuming indicator, as the ability to quantify this required extended focal observations; therefore, this was only included as a second level requirement if the pony’s mobility was impaired (Figure 1). Fecal samples were opportunistically observed, and despite not having a time limit for observations, waiting for a pony to defecate could potentially add considerable time to an individual pony observation. For those ponies that did defecate, feces were easy to score and collect without moving into a pony’s flight zone. Pictures were taken of defecating ponies along with landmarks (e.g., protruding rocks and vegetation) in proximity to the fresh feces. Photos were then used to locate feces and assess fecal consistency.Ponies were photographed during initial observations, and their GPS locations were recorded to enable the identification of ponies on subsequent visits. All ponies were easily located on each visit, with some nursing mares using close to the same grazing area during both visits. All ponies were approachable with varying flight zones, the majority of which were between three and nine meters. The Carneddau Mountains are a tourist destination attracting walkers, fell runners, and mountain bikers, and the ponies are exposed to regular human–animal disturbance. Ponies were never given less than three meters of space on approach; however, during observations, several ponies came within a meter or less of observers, as foals were curious, and adults would approach while grazing. Binoculars were used to assess several the indicators regardless of distance and were required for the assessment of wounds, nasal discharge, ocular discharge, skin condition, and hoof shape. Both eyes and nostrils were not always visible on first approach as due to the position of the pony. However, the observer could easily move around the pony to achieve the best vantage point. The upland conditions provide open spaces, so the pony moving out of sight of the observer was not a factor. Binoculars were also used to identify bands from a distance when locating individual ponies. The most important aspect of the assessment of the welfare indicators in-range was allowing for ample time to assess. It was often necessary to wait until a pony moved to an area with open patches of grass or to one of the paths or roads (which they frequently accessed) to observe lower limbs.2.5. Modification and Pre-Testing of Welfare Indicators for In-Range Assessment2.5.1. Lameness and MobilityLameness and compromised gait scoring are animal-based indicators found in numerous equine welfare assessments for sport and working horses and donkeys, e.g., [22,23,25,26,33]. A typical equine lameness examination includes horses observed in the walk and trot on flat even ground, in a straight line, and circling on a hard surface [38]. Like many environments in which free-roaming ponies inhabit, the Carneddau Mountain terrain is steep, uneven, and features tall vegetation and varying substrates. Free-roaming horses are also not frequently handled, which makes an in-hand lameness evaluation prohibitive. Thus, a standardized lameness grading protocol would not be feasible in-range, meaning a scoring system focusing on the pony’s mobility rather than identifying a specific type of lameness was required.The mobility scoring system used here was modified from the Agriculture and Horticulture Development Board, UK (AHDB) Dairy Mobility Score system [39], which indicates a cow’s ability to move comfortably in a walk during normal, unaided locomotion for 6–10 strides, observed both from the side and behind the animal. The system scores each cow on a 3-point scale; cows with good mobility (no impairment) are scored as zero, imperfect mobility (steps unevenly) is scored as 1, and cows with impaired mobility (cannot keep up with herd and limb lameness is visible) are scored as two. Similarly, we used a 3-point scale from 0 to 2, with a score of two indicating no signs of abnormality in mobility, 1 indicating walking with abnormality in gait and not even in rhythm (weight-bearing), and zero signaling severely impaired mobility (pony unable or unwilling to move forward, unable to stay with herd). Individual ponies were observed from a suitable distance (respecting the ponies’ flight zones) that was not less than three meters. In practice, this was usually less than nine meters (binoculars could also be used as required), and for the majority of ponies, this was just over three meters. The ponies were observed resting and then in a natural walk, with the assessor viewing from the side and rear of the pony. Assessors did not push the pony forward by entering the flight zone; instead, they observed until the pony walked freely and naturally and scored based on findings. There was no time limit for the observation. The observer then used the decision tree (Figure 2) to score the pony. A score of zero required immediate action, as previously indicated. If the pony was not moving for a non-musculoskeletal issue such as colic or being stuck (e.g., in a ditch), then the mobility was not scored and immediate action was taken.2.5.2. Body Condition Score (BCS)Equine body condition is commonly assessed using either a 5-point [40] or 9-Point [41] scale. Scoring is commonly performed both visually and using palpation to evaluate body fat and muscle covering specific areas of the body, including the neck, shoulder, ribs, abdomen, and rump. A BCS is a valid, reliable and repeatable measure to assess on-farm nutrition [19]. However, palpation on free-roaming ponies is often not feasible. Body condition scoring is sometimes carried out by comparing the equine in question with pictures [42,43] or observations without palpation [44]. Data collection commenced with an attempt to validate BCS by comparing assessor scores of the same pony with and without palpation. Two groups of conservation grazing ponies were selected n = 22; both facilities had race and crush systems in place for safe handling and had previously handled the ponies. However, during our initial assessment, it became evident that the ponies had not been sufficiently handled to enable safe palpation (for pony/assessor), nor was the crush system adequate for the ponies at one facility. The decision was made to suspend the validation of BCS scoring without palpation and to focus on the inter-reliability of assessors using a modified (no palpation) approach. Before the commencement of the formal study, we conducted a pre-trial of BCS without palpation by two assessors (a primary investigator and an equine veterinarian) on thirty-four free-roaming and conservation grazing ponies between February and March 2020. The observer approached the pony slowly and quietly without entering the pony’s flight zone (if the pony turned or moved away, the observer stopped immediately, as this indicated they had entered its flight zone). Observations were typically about three meters from the pony, although a handful of ponies had slightly larger flight zones (both observers viewed from the same point). The observer began with an initial visual inspection from the side of the pony to examine muscle and fat cover of the ribs, neck, shoulder, back, abdomen, and pelvis. The observer then slowly and quietly stood behind the horse and evaluated fat deposits around the tail bone while observing the shape of the croup (the point of the buttock). Finally, the observer assessed the visibility of the spine and hip bone. The observer then referred to the body condition score template [41] using a scale from one (extremely emaciated) to nine (extremely fat) to score the pony’s body condition. This score was then categorized as a 3-point score (from 0 to 2). Ponies scoring a zero in the modified BCS equated to a Henneke score of 1–2 (emaciated/poor) or 8–9 (fat/extremely fat); as both emaciation and obesity have welfare implications, a score of one was equal to thin–moderately thin (Henneke score: 3–4) and a score of 2 equated to moderate (Henneke score: 5–7) [41]. Pictures for the later identification of each pony were taken during the BCS scoring.2.5.3. Horse Grimace Scale (HGS)Pain in animals is an unpleasant sensory and emotional experience, and it is a significant welfare concern [45,46]. Methods of the assessment of pain are essential for animal management. Facial expressions as an indicator of pain have been defined for several species including sheep [47], rabbits [48], rodents [49], and horses [31]. The HGS was validated for use in horses undergoing routine castration and has been used in the AWIN Welfare Assessment for Horses [33]. The HGS was developed into an application that is available for most mobile phone platforms (e.g., IOS and Android) and includes in-built training for users. In this welfare assessment trial, HGS was indicated as a second level assessment for ponies scoring a zero for the following AB indicators: mobility, ocular injury/discharge, wounds, fecal consistency, and BCS. If a zero was scored, the HGS mobile application was accessed and the assessor was asked to answer all questions in the HGS 13-point scale, which resulted in a possible HGS score from 0 to 12. A score of zero equated to no facial indicators of pain present and a score of 12 indicated that all equine facial indicators of pain were obviously present to the assessor. The assessor was then asked to classify the HGS score according to a 3-point scale as follows: an HGS score of 0–3 received a score of 2, an HGS score of 4–7 received a 1, and an HGS score of 8–12 resulted in a score of zero.2.6. Testing the Welfare Indicators for Reliability and RepeatablilityThe final template comprised 17 questions that received a numerical score between 0 and 3, two categorized questions where the observer applied a letter score (hoof condition and reproductive status), and one that used a mixed method (fecal consistency) for a total of 20 questions. In July and August 2020, the prototype assessment template was trialed on 35 free-roaming Carneddau mountain ponies across two grazing areas on the Carneddau mountains: Abergwyngregyn/Llanfairfechan (53.2286705, −3.9842733) and Conwy (53.272243, −3.868534) common areas. Individual ponies were concurrently evaluated by two people (a primary investigator and either an equine veterinarian or a behavior and welfare scientist with equine experience) to test the reliability of each indicator. Of the thirty-five assessed ponies, twenty-six were adults, six were sub-adults, and three were foals, with a sex ratio of twelve males to twenty-three females, representing six bands, a bachelor group, and a mare and foal pair. The female bias was a result of the composition of the study population, as this comprises approximately fifteen stallions and only a few bachelor bands. All assessments of welfare indicators were conducted between 08:30 and 18:00 h and were not carried out during heavy rain to avoid conditions that would affect visibility and assessor safety on the mountain, as permissible within the mountain safety advice for Snowdonia National Park.In addition, the welfare indicators were tested and repeated on twenty ponies by the primary investigator within 7–17 days of the original assessment to test for the repeatability of each indicator. Ponies were photographed for identification at the commencement of the evaluation to ensure that ponies could be recognized (by individual markings) on future visits to test for repeatability. All repeat assessments were conducted on ponies in the Abergwyngregyn and Llanfairfechan common area.2.7. Statistical Analysis All statistical analyses were carried out using the R 4.0.2 platform (Vienna, Austria) [50] and MedCalc for Windows, version 19.2.3. (MedCalc Software, Ostend, Belgium) ([51] Reliability and agreement considerations were selected as advised in the literature for assessing inter and intra-rater variability [52,53,54,55].2.7.1. Preliminary Testing of Modified BCS and Mobility Descriptors for Inter-Assessor ReliabilityAn interrater reliability analysis using the Cohen’s kappa (k) and weighted kappa (kw) test statistic was performed to determine consistency among assessors scoring each pony in the preliminary testing phase using the modified Henneke BCS system and the mobility score system. The kappa statistic compares the observed agreement between two assessors on an ordinal scale using a chance-adjusted indicator, and it considers the matches on the main diagonal [52]. The weighted kappa is modified such that it considers off-diagonal differences between observers and considers the degree of disagreement between observers rather than treated as equal, and it is preferable for ordinal data. Linear weighting based on agreement was used because the questions were all based on a 3-point scale. Linear weighting places the same importance on the difference between the first and second category as the difference between the second and third category [53]. Values were interpreted according to the work of Altman [54] with values of 0.81–1.00 considered very good, 0.61–0.80 considered good, 0.41–0.60 considered moderate, 0.21–0.40 considered fair, and <20 considered poor agreement. Kappa estimates and their 95% confidence intervals were calculated using the “psych” package in R [56].2.7.2. Final Testing of All Welfare Indicators (Prototype Template) for Inter-Assessor and Test/Retest ReliabilityAnalysis for inter-assessor agreement between the primary assessor and one of three additional assessors using the prototype welfare assessment (n = 35) was conducted using the linear weighted Cohen’s kappa and the percentage of agreement. The test/retest analysis of the prototype welfare assessment (n = 20) by the primary assessor was performed using Cohen’s unweighted kappa test statistic and the percentage of agreement. Values were interpreted according to the work of Altman [54] with values of 0.81–1.00 considered very good, 0.61–0.80 considered good, 0.41–0.60 considered moderate, 0.21–0.40 considered fair, and <20 considered poor agreement. Kappa estimates and their 95% confidence intervals were calculated using the “psych” package in R [56].3. Results3.1. Preliminary Testing of Individual Indicators: BCS and MobilityThere was a good degree of reliability between the two observers for BCS of the 34 ponies tested in the preliminary phase using the modified Henneke score. Cohen’s weighted kappa score was 0.78 (95% CI: 0.78–0.78; (Table 2). The reliability of the mobility scoring using a 3-point scale was very good, with 100% agreement between the two observers in their scoring of the 34 ponies and a κw of 1 (Table 2).3.2. Inter-Assessor ReliabilityIn this study, several of the welfare indicators in the audit were not witnessed (alopecia of mane or tail, nasal discharge, and coughing). They were therefore not included in the results because their reliability could not be determined. The HGS [31] was only infrequently indicated as the prescribed criterion for conducting the HGS required the observed pony to score a zero in one or more of the following categories: mobility, ocular injury/discharge, wounds, fecal consistency, and BCS. Therefore, assumptions about reliability could not be made. Hoof shape/condition assessment was only indicated in two instances where mobility was impaired. The agreement in these instances was 100% between assessors, but not all specific categories were observed (A = overgrown and B = cracked/chipped) and therefore could not be tested for reliability. Data on fecal consistency were opportunistically collected, and only 12 of the 35 ponies defecated. Assessors reached 100% agreement in terms of fecal consistency assessment, but all ponies had normal feces; other categories (0 = watery, and 1 = abnormal) were not observed, and thus assumptions about reliability across all categories could not be confirmed. Water quality scored three (fresh spring forming a stream, pond, or lake) by all assessors for n = 35 ponies. As the categories 1 = no water detected and 2 = stagnant pool/ puddle were not identified, this indicator could not be further tested for reliability. Ocular discharge and/or swelling were only infrequently encountered, and while all assessors had 100% agreement, category 1 = discharge with an open eye was not witnessed. Environment ease of movement (people/bikes/dogs) had 100% agreement between all assessors; however, there were no scores in category 1—high footfall. This was also true for wounds and swelling, where no ponies received a score of 0 = open wound involving deeper tissue/muscle (acute).The indicators that could be fully evaluated showed mixed reliability across the assessors (Table 3). Assessor 1 (primary investigator) and Assessor 2 (equine surgeon), with n = 18 assessments, had a very good agreement between five of the seven remaining indicators (BCS, ease of movement, social contact, and human approach), moderate agreement for comfort around resting, and poor agreement for thermal comfort. Assessors 1 and 3 (scientist), with n = 11 assessments, achieved good to very good reliability for social contact and human approach. However, for ease of movement, comfort around resting, and thermal comfort, both assessors assigned scores in only one category, meaning that although they had 100% agreement, assumptions about reliability across all categories could not be made. Finally, Assessors 1 and 4 (scientist), with n = 6 assessments, had good to very good agreement (BCS, social contact, and human approach), whilst thermal environment and comfort around resting had kappa scores of 0, indicating no agreement better than chance. Indicators that achieved 100% reliability between assessors were not further investigated; details are provided in Table 4.3.3. Intra-Assessor Reliability (Test/Retest)As with the inter-assessor trial, during the test/retest phase, some of the welfare indicators were not observed (alopecia of mane or tail, nasal discharge, and coughing). The HGS and hoof condition were not warranted because no ponies scored a zero in any of the indicators that triggered the need to carry out the HGS. Additionally, there were no mobility scores of 0 = immobile or 1 = minor impairment during the test/retest phase. As with the inter-assessor reliability assessment, during the test/retest phase, all ponies scored a three for water quality and availability; therefore, all categories could not be tested for reliability but repeatability was confirmed. For the remaining indicators, the reliability of the repeated observations by the primary investigator of the twenty ponies had mixed results. BCS, reproductive status, and skin condition all had a kappa estimate of 1.0 (very good agreement). Wounds and swelling had fair to good agreement, as did the ease of movement (dogs/bikes/people). Social contact had a kappa score of 0; however, the agreement was 95%, with 19 of the 20 ponies receiving the same score in the initial and repeated welfare assessments. Similarly, comfort around resting had a percentage agreement of 90% and no agreement for kappa, with 18 of 20 ponies receiving the same score in the test/retest phases. All other indicators attained moderate reliability (Table 5).4. DiscussionThis project involved testing AB and RB welfare indicators to determine their feasibility, reliability, and repeatability. This was carried out to address the need for reliable in-range welfare indicators for free-roaming and extensively managed horses, which, to date, has not been readily available. The results of this preliminary study indicate that many of the indicators trialed here were successfully repeated over two assessments and had good inter-assessor reliability when trialed by the primary investigator, equine veterinarian, and two animal behavior and welfare scientists. Many of the indicators reached the predefined thresholds for reliability (good to very good) between assessors in numerous categories (human approach, BCS, ease of movement, and social contact), and several indicators, such as skin/coat condition, had 100% agreement between assessors (Figure 3).The resource-based indicators for ease of movement (hazards), comfort around resting, and thermal environment had variable reliability. Each of these questions asked the assessor to view a 500 m circular radius to identify artificial hazards, suitable area for resting, level of anthropogenic disturbance, and access to shelter or shade (microclimates), as well as to use a decision tree-style question to determine the most appropriate answer. Unlike most of the welfare indicator questions, no images were included, which may have allowed for a greater level of subjective interpretation. A stone wall may act as a wind protector and provide a shaded area but is an atypical ‘shelter’ and may be discounted if the observer is only considering more traditional means of shade, e.g., trees. Additionally, without explicitly providing the number of people that quantifies ‘high public footfall,’ the questions are open to subjective interpretation. The addition of visual cues for these questions and quantitative values may be beneficial to improve agreement among assessors and can be easily rectified in future trialing of the assessment. As recreation can negatively and directly affect wildlife, resulting in altered behavior and temporal avoidance of paths and trails [57], this is an important indicator for free-roaming horses that inhabit tourist destinations. Further refinement is required to ensure reliability among observers.Several indicators were only infrequently assessed (hoof shape/quality, eye discharge, and second level HGS) or not witnessed (alopecia mane or tail, nasal discharge, and coughing), or scores were not obtained across all categories (water quality, fecal consistency, and wounds and swelling). The reliability of these indicators was therefore difficult to fully ascertain in this study. However, we were able to view each pony sufficiently to determine their presence or absence. Occurrences of some of the health indicators in free-roaming populations may also have low prevalence. For example, coughing is a clinical sign typically observed in domestic horses and ponies with equine asthma or pleuropneumonia [37]. The former is often attributed to poor ventilation and increased environmental allergens found in intensively managed environments. The absence of coughing observed in this study reflects the low prevalence of respiratory disease in this population and may be attributed to the open-air environment and low levels of aerosolized allergens. Eye abnormalities were only infrequently observed in the Carneddau population. In contrast, this was the most frequent physical health indicator observed among 75 long-lined tethered horses and in 28% of a population of 112 free-roaming domestic horses kept on public land in South Wales [32]. Where eye injuries do occur, they are painful; corneal ulcerations are common in horses, and bacterial and fungal keratitis may present with initially mild clinical signs, but serious ocular complications may occur without action [57]. The identification of these health indicators, or at least the confirmation of the absence of injuries, is therefore important for the management of free-roaming and conservation grazing ponies. Therefore, one of the important outcomes of this preliminary trial was that assessors were able to view both eyes of each of the observed ponies using binoculars with distances of three-to-nine meters. Assessors at this distance identified both normal eyes and ponies with discharge with the partial or complete closure of the eye with 100% agreement. Another indicator with low variance was water quality. All ponies had access to numerous clean water points comprised spring-fed streams and ponds. Precipitation totals during July and August 2020 were 110.1 and 163.4 mm, respectively [58]. However, the lowest rainfall on record since 1873 was recorded in May 2020, with 12.9 mm falling North West England and Wales [59,60]. Assessments in May could therefore have yielded a different result. Indeed, free-roaming environments are characterized by seasonal variation, and indicators that can identify welfare change across periods impacted by climatic conditions, food availability, and seasonal biological factors (e.g., breeding) are essential. In conclusion, the lack of the incidence of the specific categories in some welfare indicators does not necessarily diminish their value as part of the assessment but rather indicates that some conditions may have low prevalence in this population. Further trialing across different horse populations and seasons in later stages of this project could enable validation in all categories. While testing for reliability was difficult due to homogeneity in some scores, the same difficulties have arisen in equine assessment trials in domestic horses, e.g., [25]. Conversely, the lack of negative scores for indicators of systemic illness or injury may be a positive sign that the population was mostly experiencing higher welfare during the assessment period.Statistical parameters indicated moderate to very good reliability across the majority of individual indicators in the test/retest phase. A limitation of the Cohen’s weighted kappa coefficient statistical test for reliability is related to the prevalence of the condition under consideration and that low prevalence (skewing of data) affects kappa estimates [61,62,63]. Comfort around resting and social contact had no agreement when using the kappa calculation, but both had high percentages of agreement (90% and 95%, respectively). Skewed scoring, where there is only one score or a low occurrence of other scores, appears to be commonplace in welfare assessment [3,25,64]. Though the percentage of agreement does not correct for any agreement that could occur by chance, all assessors were trained and experienced, therefore reducing the likelihood that a rater would guess and a result would be overestimated. Percentage agreement has been used in other welfare assessments in conjunction with additional reliability tests, as presented in this study [3,25]. This is a reason why the interpretation of kappa values must be carefully considered and why presenting kappa estimates along with the percentage of agreement for the context of how the assessors scored each indicator is useful [55]. However, we could also explain the lack of agreement with comfort around resting and social contact in a test/retest setting because of changes in the dynamic of the group, not as a failure of the indicator to be reliably repeated. For example, in an initial assessment of one stallion, he was observed away from his band overmarking feces, whilst in the follow-up assessment, he had re-joined his band. In this case, it was a change in condition rather than a lack of reliability. Though Carneddau ponies maintain a small, consistent home range of around 1.5 km2 [30,65], individuals are unlikely to be in exactly the same area for both assessments.Following an initial trial carried out on 34 ponies, the resulting modified BCS without the palpation criterion showed a very good level of reliability between assessors, with two assessors and the lead investigator achieving 100% agreement. Within the context of this assessment, BCS is particularly relevant due to the absence of other rapid resource-based indicators to evaluate nutritional state. Thus, the BCS indicator is an integral aspect of the effectiveness of this assessment. However, further testing across the late autumn and winter seasons is required to ensure that a similar agreement is achieved when the ponies have their winter coats. This is typically the time when the palpation of the animal is necessary to accurately determine body condition [41] (Henneke et al. 1983). However, in total, a total of 69 ponies were BCS assessed with good reliability, and 34 of the 69 trials were conducted during February and March when winter coats were not fully shed.Our results showed that an initial trial of this prototype assessment achieved good reliability between assessors for many of the indicators. One limitation of the study was that our assessors only assessed the reliability of the welfare indicators in weather that did not include heavy rain or snow. This was due to time of year and safety considerations relating to visibility on the mountains. Further testing should be carried out under different weather conditions permissible within the mountain safety advice for Snowdonia National Park.Another limitation of this study was that our trial was limited to one population of free-roaming Carneddau Mountain ponies and a relatively small sample size of individuals. Sample sizes have varied across similar studies; Viksten et al. 2017 tested their HWAP draft protocol for reliability on 37 horses from two Swedish riding schools using one assessor, while the Animal Welfare Indicators (AWIN) horse welfare assessment protocol was tested by researchers for reliability at 10 German horse farms, with a total of 435 horses being assessed by two assessors [35]. Nonetheless, this sample size reflected 16% of the total population, and this study’s aim of testing of AB and RB indicators for feasibility, reliability, and repeatability was therefore mostly achieved using this sample. This study therefore provides a useful baseline understanding for future testing and suitable evidence to be evaluated prior to the inclusion of these welfare indicators in welfare assessment templates for free-roaming horses. To assess the true potential of these indicators for free-roaming horses, further trialing will continue across different horse and pony populations. Additionally, testing will be carried out with practitioners (stock persons and grazing scheme managers), as the tool is ultimately intended for those working directly with the animals.5. ConclusionsA range of welfare assessment protocols has been developed to determine the welfare of farm, companion, laboratory, and zoo animals. These objective appraisals and formal recordings of the welfare state of an animal have been found to contribute to welfare improvements in a myriad of species. As ours was a preliminary study, the focus was on the reliability, feasibility, and repeatability of indicators rather than an assessment of the ponies’ welfare per se. This study is therefore an initial step towards designing a much-needed tool for the assessment of free-roaming native ponies. Our preliminary assessment has demonstrated that many of the trialed welfare indicators were repeatable, with many featuring good reliability. It has also established that with minor modifications, animal-based indicators commonly used to assess equines in intensive conditions could be applied in free-roaming ponies. While the further refinement of the decision trees is required, the initial trial of the selected welfare indicators has enabled the identification of potential risks for welfare compromise in free-living Carneddau ponies. Results were presented to members of the Carneddau Pony Society to aid them in the necessary mitigations protocols to safeguard pony welfare.

animals : an open access journal from mdpi

10.3390/ani11092670

PMC8472833

Animal signals can convey information about the animal’s state, but these signals can also be used to influence the behavior of others through emotional contagion. Music can influence the emotional state of human listeners and has also been used therapeutically with a variety of captive species including pets. However, the successful use of music to influence the well-being of animals must be based on an understanding of the natural communication signals of the species including the frequency range and tempos of its own communication signals. Furthermore, different types of music can induce different emotional states. In this paper, I review work using music to influence animal emotion, physiology and behavior, and I outline a theory of emotional induction that predicts what types of music stimuli are likely to influence different emotions and behavior. I will illustrate this with some examples of animal-based music. The use of music to influence the emotional well-being of our pets, farm animals and in zoological parks depends on our understanding the communication system of other species and the variety of emotional states that can be induced through different types of music. My goal is to help those managing animal facilities or advising pet owners to be more aware of the issues involved in using music with animals, as well as provide advice to researchers investigating effects of music on animals.

Playing music or natural sounds to animals in human care is thought to have beneficial effects. An analysis of published papers on the use of human-based music with animals demonstrates a variety of different results even within the same species. These mixed results suggest the value of tailoring music to the sensory systems of the species involved and in selecting musical structures that are likely to produce the desired effects. I provide a conceptual framework based on the combined knowledge of the natural communication system of a species coupled with musical structures known to differentially influence emotional states, e.g., calming an agitated animal versus stimulating a lethargic animal. This new concept of animal-based music, which is based on understanding animal communication, will lead to more consistent and specific effects of music. Knowledge and appropriate use of animal-based music are important in future research and applications if we are to improve the well-being of animals that are dependent upon human care for their survival.

1. IntroductionStudents of communication in nonhuman animals have developed two main models for what animals are communicating. The information model [1] states that animals are communicating about their internal states or about events in their environment and, therefore, providing information to recipients that will be valuable to the recipients. Signals may inform others about the caller’s internal state, what the caller may do next, or about the presence of food or potential predators in the environment. The manipulation/management model [2,3,4], in contrast, suggests that animals use communication to manipulate or manage the behavior of recipients. In this view, communicators are attempting to change or manage the behavior of recipients to the benefit of the caller, but not necessarily providing information that benefits the recipient.These models need not be mutually exclusive since both information and manipulation can be present in an organism’s communication system. However, the manipulation model has been used to predict specific acoustic features that should be effective with inducing behavior change in listeners [4]. Thus, a series of short rapid calls generally have an arousal effect, and long tonal calls have a calming effect. Dissonant or noisy calls induce fear or aggression, whereas harmonic calls induce calm or affiliative behavior.Clear evidence for the effects of different acoustic structures on inducing behavior has been shown in studies of how humans communicate with nonverbal organisms, both nonhuman animals and babies. McConnell [5] looked at how animal handlers communicated with sheep herding dogs, with horses and other working animals. Across many cultures and linguistic groups, humans used rapid staccato notes with increasing pitch to arouse animals, long, slow descending notes to slow or calm animals, and a short, sharp plosive note to stop an animal’s movement. At the same time, Fernald [6] showed that similar types of sounds, embedded in the prosodic (or musical) contours of speech, were used by parents to communicate with infants of several different cultures with the same outcomes (short upwardly rising speech, led to increased arousal, long slowly descending speech calmed the infants, and a sharp plosive sound inhibited behavior).These features appear to be auditory inducers of emotion not only in humans but in other species. Since it is unlikely that animal handlers or parents of infants are experiencing the emotions relating to these calls, the best explanation is that they are trying to manage or change the behavior of animals or infants, supporting the management/manipulation view of communication.Many of these same structures are seen in the emotional features used in human music: short quick notes are arousing, long harmonic notes are calming, dissonance induces feelings of anger or fear and harmonic patterns lead to feelings of calm and relaxation. These emotional components to music have been hypothesized by musicologists and biologists [7,8] to be present in nonhuman species and serve as the original functional origins of music. Neuropsychological studies of brain activity [9] suggest that these emotional structures of music have different and specific effects on the different brain areas associated with processing different emotions.Music has often been used by pet owners and by those involved in managed care of animals in research environments, zoological parks, and farms. However, the results of published studies have been inconsistent in the effects of music on animals. Thus, it is important to understand the reasons behind these inconsistent effects and to probe at greater depth how and when music can be used to promote the well-being of animals and when it might be detrimental.I begin by reviewing many of the studies where music has been used in attempts to alter the behavior or physiology of animals in order to attempt a synthesis of what does and does not work. I will suggest two common problems, failure to match music to animal sensory systems and being unaware that different types of music may lead to different emotional reactions. I will then review some studies that show the importance of considering and eliminating these two problems. I will conclude with some ideas about how and why music evolved and how it can be important in managing our own behavior as well as that of other humans, our pets and other animals that are dependent upon humans for their well-being.2. Studies Involving Music and Animal ResponsesI conducted a Web of Science search on 16 June 2021 using the terms “music” and “animals” and examined the reference sections of publications to discover additional studies. I present a summary of the results in Table 1. The table summarizes the results from 58 publications ranging from 1989 through June 2021. The table is organized by species with the types of stimuli presented, and the outcomes observed. Studies were included if music or musical tones were presented to animals with outcome measures of preference, discrimination, improved welfare (reduced stress behavior or increased positive behavior), or altered physiological responses (hormones, neurotransmitters, weight). These outcome measures are shown in the table adjacent to the species names.Several observations can be derived from the table. The authors do not always specify the specific music used but 36 of 58 studies (62%) reported that classical music was among the stimuli used and of these 36 nearly half specified Mozart as the composer (17 of 36). Other genres were occasionally used including Indian, African and Japanese music, Romanian folk music, hard rock, easy listening, country music and jazz. The effects of music often differ within the same species which I will summarize by general taxonomic category.Among apes, orangutans preferred silence to music [10] and four studies in gorillas played natural forest sounds, with one showing increased calming, another showed increased stereotypic behavior and two showed no change [11,12,13,14]. Rock music had no effect on gorillas and in one study slow tempo classical music reduced anxious behavior [14], whereas two studies recorded no significant effects of classical music [12,13]. Chimpanzees preferred rock music to silence in one study [15]. In another study, instrumental music increased affiliation while vocal music reduced agonistic behavior [16]. Unspecified music was said to reduce aggression and exploration while increasing social grooming and resting [17]. Chimpanzees preferred Indian and African music to silence [18], and a study of a single young chimpanzee reported a preference for consonant over dissonant music [19]. In gibbons, a mix tape of classical music produced no behavioral change in Moloch gibbons [20], whereas a study playing back the animal’s species typical songs led to increased activity in Lar gibbons [21].In monkeys, oldies radio music led to no change in blood pressure in baboons, but did reduce heart rate and activity [22]. Classical music, along with silence, produced no changes in response times of rhesus monkeys to neutral stimuli versus emotional stimuli, whereas playing noise led to delayed response time to all stimuli [23]. Marmosets and tamarins preferred Mozart to heavy metal, but preferred silence overall [24]. A study of elephants found unspecified classical music reduced behavioral stereotypes [25].In dogs, classical music increased sleep and rest behavior in three studies, whereas rock music led to increased activity and barking in three studies [26,27,28]. However, two other studies reported no effects of music on behavior [29,30]. Thus, the effect of music on dogs is quite mixed.Researchers have studied a variety of farm mammals and these also show mixed results. In lambs, “easy listening” music calmed animals but playing noise led to increased weight gain [31]. In cattle, country music led to increased approach to milking stalls, but no other types of music were tested [32]. Ponies showed no significant behavioral change to a variety of musical genres, but there was a nonsignificant tendency toward increased feeding with country music [33]. One study in piglets found no effects of music [34], whereas two other studies in piglets found that fast tempo music increased activity, play, and tail wagging [35,36].Many studies have been carried out in rodents. Studies involving playing Mozart perinatally to mice and rats improved learning in young even though the music is not in the optimal frequency range of hearing for these animals [37,38]. Possibly the rhythmic pulse of the music could be detected. Mozart’s music decreased blood pressure in hypertensive rats, whereas the 20th-century music of Ligeti increased blood pressure in one study [39] but no differences were seen in another study that compared the music of Bach with 20th-century music by Stravinsky [40]. Photoperiod influenced the effects of music [41]. One study played pure tones and found increased ghrelin levels and weight gain. Six studies on adult rats have played music by Mozart [41,44,45,46,47,48], with three reporting decreased fear, stress and anxiety [41,45,47]; two studies showed increased spatial learning and increased levels of Brain-Derived Neurotropic Factor that facilitates spatial learning [45,46], and one showed decreased blood pressure in hypertensive rats [48]. Interestingly, this last study found that the effects were the same when all frequencies below 4 kHz were filtered out, meaning that most of the music in the range of human communication was irrelevant.Several studies have looked at the effects of music on poultry. In hens, one study found that music increased stress behavior [49]. In another study, music had no effect, but noise was aversive [50]. In another study, sitar music increased synaptic binding proteins [51], increased calcium binding proteins and increased neuronal density and volume [52], whereas loud noise led to decreased neuronal density and volume [53]. Loud music enhanced auditory function in comparison with loud noise [54] and Mozart’s music reduced stress physiology, but also reduced growth [55]. Chickens preferred consonant to dissonant music [56], a finding seen only in one young chimpanzee in other animal studies [19]. In a review of several studies on chickens, the playing of happy music increased “happy” behavior, whereas playback of sad music increased “sad” behavior. Both types of music decreased anxiety and anger. Happy music also induced behaviors similar to those induced by oxytocin, a hormone thought to be involved in affiliative bonding behaviors. There was also an increase in brain levels of the reward neurotransmitters, dopamine and norepinephrine. The authors of this review had previously done similar studies in rats with no effects of music on behavior and observed that most human-based music was outside the range of natural communication in rats, but not of chickens [57].In other birds, pigeons could discriminate between Bach and Stravinsky and could generalize this discrimination to other similar composers [58]. Some Java sparrows could learn to discriminate between Bach and Schoenberg and generalize to composers similar to that which was reinforced [59], and some sparrows showed a preference for Bach [60].Finally, research in fish has shown a variety of effects of music. In carp, both Mozart and Romanian folk music increased growth and reduced stress [61,62]. Mozart also increased growth in seabream [63]; but in goldfish, lute violin music had no effect on growth or weight [64]. Goldfish could discriminate between Bach and Stravinsky but preferred neither [65]. Finally, music presented at different tempos to turbot showed increased growth with slow tempos and impaired growth with fast tempos [66].In summary, the results of this survey indicate that music has a variety of effects that vary within and between species. There are few consistent results and it would be difficult to find any consistent effects of music, despite the large number of studies. In the next section, I will suggest why this may be the case and suggest some new ways to think about the use of music in animal well-being.3. A Critique of Past Research and Suggested SolutionsThere are two main reasons for the failure of music to produce consistent results when played to animals: (1) animal sensory and communication systems differ from our own and vary according to species; (2) music is highly variable both between and within genres. So music must be selected carefully to produce the desired emotional effects. I will consider each of these in turn and then provide some examples of successful development of species—appropriate music in tamarin monkeys and cats.3.1. Matching Sensory SystemsIt is widely known that some animals perceive and communicate about the world in very different ways from humans. Some well-known examples are the use of infrasound (frequencies below the range of human hearing) by elephants and many cetaceans (dolphins and whales) and the use of ultrasound (frequencies above the range of human hearing) by bats and many rodents. Thus, it might be obvious that for music to be effective for these species, it would need to be outside the range of human hearing. What is less obvious is that even within the range of human hearing ability, different species may use frequencies very different from what we use for human communication and music (see [67]).A couple of compelling examples illustrate this. Rats communicate using much higher frequencies than humans and many of their vocalizations extend into the ultrasonic range. Thus, Akiyama and colleagues [48] found that when they filtered out frequencies in the human range (below 4 kHz), playing Mozart to hypertensive rats was just as effective in reducing blood pressure as when the full auditory spectrum was available. The only thing that mattered to rats was whatever was transmitted above 4 kHz. Many inexpensive speakers may not be able to produce the higher frequencies perceived by rats so some failures to find effects in rodents may be due to this. Panksepp and Bernatzky [57] deliberately switched their work from rats to chickens when they found rats unresponsive to the music they played. With chickens, they found clear effects of music on behavior and on brain neurotransmitters.Cotton-top tamarins, small monkeys from northern South America, preferred Mozart to heavy metal when tested, but preferred silence to Mozart [24]. The authors concluded that these monkeys had no interest in music. However, the natural calls of these monkeys average about three octaves higher than human speech and music, so they may be responsive to music within their frequency range. In Section 3.3, I will describe results using species-appropriate music with this species of tamarins.Tempo or pulse is often not considered in selecting music, but music with tempos that match the resting heart rate of humans tends to be calming, whereas tempos that match higher heart rates induced by exercise or dance are more arousing [68]. It seems likely that this principle would hold with other species as well and, in general, smaller bodied animals have higher resting heart rates. It is also be important to acknowledge the range of heart rates of another species when determining what tempos are likely to be most effective with that species.Studies have shown that tempos approaching natural tempos are important. For example, chicks prefer tempos that are most similar to maternal call rate [69] and are attracted more to tempos approaching patterns of natural vocalizations [70]. Some species have shown the ability to match different tempos (cockatoos [71], Java sparrows [72] sea lions [73]). Harbor seals [74] show trajectories toward adult tempos over development. A single bonobo was resistant to entrainment by experimenters, having a preferred rate of drumming [75].This critique has focused on cases where researchers have not fully considered the sensory abilities of the species they are testing. There are, of course, some other research papers where these points have been considered.3.2. Music Genres Are Not Uniform in Emotional EffectsThe second main reason for disparate results is that many researchers have assumed that a specific genre of music has a uniform emotional effect. Thus, classical music is often thought to be relaxing (or boring), with rock music being arousing. However, there is considerable variation within each genre. For example, much of Mozart’s music can be considered upbeat and arousing and the specific pieces used in many of the studies with animals would be arousing to human listeners. At the same time, some classical music can be relaxing and calming, while still other pieces may be able to induce fear or anger. There has been considerable research on how humans perceive emotions in music and Table 2 summarizes several principles that appear to be important in how humans perceive emotions in music. These principles hypothesized by [76] have been validated in experiments with musicians [77], and Western nonmusicians tested with both Western music [78] and unfamiliar Indian ragas [79].If the work on how humans communicate with working animals and with nonverbal infants can be generalized to other species, then the principles shown in Table 2 should be applicable to other nonhuman species as well. Thus, if one wants to calm an animal (perhaps a recent arrival at a shelter, or an animal in a clinic) then music should be slow, with long notes (legato), with generally descending pitches, relatively low amplitude and harmonic intervals. If one wishes to arouse an animal (such as one showing depression or lethargy), then the tempo should be fast, the changes in pitch generally ascending, with short notes that have a fast attack speed (staccato) and again harmonic intervals. Music that induces fear is usually rapid in tempo, high pitched with staccato notes, loud amplitude and dissonant intervals. Finally, threat is conveyed by a moderate tempo, low dissonant notes, moderate amplitude and dissonant intervals with a fast attack speed. Most people working with captive animals or pets would probably wish to avoid inducing these last two states in animals, but it is important to be aware that the use of these features may possibly induce unwanted behavior. Table 1 gives some examples of unspecified slow music calming animals and faster music arousing animals (see, for example, the entries for pigs).Since consonant music is hypothesized to be therapeutic for most uses with animals, it is important to consider some additional literature on how animals perceive consonance. A single chimpanzee was shown to prefer consonant to dissonant music [19] and marmosets and tamarins preferred the consonant music of Mozart to heavy metal [24]. Birds have also been shown to prefer consonant music [56,81] but although rats can be trained to discriminate between consonance and dissonance, they do not generalize to other stimuli nor show the processing advantage that humans show to consonant versus dissonant music [82,83]. These data show that birds and primates do prefer consonant music, and it is not clear for other taxa. I suggest that consonant music be used with all music provided for enrichment.Thus, the specific structure of the music to be used must be chosen to match the goals of those working with animals. If one is using music that is within the perceptual range of the species and music that has the specific structural features that are predicted to induce the desired behavior, then music may be used successfully [84]. In the next section, I describe research on music and behavior of animals that incorporated these points into the experimental design.3.3. Effects of Species-Appropriate MusicTo test these ideas, musician and composer David Teie and I have collaborated on two studies using species-appropriate music. The first study [85] tested captive cotton-top tamarins, previously shown by McDermott and Hauser [24] to prefer Mozart to heavy metal, but silence to either type of music. These monkeys are small (c.a. 500g), and in the wild live in forested habitats in Northern Colombia and use vocalizations as their main communication modality. Since the normal communication range of these animals is about three octaves higher than human speech and music and their resting heart rate is nearly three times that of an adult human, we composed the species-appropriate music with these facts in mind. We tested two examples of each of two types of music. The first set of compositions used the principles hypothesized to be involved in creating calming music—long notes, harmonic structure, lower amplitude, and slow attack speed with tempo at about the resting heart rate of tamarins. We designed the second set of compositions to be highly arousing and fear inducing with short staccato notes, dissonant or noisy features, and rapid tempos at twice the resting heart rate of tamarins. We tested two additional sets of human-based music, one pair with calming features and the other set, typical of heavy metal that would be arousing to humans. Thus, there were eight different pieces of music in total.We tested pairs of tamarins with each of the eight musical examples, each edited to 30 s duration. Each piece was used only once with each pair in order to avoid habituation. In each session, we waited for the animals to settle after the observer entered the colony rooms and then conducted baseline observations for five minutes. Then, the observer played a 30 s. sample of music and recorded behavior for an additional five minutes. Data were analyzed by pairs since the response of one animal is not independent of its mate. We compared responses to calming versus arousing tamarin music in the five minutes after the play back, compared the five minutes baseline response to the behavior following the music, and compared responses to species-appropriate music to the similar types of human-based music.The tamarins exhibited less motor activity and engaged in more eating and drinking behavior after the tamarin calming music, both compared with baseline behavior and compared with the arousing music. In contrast, tamarins showed increased activity, increased levels of anxiety behavior and increased huddling and grooming (social reassurance) with their mates after hearing the tamarin arousing music. In contrast, the tamarins showed few significant behavioral responses to human-based music. The only exception was some decreased locomotor behavior after being presented with human heavy metal music. This result seemed paradoxical at first, until we realized that the tempos of the human heavy metal music matched the resting heart rate of the tamarins. Thus, music that is arousing for us could actually be calming for the tamarins.This study illustrates the two main points: the tamarins were generally uninterested and unaffected by the presentation of human-based music, showing that the musical range must match the frequency range and tempos of the species being tested. Second, the structures of music that affect different emotional responses in human listeners are also effective in nonhuman animals, if and only if, there is a perceptual match to the animal’s sensory system. Importantly, this was not a test of preference. We do not know if the animals enjoy or would avoid this music, but what we do know is that their emotional tone was affected by the music. A final observation is that most humans who have listened to this music have found it aversive. Thus, just as tamarins may find human music aversive, so humans may find tamarin music aversive.Our second study looked at cats [86]. After the tamarin study was published, we heard from many people who use music as a putative therapy with their pets. However, as we talked with these people, we learned that each person was convinced that their pets liked the same music as they enjoyed; one person liked classical, another liked heavy metal, a third liked country music and that was what they presented to their pets. Do pets respond to human music and, if so, how? Do they find their owner’s preferred music pleasant, aversive or irrelevant?We chose to study cats rather than dogs since the body sizes, heart rates and voices of cats are more homogeneous across breeds than dogs. The study design was similar to the previous study. Two pieces of music were composed for cats. The main theme averaged an octave higher than human music. Glides are important in natural cat vocalizations and were also incorporated in the music. The tempos of purring and suckling formed the basis of the two cat music pieces and two classical music examples were selected that were similar in structure from a human perspective. We tested cats in their own homes where we would place two speakers 1 m to the left and right of our playback computer. Each musical piece was three minutes long. We played the four selections counter-balancing the order of presentation and the side of the speaker broadcasting the music. We compared the responses of cats to the cat music compared to the human music. The cats responded with significantly shorter latency and showed significantly more interest (orienting, approaching speaker, rubbing speaker) in the cat music than in the human music. Cats that had been restless during the baseline period became calmer upon hearing the cat music. Several people have told us anecdotally in response to the paper that the cat music was especially useful with helping their shelter- and feral-adopted cats to become more relaxed and interactive with their owners. A recent study [87] compared the effects of cat music, classical music, and silence on stress-related behaviors of cats during veterinary clinic examinations and found that the cat music significantly reduced stress behavior when compared to both silence and classical music. In our study the cats were essentially indifferent to the classical music excerpts, showing neither positive nor negative reactions. Thus, there is likely no harm in playing human classical music to cats, but there is no benefit either. We do not know about the effects of other types of music.Taken together, these studies suggest that species-appropriate music can lead to behavioral change in nonhuman animals, and that the specific features of the music can have either calming or arousing effects, so that care must be taken to assure that the music presented is structured to produce the desired results.4. DiscussionMusic, as we know it today, is a structurally complex system of notes, chords, tempos, themes and variations, with a complexity that some authors feel makes it similar to language [88]. However, music also evokes strong emotions (joy, sadness, anger, fear, calmness, happiness, surprise, disgust), with brain activation in areas associated with these emotions being activated by music [9,89,90]. Although the structural complexity of modern human music may have co-evolved with language, the emotional components of music may have a long evolutionary history [7]. It is likely that these emotional components have made music adaptive to humans. The induction of emotions in listeners can foster cooperation, increase social activity, create a sense of group cohesion and induce empathy. An experimental study of four-year-old children showed that children who sang music together were more likely to cooperate and help each other than children who merely recited the same words [91]. Music has also been shown to assist in healing during illness in many studies [92,93,94]. If music has this value to humans, might it not also be helpful to animals, leading to more closely integrated social groups, an ability to coordinate actions in defense? This question has led to the many studies in Table 1 evaluating music on the behavior and physiology of animals.The vocal signals of many animals have musical components. Indeed the early studies of vocalizations before the widespread use of spectral analyses used musical notation to denote animal signals (e.g., [95]), and we call the territory defense and courtship sounds of birds “songs”. The studies of McConnell [5] and Fernald [6] on how humans use musical features to influence the behavior of working animals and nonverbal infants, respectively, provided additional impetus for the use of music with animals in managed care (zoological parks, farms, laboratories and pets). Because music written for humans is the most accessible form of music for us, it seemed quite logical to assume that human-based music would also be of benefit to animals. However, as reviewed here, it is clear that human-based music can have a variety of effects (positive, negative, and neutral) within and between species.This variety and inconsistency of effects of music on animals can be explained in part by the use of a variety of different types of music stimuli which makes direct comparisons difficult. The mixed results may also be due to failures to realize that the vocal communication systems of animals may be fundamentally different from our own with different sensitivities to both pitch and tempo. The success of playing only the frequencies of a Mozart piece that are above 4 kHz on reducing blood pressure in hypertensive mice [48], the failure of musical manipulations that worked in chickens to have any effects in rats [57], and the ability to manipulate behavior in tamarins with music in their pitch range and tempos [85] suggest that it is important to match music to the perceptual abilities of the species.It is equally important to understand the structural features of music that can lead to different emotional responses. Thus, it is important to articulate goals and use music that is appropriate toward achieving those goals. The music one uses to arouse an animal that appears lethargic or depressed will be very different from what one uses to calm a hyperactive animal, and, presumably, no one wants to use fear- or threat-inducing music very often. Additionally, it is important to be aware that within each musical genre there is a variety of emotions expressed musically. No one genre is uniformly calming, arousing, or threatening.Armed with knowledge of a species’ perceptual abilities and with a clear view of the goals of playing music, one can then locate or create the music that will be best for improving the well-being of animals in human care. Those designing future research on the effects of music on animals should also pay attention to these points. Then, we will be able to say (paraphrasing William Congreve, 1670–1729) that “music hath charms to soothe (and arouse) the savage beast”.

animals : an open access journal from mdpi

10.3390/ani11113077

PMC8614364

Oocyte cryopreservation allows female gametes to be conserved for long periods, which would be of benefit for mares of high genetic merit, but its efficiency is not satisfactory yet. Therefore, the aim of this study was to optimize a vitrification protocol for equine oocytes using a systematic approach. We performed a side-by-side comparison of different cryoprotective agents (CPAs) during the vitrification and warming of equine oocytes. In the first experiment, a fixed mixture of CPAs that enter the oocyte was used, and three sugars were compared, which cannot penetrate the oocyte but provide protection through an osmotic effect. In the second experiment, one sugar from the first experiment was selected to compare three mixtures of CPAs that enter the oocyte. Overall, the embryo development was reduced after oocyte cryopreservation when compared to fresh oocytes. Yet, we were able to produce embryos with all six cryoprotective agent mixtures, and we identified one promising combination of cryoprotectants, consisting of propylene glycol, ethylene glycol, and galactose, that resulted in blastocyst rates in the same range as the fresh control group.

Equine oocyte vitrification would benefit the growing in vitro embryo production programs, but further optimization of the protocol is necessary to reach clinical efficiency. Therefore, we aimed to perform a direct comparison of non-permeating and permeating cryoprotective agents (CPAs) during the vitrification and warming of equine immature oocytes. In the first experiment, cumulus oocytes complexes (COCs) were vitrified comparing sucrose, trehalose, and galactose in combination with ethylene glycol (EG) and dimethyl sulfoxide (DMSO). In the second experiment, the COCs were vitrified using three mixtures of permeating CPAs in a 50:50 volume ratio (ethylene glycol-dimethyl sulfoxide (ED), propylene glycol-ethylene glycol (PE), and propylene glycol-dimethyl sulfoxide (PD)) with galactose and warmed in different galactose concentrations (0.3 or 0.5 mol/L). Overall, all the treatments supported blastocyst formation, but the developmental rates were lower for all the vitrified groups in the first (4.3 to 7.6%) and the second (3.5 to 9.4%) experiment compared to the control (26.5 and 34.2%, respectively; p < 0.01). In the first experiment, the maturation was not affected by vitrification. The sucrose exhibited lower cleavage than the control (p = 0.02). Although the galactose tended to have lower maturation than trehalose (p = 0.060) and control (p = 0.069), the highest numerical cleavage and blastocyst rates were obtained with this CPA. In the second experiment, the maturation, cleavage, and blastocyst rates were similar between the treatments. Compared to the control, only the ED reached similar maturation (p = 0.02) and PE similar cleavage (p = 0.1). The galactose concentration during warming did not affect the maturation, cleavage, or blastocyst rates (p > 0.1), but the PE-0.3 exhibited the highest blastocyst rate (15.1%) among the treatments, being the only one comparable to the control (34.2%). As such, PE–galactose provides a valuable option for equine immature oocyte vitrification and should be considered for the future optimization of the protocol.

1. IntroductionOocyte cryopreservation allows conserving both female genetics and fertility. It has gained increased interest during the last few years in the horse, especially as a complement for assisted reproductive techniques (ARTs), like ovum pick-up (OPU), intracytoplasmic sperm injection (ICSI), and cloning. The application of these ARTs has increased dramatically due to the rise in their efficiency and the consistent pregnancy rates that can be achieved [1,2,3,4]. Gamete cryopreservation would further increase ICSI flexibility by allowing to postpone the decision on the fertilizing stallion. Moreover, oocytes could be salvaged from recently deceased animals in places where ICSI is not available; oocytes might be collected and stored outside the reproductive season, and good-quality germ lines of young mares could be preserved before they are enrolled in competition [5,6]. Finally, horse oocyte banks would allow the worldwide spread of valuable female genetics for commercial purposes, for breeding programs of endangered equid breeds and species, or for research in any ART in horses [5,7]. However, the mammalian oocyte is one of the hardest cells to cryopreserve [8], and, up to now, the efficiency of equine oocyte cryopreservation is limited. Only a few foals have been born resulting from mature or immature oocytes cryopreserved by vitrification [9,10]. The vitrification of immature and in vitro matured equine oocytes compromises embryo development after ICSI [11,12,13,14]. Hence, in order to exploit the application potential of equine oocyte vitrification, further optimization of the vitrification protocols is necessary (for review see De Coster & Angel-Velez, et al., 2020 [5]).During vitrification, high concentrations of cryoprotective agents (CPAs) (30–40% v/v) combined with a high cooling rate are required to prevent intra- and extracellular ice formation and achieve stable vitrification [15,16]. Both permeating and non-permeating solutes may be used as CPAs. Permeating CPAs, such as ethylene glycol (EG), dimethyl sulfoxide (DMSO), propylene glycol (PG), and glycerol, are small molecules that can penetrate the plasma membrane and form hydrogen bonds with water molecules to lower the freezing point, which, combined with high cooling and warming rates, can prevent significant intracellular and extracellular ice formation during vitrification and warming [17]. While these actions largely depend on the colligative properties of the CPAs, the effective protection provided by the specific permeating CPAs varies depending on their membrane permeability and other properties, as well as on the maturation stage of the oocyte and the temperature at which they are introduced [18]. However, high concentrations of CPAs can also have toxic effects, and an optimal balance between protection against cryoinjury and toxicity needs to be established. Combining two or more permeating CPAs can decrease the overall CPA toxicity because of the lower concentrations of each CPA used [19,20]. This was also demonstrated by Szurek and Eroglu, 2011 [21] in mouse oocytes, in which 1.5 M PG induced a significant increase in oocyte degeneration (54.2%), while it was possible to avoid the toxicity of PG by decreasing its concentration to 0.75 M and combining it with 0.75 M DMSO. Likewise, Somfai et al. [22] showed that the combination of PG and EG provided greater embryo development after the vitrification of the germinal vesicle (GV)-stage porcine oocytes than did either CPA alone. In mature human oocytes, EG has approximately half the permeability of PG and DMSO, but it is preferred due to its low toxicity [18]. In the horse, EG was used in combination with non-permeating CPAs to preserve oocytes [23]. However, it was not until combinations of EG–DMSO–sucrose [10] and PG–EG–trehalose [11,12] were used that blastocyst development was obtained from vitrified immature equine oocytes. Despite this progress, the blastocyst rates from vitrified equine oocytes remain severely affected, and the optimization of the protocol is compulsory. Only one study has directly compared protocols using different CPAs in the horse, in which PG –EG–trehalose seemed to be the most effective combination [11], but toxicity was still present and no direct comparisons between CPAs at the same concentrations were performed.Non-permeating CPAs are solutes that do not penetrate the ooplasm and remain in the extracellular compartment during cooling to promote glass formation [24]. Non-permeating CPAs include sugars (e.g., sucrose and trehalose), macromolecules (e.g., Ficoll and bovine serum albumin), and synthetic (co)polymers, such as synthetic ice blockers (e.g., polyvinylpyrrolidone, polyvinyl alcohol, and SuperCool X-1000) [25]. Non-permeating CPAs may be less cytotoxic than permeating ones, although they may result in more mechanical stress from the increased shrinking of cells. The addition of non-permeating CPAs to the vitrification media contributes to the viscosity and tonicity, allowing lower concentrations of permeating CPAs to be used without compromising the vitrification properties [26]. In equine oocyte vitrification, sucrose is the commonly used non-permeating CPA [9,10,27,28,29]. However, trehalose has recently shown promising results. Clerico et al., 2021 [14] reached an encouraging blastocyst rate of 15% (9/56) using trehalose combined with EG–DMSO. Previously, Canesin et al., 2017 [11] compared different treatments and obtained the highest vitrification efficiency with a vitrification medium containing trehalose (42% maturation (10/24), 80% cleavage (8/10), and 10% blastocysts (1/10)), whereas no blastocysts were obtained from sucrose-containing media [11], but, in the latter study, the extracellular CPAs compared had different molar concentrations. In a study in which equal molar concentrations of non-permeating CPAs were compared, a similar pronuclear formation was found for horse oocytes vitrified in a DMSO–EG medium containing sucrose or trehalose [30]. In humans and mice, the oocyte vitrification in a trehalose-containing medium was also associated with higher blastocyst rates than vitrification with sucrose [31], but the differences were not always significant [32,33], probably due to variations in the protocols and concentrations. Besides these disaccharides, which have been the only non-permeating CPAs tested in equine oocytes vitrification up to now (for review, De Coster & Angel-Velez, et al., 2020 [5]), monosaccharides may provide a valuable alternative. Monosaccharides (galactose, fructose, glucose) seem to be more effective osmotic buffers than disaccharides during vitrification and warming [34], and they have been used successfully for the vitrification of bovine [35] and feline [36] oocytes, as well as for bovine [37], porcine [38], dromedary camel [39], alpaca [40], and equine embryos [37,41,42].The warming procedure is as crucial in oocyte survival as vitrification, and relatively little attention has been paid to the warming systems in all species. During the thawing procedure, the oocyte should slowly recover its original volume to avoid osmotic shock or over-swelling [43]. Commonly, hyperosmolar solutions (1 mol/L) with mono- and disaccharides have been used in warming solutions as an osmotic counterforce in restricting water permeation into the oocyte, preventing a swelling injury [44,45,46]. However, oocytes should be removed from this solution before they start shrinking excessively, which may start (e.g., after 60 s) after sufficient CPA has left the cells and the continued efflux of CPA is accompanied by the efflux of water as a result of the extracellular hyperosmotic solution [47,48]. Direct warming to an iso-osmolar base medium (0 mol/L sugar) has been tried [12]; still, this could create an over-swelling due to the fast inward diffusion of water by osmosis [49]. Therefore, different warming concentrations should be explored to further optimize the warming procedure.A superior standard protocol describing all the aspects of oocyte cryopreservation has not yet been identified in the horse. Multiple factors influence the success of equine oocyte vitrification, and one crucial technical aspect is the type of CPA [5]. Therefore, the aim of the present study was to determine the optimal combination of CPAs for the vitrification of immature equine oocytes. To do so, we compared (1) three sugars as non-permeating CPAs (sucrose vs. trehalose vs. galactose) with EG and DMSO as permeating CPAs, and (2) three combinations of permeating CPAs (EG–DMSO vs. PG–EG vs. PG–DMSO) with galactose as a non-permeating CPA. Moreover, the efficiency of two different concentrations of galactose in the warming solution (0.3 vs. 0.5 mol/L) was assessed.2. Materials and Methods2.1. Media and ReagentsDulbecco’s Modified Eagle Medium Nutrient Mixture F-12 (DMEM/F-12), Tissue Culture Medium-199 (TCM-199) with Hanks’ salts, and Tissue Culture Medium-199 with Earle’s salts were purchased from Life Technologies, Merelbeke, Belgium. Unless otherwise stated, all other components were obtained from Sigma, Bornem, Belgium.2.2. Collection of Equine Immature OocytesEquine ovaries were obtained from a local slaughterhouse and transported in an insulated box to the laboratory at room temperature within 1 h. All follicles between 5 and 30 mm were aspirated using a 16-gauge needle attached to a vacuum pump (100 mm Hg), scraped with the aspirating needle, and flushed with prewarmed flushing medium (Equiplus, Minitube, Tiefenbach, Germany). The aspirated fluid was collected in sterilized glass bottles, and the bottom content was pipetted several times to a 100/20 mm petri dish until no more oocytes were found. All cumulus–oocyte complexes (COCs) were recovered in TCM-199 with Hank’s salts (Gibco), washed twice, and pipetted with a 200 µm denudation tip (EZ-tip, Origio, Vreeland, the Netherlands) to remove the outer cumulus cells, leaving the corona radiata. Then, COCs were randomly assigned to vitrification or in vitro maturation (IVM; control). Denuded, partially denuded, and clearly expanded COCs surrounded by a hyaluronan-rich matrix were excluded from all experiments.2.3. Oocyte Vitrification and WarmingThe composition of the vitrification and warming solutions is described below in the experimental design and summarized in Table 1. For all experiments, vitrification and warming steps were performed on a heated plate at 39 °C. The vitrification method and the device used for vitrification were based on the “short vitrification protocol” described by Ortiz-Escribano et al., [10] with minor modifications. All oocytes assigned to different vitrification treatments were transferred to a small petri dish (35/10 mm) with 4 mL of base solution (BS), which was equal for all experiments (TCM-199 with Hank’s salts supplemented with 0.4% (w/v) bovine serum albumin (BSA) (A6003)). Then, four to six oocytes at a time were placed and washed thorough two droplets of 100 µL of equilibration solution (ES) for 25 s. Finally, oocytes were transferred to a 100 µL droplet of vitrification solution (VS) for 15 s, loaded onto a custom-made minimal volume (<1 µL) cryo-device (Figure 1), and plunged into liquid nitrogen. The time between the placement of oocytes in the VS and the immersion of the device into the liquid nitrogen was 30–45 s. Oocytes were loaded using a 200 µm pipette to minimize the volume surrounding the oocytes. Moreover, extra medium surrounding the oocytes was removed with the pipette by capillarity. After at least one week of storage in liquid nitrogen, oocytes were warmed. For warming, the cryo-device was transferred into 4 mL of warming solution 1 (W1), containing BS supplemented with 0.3 or 0.5 mol/L sugar, depending on the experiment, for 5 min. Then, all oocytes were moved to BS until the warming of all oocytes was completed (Table 1).2.4. In Vitro Maturation and ICSIVitrified–warmed or fresh oocytes were transferred to maturation medium (TCM-199 with Earl’s salts (Gibco) containing 10% (v/v) FBS (Gibco), 9.4 μg/mL follicle-stimulating hormone, and 1.88 μg/mL luteinizing hormone (Stimufol, Reprobiol, Ouffet, Belgium)). Maturation was performed in groups of 20–40 COCs in 500 μL maturation medium under paraffin oil (Cooper Surgical, Venlo, The Netherlands). Oocytes were matured at 38.5 °C in 5% CO2 in air for 30 h on average (min: 27.5 h; max 32.5 h). Frozen-thawed semen from a single stallion of proven fertility was used for ICSI. Spermatozoa were selected using a 45–90% Percoll density gradient centrifugation for 40 min at 750× g at 26 °C. After removal of the supernatant, the sperm pellet was washed in 5 mL of G-MOPS (Vitrolife, Londerzeel, Belgium) and centrifugated for 10 min at 400× g at 26 °C. The supernatant was removed again, and the sperm pellet was resuspended in 300 μL of G-MOPS and kept at room temperature until used for ICSI. Mature oocytes (MII), indicated by an extruded polar body, were injected by piezo drill, and presumptive zygotes were cultured as described by Papas et al., 2021 [4]. Cleavage rate was evaluated 2 or 3 days after ICSI, and blastocyst development was monitored daily from day 7 until 10 post ICSI.2.5. Experimental Design2.5.1. Experiment 1: Effect of Sucrose, Trehalose, or Galactose as Non-Permeating CPAs in VS and in WS on Maturation, Cleavage, and Blastocyst RatesCOCs were obtained as described above and immediately vitrified in three groups using 0.5 M of one of three sugars in the VS: sucrose (S1888; n = 155), trehalose (T0167; n = 160), or galactose (G5388: n = 153). For this experiment, ES contained BS with 10% (v/v) EG (#102466) and 10% DMSO (#D2650), and VS contained BS with 20% EG, 20% DMSO, and 0.5 M of each sugar (sucrose, trehalose, or galactose). During vitrification, the groups were alternated to keep the oocyte handling time similar. All oocytes were warmed with 0.5 mol/L of the sugar (sucrose, trehalose, or galactose) in WM (Table 1). A control group (n = 173) with non-vitrified oocytes was included in every replicate (five replicates). Oocytes of each treatment were warmed consecutively and placed immediately in IVM. The order of warming was considered for the ICSI to keep the maturation duration equal for all groups.2.5.2. Experiment 2: Effect of Three Different Mixtures of Permeating CPAs and Two Different Warming Regimens on Maturation, Cleavage, and Blastocyst RatesCOCs were vitrified as described above using three different permeating CPA mixtures in a 50:50 ratio: EG–DMSO (ED), PG (#P4347)–DMSO (PD), and PG–EG (PE) (Table 1). For vitrification, ES consisted of BS with 20% (v/v) of the CPA mix, and VS of BS with 40% (v/v) of the CPA mix and 0.5 M galactose. For warming, two different concentrations of galactose were used, resulting in six groups (ED-0.5 (n = 110), ED-0.3 (n = 85), PD-0.5 (n = 115), PD-0.3 (n = 79), PE-0.5 (n = 107), and PE-0.3 (n = 90)). A control group (n = 242) with non-vitrified oocytes was included in every replicate (four replicates).2.6. Statistical AnalysisAll statistical analyses were performed using R-core (version 3.6.1; R Core Team, Vienna, Austria). The oocyte/zygote/embryo was considered as the unit of interest. Generalized mixed effects models were used to analyze the data. In the first experiment, we evaluated the effect of the three sugars as non-permeating CPAs (control vs. sucrose vs. trehalose vs. galactose) on maturation, cleavage, and blastocyst development. In the second experiment, the effects of the CPA mixture (control vs. ED vs. PD vs. PE), the galactose concentration in the WM (0.5 mol/L vs. 0.3 mol/L), and their interaction on maturation, cleavage, and blastocyst rates were assessed. Cleavage and blastocyst rates represent the percentage of cleaved embryos or blastocysts, respectively, per injected oocyte. The control group was resampled randomly to balance lower numbers of injected oocytes in cleavage and blastocyst rate of the interaction (ED-0.5 vs. ED-0.3 vs. PD-0.5 vs. PD-0.3 vs. PE-0.5 vs. PE-0.3 vs. control). For all the models, the replicate was set as random. Results are expressed as least square means and standard errors. The differences between treatment groups were assessed using Tukey’s post hoc test. The significance and tendency levels were set at p < 0.05 and p < 0.1, respectively.3. Results3.1. Experiment 1: Effect of Non-Permeating CPAs on Maturation, Cleavage, and Blastocyst Rates3.1.1. Comparison among All Vitrified and Non-Vitrified OocytesOverall, when we compare all the oocytes vitrified with EG–DMSO and different sugars as one group, they exhibited maturation rates (51.0 ± 2.5%) similar to fresh oocytes (56.9 ± 3.9%; p = 0.19). However, the cleavage (75.3 ± 4.7%) and blastocyst rates (26.5 ± 5.7%) were higher in the fresh oocytes than in vitrified ones (62.1 ± 3.19 and 5.48 ± 1.6%, respectively; p < 0.03). In general, the descriptive kinetics of development showed that the blastocysts developed from fresh oocytes occurred earlier, on average 8.5 days after the ICSI, compared with 9.3 days on average for the blastocysts developed from vitrified oocytes (Table 2).3.1.2. Effect of Sucrose, Trehalose, and Galactose as Non-Permeating CPAs during VitrificationOnce the vitrification with each sugar was assessed individually, the comparison between the different sugars showed no significant difference in the maturation rate amongst the vitrification groups (sucrose: 52.4 ± 4.1%; trehalose: 57.4 ± 4.1%; galactose: 43.1 ± 4.1%), and the control (56.9 ± 4.0%; p > 0.05) (Figure 2). However, galactose tended to result in a lower maturation rate compared to trehalose (p = 0.06) and the control group (p = 0.069). The cleavage rates were not different either among the sugar treatments (sucrose: 53.2 ± 5.6%; trehalose: 61.8 ± 5.2%; galactose: 73.4 ± 5.5%), but the cleavage rate after the vitrification with sucrose was significantly lower than that of the control (75.3 ± 4.8%; p = 0.02) and tended to be lower than that of galactose (p = 0.06). The blastocyst rates for all the vitrified groups (sucrose: 5.0 ± 2.5%; trehalose: 4.3 ± 2.2%; galactose: 7.6 ± 3.4%) were lower compared to the control group (26.5 ± 5.7%; p < 0.04) (Figure 2).3.2. Experiment 2: Effect of Three Different Mixtures of Permeating CPAs and Two Different Warming Regimens on Maturation, Cleavage, and Blastocyst Rates3.2.1. Comparison among All Vitrified and Non-Vitrified OocytesWhen all the vitrified oocytes, combining the results for the three CPA mixtures, are compared with fresh oocytes, the vitrified oocytes exhibited an overall reduction in the maturation, cleavage, and blastocyst rates (41.2 ± 2.1, 56.6 ± 3.0, and 6.3 ± 1.6%, respectively) when compared to fresh oocytes (58.6 ± 3.4, 71.1 ± 4.3, and 34.2 ± 5.1%, respectively; p < 0.01). In general, the blastocyst development from the fresh oocytes also occurred faster, on average 8 days after the ICSI, compared with 8.8 days for the blastocyst development from the oocytes vitrified with different CPAs mixtures (Table 3).3.2.2. Effect of Three Different Mixtures of Permeating CPAs on Maturation, Cleavage, and Blastocyst RatesIn the direct comparison of the three permeating CPA combinations, the maturation rate was higher in the control (58.6 ± 3.4%) than in PD (44.3 ± 3.6%; p = 0.02) and PE (42.6 ± 3.5%; p = 0.007), while the ED reached a maturation rate comparable to the control (48.7 ± 3.6%; p = 0.1) (Figure 3). The cleavage rate was similar among the vitrification treatments. In comparison with the fresh oocytes, the cleavage rate in the PE (65.5 ± 5.3%) was similar to the control (77.1 ± 4.3%; p = 0.3), while it was significantly lower in the PD (51.1 ± 5.5%; p = 0.002) and ED (53.5 ± 5.3%; p = 0.005). The blastocyst rates were lower for all the vitrified groups (ED = 6.2 ± 2.6%, PD = 3.5 ± 2.0%, and PE = 9.4 ± 3.3%) compared to the control (34.2 ± 5.1%; p < 0.01) (Figure 3). The embryo development from the vitrified oocytes occurred between 8 and 9.1 days after the ICSI (PD: 8 days; PE: 8.9 days; and ED: 9.0), while the average was 8 days in the control (Table 3).3.2.3. Effect of Two Warming Regimens on Maturation, Cleavage, and Blastocyst RatesFirst, to evaluate the effect of the warming concentration, an overall comparison between the oocytes warmed in 0.5 mol/L and 0.3 mol/L did not display differences in the maturation (46.1 ± 3.0 vs. 44.1 ± 3.1%; p = 0.6), cleavage (60.1 ± 4.0 vs. 51.8 ± 4.7%; p = 0.1), or blastocyst rates (4.6 ± 1.7% vs. 8.9 ± 2.7%; p = 0.2). The effect of the interaction between the mixtures of the permeating CPAs in the ES and VS and galactose concentration in the warming medium (0.3 or 0.5 mol/L) on the maturation, cleavage, and blastocyst rates are shown in Figure 4 and Supplementary Table S1. No differences in the maturation or cleavage rates were found among the vitrification groups. However, the PE-0.3 numerically had the highest blastocyst rate after vitrification (15.1%), which was not significantly lower than that of the control (p = 0.16), while all the other vitrification groups had significantly lower blastocyst rates compared with the control. No significant differences were obtained between the six treatment groups (Figure 4 and Table S1). The kinetics of the development of the embryos obtained from the fresh oocytes and from oocytes vitrified with different CPAs mixtures and warmed with different galactose concentrations is represented in Table 4. Although the average day of blastocyst formation seems to be faster after warming in 0.5 mol/L galactose (7.9 days) than after warming in 0.3 mol/L (9.2 days), this difference is not significant due to the low numbers of embryos (Table 4).4. DiscussionThis study is the first to perform a side-by-side comparison of three mixtures of permeating CPAs plus three non-permeating CPAs in equal concentrations for the vitrification of immature equine oocytes. Our study demonstrates that all the CPA mixtures used can result in blastocysts using a short vitrification protocol. Notwithstanding, the mixture of PG–EG allowed for the highest cleavage (65.5%) and blastocyst rates (9.4%) among the permeating CPAs, with a cleavage rate similar to the control (77%). In addition, the oocytes vitrified with PG–EG with 0.5 mol/L galactose and warmed in a base medium with 0.3 mol/L galactose resulted in the highest blastocyst rate (15.1%) after vitrification, representing the only group of which the blastocyst rate was not significantly lower than the control (34.2%). Moreover, galactose, a monosaccharide used with success in equine [24] and camelid [39] embryo vitrification, was tested for the first time in equine oocyte vitrification and resulted in the highest blastocyst rates among the non-permeating CPAs (7.6%), with a cleavage rate equal to the fresh oocytes.Only a few studies have compared monosaccharides and disaccharides for oocyte or embryo vitrification, with contrasting results [34,50,51]. Kuleshova et al., 1999 [51] determined that disaccharides appear to have a greater influence on the vitrification properties of EG–saline solutions, and Huang et al., 2008 [50] showed a higher maturation rate of porcine oocytes vitrified with sucrose than those vitrified with glucose. However, McWilliam et al., 1995 [34] demonstrated a numerically greater survival rate of murine zygotes after exposure to monosaccharides than disaccharides. In studies with human [31] and pig oocytes [52], sucrose and trehalose resulted in similar maturation, cleavage, and embryo development rates. However, in our study with equine oocytes, the cleavage rate obtained with sucrose was numerically lower than that obtained with other saccharides, and was significantly lower than that of the control, while galactose gave a cleavage rate almost identical to the control. It seems that sucrose might present some degree of toxicity, as was suggested for camelid embryos [39]. In fact, in our study, both disaccharides sucrose and trehalose gave numerically lower cleavage and blastocyst rates than the monosaccharide galactose. While these differences were not statistically significant, the contrast in the cleavage rate between galactose and sucrose was very close to being significant. One of the differences of galactose versus the disaccharides is that the disaccharides displace more water than galactose due to their larger partial molar volume. The lower water ‘concentration’ (cw) results in an approximately 6% higher molal concentration of the CPAs and other solutes, and a correspondingly higher osmolality of the ES, VS, and WM when using 0.5 mol/L disaccharides compared with galactose. Further research is needed to confirm the apparent advantage of galactose and find potential mechanisms.In humans, in which the vitrification of in vivo matured oocytes is routinely performed in clinical practice, protocols make use of a two-step increase in the CPA concentrations [53,54]. In the first step, the oocytes are incubated for a relatively long time (10–15 min) in equilibration solution, and the second step consists of exposure to the vitrification solution for 30–90 s [55,56,57]. However, for horse oocytes, despite good maturation and cleavage rates and some blastocyst development, the relatively long exposure to CPAs has been associated with CPA toxicity, affecting subsequent embryo development [11,13,58,59]. Tharasanit et al, 2006 [28] demonstrated that a short protocol did not exhibit toxicity in horse oocytes. However, this short exposure may not have provided sufficient cryoprotection since the blastocyst rate of the vitrified–warmed oocytes was low (<1%) with the lower cooling and warming rates (open-pulled straw method) that they applied. Later, studies showed that shorter CPA exposure times with minimal volume (<1 µL) could yield better blastocyst rates [12,14], and such protocols produced the only foal born to date from vitrified–warmed immature oocytes [10]. Therefore, in this study, a short vitrification protocol (less than 90 s) was selected to be optimized.Although no significant differences among the CPAs were found in the present study, the PE mixture presented a higher cleavage rate than the other CPA combinations, with PE-0.3 also giving rise to the numerically highest blastocyst rate (15%), representing the only group not significantly different from the control. Yet, the blastocyst rate of 34.2% in the control group remains numerically higher, so further improvement of the vitrification protocol is indicated. This may be based on the use of PG–EG, which outperformed the other CPA mixtures in our study. Similarly, Canesin et al., 2017 [11] only obtained a blastocyst with the combination of PG–EG–trehalose, while this was not achieved with EG–DMSO–sucrose [11]. Furthermore, Somfai et al., 2013 [22] demonstrated in porcine oocytes that the PG–EG combination also provided greater embryo development after the vitrification of germinal vesicle-stage oocytes than the sole use of either CPA. Ethylene glycol alone or in a mixture is the most used CPA since it was demonstrated that EG is one of the safest CPAs regarding toxicity [18,60]. However, the membrane permeability of EG was found in horse oocytes to be substantially lower than that of PG [18,61]. Therefore, in short protocols, the intracellular CPA concentration will be lower when using only EG and may be insufficient to ensure cryoprotection. Conversely, PG exhibits a higher toxicity, but it presents one of the highest cell membrane permeabilities among CPAs [61]. As such, both CPAs provide complementary properties, and the combination of PG and EG can improve the cryopreservation outcome. This was supported by Somfai et al., 2015 [53], who showed that the combination of PG–EG was superior to EG–DMSO in terms of the oocyte survival after vitrification and the quality of the resulting blastocysts [52]. On the other hand, Clerico et al., 2021 [14] reported a 15% blastocyst rate, representing one of the best results with equine immature vitrified oocytes using EG–DMSO–trehalose, but these results were obtained after supplementation with the antioxidant melatonin during in vitro maturation, which improved the development after the ICSI. Using the combination of EG–DMSO—trehalose without further supplementation, they reached a blastocyst rate of 9%, which is in alignment with our blastocyst rate of 8% for the oocytes vitrified in EG-DMSO and warmed in 0.3 mol/L. Our results revealed that PG–EG provides a valuable option for equine immature oocyte vitrification since 15% was obtained when the oocytes were warmed in 0.3 mol/L, and the study of Clerico et al. [14] suggests that an even higher blastocyst yield could be obtained with the supplementation of substances that reduce stress post-vitrification/warming. However, future studies need to be performed to confirm these findings and to reveal the underlying mechanisms.Routinely, warming is performed in a hyperosmotic solution of sugars, decreasing to an isotonic base solution in two to four steps to avoid osmotic shock or over-swelling [56,62]. However, simpler warming systems have been evaluated. Inaba et al. [63] found that warming in an isotonic solution (holding medium) was equally effective to warming in a standard hypertonic solution for in vitro bovine embryos. Later, Canesin et al., 2018 [12] warmed equine oocytes in an isotonic base solution and found similar maturation and cleavage rates compared with a base solution containing 0.4 mol/L of trehalose, but the blastocyst rate was affected in all the treatments. We evaluated the effect of lowering the galactose concentrations of the warming medium (0.3 vs. 0.5 mol/L), attempting to simplify the warming method but keeping one step to avoid osmotic shock. The galactose concentration in the WM did not affect the maturation, cleavage, and blastocyst rates, but the oocytes vitrified in the PG–EG and warmed in 0.3 mol/L resulted in the highest blastocyst rate among all the vitrification treatments in the present study (15.1%). This may indicate that a less hypertonic (or possibly even isotonic) warming medium might be beneficial for the warming of equine oocytes. The short incubation of the oocytes in the ES and in VS, and the presence of 0.5 mol/L galactose in VS limit the entry of the permeating CPAs [64], which lowers the risk of the oocytes swelling in a warming medium above their isotonic volume and reduces the need for a high sugar concentration in the WM to counter the swelling. In addition, the osmolality of the WM with 0.5 mol/L in the WM will cause the oocytes to be still strongly shrunken after equilibration in the WM, which would prolong the shrunken state of the oocytes and could possibly make the step from the WM to BM and the subsequent reswelling to isotonic volume too abrupt. The at least equal, and possibly even better, result in our study with 0.3 mol/L galactose in the WM could thus be interpreted and appear to be in line with the study of Canesin et al. [12], in which the only blastocyst resulted from the warming in the medium without sugar. More research is needed to evaluate the different warming systems for equine oocytes with gradually decreasing concentrations in order to further optimize equine oocyte vitrification.5. ConclusionsIn conclusion, while we were able to produce blastocysts after the vitrification of equine immature oocytes with all six CPAs, the overall developmental competence remained lower compared to the fresh control. The PE mixture presented the highest cleavage rate compared with the other CPA combinations, while the PE-0.3 also gave rise to a competitive blastocyst rate of 15%, representing the only treatment not significantly lower than the fresh control. Moreover, galactose, a monosaccharide tested for the first time in equine oocyte vitrification, resulted in the highest blastocyst rates after vitrification as well as in cleavage rates equal to those of the control. As such, PE–galactose provides a valuable option for equine immature oocyte vitrification and should be considered as an alternative for the future optimization of the vitrification protocols for equine immature oocytes.

animals : an open access journal from mdpi

10.3390/ani11072103

PMC8300126

In this work, we aimed to discover unknown genes that are important in the regulation of other genes. These genes often play an important role during the development of the embryo. By screening thousands of mice, we found a gene, namely, Nuclear Respiratory Factor 1 (Nrf1), that controls the switching on and off of other genes. Mice with a defective Nrf1 present lesser levels of the gene and embryonic delay. When the mutation is in both chains of the DNA, mice are not born and die in the uterus. Our work unveils a novel, previously unknown functionality of Nrf1 and provides a new mice model for the study of diseases caused by a defective Nrf1.

We have established a genome-wide N-ethyl-N-nitrosourea (ENU) mutagenesis screen to identify novel genes playing a role in epigenetic regulation in mammals. We hypothesize that the ENU mutagenesis screen will lead to the discovery of unknown genes responsible of the maintenance of the epigenetic state as the genes found are modifiers of variegation of the transgene green fluorescent protein (GFP) expression in erythrocytes, which are named MommeD. Here we report the generation of a novel mutant mouse line, MommeD46, that carries a new missense mutation producing an amino acid transversion (L71P) in the dimerization domain of Nuclear Respiratory Factor 1 (Nrf1). The molecular characterization of the mutation reveals a decrease in the Nrf1 mRNA levels and a novel role of Nrf1 in the maintenance of the DNA hypomethylation in vivo. The heritability of the mutation is consistent with paternal imprinting and haploinsufficiency. Homozygous mutants display embryonic lethality at 14.5 days post-coitum and developmental delay. This work adds a new epi-regulatory role to Nrf1 and uncovers unknown phenotypical defects of the Nrf1 hypomorph. The generated mouse line represents a valuable resource for studying NRF1-related diseases.

1. IntroductionMutagenesis screens for the identification of modifiers of position-effect variegation (PEV) in Drosophila melanogaster have been pivotal for the development of the field of epigenetics [1]. These screens, performed in X-ray irradiated flies, have identified PEV alleles able to modify the expression of various genes presenting variegated expression, including the white (w) locus which leads to an easily-identifiable mosaicism of white and red patches in the eyes [1]. The identified alleles have been classified as dominant enhancers or suppressors of variegation, namely E(var)s or Su(var)s [2,3], depending on whether they enhance or suppress heterochromatin-mediated gene silencing. Similarly, mutagenesis screens have been performed in mammals for the identification of PEV modifiers. For this purpose, our group has established an N-ethyl-N-nitrosourea (ENU) forward dominant mutagenesis screen in transgenic mice (Line3) expressing green fluorescent protein (GFP) in erythrocytes in a variegated manner [4,5]. The transgene is composed by the human α1-globin gene promoter and the human α-globin locus enhancer region (HS-40) which is linked to the human GFP (reporter gene) [5]. Such a construct is expressed in approximately 55% of the erythrocytes of Line3 mice [6]. In this screen, enhancers and suppressors of variegation are easily identified by flow cytometry with a drop of blood [7,8], and are denominated as MommeDs (modifiers of murine metastable epialleles, Dominant). In our laboratory, this screen has successfully identified more than 40 unknown and known genes playing a role in fundamental epigenetic processes in mammals [4,6,7,8,9,10,11,12,13,14]. Nearly all these genes are also required for correct embryonic development [6,7,8,10,13] and have been linked to a broad range of syndromes and diseases. For instance, reduced levels of Structural Maintenance of Chromosomes Flexible Hinge Domain Containing 1 (SMCHD1) have been associated to the human facioscapulohumeral dystrophy type 2 and cancer [15,16], a loss-of-function mutation in Wiz has been associated with anxiety-like behavior in mice [14] and loss of Rearranged L-Myc fusion (RLF) has been shown to cause heart defects in mice [17].Herein, we report the identification and characterization of a novel mutation in the Nuclear Respiratory Factor 1 gene (Nrf1). The mice line is named MommeD46. NRF1 is a transcription factor initially found to bind the cytochrome c promoter and activate transcription [18]. NRF1 regulates respiration, mitochondrial biogenesis and mitochondrial DNA transcription and replication [19,20,21]. Consistent with being identified in our mutagenesis screen, Nrf1 plays a role in the maintenance of DNA hypomethylation, because the under-functioning Nrf1 induces hypermethylation. Moreover, the novel Nrf1 mutation is homozygous lethal and presents a characteristic hereditability consistent with paternal imprinting and haploinsufficiency. This work adds a new epi-regulatory role to Nrf1 and discovers unknown embryonic effects of a novel Nrf1 hypomorph. The generated mouse line could be used for modeling and studying diseases caused by an underfunctioning NRF1.2. Materials and Methods2.1. AnimalsAll animal experiments were performed in accordance with protocols approved by the Animal Ethics Committee of QIMR Berghofer Medical Research Institute. All the animal experiments comply with the ARRIVE guidelines. Pups were weaned at three-weeks-old and bred when males and females were eight and six-weeks-old, respectively. The ENU screen was performed in FVB/NJ inbred transgenic mice, namely Line3, homozygous for the green fluorescence protein (GFP) transgene under the control of the α-globin promoter, as previously described [4]. MommeD46 was maintained in this mice line and crossed for at least five generations with wild-type FVB/NJ mice, down from the mutant founder. Line3C mice were used for the linkage analysis (Figure 1a). Line3C was obtained by crossing Line3 with mice of strand C57BL/6J for ten generations and selecting for the offspring carrying the GFP transgene by flow cytometry. C57BL/6J mice were purchased from the Animal Resources Centre (Perth, WA, Australia). Sperm of MommeD46 is cryopreserved at the Australian Phenome Bank at the Australian National University (2601 ACT, Australia).2.2. Embryo DissectionsHeterozygous MommeD46 mice were crossed and females were daily monitored for vaginal plugs. A vaginal plug was counted as 0.5 days post-coitum (dpc). Pregnant females were sacrificed by CO2 inhalation, dissected and embryos isolated and visualized under a binocular dissecting microscope.Determination of GFP Expression in ErythrocytesA tiny cut in the tail tip was performed in manually strained (three-weeks-old) mice to extract a drop of blood of three-weeks-old mice to be then dissolved in Osmosol buffer and analyzed using the benchtop flow cytometer Guava easyCyte HT (Merck/Millipore, Darmstadt, Germany). A gating was set to exclude 99.9% of erythrocytes not expressing the GFP.2.3. Linkage Analysis by SNP ChipDNA from tail tips from 15 mice phenotypically expressing GFP in 55% of the erythrocytes (wild-type) and 15 expressing lower GFP percentages (MommeD46) from the backcross between heterozygous MommeD46 with Line3 background (FVB/NJ strain) and Line3C (C57BL/6J strain) mice were included in the linkage analysis (Figure 1a). Mice were at least three-weeks-old. The Mouse Medium Density Linkage Panel from Illumina (San Diego, CA, USA) was used for single nucleotide polymorphism (SNP) genotyping. The panel contains 766 measurable differential SNPs between C57BL/6J and FVB/NJ strains. Wild-type mice for the mutation should only have C57BL/6J-specific SNPs in the linked region while heterozygous MommeD46 mice can have both C57BL/6J and FVB/NJ-specific SNPs. SNP Chip was performed according to Illumina instructions and SNP calling was carried out using the Genotyping module of the GenomeStudio v1.1 software (Illumina). Only call rates higher than 95 were accepted. Logarithm of odds (LOD) scores across the entire mouse genome were calculated and linked intervals identified (LOD scores ˃ 2, Figure 1a and Figure 2). Further fine mapping using additional SNP markers (Figure S1) was performed to reduce the linked interval.2.4. Whole Exome SequencingWhole exome sequencing for the identification of the underlying mutation in MommeD46 was carried out as previously described [7].2.5. Methylation Analysis in the Transgene HS-40 EnhancerDNA samples from three wild-type, three heterozygous and three homozygous MommeD46 9.5 dpc embryos were converted with bisulfite using the EpiTect Bisulfite Kit (Qiagen, Doncaster, VIC, Australia) in accordance with the manufacturer’s instructions. Bisulfite-converted DNA was amplified (first PCR) using specific oligonucleotides (Integrated DNA Technologies, IA, USA) for the HS-40 enhancer region, as follows: 5′-AAAATAAAATTTTTGGATTGTTATTATTATAA-3′ (Forward 1), 5′-ATATTTGTAATTTTAGTATTTTGGGAGGTT-3′ (Forward 2) and 5′-AATCTCTACTCACTACAAACTCCATCTC-3′ (Reverse) with the following cycling conditions: one cycle at 94 °C for 2 min; 35 cycles at 94 °C for 30 s, 60 °C for 30 s, 72 °C for 45 s; one cycle at 72 °C for 6 min. Next, a semi-nested PCR was performed with the first PCR product using the oligonucleotides Forward 2 and Reverse. The product was then ligated into a pGEM®-T Vector (Promega, Madison, WI, USA) as per Promega instructions. DH5α competent E. coli were transformed using the heat shock method and plated in LB agar plates containing ampicillin, x-gal and IPTG. Positive colonies (white) were picked and DNA was amplified to determine transgene insertion using the following oligonucleotides: 5′-ATTTAGGTGACACTATAG-3′ (Forward, SP6) and 5′-TAATACGACTCACTATAGGG-3′ (Reverse, T7) and PCR conditions according to the MangoTaq DNA polymerase protocol (Bioline Reagents Limited, London, UK). Plasmid DNA from positive colonies was then extracted with Qiagen Miniprep Kit (Germantown, MD, USA) and sequenced. The BIQ Analyser software was used to determine the quality of the conversion and for the analysis. Sequences with <98% of conversion rate and <80% of sequence matching as well as identical clones were excluded from the analysis. White and dark circles-containing diagrams, obtained from introducing the BIQ results’matrix to the software freshteapot, were used to represent the unmethylated and methylated CpG sites in the HS-40 enhancer, respectively.2.6. GenotypingGenotyping of MommeD46 was performed by Sanger sequencing or by restriction enzyme digestion followed by gel electrophoresis. For both genotyping methods, DNA from tail tips (at least three-weeks-old) and embryos was extracted using a saturated salt solution containing 50 mM Tris pH 6.8, 20 mM EDTA pH 8, 2% SDS and 6.25 µg/mL proteinase K. DNA was precipitated with isopropanol and washed with 75% and 100% ethanol. DNA pellets were air dried and treated with 200 µL of RnaseA (10 mg/mL). Sanger sequencing was performed for 9.5 dpc embryos genotyping (Figure 3b). Wild-type, heterozygous and homozygous MommeD46 embryos (Figure 3b) were homogenized. DNA was isolated and a genomic region of 321 base pairs (bp) covering the causative mutation in Nrf1 was amplified by PCR using the oligonucleotides 5′-GGCTCAGCCAGTGTTTTCTTA-3′ (forward) and 5′-TTTGATCCCCAGCAGTGAA-3′ (reverse). The PCR product was then sequenced using the BigDye Terminator Sequencing Kit (Applied Biosystems, Foster City, CA, USA) and the previous forward oligonucleotide. The restriction enzyme digestion was performed for both adult and embryo genotyping. To do so, the same obtained above PCR product was subjected to digestion with HpaII (New England BioLabs, Ipswich, MA, USA). In wild-type MommeD46 the enzyme did not cut, in heterozygous MommeD46 the enzyme only cut one DNA strand, producing fragments of 321 bp, 173 bp and 148 bp, and in homozygous MommeD46 the enzyme cut both DNA strands, producing fragments of 173 bp and 148 bp.2.7. Determination of Nrf1 mRNA Levels by Quantitative Real-Time PCRThe determination of Nrf1 mRNA levels was performed in three wild-type, three heterozygous and three homozygous MommeD46 of at least three weeks of age, following a three-step process which included RNA isolation, cDNA conversion and quantitative real-time (qRT)-PCR. For RNA isolation, 9.5 dpc embryos were mechanically homogenized and treated with Trizol® (Qiagen, Doncaster, VIC, Australia) and chloroform. RNA was precipitated with isopropanol and high salt solution (0.8 M NaCitrate and 1.2 M NaCl). RNA was washed with 75% ethanol, air dried and dissolved in DEPC water. RNA was converted into cDNA using SuperScriptIII reverse transcriptase and random hexamer primers (Invitrogen, Carlsbad, CA, USA). For the qRT-PCR, the forward oligonucleotide for Nrf1 was designed at the boundary of the first exon and the second exon (5′-TGAGGTCGAATGGTATGTGGT-3′), and the reverse oligonucleotide 78 bp downstream (5′-ACACCCATGTTCATAGCAGCA-3′), leaving a product of 121 bp. The DNA amplification was performed using the Platinum SYBR Green qPCR Super Mix—UDG (Invitrogen, Carlsbad, CA, USA). The PCR was run in the Viia7 (Applied Biosystems, Mulgrave, VIC, Australia) and analyzed using the Viia7 software. Hprt was used as housekeeping gene, using the qRT-PCR oligonucleotides 5′-GGCCAGACTTTGTTGGATTT-3′ (forward) and 5′-ACTGGCAACATCAACAGGACT-3′ (reverse).2.8. Determination of Nrf1 Protein Levels by Western BlotWhole-protein extracts were prepared from three wild-type, three heterozygous and three homozygous MommeD46 9.5 dpc embryos. Embryo tissue was mechanically homogenized and lysed with lysis buffer containing urea 8 M, 1% sodium dodecyl sulfate, 10% glycerol, 0.5 mM dithiothreitol and 10 mM Tris pH 6.8, complemented with the serine protease inhibitor phenylmethylsulfonyl fluoride and half tablet of the protease inhibitor cocktail cOmplete® from Roche (Swedesboro, NJ, USA) for 5 mL of lysis buffer. Next, samples were sonicated and quantified using the Pierce BCA Protein Assay Kit (ThermoFisher Scientific Australia Pty Ltd., Scoresby, VIC, Australia). Equal amounts of protein (30 µg) were loaded in a Mini-Protean precast 7.5% polyacrylamide gel from Bio-Rad (Hercules, CA, USA) and run at 100 V. Proteins were transferred into PVDF membranes following the wet transfer protocol, overnight at 4 °C at 20 V. Membranes were blocked with 5% of skim milk and 1% BSA in PBS-T. The anti-Nrf1 antibody from Novus Biologicals (Littleton, CO, USA) was incubated in the same solution overnight at 4 °C, 1:1000 dilution. A goat anti-rabbit secondary antibody conjugated to horseradish peroxidase (HRP) from DAKO, 1:5000 dilution in PBS-T, (Santa Clara, CA, USA) was added for one hour at room temperature. Clarity Western ECL substrate (Bio-Rad) was added in the membrane and the chemiluminescence signal was captured with the MF Chemi-Bis (DNR Bio-imaging systems). The γ-Tubulin antibody from Sigma-Aldrich (St Louis, MO, USA), 1:1000 dilution, was used as loading control.2.9. Statistical AnalysisAll the experiments were done in at least three biological replicates. Normality of data was assessed by Saphiro–Wilk or Kolmogorov–Smirnov tests, depending on the sample size. When the data were satisfied with normality assumption, statistical significance between mean differences of the groups was assessed by two-tailed Student’s t-test or one-way ANOVA and p < 0.05 was considered significant. Otherwise, the non-parametric Mann–Whitney U test was applied and p < 0.05 was considered significant. Chi–Square test (χ2) or Fisher’s exact test (F) were used to compare the observed proportion of genotypes with the expected Mendelian ratios when the number of events in a cell was ≥5 or <5, respectively, and a p value < 0.05 was deemed as significant. Levene’s test or F-test of equality of variances was used to determine variance differences between wild-type and heterozygous distributions, holding significance when p < 0.05. In this figure, two-way ANOVA was also employed. All analyses were performed using the software IBM SPSS Statistics version 24.0 (Chicago, IL, USA).3. Results3.1. MommeD46 Is an Enhancer of VariegationThe causative mutation in MommeD46 induces a change in the GFP expression in erythrocytes. Wild-type MommeD46 mice express the GFP in approximately 55% of red blood cells (Figure 1b), as previously shown for Line3 mice [7]. Heterozygous MommeD46 mutants present 44 ± 5% of GFP-expressing erythrocytes, which is significantly less (p = 2.15 × 10−28) than the percentage observed in their wild-type counterparts (53 ± 4%). Figure 1c shows the flow cytometry GFP profiles of wild-type and heterozygous mutant MommeD46. Mutant mice show a lesser count. Therefore, MommeD46 is an enhancer of variegation.3.2. MommeD46 Carries a Point Mutation in the Nrf1 GeneTo elucidate the underlying mutation in MommeD46, we followed different strategies. First, we performed SNP Chip to determine the linked genomic interval where the mutation was located. We identified two linked regions with LOD score of 2 or higher in Chromosome (Chr) 6 and Chr 15 (Figure 2a). As the highest LOD score was in Chr 6, we narrowed down the linked interval in that Chr, comprising a region of 123.3 Mb, from the genomic location 4.5 Mb to 127.8 Mb (Figure S1). We exploited nine differential SNP markers between FVB/NJ and C57BL/6J at 4.5, 21.9, 34.7, 38.1, 41.4, 48.7, 75.5, 107.8, and 127.8 Mb. We found a number of supporting genotypically heterozygous mice, and thus, with C57BL/6J-specific SNPs, concentrated at 21.9 and 34.7 Mb. At the same locations, we also found genotypically wild-type mice (with FVB/NJ-specific SNPs) (Figure 2b). This observation indicated that the underlying mutation in MommeD46 was located somewhere between 21.9 and 34.7 Mb. Later, we performed exome deep sequencing and we identified the causative mutation to be a missense mutation at the end of the third exon of Nrf1 (Chr 6:30,040,135; USC Genome Browser) (Figure S2). This was confirmed by Sanger sequencing in 9.5 dpc embryos (Figure 3b). The mutation is a T ˃ C substitution that induces a Leucine (L) ˃ Proline (P) change at position 71 of the amino acid sequence of Nrf1, located before the predicted Nrf1 DNA binding domain. The affected amino acid is conserved between species (Figure 3a). The point mutation in Nrf1 causes a decrease in the Nrf1 mRNA levels as seen by qRT-PCR. In particular, Nrf1MommeD46/+ and Nrf1MommeD46/MommeD46 9.5 dpc embryos showed a 15% (p = 0.014) and a 44% reduction (p = 0.0004) in Nrf1 mRNA levels, respectively, compared to Nrf1+/+ (Figure 3c). In contrast, the mutation did not induce a noticeable change in the Nrf1 protein levels as observed by Western blot (Figure 3d). In addition, the band quantification of the Western Blot did not show significant differences of band intensity between the groups when normalized to γ-Tubulin (Figure S3).3.3. Nrf1 Mutants Display Enhanced DNA MethylationTo further prove a role of Nrf1 as epigenetic regulator in mammals, we performed bisulfite sequencing of the HS-40 enhancer region of the GFP transgene in wild-type, heterozygous and homozygous MommeD46 9.5 dpc embryos. We observed that Nrf1+/+ embryos presented 57% of the CpG islands methylated. This percentage increased up to 66% in Nrf1MommeD46/+ embryos, and up to 71% in Nrf1MommeD46/MommeD46 embryos (p = 0.049) (Figure 4). This result suggests that Nrf1 is involved in DNA methylation and a functional Nrf1 is essential to maintain a hypomethylated state, which is consistent with the fact that Nrf1 is an enhancer of variegation.3.4. The Observed Hereditability of the Nrf1 Suggests Paternal Imprinting and Haploinsufficiency, and Normal Nrf1 Is Required for Normal Embryonic DevelopmentHerein, we wanted to determine whether the newly discovered mutation in Nrf1 was responsible for any phenotypic change or defects in mice. To do so, we performed different intercrosses between adult Nrf1MommeD46/+ mice. We observed proportions of Nrf1+/+, Nrf1MommeD46/+ and Nrf1MommeD46/MommeD46 to be significantly different than that expected by Mendel at weaning (p < 0.001) (Figure 5a). Interestingly, neither homozygous mice were observed at weaning nor at the embryonic stage of 14.5 dpc indicating an earlier embryonic lethality of the mutation in homozygosis. In the earlier embryonic stages of 8.5, 9.5 and 10.5 dpc, homozygous embryos were observed at decreased ratios than expected (not significant). When further analyzing the homozygous embryos, we detected the presence of empty decidua in some of them, all presumably being Nrf1MommeD46/MommeD46, while the rest resembled embryos (Figure 5b). However, some of them looked abnormal. Male and female heterozygous mice showed similar fertility, producing an average of 7.9 ± 2.1 pups per litter when the heterozygous progenitor was a male, and 7.7 ± 2.5 pups when the heterozygous progenitor was a female. As expected, the heterozygous intercross produced less pups per litter, specifically 6.8 ± 1.3 (Figure 5c). To determine any possible progenitor effect in the hereditability of the mutation, we genotyped the offspring of two types of intercrosses: a heterozygous mutant male crossed with a wild-type female, and a wild-type male crossed with a heterozygous mutant female. The expected Mendelian proportions for both intercrosses were 50% wild-type and 50% heterozygous mutants. However, we observed different proportions in both intercrosses, with the ratios significantly different than expected (p < 0.001) for the intercross including a female wild-type. Moreover, the percentage of wild-type pups obtained from the intercross containing the female wild-type was significantly greater (75%) than the percentage obtained from the intercross having the male wild-type (62%) (p < 0.020) (Figure 5d). Altogether, these results suggest an accentuated maternal hereditability and some degree of haploinsufficiency of the described mutation in Nrf1 which is further discussed in the next section. Finally, to ascertain any possible embryonic or adult defects caused by the mutation in Nrf1, we determined the body weights of Nrf1+/+ and Nrf1MommeD46/+ mice at weaning and explored the morphology of Nrf1+/+, Nrf1MommeD46/+ and Nrf1MommeD46/MommeD46 embryos at different embryonic stages. As seen in Figure 5e, heterozygous mutant mice weighed less than their wild-type counterparts in both sexes (not significant), with significantly different variances found in males (Levene’s F-test, p < 0.05). When examining the embryos at 8.5 and 9.5 dpc, we observed that homozygous mutant embryos were clearly smaller and less developed than the wild-type or heterozygous mutants (Figure 5f). However, no differences were detected between wild-type and heterozygous mutant embryos. This observation highlights the importance of Nrf1 for correct embryonic development.4. DiscussionBy deploying a forward dominant ENU mutagenesis screen in mice, we have discovered a novel missense mutation in Nrf1 capable of altering the epigenetic state. Additional mutations were found in the link obtained by SNP-Chip but outside the fine mapped interval (Figure S4), which reassured us that the mutation in Nrf1 was causative of the phenotype. Nrf1 is an enhancer of variegation as seen by the decrease in the percentage of GFP positive erythrocytes in heterozygous mutant mice. The exonic mutation in Nrf1 is responsible for the amino acid (aa) change L ˃ P at the position 71 of the Nrf1 protein (total length: 503 aas) and is located in the aa region responsible for Nrf1 dimerization. This region ranges between 1 and 78 aas as the previously generated 1-to-78-deletion mutant has been shown to be entirely monomeric [22]. This same mutant also binds DNA less robustly compared to the wild-type protein [22], consistent with a possible overlap with the Nrf1 DNA binding domain, predicted to be between 79 and 172 in the chicken Nrf1 [23]. This suggests that the 71L ˃ P substitution of MommeD46 could perturb both the homodimerization and DNA binding ability of Nrf1. Nrf1MommeD46/+ and Nrf1MommeD46/MommeD46 embryos show substantial reduction in the Nrf1 mRNA levels indicating that the mutation could affect mRNA synthesis and processing. In contrast, Nrf1 protein levels do not seem to be altered in the mutant embryos. This could be due to increases in the half-life of the mutated Nrf1 protein or an increase in the Nrf1MommeD46 mRNA translation, processes that have been reviewed in Liu et al. [24]. As for studies with Nrf1, Dhar et al. found a drastic reduction in Nrf1 mRNA levels by qRT-PCR accompanied with a not so accentuated decrease in protein levels by Western blot in murine neuroblastoma cells [25].We are the first to demonstrate a role of Nrf1 in maintaining DNA hypomethylation in vivo as seen by the significant hypermethylation of the α-globin HS-40 enhancer region observed in Nrf1MommeD46/+ and Nrf1MommeD46/MommeD46 embryos in comparison with wild-type embryos. Interestingly, only this present MommeD46 and the previously published MommeD28 (Rlf null mutant) [7] show hypermethylation in the HS-40 enhancer. This is consistent with the decreased expression of the GFP transgene. Although the preference of NRF1 to bind un-methylated regions [26,27] and methylated sequences [28] has been studied in depth, little is known about the role of Nrf1 in directly controlling DNA methylation. In this line, one study has found that Nrf1 transcriptionally activates Euchromatic histone methyltransferase 1 (EHMT1) during in vitro meiosis, and that its DNA binding ability is negatively regulated by CDK2-mediated phosphorylation [29]. However, the authors did not assess DNA methylation changes. In another work, the authors did not find changes in DNA methylation in the gene promoters of the six Nrf1 target genes analyzed in CD9+ spermatogonia from Nrf1 germ cell-conditional knockout (Nrf1f/f) mice [27]. In contrast, in our study, Nrf1 was isolated from a genome-wide genetic screen in search of epigenetic modifiers and we found methylation changes in the HS-40 enhancer in mutants with dysfunctional Nrf1. Besides all this, it could be that NRF1 indirectly maintains DNA hypomethylation. The mutation L71P in Nrf1 could reduce the ability of Nrf1 to bind the DNA, such that Nrf1 is then unable to protect the CpG sites from methylation, leading to a methylation gain as previously described [26].According to our embryonic and adult examinations of MommeD46, Nrf1MommeD46/MommeD46 mice are not viable at birth and die before 14.5 dpc. Also, the offspring of the heterozygous intercross present significantly altered Mendelian ratios at weaning. Similarly, the Nrf1 full knockout has shown to be embryonic lethal between 3.5 and 6.5 dpc [30]. Regarding mice fertility, our mutation in Nrf1 does not seem to affect fertility as seen by the similar average litter sizes obtained from the cross Nrf1+/+ × Nrf1MommeD46/+ for both progenitor genders. However, in the literature, conditional ablation of Nrf1 in gonocytes has been shown to lead to infertility in male mice [27]. This contrasting evidence could be due to the subtle nonobvious effect that our mutation has on fertility. We observed that the wild-type Nrf1 allele is significantly more present than expected in the offspring when the progenitor is a wild-type female rather than when it is a wild-type male (75% versus 62%). This could be explained by paternal imprinting of the mutated allele in the mating ♀Nrf1+/+ × ♂Nrf1MommeD46/+, by maternal imprinting of the mutated allele in the mating ♀Nrf1MommeD46/+ × ♂Nrf1+/+, by haploinsufficiency or by the existence of an additional ENU mutation having a cis- or trans-acting effect on the wild-type or the mutated allele. Maternal imprinting or inhibitory cis-/trans-acting effects on the mutated allele does not occur because the proportion of Nrf1MommeD46/+ progeny from the mating ♀Nrf1MommeD46/+ × ♂Nrf1+/+ is not significantly different from the expected Mendelian ratios, 38% versus 50%. In contrast, the possibility of activating cis- or trans-activating effects on the wildtype allele due to another ENU mutation cannot be ruled out, but, at the same time, is very difficult to prove. Regarding Nrf1 inheritance, the literature shows that there is no maternal inheritance of Nrf1 gene products as wild-type embryo eggs (Nrf1+) from the Nrf1null/+ heterozygous intercross were seen at the stage before the first meiotic division [30]. Regarding paternal imprinting, NRF1 has previously been shown to positively regulate the small nuclear ribonucleoprotein polypeptide N (SNRPN) locus which is associated with the imprinting center that regulates the Prader–Willi syndrome domain (PWS-IC), suggesting a relevant contribution of NRF1 in PWS-IC activity [31,32]. In other words, NRF1 orchestrates the characteristic loss of expression of the paternally-inherited allele of an imprinted gene cluster, that includes SNRPN, in the PWS. This suggests that the mutation in MommeD46, which affects the dimerization and DNA binding capabilities of Nrf1, could generate imprinting defects and be responsible for the increase in the expression of the maternal wild-type allele observed. Another possibility that could explain the greater presence of the wild-type allele in the offspring when the progenitor is a wild-type female is the haploinsufficiency caused by the mutated Nrf1 allele in the heterozygous mice. The early lethality of the homozygous mutants supports this notion.Regarding the effects of the mutation in embryos and adult phenotypes, we observed a decrease in size and developmental delay in 9.5 and 10.5 Nrf1MommeD46/MommeD46 embryos compared to wild-type embryos. Moreover, both male and female Nrf1MommeD46/+ mice display a slight decrease in body weight compared to wild-type mice at weaning, with significantly different variances in males. Similarly, a previous study demonstrated defective development in vitro of all Nrf1-null and some Nrf1 heterozygous blastocysts as opposed to their wild-type counterparts [30]. In mice, Nrf1 is highly expressed in thymocytes, B-cells and retina, and in humans, in natural killer cells, T cells and skeletal muscle. Consistent with that, retinal progenitor cell and postmitotic rod photoreceptor cell-specific Nrf1 conditional knockout mice exhibit rod photoreceptor cell degeneration as a result of mitochondrial impairment [33]. In addition, Drosophila erect wing (ewg) (Nrf1 homologue) mutants show aberrant axonal projection and absent or reduced indirect flight muscles [34], and zebrafish nrf (Nrf1 homologue) mutants exhibit loss of retinal photoreceptors and reduced brain size [35]. Also in the neural context, reduced transcriptional activity of NRF1 has been associated with fragile X mental retardation syndrome [36] and with GABA-associated neuronal disorders [37], and functional NRF1 has been shown to be necessary for neurite outgrowth of human neuroblastoma cells [38,39]. Other non-neuronal roles attributed to NRF1 are the regulation of oncogenesis [40,41,42,43], stem cell aging [44], telomere transcription [45] and oxidative stress protection and lipid homeostasis in the liver; liver-specific Nrf1 conditional knockout mice develop steatohepatitis and spontaneous liver cancer [43]. In addition, low levels of (mRNA) Nrf1 have been observed in Cystathionine β synthase (CBS) heterozygous-null mice modeling hyperhomocysteinemia, a syndrome characterized by skeletal muscle weakness and fatigability [46]. However, it remains unknown whether Nrf1 is a causative factor or a consequence of the disease. Interestingly, CBS+/− mice exhibit significant weight loss [47], and homocysteine treatment, as a way to model the disease in vitro, increases the expression of DNA methyltransferase (DNMT)3a and DNMT3b and consequent global methylation levels in muscle C2C12 cells [46], which is in agreement with our findings of Nrf1 controlling the epigenetic state and the decrease in body weight observed in Nrf1MommeD46/+ mice. Alternatively, given that Nrf1 controls muscle energy production [48], it could be that the decrease in body weight in the heterozygous Nrf1 mutants is due to the reduction in muscle function and subsequent muscle mass. Further investigations will be undertaken to unravel other phenotypic traits or defects caused by the 71L ˃ P substitution in Nrf1.5. ConclusionsBy using a genome-wide mutagenesis screen, we have identified Nrf1 acting as a controller of the epigenetic landscape in mammals, through the maintenance of DNA hypomethylation in vivo. This novel MommeD46 causes a reduction in the Nrf1 mRNA levels and severe embryonic developmental defects. Additionally, we noted a peculiar hereditability of the mutation which is consistent with a paternal imprinting of the Nrf1MommeD46 allele and haploinsufficiency. Our study unveils an important new epi-role of NRF1 in addition to its extensive transcriptional regulation throughout the animal life. Furthermore, the variant mouse produced in our laboratory could be used to model or elucidate unknown syndromes driven by a reduced function of NRF1.

animals : an open access journal from mdpi

10.3390/ani13050808

PMC10000129

In this study, bioinformatics approaches were used to better understand the genetic architecture of internal organ weights in three-way crossbred commercial pigs and to map genetic markers and genes. For this purpose, we used single-trait and multi-trait genome-wide association studies (GWASs) followed by a haplotype block analysis. We explored the key genetic markers and genes from the internal organ weight genome-wide association study results of three-way crossbred commercial pigs. In this manner, five genes, TPK1, POU6F2, PBX3, UNC5C, and BMPR1B, were defined as central in affecting internal organ weight in pigs. Moreover, APK1, ANO6, and UNC5C were identified to be pleiotropic in multi-trait GWASs. These results can be applied to various types of genomic studies of pigs.

Internal organ weight is an essential indicator of growth status as it reflects the level of growth and development in pigs. However, the associated genetic architecture has not been well explored because phenotypes are difficult to obtain. Herein, we performed single-trait and multi-trait genome-wide association studies (GWASs) to map the genetic markers and genes associated with six internal organ weight traits (including heart weight, liver weight, spleen weight, lung weight, kidney weight, and stomach weight) in 1518 three-way crossbred commercial pigs. In summation, single-trait GWASs identified a total of 24 significant single- nucleotide polymorphisms (SNPs) and 5 promising candidate genes, namely, TPK1, POU6F2, PBX3, UNC5C, and BMPR1B, as being associated with the six internal organ weight traits analyzed. Multi-trait GWAS identified four SNPs with polymorphisms localized on the APK1, ANO6, and UNC5C genes and improved the statistical efficacy of single-trait GWASs. Furthermore, our study was the first to use GWASs to identify SNPs associated with stomach weight in pigs. In conclusion, our exploration of the genetic architecture of internal organ weights helps us better understand growth traits, and the key SNPs identified could play a potential role in animal breeding programs.

1. IntroductionBody weight, which can reflect growth performance and thus affect economic efficiency, has attracted a lot of attention in animal breeding programs. The body weight of cattle is the sum of various elements, including fat weight, internal organ weight, muscle weight, and bone weight, among others. Of these components, internal organ weight constitutes 14% of the total body weight of cattle [1]. The weight and size of an organ are salient features that serve as dependable predictors of its developmental progression, wherein an augmented organ mass typically alludes to a heightened degree of maturation. Accelerated organ development leads to a smoother coordination of internal organs during vital biological processes such as oxygen transport, blood circulation, lipid metabolism, and digestion. This refinement of these processes can positively impact growth and economic traits. Previous studies have shown that the internal organ weights of crossbred steer calves are strongly correlated with carcass growth rate [2]. Moreover, in humans, internal organ weights have been shown to be positively correlated with body weight and height in normal Zambian adults [3]. Thus, comprehending the genetic architecture of heart weight (Heart WT), liver weight (Liver WT), spleen weight (Spleen WT), lung weight (Lung WT), kidney weight (Kidney WT), and stomach weight (Stomach WT) will propel genetic progress and facilitate the successful implementation of breeding programs.Genome-wide association studies (GWASs) are widely used to identify quantitative trait loci (QTL) and candidate genes associated with complex traits in animals and plants. To date, the number of QTL associated with Heart WT, Liver WT, Spleen WT, Lung WT, and Kidney WT are 29, 31, 19, 5 and 8, respectively, and no QTL have been reported to be associated with Stomach WT in the pig QTL database [4] (accessed on 15 November 2022). Previous studies reported 39 QTL to be associated with internal organ weight in four local pig populations and one commercial population [5]. For instance, Zhang et al. [6] showed that a 2 cM QTL on Sus scrofa chromosome 2 (SSC2) was significantly associated with Heart WT, and three QTL were associated with Liver WT, Lung WT, and Spleen WT. Although several studies have identified QTL to be associated with internal organ weight [7,8], the process of genetic improvement remains slow.The difficulty (and high cost) of obtaining phenotypes for internal organ weight studies has led to fewer studies on its genetic architecture. Moreover, previous studies conducted single-trait GWASs for internal organ weight to map the genetic markers and genes; however, internal organ development is mutually coordinated by each different organ, and the single-nucleotide polymorphisms (SNPs) in the genome may act on multiple organs at the same time. Therefore, it is difficult to identify SNPs and candidate genes that affect multiple internal organs simultaneously using single-trait GWASs. Therefore, herein, we performed multi-trait GWASs to identify polymorphic SNPs and improve statistical efficiency, which mainly depends on the genetic correlation between traits [9,10]. In this manner, it was observed that the statistical efficiency was improved in the case of low trait correlations [11,12].Previous studies demonstrated the superiority of conducting multi-trait GWASs in terms of uncovering the genetic architecture of complex traits in animals. For instance, Zhou et al. [13] performed multi-trait GWASs to identify 21 pleiotropic SNPs that were not detected via single-trait GWASs in three body size traits. In Simmental beef cattle, An et al. [14] detected 29 pleiotropic SNPs that were functional in all three growth periods using multi-trait GWASs. To date, there are no studies that use multi-trait GWASs to analyze the genetic architecture of visceral weight. Herein, we performed multi-trait GWASs to compensate for the deficiencies associated with single-trait GWASs and to provide new insights into the genetic mechanisms of multi-organ co-development.The aim of this study was to map the genetic markers and candidate genes associated with internal organ weight in pigs. To this end, we conducted single-trait and multi-trait GWASs for six internal organ weight traits in 1518 crossbred commercial Duroc × (Landrace × Yorkshire) DLY pigs. The results from the current study advanced our understanding of the genetic basis for internal organ weight and further revealed the complexity of the genetic architecture of internal organ weight in pigs. Integrating SNP results from GWASs as a source of prior biological information in the improvement program enhances the selection process by assigning higher weight to key SNPs that are critical for improving internal organ weight traits.2. Materials and Methods2.1. Ethical StatementAll animals used in this study were treated in accordance with the guidelines for the use of laboratory animals of the Ministry of Agriculture of China and with the approval of South China Agricultural University (Guangzhou, China), No. 2018F089.2.2. Animal Samples and Phenotype CollectionExperimental animals were selected from a DLY three-way crossbred commercial line with no overlapping blood relations, through random selection based on genealogy, in which 89 Duroc boars were mated with 397 Landrace × Yorkshire sows to produce a large number of offspring. All pigs were raised in four farms of the Guangdong Wens Food Group Co., Ltd. (Guangzhou, China). In brief, a total of 1518 individuals (757 boars and 764 sows) were reared with free access to water and feed and were fattened to 115 kg. They were euthanized in 13 batches with a 24 h interval between each batch and had an average slaughter age of about 7 months. After the pigs were euthanized, their phenotypes were recorded, and their internal organs were excised, emptied, flushed, blotted dry, and weighed immediately using an electronic scale with a range of 0.0 kg to 300 kg and accuracy of ±100 g. The scale was calibrated using the linear calibration method with 20% MAX or 60% MAX weight. The organ distribution is shown in Figure 1. R 4.2.1 software was used to test the normal distribution of the descriptive statistics of the internal organ traits.2.3. Genotyping and Quality ControlEar samples were collected from all 1518 individuals, and genomic DNA was extracted from the ear tissue of each pig using a standard phenol–chloroform method and subsequently diluted to 50 ng/μL for the genotyping procedure, controlling the quality OD260/280 between 1.8 and 2.0. The 1518 DLY pigs were genotyped using the GeneSeek Porcine 50K SNP BeadChip (Neogen, Lincoln, NE, USA), which contained 50,703 SNPs. After genotyping, to ensure the accuracy and validity of the GWAS results, we performed a quality control (QC) procedure using the PLINK v1.07 software [15] with the following parameters: individual call rate > 95%; SNP call rate > 99%; minor allele frequency > 1%; and p > 10−6 for the Hardy–Weinberg equilibrium test. Moreover, SNPs in sex chromosomes and unmapped regions were excluded. After QC, a final set of 31,941 eligible SNPs remained for subsequent single-trait and multi-trait GWASs.2.4. Population Structure and Linkage Disequilibrium (LD) EstimationPCA was conducted using the GCTA software [16] to assess the population structure, and PLINK v1.07 was used to calculate the LD decay distance, which was evaluated as the squared correlation of alleles (r2) with a window size of 1000.2.5. Single-Trait and Multi-Trait Genome-Wide Association StudiesThe GEMMA software [17] was used to implement the linear mixed model (LMM) for the single-trait GWAS of each internal organ weight trait, including heart weight, liver weight, spleen weight, lung weight, kidney weight, and stomach weight. GEMMA calculated the genomic relatedness matrix (GRM) between individuals to account for the population structure. The mixed linear model was as follows:y=Wα+Xβ+u+ε where y is a vector of phenotypic values for each internal organ weight; W is the correlation matrix of covariates (fixed effects), including the top five eigenvectors of PCA, farm, sex, and slaughter lot; α is a vector of corresponding coefficients including the intercept; X is the genotypic vector of the SNP markers; β denotes the effect size of the SNP markers; u is a random effects vector, u~MVNn (0, λτ−1K); ε is the residual vector, ε~MVNn (0, τ−1In); λ is the ratio of the specified variance components; τ−1 is the variance of the residuals; K denotes the kinship matrix; I is the unit matrix; n is the number of individuals in the DLY population; MVNn denotes the multi-dimensional normal distribution.Moreover, the GEMMA software [17] was used to implement the multivariate linear mixed models (mvLMMs) [18] for multi-trait GWASs to assess pleiotropic SNPs. The mvLMMs and LMMs were both implemented as described in previous studies [19]. In the current study, the LMMs and mvLMMs in the single-trait GWAS and the multi-trait GWAS utilized the same covariates. The multivariate linear mixed models were as follows:Y=WA+xβT+U+E; G ∼MNn×d(0, K,Vg), E ∼MNn×d(0,In×n,Ve) where Y is a matrix of six internal organs for 1518 individuals; W is a covariable matrix (fixed effects); A is a matrix of the corresponding coefficients; x is a vector that marks the genotypes; β is a vector of marker effect sizes for six internal organs’ weights. U denotes the random effects; E is a matrix of errors; K denotes the kinship matrix; Vg denotes symmetric matrix of genetic variance component; I is an identity matrix; Ve denotes a symmetric matrix of the environmental variance component; MNn×d(0,V1,V2) denotes the n×d matrix normal distribution with mean 0; V1 denotes row covariance matrix; V2 denotes column covariance matrix.Furthermore, the Bonferroni correction can lead to an overcorrection and can be too conservative, this can result in a limited number of labeled association p-values that meet the standard across the genome. This can lead to a high false-negative rate. To address this issue, the false-discovery rate (FDR) was employed as a correction to the threshold [20]. Thus, the threshold p-value was calculated as P=FDR∗NM; the FDR was set to 0.01, N is the number of SNPs with p-value less than 0.01, and M refers to the total number of SNPs after quality control. Moreover, quantile–quantile (Q–Q) plots were constructed for the six internal organ weight traits to further assess the population structure.In addition, the PLINK v1.07 and Haploview v4.2 software [21] were implemented to perform the haplotype block analysis in chromosomal regions with multiple significant SNPs. The default parameters of Haploview 4.2 [22] (MAF > 0.05, Mendelian error < 2, and p-value < 10−3 for the HWE test) were used to define the linkage disequilibrium (LD) blocks of SNPs.2.6. Estimation of Heritability and Phenotypic VariationIn the present study, the restricted maximum likelihood (REML) method was used to assess the SNP-based heritability of each internal organ weight trait, and the percentage of phenotypic variation that could be explained by significant SNPs was calculated using GCTA software. SNP-based heritability and the percentage of phenotypic variation explained by significant SNPs were calculated as follows [23]:y=Xβ+g+ε with var(y)=Agσg2+Iσε2 where y is the phenotypic value of each internal organ weight trait; β is the vector of fixed effects, including the top five eigenvectors of PCA, farm, sex, and slaughter lot; X is an association matrix; g is the vector of total genetic effect of all the qualified SNPs for the 1518 DLY pigs; Ag is the genomic association matrix between different individuals; σg2 is the additive genetic variance captured by either the genome-wide SNPs or the selected SNPs; σε2 refers to residual variance.2.7. Candidate Gene Search and Function AnalysisOur previous studies on this population showed that the average r2 of 0.2 is about 200 kb apart [24]; the range for searching for the functional gene closest to the position of the significant SNP is determined based on the LD decay distance (r2 = 0.2) of the populations [25]. We used the “biomaRt” package [26] in R, based on the Sus scrofa 11.1 genome version database (http://ensemble.org/Sus_scrofa/Info/Index, accessed 20 September 2022). Genes nearest the significant SNPs are list in Tables. We conducted a search of both PubMed and the relevant literature to examine the correlation between the nearest peak SNPs of all the candidate genes and the internal organ weight traits being analyzed.3. Results and Discussion3.1. Phenotype Statistics and Heritability EstimationThe descriptive phenotypic statistics and estimated heritabilities (h2) for analysis of the internal organ weights are listed in Table 1. The weight of internal organs is a crucial indicator of internal organ development and has a significant impact on organ function. In the current study, the average Heart WT, Liver WT, Spleen WT, Lung WT, Kidney WT, and Stomach WT in DLY pigs were 455.57 g, 1763.61 g, 212.54 g, 1020.54 g, 0.41 kg, and 727.68 g, respectively. The estimated heritabilities of Heart WT and Lung WT were the lowest at 0.21 ± 0.04 and 0.28 ± 0.04, respectively, and all other organ weights had had moderate to high estimated heritabilities, ranging from 0.36 ± 0.04 to 0.49 ± 0.04. Similar to the results of a previous study, the estimated heritabilities of Heart WT, Liver WT, Spleen WT, and Kidney WT were between 0.35 and 0.54, which were moderate to high estimations [5], indicating that the estimated heritabilities of the weight of an internal organ is generally high in pigs and there is considerable room for improving the genetic contribution through breeding. Furthermore, the coefficients of variation were the lowest for Lung WT and all other traits were relatively high, indicating individual heterogeneity, low trait selection intensity, and high breeding potential.Moreover, the genetic and phenotypic correlation coefficients among Heart WT, Liver WT, Spleen WT, Lung WT, Kidney WT, and Stomach WT are listed in Table 2. The results revealed moderate to low genetic correlations among the six internal organ weight traits. Heart WT had moderate genetic correlations with Liver WT, Lung WT, and Kidney WT, suggesting that these traits could be improved together in pig breeding programs. On the other hand, Stomach WT showed close to 0 genetic correlations with most of the other traits, indicating Stomach WT traits are less influenced by other traits when they are inherited. Therefore, reasonable breeding strategies need to be designed to improve internal organ weight traits. The phenotypic correlation results showed that the correlation coefficients between the phenotypes were at moderate to high levels, excluding the low phenotypic correlation coefficients between Lung WT and Liver WT, and Spleen WT and Kidney WT, especially the phenotypic correlation coefficients of Liver WT and Kidney WT were as high as 0.62. When selecting for a certain phenotype in pig breeding, it is advantageous to also consider other related traits.3.2. Population Structure and LD decayPopulation stratification is known to lead to false-positive results in GWASs. To detect potential population stratification, we performed PCA and added the first five principal components to the covariates of the GWAS model to correct for the population structure. Moreover, our previous study showed that the LD decay coefficient of the analyzed DLY pig population with r2 decayed to 0.2 at a physical distance of 200 kb [24], indicating that the DLY population is diverse with a weak linkage between loci, which facilitates the detection of key SNPs for internal organ weight traits. In addition, Q–Q plots were generated for Heart WT, Liver WT, Spleen WT, Lung WT, Kidney WT, and Stomach WT to further assess population stratification (together with the Manhattan plots: Figure 2). The expansion coefficients (lambda) of the Q–Q plots for all six internal organ weight traits were close to 1, and no overall systematic bias was observed, signifying a negligible effect of the DLY pig group structure on GWASs.3.3. Single-Trait GWASsSingle-trait GWASs were performed for the weight of the heart, liver, spleen, lung, kidney, and stomach. The results showed that 6, 4, 3, 4, 3, and 4 SNPs were significantly associated with the weight of each organ, respectively. The results of these single-trait GWASs are presented in Figure 2 and Table 3. Notably, it is the first time that significant SNPs associated with Stomach WT have been identified in pigs. Furthermore, on the basis of the LD decay map, a region of 200 kb before and after the key SNPs was defined as a region to screen for candidate genes [24]. For heart weight, six significant SNPs were identified, located on SSC5, 6, 7, 12, and 14. These six SNPs surpassed the significance threshold of 1.01 × 10−4. Figure 2B shows an expansion coefficient lambda (λ) of 1.006. Details of the significant SNPs are listed in Table 3. The most significant SNP, WU_10.2_12_6703865 on SSC12, explains 1.70% of the phenotypic variation and is about 44 kb downstream of the CD300LB gene. The CD300LB gene is a triggering receptor expressed on bone marrow cells that regulates the cytosolic process of bone marrow cells [27], and the CD300LB protein stimulated by T cells regulates DNMT3A mutation and alters immune cells in heart failure [28].For liver weight, four significant SNPs were detected on SSC4, SSC9, and SSC10 with a λ of 0.999 (Figure 2C,D and Table 3). These four SNPs surpassed the significance threshold of 1.04 × 10−4. The top SNP, H3GA0028070, accounted for 2.10% of the phenotypic variance and is located within the TPK1 gene. A significant SNP, named ASGA0044340, 12 kb upstream of H3GA0028070, also located on TPK1, explained 0.82% of the phenotypic variation. According to reports, TPK1 is a cofactor of certain enzymes associated with the glycolysis and energy production pathways. It is involved in the metabolism of water-soluble vitamins and cofactors and the thiamine metabolic pathway, and mutations in TPK1 can cause thiamine metabolic dysfunction syndrome [29]. In addition, knockdown of this gene can lead to glycogen storage dysfunction [30]. However, no studies have shown TPK1 to be directly associated with liver development and weight in pigs.The GWAS results of Spleen WT identified three significant SNPs, located on SSC3, SSC9, and SSC18, with a λ of 0.972 (Figure 2E,F and Table 3). All three SNPs surpassed the threshold of significance (p < 1.20 × 10−4). The top SNP, ALGA0098928, explained 2.22% of the phenotypic variation and is located within POU6F2. POU6F2 is a suppressor associated with nephroblastoma (WT) that regulates cell proliferation and specific differentiation [31]. According to the RT-qPCR results, the expression of POU6F2 is associated with renal morphogenesis [32], suggesting that POU6F2 may be closely associated with spleen weight traits.For lung weight, four significant SNPs were detected on SSC1, SSC7, and SSC11 with a λ of 1.008 (Figure 2G,H and Table 3). These four SNPs surpassed the significance threshold of 1.04 × 10−4. An SNP named ALGA0110225 explained 1.81% of the phenotypic variance and is located 97 kb downstream of PBX3, indicating that PBX3 and ALGA0110225 may both play a role in Lung WT. A literature review revealed that PBX3 is directly regulated by targeting NBPF10, miR-144, and miR-224, which are directly associated with lung cancer cell proliferation [33]. In addition, overexpression of PBX3 promotes the proliferation of A549 cells (lung cancer histiocytes) [34]. Therefore, we believe PBX3 to be a promising candidate gene for influencing Lung WT, and the regulatory mechanism needs further investigation.We performed GWASs with the Kidney WT trait in DLY pigs and detected three SNPs that were above the significance threshold (p < 1.03 × 10−4) (Figure 2I and Table 3). Figure 2J shows that the lambda is 0.987. WU_10.2_15_153747936 on SSC15 explains 1.97% of the phenotypic variation and is located 341 kb downstream of the HDAC4 gene. Because the LD decay distance is 200 kb, it follows that HDAC4 may not have a significant effect on Kidney WT traits.QTL and significant SNPs have not been previously reported in relation to Stomach WT. Thus, this study is the first GWAS on pig Stomach WT. Herein, four significant SNPs were identified for the DLY pigs that were above the significance threshold of p < 1.17 × 10−4 (Figure 2K and Table 3). Of the four significant SNPs, three SNPs were simultaneously located on SSC8 and both MARC0052872 and ALGA0106192 were located within the UNC5C gene. Furthermore, the distance between MARC0052872 and ALGA0106192 was only 21 kb, explaining 2.53% and 2.62% of the phenotypic variation, respectively. ASGA0101191 was 100 kb away from the aforementioned SNPs with a phenotypic variation value of 2.14% and was located within BMPR1B. A literature review revealed that UNC5C plays a dominant role in netrin-1/UNC5C-mediated axonal rejection [35] and that its promoter region sequence binds to p53 and acts as a target of p53 to regulate apoptosis [36]. As regards the BMPR1B gene, it has been shown that the BMP family is expressed in the early organ and tissue formation during mouse embryonic development [37]. However, neither the BMPR1B nor the UNC5C gene is directly associated with internal organ weight traits.The above GWAS results show that none of the SNPs associated with internal organ weight overlapped with those previously reported QTL documented in the pig QTL database [1]. This may have been due to the fact that most studies focused on the breeding of native Chinese pigs, and fewer studies were conducted on DLY three-way crossbred commercial populations with significant breed differences. Moreover, the significant SNPs did not overlap in the six traits, i.e., none of the SNPs were polymorphic, which may be related to the low genetic and phenotypic correlation between traits and the low density of genetic markers, which was further verified by multi-trait GWASs.3.4. Haplotype Block AnalysisFigure 3 shows the LD pattern of significant SNPs associated with Stomach WT. In this study, multiple SNPs associated with Stomach WT were in close proximity to each other, with two significant SNPs on SSC8, which is located in a 21 kb region within the UNC5C and BMPR1B genes (the gene function is described above). The insufficient density of 50K microarray markers resulted in a low number of SNPs with linkage disequilibrium, which limited the resolution of the genetic architecture of key SNPs for the trait to some extent.3.5. Multi-Trait GWASsIn order to improve the statistical effect, multi-trait GWASs were individually performed for each SNP by combining the joint analysis of six internal organ weight traits. This revealed the genetic factors with significant interactions among different traits in the same individual under the same environment. Manhattan plots of the multi-trait GWASs are shown in Figure 4.The multi-trait GWASs combining six internal organ weight traits identified four significant SNPs with polymorphisms affecting the phenotypes, ALGA0032998, H3GA0028070, MARC0052872, and ALGA0106192 (Figure 5 and Table 4). SNP ALGA0032998 explained 1.36% of the phenotypic variation and is located within the ANO6 gene. The overexpression of CCR7 was observed to enhance the migration of BxPC-3 cells under the induction of the ANO6 gene, which is a potential mediator of ANO6 expression through the ERK signaling pathway. This promotion of migration was also seen in pancreatic ductal adenocarcinoma cells [38]. The single-trait GWAS described the effects of three SNPs (H3GA0028070, MARC0052872, and ALGA0106192) located on TPK1 and UNC5C genes on the weight of the liver and stomach. These SNPs were not only significant in the single-trait GWAS but were also found to be simultaneously associated with the weight of all six internal organs, suggesting that these four SNPs have pleiotropic effects. Furthermore, no additional SNPs, independent of the single-trait GWAS results, were found. Similar results were previously reported by Guo et al. [39], in which no additional SNPs, independent of the single-trait GWAS results, were detected in the multi-trait GWASs for backfat thickness, carcass weight, and body weight in the DLY and Duroc populations. The reasons for this situation are manifold. For example, the complexity of the genetic architecture of the internal organ weight trait and the low marker density result in a low number of SNPs reaching significant levels. This renders LD detection insufficient and increases the difficulty of screening for co-dominant SNP or QTL regions. Thus, a larger sample population and a higher marker density are required to screen for loci associated with internal organ weight.4. ConclusionsIn this study, we conducted single-trait and multi-trait GWASs on the internal organ weights of 1518 DLY pigs. A total of 24 significant SNPs were detected in the single-trait GWAS results for six internal organ weight traits. The four significant pleiotropic SNPs identified via multi-trait GWASs were associated with six internal organ weight traits, confirming the results of the single-trait GWASs and improving our ability to reveal the genetic architecture of organ weight traits. TPK1, POU6F2, PBX3, UNC5C, and BMPR1B were highlighted as potential genes responsible for differences in Liver WT, Spleen WT, Lung WT, and Stomach WT among individuals according to their gene functions. In summary, the results of this study contribute to our understanding of the genetics of internal organ weight traits in DLY pigs by assigning higher weights to relevant SNPs and key genes in the genome.

animals : an open access journal from mdpi

10.3390/ani11051373

PMC8151839

Consumers are increasingly concerned about the sustainable production of food, leading producers and scientists to evaluate farming practices that preserve environmental resources, provide adequate production, and are economically viable. However, there are challenges to synthesize these results and apply them on-farm in a holistic nature. Simulation modeling of farm systems, such as the dairy system, can allow producers, industry members, and policy makers to prioritize interventions that improve sustainable outcomes. We introduce the Animal Module of the Ruminant Farm Systems (RuFaS) model—a whole farm dairy system model—and describe its use to assess the environmental impact of improved feed efficiency in dairy cows. By decreasing the amount of feed intake required to produce the same amount of milk, the RuFaS model provides estimates of the reduction in feed use, enteric methane, and manure production.

Dairy production is an important source of nutrients in the global food supply, but environmental impacts are increasingly a concern of consumers, scientists, and policy-makers. Many decisions must be integrated to support sustainable production—which can be achieved using a simulation model. We provide an example of the Ruminant Farm Systems (RuFaS) model to assess changes in the dairy system related to altered animal feed efficiency. RuFaS is a whole-system farm simulation model that simulates the individual animal life cycle, production, and environmental impacts. We added a stochastic animal-level parameter to represent individual animal feed efficiency as a result of reduced residual feed intake and compared High (intake = 94% of expected) and Very High (intake = 88% of expected) efficiency levels with a Baseline scenario (intake = 100% of expected). As expected, the simulated total feed intake was reduced by 6 and 12% for the High and Very High efficiency scenarios, and the expected impact of these improved efficiencies on the greenhouse gas emissions from enteric methane and manure storage was a decrease of 4.6 and 9.3%, respectively.

1. IntroductionThe FAO [1] predicts the world population will reach 9.8 billion people by 2050, requiring significant changes to food production to provide nutritious products to all people. Expanding populations with increased disposable income has led to a shift in dietary preferences, with emerging markets favoring animal products [2]. Animal products are important food, industrial, and health sources but are associated with increased environmental costs [3,4]. While some markets favor the production of more animal proteins, others are increasingly concerned about environmental impacts and sustainable production [5]. Sustainable production requires a systems approach to account for food security, environmental stewardship, and societal impacts [6].Managing the need for sustainable food production requires a careful allocation of limited farm resources. One method to guide policy, support farm decisions, and evaluate novel technologies is the use of simulation models. The effect of implementing new technologies on-farm or changing production management can be evaluated on multiple outcomes, generally providing a more robust evaluation than traditional research experiments. Existing whole-farm models for dairy production include the Integrated Farm Systems Model [7], DairyMod [8], DyNoFlo [9], and SIMS(DAIRY) [10], but these models have limitations that prevent their wide-scale applicability in current and future scenarios. Incorporating vast amounts of data from farm records or adapting to new technologies is often impossible with existing models due to the model structure or restrictions in the model code. Thus, we saw the need to develop a new farm simulation model that can adapt to changing technologies and support sustainable dairy production [11]. The Ruminant Farm Systems Model (RuFaS, Figure 1) incorporates modern computer coding practices centered around clarity and adaptability to respond to evolving technologies in the dairy industry. RuFaS embraces the key characteristics for next-generation agricultural systems models described by Jones et al. [12]: “technological advances; open, harmonized data; transdisciplinary collaboration; modularity and interoperability; user-driven data and model development”.Within the RuFaS framework, the environmental impacts of dairy production can be determined from each of the four main parts of a dairy farm: animals, manure, field, and feed storage. Previous life-cycle assessments (LCA) of dairy production have estimated the total environmental impacts of dairy production and the relative contributions of each part of the dairy farm and supply chain (e.g., [13,14]). Thoma et al. [13] attributed 73% of GHG emissions from US dairy production in 2008 to on-farm sources with 26, 34, and 33% of the on-farm emissions attributed to feed production, enteric methane, and manure emissions, respectively. The works by Capper et al. [15] and Capper and Cady [14] demonstrate the significant reductions in farm-gate GHG intensity (kg CO2 eq/kg milk) achieved in past decades by comparing LCA estimates of US dairy GHG footprints in 1944, 2007, and 2017.Much of the reduction in GHG emissions per kg of milk is attributable to improvements in feed efficiency. In particular, Capper and Bauman [16] highlight the impact of increased milk production per cow that results in the dilution of maintenance effect, whereby the fixed cost (and GHG production) of maintaining an animal’s basic life function is reduced per unit of product.Although the dilution of maintenance concept describes improvements in individual animal efficiency with increases in milk output, many other factors will influence feed efficiency and the subsequent environmental footprint over an animal’s lifetime and at a farm-system scale. In addition to some of the reproduction and life-cycle management impacts mentioned by [14], feed efficiency can be improved by selecting cows that consume less feed while producing the same amount of milk. Conceptually, this is equivalent to reducing the denominator in the feed efficiency equation, whereas the dilution of maintenance is focused on increasing the numerator. While we have been effective at increasing the numerator increasing the temporal milk production per cow, we have been less effective at improving efficiency by reducing the feed required to produce a given quantity of milk. The work in [14] illustrates this point with the equivalent nutrient requirements for milk production between 2007 and 2017. Both mechanisms for improving feed efficiency (increase in milk production per animal and reduction in feed intake per kg of milk) are influenced by animal genetics and diet composition, among other things. However, the methods for measuring and influencing these mechanisms on-farm are quite different.Residual feed intake (RFI, kg DMI/cow/d) is a metric used to evaluate differences in feed efficiency between cows with the goal of identifying animals with improved feed efficiency due to reduced feed consumption for an equivalent production. There is a long history of interest in using RFI as the phenotypic expression of feed efficiency to inform breeding programs [17,18,19], but implementation and adoption has been slow, in part because it is more difficult to measure individual animal intake than milk production. New technologies to measure or estimate feed intake and advances in genomic selection methodologies make selecting for improved feed efficiency through reduced RFI [19,20] more accessible. To advance research into, and the adoption of, different methods to improve feed efficiency, it is important to be able to estimate the expected system-wide impacts of the proposed methods.Whole-farm models like RuFaS can be used to estimate system outcomes to management or biological changes, but most existing models cannot easily represent the impact of a change in RFI. The aggregate or static representation of animal traits and simplifications in the connection between diet and feed consumption, animal life-cycle, and methane and manure production mean that the impact of a reduced individual animal RFI on whole-farm feed consumption and downstream methane and manure production are difficult to include in most farm systems models. We have intentionally designed the Animal Module of RuFaS to enable the investigation into a wide variety of current and future precision management practices that are expected to influence whole-farm efficiency and environmental outcomes [11]. As a result, the RuFaS animal module is easily adapted to address the question of how changes in feed efficiency influence herd and whole-farm environmental outcomes.The objective of this work is to document the advances in dairy farm simulation modeling made by the Animal Module as part of the RuFaS modeling ecosystem and to demonstrate its utility through a case study of the environmental impacts of improving feed efficiency. Thus, our first objective is to provide detailed documentation of the algorithms that define management, production, and nutrient flows in a dairy herd. For our second objective, we describe and implement methods to add the representation of variable RFI and use this altered model to compare the impact of RFI on feed intake, enteric methane production, and manure production in a simulated herd with 1000 lactating Holstein cows. The future application of the RuFaS model will support scientific investigation and on-farm decision support related to precision-feeding, breeding, and other dairy herd management practices.2. Materials and MethodsThe Ruminant Farm System (RuFaS) model consists of four biophysical modules: animal, manure handling, crop + soil, and feed storage (Figure 1). The simulation inputs include time of simulation, herd characteristics, crop characteristics, and other farm decisions. The required inputs follow a tiered file structure that separates inputs that designate the whole-farm and simulation structure from inputs specific to each of the modules with increasing level of detail associated with inputs at lower tiers. The model uses a daily time-step and is programmed in Python, an adaptable and easy-to-read computer programming language.The model inputs span decisions at the farm, herd, and animal levels (Appendix A). The first set of inputs include the dates of simulation and corresponding weather information. At the farm level, the user can define housing specifications and feeds available for ration formulation and target herd size. At the herd level, the user can specify the breed, reproduction protocols, and lactation curve parameters. Inputs that define animal characteristics include parameters that define the bodyweight distribution, reproductive efficiency, and probability of disease. The model outputs are exported to CSV and graphic images.The animal module simulates the individual animal from birth to culling, accounting for farm management decisions and individual animal responses to those decisions. Animals are simulated individually, and outputs such as animal growth, animal production, and manure production are estimated daily for each animal. The main routines of the animal module are the animal life cycle, animal nutrition, and manure excretion (Figure 2). The animal management class manages the animal routines—performing the algorithms to account for all animals and animal classes in the simulation.The animal life cycle simulates the animal from birth to culling, encompassing weaning, first reproductive cycle, first lactation, and subsequent calving and lactation cycles. There are five animal classes: calf, heifer I, heifer II, heifer III, and cow. Calves transition to heifers based on weaning day. The heifer stage is divided into three categories: heifer I (from weaning to breeding period), heifer II (first estrus to transition period), and heifer III (transition period, default one month before calving). Cows include lactating and dry cows. Animals are culled from the herd or added to the herd depending on the management inputs to maintain the herd size specified in the user input.The life cycle uses stochastic, Monte Carlo methods incorporating the probability of events and parameter averages, distributions, and standard deviations to simulate random variables [21]. The probabilities of each outcome, such as the calf’s sex, conception, or culling, are inputs to the model. Each event’s probability is compared to a randomly generated number between 0 and 1, and, if the probability is greater than the random number, the event occurs. To determine the outcome of the random variables in the model, such as bodyweight, a random number is drawn from a user-defined distribution for that parameter. To prevent extreme values, upper and lower bounds are defined based on recommendations from the literature or biological principles.Calves are the first animal class initialized, and the success of a calf’s birth, its sex, and its longevity on the farm are determined by stochastic processes as described above. A target average daily gain (ADG) is used to estimate growth until the weaning day is reached. The daily update for the calf includes a check to see if it has reached its weaning day and a method to record bodyweight changes.The three classes of heifer are used to organize reproduction and life events during the heifer period. Heifer I includes growth until the breeding start day is reached. If timed artificial insemination (TAI) is used, no estrus simulation occurs. The target ADG is set to achieve 55% of mature shrunk bodyweight (BW) at first pregnancy [22]. The heifer II stage incorporates reproductive protocols for conception and pregnancy success. The length of the estrous cycle is determined stochastically, and the simulated day of estrus determines the breeding day. The reproduction programs include timed artificial insemination, estrus detection, and synchronized estrus detection [23]. If the conception was successful (determined stochastically), a pregnancy update determines the gestation length. Stochastic simulation events simulate pregnancy checks at conception and at two other time points to confirm pregnancy. The loss of pregnancy occurs based on a pregnancy loss rate. Pregnancy loss results in an abortion day, and the heifer can be rebred according to rebreeding protocols. Conceptus growth is assumed to be zero prior to 51 days in pregnancy. The total conceptus growth is calculated as:Total conceptus weight = 0.0148 × gestation length − 2.408 × calf birth weight(1)Conceptus growth from 51 to the end of gestation is calculated as Conceptus growth = 3 × (total conceptus weight1/3/gestation length − 50)3 × (days in pregnancy − 50)2(2)A heifer moves to the Heifer III class when they are within the user-specified pre-fresh period. Heifers are culled from the herd if they exceed a user-specified heifer reproduction cull time, i.e., if the heifer is not pregnant by the specified age. The target ADG is adjusted to reach 82% of mature shrunk body weight (BW) at the first parturition.Cows are managed from first calving to culling from the herd. When the days in pregnancy equal the gestation length, the cow calves reset the days in pregnancy to zero and initiate milk production. The milking update estimates milk production from a Wood lactation curve [24,25]. The end of the lactation cycle is a user input, based on the days in pregnancy when the cow is dry. The reproduction program for cows restarts after calving with options for estrus detection, estrus detection with timed artificial insemination, or timed artificial insemination. The start times, pregnancy check times, and success rates of the reproduction protocols are specified by the user. Cows are targeted to grow to their mature body weight by the end of their second lactation. The target average daily gain for cows is set to reach 92% of mature shrunk BW at the end of the first lactation and full mature shrunk body weight by the end of the second lactation. Estimates for lactation-related body weight changes represent the tissue change due to lactation and can be positive or negative. This value is only estimated for lactating cows and is assumed to be 0 for dry cows. Lactation BW change =−20/65 × exp(1-days in milk/65) + 20/(65^2) × days in milk × exp(1-days in milk/65) if parity = 1Or −40/70 × exp(1-days in milk/70) + 40/(70^2) × days in milk × exp(1-days in milk/70) if parity > 1(3) where BW is bodyweight (kg). The daily body weight change of a cow is the sum of the target ADG, conceptus weight change (same as heifer pregnancy), and bodyweight tissues.The user input herd number is used to maintain the herd size. If the number of heifers is greater than the herd needs, Heifer IIIs will be sold. If the number of heifers is less than the herd needs to maintain the herd size, Heifer IIIs are purchased from the replacement market. Animals can be culled at any life stage depending on user inputs. Male calves can be sold, heifers culled for reproduction failures, and cows for reproduction or milk production issues. Health reasons for culling include lameness, injury, mastitis, disease, udder problems, and unknown issues.The animal nutrition routine automatically formulates a ration to meet animal nutrient requirements using the feeds specified by user inputs. Feed nutrient composition is based on feed tables in the 2001 NRC with new 2021 starch values added for starch concentrations (Table S1). The maximum feed intake of feeds can be specified by the user, and default settings for certain feeds are set based on the literature.For calves, the feed intake of milk or milk replacer is assumed to be 10% of the birth weight, a common industry practice (E. Miller-Cushion, personal communication) [26]. The intake of calf starter was estimated from data reported by Khan, M. A., D. M. Weary, and M. A. von Keyserlingk [27] using a broken line regression:Starter intake (kg) =−0.24783 + 0.0049567 × body weight if body weight ≤ 69.365Or −6.2263 + 0.091145 × body weight if body weight > 69.365,(4) where the body weight is in kilograms. During weaning, the length of the weaning period is used to calculate the reduction in milk intake each day. Calf feed intake is used to determine the energy allowable or protein allowable gain. The minimum can be used for determining the daily calf growth as a mechanistic alternative to the ADG estimates set in the life-cycle update, based on the length of the weaning period.The ration formulation for heifers and cows includes four key processes: requirement calculation, compiling types and amounts of available feeds, nonlinear program optimization, and ration reporting. Rations for heifers and cows are formulated to meet the energy, protein, Ca, and P requirements provided in the NRC [22]. Individual animal requirements are calculated based on animal bodyweight, milk production, growth, and the environment. Individual requirements are then averaged by pen before completing the optimization to generate a ration that meets the average nutrient requirements of each pen. Pen-level summaries of nutrient requirements can be adjusted to formulate diets that meet or exceed the requirements of a larger proportion of the animals within the pen. Net energy, metabolizable protein, and mineral supply are calculated based on the nutrient composition in the feed library and the recommendations in the NRC [22].Automated ration formulation currently uses an algorithm for least-cost non-linear optimization similar to that described by Rotz [28]; however, we plan to implement additional single and multi-objective optimization methods in future iterations of RuFaS (e.g., goal-programming and multi-objective optimization that include environmental outcomes [29]). In the current version, the optimization routine uses sequential quadratic programming to minimize the price of the ration while meeting nutrient requirements:Min z1 = Σj = 1 cjxj, xj ≥ 0,(5) Σj = 1 aijxj ≥ bi, i = 1,…, m(6) where z1 is the price of the ration ($), cj is the price of feed j ($/kg), xj is the amount of feed j (kg), and aij is the nutrient concentration of the feed (Mcal/kg or %). Sequential quadratic programming is a non-linear optimization method capable of handling both discrete and continuous constraints [30]. In addition to the nutrient requirements for net energy, metabolizable protein, Ca, and P as described above, we include constraints for a minimum dietary NDF concentration of 25% (DM basis), maximum NDF concentration of 40%, minimum dietary forage NDF concentration of 19% (DM Basis), and a maximum dietary fat concentration of 7% (DM basis) [31]. If the algorithm reaches the 100-iteration limit and does not satisfy all constraints, we reduce the estimated daily milk production by 0.5 kg and reenter the optimization routine.The ration formulated for each pen is applied to each animal for the user-defined period between ration formulations, and the total amount of feed consumed by each pen is recorded daily. The manure subroutine calculates the animal manure excretion and sums the manure excretion by both animal class and pen. Total manure (kg as-excreted), total solids (kg DM basis), and methane emissions (g/d) are calculated for each animal and for all animal classes based on animal and dietary characteristics.For calves, total manure and manure solids are calculated according to animal bodyweight [32]. The average methane emissions from calves reported by Pattanaik, A. K., V. R. B. Sastry, R. C. Katiyar, and M. Lal [33] are used to estimate the methane production from calves:Methane emis = (0.013 × bw0.75 × 4.184)/0.05565,(7) where methane emis are methane emissions (g/d) and bw is animal bodyweight (kg).For all classes of heifers, total manure, manure solids, and volatile solids are calculated from ASABE [32] equations. The empirical relationship between methane production (L/d) and dry matter intake described Boadi, D. and K. Wittenberg [34] is used to calculate the methane emissions for heifers.Manure production and methane emissions of cows are divided into lactating and dry cows. For lactating cows, fecal water, total solids, urine excretion, manure excretion, volatile solids (separated into degradable and nondegradable volatile solids) are calculated according to [35] using dry matter intake as a predictor variable. The user has the option to select from three enteric methane emission calculations: (1) the US animal model described by Niu et al. [36], (2) the Mills et al. [37] Mitsherlich Model 3, or (3) the IPCC [38] Tier 2 model. The Mills et al. [37] and IPCC [38] models were selected to provide a comparison to other models commonly used in the dairy industry and environmental sciences. For instance, the Mills et al. [35] Mitsherlich 3 equation is used in the Cornell Net Carbohydrate and Protein System [39] and Integrated Farm System Model [28]. The dietary gross energy is calculated for the IPCC [38] Tier 2 model according to [31] (Equation (S3)). For dry cows, manure excretion, total solids, and volatile solids are calculated according to ASABE [32]. The methane production from dry cows is calculated according to Mills et al. [35]. The animal model described by Niu et al. [34] includes milking parameters, making that equation a poor choice for calculating dry cow methane production. However, most methane emission models are developed and evaluated with lactating cow data, leaving inexact predictions for dry cows. The IFSM model uses the Mills et al. equation [37] to calculate methane emissions from all cows, and thus that approach was used in this model, but future work to improve the methane emission prediction from dry cows is needed.2.1. Running the SimulationThe user has the option to create a new initialization herd for each simulation. Starting a simulation with an initialization herd substantially reduces the time for the model to stabilize. The initialization herd creates a database of animals that are used to populate the herd and replacement heifer market on day 1 of the simulation. The replacement heifer market is a large number of Heifer III class animal instances that can be selected to meet the herd number targets after cow culling. When the user chooses to create a herd initialization database, a minimum of 1000 calves are created and simulated for a minimum of 5000 days. The initialized herd requires the same user inputs as the simulated herd to create a large herd population with the same characteristics (e.g., breed, BW, etc.) as the animals in the desired simulation. An SQLite database file is created to store the simulated animals at the end of the initialization of the program. The simulated animals represent the five animal classes plus a replacement herd of instances of the Heifer III class (Table 1). Random draws from the initialized database of animals form the herd used for the simulation. The number of calves, heifers, and cows along with the total lactating herd size is a user input.Each animal is updated daily through the execution of a daily update function that is specific to each animal class. The function updates each animal according to its life cycle functions, calculates the daily manure excretion, and, if it is the end of the ration interval, calculates the animal nutrient requirements, redefines lactating cows’ pen assignments, and creates a new ration. Animals are culled from pens, and new animals are selected from the replacement during the daily update.2.2. Feed Efficiency Case StudyTo demonstrate the utility of the RuFaS Animal Module, we developed a method to alter the animal-level feed efficiency and used the model to estimate the expected outcomes of a herd that has been bred for improved feed efficiency. We created a feed-efficiency random-variable (ρi) that represents the proportional change in DMI of animal i. Because the biological mechanisms that drive feed-efficiency are still not well understood, the objective of the parameter ρ is to modify the cow’s DMI without changing the ration formulation or predicted milk production. During the initialization of each animal, a random draw from the user-defined distribution for this parameter is assigned to that animal’s record. After a ration has been formulated for each pen, when the ration is assigned to each animal’s record, the expected intake for that animal is modified by multiplying the previous model-predicted DMI by ρi. The individual animal DMI that has now been adjusted to reflect that animal’s feed efficiency is then used as the input to predict the animal’s enteric methane and manure production. The result is a representation of the RFI phenotype that is commonly measured and used to study feed efficiency.To illustrate how changes in the distribution of a herd’s RFI are expected to influence environmental outcomes, we simulated herds with 3 different distributions for ρ and compared the predicted outcomes with each other. We based our feed efficiency distributions on recent studies that measured and reported RFI [40]. Based on these studies, we set the standard deviation of the RFI distribution to 6% of the intake for all scenarios and set the mean RFI to 0, −1 SD, and −2 SD, so that the variability in feed efficiency is similar between scenarios but there is a mean shift increasing the mean population efficiency. The ρ distributions for each scenario are shown in Figure 3 and are designed to represent (1) a Baseline scenario in which the herd has an average RFI that is similar to present-day efficiency, (2) a High efficiency scenario in which the herd’s mean efficiency shifted to the 16 percentile of present-day efficiencies, and (3) a Very High efficiency scenario in which the herd’s mean efficiency shifted to the 2.5 percentile of present-day efficiency.For each scenario, we simulated a 1000-cow Holstein herd for 1 year and held all inputs constant (Table A1, Table A2 and Table A3), with the exception of the input to determine the mean of the distribution of ρ. To support the comparison between scenario runs, we set a common seed for the random number generators so that the outcomes of all Monte-Carlo processes are constant between model runs except where the ρ input distribution was altered.3. Results3.1. Herd DemographicsAs expected, the herd demographic outcomes from all three simulations were the same because the inputs defining the number of animals, grouping management, and reproduction and culling protocols were held constant and the differences in ρ are not expected to influence the animal life-cycle outcomes. The numbers of animals in each animal class (Calves, Heifer I, Heifer II, Heifer II, and Cows divided into lactating cows and dry cows) are shown in Figure 4. This figure demonstrates the capability of the life-cycle model to maintain a consistent number of animals in each animal class across the timeline of the simulation. All female calves were kept in this simulation (A1. “keep female calf rate” = 1), which resulted in a large number of animals in the Heifer I and Heifer II classes. In this situation, the life-cycle algorithms maintain a constant herd size by deciding when to keep a pregnant heifer (Heifer III) or sell her as a replacement animal. During this year of simulation, the model removed 666 heifers from the simulated herd, to be sold as replacement animals. In addition to selling replacements, the life-cycle algorithms simulate the culling of animals for a variety of health and production reasons. The numbers of animals culled for each reason during the 365-day simulation are given in Table 2.3.2. Feed Efficiency Case StudyWe successfully implemented a feed efficiency parameter that influences the individual animal intake and reflects the variation in RFI seen in commercial and research herds. The simulated feed efficiency outcomes are given in Table 3 to demonstrate that we achieved the target mean feed efficiencies for the Baseline, High Efficiency, and Very High Efficiency scenarios. The simulated RFI for the Baseline feed efficiency scenario was very close to 0, as desired, whereas the High and Very High Efficiency scenarios had average simulated RFIs of 1.4 and 2.71 kg DM/cow/day, which correspond to a ratio of the simulated DMI:expected DMI of 0.94 and 0.88, respectively. By tracking the simulated intake and comparing it with the expected intake, we confirm that we achieved the desired effect on herd level feed efficiency through the alteration of individual animal outcomes.The expected impact of feed efficiency on herd production, intake, enteric methane, and manure production is shown in Figure 5. Figure 5a shows that each scenario achieved an equivalent milk production on every day of the simulation, as expected. Panel (a) also highlights oscillations in milk production throughout the year-long simulation, as the total number of lactating animals oscillates with individual animal reproduction and lactation events. We highlight some of the simulated manure characteristics in Figure 5b, and the impact of each feed efficiency scenario on manure outcomes is demonstrated. The total pen DMI shown in Figure 5c represents the daily sum of intake of a ration that is formulated at the user-defined ration interval of 3 days. The composition of the ration at each interval will vary in response to the varying nutrient requirements of the herd at the start of each ration formulation interval. The average and SD of the inclusion rate of each feed are the same for each scenario because the feed-efficiency parameter (ρ) was designed to create a distribution in intake responses to the same formulated ration. Thus, the ration formulation stays the same but the simulated feed consumed is altered. The average and SD of the inclusion rates for each feed are given in Table 4. When the animal module is integrated with the rest of the RuFaS simulation platform, the associated GHG emissions and environmental footprint of producing each of these feeds will be included in the model outputs and separated by farm-grown and purchased feeds.We display the daily simulated outputs for enteric methane, manure N, milk production efficiency, and milk:manure ratio in box-plots in Figure 6, to facilitate the comparison of the expected distribution of outcomes from each feed efficiency scenario. The variation in outcomes shown in Figure 6 is a reflection of the multiple Monte-Carlo processes used in the Animal Module and the day-to-day variability in the combined number of animals in each stage of lactation. These distributions of responses are more representative of the expected outcomes in a commercial farm setting and provide a better basis for inference and comparison than deterministic mean estimates.Although the RuFaS Animal Module is not yet integrated with the process-based manure management simulation, we provide an empirical estimate of the carbon footprint of the predicted enteric methane, manure volatile solids, and manure nitrogen using the methods described by IPCC [41]. For manure-based emissions, we used the methods for estimating GHG production from an anaerobic manure lagoon in North America. We chose to use a 100-yr CO2-eq of 30 for CH4 and a CO2-eq of 298 for N2O for the purpose of illustrating the comparative impacts [41]. Based on these estimates, the expected decrease in manure excretion from a herd with lower average RFI will result in an even larger reduction in CO2-eq than the expected benefit from reduced enteric methane (Table 5).4. DiscussionThe Animal Module of RuFaS offers a comprehensive representation of individual dairy cow life events, feeding, milk production, and manure production, as well as the application of a wide variety of herd management options to simulate dynamic, holistic outcomes from the animal part of a dairy farm system. These holistic outcomes and their expected response to different management scenarios are essential information for scientists and industry members to guide sustainable development in the dairy sector. Although the static, deterministic estimates of environmental outcomes offered by LCAs [13,14] are useful signposts, the integration of scientific knowledge through dynamic, stochastic process models provides a tool for comparison across systems, practices, and farm inputs that national averages lack. As a tool for estimating farm-level outcomes, it is essential that RuFaS be both capable of representing commonly used management practices and adaptable to new technologies and practices as they develop. For this reason, we built the RuFaS Animal Module to be flexible in how it represents herd dynamics and included the ability to alter both animal and herd level outcomes in response to simulated changes in management or biology. For example, the representation of reproductive protocols that can be applied separately to heifers and cows enables the model to probe questions about how reproductive efficiency and herd dynamics influence expected environmental outcomes. Although the ration formulation is based on a least-cost optimization, the user can define any number of feeds to be included and set inclusion limits to direct the algorithm’s ration formulation. One more example of built-in flexibility is the housing structure in which the user can define the type of housing, which is similar to other models, but also allows the user to separate the animals into any number of pens and define pen-specific distances to the milking parlor, bedding, and manure management characteristics. All of these model attributes combine to provide a flexible dairy animal and herd simulation model that is more representative of, and adaptable to, the management practices in use today than many existing farm models. Among some of the most comprehensive whole-farm models for US systems (e.g., IFSM [7], COMET-Farm, and Manure-DNDC), RuFaS is the only one that simulates individual animals, and this is a core feature from which much of the Animal Module’s flexibility stems.The feed efficiency case study highlights the utility of simulating individual animals instead of groups or classes of animals and the adaptability of the model to target specific study questions, which is another key feature of the RuFaS model. A single stochastic parameter to represent the observed animal variation in RFI, and thus the feed efficiency, was easily added to the model code base and facilitated the comparison of multiple outcomes of interest with a single input value. This case study was selected because improving feed efficiency through the use of metrics like RFI has gained increased attention in recent years in the scientific literature [20,40,42] and in industry breeding programs [43,44]. The importance of evaluating the expected outcomes of RFI breeding programs at multiple points in the dairy farm system is highlighted in the model outputs listed in Table 5. The ratio of feed intake between the improved efficiency and baseline scenarios matches the mean input values for the feed efficiency parameter ρ, as intended, but the methane and manure estimates each have their own distribution, that is not centered at the mean of the feed efficiency parameters, as might be expected. This demonstrates that the whole-farm GHG emission reductions from a herd that reduced the feed intake by 6% cannot be assumed to be 6% of each outcome of interest. To estimate the combined impact on the CO2-eq from enteric methane and manure production, we summed the CO2-eq from Table 4 and took the ratio of the improved feed efficiency scenarios to the baseline scenario. We found that the High efficiency scenario reduced the total CO2-eq by 4.6% and the Very High efficiency scenario reduced the total CO2-eq by 9.3%. This preliminary assessment of the impact of the improved feed efficiency on enteric and manure GHG emissions can be used to support and set targets for breeding programs. In future work, we plan to add the representation of genetic inheritance between dams and calves to simulate not just the target of breeding programs but the speed with which different breeding and management programs will be expected to get there. The cumulative outcomes and environmental benefits of breeding for improved efficiency over the years can then be estimated.In addition to adding genetic inheritance, future work to fully integrate the Animal Module with modules that represent manure processing and storage, field management, and crop harvest and storage will complete the dairy farm nutrient cycle to allow for the dynamic feedback of animal management on the whole farm and enable process-based estimates of environmental outcomes both down and upstream from the dairy barn. Upon completing version 1 of the whole-farm model, we will publish the code and underlying documentation in an open-source repository, so that scientists and industry professionals may engage with the model directly. The fully integrated model will have many applications in science to probe questions about the outcomes of single and combinations of management changes. The RuFaS model will also have applications in industry as a tool to estimate whole-farm environmental impacts, and expected changes to the adoption of novel practices, so that producers can meet market and societal demands for improved sustainability.5. ConclusionsWe have presented a description of the Animal Module of the Ruminant Farm Systems modeling platform, described its structure and advantages over other existing models, and used the example of feed efficiency to illustrate some of its functionalities and applications. The results of the case study demonstrate the milk, manure, methane, and herd demographic outputs available from the RuFaS Animal Module and provide preliminary estimates of the multiple environmental benefits of improving feed efficiency by reducing RFI. The use of the Animal Module within the larger RuFaS modeling platform will enable an even more comprehensive assessment of how improvements in management and biological performance will affect the environmental impacts of dairy production.

animals : an open access journal from mdpi

10.3390/ani13050795

PMC10000050

Weaning (that is, removal from the sow) and the following two months are the riskiest periods in a pig’s life, especially for pig’s gastrointestinal health. The change in diet due to the suspension of the mother’s milk, accompanied by an acceleration of both morphological and enzymatic maturation of the intestinal mucosa of the piglets, can worsen digestion and absorption. In this context, the protein requirement of piglets, which are in a phase of rapid growth, may be greater than the intestine’s ability to digest proteins. Undigested proteins are the best pabulum for the proliferation of the pathogenic bacterial flora that causes diarrhea. Since these problems can no longer be resolved with prophylactic use of antibiotics, the best balance between intestinal health and growth performance must be found. A diet low in crude protein and supplemented with synthetic amino acids can help achieve this goal.

The aim of this research has been to assess the effect of the dietary protein level on piglet growth and post-weaning diarrhea (PWD) incidence. Piglet fecal microbiota and feces composition were also assessed. The experiment was carried out on 144 weaned piglets (Duroc × Large White; 72 piglets per treatment) and lasted from weaning (at 25 days of age) until the end of the post-weaning phase (at 95 days). Two dietary protein levels were compared: high (HP; 17.5% crude protein on average, during the experiment) and low (LP; 15.5% on average). Lower (p < 0.01) average daily gain and feed conversion ratio were observed in LP piglets in the first growth phase. However, at the end of the post-weaning period, the growth parameters were not significantly different in the two diets. Diarrhea scores were lower in piglets fed LP diets than in piglets fed HP diets (28.6% of the total vs. 71.4% in the HP piglets). Fibrobacteres, Proteobacteria, and Spirochaetes were more abundant in the feces of the piglets fed LP diets. Feces nitrogen content was lower in piglets fed LP diets. In conclusion, low protein levels in the diet can reduce the incidence of PWD while only marginally affecting growth parameters.

1. IntroductionThe period between weaning (i.e., removal from the mother) and reaching the bodyweight for transfer to the fattening boxes (at about three months of age and 35–40 kg of body weight) is more delicate from the point of view of health and functioning of the piglet digestive system. Among the numerous factors that can intervene to destabilize the delicate balance of the piglet’s intestine, linked to an acceleration phase of both morphological and enzymatic maturation, diet-linked factors undoubtedly play a fundamental role [1].Among the dietary factors, the quantity of protein fed to piglets plays a leading role. In fact, in this phase, the coverage of nutritional needs requires a protein level higher than the digestive potential of the piglet. According to Kim et al. [2], a protein level of between 21.5% and 24% is required for modern fast-growing lines, a level that, in fact, is higher than the piglet’s digestive capacity; these authors suggest a protein level not higher than 18% in the first days after weaning and with a consistent addition of synthetic amino acids. De Lange et al. [3] pointed out that low protein levels are beneficial for the gut health of piglets because the presence of undigested proteins, as it can occur with high dietary protein levels, allows the proliferation of a bacterial flora producing toxins capable of altering the intestinal barrier. This alteration implies: the colonization of the intestinal epithelium by pathogenic microorganisms; the acceleration of the production of enterocytes that, being immature, have an exudative rather than absorbent attitude; greater ease of crossing the cellular barrier by specific bacterial toxins (edema disease). Zhang et al. [4] highlighted that high-protein diets increase the microbial fermentation of proteins, peptides, and amino acids. According to Gao et al. [1], high protein levels favor the production of ammonia and branched-chain-fatty acids and, therefore, the proliferation of pathogenic bacterial flora, while low protein levels favor the production of short-chain fatty acids (SCFAs), primarily butyric acid, which favors the proliferation of beneficial bacterial flora. The same authors, comparing two protein levels (17% vs. 30%), both obtained exclusively with casein, highlighted that with a high protein level, the bacterial diversity of the microbiota is reduced. The increase in ammonia can negatively affect the formation of intestinal epithelial cells [5]. The reduction of beneficial Lactobacilli that accompanies the maturation of the pig’s intestine and the variations in the buffering effect of pH due to protein fermentation can make the intestinal environment more susceptible to the emergence of opportunistic pathogens, such as Bacteroides and Clostridium species [6]. Opapeju et al. [7] compared four diets administered to piglets with an initial weight of about 6.5 kg: control feed with 21% crude protein (CP); feed with 19% CP and deficient in isoleucine; 19% raw protein feed supplemented with synthetic isoleucine to reach the isoleucine level of the control feed; 17% raw protein feed supplemented with isoleucine and valine to reach the ratio indicated by the ideal protein. The control group showed better production performance in terms of growth and conversion index, although they showed softer feces, a greater amount of ammonia in the feces, and a greater depth of the crypts of the intestinal mucosa, indicating an acceleration of production of enterocytes; essentially, better production performance, but greater susceptibility to the onset of a syndrome affecting the gastrointestinal system.This situation of precarious equilibrium could be easily kept under control only with targeted antibiotic prophylaxis, which, however, is no longer allowed, and even the spaces for metaphylaxis, albeit careful, become increasingly restricted. Therefore, it is necessary to identify feeding strategies that reduce the risk of the appearance of alterations in the gastrointestinal function of the pig [8] and reserve the use of antibiotics for clinically overt pathological situations.Knowledge of the relationship between pig microbiota and diet can be used to orient the intestinal microbial dynamics in the desired direction by diet manipulation [9]. The microbiota of healthy piglets susceptible to post-weaning diarrhea (PWD) has been the subject of numerous investigations [10,11], which highlighted how the state of health of the pig and its susceptibility to diseases, such as PWD, can be related to the change in the composition of the microbiota during the early stages of growth.All factors affecting PWD susceptibility also affect microbiota composition. Among these, pig feeding plays a primary role. Heo et al. [12] observed a reduction in PWD in piglets challenged with an enterotoxigenic strain of Escherichia coli, when fed with lower protein levels. Rist et al. [13] suggested reducing proteins and increasing fermentable carbohydrates in the diet to reduce harmful protein fermentations. Luise et al. [14] showed that lower dietary protein levels could reduce the intestinal fermentation of undigested proteins and the consequent risk of diarrhea. The aim of this research has been to assess the effect of the dietary protein level on piglet growth and PWD incidence. Piglet fecal microbiota and feces composition were also assessed to support the understanding of the results.2. Materials and Methods2.1. Animal EthicsAll animal procedures were performed in strict accordance with the Code of Ethics of the World Medical Association (https://ec.europa.eu/environment/chemicals/lab_animals/legislation_en.htm. Accessed on 19 February 2023).2.2. Experimental DesignThe experiment was carried out at our experimental pig farm in San Cesario sul Panaro (Modena, Italy), on Duroc Italiana × Large White Italiana crossing lines. A total of 144 weaned piglets was used, half barrows and half females. The experiment lasted from weaning, when the age of the piglets was 25 ± 1.5 days, until the end of the post-weaning phase, corresponding to 70 days of trial and 95 days of age of the animals (Table 1).Two dietary protein levels were compared (Table 2): high (HP) and low (LP), in two growth phases characterized by different diet compositions. The diet composition (Table 3) was changed 25 days after the start of the experiment, at the expected piglet body weight of 15 kg (and actual weight of 17 kg). Synthetic amino acids were also supplemented to ensure a balanced feed formulation. Specifically, total lysine levels in the first and second feeding periods were set at 1.40% and 1.20% of the feed, respectively. The percentages of methionine, cystine, threonine, and tryptophan were balanced according to the proportion of the ideal protein [15] with the addition of synthetic amino acids.After separation from the mother, the piglets were housed in cages of 12 individuals each, distributed as evenly as possible within each cage by body weight, age, and litter of origin. Since it was not possible to accommodate 12 cages (3.3 m2 each) within the same room, the males were housed in one room (6 cages) and the females in another (6 cages). In this way, it was necessary to accept that we were dealing with a confused effect (room and sex), while the factor of interest of the experiment (protein level) was homogeneously represented in both rooms. Each cage was equipped with a hard-plastic floor. The complete feed was administered ad libitum in a hopper feeder with 4 places to ensure sufficient access to feed for all piglets; in each cage, there was a nipple drinker. The temperature of the air in the rooms was 23 °C ± 1 °C, whereas the humidity was not controlled.Forty-six days after weaning, the piglets were moved to larger pens in the fattening area and housed in 12 pens (9 m2 each) of 12 piglets each. The floor was thermally insulated concrete. The pens were arranged in 2 rows of 6 pens (3 adjacent pens for males and 3 adjacent pens for females), one row for each protein level. The animals were given 1 kg of feed per animal and per day in two daily meals. In this phase, the administration of feed was limited to reduce the risk of diarrhea related to the stress of the change of housing. In the first two days, the meal was dry and distributed on the ground, gradually passing to a wet meal in the trough over the next three days. In each pen, the water was still available through a nipple drinker. The assignment of the treatment to the pens followed the criterion of minimizing the possibility of mixing feces from pigs fed different protein levels. In the fattening area, 21 °C was always ensured during the experiment.The effects of dietary protein level on piglet growth, health status, and feces microbiota were considered for the following periods (Table 1):-Period I: From the start of the experiment until the change of feed;-Period II: From the change of feed until the change of housing;-Period III: From the change of housing to the end of the post-weaning period (end of the experiment).2.3. Growth ParametersThe animals were weighed individually at the start of the experiment, at the day of diet change, at the day of change in housing, and at the end of the experiment. For each period, average daily gain (ADG = [body weight at the end of the period—body weight at the beginning of the period]/day), average daily feed intake (ADFI = feed consumption in the period/day), and feed conversion rate (FCR = ADFI/ADG) were calculated. Since feed consumption was known only at the cage/pen level, the values of the ADFI and FCR variables were not known at the individual level and thus were calculated at the cage/pen level. For homogeneity, the ADG values, although individually known, were also processed at the cage/pen level, with six replicates in total for each treatment.2.4. Diarrhea Scores and Corrective InterventionsThe health of the piglets was monitored daily. A score was assigned to each cage/pen based on the number of cases and the extent (mild, medium, severe) of diarrheal phenomena in piglets, visually assessed from the consistency of the feces (Table 4). For each growth period, individual diarrhea scores were summed for each treatment and related to the period’s total diarrhea score.Piglets suffering from diarrhea before the change of diet (Table S1) were treated (Table S2) parenterally with the antibiotics enrofloxacin or marbofloxacin. No antibiotics were administered between the time of the diet change and the date of transfer to the pens because there were no cases of diarrhea. After the transfer, all individuals were treated orally with colistin sulfate, starting 5 days after the transfer and for 8 days. It was decided to resort to oral mass therapy because the appearance of overt diarrhea (score 4 or higher) occurred five days after the moving, in four out of six boxes of the HP treatment and in one out of six boxes of the LP treatment (Table S3). No treatment was applied in the last 12 days before the third and final sampling of feces.2.5. Feces Chemical CharacterizationFeces samples were collected at the end of each growth period by piglet rectal ampoule stimulation. For each cage/pen, individual samples were pooled into a composite sample. Samples were immediately frozen and stored at −20 °C until analysis.Dry matter, organic matter, total (Kjeldahl) nitrogen and ammonium nitrogen contents, and pH, were determined according to the APHA methods [16]. Crude fiber (CF) and fiber fractions were determined on fecal samples dried at 60 °C. Crude fiber was determined according to [17]. Hemicellulose and cellulose concentrations were estimated by determining neutral-detergent (NDF) and acid-detergent (ADF) fiber fractions and acid-detergent lignin (ADL) according to [18]. The difference between NDF and ADF is an estimate of the hemicellulose content; that between ADF and ADL of the cellulose content.For volatile fatty acid determination, 1 g of the sample was diluted with 3 mL distilled water, then centrifuged at 4000 rpm for 15 min. The supernatant was used for the analyses. Half mL of sample supernatant was added to 0.25 mL 4% H3PO4 and 0.25 mL internal standard to a final volume of 1 mL. A microliter of this mixture was injected in the injection port of the gas-chromatograph (Shimadzu GC 2010 Pro), equipped with a NukolTM capillary column (Supelco, cat. no. 24107), 30 m × 0.25 mm internal diameter, 0.25 μm film thickness. Total volatile fatty acid content (mg L−1) was calculated as the sum of the individual concentrations of acetic, propionic, butyric and iso-butyric, valeric and iso-valeric, and caproic and iso-caproic acids.2.6. DNA Extraction, Library Construction, and SequencingTotal DNA was extracted from the fecal samples after thawing using the QIAamp PowerFecal Pro DNA Kit (QIAGEN, The Netherlands) according to the manufacturer’s instructions. The V3-V4 regions of the 16S rRNA gene were amplified using the 341F (5′-CCTACGGGNGGCWGCAG-3′) and 805Rmod (5′-GACTACNVGGGTWTCTAATCC-3′) (based on [19], with degenerate bases) primers.Library construction and sequencing were performed at the Sequencing Platform, Fondazione Edmund Mach, Italy. More in detail: each sample was amplified by PCR using a 25 µL reaction mixture with 1 µM of each primer. More in detail, 12.5 µL of 2× KAPA HiFi HotStart ReadyMix and 10 µL forward and reverse primers, were used in combination with 2.5 µL of template DNA (5 ng/µL). The PCR reactions were carried out by GeneAmp PCR System 9700 (Thermo Fisher Scientific) and the following cycling conditions: initial denaturation step at 95 °C for 3 min (one cycle); 25 cycles at 95 °C for 30 s, 55 °C for 30 s, 72 °C for 30 s; final extension step at 72 °C for 5 min (1 cycle).The amplification products were checked on 1.5% agarose gel and purified using the CleanNGS beads (CleanNA, Waddinxveen, The Netherlands) following the manufacturer’s instructions. Afterward, a second PCR was used to apply dual indices and Illumina sequencing adapters Nextera XT Index Primer (Illumina), by 7 cycles PCR (16S Metagenomic Sequencing Library Preparation, Illumina). The amplicon libraries were purified using the CleanNGS beads (CleanNA, The Netherlands), and the quality control was performed on a Typestation 2200 platform (Agilent Technologies, Santa Clara, CA, USA). Finally, all barcoded libraries were pooled in an equimolar way and sequenced on an Illumina® MiSeq (PE300) platform (MiSeq Control Software 2.5.0.5 and Real-Time Analysis software 1.18.54.0). A total of 3.093.548 raw reads were detected across the samples by the Illumina MiSeq sequencing platform (PE300) (Illumina, Santa Monica, CA, USA).2.7. Statistical AnalysisThe estimate of the effect of the protein level in the diet on growth performance was carried out by means of a one-factor analysis of variance. The effect of the sex/room factor was included in the block effect. Values of F with p > 0.05 were considered not significant (NS).The analysis of variance (ANOVA) for the effect of diet on diarrhea incidence was applied to the sum of the scores of the period for each treatment replication. Since in the second period the sum of the scores was very low, for the purposes of ANOVA, the first two periods were merged into a single period.Two-way ANOVA (fixed sources of variation: time of sampling, protein level, time of sampling × protein level) was applied to operational taxonomic units (OTU) percentages to estimate the effect of diet and growth period on the relative abundance of phyla, families, and genera in the analyzed samples, using the procedure MIXED, SAS language [20]. A threshold equal to 0.1% of the total reads was adopted to include the phylum, family, or genus in the analysis. Mean multiple comparisons were performed using the statement LSMEANS and the Tukey HSD test.The same two-way ANOVA design was applied to the statistical analysis of feces composition.2.8. Bioinformatic AnalysesData were pre-processed using the MICCA v. 1.7 [21] pipeline and rarefied to an equal depth of 45.225 reads per sample. OTUs were created de novo by clustering sequences with 97% sequence identity and classified using the RDP [22] software version 2.11.The alpha diversity of the populations, the relative abundance (%) of microbial components down to the family and sex level, and their grouping based on the sources of variation were estimated. Alpha diversity is an index of the richness (number) and diversity (relative abundance) of OTUs in a population. The richness of species is indicated by the total number of OTUs (“Observed”) in the microbial community: the higher the number, the more species are present. The CHAO1 index estimates the richness of species, giving more weight to the less abundant ones. The value of CHAO1 is at least equal to “Observed” and increases as the number of rarer species increases. It can be calculated as follows:CHAO1 = Sobs + F1(F1 − 1)/(2 × (F2 + 1))(1) where Sobs is the number of observed species and F1 and F2 are the count of singletons and doubletons, respectively.The Shannon index is calculated as follows:Shannon Index = −∑(pi ln(pi))(2) where: Σ is the summation from 1 to the total number of OTUs, and pi is the proportion of the community represented by the OTU i. It increases with increasing species richness, uniformity, and uncertainty of the estimate.The samples were grouped for compositional similarity (beta diversity) using Principal Coordinate Analysis (PcoA), which is a multivariate method of data analysis used to explore and to visualize similarities or dissimilarities of data [23].3. Results3.1. Growth PerformanceThe piglets fed HP level showed a greater ADG (p < 0.01) than those fed LP (Table 5) in the first 25 days of the experiment until the change of diet, while the increases were the same for the two protein levels in the period from the change of diet to the change of housing. The difference remained significant (p < 0.05) during the entire phase in the cages. The FCR value was also significantly better in HP. Conversely, ADFI was not significantly different in piglets fed different protein levels. In the pen housing, no differences were detected between the two treatments as the feed was rationed. The body weight of the piglets at the end of the post-weaning period was not different in the two diets.3.2. Health StatusDiarrhea cases started to occur in the early post-weaning period (Table S1). They were concentrated in the period between weaning and change of feed (Period I; score summation: 107 out of 234, i.e., 45.7% of the total; Table 6) and in the period between change of housing and the conclusion of the post-weaning period (Period III; 121 out of 234: 51.7% of the total). In Period II, overall, feces had normal consistency, as well as in the first and last part of Period III (Tables S1 and S3).The overall incidence of diarrhea in the LP treatment was 28.6% of the total (67 out of 234 scores) and 71.4% in the HP treatment. Specifically, from weaning until feed change (Period I), the diarrhea scores in the LP treatment were 15.0% of the total (16 out of 107); after feed change (Period II), they were 16.7% of the total (1 out of 6), while after the change of housing (Period III) they rose to 41.3% (50 out of 121). The effect of the dietary protein level on diarrhea score summations was highly significant (p < 0.01) until change of housing (Period I + Period II), whereas it was not significant (p = 0.140) from the change of housing until the end of the experiment (Period III).Poor feces consistency was noted in all the boxes in the HP treatment, starting from the twelfth day of the experiment (Table S1). The phenomenon initially extended to the whole pen in three out of six pens and then progressively became reduced, presumably due to therapeutic interventions (Table S2), until it disappeared two days after the change of feed.3.3. Composition of the Fecal MicrobiotaThe alfa diversity of the microbiota (intra-sample diversity) increased after the change of feed and further increased after the change of housing (Figure 1a), while it was not influenced by the protein level in the diet (Figure 1b).The analysis of beta diversity (inter-sample diversity) using the PCoA tool allowed the clustering of the fecal samples into three groups, corresponding to the three times of sampling (Figure 2). The protein level in the diet allowed only partial separation of the treatments.The most represented phyla in the piglet fecal samples (Table 7) were Bacteroidetes and Firmicutes, which together accounted for 90.6% of reads. The relative abundance of phyla changed in relation to the sampling date: at the first sampling, 25 days after weaning, Firmicutes and Actinobacteria were more abundant than in the two subsequent samplings, whereas the other classified phyla: Bacteroidetes, Spirochaetes, Proteobacteria, and Fibrobacteres were more abundant in the two samplings following the first, without significant differences between the second sampling, at the end of a period characterized by a change of feed, and the third sampling, at the end of a period that had begun with the change of housing. Fibrobacteres, Proteobacteria, and Spirocheaetes were, on average, more represented in the feces of piglets on a low-protein diet, whereas Firmicutes were more abundant in the feces of piglets on a high-protein diet.Overall, the most represented bacterial families were: Prevotellaceae, Lachnospiraceae, Ruminococcaceae, and Porphyromonadaceae (65.1% of the total reads). While the influence of the sampling time was evident, the same cannot be said for that of the treatment (Table S4). Low protein levels in the diet were associated with a higher abundance of Fibrobacteraceae, Succinivibrionaceae, Sutterellaceae, and Spirochetaceae, whereas Firmicutes-belonging families prevailed in the HP fecal samples: Lachnospiraceae and Eubacteriaceae were more abundant along all the post-weaning period, whereas Erisipelotrichaceae, Clostridiaceae 1, and Peptostreptococcaceae were more abundant only in the first sampling event. Lactobacillaceae were the only Firmicutes more abundant in the feces of piglets on a low-protein diet and only in the first sampling event (significant interaction effect: Sampling time × Protein level). A few classified families were little or not at all influenced by both the period of growth and the protein level; among these, Enterobacteriaceae.Among the classified genera, the most abundant were Prevotella (27.2%), followed by Clostridium sensu strictu (3.9%), Lactobacillus (3.3%), Alloprevotella (3.2%), Treponema (3.1%) (Table S5). Selected genera were more represented in the families significantly influenced by the dietary protein level, alone or in interaction with the sampling time (Table 8). Lactobacillus and Treponema were more represented in the LP fecal microbiota, whereas Roseburia, Blautia, and some Clostridium genera were more abundant in HP.3.4. Feces CompositionAlmost all the traits analyzed varied significantly over time (Table 9). More specifically, the concentration of dry matter, organic matter, crude fiber, and fiber fractions increased over time, whereas total and ammonium nitrogen concentrations decreased. The pH slightly increased after the feed change. The dietary protein level also affected the feces composition. Piglets fed low protein had feces that were richer in dry and organic matter and lower in total and ammonium nitrogen all over the experiment period. Even though crude fiber concentration was not affected by the dietary protein level, however, the fiber fractions were. In fact, hemicellulose and cellulose concentrations were higher in the LP fecal samples.As for volatile fatty acids, the concentrations of propionic, valeric, and especially isovaleric acid slightly increased after the change of feeding. No effect on the protein level was detected, except for the isovaleric acid concentration, which tended to be lower in low-protein diets.4. Discussion4.1. Dietary Protein Level, Growth Performance, and Susceptibility to PWDIn our experiment, the low protein level negatively affected the growth of the piglets (ADG and FCR) only in the first post-weaning period (up to the change of diet), whereas it had a clear positive impact on diarrhea intensity reduction. Literature on the matter reports contrasting results, as also highlighted by Wang et al. [24]. Heo et al. [12] reported a higher incidence of PWD when feeding weaned piglets at a high (24.3%) compared with a low (17.3%) CP diet. Wen et al. [25] observed a higher incidence of PWD for piglets fed higher protein levels (up to 23%). Rattigan et al. [26] obtained contrasting results by working in sanitary vs. unsanitary conditions. Reducing the dietary CP level did not affect growth performance; however, in sanitary conditions, it increased the Enterobacteriaceae abundance in the colon and the incidence of diarrhea occurrence, whereas the opposite occurred in unsanitary conditions. Limbach [27] tested the effect on growth and PWD incidence in piglets fed soybean-maize diets at three protein levels (22% AA-balanced, 19% AA-balanced, and 16% protein AA-unbalanced) and concluded that low CP diets may be used for the initial post-weaning period to reduce piglet susceptibility to PWD without largely impacting growth performance. Lynegaard et al. [28] overall reduced CP in the first weaning phase (6–9 kg) from 19.1% to 16.6 and 14%, and from 18.4% to 16.2–17.4%, in the second phase (9–15 kg), depending on the treatment, while they left the protein almost constant in the third phase (15–30 kg) of growth. Under these conditions, they observed a decrease in PWD incidence for treatments with reduced dietary protein, as well as lower ADG and FCR values in piglets fed lower protein levels up to 15 kg body weight. The disadvantage, especially in terms of ADG, also remained in the following period. In their experiment, however, the reduction of CP in the early post-weaning growth period was more pronounced than in our experiment. According to [29], CP levels too low (i.e., 3% below average: 17%) are considered harmful in consideration of maladaptive changes to small intestinal morphology and pepsin activity in weaned piglets.4.2. Piglet Fecal Microbiota in the Post-Weaning PeriodThe observed increase in the microbiota’s alfa diversity during piglet growth has also been reported by other authors [30], and it can be linked to changes in diet as well as to the progressive maturation of the intestinal system [31]. On the contrary, dietary protein levels have been reported not to influence alpha diversity in pigs [32,33]. Our results confirm these findings.The most represented phyla (Table 7) are Bacteroidetes, Firmicutes, Spirochaetes, and Proteobacteria (96.9% relative abundance), and the most abundant bacterial families: Prevotellaceae, Lachnospiraceae, Ruminococcaceae, and Porphyromonadaceae in the piglet fecal samples of this experiment are those recurring in the feces of healthy piglets [34,35]. Clostridium sensu stricto, Roseburia, Paraprevotella, Clostridium XIVa, and Blautia have been reported as major representative genera after weaning [36,37].In general, the results in the literature refer to piglets stressed on purpose or younger than those in this experiment. The considered microbiota is more often that of the intestinal system, different from the fecal one [38]. In our experiment, we considered the microbiota of the feces of piglets raised in protected conditions, which should be the most likely or, in any case, desired in real farms. This protection has been applied both at the environmental level (attention to the absence of causes of stress) and at the health level (preventive interventions with drugs). This may explain the low presence of Enterobacteriaceae, which is normally associated with stressful conditions.4.3. Dietary Protein Level, Composition of the Fecal Microbiota, and PWD SusceptibilityThe effect of dietary protein level on the composition of gut microbiota has mainly been studied for finishing pigs. Moderate protein levels in the diet have been found to modify the gut microbiota composition and to improve the ileal barrier function [39,40]. On the contrary, information on piglets is scarce, especially when referring to the composition of the fecal microbiota. In our experiment, differences in the composition of the fecal microbiota due to the protein level in the diet were found above all in the first sampling, that is, at the end of the first period after weaning and before the change of feed, just when the effect of protein intake on the incidence of diarrhea phenomena was most evident. Given this coincidence, we can think of associating the higher susceptibility of the piglets to PWD with the greater presence of Firmicutes-belonging families (with the exclusion of Lactobacillaceae), which were more abundant in T1 in the fecal microbiota of piglets fed HP diets. Conversely, a lower incidence of PWD can be associated with the prevailing abundance in LP diets of families belonging to Lactobacillaceae (gen. Lactobacillus), Fibrobacteraceae (gen. Fibrobacter), Succinivibrionaceae, and Spirochetaceae (gen. Treponema).Yang et al. [30] compared the fecal microbiota of healthy piglets and that of diarrheal piglets in three stages of growth: lactation, intermediate stage, and weaning (solid diet for piglets) and noted that, with the transition to solid feeding, the incidence of Lactobacillus and E. coli decreased while that of Prevotella increased. They related reduced numbers of Bacteroides, Ruminococcus, Bulleidia, and Treponema, which are responsible for the digestion of solid foods, to the onset of post-weaning piglet diarrhea.The genera prevalent in the fecal microbiota of piglets fed HP diets, such as Roseburia, Blautia, Eubacterium, and selected Clostridium species, are commonly found in piglet fecal microbiota and are all fermentative. Their greater abundance in the HP feces can be the consequence of the incomplete digestion of complex fermentable substrates because less digestible feed components are more available for microbial activities in the gut terminal tract. The greater presence of Clostridia in the microbiota of piglets fed HP diets can be related to their ability to metabolize amino acids [4]. Conversely, the increased presence of Fibrobacter in low-protein diets can be related to the availability of higher amounts of undigested cellulose in the final intestinal tract. In fact, Fibrobacter is a fibrolytic bacterial species commonly present in the pig intestinal microbiota [41], although it is a typical colonizer of the rumen. Fibrobacter uses glucose, cellulose, and cellobiose as carbon and energy sources with the production of succinate, acetate, and formate [42]. Fibrobacter and Treponema have been reported as more abundant in diets of ruminants fed with higher amounts of lignocellulosic components [43].4.4. Dietary Protein Level and Feces CompositionFeces composition clearly depends on the type of food ingested and on the use that the animal can make of the food. The less digestible the food, and the worse the digestive abilities of the animal, the higher the quantity of feed that remains undigested in the feces. Higher crude fiber contents accompanying the change of feed can be the reason for worse digestibility: a higher crude fiber content was, in fact, found in the piglet feces at the second and third sampling (i.e., at the change of housing and at the end of the post-weaning period).An important side effect of the reduction in dietary protein was the lower concentration of nitrogen in the feces, possibly because a greater quantity of ingested nitrogen was assimilated instead of wasted. This result has been reported by several authors. Zhao et al. [44] found a significant reduction in N excretion in 90-day-old pigs fed protein diets 3.5% lower than standard. Yang et al. [45] observed a linear decrease in fecal total nitrogen for decreasing dietary protein levels in the diet of growing pigs.Other components in the diet may interfere with nitrogen metabolism. In fact, in maize, the nitrogen-free extractives consist of starch, whereas in soybeans, the starch is less than 1%, and soluble carbohydrates are mainly sucrose, raffinose, and stachyose [46]. High amounts of raffinose and stachyose from soybean in the diet are supposed to reduce the digestibility of nitrogen and amino acids in growing pigs [47]. Therefore, in HP-fed piglets, the greater nitrogen excretion can also derive from a lower digestibility due to the higher percentages of soybean oligosaccharides. Interestingly, Zhang et al. [48] showed a higher incidence of diarrhea cases in piglets fed with soybean flour as is or with added stachyose compared to the control consisting of corn with concentrated soybean protein.Short-chain fatty acids are final metabolites of the intestinal microbiota, produced mainly in the large intestine, where they are used by mono-gastrics as a source of energy. The energy contribution of the SCFA in pig metabolism is important. Weaning has been reported to affect the concentrations of SCFA in the intestine [49]. In our experiment, the dietary protein level did not affect the SCFA content, apart from the case of isovaleric acid, which was less abundant in LP diets.Some confirmations and some interesting insights emerge from this experience. The first of these is whether the body weight differences found in the experiment cancel out during the subsequent stages of growth of the animals and, once this has been established, how far one can go with the reduction of the protein content in particularly critical phases (first 15 days after weaning, sudden changes in the environment) without compromising the subsequent productive career. The second interesting point is to investigate the evolution of the microbiota to identify moments in which an analysis of this can serve as an indicator of the evolution of the state of health of the gastrointestinal system. From our results, it appears that the phase immediately following weaning is the most subject to changes in the fecal microbiota composition and most suitable for earlier identification of possible stress conditions. Everything must be understood in terms of optimizing individual and mass therapeutic interventions in order both to improve the profitability of breeding and to reduce the risk of the appearance of antibiotic-resistance phenomena.5. ConclusionsThis research confirms that a reduction in the protein content of feed can reduce the appearance and severity of gastrointestinal syndromes in piglets in particularly stressful stages of rearing (removal from the mother, change of housing) while only marginally affecting growth performance. Low-protein diets, resulting in excreta with lower quantities of nitrogen than those of standard diets, may allow for potential environmental benefits.The variations in the microbiota are largely determined by the growth phase, which in turn is accompanied by an evolution of the diet. The different protein levels at the same age caused slight but significant variations in some components of the intestinal microbiota. Further research is needed to ascertain whether microorganisms found in low-protein diets can be considered indicators of lower susceptibility to diarrhea in piglets.

animals : an open access journal from mdpi

10.3390/ani13061076

PMC10044542

Skeletal muscle satellite cells (SMSCs) serve as the source of myogenic cells and can afford to differentiate into myotubes as well as act as a model for exploring myogenesis in vitro. In this study, the transcriptional profile of ovine skeletal muscle satellite cells was constructed via the RNA-Seq method. A total of 1954 DEGs, 1479 AS, and 253 TFs were detected during the proliferation and differentiation of SMSCs. GO and KEGG analyses showed that the MAPK signaling pathway, PI3K-Akt signaling pathway, Wnt signaling pathway, and Ras signaling pathway were enriched. Together, our study provides novel insights into the transcription regulation of SMSCs during proliferation and differentiation at the transcriptional level, and provides a valuable resource for understanding the molecular mechanism of myogenesis and muscle development.

Skeletal muscle satellite cells (SMSCs), which are highly multifunctional muscle-derived stem cells, play an essential role in myogenesis and regeneration. Here, the transcriptional profile of SMSCs during proliferation and differentiation were constructed using the RNA-Seq method. A total of 1954 differentially expressed genes (DEGs) and 1092 differentially alternative splicing genes (DAGs) were identified including 1288 upregulated genes as well as 666 downregulated genes. GO and KEGG analyses showed that the DEGs and DAGs were enriched in the MAPK (mitogen-activated protein kinase) signaling pathway, the PI3K-Akt (phosphatidylinositol-tris-phosphate kinase 3/protein kinase B) signaling pathway, the Wnt signaling pathway, and the Ras signaling pathway. In total, 1479 alternative splice events (AS) were also identified during SMSC proliferation and differentiation. Among them, a unique AS event was the major per-mRNA splicing type, and SE was the predominant splicing pattern. Furthermore, transcription factors with AS were scanned during SMSC differentiation such as myocyte enhancer factor-2C (MEF2C) and the nuclear receptor subfamily 4 group A member 2 (NR4A2). Our results imply that MEF2C and NR4A2 can interact, and we speculate that NR4A2 and MEF2C might regulate the myogenesis of ovine SMSCs through interaction. Together, our study provides useful information on the transcriptional regulation of SMSCs during proliferation and differentiation at the transcriptional level, and provides a valuable resource for understanding the molecular mechanism of myogenesis and muscle development.

1. IntroductionSkeletal muscle is composed of multinucleated and nondividing muscle cells (fibers). Myogenesis is a highly ordered and complex process, whereby myoblasts fuse in a manner finely regulated by various myogenic regulatory factors (MRFs) such as Myf5, MyoD, MRF4, and myogenin [1,2]. MRFs serve as master transcription factors that are upregulated during myogenesis and cause stem cells to differentiate into myogenic lineage cells [1,3,4,5,6]. Skeletal muscle satellite cells are a type of muscle-derived stem cell lying between the myofiber sarcolemma and basal lamina, which are generally quiescent and can be induced to differentiate into adipocytes, osteocytes, and myoblasts in vitro [7,8,9]. Satellite cells serving as the only source of myogenic cells can afford to differentiate into myotubes [10]. It has been shown that damage to muscles or muscle diseases can activate skeletal muscle satellite cells [11,12,13]. The absence of satellite cells severely hampers myogenic differentiation during the initial wave of muscle regeneration [14]. Furthermore, studies have been reported on myoblast proliferation and differentiation study of the transcriptome of many species such as mice [15,16], geese [17], goats [18], and pigs [19], but there are few transcriptome studies involving ovine skeletal muscle satellite cells.Alternative splicing (AS) is one of the most important contributors of different protein isoforms produced from the same gene, resulting in the high complexity of eukaryotic transcriptomes [20]. The AS events of precursor messenger RNA (pre-mRNA) occur due to the involvement of five small nuclear ribonucleoprotein (snRNP) complexes in the removal of introns [21]. In general, AS events display five patterns: exon skipping (ES), intron retention (IR), mutually exclusive exons (MEE), alternative acceptor site (AAS), and alternative donor site (ADS) [22]. Previous studies have indicated that the AS is associated with the regulation of proliferation [23], differentiation [24,25], and apoptosis [26] of the cell. In addition, a recent report has demonstrated that AS events result in the proliferation and differentiation of myoblasts. A typical example is the gene encoding the myocyte enhancer factor 2 (MEF2) family of transcription factors that form isoforms by extensive alternative splicing. There are four isoforms (MEF2A, MEF2B, MEF2C, and MEF2D) in the MEF2 family, which play an important role in cell proliferation and differentiation [27].In this study, we performed RNA-Seq analysis on ovine skeletal muscle satellite cells at the proliferation and differentiation stages (7 days of differentiation), and screened differentially expressed genes, alternative splice events, and transcription factors associated with muscle development. This study identified potential DEGs and DAGs associated with proliferation and differentiation in SMCSs, and suggests their potential roles in the skeletal muscle development of sheep. These results will provide new insights that are useful for further studies on the molecular mechanism of myogenesis.2. Materials and Methods2.1. Culture and Induced Differentiation of Sheep Muscle Satellite CellsAs previously described, ovine skeletal muscle satellite cells (SMSCs) were isolated from the gastrocnemius muscle of fetal sheep via the 2-step digestion method [9]. In the proliferation stage, the SMSCs were cultured in growth media containing DMEM/F12 (Invitrogen, Carlsbad, CA, USA) with 20% fetal bovine serum (Invitrogen, Carlsbad, CA, USA), 10% horse serum (Invitrogen, Carlsbad, CA, USA), and 1% penicillin-streptomycin (Invitrogen, Carlsbad, CA, USA). In the differentiation stage, the SMSCs were treated with differentiation medium including 2% horse serum upon reaching about 60% confluence. The SMSC samples from the proliferation stage and after 7 days of differentiation were collected for subsequent experiments. All of the samples were kept at −80 °C before RNA extraction.2.2. Immunofluorescent AnalysisThe SMSCs were fixed immediately with 4% paraformaldehyde for 1 h at room temperature, then washed with cold PBS (Thermo Fisher, Waltham, MA, USA) three times. Next, the SMSCs were permeabilized with 0.25% Triton-X (Sigma-Aldrich, St. Louis, MO, USA) for 1 h and then blocked with 3% normal goat serum (Bioss, Wuhan, China) for 1 h and stained with the indicated anti-Pax7 (1:100, Bioss, Wuhan, China) and anti-MyHC (1:100, Abmart, Shanghai, China) at 37 °C for 1 h. The secondary antibody (1:200, Abcam, Cambridge, UK) was incubated for 1 h after washing with cold PBS three times. Finally, the slides were repeatedly rinsed with cold PBS and counterstained with DAPI for 15 min. The slides were imaged under a confocal microscope.2.3. Library Construction and SequencingThe total RNA of six samples was extracted from the SMSCs using the Trizol reagent (Invitrogen) following the manufacturer’s instructions. The concentrations of RNAs were detected by NanoDrop 2000 and the integrity and RNA Integrity Number (RIN) of RNA were assessed by RNase-free agarose gel electrophoresis and an Agilent 2100 Bioanalyzer, respectively. After the rRNA-depleted RNA was fragmented, the RNA-Seq library was constructed using enriched poly (A)-tailed of messenger RNA (mRNA) by magnetic beads with Oligo (dT) (Invitrogen). The enriched mRNA was broken into the fragments and reversely transcribed into first-strand cDNAs. Second-strand cDNAs were obtained using DNA polymerase I (Thermo Fisher) and RNase H (Thermo Fisher). The fragmented mRNA was purified and PCR amplification was performed. Finally, six libraries were sequenced on the lllumina NovaSeq 6000 platform with the PE150 model.2.4. Transcripts AssemblyFirst, the adaptor, contaminated reads, ploy-N reads (with quality less than 3), low-quality reads, the reads with a length less than 50 bp, and duplicated reads were removed from the raw data using the fastp software. Then, the rRNA was removed from the clean data by mapping the silva database using bowtie2 software [28], and clean data with removed rRNA were mapped to the ovine reference genome (Oar_rambouillet_v1.0) by hisat2 software [29]. The fragments per kilobase of transcript per million mapped reads (FPKM) value of the six samples were calculated to determine the gene expression level using StringTie software [30].2.5. RT-qPCR ValidationThe total RNA of the six samples was reverse transcribed to cDNA using the HiScript® III All-in-one RT SuperMix Perfect for qPCR (Vazyme, Nanjing, China) following the manufacturer’s instructions. Then, qPCR was performed using the Taq Pro Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) according to the manufacturer’s instructions. Six genes were randomly selected for the validation of the RNA-Seq results including THBS2, SACS, MCM4, COL1A1, ACTC1, and MFAP4; the GAPHD and ꞵ-Actin genes were used as the internal reference gene and then calculated using the 2−∆∆Ct method. Each qPCR reaction was performed in a 20 μL reaction mixture that included 2 μL template cDNA, 0.8 μL of 10 uM forward and reverse primers, 10 μL Taq Pro Universal SYBR qPCR Master Mix, and 6 μL RNase-free water. The qPCR amplification contained an initial denaturation step (95 °C for 10 s) and 40 cycle stages of 10 s at 95 °C and 30 s at 60 °C. All primers used in the RT-qPCR are shown in Table 1.2.6. Alternative Splicing Gene AnalysisAlternative splice events were identified using the rMATS software [31]. In brief, five major AS events were identified in the aligned BAM file and merged with the reference GTF file including skipped exon (SE), alternative 5′ splice site (A5SS), alternative 3′ splice site (A3SS), mutually exclusive exons (MXE), and retained intron (RI). The differentially expressed genes with alternative splice (DAGs) were screened into two groups with a false discovery rate (FDR) of <0.01.2.7. Differential Expression and Functional Enrichment AnalysisThe differentially expressed mRNAs were identified according to a |log2 (Fold Change)| > 2 and q-value < 0.01 using the R package in edgeR. To explore the functions of differentially expressed mRNAs, we performed Gene Ontology (GO, https://biit.cs.ut.ee/gprofiler/gost accessed on 8 January 2023) and Kyoto Encyclopedia of Genes and Genomes (KEGG, http://kobas.cbi.pku.edu.cn/genelist accessed on 8 January 2023) analyses. GO terms and KEGG pathways with a p-value (p < 0.05) were considered significantly enriched terms and pathways.2.8. PPI AnalysisThe top 50 most upregulated genes and downregulated genes were calculated using the STRING database (https://cn.string-db.org accessed on 26 February 2023), and the protein−protein interaction (PPI) network was inferred and visualized using Cytoscape (v3.9.1). The node colors represent the corresponding interaction genes, and the node sizes represent the upregulated and downregulated DEGs.3. Results3.1. Identification and Overview of Transcriptome Sequencing of Ovine Skeletal Muscle Satellite Cell at Different StagesOvine skeletal muscle satellite cells (SMSCs) were isolated from fetal sheep skeletal muscles by the collagenase digestion method. As expected, a 90% proliferation confluence of ovine SMSCs was observed by day 2 in the in vitro culture (Figure 1A). When the cell density reached about 70%, they were induced into differentiation by a 2% horse serum medium, which resulted in myotubes (Figure 1A). To further characterize the cell proliferation and differentiation, we performed an immunofluorescent assay on SMSCs during the proliferation and differentiation phases. The results showed that Pax7 was positive in the satellite cells at the proliferation stage and MyHC was highly expressed at the differentiation stage (Figure 1A), which indicates that the ovine SMSCs were successfully induced into differentiation.Next, to investigate the gene expression profiles in the proliferation and differentiation stages of SMSCs, transcriptome sequencing analysis was performed. More than 69.57 million clean reads were acquired for the sequencing analysis, and above 92% of the clean reads were successfully aligned on the sheep reference genome for each sample. The Q20 and Q30 values of samples were higher than 90%, and the GC content ranged from 47% to 48% (Table 2). Additionally, the principal component analysis (PCA) demonstrated that the SMSC samples in the proliferation stage were distinct from those in the differentiation stage (Figure 1B). The box diagram shows that higher gene expression levels were observed in the proliferation stage than in the differentiation stage (Figure 1C).3.2. Transcriptional Profiling of mRNA Expression in SMSCs at Proliferation and Differentiation StagesIn total, 1954 differently expressed mRNAs (DEGs) were identified in SMSCs at the proliferation and differentiation stages including 1288 upregulated and 666 downregulated DEGs (Figure 2A,B). The clustered heatmap shows that the DEGs were clearly distinct between the differentiation and proliferation stages (Figure 2C), indicating that dynamic changes in the genes occurred during the proliferation and differentiation periods. The results showed that the DEGs weakly expressed in the cell proliferation stage were highly expressed to regulate cell differentiation such as MYH1, MYBPC1, ACTC1, MEF2C, and IGFBP5. Downregulated DEGs were highly expressed during proliferation and expressed at low levels in differentiation such as AREG and IL11 (Supplementary Table S1). Therefore, we speculated that the majority of upregulated DEGs might promote ovine skeletal muscle satellite cell differentiation.Next, the GO and KEGG pathway enrichment analyses were performed on the DEGs. Regarding the upregulated DEGs, the GO analysis showed that the DEGs were mainly enriched in the biological processes related to muscle development such as muscle system processes, cell periphery, muscle tissue development, actin binding, and muscle cell differentiation (Figure 2D). The KEGG pathways of DEGs mainly included cardiac muscle contraction, the TGF-beta signaling pathway, MAPK signaling pathway, Hippo signaling pathway, Rap1 signaling pathway, PI3K-Akt signaling pathway, and cAMP signaling pathway (Figure 2F, Supplementary Table S2). For the downregulated DEGs, the GO results show that DEGs were enriched in protein binding and DNA binding (Figure 2E). The KEGG pathways of DEGs included cell cycle, the MAPK signaling pathway, PI3K-Akt signaling pathway, and the Ras signaling pathway (Figure 2G, Supplementary Table S3).3.3. Protein–Protein Interaction (PPI) Network of Differentially Expressed Genes in SMSCs at Proliferation and Differentiation StagesAs shown in Figure 3, the PPI network based on the top 50 most upregulated DEGs and the top 50 most downregulated DEGs consisted of 28 nodes and 29 interaction pairs. The PPI network included eight interaction pairs for downregulated DEGs, 10 interaction pairs for upregulated DEGs, and 11 interaction pairs between the upregulated and downregulated DEGs. The results showed that IL6 correlated with nine other genes including IGF1, IL11, and AREG, and MYBPC1 correlated with three other genes including MYH1, which is involved in the proliferation and differentiation of SMSCs.3.4. Validation of the Differentially Expressed GenesIn order to determine the accuracy and validity of the RNA-Seq results, six differentially expressed genes (four upregulated and two downregulated genes) including THBS2, COL1A1, MFAP4, ACTC1, SACS, and MCM4 were detected using the RT-qPCR method. The GAPDH and ꞵ-ACTIN genes were used as the internal reference gene. Consistent with the RNA-Seq results, THBS2, COL1A1, MFAP4, and ACTC1 were upregulated at the differentiation stage, and SACS and MCM4 were downregulated at the differentiation stage (Figure 4).3.5. Alternative Splicing Analysis of mRNA in SMSCs during Proliferation and DifferentiationMost of the multi-exon genes underwent alternative splicing, generating complex and diverse transcripts [32,33]. It has been extensively demonstrated that alternative splicing events in myogenesis are important to proper skeletal muscle development [34,35]. To evaluate the extent and significance of AS in myogenesis, we performed alternative splicing analysis on SMSCs during proliferation and differentiation. A total of 1479 differentially expressed AS events were identified in 1092 genes including 1050 skipped exons (SE), 98 remained introns (RI), 128 alternative 3′ splice sites (A3SS), 100 alternative 5′ splice sites (A5SS), and 103 (MXE) mutually exclusive exons, of which SE accounted for 70.99% of these AS (Figure 5A). Furthermore, genes showing a unique AS event accounted for 78.3% of the total, making this major type of AS event during myogenesis differentiation, followed by 154, 54, 16, and six genes displaying two, three, four, and five AS events, respectively (Figure 5B). Among these, 58 genes were significantly differentially expressed in the genes with alternative splice sites (Figure 5C,D), of which several genes have been reported to be related to myogenesis or muscle development such as MEF2C, MYOM3, MYO5B, CEMIP, and ITGB6 (Supplementary Table S4). To further understand the function of these differentially expressed genes with alternative splice sites arising in SMSCs during proliferation and differentiation, we performed GO and KEGG analyses of 58 overlapping genes. KEGG analysis showed that these overlapping genes were most significantly enriched in cardiac muscle contraction, the MAPK signing pathway, PI3K-Akt signing pathway, and the processes of hormones (Figure 5E). Collectively, these results revealed that the unique AS event was the major per-mRNA splicing type and SE was the predominant splicing pattern, which may be involved in regulating the proliferation and differentiation of SMSCs.3.6. Transcription Factors Identified during SMSCs Proliferation and Differentiation StagesAs transcription factors play important roles in the regulation of gene transcription, we identified those that were differentially expressed transcription factors. A total of 253 DEGs were defined as transcription factors (TFs), comprising 132 up-regulated and 121 down-regulated TFs in myoblasts and myotubes (Figure 6A). In addition, we found 10 TFs (including 5 up-regulated and 5 down-regulated) with alternative splice sites, of which SSRP1 (structure-specific recognition protein 1) and MEF2C (Myocyte enhancer factor 2C) were most significantly differentially expressed (Figure 6B,C). Among the TFs, MEF2C and NR4A2 (nuclear receptor subfamily 4 group A member 2) belonged to significantly differentially expressed genes showing alternative splicing. Further analysis of the PPI network of overlapping DAGs and TFs genes analysis demonstrated that transcription factors interact with DAGs. Among these, SSRP1, FOXM1 (Forkhead box protein M1) and MEF2C can be identified as hub genes (Figure 6D).4. DiscussionMyogenesis is a complex biological process that involves multiple gene regulation mechanisms. Transcriptomic analyses of muscles across multiple developmental stages and myoblasts during proliferation and differentiation have also been reported in different species [18,36]. However, there have been a few reports addressing the regulation mechanism at play in ovine skeletal muscle satellite cells. In this study, we constructed the expression profiles of SMSCs in the proliferation and differentiation stages using the RNA-Seq method. The differentially expressed genes (DEGs), alternative splices (AS), and transcription factors (TFs) involved in ovine skeletal muscle satellite cell proliferation and differentiation were identified. A total of 1954 DEGs, 1479 AS, and 253 TFs were detected during the proliferation and differentiation of SMSCs. We found that the expression levels of all genes during proliferation were higher than those during the differentiation of SMSCs, which is consistent with the results of previous studies [17].Compared with the proliferation stage, we found 1288 upregulated mRNAs and 666 downregulated mRNAs in SMSCs during the differentiation stage including MYH1, MYBPC1, ACTC1, THBS2, COL1A1, IL11, AREG, and IL6. Consistent with previous studies, our results showed that these mRNAs were enriched in multiple signaling pathways that are relevant to muscle development including the Hippo signaling pathway [37,38], MAPK signaling pathway [39], PI3K-Akt signaling pathway [40], and Ras signaling pathway [41]. Among these, we also observed two KEGG pathways, the MAPK signaling pathway and PI3K-Akt signaling pathway, as the most commonly enriched pathways in the upregulated and downregulated genes. It has been reported that the MAPK pathway is composed of a series that operates in the protein kinase cascade, which plays an important roles in the regulation of cell proliferation [42] and differentiation [43]. Furthermore, these DEGs are closely related to the proliferation and differentiation of muscle cells such as ACTC1. Several studies have demonstrated that these genes may profoundly contribute to the expansion and differentiation of SMSCs. For instance, ACTC1 (actin alpha cardiac muscle 1) is involved in the differentiation of myoblasts, and ACTC1 deficiency led to severe structural and functional perturbations in the heart [44,45]. The knockdown of COL1A1 inhibits the proliferation of bovine skeletal muscle satellite cells [46]. Therefore, these DEGs are indispensable to the proliferation and differentiation of SMSCs and to muscle development.In our study, protein–protein interaction (PPI) network analyses of the top 50 most upregulated and downregulated genes showed that DEGs played a significant role in the proliferation and differentiation of SMSCs. For example, interleukin (IL)6 and MYBPC1 (slow skeletal muscle myosin-binding protein-C) have been found to be significantly associated with the proliferation and differentiation of SMSCs. Although IL6 is principally defined as a proinflammatory cytokine, it also potentially triggers and controls the distinct activities of satellite cells throughout the myogenic process [47]. The knockout of IL6 in differentiated C2C12 myoblasts can impair the myotube fusion [48]. Among them, IGF1, AREG, and IL11 were predicted to interact with IL6, which is reportedly involved in satellite cell proliferation and differentiation [49,50,51]. With regard to the upregulated genes, MYBPC1 has been reported to play an important role in normal muscle growth and development processes [52,53,54]. For example, MYBPC1 mutations exert negative effects on muscle function, resulting in an embryo with mild curvature and impaired mobility [55].Moreover, this study also analyzed the AS events at play during ovine myoblast proliferation and differentiation based on transcriptomic data. In order to more effectively explore the AS events involved in regulation during myoblast differentiation, we highlighted the overlapping genes between DEGs and DAGs such as MEF2C [56]. Two main isoforms of MEF2C have been found. In zebrafish, MEF2Cb is a MEF2C paralogue that predominates during somitogenesis, and the overexpression of MEF2Cb leads to the ectopic expression of both cardiac and skeletal muscle related genes [57,58]. The MEF2C isoform is reported to induce cell cycle reentry and the development of heart failure in cardiomyocytes [59]. Additionally, these overlapping genes are mainly enriched in the MAPK signaling pathway. Previous studies have demonstrated that some pathways can indirectly impact the biological functions of cells though affecting splicing factors such as the MAPK signaling pathway [60]. For example, MEK1b and ERK1c are the alternatively spliced isoforms of MAPKKs and MAPKs, and MEK1b and ERK1c can form an independent signaling pathway to regulate cell fate [60].In the development of multicellular organisms, transcription factors determine the fate of individual cells. For the PPI network with the overlap in DAGs and transcription factor, we identified MEF2C, SSRP1, NR4A2, FOXM1, FANCA, and KLHL13 as hub genes in networks, which may be related to myogenesis. Previous studies have shown that hub genes are essential for regulation of the cell cycle [61], skeletal muscle differentiation, and growth [62]. For example, the loss of FOXM1 contributes to cell communication and non-autonomous satellite cell activation in zebrafish skeletal muscle [61]. The interaction of SSRP1 with SRF dramatically increases the DNA binding activity of SRF, and SSRP1 interacts with myogenin and promotes myoblast differentiation-specific muscle gene expression [63,64]. Additionally, we also found that MEF2C and NR4A2 were overlapped in DEGs, DAGs, and TFs. Among these, MEF2C, a member of the MEF2 family, exhibits a strong interaction and co-expression pattern with other genes, and it can promote myogenic differentiation [65,66]. NR4A2 (also known as NURR1) is upregulated in response to muscle exercise. It was been reported that NURR1 binds to a site on the MEF2, and Nurr1 activation controls systemic energy homeostasis in skeletal muscle [67]. Our results showed an interaction between MEF2C and NR4A2, indicating that there might be a NR4A2 binding sites on the MEF2C gene, which is involved in myogenesis regulation. However, the potential functions and molecular mechanisms of MEF2C and NR4A2 should be investigated in future studies.5. ConclusionsIn conclusion, our study provides the expression profiles of the mRNA, alternative splice, and transcription factors of the ovine skeletal muscle satellite cells during proliferation and differentiation. Moreover, we elaborated the roles played by many important gene such as MEF2C, ACTC1, FOXM1, SSRP1, IL6, IL11, and NR4A2 in muscle growth or the proliferation and differentiation of satellite cells, and we speculated that NR4A2 and MEF2C might regulate the myogenesis of ovine SMSCs through interaction. In future, these genes should be assessed to explore the function and regulatory mechanisms of myogenic differentiation in vivo and in vitro. The present study helps to improves our understanding of the differences in gene expression during myoblast proliferation and differentiation, and provides insights into muscle development, which can offer guidance to meat production in the future.

animals : an open access journal from mdpi

10.3390/ani11102779

PMC8532812

The wearable wireless sensor system plays a crucial role in providing behavioral and physiological data for each individual in precision livestock farming. This article reviewed the most types of sensor systems available in the market and summarized detailed information on these systems. Additionally, through meta-analysis, the accuracy of the parameters generated by the sensor system was verified. As a result, it has been shown that there are more than 60 sensor systems of various types have been developed and sold. Most of them generate behavioral and physiological parameters of cattle with excellent performance (e.g., eating time, ruminating time, lying time, standing time, etc.), with the exception of a few parameters (e.g., drinking time and walking time). In this review, it was also investigated that the same parameters predicted by sensor systems of the same brand showed different accuracies, but it was not possible to confirm where this difference originated because the additional experimental conditions presented in the literature were not detailed. Therefore, this review suggested that guidelines for evaluation criteria for research evaluating sensor performance are needed.

The review aimed to collect information about the wearable wireless sensor system (WWSS) for cattle and to conduct a systematic literature review on the accuracy of predicting the physiological parameters of these systems. The WWSS was categorized as an ear tag, halter, neck collar, rumen bolus, leg tag, tail-mounted, and vaginal mounted types. Information was collected from a web-based search on Google, then manually curated. We found about 60 WWSSs available in the market; most sensors included an accelerometer. The literature evaluating the WWSS performance was collected through a keyword search in Scopus. Among the 1875 articles identified, 46 documents that met our criteria were selected for further meta-analysis. Meta-analysis was conducted on the performance values (e.g., correlation, sensitivity, and specificity) for physiological parameters (e.g., feeding, activity, and rumen conditions). The WWSS showed high performance in most parameters, although some parameters (e.g., drinking time) need to be improved, and considerable heterogeneity of performance levels was observed under various conditions (average I2 = 76%). Nevertheless, some of the literature provided insufficient information on evaluation criteria, including experimental conditions and gold standards, to confirm the reliability of the reported performance. Therefore, guidelines for the evaluation criteria for studies evaluating WWSS performance should be drawn up.

1. IntroductionTo increase the sustainability of the dairy industry, there has been an increased need for replacing traditional group-level management with precision dairy farming, which continuously monitors and manages individual productivity and health issues [1]. However, individual monitoring through direct observation of farm staff or video recordings is time-consuming, labor-intensive, difficult to detect accurately, and practically impossible on large-sized farms. Therefore, wearable wireless biosensor systems have been introduced for individual cow monitoring, and research on these systems has been actively conducted in the last 40 years [2].The wearable wireless biosensor system is composed of a battery, a data transmitter, and one or more sensors (tri-axis accelerometer, thermometer, pH electrode, microphone, etc.), which are mounted on the cow’s body to measure and collect biometric data. These sensors can be divided into eight types (ear tags, halters, neck collars, reticulo-rumen bolus sensors, leg tags, tail tags, tail head tags, and vaginal tags) according to their location on the dairy cow’s body [3]. They are used to collect and transmit biometric data, such as acceleration, temperature, pH, and pressure at specified time intervals. These raw data from the sensors are then computed into physiological and behavioral parameters (such as the number of steps, activity level, time spent for eating, ruminating, or lying) by algorithms in the sensor, by the PC software, or through clouding computing. Additionally, these parameters are used as the predictor variables for the diagnosis model for detecting physiological and health status (e.g., estrus events, calving, and illness).The literature reviews about cattle biosensor systems have primarily focused on the performance of diagnostic models and detection alarms [2,4]. However, the parameters generated from the sensors are important, not only for ensuring high performance of the detection alarms of their diagnostic models, but also for obtaining ‘big data’ of physiological status and behavior of individual cattle. Therefore, it is important to investigate how accurately the parameters generated from the sensors can represent animal physiological and behavioral parameters. Thus, the purposes of this review paper were to (1) collate commercially available wearable wireless biosensor systems for cattle farms and (2) review the literature focused on evaluating the accuracy of the parameters obtained from these biosensor systems in predicting the actual condition of animals.2. Currently Available Wearable Wireless Biosensor Systems2.1. Search Strategy and Quality Evaluation of the Constructed DatabaseIn this review, we collected all the information about the currently available wearable wireless biosensors for cattle, summarizing the basic features of these sensors. Our comprehensive search was performed through a web-based search on Google, and the search terms were as follows:cattle AND sensor AND (ear OR halter OR neck OR rumen OR leg OR tail OR vagina)The inclusion criterion was that the product must be currently commercially available. The availability of the sensors was confirmed based on the information obtained from the respective web pages. The products marked as ‘in development’ or ‘to be released soon (concept solutions and prototypes)’ were excluded from this study, i.e., only the products currently available in the market were included in the study. The initial search lasted for three months (August 2019~October 2019). It was conducted extensively and meticulously to obtain a comprehensive market inventory and minimize the risk of missing any relevant products. While writing this review, the search process was re-conducted to prevent the omission of newly released products (~April 2020). During this iterative process, we double-checked if there were any missing products in the existing database.Technical specifications and information on vendor websites were our primary sources of information, and business reports and research papers were additional sources for this review. If we found any further information about a product in scientific articles, we used this information to update our product information. For an objective evaluation of database quality, our database was compared with another independent database, the sensor product database for dairy cattle provided by the Data Driven Dairy Decision for Farmers (4D4F) project (https://www.4d4f.eu/, last updated on 23 August 2019) funded from the European Union’s Horizon 2020 research and innovation program.2.2. Wearable Wireless Biosensor Systems by Type and Mounting Location2.2.1. Ear Tag and Halter TypeSeveral wearable wireless biosensors that can be mounted externally on the animal body, such as on ears, necks, legs, and tails, have been developed. Among these, ear-mounted sensors are mainly equipped with sensors that measure temperature and activity. They are mostly mounted in the middle of the ear and used to check the animals’ health status using temperature data. Most ear tag products equipped with three-axis accelerometer sensors can additionally check the animal’s ruminating, eating, resting, and activity. The management system connected to the sensor uses these data to diagnose an animal’s estrous cycle and health issues.Halter type sensors are attached to the cow’s head, and they measure the cow’s eating and ruminating behavior through a noseband pressure sensor and a three-axis accelerometer sensor. The currently available ear tag and halter type sensors are listed in Table 1 and Table 2.2.2.2. Neck CollarsThe neck collar sensor system consists of a device with sensors attached to the strap hanging on a cow’s neck. This type of sensor is the most commonly used in dairy farms; many companies manufacture it. Generally, neck collars have been widely used to control the amount of feed or measure individual feed intake through radio-frequency identification technology. Recently, accelerometer and microphone sensors have been added to neck collars to measure eating time, rumination time, and activity level. Some are equipped with temperature sensors to measure an animal’s body temperature. These sensors provide farm managers with a cow’s health and estrus information. Some neck collar sensors are used in combination with automatic milking systems. The currently available commercial neck collar tag sensors are listed in Table 3 and Table 4.2.2.3. Reticulo-Rumen Bolus SensorsA rumen bolus system is inserted orally and placed in the reticulum, where it will remain throughout the animal’s life. It is designed to continuously monitor a few rumen parameters (temperature and pH) and an animal’s activity throughout the day. The bolus is equipped with an internal battery, a temperature sensor/pH sensor/accelerometer, and a transmitter for data transmission. Its battery can last for months to years and can transmit the data wirelessly at adjustable time intervals.Bolus sensors are primarily designed to sense ruminal temperature changes, which can signal a shift in animal physiological states. A decrease in ruminal temperature reflects drinking and eating events, and its increase coincides with increased body temperature [5,6,7]. Monitoring changes in the ruminal temperature and activity can facilitate early detection of abnormal behavior, estrous cycle, and illnesses. Unfortunately, the pH sensor is mostly unequipped due to its relatively short lifespan. The currently available commercial bolus sensor systems with a pH sensor have an operational lifetime of no more than a few months since the stability of the pH probe is limited. Thus, rumen bolus systems with a pH sensor are mainly considered as research tools. The currently available commercial bolus products are presented in Table 5 and Table 6.2.2.4. Leg TagsAlong with neck collar sensors, leg tag sensors are a popular sensor technology used in farms. Leg tag sensors are mainly equipped with three-axis accelerometers, which can measure animal activity, walking time, lying time, standing time, and the number of steps. They also provide farm managers with a cow’s health and estrus information. Similar to the neck collar system, some leg tag sensors are used in combination with automatic milking systems. The currently available commercial leg tag products are presented in Table 7 and Table 8.animals-11-02779-t001_Table 1Table 1Information about currently available ear-tag and halter type sensor.ProductCompany(Parent Company)CountryManagement SoftwareMobile ApplicationDimensions (mm × mm × mm)Weight (g)Battery LifeRange(m)Built-in SensorsEar tag SmartbowSmartbow GmbH (Zoetis Services LLC.)ATHerd Monitoring Software ○ 52 × 36 × 17342 years300AccelerometerTemperature sensoreSense Flex tagSCR Engineers Ltd. (Allflex Europe SA)ILSenseHub™/Heatime® Pro+ ○ 68 × 38 × 15253 years200 × 500 *AccelerometerCowManager SensOorAgis Automatisering BVNLCowManager System ○ 60 × 50 × 22325 years-AccelerometerTemperature sensorTekSensor tagTekVet Technologies Co.NLTekAccess™×----Temperature sensorCalf TagFeverTags LLCUSTempVerified×-142 years-Temperature sensorData Collection TagFeverTags LLCUS-×----Temperature sensorHalter RumiWatch Noseband SensorITIN + HOCH GmbHCHRumiWatch Manager/RumiWatch Converter×--2 years-AccelerometerTemperature sensorPressure sensor* Area coverage.animals-11-02779-t002_Table 2Table 2Output data and detection items of the wearable wireless biosensor systems (ear tag and halter type).Product (Module)Output DataData Reporting FrequencyDetectionEar tag SmartbowHigh activity/Activity/Inactivity/Ruminating time/LocationEvery hourHeat/Health disordereSense™ Flex tagActivity/Ruminating time/Heat indexEvery 2 hHeat/Health disorderCowManager SensOor(Find my cow)High activity/Activity/Inactivity/Ruminating time/Eating time/Temperature/(Location)Every hourHeat/Health disorderTekSensor tagTemperatureEvery hourHealth disorderCalf TagTemperatureEvery 15 minHealth disorderData Collection TagTemperatureEvery 15 minResearch purpose(Data acquisition only)Halter RumiWatchNoseband SensorRaw activity/Other chewing activity/Ruminating time/Regurgitated boli counts/Ruminating chew counts/Chews per bolus/Chews per minute/Eating time/Eating chew counts/Drinking time/Drinking gulp count/Temperature (ambient)Every minute/Every hourResearch purpose(Data acquisition only)animals-11-02779-t003_Table 3Table 3Information about currently available neck collar type sensor.ProductCompany(Parent Company)CountryManagement SoftwareMobileApplicationDimensions(mm × mm × mm)Weight(g)BatteryLifeRange(m)Built-In SensorsCowScout NeckGEA FarmTechnologies, Inc.DECowScout Activity monitoring system ○ --5 years100–1000AccelerometerRescounter III NeckGEA FarmTechnologies, Inc.DEDairyPlan C21×----AccelerometerAxel CollarMedria Inc.FRFarm’Life® ○ 100 × 48 × 30160-1000AccelerometerSmart CollarHerdInsightsIEHerdInsights Software ○ --5 years-AccelerometerMoocall HeatMoocall Ltd.IEMoocall Breedmanager ○ --60 days3G coverage-MooMonitor+Dairy MasterIEDairymaster MooMonitor ○ --10 years1000AccelerometerSmartTag NeckPearson International LLCIEPearson Heat Detection with Health Monitoring system ○ --10 years-AccelerometercSense Flex tagSCR Engineers Ltd.(Allflex Europe SA)ILSenseHub™ Dairy/SenseHub™ Beef/Heatime® Pro+ System ○ 84 × 64987 years200 × 500 *AccelerometerSCR H-LDSCR Engineers Ltd.(Allflex Europe SA)ILHeatime® HR System(Independent device)/Heatime® Pro+ System (PC) ○ 84.1 × 64.5987 years200 × 500 *AccelerometerSCR HR-LD/SCR HR-LDnSCR Engineers Ltd.(Allflex Europe SA)ILHeatime® HR System(Independent device)/Heatime® Pro+ System (PC) ○ 84.1 × 64.5987 years200 × 500 *Accelerometer/MicrophoneQwes ISO LD/LD SmarttagLelyILLely T4C management system ○ ---75AccelerometerQwes H-LDLelyILLely T4C management system ○ ---500AccelerometerQwes HR-LDnLelyILLely T4C management system ○ ---500Accelerometer/MicrophoneAfiCollarAfimilk Agricultural Cooperative Ltd.ILAfiFarm Software/Afi2Go Pro Mobile App ○ ---200–800AccelerometerMilkrite|InterPuls Neck Tagmilkrite | InterPulsITMyFarm ○ ---75–500AccelerometerSmarttag NeckNedap livestock managementNLNedap CowControl ○ --10 years75AccelerometerSmarttag Neck/All in OneCRV international B.V.NLOvalert ○ ----AccelerometerActivity meter systemDeLaval International AB Inc.SEAlPro/DelPro Farm Management systems ○ -17010 years200AccelerometerCowlarCowlarUSCowlar×110 × 62 × 332426 months>3000Accelerometer/Temperature sensorHeatSeeker II NeckBouMatic LLCUSHerdMetrix™ ○ -1357 years100–750AccelerometerRealTime SmartTagBouMatic LLCUSHerdMetrix™ ○ ----Accelerometer* Area coverage.animals-11-02779-t004_Table 4Table 4Output data and detection items of the wearable wireless biosensor systems (neck collar type).Product (Module)Output DataData Reporting FrequencyDetectionCowScout NeckActivity/Inactivity/Ruminating time/Eating timeEvery 2 hHeat/Health disorderRescounter III NeckActivityEvery 2 hHeatAxel Collar(Feed’Live/Heat’Live/Time’Live)High activity/Inactivity/Ruminating time/Eating time/Lying time/Standing time-Heat/Health disorderSmart CollarActivity/Inactivity/Ruminating time/Eating time/Heat indexEvery hourHeat/Health disorderMoocall Heat--HeatMooMonitor+High activity/Activity/Low activity/Inactivity/Ruminating time/Eating timeEvery hourHeat/Health disorderSmartTag NeckEating time/Not eating time-Heat/Health disordercSense Flex tagActivity/Ruminating time/Heat indexEvery 2 hHeat/Health disorderSCR H-LDActivity/Heat indexEvery 2 hHeat/Health disorderSCR HR-LD/SCR HR-LDnActivity/Ruminating time/Heat indexEvery 2 hHeat/Health disorderQwes ISO LDActivityEvery 2 hHeatQwes ISO LD Smarttag(CowLocator)Activity/Ruminating time/(Location)Every 2 hHeat/Health disorderQwes H-LDActivityEvery 2 hHeatQwes HR-LDnActivity/Ruminating timeEvery 2 hHeat/Health disorderAfiCollarActivity/Ruminating time/Eating time-Heat/Health disorderMilkrite|InterPulsNeck TagActivity/Ruminating time/Eating time/Location-Heat/Health disorderSmarttag Neck(Cow positioning)Activity/Inactivity/Ruminating time/Eating time/Eating bouts/(Location)ContinuouslyHeat/Health disorderSmarttag NeckEating time/Not eating timeContinuouslyHeat/Health disorderSmarttag All in One(Cow positioning)Inactivity/Ruminating time/Eating time/Not eating time/(Location)ContinuouslyHeat/Health disorderActivity meter systemActivity/Heat indexEvery hourHeat/Health disorderCowlarActivity/Ruminating time/Eating time/Step counts-Heat/Health disorderHeatSeeker II NeckActivityEvery 2 hHeatRealTime SmartTag(Activity/Rumination& Localization)Activity/Inactivity/Ruminating time/Eating time/(Location)Every 2 hHeat/Health disorderanimals-11-02779-t005_Table 5Table 5Information about currently available rumen bolus type sensors.ProductCompany(Parent Company)CountryManagement SoftwareMobileApplicationDimensions(mm × mm × mm)Weight(g)BatteryLifeRange(m)Built-In SensorssmaXtec classic/pH Plus BolussmaXtec Animal Care Inc.ATsmaXtec Messenger 4.0 ○ 105 × 35132 × 35-4 years10–30Accelerometer/Temperature sensor/(pH sensor)San’PhoneMedria Inc.FRFarm’Life® ○ ---1000Temperature sensorMoow Rumen BolusMoow Farm Ltd.HUMoow system ○ --3 years-Temperature sensor/pH sensorSmart Rumen Bolus (Temp/Temp + Activity/Temp + Activity +pH)Moonsyst IndustrialTechnologies Ltd.HUMoonsyst system ○ --6 years-Temperature sensor/(Accelerometer)/(pH sensor)LiveCareuLikeKorea Co., Inc.KRLivestock HealthCare Services ○ 110 × 25-6 years-Accelerometer/Temperature sensor/(pH sensor)eBoluseCow Ltd.UKeCow Software×135 × 271505 monthsHandheld antennaTemperature sensor/pH sensorHerdStrongDVM Systems Co.USHerdStrong® Tru-Core system ○ 114 × 33 × 31-5 years137Temperature sensorSmartStockSmart Stock Ltd.USHealthy Cow Dairy×85 × 301205 years91–182Temperature sensoranimals-11-02779-t006_Table 6Table 6Output data and detection items of the wearable wireless biosensor systems (rumen bolus type).Product (Module)Output DataData Reporting FrequencyDetectionsmaXtec classic/pH Plus BolusActivity/Temperature/(pH)Every 10 minHeat/Health disorder/CalvingSan’PhoneTemperature-Research purpose(Data acquisition only)Moow Rumen BolusTemperature/pH-Health disorderSmart Rumen Bolus(Temp/Temp + Activity/Temp + Activity +pH)Activity/Temperature/(pH)-Heat/Health disorderLiveCareActivity/Drinking bouts/Temperature/(pH)Every hourHeat/Health disorder/CalvingeBolusTemperature/(pH)Every 15 minResearch purpose (Data acquisition only)HerdStrongTemperatureEvery 15 minHeat/Health disorder/CalvingSmartStockTemperatureCustomizableHealth disorderanimals-11-02779-t007_Table 7Table 7Information about currently available leg-tag type sensor.ProductCompany(Parent Company)CountryManagementSoftwareMobileApplicationDimensions(mm × mm × mm)Weight(g)BatteryLifeRange(m)Built-In SensorsRumiwatch pedometerITIN + HOCH GmbHCHRumiWatch Manager/RumiWatch Converter ○ -2 years-Accelerometer/Temperature sensorCowScout LegGEA Farm Technologies, Inc.DECowScout Activity monitoring system ○ --5 years100–1000AccelerometerRescounter III LegGEA Farm Technologies, Inc.DEDairyPlan C21×----AccelerometerIceTag/IceQube(for research)IceRobotics Ltd.UKIceReader & IceManager×65 × 60 × 3096 × 81 × 311302 years-AccelerometerIceQubeIceRobotics Ltd.UKCowAlert ○ 96 × 81 × 311302 years-AccelerometerBreeder TagGenus Breeding Ltd.UKBreeder Tag System ○ --5 years700AccelerometerSmartTag LegPearson International LLCIEPearson Heat Detection with Health Monitoring system ○ --10 years-AccelerometerAfiAct IIAfimilk Agricultural Cooperative Ltd.ILAfiFarm Software/Afi2Go Pro Mobile App ○ -5 years200–800AccelerometerTrack A CowENGS SystemsILEcoHerd Software×68 × 50 × 261246 years700–2000Accelerometermilkrite|InterPuls Pedometermilkrite | InterPulsITMyFarm ○ ---75–1000AccelerometerGyuho (cow step) SaaSFujitsuJPGyuho SaaS system×----AccelerometerSmarttag LegNedap livestock managementNLNedap CowControl ○ --10 years75AccelerometerSmarttag LegCRV international B.V.NLOvalert ○ ----AccelerometerHeatSeeker II LegBouMatic LLCUSHerdMetrix™ ○ -1357 years50Accelerometeranimals-11-02779-t008_Table 8Table 8Output data and detection items of the wearable wireless biosensor systems (leg-tag type).Product (Module)Output DataData ReportingFrequencyDetectionRumiwatch pedometerRaw activity/Lying time/Standing time/Walking time/Stand up bouts/Lie down bouts/Step counts/Temperature (ambient)Every minute/Every hourResearch purpose (Data acquisition only)CowScout LegActivity/Lying time/Standing time/Walking time/Stand up bouts/Step countsEvery 2 hHeat/Health disorderRescounter III LegActivityEvery 2 hHeatIceTag/IceQube (for research)Activity/Lying time/Standing time/Stand up bouts/Lie down bouts/Step countsCustomizableResearch purpose (Data acquisition only)IceQubeActivity/Lying time/Standing time/Stand up bouts/Lie down bouts/Step countsEvery 15 minHeat/Health disorderBreeder TagActivity/Lying time/Step countsEvery 15 minHeat/Health disorderSmartTag LegInactivity/Lying time/Standing time/Step counts-Heat/Health disorderAfiAct IILying time/Lie down bouts/Step countsEvery hourHeat/Health disorder/CalvingTrack A CowLying time/Standing time/Step countsEvery 6 minHeat/Health disordermilkrite|InterPuls PedometerActivity/Lying time/Standing time/Walking time/Stand up bouts/Step countsNAHeat/Health disorderGyuho (cow step) SaaSStep countsEvery hourHeatSmarttag LegActivity/Lying time/Standing time/Walking time/Stand up bouts/Step countsContinuouslyHeat/Health disorderSmarttag LegLying time/Stand up bouts/Step countsContinuouslyHeat/Health disorderHeatSeeker II LegActivityEvery 2 hHeat2.2.5. Tail and Vagina Mounted TypesBoth dystocia and stillbirth significantly impact on animal productivity and farm profitability, often requiring a skilled assistant and immediate intervention at the moment of delivery [8]. In order to reduce the reliance on labor and aid animal management, sensors detecting the calving time without physical observation have been developed. These sensors are attached to the tail (or tail head), and they measure tail movement patterns triggered by labor contractions.Among the sensors used to detect calving, some sensors are inserted directly into a cow’s vagina. Using the principle that a cow’s body temperature decreases before calving [9,10,11], vaginally inserted sensors detect a reduction in a cow’s vaginal temperature and provide a calving alarm to farm managers. Another type of vaginally inserted sensor detects light. When the device is pushed out of the vagina by a cow’s water break, it is recognized that the device is out of the cow’s body through detecting light. At this time, the device sends a text message to the farm manager to notify the start of calving. The currently available commercial products of the abovementioned types are presented in Table 9.3. Literature Review on the Evaluation of Parameters Generated by Wearable Wireless Biosensor SystemsWearable wireless wearable biosensors provide farm managers with physiological and behavioral data, such as eating, rumination, walking, and lying time. These data are generated by computing raw data measured by the sensor using a specific algorithm. The units of the generated values depend on the sensor type and the algorithm used. As the computed physiological and behavioral parameters are used as predictor variables in health and estrus diagnostic models, they should accurately represent the actual state of individual animals. Several studies have been conducted to verify the performance of different sensors. The majority of these studies conducted correlation analyses between the sensor data and the gold standard (actual observations) and performance analyses (i.e., sensitivity, specificity, accuracy, and precision). We reviewed the literature on the evaluation of physiological and behavioral data generated by wearable wireless biosensors.3.1. Search Strategy, Study Selection, and Quality AssessmentA literature search was conducted by a keyword search in Scopus. To avoid an excessive number of search results, we used specific keywords. The final query used to search for articles in the databases was (TITLE-ABS-KEY (correlation OR correlated OR regression OR sensitivity OR specificity OR precision OR accuracy)) AND (TITLE-ABS-KEY (cow OR cattle OR calf OR heifer OR buffalo)) AND ((TITLE-ABS-KEY (sensor* AND NOT sensory)) OR (TITLE-ABS-KEY (automat* OR *meter OR device OR tag))) AND (TITLE-ABS-KEY (detect* OR monitor* OR record*)) AND NOT (TITLE-ABS-KEY (genetic* OR chromatography OR follicle OR muscle OR meat OR DNA OR antibody OR serum OR patient OR assay OR spectro*)) AND (LIMIT-TO (DOCTYPE, ‘ar’)) AND (LIMIT-TO (LANGUAGE, ‘English’)). A total of 1875 articles were retrieved using this query (search date: 26 April 2020).After the initial database search was completed, we screened the title and abstract of each selected article and made decisions on the suitability of each study for inclusion in this review. Articles were included in the final database if they (i) investigated the performance of wearable wireless biosensors for beef or dairy cattle, (ii) evaluated variables related to feeding behavior, moving behavior, or rumen status generated by the sensors, (iii) tested the performance of the sensors with other independent reference measurements (a.k.a. the gold standard), such as real-time or recorded visual observations for the behavioral activities and manual pH or temperature measurements, and (iv) presented at least one or more quantitative evaluation measures, such as correlation, accuracy, precision, sensitivity, and specificity. A total of 46 articles met the above criteria and were selected for our systematic review. These studies evaluated the sensor’s performance in monitoring the following three parameters: feeding behavior, activity behavior, and rumen status. The following information was extracted from the selected papers: target behavioral and physiological parameter (i.e., feeding behavior: eating time, ruminating time, drinking time; activity behavior: lying time, standing time, walking time, step count, active time, inactive time; rumen statue: rumen pH and rumen temperature), sensor information (i.e., mounting position, product name, company, country), animal information (i.e., breed, gender, physiological stage), housing information (i.e., barn type, feeding method), gold standard information (i.e., method, number of observers, reliability between observers), data quantity (i.e., number of animals, total collection time, mean collection time per animal), and evaluation results (i.e., correlation coefficient: Pearson, Spearman, Concordance; diagnostic accuracy: sensitivity, specificity, precision, accuracy).3.2. Evaluation of Wearable Wireless Biosensor SystemsIn this study, feeding behavior was classified as eating, ruminating, or drinking. Feeding behavior is usually measured by a sensor located on the head of the cow, such as an ear tag, halter, or neck collar. Activity behavior was classified as lying, standing, walking, active, or inactive (resting). These activities are usually measured by leg tag sensors; however, there are other types of sensors (e.g., ear tags and neck collars) capable of recording daily active and inactive time. As the gold standard for evaluating the sensor, the total duration of the target behavior quantified through visual observation of an observer is used for the behavioral activities, while independent measurements are used for physiological parameters (rumen pH and temperature). During observation, the trained observer records the start time and end time of the target behavior and calculates the duration of target behavior based on this record. The target behavior is defined through an ethogram, and the observer is trained to identify the animal’s behavior based on this definition before observation. Visual observation of an observer includes both real-time (live observation) and non-real-time (video recordings) observations. The case where values derived from other wearable wireless sensors were used as the gold standard were excluded from this study.The correlation results, i.e., the values of Pearson’s correlation coefficient (PCC), Spearman’s rank correlation coefficient (SCC), and Lin’s concordance correlation coefficient (CCC) were graded using the criteria of Hinkle et al. [12]. The grades were negligible (0.00–0.30), low (0.30–0.50), moderate (0.50–0.70), high (0.70–0.90), and very high (0.90–1.00). PCC and SCC can describe a linear relationship between a measured value and a value to be compared, and CCC can additionally explain the degree of agreement with the measured value as well as the linear relationship. In this review, along with correlation and CCC, the results of binary classification tests based on 2 × 2 contingency tables (true positives, false negatives, false positives, and true negatives) of the sensors presented in the articles are also discussed. The following performance results were considered: sensitivity (Se; true positives out of the sum of true positives and false negatives), specificity (Sp; true negatives out of the sum of true negatives and false positives), accuracy (Acc; true positives and true negatives out of the total number of tests), and precision (Pre; true positives out of the sum of true positives and false positives; positive predictive value).3.3. Statistical AnalysisA meta-analysis was performed for the reported correlation coefficients (PCC, SCC, and CCC) and diagnostic accuracy (i.e., Se and Sp). The mean and 95% confidence intervals of the statistics were estimated through a random-effects model based on the DerSimonian–Laird estimator [13], which was generally considered as the standard procedure in the meta-analysis. Since the animal types, physiological stages of animals, feeding and housing conditions, and sensor products were varied among the studies included in the meta-analysis, the random-effects model was selected instead of a fixed-effects model. Given the non-normality of correlation coefficients, point estimates were variance-stabilized using Fisher’s z-transform [14]. The mean value from each study was weighted based on the inverse variance method using the study sample size (number of animals). We treated evaluations conducted under different conditions within the same article as separate individual studies. The analysis was not performed if there were no more than two independent study samples for one behavior. Heterogeneity was examined using τ2, I2, and Cochran’s Q statistic, where τ2 = 0 suggests no heterogeneity, and I2 values of 25, 50, and 75% correspond to cut-off points for low, moderate, and high heterogeneity, respectively [15]. The differences in the correlation between sensor types were analyzed using analysis of variance. All the procedures of the meta-analysis were performed using the ‘metacor’ function in the ‘meta’ package of R version 4.0.3 [16]. Statistical significance was set at p < 0.05, and the results characterized by 0.05 ≤ p < 0.1 were considered trends.3.3.1. Feeding BehaviorEating TimeEating time refers to the amount of time that an animal spends consuming feed per day. This variable was evaluated in both indoor intensive farming systems (such as free-stall barns or tie-stall barns) and pasture systems (Supplementary Tables S1 and S2). In intensive farming, eating behavior is defined as the chewing or licking movement occurring when the animal’s muzzle is located in or above the feed bunk [17,18,19,20,21,22,23,24,25,26,27,28]. In pasture systems, eating behavior is defined as the process of biting or chewing grass when the cow’s muzzle is located near or above the grass [29,30,31,32,33,34]. PCC and SCC values based on 18 independent study samples from 15 articles (12 for PCC and seven for SCC) showed that the correlation between the eating time recorded by sensors and actual observations was very high, regardless of the sensor type (PCC = 0.90, n = 263, I2 = 51%; SCC = 0.92, n = 178, I2 = 61%; Figure 1 and Supplementary Table S1) [17,18,19,21,22,24,25,27,28,29,30,31,32,33,34]. Moreover, the CCC value based on 12 independent study samples from 10 articles was high (0.88, n = 271, I2 = 67%; Figure 2 and Supplementary Table S1) [17,18,19,22,26,29,30,31,32,34]. The sensor products used between the studies were the same except for the neck collar type (Supplementary Table S1), and the animal type and feeding method were different but showed moderate heterogeneity overall (I2 = 60% and τ2 = 0.25). Among the different types of sensors, on average, the eating time measured by the halters and neck collar tags showed higher correlation with the visual observations (halters, PCC = 0.91 and CCC = 0.96 [24,25,26,31,33]; and the neck collars, PCC = 0.96 and CCC = 0.95 [18,28,31,33]) than that measured by the ear tag sensors (PCC = 0.86, p = 0.07; and CCC = 0.79, p < 0.01) [17,18,21,22,26,27,30]. The results of a binary classification test for the performance of sensors for eating time obtained from 10 independent study samples from seven articles (10 for Se, nine for Sp, seven for Acc, and nine for Pre; Table 10 and Supplementary Table S2) showed an Se of 85% (n = 220), an Sp of 96% (n = 210), an Acc of 91% (n = 184), and a Pre of 87% (n = 210) [20,23,25,26,28,29,32].Rumination TimeRumination time is a variable that represents the amount of time a cow spends ruminating per day. In the literature, ruminating behavior is defined as a behavior that includes regurgitation, rhythmic chewing, and swallowing of the bolus [17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43]. PCC and SCC values based on 33 independent study samples from 25 articles (26 for PCC and eight for SCC; Supplementary Table S3) showed that the rumination time recorded by sensors was highly correlated with visual observations regardless of the sensor type (PCC = 0.88, n = 400, I2 = 82%; SCC = 0.93, n = 210, I2 = 78%; Figure 3) [17,18,19,21,22,24,25,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43]. The CCC value based on 15 independent study samples from 12 articles was also high (0.88, n = 297, I2 = 89%; Figure 4) [17,18,19,22,26,29,30,31,32,34,39,40]. The sensor products, animal types, and feeding methods used were all varied between studies included in the meta-analysis (Supplementary Table S3), and as a result, overall high heterogeneity was observed (I2 = 83% and τ2 = 0.36). The data recorded by the halter sensors showed a very high correlation with the actual observed durations of rumination time (PCC = 0.94, SCC = 0.94, and CCC = 0.97) [20,24,25,28,31,33,38,40]; similarly, the data from the ear tag and neck collar sensors showed a high correlation with the actual observed durations of rumination time (ear tag, PCC = 0.89 and CCC = 0.78 [17,18,21,22,26,27,30,41]; and neck collar, PCC = 0.83, SCC = 0.91, and CCC = 0.91 [19,29,32,34,35,36,37,39,42,43]) (Supplementary Table S3). However, there was no significant difference in the correlation between the sensor data and the visual observation data of rumination time among the different sensor types (p > 0.05). The mean diagnostic accuracy of wearable biosensors based on 10 independent study samples from seven articles (nine for Se, eight for Sp, six for Acc, and eight for Pre; Table 10 and Supplementary Table S4) showed an Se of 92% (n = 205), an Sp of 95% (n = 195), an Acc of 94% (n = 169), and a Pre of 87% (n = 195) [20,23,25,26,28,29,32].Drinking TimeDrinking time is a variable that represents the amount of time a cow spends drinking water per day. In the literature, drinking behavior is defined as the behavior that cows exhibit when they put their muzzles into water bowls and swallow water [23,24,25,28,33]. The SCC value based on four independent study samples from three articles showed that the drinking time recorded by the sensors was poorly correlated with the actual observations (0.50, n = 142; Figure 5 and Supplementary Table S5) [24,25,28,33]. The same sensor product was used for the analysis of drinking time, but there were some differences in the animal type and feeding method (Supplementary Table S5), which showed high heterogeneity (I2 = 79% and τ2 = 0.14). The mean diagnostic accuracy of the wearable biosensors based on four independent study samples from three articles (four for Se, Sp, Acc, and Pre; Table 10 and Supplementary Table S6) showed an Se of 21.9%, an Sp of 99.9%, an Acc of 98.8%, and a Pre of 30.8% (n = 149); notably, Se and Pre were lower than those relative to other feeding behavior variables [23,25,28].3.3.2. Activity BehaviorLying TimeLying time is a variable that indicates how long an animal is lying on the ground per day. In the literature, lying time is defined as the time during which the body is not supported by the legs and is in contact with the ground [18,31,32,33,37,44,45,46,47,48,49,50]. The PCC and SCC values based on 10 independent study samples from eight articles (six for PCC and four for SCC; Supplementary Table S7) showed that the lying time recorded by the leg tag sensors was very highly correlated with the actual observations (PCC = 0.99, n = 180, I2 = 0%; SCC = 1.00, n = 53, I2 = 97%; Figure 6) [18,31,33,37,44,45,46,49]. The CCC value based on six independent study samples from three articles was also very high (1.00, n = 168, I2 = 90%; Figure 6) [18,31,48]. Both the sensor product and the animal housing condition were different among the studies included in the meta-analysis (Supplementary Table S7), and very high heterogeneity was observed (I2 = 94% and τ2 = 1.69), with the exception of the analysis for PCC. The mean diagnostic accuracy of the wearable biosensors based on five independent study samples from three articles (five for Se and Sp and four for Pre; Table 10 and Supplementary Table S8) showed an Se of 99.8% (n = 53), an Sp of 99.9% (n = 53), and a Pre of 99.9% (n = 44) [32,47,50].Standing TimeStanding time is a variable that represents the amount of time an animal spends standing per day. In the literature, standing behavior is defined as an animal’s behavior when it is in an upright position with support from the legs but is not walking [31,33,44,45,47,48,50,51]. The SCC value based on four independent study samples from four articles showed that the standing time recorded by the leg tag sensors was very highly correlated with the actual observations (0.93, n = 56, I2 = 57%; Figure 7 and Supplementary Table S9) [31,33,44,45]. In addition, the CCC value based on three independent study samples from two articles was 1.0 (n = 28, I2 = 87%; Figure 7 and Supplementary Table S9) [31,48]. The sensor products and animal housing conditions used were different between the studies included in the meta-analysis of standing time (Supplementary Table S9), and moderate heterogeneity was observed (I2 = 72% and τ2 = 0.63). The mean diagnostic accuracy of wearable biosensors based on four independent study samples from three articles (four for Se and Sp and three for Pre; Table 10 and Supplementary Table S10) showed an Se of 95% (n = 53), an Sp of 98% (n = 53), and a Pre of 98% (n = 44) [47,50,51]. Only one study tested the performance of a neck sensor in estimating the standing time. The reported sensitivity of a neck sensor was approximately 30% lower than that of a leg sensor (Se = 63% and Sp = 98%) [51].Walking TimeWalking time is a variable that represents the amount of time in which the animal walks per day. Walking time is typically defined as a period characterized by at least three consecutive strides in the forward or backward direction [31,32,33,44,45,47,48,50,51]. The SCC value based on four independent study samples from four articles showed that the walking time recorded by the sensors was highly correlated with the actual observations (0.83, n = 56, I2 = 75%; Figure 8 and Supplementary Table S11) [31,33,44,45]. The CCC value based on three independent study samples from three articles was also high (0.80, n = 28, I2 = 49%; Figure 8 and Supplementary Table S11) [31,32,33,44,45,48]. There were differences in the sensor products and the housing conditions used among the studies included in the analysis of the walking time (Supplementary Table S11), but the heterogeneity was moderate (I2 = 62% and τ2 = 0.21). The mean diagnostic accuracy of the wearable biosensors based on five independent study samples from four articles (five for Se and Sp and four for Pre; Table 10 and Supplementary Table S12) showed an Se of 34% (n = 53), an Sp of 98% (n = 53), and a Pre of 27% (n = 44); the Se and Pre were lower than those relative to other activity behavior variables [32,47,50,51].Step CountStep count is a variable that represents the number of steps a cow makes per day. A step is defined as the phenomenon occurring when the rear foot is lifted completely off the ground and returned to the ground in any location with or without the movement of the entire body [45,48,52,53,54]. The CCC value based on three independent study samples from two articles showed that the step count measured by the sensors was moderately correlated with the actual observations (0.69, n = 22, I2 = 0%; Figure 9 and Supplementary Table S13) [48,54]. Although there were differences in the sensor product, animal type, and housing condition among the studies included in the analysis of the step counts (Supplementary Table S13), no heterogeneity was observed (I2 = 0% and τ2 = 0).Active TimeActive time is a variable that represents the total active time of a cow per day. It should be noted that the definition of active behavior varies in the literature. Bikker et al. [17] and Pereira et al. [30] defined active behavior as the process of moving the head or body and walking. Elischer et al. [37] defined active behavior as standing or walking behavior. Zambelis et al. [27] defined active behavior in detail as follows: exploring, drinking, urination, defecation, rising, lying down, head swinging, self-grooming, and social interaction. Swartz et al. [49] defined active behavior as a step activity in which the right rear leg is lifted off the floor while standing. The PCC and SCC values based on 10 independent study samples from eight articles (seven for PCC and four for SCC; Supplementary Table S14) showed that the active time recorded by the sensors was highly correlated with the actual observations (PCC = 0.80, n = 98, I2 = 77%; SCC = 0.92, n = 146, I2 = 0%; Figure 10) [17,25,27,28,30,31,37,49]. However, the CCC value based on three independent study samples from three articles showed that such correlation was moderate (0.57, n = 51, I2 = 81%; Figure 10 and Supplementary Table S14) [17,30,31]. There were differences in the sensor products and the housing conditions used between the studies included in the analysis of active time (Supplementary Table S14), and high heterogeneity was observed (I2 = 79% and τ2 = 0.33), with the exception of SCC analysis. Unlike the other sensor types, the halter sensors (RumiWatch Noseband sensors) record active time in terms of movement of the muzzle that is not related to ingestion and drinking [25,28,31]. The active time variables evaluated in these studies showed a high correlation with the actual observed values (PCC = 0.87, SCC = 0.92, and CCC = 0.90) [25,28,31]. The diagnostic accuracy of the halter sensors based on three independent study samples from two articles (three for Se, Sp, Acc, and Pre; Table 10 and Supplementary Table S15) showed an Se of 93.1%, an Sp of 93.4%, an Acc of 93.4%, and a Pre of 89.9% (n = 134) [25,28].Inactive Time (Resting Time)Inactive or idle time is a variable that represents the amount of time in which cows are not active per day. Inactive time is defined as the time of lying or standing while resting without performing any action, that is, rumination, eating, or drinking [17,19,21,27,29,30,32]. The PCC value based on seven independent study samples from seven articles was very high (0.94, n = 107, I2 = 84%; Figure 11 and Supplementary Table S16) [17,19,21,27,29,30,32]. Although slightly lower than that of the PCC, the CCC value calculated from five independent study samples from five articles was also high (0.85, n = 81, I2 = 83%; Figure 11 and Supplementary Table S16) [17,19,29,30,32]. There were differences in the sensor products used and the animal housing conditions between the studies included in the analysis (Supplementary Table S16), and high heterogeneity was observed (I2 = 84% and τ2 = 0.42). The mean diagnostic accuracy of the wearable biosensors based on three independent study samples from two articles (three for Se, Sp, and Pre; Table 10 and Supplementary Table S17) showed an Se of 59% (n = 53), an Sp of 98% (n = 53), and a Pre of 89% (n = 44) [29,32].3.3.3. Rumen StatusRumen pH and rumen temperature are variables measured using reticulo-rumen bolus sensors. In the case of rumen pH measured by the bolus sensors, the pH of the rumen fluid measured by a pH meter is used as the gold standard [55,56,57,58]. The PCC value of the correlation between the pH measured by these sensors and actual observations, based on six studies from four articles, was high (0.79, n = 40, I2 = 0%; Figure 12) [55,56,57,58]. However, the CCC value based on two articles (four independent studies) indicated an only moderate correlation (0.62, n = 32, I2 = 0%; Figure 12) [55,57]. There were differences in the sensor product and gold standard used between the studies included in the analysis (Supplementary Table S18), but heterogeneity was not observed (I2 = 0% and τ2 = 0). In the literature, the rumen temperature measured by the bolus sensors was compared with the rectal temperature measured using digital thermometers [56,59,60,61,62]. The PCC value from five articles (contributing to five independent study samples) showed that the rumen temperature measured by the bolus sensors was moderately correlated with the actual observations (PCC = 0.67, n = 456; Figure 12) [56,59,60,61,62]. There were differences in the sensor products between studies included in the analysis (Supplementary Table S18), but low heterogeneity was observed (I2 = 42% and τ2 = 0.01).Figure 12Forest plot of the correlation coefficient of rumen status (pH and temperature) between wearable sensors and visual observation. (A,B) show Pearson’s correlation coefficient and concordance correlation coefficient of rumen pH, respectively. (C) shows the Pearson’s correlation coefficient of rumen temperature. Numbers in parentheses indicate individual studies applying different evaluation conditions within the same article. ‘Total’ means the sample size of each study and ‘Weight’ means the weight for the mean based on the sample size.animals-11-02779-t010_Table 10Table 10Meta-analysis results of diagnostic accuracy of feeding and activity behavior variables from wearable sensors. Diagnostic Accuracy 1,2 SensitivitySpecificityAccuracyPrecisionVariableStudy No. n % (95% CI)Study No. n % (95% CI)Study No. n %(95% CI)Study No. n % (95% CI)Feeding behavior Eating time1022084.9 (70.0–92.7)921096.3 (91.7–98.4)718490.8 (86.3–93.9)921087.3 (72.9–94.3)Ruminating time920592.2 (85.6–95.9)819595.4 (91.0–97.7)616993.9 (91.0–95.1)819587.0 (77.7–92.5)Drinking time414921.9 (5.5–37.1)414999.9 (99.7–100)414998.8 (98.0–99.3)414930.8 (15.0–45.1)Activity behavior Lying time55399.8 (98.2–100)55399.9 (99.6–100)---44499.9 (96.6–100)Standing time43895.3 (87.9–98.2)43898.3 (94.7–99.4)---32997.9 (86.7–99.7)Walking time54833.8 (1.1–60.0)54898.0 (96.0–99.0)---43926.6 (10.5–57.1)Active time313493.1 (90.3–95.1)313493.4 (90.8–95.3)313493.4 (90.7–95.3)313489.9 (85.7–92.9)Inactive time32859.2 (22.7–81.1)32898.2 (95.6–99.3)---32889.3 (75.7–95.5)1 Study No.: number of studies; evaluation results analyzed under different conditions within the same article are counted as individual studies. 2 n, sample size; number of animals.4. Summary and ImplicationsA wide variety of wearable wireless biosensor systems for health or estrus detection are currently available in the market. Most of these sensor systems measure acceleration using a three-axis accelerometer and convert this into a numeric value to quantify specific physiological parameters, such as eating time, rumination time, and resting time, using a customized algorithm. The reporting methods (reporting frequency, data units, etc.) of the information generated by the sensors are also diverse. Important basic information on the sensors, such as the frequency of data measurement and the algorithm used for calculating the value of a specific variable from acceleration, was largely undisclosed because of company confidentiality.To date, several studies have evaluated different parameters related to feeding behavior, moving behavior, and rumen status that were measured and calculated using sensor systems. These sensor systems showed a high performance in measuring most of the physiological parameters. However, the sensor performance for some parameters (e.g., drinking time and walking time) needs to be improved [23,24,25,28,32,47,50,51], and a specific sensor showed low performance for a particular behavior (i.e., walking time measured with a neck sensor) [32,51]. Moreover, it seems that the mounting position of a sensor using an accelerometer is critical to detect a cow’s specific behavior of interest, which is consistent with a previous report [63]. In particular, feeding behavior was classified more accurately by a neck-mounted than a leg-mounted accelerometer (Se 96 versus 80% and Pre 88 versus 79%, respectively), but the opposite was true for lying behavior (Se 95 versus 96% and Pre 82 versus 97%, respectively) [63].A standardized guideline for reporting sensor evaluation is required. Different performance levels were reported under different conditions, which was reflected in the considerable heterogeneity of the meta-analysis (average I2 = 76%). In some cases, the same brand of sensor was evaluated very differently in the literature, even under the same feeding and housing conditions [18,22,27,32,36]. Unfortunately, a number of literature sources provided insufficient evaluation criteria, which makes it impossible to ascertain which evaluation factor caused such differences in performance between the sensors. In order to clarify the factors affecting the difference in the accuracy of these sensors, more detailed information is required as follows: animal information (species, gender, physiological status, etc.), housing information (stall type, pen size, stocking density, etc.), data information (observation time per animal, number of observation points per day, total collection days, etc.), and gold-standard information (method, reliability within and between observers, etc.). In the medical field, there is a guideline for writing papers that report the accuracy of a diagnostic method called a Standards for Reporting of Diagnostic Accuracy (STARD) statement [64]. This guideline contains a list of essential reporting items that can be used as a checklist to ensure that a report of a diagnostic accuracy study contains the necessary information. Performing a meta-analysis using articles written using this guideline enables a detailed discussion of bias and heterogeneity among the studies. Therefore, it is necessary to establish reporting guidelines including the above-mentioned factors (i.e., animal, housing, gold standard, etc.), such as the STARD statement, for papers reporting the accuracy of wearable wireless biosensors.5. ConclusionsIn conclusion, the present study showed that the wearable biosensors tested in the literature predict targeted behavioral information with high accuracy. However, the algorithms used to generate some types of information, such as drinking time and walking time, need to be improved. Furthermore, since the accuracy of behavioral information changes sensitively depending on the evaluation conditions, it is recommended to evaluate each sensor using adequate and validated criteria and report the evaluation criteria in detail.

animals : an open access journal from mdpi

10.3390/ani12070805

PMC8996840

Human–animal interactions consist of many relationship types. The human–dog bond is one such example. This study reviewed how we measure the human–dog bond through questionnaires and found a lack of questions related to the dog’s investment in said bond. To rectify this, twelve semi-structured interviews were carried out with a variety of dog guardians to investigate their views on how their dogs showed that they shared a bond with them. The common themes that emerged included ‘affirmation’, ‘understanding of a dog’s preferences, likes, and dislikes’, and ‘adaptation’. These themes provide a useful foundation from which to design new questions within human–dog bond questionnaires. This would allow better representation of a dog’s investment in the bond and, therefore, help create tools that reflect the reciprocal nature of this relationship.

Dogs play an important role in many western societies, providing companionship, emotional support, and assistance, as well as other more specialist roles. The literature reveals that many human–animal interaction (HAI) questionnaires exist to measure the human–dog bond (HDB). The first part of this study assessed how far existing questionnaires went in measuring HDB (defined as the unique, dynamic and reciprocated relationship between a person and dog, one in which each member can influence the other’s psychological and physiological state). A systematic literature review revealed that a common limitation in HDB questionnaires was a lack of questions based on the dog’s investment in the bond and, therefore, a failure to measure the two-way characteristic of the HDB. This led to the second part of the study: to identify novel themes relating to dog investment in the HDB from which new tool questions could be developed. This was investigated qualitatively using twelve semi-structured interviews on HDB, undertaken with participants from a variety of dog–guardian relationship types. HDB themes that emerged included ‘adaptation’, ‘understanding of a dog’s preferences, likes, and dislikes’, and ‘affirmation’. Subthemes included ‘boundaries’ and ‘expectations’ (within adaptation), ‘excitement’, ‘proximity’, ‘affection’, and ‘recall’ (within affirmation). The themes that arose provide a foundation from which to build new lines of questioning within HDB tools. Such questioning can better represent a dog’s investment in the HDB and, therefore, help create tools that reflect the reciprocal nature of a bond more accurately.

1. IntroductionWithin human–animal interaction (HAI) research, the concept of bond formation is common, particularly in discussions of the human–dog bond (HDB). For over 15,000 years, dogs have been co-evolving to meet our needs as “man’s best friend” [1] and the meaning of our relationships with them has continued to evolve too. To understand the human–dog bond (HDB), it is first important to understand what a bond is. In this paper, we define a human–animal bond (HAB), specifically the HDB, as “a unique, dynamic and two-way (reciprocated) relationship between a person and an animal, one in which each member can influence the other’s psychological and physiological state”. This definition was created by amalgamating the most suitable characteristics of several HAB definitions (see Table 1) and by removing any similarities or overlap with the definition for attachment (a lasting psychological connectedness [2]). The key difference between “bond” and “attachment” is that attachment does not need to be reciprocated between the two parties. Table 1 demonstrates the occasionally contradictory nature of existing HAB definitions.This confusion between HAI and HAB does not end at definitions. The interchangeable use of HAI terms and meanings has led to the design of HAI questionnaires, referred to throughout as tools, which may not always meet their intended purpose (construct validity) [8]. Although such tools are human-centric by nature of their intended users (humans), for those intending to measure the HDB, there is often a notable lack of question representation to assess the dog’s side of the relationship [9,10]. By not assessing the dog’s investment into a bond, tools will only assess the human attachment rather than the bi-directional bond.Within HAI research, much has been written on the benefits of human guardianship and interaction with companion animals, whether through companionship, emotional support, health benefits, or assistance with tasks [11,12]. While not unreasonable to hypothesise that those animal species that humans form bonds with may feel the same way, we must avoid assumptions without objective evidence. In addition, the wellbeing of dogs within either companion or working relationships needs to be assessed as part of bond evaluation.Many researchers have developed tools to help assess the value, strength, and benefit of human–dog relationships (for examples, see Table S1 in the Supplementary Materials). While advantageous in being quick and easy to use to assess relationships, the disadvantage of such tools is that they can only ask one party: the human in the relationship. The unequal representation of both parties’ investment in the bond can leave assessments vulnerable to anthropocentric bias.The first aim of this study was to assess how far existing HAI tools go in measuring an HDB and, therefore, how well they attempt to represent the dog’s side of the relationship. These first investigations highlighted limited evaluation of the HDB based on the dog’s investment in the bond. Thus, the second aim of the study was to identify novel themes based on dog investment in the HDB from which new HDB questions could be developed within HAI tools. This second aim was achieved through use of qualitative interviews with a variety of dog guardians. Due to the dual nature of this study, this manuscript is presented in two parts.2. Materials and Methods2.1. Review of Current HAI and HDB Assessment Tool Questions (Part I)Based upon the results of systematic literature reviews by Samet et al. [13] and Wilson and Netting [8], 569 HAI questions from 170 tools were identified (some of these questions and tools are referenced in Supplementary Materials, Table S1, the rest can be found in Samet et al. [13] and Wilson and Netting [8]). Using the suitability criteria listed in Table 2, questions which were suitable for use in assessing the HDB were agreed (passed) independently by two researchers (LS and HVW). Any questions with contrasting answers were discussed. “Pass”, “Unsure”, or “Reject” labels were used to determine question suitability; pass/unsure was passed, unsure/reject was rejected, while unsure/unsure or pass/reject results were reviewed and discussed again by both researchers and a final decision was agreed upon. These final questions were then independently categorised for content by the two researchers and grouped into broad themes.2.2. Identifying Novel HDB Themes for Future HDB Questionnaires (Part II)The themes identified in Part I of the study led to further investigation of novel HDB question themes in Part II via the use of semi-structured interviews with a range of dog guardians representing various human–dog relationship types.2.2.1. Participant RecruitmentParticipants were recruited via convenience sampling of professional and personal networks of the researcher (LS). Although sampling was convenience based, interviewees were recruited based upon their different HDB relationship circumstances to allow for diversity in representation of HDBs and to avoid bias in views (Table 3 provides a list of the types of relationships explored). All interviewees were aged 18 years or over and written informed consent was obtained from all participants prior to interviews. Consent forms highlighted that the interviews were voluntary, not incentivized, and participants could withdraw from the interview and data collection process at any time.The study aimed to provide preliminary data for a larger project and, therefore, whilst every HDB can be unique and worth investigating, as a time-limited qualitative study, a purposive sample was sought rather than random sampling. Data were collected until saturation was reached (i.e., in line with accepted qualitative research approaches (e.g., [14]), interviews were conducted until similar themes were identified multiple times between interviewees and additional sampling was unlikely to lead to new information being collected). This equated to 12 semi-structured interviews with dog owners or guardians, both of which are referred to as guardians from here on.2.2.2. Interview ProcessThe semi structured interviews were based around a written interview guide of predetermined but open-ended questions (Supplementary Materials, Table S2) to allow for broad topics of interest to be covered in question responses, while enabling participants the freedom to articulate their experiences in their own terms [15]. Participants were provided with an overview of question content and the written HDB definition before the interview began, which was repeated again verbally at the start of the interview. The semi-structured interview guide was piloted with eligible Dogs Trust staff (who were not participating with interviews) to refine content before use. Interviews were conducted either in person or over the telephone, and audio was recorded with participant consent before digitising by the researcher (LS). Pseudonyms were created for each participant and any dog/s or family members that were discussed.Twelve dog guardians were interviewed of ages ranging between 25 and 65. Female guardians were over-represented (75%), as is common in HAI research [16]. Interviews lasted between 20 and 90 minutes with a mean duration of 55 minutes. Relationships between dog and interviewee included companion, working, assistance, social facilitator, breeder, and temporary guardian. Length of relationships varied but all had been established for over 12 months.2.2.3. Thematic AnalysisNVivo (v.12, QSR International, Melbourne, Australia) was used for inductive thematic analysis of interview transcripts. Themes were then inductively identified from semantic and latent codes by LS [14]. Through systematic data coding, the key themes around guardians’ perceptions of dog investment into the HDB emerged.2.2.4. Ethical ApprovalEthical approval for this study was granted by the Dogs Trust Ethical Review Board (Reference Number: ERB023). Informed consent was obtained from all subjects involved in the study.3. Results3.1. Review of Current HAI and HDB Assessment Tool Questions (Part I)From the 569 HAI questions, the researchers independently agreed that 151 questions were suitable for HDB assessment. A further 77 questions were agreed as suitable following discussion (42 initially rated as unsure/pass, 31 as unsure/unsure, and four as pass/reject). Both researchers independently rejected 321 questions, with a further 20 questions being rejected after discussion (18 initially rated as reject/unsure and two as unsure/unsure). This left 228 HDB questions to be independently categorised for content by the two researchers (the 228 questions can be seen in Supplementary Materials, Table S1; meanwhile, Table 4 shows each researcher’s content categorization results). Question content categories formed ten broad themes (Table 5). Only one of the ten HDB question themes was exclusively dog-centred (labelled “dog investment” in the HDB, i.e., emotional investment/attachment), which equated to 18 questions in total (8% of all HDB questions). This indicated that dogs’ perspective of the bond was underrepresented in many current HDB tools. Part II of this study was conducted to begin exploration of dogs’ investment in an HDB in the hope of expanding future question representation from the dog’s point of view.3.2. Identifying Novel HDB Themes for Future HDB Questionnaires (Part II)Within the semi-structured interviews, three themes emerged from the analysis. The first theme was related to humans’ perceptions of HDB affirmation from their dogs and included the subthemes excitement, proximity, affection, and recall. The second theme was related to an understanding of the dog’s unique preferences, likes, and dislikes (this related somewhat to owners’ attitudes towards whether dogs have, or should have, agency, and whether a guardian had respect for that; this theme often related to the human’s understanding and accommodation of their dog’s emotional needs). The third theme considered the adaptation of both dog and human to that specific HDB, and included the subthemes of a relationship boundaries and expectations from both parties.3.2.1. Affirmation of HDB from the Dog (Theme 1)A common theme to emerge from interviews when guardians where asked how they think their dog showed they felt bonded to them was related to some of the responses dogs had to humans. Whether they involved greeting a guardian (excitement), coming when called (recall), choosing to position themselves closely to humans (proximity), and/or affection with said guardian, these responses evoked emotion in guardians (guardians tending to feel positive emotions when related to dogs’ expression of a bond presence, but negative emotions when the opposite was true). From the dog’s perspective, however, many factors other than investment in a bond could play a role in their behaviours described by guardians as affirming; for example, prior training, training methods used, emotional state, temperament, physical comfort, and learnt associations with each behaviour or stimulus. Affirmation was recognized here as a theme because it came up so often in interviews; however, care must be taken to address whether this theme relates directly to a dog’s investment in the HDB in future research.Excitement (Theme 1 Subtheme)Several interviewees cited their dogs’ showing excitement at reuniting with them after time apart, feeling this suggested a dog had missed them or was pleased to see them as opposed to emotional neutrality which may be assumed with a stranger. This participant describes a time when they felt or feel their bond is strongest with their dog: “When I come home from work, he gets what we call the crazies and basically that means it’s just a crazy five-minute period where he’s just like “mum’s home! I’m so excited!” he’s just running around the house, then we have a lil’ tug of war and everything. And I’m just like… it just makes me so happy!”(Guardian of a newly adopted dog) “He was still massively bonded to his previous owner. His previous owner came to visit us about 18 months after he came to us, just kind of to say, “hi and how’s he doing?” and [Dog] went absolutely berserk when he arrived! He was so excited! I mean, literally did laps, did zoomies [sic], around the garden, you know, he was just so excited to see him!”(Current dog’s owner describing the dog being reunited with their old owner) “He is very over the top. Yeah, no you almost have to kind of actively discourage the over enthusiasm when he walks in or when you get home.”(Guardian in a single-dog household)This type of behaviour was also mentioned in relation to dogs reuniting with other familiar dogs: “We always have a whole re-greeting dynamic even if one’s just been away for 10 min.”(Guardian to a multiple-dog household)Guardians tended to assess the absence or opposite of excited behaviour, i.e., a lack of excitement when being reunited, as not indicating a bond, or there being a possible problem with the relationship. Here, a guardian describes their concern when they initially rehomed a dog: “So, at first when we were at home with him, he would sit in his crate, he didn’t really care about being near to us, and so that was part of the reason I thought maybe he’s doesn’t like [us], maybe he’s not happy?”(Guardian of a newly adopted dog)This quote also indicates how the owner felt it was personal whether the dog chose to spend time with them, and how a dog opting to spend time with them was an important signal that the dog liked them and enjoyed spending time with them. The same guardian commented: “One evening he just jumped up on the couch and sat between us and we were like ‘Woah, he’s done it! he’s done it—he’s happy!’”(Guardian in a single-dog household)Proximity (Theme 1 Subtheme)Proximity or choosing closeness with a guardian also featured in comments about relationship type: “Some of them have been very tactile, where the dog has come to me for lots of fuss. And I’ve also been able to, like, they trust me to touch them pretty much anywhere… …whereas some of them have been more relationships of, for instance, just being nearby, so it’s more of just a companionship rather than a full-on physical relationship.”(Guardian of multiple dogs)While proximity seems important to some dog guardians’ affirmation of the bond, some dogs may have more complicated feelings around human company. Even if a bond is shared between the human and animal, an understanding of the need for space may be favourable to a dog’s social preferences and wellbeing, which may determine how favourably they see their relationship with that human: “She [the dog] was quite standoffish in her behaviour, in that she didn’t seek physical contact with you very, very often, or not to start with anyway, she does more now, but at the start she was very self-contained and did her own thing, and would actively at times, she’d sit with you in the room for a little while.”(Guardian of an assistance dog) “He’ll go and lie by himself up the stairs or at the end of the couch beside you… … He would sometimes come up sit with me, but other times go and sit by himself.”(Guardian of an older dog in a multi-dog household)This understanding of agency could aid empathy towards dogs as individuals with a choice in being part of an HDB. These themes crop up again in the Understanding of Dogs’ Preferences, Likes, and Dislikes section below.Affection (Theme 1 Subtheme)Many dog guardians appeared to enjoy the affection and tactile closeness that they had with their dogs, indicating that they believed their dog enjoyed this aspect of their relationship too: “I know everyone says dogs hate hugs, which is rubbish when it comes to [dog’s name], because [dog’s name] will wrap his arms and legs around you... … [dog’s name] sort of pulls you into him.”(Guardian discussing the differences between the dogs in their household) “…and in the mornings, he likes to loll[sic] on me, he likes a cuddle.”(Guardian describing their daily routine with their dog)Dogs that enjoy affection may also enjoy it in different ways depending on the relationship they have with the provider/recipient. Here, a guardian describes the difference his and his dog’s bond makes to the dog’s tolerance of affection from other people: “Some dogs enjoy that tactile contact, and when it’s a new person she’ll enjoy it with you for a few moments, whereas I could probably go on [with tactile contact] for minutes and minutes at a time with her, with her cuddling with me on the sofa and [me] stroking her.”(Guardian in a multiple dog household)Meanwhile, this guardian describes sharing different affectional bonds with each of their five dogs: “I think I am more snuggly with my boys, and they are more snuggly with me, than the bitches are.”(Guardian in a multiple dog household)For some dogs, as with proximity, exchanges of affection are on their terms and not the guardian’s: “I think with [dog’s name] it’s probably kind of late in the evening when she’s got into a comfortable spot. If she’s close to you, and she’s kind of in a spot where she will request, you know, that you stroke her head in a certain way, and be very specific about what she wants and she literally will just kind of cuddle in, and just relax completely, and just as long as you keep doing what she wants you to do then she’s totally happy, and that’s lovely from my point of view in terms of going, you know, feel like she’s relaxing and content where she is, and knowing she’s comfortable.”(Guardian of a rehomed dog)Recall (Theme 1 Subtheme)Recall came up surprisingly often in discussion of dogs’ behaviours within the HDB. Poor recall appeared to be related to a guardian feeling frustrated about their dog’s ‘defiance’ of a command that they believed the dog knew the meaning of. Here, a guardian was asked whether there was anything that their dog could do, or that could happen, that might affect their HDB: “When she eats cat poo! ..... I don’t like her very much (laughs). No, I don’t think so. I think, her recall’s really bad, so when she doesn’t come back, that makes me more frustrated rather than worried because she just acting like an idiot.”(Guardian of an older dog) “I would say he’s one of the most difficult cockers I’ve ever had to date, or worked with to date, and I would say that includes other people’s cockers. One minute he’s there and then the next he’s out and off! Hence the reason he has his own hashtag!”(Guardian of a multiple-dog household, who also breeds dogs)This guardian was asked to describe a time they felt the bond had been broken due to poor recall: “When walking the dog and a herd of runners ran by and that frightened the dog and the dog panicked and ran off. She behaved in a manner unbecoming of her you know, she freaked out and it took me a while to recover her, for her to come back to me, so I’d lost, I lost the bond in that instance.”(Male guardian of a single dog)Conversely, good recall whilst out together appeared to be associated with the theme of proximity, good behaviour, and of a dog choosing to interact with their guardian when in an outside environment, whether via preference or obedience in recall. These guardians described a characteristic of the relationship they share with their dog when on walks: “Walks he always wanted to be near you, around you, he would always come back from other dogs.”(Guardian of a single dog) Interviewer: “And when he’s out, does he retain a bond with you, or does he have any behaviours that indicate that?” Participant: “He always loops, I mean they go far and wide pointers, they just disappear into the undergrowth and out of sight but [Dog] does specifically loop round every few minutes to check in. He’s not the pointer that kind of disappears over the horizon completely. So, he literally does come [back] and he will literally come running back to you with eye contact kind of like going “I’m here, I’m here. Are you okay? Ok right, I’m off again.” “Yeah he will look directly at you.”(Guardian of a multiple-dog household) “Yeah. I actually learned a lot from him because he was very much the dog, you know, if you were out walking and you had him off lead, I’d call to him he’d come straight back to me, I’d pop him on the lead and someone’s off lead dog would be coming bounding up and I’d be going “Please put your dog on a lead!” and the person’s going “but he’s friendly” and I’d be going “but mine isn’t!”(Guardian of a multiple-dog household)Obeying a recall command was seen as good behaviour by guardians. A well-behaved dog appeared to be satisfying or pleasing for the owner who trained them. This owner describes their expectation of their dog to be well behaved but also cites factors that may influence that (e.g., age, play): “In return, I expect him to, yeah, I mean, be fairly well behaved, umm, which but that’s through training and obviously I can’t expect him to be well behaved without him being trained but I mean he does all that anyway, he’s still young and he likes to play but he is generally well behaved—very well behaved actually I’d say. And when he’s out, he’s no trouble at all, he comes back when he’s called, he’s a pleasure to have as a dog actually.”(Co-guardian to a single dog)When this guardian was asked which of their dog’s behaviours showed them that their dog was bonded to them, they responded: “Well, I don’t think it is any one behaviour. I like it when he comes when I tell him to, and he likes to play with me, and I like to play with him, and I can see he’s happy because he wags his tail, and he’s happiest of all when I give him food, so it’s all aspects really, it’s his demur [sic].”(Co-guardian to a single dog)Poor recall out on walks may have had an additional element of frustration, or negative emotional impact to guardians, because of the public nature of the “disobedience”. Judgement by other dog walkers appeared to hold value to some participants, as this guardian explained: “I think the thing I don’t like, or don’t like the idea of, is if people think I beat him [in response to the dog not coming back when called]. I’d hate to go into a park, and I call him, and he sort of like puts his head down for some reason, which he could do, and people think “oh look at that poor dog.”(Male guardian of a single dog)Another participant’s thoughts corresponded to this sentiment when discussing the need for good recall from their dogs when out and about: “I’m not horrible to my dogs, but you must have, for their safety, you need certain boundaries or certain stringent things that are in place because their safety is important. So, if it means that I’m going to shout, somebody might interpret that as “oh you’re being a bit mean”, but I don’t want that dog going in somewhere unsafe.”(Guardian of multiple dogs)3.2.2. Understanding of Dogs’ Preferences, Likes, and Dislikes (Theme 2)An understanding of a dog’s greatest positive reinforcers and motivators for training, happiness and wellbeing, commonly featured within the semi-structured interviews. It was particularly important for successful training and a positive working relationship between dog and guardian, where there was a strong functional aspect to the bond.“The two boys, I think they get as much out of physical contact and being with me as I do, and [Dog] is a classic example in that he will work much more for physical contact and praise than treats, and I train with treats, I use treats, BUT he’s definitely not a foodie orientated dog. No, he loves you—basically just give him a big physical fuss.”(Guardian to multiple dogs)However, for some dogs, their greatest reinforcers or preferences may be unrelated to anything a single human relationship can uniquely provide them, as this interviewee commented: “Sometimes you can get dogs where they’re so they can provide reinforcement to themselves from the environment, and from so many other things, that potentially they don’t pay us [humans] [any attention] in [that] they are not too in tune with the owner, or the owner doesn’t necessarily play a really pivotal role in [their] development, or welfare. Do you see what I mean? Like they [guardians] feed them, and the dog knows that, but they [the dog] [would] also potentially just as happily go off with somebody else. It’s not necessarily a really unique bond in that sense.”(Dog guardian who also works with dogs)Knowing what could scare a dog or when the dog might be uncomfortable could allow guardians to offer support and advocate for their dog when needed: “I remember one dog would scare her a lot so my bond would become more supportive. An example might be thunder or fireworks when then I became another role of protector.”(Guardian of an assistance dog) “We didn’t live right next to the main road, but it was just off our road and up another one and it was there. And it was connecting between two cities, so it was quite busy and a lot of like farm trucks and stuff going past, and I was like we’ll have to do lots of desensitisation and counter conditioning to noisy trucks—didn’t bat an eyelid and there’s lots of buses as well. Doesn’t like cars splashing through puddles though; startles at the noise, of like any sort of noise like the splash of a puddle, or when a bird suddenly takes off from a bush and it’s that kind of like “spshhhh” noise, but not bangs and not claps. It is very specific, like, a specific kind of “spshhh” kind of noise.”(Guardian of a rehomed dog)Knowledge of a dog’s preferences may also lead to other forms of supportive behaviour changes from guardians or encourage guardians to change their behaviours according to the dog’s preferences.“He had been an outdoor [dog] when we got him, we realised quite quickly that he preferred to have access to the garden as much as possible. And we did, we actually gave him a kennel in a reasonably sized garden, so we actually had a kennel in the garden and we just basically kind of let him move around.”(Guardian of a rehomed older dog)“I think one of the reasons I don’t particularly like staying away from home is because lots of hotels and Airbnb’s are really strict about not having the dog on the bed. Which she [the dog] doesn’t like and I really look forward to getting home and having a cuddle with [Dog] in the morning [on the bed] with a cup of tea.”(Guardian of an assistance dog)Lack of understanding about a dog’s behaviour, especially an unwanted one, can highlight to guardians an element of the relationship with their dog that they do not have control of or cannot explain. The size of this element, and the issue it causes for the owner, can place challenges on the bond. Guardians may face the challenge of being able to support their dog with behaviours that are routed in the dog being uncomfortable, upset, or lacking in something needed for good welfare. This could create a characteristic of that bond.“You know he’s always been like that and that’s about the only thing about him. I think, why do you do that? and I don’t know. And I don’t think I’ll ever know coz he can’t tell me.”(Guardian of an assistance dog)“[Dog 1] has had her own different tricky areas, which has taught me so many things, but it’s not taught me to decipher them, just like, “oh, how do I cope with this? So, what do I need to do to work with that”, you know, if I brought [Dog 2] into a new house she’d settle and lie down quite easily. [Dog 1] would to an extent, but I know I could leave [Dog 2], whereas [Dog 1] would be like, “I can’t be left”. So, there is certain scenarios that [Dog 1] needs a bit more work on, she’s reliant on me because she doesn’t like being left.”(Guardian of an assistance dog)Accrediting a dog with agency seemed to support dog guardians’ understanding of their dog and empathy for their feelings and different responses in certain situations: “It’s a bit like a mum-child relationship, she definitely gets angry with me, well not angry just stroppy, but at the same time if she’s feeling poorly, or wants something, I’m the first person she comes to.”(Guardian of a single dog)The same guardian remarked: “I think [Dog] gets angry with me sometimes—oh she does get cross! She definitely gets cross. If we go somewhere, she doesn’t want to be or something, she will absolutely tell me and she’ll strop about it afterwards.”An acceptance and an understanding of these preferences suggested an owner had a respect for that agency. One participant stated: “I think she would be very specific about who she engages with and she doesn’t engage with so she would decide whether she wants to be with you or attached to you. And you’ve got to, well we’ve learnt, or we have figured out that you have to kind of respect her choices or ways and just let her, let her kind of direct things to make her comfortable. So, she could be anybody’s as long as they were prepared to respect that. I don’t think she could cope with a house full of children for example, no way. And as we know that she’s worried about men she actively avoids them. Now, with my husband, she’s as cuddly with him as she is with me but that’s taken five years to kind of learn.”(Female owner of an older rehomed dog)However, once again, participant’s answers showed an acknowledgement of other people’s views in relation to this: “I am conscious people have said to me “oh you shouldn’t allow that” and I’m like “why? that’s their dynamic”. Now I don’t believe they’re a pack. There is definitely a movable dynamic in our house as to, you know, who goes where, who’s allowed to sleep where, who has what toy and that kind of thing, and I allow them that”(Female owner of multiple dogs)Consistency in a guardian’s preferences appeared to help dogs understand their human’s likes and dislikes, and how to respond to those likes and dislikes: “I think what dogs like is rules and they want to know what they are allowed to do and what they’re not allowed to do.”(Guardian of an assistance dog) “I think he expects sometimes, he will expect me to tell him off. Like if he picks up my shoe—and he likes to show off right—If he picks up my shoe, he knows if he sees me he’ll drop it because he knows I’m going to tell him off… …Whereas if [female owner] was to walk into the room and he’s got my shoe, and she says “stop it” he’ll sort of run around, probably doing a round of the room you see, but once he sees me he’ll drop it, and he lays down as if to say “I’ve done something wrong, I shouldn’t of done that”. So, I’m just generally harder with him [than the female owner], you know, because I want him to be a well-behaved dog.”(Guardian of a young companion dog)3.2.3. Adaptations of Dog and Human to the Relationship and an Evolution of the HDB (Theme 3)Within evolutionary theory, a key aspect to creating a unique and successful species is the ability of said species to adapt. If we consider the HDB as the unique entity between owner and dog, then the ability of the two parties to adapt to one another can perhaps predict its success (schematic example in Figure 1). Whether guardians receive, or need to receive, clear and regular affirmations from their dogs, whether guardians are understanding and/or are respectful of their dog’s preferences (and whether this latter aspect impacts the dog’s feelings towards their owner), all feed into the HDB. This will be affected by both parties’ previous experience, personality and temperament type, and expectations, but above all else, whether each individual can adapt to what the other offers in terms of characteristics around these factors. The final theme noted from the interviews was owner and dog adaptations to that unique relationship, which included both parties’ boundaries and expectations in and of the HDB. This quote from the guardian of an assistance dog highlights how these three elements are interconnected: “I think I just put some of the rules out of the window—with the assistance dog charity’s permission as well—and she was allowed on the bed and we didn’t worry about her pooing in a specific place and then we started to really bond, and we’ve [had] a really good bond since.”Expectations (Theme 3 Subtheme)It is likely that humans have several conscious and unconscious expectations from a dog and an HDB in any relationship. This may relate to previous knowledge and/or experience of dogs, such as the example below: “I mean, I guess collies and German shepherds are traditionally kind of bonded dogs but again, I don’t know whether that was a breed related thing or whether that was just because of his background and the way he’d been kind of raised.”(Guardian of an older rehomed dog)Humans may also have expectations they place on themselves as guardians. An expectation of the ease of training is one such example. A canine behaviourist describes a common misconception from dog guardians: “[many guardians believe that] this is brand new, but I should be Einsteining[sic] it by the end of this training session, which we know is not the case because they are individuals. And actually, it is understanding that we need to do a lot of work to buy the dog into what we’re doing, but then we can fade out the use of any tangible rewards and reinforcers if needs be.”(Dog guardian who also works with dogs)Preconceptions about characteristics of the dog–human relationship could be seen as disappointing, or even a deal breaker in some cases, if guardians were not willing to accept the reality of the relationship. It would be difficult to measure whether a dog’s expectations are met in a relationship; however, it could be argued that it can be measured if a dog’s expectations are exceeded by signs of an improvement in trust of a guardian/s over time: “I think when we first got her, he [the partner] was very much hoping that she [the dog] would be if you like “his” dog, as historically quite a few [of our dogs] have kind of gravitated to me, mainly because they had more time with me because of our working pattern. So originally, I think he was anticipating, or hoping, that she would be his dog, more than my dog, but actually as it turned out, she was very, very, wary of us both—but him particularly—just being a man, so it did take him longer, him and her longer, to kind of figure each other out and decide what they like and what they don’t like, but now I would say she’s fully equally bonded with both of us, just in slightly different ways.”(Guardian of a rehomed dog)Letting go of expectations, arguably a form of adaptation to the reality of the HDB, appears to strengthen the bond between some humans and dogs, as this guardian of a working dog describes: “[Some days I think] “I’m going to let you rest, let you be a dog, let’s go for a walk”. And I don’t need to do any more with [her] or don’t want to do anything today. And actually, it’s probably made [our]—in terms of a working relationship—I’ve always had a really good, strong working relationship with [dogs] but probably the working relationship with [the dog] is even stronger [now] because when it does happen, it’s probably not as repetitive, as boring, or the pressures not [there].”(Guardian of an agility dog)For dogs, expectations from relationships are likely to come from experience and the repetition and association of certain human’s care-giving roles within their life. This may be a positively reinforced expectation (e.g., feeding or walking if a dog enjoys or is motivated by these things) or a negatively reinforced association (e.g., veterinary visits if a dog dislikes medical procedures or car journeys). These different associations/expectations may form nuances in the type of HDB a dog shares with a human. Here, an interviewee described her dogs’ behaviours around different humans within their social network, who provided different resources to the dogs: “I would say that the one [bond] they had with my partner were very affectionate and they were quite happy to chill out with him. I would say that with my dog walker… …they tend to get a lot more excitable when she comes in the house. Rather than just “oh your home hiya”, they’re like “OH MY GOD! A walk!” And it’s kind of the same with my mum, when she comes over to visit, she’s really bad for feeding them treats—no matter how often I ask her not too! If mum’s there, they’ll sit at her feet. It’s a learned behaviour but they expect it so they’re a lot more excitable…”(Guardian of a rehomed dog who also works with dogs)Expectation can also form trust from a dog; the expectation that someone will or will not hurt them can impact the HDB. Whether a dog trusts and is comfortable with a human can usually be read from a dog’s behaviour. The opposite is also true. Here, a woman describes the change to a bond she encountered with a dog she cared for: Participant: “We used to walk him every day, got on really well with him, he got really excited bouncing back in his kennel. We were working on doing a grooming programme with him because he had really bad matts in his ears, but he wasn’t great at handling, so we had weeks and weeks of giving him cuddles, introducing him to the type of scissors we were going to use, making the noise around his head, and then when it came to actually touching him with the scissors—not to cut, just to touch—the minute I touched his ears with the scissors he turned, ran at me, growling, barking, growling at me, and obviously I put the scissors away and everything calmed down. But after that I’d probably say for about a week, he wasn’t as excited to see me anymore, he walked to the back of the kennel when I came towards the kennel to get him out. We did get that bond back, but I would say that he was a lot more worried about me being around him because he thought I might have the scissors again.” Interviewer: “That’s really interesting, and quite unusual circumstances so an interesting example. Do you think it took you longer afterwards to build that bond back up with him?” Participant: “yes definitely, he would do a lot of body language, like, he would kind of give me side-eye… …so it did take him a long time to realise I wasn’t going to do it again but to trust me again I would say.” Interviewer: “What behaviours did he show when you got that trust back?” Participant: “He was more excitable and waggy in his kennel, excitable when I came in, he stopped showing me side-eye… a lot more relaxed.”(Guardian of multiple dogs who worked with dogs too)This last quote is also an example of a boundary in the relationship set by the dog and could, therefore, also be included in the section below on boundaries.Boundaries (Theme 3 Subtheme)Boundaries link closely into the previous section’s theme of dog and human dislikes. Knowledge of one another’s dislikes may form boundaries in the HDB; however, if each party is respectful of them, they can build trust, a feeling of safety, and sometimes more dependence in a relationship. Often, it is the dog’s boundaries in a relationship that lead guardians to believe not everyone could be their owner.Interviewer: “so almost because she needs that little bit of extra understanding? And you have that, do you think that makes you feel closer?”Participant: “definitely you almost have to have all eyes on her in certain scenarios… …not everyone could be her owner… …I think she would be a tricky dog. For some people in those scenarios, especially taking her to the vet, because of her pain she will react… …But, you know, it doesn’t faze me, and I know that the pain is there, but I try and make [it] as pain free as possible with her.”(Guardian in a multiple-dog household)“[Dog] has had her own different tricky areas and in there as well which has taught me so many things, but it’s not taught me to decipher them, just like, oh, how do I cope with this?”(Guardian of an assistance dog)Knowing and adapting to each other’s boundaries can assist the relationship, as this participant suggests: “I think what dogs like is rules and they want to know what they are allowed to do and what they’re not allowed to do, and so my partner moved in, and he didn’t want the dog to sleep on the bed with us. Fair enough it’s not a massive bed there’s not a load of room so suddenly the rule has changed for [dog] and she doesn’t get that, and she’s struggling as to why she now needs to sleep on the floor in her bed and not in my bed and I think that was the big problem, and so I had to compromise with my partner I said you know because you know he didn’t want the dog on the bed full stop. And I had to work with him to change that, you know not to change him but to say It’s actually really important for [dog] that she’s allowed on the bed sometimes and so every morning now if I’m not working [partner] leaves and [dog] gets on the bed and we have a cuddle—and that’s a really important part of our week.”(Guardian of a single dog)Management of boundaries can sometimes cause extra lifestyle adaptations and effort for the guardian: “[Dog] would not like unknown dogs around him so that meant in the house sometimes we had, not issues, because with his own group he was fine. But as [the dog] got older, basically when I first got [other dog] I couldn’t have [other dog] and [first dog] in the house together… …I had to do lots of management.”(Guardian of a multi-dog household)How far an owner is willing to adapt to meet a dog’s need may be influenced by many environmental factors. The cost-to-benefit ratio of adaptation is worthy of exploration in future studies.4. DiscussionThe overall aim of this work was to explore the current limitations in lines of questioning around the HDB, and to propose potential themes for improvements to current tools for a more complete assessment of the HDB. The work here highlights how dog investment in the HDB is currently underrepresented in HDB tools. Yet, understanding dog investment in HDB can provide us with greater insights into how and why some relationships are successful while others fail, increasing relinquishment risk for the dog (and potentially the risk of euthanasia) [17,18]. As social mammals, a positive HDB can provide benefits to both parties and contribute to both the dog and human’s wellbeing [19,20,21], the latter typically being where most research is focused. As an adaptable species, dogs are regularly used for assistance, therapy, and training or teaching, and it is, therefore, important that we can measure their experience of the relationships in their lives to ensure that they gain as many of the benefits from these interactions as humans do.A review of the literature (Part I) highlighted the imbalance between dog versus human investment questions in current HDB tools and, therefore, identified a potential flaw in construct design: not measuring both sides of the relationship equally. Therefore, it can be questioned whether the true definition of bond is actually being measured. This imbalance was noted previously in work by Payne et al. [10] and Horowitz [9], while misuse of HAI terms, formerly noted by Wilson and Netting [8], can also lead to a failure to measure a bond, instead commonly measuring human attachment. Payne et al. [10] pointed out that if the HDB is supposedly symbiotic, then affective benefits to the dog, through attachment or otherwise, should also be considered. In other words, failure to measure both parties’ interactions in the relationship is a failure to measure the bond.Behavioural measurement of a dog’s investment in a relationship has been attempted using the Ainsworth Strange Situation Test (ASST). This test has been used to assess attachment styles of dogs, with some authors also using it to help validate human-centric tools (e.g., [22,23,24,25]). Behavioural validation is essential to ensure that anthropocentric evaluation of canine state accurately reflects the reality for dogs themselves. Yet, these tests alone may not capture all the nuances of a relationship, and use of other approaches in addition could allow for a wider perspective of the relationship, facilitating a more in depth understanding [23].The themes that emerged from qualitative interviews are summarised in Figure 1. Adaptation by both dog and human to the HDB appears to be a cornerstone in determining the type of relationship said dog and human will have (as shown in the grey box). Boundaries and expectations within all aspects of the HDB may create limitations or opportunities within a relationship to improve its quality, as might an understanding of a dog’s preferences, likes, and dislikes. Whilst not necessary for a relationship, these factors, if understood and respected, could have an overarching impact on all aspects of the HDB and the distinct factors within it. Meanwhile, the “signs of affirmation”, which guardians perceive indicate bond existence, are not actually needed for an HDB but are reaffirming to the human. That is, reaffirming that the relationship, and the positive feelings about it, are mutual and that their investment in the relationship is being rewarded by return. The themes identified can facilitate future research exploring new dog-investment-based HDB questions using this foundation of preliminary work (such questions would require validity and reliability testing in tools). The themes that emerged here reflect how and why a dog might invest in the HDB and allow development of hypotheses of different bond types depending on both parties’ investment (as has been suggested within human–dog attachment [26]).There are likely to be many other factors within the larger oval in Figure 1 (that encompasses adaptation, expectations, and boundaries) that influence each HDB in a unique way and could be explored further in future research. These other factors include training methods [27,28], individual differences or each party’s personality, previous experiences, aspects of the environment, other relationships and social support networks, husbandry responsibilities, time spent together, and quality of experiences shared. The three main themes that emerged in this study are discussed in more depth below.4.1. HDB AffirmationPerceived HDB affirmation behaviours cannot be assumed to be related to a dog’s investment in the bond but were recurrently reported in Part II of this study as important to humans within bonds. Nagasawa et al. (2008) presented physiological evidence that guardians can experience an increase in urinary oxytocin from their dog simply gazing at them [29]. Therefore, it is likely that perceived bond affirmation behaviours from dogs are physiologically very reinforcing to guardians. This has not gone unnoticed in current HDB assessment tools with the inclusion of questions such as “My dog gets excited when I come home” [30], and “My dog is constantly attentive to me” [31] in attempt to obtain a reading of how the dog might exhibit bond-affirming behaviours towards their humans. However, from a dog’s point of view, several emotional states could be associated with these contexts, e.g., excitement, fear, anxiety, or multiple conflicting states associated with previous learnt experience [32]. Alternatively, they may be related to previous positive reinforcement history, such as maintaining proximity to the human that feeds them. Human interpretation of affirmation behaviour does not necessarily indicate the type of emotions, relationship, or bond a dog has with a human; however, based upon these interviews, they are important to dog guardians in helping them gauge whether their dog is bonded to them.Successful recall and obedience were commonly mentioned by dog guardians as indicative or affirming of a bond, although these responses will be influenced to a great extent by prior training and the dog’s interest in other aspects of an environment. Poor recall was often viewed by guardians as frustrating, embarrassing, or insulting. However, to the contrary, within parent–child secure-base attachment theory, children that were more likely to confidently explore the world away from the parent were considered to feel more secure in their bond [33], and similar secure-base effects have been noted between securely attached dogs and their guardians (e.g., [34]). This demonstrates what guardians perceive as indicative of a bond might not be so, and vice versa. It has been noted in past studies that guardians often have a limited understanding of the relationship between a dog’s emotions and behaviour [35,36,37,38,39,40]. Coupled with the possibility of anthropomorphism and the availability heuristic (i.e., using the most easily recalled information to provide context for decision-making processes [35,41,42]), HDB surveys may struggle to tease apart the dog’s true investment unless objectively relevant dog behaviour questions are included to assess HDB.4.2. Understanding of Dogs’ Preferences, Likes, and DislikesAn understanding and respect for a dog’s unique preferences, likes, and dislikes could be interpreted as an accurate knowledge of a dog’s behavioural motivators and reinforcers. These are key concepts to recognise when working with, training, and promoting a harmonious relationship with any animal; how this knowledge is utilised, respected, and understood could be the difference between simply repeated HAIs and forming a HDB.Acknowledgement of dogs’ preferences can be found in some HDB questionnaires (e.g., “My dog usually plays by himself/herself or someone else instead of me, even when I’m around” [31]). However, dog-preference-related questions were scarce in contrast to guardian-preference questions in which “likes” were heavily represented (see Table S1, Supplementary Materials). This limited representation does not ensure that dog preferences are recognised in HDB tool question content.It is likely that time, proximity, and shared experiences in a variety of situations together assist guardians in learning more about a dog and their responses, reactions, and preferences in a multitude of contexts. This phenomenon is known as behavioural or local synchronicity [43,44,45,46]. For learning to occur however, the human party would still need to observe, acknowledge, and understand the emotions behind a behaviour within a situation [36]. For example, interpreting a dog barking as being badly behaved rather than recognising that the dog may be experiencing fear, anxiety, or frustration, would be a direct barrier to the guardian responding appropriately. Attitudes to animals and empathy for a dog’s feelings are just as important and may be the reason why so many current HAI surveys address these factors.A guardian’s knowledge of situations in which their dog may experience anxiety or fear can allow them to support their pet (e.g., comforting a dog that fears fireworks). In demonstration of this, one of the interview participants said that their dog sought proximity with them when feeling unwell, even referencing a time when the dog had entered into the shower with them when seeking comfort for a digestive upset. If a human can ease a dog’s negative emotional state by providing a source of positive emotion (e.g., safety, comfort, and companionship), then this is likely to impact how the dog feels about that human, their general wellbeing, and their emotional state around that human. Whether the guardian is a “unique resource” to the dog, i.e., whether they can offer the dog something emotionally, which no one else can, may also impact the importance of the relationship to that animal. Dogs’ emotions related to these factors are likely to create a different type of HDB than those of a dog that cannot rely on a guardian for emotional support but may still choose to sit near them when in a room together. Dogs’ choice for proximity is a common question in HDB tools, e.g., “My pet is constantly at my side” [47,48] and “I feel as if my dog often stays physically closer to another family member or a friend than me” [31]. However, this could be influenced by time of day, anxiety, and/or which resource that human may be associated with, rather than sign of a bond.Numerous tools offer questions about the dog being a source of comfort to the human (e.g., “When upset or anxious I turn to my pet” [49]) while none appear to investigate the human as a source of comfort to the dog. This may be related to these questions being removed during tool design if question reliability was poor, or else it could be argued that questions exploring a dog’s choice of proximity to a human begin addressing this as a topic. Without HDB tools simultaneously asking about training methods, a dog’s preferences, and a human’s response to those preferences, alongside verification through behavioural trials, it is difficult to interpret how a dog may feel about the relationship with their guardian. For example, whilst questions such as “My dog does not look at me often” [31] and “My dog is constantly attentive to me” [31] were highlighted as indicative of a dog’s investment in an HDB (Table S1), a similar question “My dog always pays attention to me and obeys me right away” [31] was not included because this appears much more linked to a dog’s training history rather than emotional investment in an HDB (see [27,28]). Future questionnaires would benefit from exploring the use of training method-based questions, alongside trialling question themes based on where a dog seeks comfort.4.3. Adaptation in the HDBA key theme to emerge from thematic analysis was adaptation (Figure 1) by both dog and human to support and encourage a relationship to be established and maintained. Sub themes for this category were ‘boundaries’ and ‘expectations’, with all three themes fitting into the basic outline for prosocial behaviours (i.e., helping each other, obeying rules, conforming to socially acceptable behaviour, and cooperating with each other [26]). In their argument for the prosociality of dogs towards humans, Bräuer suggested that most dogs were motivated to please humans, but they often had problems understanding what was being asked of them [26]. Here, we suggest that the reverse could also be true; humans can also be motivated to please their dogs but can find difficulty in understanding them. Criteria for adaptation, which appeared during the interviews, included behavioural flexibility, forgiveness, gaining knowledge or understanding, empathy, respect, and not crossing boundaries, along with occasional acceptance that a dog may not be able to “invest” as much into a bond, or behave within a relationship, quite as the human party might like.One example of beneficial adaptation for dog–human affiliation is behavioural synchronization [43,44,45,46]. Duranton and Gaunet (2016) defined behavioural synchronization as both parties doing the same thing, at the same time, in the same place [46]. Although it is a reasonably new area of research within canine cognitive science, questions relating to behavioural synchronization have made appearances in past HDB tools. Such questions were often based on sleep (e.g., “He/she is encouraged to sleep on my bed at night” [49]), rest (e.g., “My pet and I watch TV together frequently” [47,48]), play (e.g., “I enjoy playing with my dog”, “I play fetch with my dog often”, “My dog often is not interested in playing with me”, and “My dog usually plays by himself/herself or someone else instead of me, even when I’m around” [31]), and walking arrangements (e.g., “I take my pet along when I go jogging or walking” [49]). However, existing questions on synchronisation are limited in how they address the dog’s point of view regarding the time spent together, not necessarily revealing the quality of the time spent together (i.e., the level of enjoyment and or enthusiasm each party experiences during the activity). Many questions also fail to address active engagement and agency between the dyad when participating in a shared activity. For example, when walking a dog, it can be a very two-way process in which there is engagement between both parties and both the dog and human make choices about the route, or it can be very one sided, human-led, with little true choice in the behavioural synchronicity. Exceptions would be questions that specifically describe an activity that requires engagement (e.g., throwing a ball for a dog on a walk) or an active choice to participate in the same activity, such as sleeping in the same bed at the same time. Equally, forced behavioural synchronization (e.g., placing a restless dog in a crate to “rest” when the owner rests) is not necessarily an example of synchronization as there is no agency involved on behalf of the dog. Future HDB tools would benefit from including questions relating to true forms of behavioural synchronicity between dog and human, particularly those in which humans have adapted to a dog’s behaviours or preferences on how they spend their time.In general, adaptation occurs in response to a change in environment, situation, or routine. Changes result in a stress response, and behavioural adaptation leads to a reduction in this stress. Adaptive responses to change may not always be rapid or successful and may result in prolonged stress responses. For an animal (human or otherwise) to adapt, an initial stressor or stimulus is required. As the name implies, a stressor may be stressful to said animal before it adapts; therefore, refusal, failure, or a long period of time taken to adapt may have knock-on effects on an animal and/or their relationships. It is known that stress leads to elevated glucocorticoids, which have been shown to reduce social tolerance [50,51]. Chronic stress can impact social interactions, and this could further disrupt the relationship or bond formation between dog and owner. Hence, the ability of owner and dog to adapt rapidly to changes in the nature and context of their relationship seems to be an important element in the development and maintenance of a positive bond. Adaptation (including boundary and expectation changes) can be promoted by the guardian’s knowledge of their dog’s preferences—ideally, guardians should be able to distinguish between when their dog has adapted to something and is comfortable, and when their dog is merely tolerating it.Adaptation is not a new theme in qualitative studies on HDB (e.g., [35]), and it addresses whether an animal (human or otherwise) would or could modify their behaviours, routines, and/or own preferences to suit another’s preferences, or benefit a bond (i.e., whether an altruistic bond could be formed). However, adaptation appears to be underrepresented in current HDB tools; although, some questionnaires have tried to ascertain relationship limits (a form of boundary) e.g., “If a 3-month-old puppy or kitten was having problems with destructiveness I would get rid of it” [52]. Boundaries and expectations within all aspects of the HDB may create limitations or opportunities in adapting to each other during the creation or maintenance of a bond (see Figure 1). These are likely to be unique to each dyad depending on environment, personality type, attitudes towards animals, communication style (and success), previous experience, etc., highlighting the unique nature of many HDBs. uture questions on adaptation, whether based on life examples or fictional scenarios, could be incorporated into questionnaire themes on dogs’ preferences, likes, and dislikes.4.4. LimitationsThis paper identifies a lack of dog-centric questioning in current HDB tools and presents the results of a preliminary investigation that identifies potential themes to develop questions to address this gap in future. Any new questions arising from these themes would still require full validity and reliability testing within a questionnaire format to ensure they meet the criteria needed to report accurate information within a tool. Whilst current tools lack lines of questioning on dog investment within the HDB, it is unknown whether this has arisen from HDB tool authors exploring and removing questions that lack validity or reliability during testing. This study used qualitative methods and a diverse purposive sample was sought (rather than random sampling); however, the themes generated in this study could be limited or biased to the participants that took part.5. ConclusionsThis study highlights an area for improvement in current HDB tools: the need for further consideration and representation of canine investment in the HDB. If a bond is to signify a two-way shared relationship, then an understanding of the investment from both parties is required. While there is no shortage of HDB tools (e.g., [53,54,55,56,57,58,59,60,61,62,63,64,65]), there remains a lack of true interpretation of the bond from both parties and little discussion on the types of bonds that may exist outside of relative strength. Until a more rigorous examination of dog investment into the HDB is completed, the content of the questions used for measurement remains scarce and they remain unreliable in their ability to decipher the dog’s investment in a bond. Consequently, existing measurements must be utilised with caution [66,67]. Here, we suggested question themes that could be expanded on in future studies to address the deficit of questions assessing the canine investment in a bond. This paper is focussed on the HDB; however, future work could investigate similar trends of underrepresentation in other non-human species within HAI questionnaires.Better representation of the animal’s investment within human–animal relationships could further our understanding of HAIs and increase our awareness of factors that threaten bonds for one or both parties. Equally, it could provide greater knowledge on distinct types of bonds and the factors that solidify or strengthen them (which could benefit those adopting/adopted dogs [68]), as well as the overall impact bonds have on either invested parties’ wellbeing. It is hoped that the more informed dog guardians are on their dogs’ needs in various types of HDB, the greater the number of interventions and support measures that can be put in place to avoid their break-down. This could reduce the risk of dog relinquishment and/or poor welfare outcomes that may be associated with relationships breaking down.

animals : an open access journal from mdpi

10.3390/ani13071220

PMC10093145

Sentient animals have moral rights. This follows from the best justification for human rights that we can give. However, that does not mean that animals have the same rights as we do. First, they have partially different interests. Second, humans have special relationships with each other, from which special duties also follow. Humans live in states and are subject to compulsory laws. However, we have also made many animals existentially dependent on us through subjugation. A just coexistence with such animals is only possible if we also grant them political membership rights.

Do animals have moral rights? An affirmative answer follows from the best justification for human rights that we can give. The moral status not only of humans but also of animals consists in an egalitarian right to have rights. From this equal status, however, substantially equal rights follow only if the morally relevant interests are equal. A reasonably broad and differentiated understanding of our own, human animal nature reveals which interests we share with many other animals. Thus, sentient animals have basic rights to life and well-being, including volitional activities and access to beneficial social relationships. Further rights arise from special human–animal relationships that are also politically relevant. By subjecting animals and, thus, making them existentially dependent on us, we owe them more than mere protection and help in easily remediable emergencies. We thereby also assume associative duties, as they exist among fellow citizens. Therefore, we should open our understanding of the common good to the reality of species-mixed communities and represent animals politically.

1. IntroductionAll humans have rights understood as morally valid claims that are important enough to deserve legal protection. Due to their moral justification, no state is legitimate that systematically disregards them. Do some or all other animals also have rights in this particularly strong sense? Although in more recent times some legal scholars have argued for legal animal rights [1,2,3], the prevailing view among lawyers is that animals lack legal personhood. Consequently, most defenders of animal rights concentrate on the possible moral foundations in order to show that animals also deserve the strong legal protection that humans already enjoy.The text is organized as follows. First, I outline the function and structure of rights. Then I sketch a model of moral reasoning in which we play the dual role of subjects and objects of moral consideration. Although animals are not capable of grasping the content of moral reasons, they share with us morally significant characteristics in which they basically deserve equal concern. Contrary to the prejudice that animals cannot have rights, I then show that at least the protective function of rights can also apply to them. All sentient animals have basic rights to life and well-being, including volitional activities and access to beneficial social relationships. Further rights arise from special human–animal relationships that are also politically relevant. Not all, but some animals thus have a valid claim to membership in a political community.2. The Concept and Justification of Moral RightsMorality is about duties. Not all, but central duties arise from rights. They follow from the valid claims of others. Whoever violates them does not simply act wrongly; they act wrongly by doing a wrong to others. They violate directed duties whose fulfillment would be owed to some other party [4]. Certainly, rights in this sense are not metaphysical dowries. They are not natural or metaphysical properties of individuals, as the natural law rhetoric of ‘rights inherent to all human beings’ might have us believe. This rhetoric would be better understood to mean that our fundamental rights do not first come into the world through political or social acts of bestowal. Rather, they emerge from moral reasoning itself.A moral right is a special case of valid claims. It is a claim backed up by reasons that can be shared by all normatively responsible persons. I presuppose that at least human rights satisfy this condition. Since human rights, unlike animal rights, are in principle not very controversial, they shall serve me as a model case for the following conceptual explication. I will methodically progress from the more familiar to the less familiar by showing that from the best understanding of human rights claims it follows that many animals also have moral rights [5].Claims always have a three-digit structure. A (a subject) has a claim to X (a good) against B (a bearer of the duty or obligation). The relation can also be logically reversed: B has a duty with respect to X against A. This fact supports the so-called correlativity thesis: no rights of one party without substantively corresponding duties of another [6]. This is not entirely wrong, but it must not obscure an essential asymmetry: Claim rights give rise to the duties that correspond to them. B is obligated to do or omit something because A has the practical authority [7] to demand this of B. B owes A the fulfillment of the obligation. The normative relationship emanates from A, the holder of the right [8]. The latter may directly and for their own sake expect B to do or not to do something. Without this practical authority embodied by A, the duty would not exist.Human rights are a subclass of claim rights. They are morally justified claims that every human being possesses as such. A minimal consensus states that every born and not (entirely) brain-dead individual member of the species Homo Sapiens is necessarily and inalienably endowed with certain rights, including the rights to life and physical integrity. Normatively, they apply even if the positive law of a state does not provide for them. The convincing moral justification, not the actual legal implementation and observance, constitutes the condition of the existence of human rights. We react specifically morally, for example with indignation [9], where we believe them to have been violated; and our indignation would not diminish if we learned that those in power did not believe in human rights and that the laws of the state therefore did not entail them. Rather, we would see this as an additional reason for moral resentment.What might a convincing moral justification for human rights look like? Let us imagine a moral discourse about what we owe to each other as human beings. Such a discourse has a practical purpose: it aims at reasons for norms of action that all normatively responsible persons should see as categorically binding. Every sane person should incorporate a validly grounded norm of morality into their conscience. The reasons that speak for the validity of the norm must therefore not only be valid relative to particular social positions. They must be acceptable for all independent of their bargaining power and other properties that separate agents from each other. Valid norms of morality are based on reasons all moral agents can share [10].However, human persons are not only addressees of moral responsibility. They also require moral consideration from all other moral actors. They are not only moral agents, but also moral patients. We should therefore presuppose that any participant in moral discourse also regards themselves as a vulnerable being worthy of consideration. Each of them has their own point of view of the world, makes their own experiences, and is irreplaceable in the conduct of their life, the success of which is of ultimate value for any of them. Any agent will therefore want to reconcile the norms of morality with the pursuit of a successful life. In Kantian terms: Any moral agent sees themselves as an end and never merely as a means [11].This rules out a purely aggregative view of morality, as represented by utilitarianism. Any moral agent will reasonably attach importance to not letting their most important interests fall under the wheels of collective goal pursuit. We will all demand guarantees that our individual good cannot simply be sacrificed for a supra-personal overall good. The moral status therefore consists in a right to have rights [5]. This concept, redundant only in appearance, emphasizes the difference from a purely aggregative moral conception. Moral agents not only have the right to appear with their utility functions in an ultimately impersonal moral calculus. The right to have rights gives each member of the moral community the normative authority to demand from all others a consideration owed to him or her. Otherwise, a normative order would not merit the consent of every participant in a moral discourse. The normative order must be characterized by the form of subjective rights. In addition, since the participants in the moral discourse are all human beings, they will grant each other rights as human beings.3. Rights beyond AnthropocentrismThe picture that emerges is apparently quite anthropocentric. This is true of the criterion of universal consent and equally true of the circle of rights holders that results from it. However, this is simply because I have started from a discourse in which moral agents seek to work out together what they owe to each other generally and reciprocally. Since only human persons can recognize and critically review moral duties as such, there is no alternative to an anthropocentrism of morality in this epistemological respect. However, a genuinely normative question distinct from this is who should count as a moral patient. Who belongs to the community of beings who deserve consideration for their own sake in the form of subjective rights?To answer this second question, participants in the discourse must refer to and reflect upon their own moral neediness as vulnerable individuals. They should ask themselves what aspects of their vulnerability give them valid reasons for claiming consideration in the form of human rights. In doing so, they should not be distracted by who else might benefit from the consideration owed to them. This follows from the formal generality of morality. It requires us to give equal consideration to the same morally significant properties in otherwise equal circumstances.Only the content of a valid justifying reason therefore determines the condition of moral belonging in the respect specified by this very reason, and only the facts decide who fulfills the condition of belonging. If respect is due to us because of our disposition to autonomy, then in that regard only individuals belong to it who are at least potentially normatively sane and morally responsible. With regard to the sensitivity to pain, on the other hand, all who can feel pain like we do belong to the community of beings to whom no unnecessary pain may be inflicted. After all, at least one reason that we (should) consider morally significant applies to them as well as to us [12].This suggests a pluralistic picture of interests relevant to human rights [3]. It is not determined solely by the higher capacities that distinguish us as normatively sane persons and morally capable actors. Certainly, such specifically human capacities play a role in making some human rights intelligible. Only those who can reason and articulate linguistically have something to gain directly from the freedom of linguistic expression. Only those who can understand what offices and authorizing acts are can meaningfully exercise a political right to vote. Only those who can understand their own moral status have to worry about their self-respect. Even violations such as torture are also abhorrent in human rights terms because the victims are humiliated by them, and their will as autonomous beings is to be broken.However, that is hardly all that makes torture terrible and utterly reprehensible. Shall we give no weight of its own to the aspect of body-bound torment? After all, severe pain is bad in itself. It is so regardless of whether the victim is a brilliant intellectual or a one-year-old child, whether they can speak or only scream. Torture tends to reduce even mature people to mere creatures trembling for their lives and writhing in pain, and it makes them experience in an extreme way that they are creatures capable of suffering.Perhaps we suffer pain in a special way because we can also ask the question about its meaning and possible justification. In addition, we certainly do not interpret every intentional infliction of pain, no matter how severe, as torture. The interpretation of an infliction of suffering as torture implies the accusation that the procedure is at least prima facie wrong. However, it would certainly also be wrong to pull out the nails of a newborn child without compelling reasons. Moreover, it would be absurd to assume that the only decisive factor for the type and severity of the pain is whether the victim is aware of the injustice that he or she may experience through the treatment.Thus, although we started from human persons as participants in moral discourses, interests also become recognizable as relevant to human rights, which the participants have in common with many other creatures. Not only human beings who can relate to reasons as reasons and can obey duties out of insight can have a more or less good or even only bearable life. Immature and severely mentally impaired people can also experience their existence as more or less gratifying, and the same is true for many other animals. They all possess not only physical but also psychological characteristics that grant them their own perspective on their existence in the world. They are all irreplaceable individual subjects experiencing, or suffering, some of their life processes. This ontological feature separates humans and all sentient animals from inanimate objects, but equally, as far as we know, from simple animals such as nematodes, from all plants, fungi, and microorganisms [13].This is also normatively relevant because all sentient beings possess interests that are eligible for morally owed protection and morally demanded promotion. In our own case, this protection and promotion take the form of human rights. Yet, we share some interests relevant to human rights in this way or in a similar way with many other animals capable of feeling and experiencing. After all, we are not only rational, linguistically gifted, and moral beings, but also bodily existing finite creatures that can suffer and need bonding. We should therefore also grant morally justified rights to all other such creatures. They share with us the fact that their lives matter to them subjectively and that the success of their lives is of ultimate value to them. Thus, they should all be recognized as ends and never merely as means [14].It is worth noting that the argument sketched out so far is not identical to the notorious argument from human marginal cases (AHMC) [15]. The AHMC also starts with the basic moral imperative to treat like cases alike. We establish the rights of humans based on characteristics that we consider morally significant. We share some of these characteristics with many other animals. If we nevertheless exclude them or put them in a worse position, it seems to be only because of their different genetic makeup. Alternatively, we could try to base human rights solely on such properties that no other animal shares with us. This counts in favor of recourse to our ‘higher’ capacities such as linguistic agency and normative sanity. If only these higher abilities were relevant to our rights, we would indeed be allowed to exclude all other animals.Unfortunately, however, we would thereby also exclude many fellow human beings. Some born and not (completely) brain-dead members of our species are not even potentially capable of propositional language and normative sanity because they are, for example, severely mentally handicapped from birth. Do they therefore have no human rights? Those who consider this counterintuitive and downright abhorrent, however, seemingly have to concede that some nonhuman animals have the fundamentally same rights as comparably competent—or impaired—humans do. In short, our ‘higher’ traits may constitute human monopolies, but not all humans possess them. Such properties, on the other hand, which are possessed by just about all born and non (whole) brain dead humans, are not human monopolies; we share them with many other animals. Thus, we need to include these other animals as well. This simple ‘moral set theory’ is the essence of the AHMC.However, my argument presented above is not dependent on the truth of the AHMC. It does not refer specifically to mentally impaired people. Instead, it relies on an appropriately broad understanding of the animal nature of even all mature human beings. Such an understanding counts in favor of a pluralistic conception of the morally relevant interests of humans. It thus simultaneously forms a bridge across the species barrier. It allows us to decide, free of arbitrariness, which interests of animals also morally deserve protection and promotion. Many animals possess interests that we among us humans consider sufficient to protect and promote them by means of moral rights. Accordingly, it follows from the basic moral precept of equal treatment of equal cases that an animal possesses a right in precisely that respect in which it sufficiently resembles a human being who possesses a right in this very regard.A possible objection to this last claim is that it is inappropriately anthropocentric. It might suggest that the more an animal resembles us, humans, the more it is morally worthy of protection and promotion. However, an interest does not have to be the same in all respects for all people and all animals in terms of content. Different individual animals of different species have different needs and abilities and can therefore be hurt in different ways [13]. For example, sensory deprivation means something different for an animal that orientates itself primarily by means of the sense of smell than it does for an animal that uses an echo sounder.Once we recognize that we share a basic dimension of interest with other animals, we should be open to morally relevant differences within that dimension. Thus, it is only the epistemic starting point of morality which is necessarily anthropocentric: in moral judgments, we cannot help but begin with a substantive respect in which we owe something to each other in order to take the step across the species barrier from there. Without any similarities to us, we simply could not decide whether and with regard to what we should consider an animal. However, this does not force us to anthropomorphize those animal interests, which we in this way reveal as morally relevant.4. Objections to the Possibility of Animal RightsNotwithstanding the partial overlap of morally relevant interests, some philosophers and lawyers argue that animals cannot have rights. Three such arguments should be considered. A first one starts with the correlativity of rights and duties. Almost all animals except humans are incapable of insightful compliance with the rights of others. They lack the ability to accept an obligation from insight into the validity of the reasons justifying it. However, the view that only those who can observe moral obligations arising from rights can have rights themselves is already implausible with regard to human rights. Even small children or people with severe dementia cannot insightfully respect other people’s rights.A second argument refers to a distinction drawn by the legal philosopher Joel Feinberg [16]. Those who confidently make use of their normative possibilities as rights holders thus testify to their self-respect. By claiming rights, they can value themselves as holders of valid claims. Feinberg argues that this special self-relation is only conceivable in a world with rights. He illustrates this with the counter-image of a world only with duties. In this counter-world—Feinberg calls it Nowheresville—a considerate interaction on all sides would probably be possible. Perhaps no one would have to fear for their life, their health, or their basic freedoms. Perhaps everyone would treat each other nicely. However, no one could say that certain actions were owed to her. No one could demand that others do or refrain from doing something for her sake. In this important respect, Nowheresville would remain normatively deficient: it would be a moral world without self-respecting subjects.However, if rights enrich the moral universe by enabling self-respect, does this not confirm the thesis that only self-conscious subjects can have rights? After all, only they have a self-respect to gain or lose. Animals cannot conceptualize themselves as rights-bearers and cannot derive any self-respect from this status. There is a conceptual reason, however, why enabling self-respect cannot be the very ground of rights. Self-respect is not a good that we are entitled to because we desperately need it. It is a genuinely normative good. Those who respect themselves thereby testify that they know about and affirm their own moral status as subjects of rights.However, this means that the commanded respect logically precedes self-respect. Nobody has a claim to respect herself for something which is not already due to her independently of this self-relation. The fundamental claim can therefore only be the claim to respect itself. Self-respect is systematically secondary to it. This does not invalidate Feinberg’s insight that normative orders with rights, unlike those that entail only duties or even virtues, produce subjects who can respect themselves. This is an additional advantage of rights-oriented orders, but it does not give rise to the moral status on which self-respect is based.Feinberg himself denied that only possible subjects of self-respect could also be subjects of rights. Even though rights can be claimed, the person who claims something and the individual in whose name she claims it do not have to be identical with each other. According to Feinberg, every individual with their own interests can be considered a right holder. That applies to young children and likewise to many animals [17]. Others could exercise the claims as proxy representatives for them.However, does this not mean that something essential is lost? A merely advocatory use of rights may be conceptually possible, but it seems to be part of the special value of rights that they give us possibilities for the self-confident shaping of normative relationships. The connection between rights and self-respect points to the emancipatory value of valid claims. This added value of rights is emphasized by supporters of the will theory, which, as the name already suggests, accentuates the willpower of the rights holder [18]. The latter can use their rights as they see fit and thereby also change normative relationships. Thus, an owner can reclaim a loan, appeal to the courts, or even make others into owners with their consent. Holders of rights have freedoms and powers that they would not possess in a world with only duties or even virtues. Subjective rights in the sense of will theory are thus an important embodiment of the modern basic value of personal self-determination or autonomy.From the will theory emerges the third reason why animals could not have rights. Almost all of them, after all, are incapable of self-conscious and self-determined use of claims. They possess at best a rudimentary concept of moral claims and normative relations as such. However, do rights really always include freedoms such as the use of a thing at one’s own discretion and powers such as to transfer property? There are two reasons against such a generalization [19].First, the will theory leaves no room for inalienable rights. An inalienable right is a valid claim that the holder may not relinquish or cede to others. Thus, the right not to be enslaved arguably includes the prohibition against offering oneself for enslavement. Consent to an enslavement contract would be normatively void: no one would possess the competence to enter into such a contract. Similarly, procedural rights such as the right to a fair trial are not primarily there to bring the free will of the accused to bear. The latter may have the freedom to choose a defense counsel or a particular defense strategy, but she does not therefore also have the freedom to forego overall fair treatment in court. Second, the will theory has a socially exclusive effect. It misses all subjects who are not yet, no longer, or never will be capable of self-conscious self-determination. This is not a marginal problem, and it does not only concern animals. It affects all human beings in the early, and many in the later stages of their lives. Once again, the solution lies in the possibility of using rights by proxy.The will theory is therefore not a good general framework for a conception of claim rights. It tells us how self-aware subjects can use some of their rights, but it tells us too little about what rights might be good for. That is rather the domain of the interest theory. The latter emphasizes the substantial benefits that rights offer us. According to the interest theory, an individual X can have rights if some aspect of X’s well-being (an interest) is weighty enough for holding some moral agents under a duty [20]. Sentient animals are capable of subjective well-being and can therefore be considered rights bearers in the sense of the interest theory.To be sure, animal rights lack the kind of emancipatory added value that will theorists emphasize. Their central purpose is protection, not personal autonomy; however, the protective function of the rights is essential, especially for animals. Today, many of them are even excluded by definition from the space of rights-bearers. They are labeled as ‘livestock’ or as ‘pests’ and thus reduced to possible contributions to the fulfillment or thwarting of human purposes. Likewise, humans dispose of animal habitats without appreciably considering their importance to the well-being and flourishing of untold numbers of animals. Humans decide on new settlements, roads, power lines, cultivation areas, or spoil heaps and regard animal territories as terra nullius [21].In the meantime, many states have enacted laws to protect animals. German animal protection law, for example, stipulates that animals must be stunned before slaughter. However, the law basically serves to regulate the use and also killing of animals for foreign, primarily human purposes. The so-called farm animals are only in the world to be used or consumed by humans. They have no independent right to their own life. Our interests of use form the framework of their consideration by animal protection norms. These norms should ease the animals’ lot after it is already determined that they should give us meat, milk, eggs, leather, or other products and may be locked up and slaughtered for it.However, sentient animals are not our natural servants. They are individuals with a life of their own, which for each of them is the only one they have. To recognize them as holders of rights would mean to respect them as ends instead of seeing them merely as means. In addition, even if very few rights apply absolutely, they do imply strict justification obligations for any violation of basic goods [22]. Most animals would objectively benefit greatly if at least their painful and life-shortening use for comparatively trivial purposes such as meat-eating were excluded and if they were no longer allowed to be mere material for medical experiments. Animals would even benefit from rights in a special way: they are particularly vulnerable inhabitants of a world designed by humans for humans, they cannot defend their interests themselves through collective action, and the vast majority of them cannot hope for our spontaneous sympathy, as dogs, cats, or canaries might.5. What Rights Do Animals Have?In order to have a morally relevant interest, an animal must only be sentient. Sensations are not value-neutral, for they have a more or less pronounced positive or negative valence such as joy or pain. In addition, sentient animals actively deal with their environment and care about their own well-being and survival in at least a behavioral sense. Therefore, we should not only regard them as sentient but also as capable of experiencing in a broader sense, including strivings and volitions [13].We can therefore give a broad outline of the content of animals’ moral claims: All sentient animals capable of experience have at least interests in the dimension of well-being, which includes pleasant sensations and pleasurable experiences as well as scope for volitional activities. Negatively, this means we must neither let animals suffer without necessity nor prevent them from engaging in activities that are promising for them. In addition, as far as they are social beings, like all animals domesticated by us, they must be able to establish contact with conspecifics or other—human or non-human—animals in forms that fit their social dispositions. I think, moreover, what I can only assert here, that we should also grant animals capable of experiencing a fundamental interest in their own further life. After all, life is a condition of the possibility of any beneficial experiences and activities, and we should not deprive animals of this possibility without necessity by killing them.Last but not least, we also know how strong animal legal claims are in principle: they are just as strong as the claims of any human being based on similar interests. This, in turn, follows from the basic requirement of equal treatment of equal cases: equal morally relevant interests count alike, no matter whose interests they are. In this sense, all the bearers of moral rights possess fundamentally the same status irrespective of species affiliation.However, it does not follow from this that all animals that have any moral rights at all, therefore have the same rights as humans do [22]. The assumption of a fundamental equality of status is, first, compatible with the fact that animals do not have certain rights that humans have, because they do not bring them any advantages. Incidentally, this is again already true among humans. A political right to vote for very young children makes no more sense than a right to vote for foxes. Animals may benefit from how people exercise their right to vote who can exercise it intelligently, but they are then only indirect beneficiaries of a right that they themselves could not use with rhyme and reason.Rights to political and also personal autonomy are, generally speaking, only directly significant for individuals who can bear a normative responsibility for living together and for leading their own lives. For the vast majority of non-human animals, this does not apply even in a rudimentary sense. Likewise, protection against discrimination or symbolic degradation is at most indirectly significant for them, since they cannot understand their own valid claims as such and relate them to the valid claims of others. This means, for example, that animals cannot suffer as such from the fact that humans somehow dominate them [23]. In fact, it seems to me that human animal husbandry is permissible precisely when it is consistent with the principle of equal regard for all the interests that animals actually have: in continued life, in well-being, and also in volitional activities and access to beneficial social relationships. Conversely, however, this means that almost all commercial and, without exception, industrial animal husbandry is incompatible with this principle and should therefore be abolished.A second type of differentiation is permissible where animals have similar morally significant interests as we do, but these interests are less strong in their case. This is certainly the more complicated case, and we must not conclude from the fact that, e.g., a restriction of freedom of movement affects an animal and not a human being that it is therefore less severe. Nevertheless, some restrictions will affect a more complex experiencing animal more strongly and in more dimensions than a less complex experiencing animal, and human persons are the most complex animals we know.Maybe we should therefore also assume that human persons, or self-conscious subjects in general, have a uniquely strong interest in their own survival. After all, they can refer to their own existence in a forward-looking or also in a retrospective way, wish for its continuation, and pursue longer-term projects, which only consciously but not self-consciously living animals are not able to do. To argue for such a position on the gradability of the interest in life would lead us into our very own philosophical depths [24]. Here it may suffice to point out that the presence of an interest across species does not necessarily mean that it is present everywhere in the same strength. Moreover, to consider unequal interests unequally would not be arbitrary in itself. It would certainly be arbitrary, however, if we killed an animal for a minor advantage such as the pleasure of a palate. At least in societies like ours, all people could exist well and healthy without products for which animals live miserably and die violently.6. Social Relations and Membership RightsI arrived at this result of a moral status equality of humans and (other) animals by inserting the principle of equal consideration of interests into the framework of a conception of moral rights. However, the obligations of moral agents toward an individual do not sufficiently emerge from the latter’s moral status. Also relevant is the relationship between the duty bearers and the bearers of the rights. This consideration sets limits on the conceptional continuity between human and animal rights. Human rights are not only moral ‘possessions’ of individual human beings, but they also regulate social relations and should enable social participation and joint action.As far as possible, all people would have to have, first, and fundamentally, institutionally secured access to as many human rights goods as possible [25]. Human rights therefore require more from us than just the renunciation of harmful behavior by moral actors, effective protection against such behavior, and reparation or compensation after such behavior. They require us to respond effectively, comprehensively, and even proactively to whatever threats to basic goods may arise. Consequently, we encounter human rights obligations on two different levels: on a directly goods-related level involving respect, protection, and assistance (if possible, for self-help); on a logically higher level as obligations to form or strengthen institutions in order to be able to fulfill obligations on the first level as effectively as possible.To be sure, we also need institutions to prevent people from hunting animals in the wild or harming them in other ways, such as by destroying the environment. Animal rights then form a protective belt to shield animal habitats and life forms against the negative consequences of human intrusion and encroachment. However, humans are not the only beings and forces that threaten animals’ basic interests. Wild animals are also at risk of starvation, thirst, freezing, or being killed by other animals regardless of the consequences of human activities. If we wanted to effectively protect them from all these dangers, we would have to cover their habitats with institutions that would turn the wilderness into a kind of giant wildlife park. To be sure, virtually every wilderness today is already heavily infused and shaped by human activities [13]; and to the extent that we have harmed wildlife as a result, we owe them reparation or compensation [26]. However, compensation cannot reasonably consist of moralizing and transforming the last remnants of wilderness according to human conceptions. Insofar as wild animals are to remain wild, the human rights ideal of a humane and justly regulated world cannot apply to their habitats.A second, related difference is that human rights, beyond their defensive and protective function, should secure the conditions of effective inclusion in political and social communities. They are meant to enable cooperative, solidary, and caring relationships among free and equal citizens. Some political philosophers even consider the membership function of human rights to be paramount [27]. However, once again, this finds no counterpart in our relationship with wild animals, as envisioned by most animal rights advocates. The prevailing view is that we should leave wild animals alone as far as possible and reasonable. Rights then have the primary function of protection from human encroachment [28]. At most, limited and predominantly interactional duties to help in emergencies are compatible with them.This does not mean, however, that our relationship with wild animals must be paradigmatic for our relationship with animals in general. Representatives of contextualist [26] and political [5,21,29] theories of animal rights emphasize that we have brought many animals directly into our ways of life through breeding and husbandry. Many domesticated animals, in particular, could not stay alive, or at least not have good lives, without regular human attention. A possible alternative to letting such animals slowly but surely die out would be to recognize them as members of our communities. Their rights would then also acquire a membership function. It would include social, and in the case of working animals, also labor rights [30].Anyone who keeps animals thus has extensive support obligations towards them. The animals have the right to active attention to their welfare through appropriate housing, nutrition, supervision, and care. In addition, we must all ensure, through our state, that the animal keepers fulfill their guarantor obligations. One might spontaneously find the idea of compulsory health insurance for domestic pigs or state-guaranteed pension insurance for guide dogs laughable. However, the only normatively acceptable alternative would be not to keep and use such animals as companions or cooperation partners in the first place. Once such animals are there, we cannot excuse ourselves by insisting that they should see to it themselves how they might cope in case of illness or in old age. After all, we have forced them, directly or through our legislation, into circumstances in which they could hardly care for themselves.As citizens, we are responsible for the laws that permit the keeping and use of animals. This, in turn, speaks to a socially expanded understanding of the common good. We should make political decisions with the awareness that the collective for which they claim binding force includes humans and animals. For such a species-mixed community, what Ronald Dworkin understands as the sovereign virtue of political morality must apply: no state is legitimate that does not pay fundamentally equal consideration to the fortunes of all its members [31].Although Dworkin limits this basic norm to fellow human citizens, we should apply it to all animals we subject through our institutional orders. We owe these animals the fulfillment of duties of social justice. The basic social order regulated by law and armed with coercion must be acceptable from the perspective of really all individuals who belong to it or are existentially dependent upon it. Each individual must be able to reconcile the norms that regulate coexistence with their fundamental claim to live their own life, according to their basic needs and capabilities. This fundamental claim is also possessed by animals that cannot take a stand on moral and political norms themselves. Every norm that concerns them must also be acceptable from their perspective. This follows from the extension of an all-subjected principle to human–animal relations.I therefore propose a political membership status for all animals whose living conditions we comprehensively control and who could not live well or at all without regular human attention. It would include the right to live on a state territory and to return to it. Likewise, the interests of animals should receive fundamentally equal consideration in determining the common good [21]. To this end, animals require representation by human proxies in the political process. The consideration of animal interests is a cross-cutting political task that should be anchored at the macro-level of political decision-making.However, how can we know which of the animals’ interests their human representatives should advocate for? The general answer is that animals could give us valuable clues about their needs and preferences through their own activity [21]. Many animals can communicate with us nonverbally. They can tell us what is important for their well-being. For this reason, animal members should have the right to become visible in our midst and to develop as unhindered as possible and feasible. For the same reasons, animals should also be allowed to express their own will in working relationships. We must first respect the fact that they have a will that we may influence, through offers and incentives, but which we must not break. Secondly, we must learn to interpret the animals’ expressions of will with a view to possible consequences for their welfare. Thirdly, we must offer them possibilities of evasion and refusal without penalty. Subjecting animals to an alien purpose against their stable will is a tyrannical form of exploitation.7. ConclusionsThe moral status not only of humans but also of animals consists in an egalitarian right to have rights. From this equal status, however, substantially equal rights follow only if the morally relevant interests are equal. We do not need to resort to the argument from human marginal cases for this purpose. Already a reasonably broad and differentiated understanding of our own, human animal nature reveals which interests we share with many other animals. These, too, possess interests in the dimensions of life, well-being, and volitional activities. By contrast, at least the vast majority of animals have no interest in autonomy and self-respect.I consider justly regulated lasting relationships with some animals to be possible and also desirable. However, by subjecting animals and, thus, making them existentially dependent on us, we owe them more than just protection and help in easily remediable emergencies. We thereby also assume associative duties, as they exist among fellow citizens. Therefore, we should open our understanding of the common good to the reality of species-mixed communities and represent animals politically. However, we should essentially let wild animals be wild, even if this means that we restrain ourselves in our capacity to intervene.

animals : an open access journal from mdpi

10.3390/ani11041062

PMC8069219

Evaluation of Fogera cattle breed during the different seasons would help to determine their response to heat stress. This would assist in designing intervention strategies for the anticipated climate change. Therefore, this study aimed to determine physiological, hematological, biochemical, and growth parameters of Fogera cattle calves to heat stress in different seasons. We observed that the heart rate and respiration rate were increased by six beats per minute and four breaths per minute when the temperature-humidity index increased from 66 to 78, respectively. The positive relationship between the temperature–humidity index and physiological parameters further confirms that calves were good at thermoregulation at a temperature–humidity index value of 66. Thus, the temperature-humidity index value of 66 can be considered as optimum for high weight gain and normal physiological response to heat stress in Fogera cattle calves under their current production system. However, some more amelioration strategies such as better nutrition, availability of shade, and routine health management practice will further strengthen the resilience of the breed to heat stress in the future.

Fogera cattle are among indigenous breeds of cattle in the northern part of Ethiopia. However, their response to heat stress (HS) under different seasonal variations has not been well investigated. This study was aimed to determine physiological, hematological, biochemical, and growth parameters of Fogera cattle calves to HS during dry season, short rainy, and long rainy. A total of 72 calves (24 for each season) that were 6 months of age with an equal number of males and females were evaluated for physiological, hematological, biochemical, and growth parameters. Daily ambient temperature (AT) and relative humidity (RH) were recorded two times per day during the study periods from which the daily average temperature–humidity index (THI) was calculated. The study revealed higher AT and THI during dry and short seasons while higher RH was observed during the long rainy season. Physiological parameters except rectal temperature were affected by the seasons. Hematological parameters were also affected by season except for packed cell volume. Biochemical and growth parameters were also significantly affected by the seasons. THI was positively related with physiological but negatively with growth parameters. Thus, the THI value of 66 can be considered as optimum for high weight gain and normal physiological response to HS in Fogera cattle calves under their current production system.

1. IntroductionCattle are generally the livestock species most susceptible to water and nutritional stresses engendered by climate change [1]. High ambient temperature (AT) in combination with relative humidity (RH) compromises the animals’ ability to lose heat to the surroundings resulting in heat stress (HS). HS lowers the feed intake of an animal, which reduces their productivity in terms of milk yield, body weight, and reproductive performance. Besides, HS exerts a negative effect on the dry matter intake (DMI) and growth performance of calves and heifers. It was reported that dairy calves born in summer tended to have a lower average daily gain (ADG) than those born in winter [2].As calves consume agiven volume of milk or milk replacer and starter ad libitum daily, the main effect of HS on DMI for calves might lie in the starter [3]. Baccari et al. [4] reported that lower feed intake, ADG, and feed efficiency of Holstein heifers under HS conditions (32.5–34 °C environment) compared with cooler conditions (18–20 °C environment). Nonaka et al. [5] also found that daily dry matter intake and ADG of prepubertal Holstein heifers at 33 °C environment dropped by 9% and 22%, respectively, compared to those raised at 28 °C environment. However, water intake increased by 23% due to additional evaporative water loss, such as sweating. Neuwirth et al. [6] reported that bull calves responded with higher heart rate (HR), arterial pressure, skin temperature, plasma cortisol, and thyroxine concentrations to acute HS only above 32.2 °C at 60% relative humidity, which corresponds to a THI of 80.6.The temperature–humidity index (THI) is the widely used index to measure the magnitude of HS in animals [7]. However, some studies show that THI values only serve as a rough measure of HS effect on production [8]; they call for necessary adjustments because the environmental stimulus includes other factors such as wind speed and solar radiation [9]. Moreover, the THI threshold for calves and heifers remains unknown because of the very limited information available related to THI and HS on calves. Therefore, more studies will help to quantify THI for calves and even explore new indices to indicate the level of HS [3]. Under HS, the animals exhibit various behavioral, physiological, hematological, biochemicals, and endocrine adjustments to reduce stress inflicted. HS can be assessed by changes in physiological, hematological, and biochemical responses. These are the values that help in determining the adaptation of the animal to the existing environment [10,11,12]. Metekel ranch is located in northwestern Ethiopia characterized as a sub humid environment, meaning very wet and dry seasons [13]. Thus, the young stock population existing in this ranch was expected to undergo HS, particularly during the dry season. Therefore, the hypothesis of the proposed study was based on the assumption that seasonal changes in AT and RH induce heat stress in animals, which may be reflected by alteration in physiological, hematological, biochemical, and growth responses.Fogera cattle breed is among the indigenous breeds of cattle in the northern part of Ethiopia and has been raised at Metekel ranch since 1986 for breed improvement and conservation. Tesfa et al. [14] reported that the current population size of the breed is about 55,646. However, it is declining through time due to its wide area distribution in the region and genetic admixture. The same author also reported that the overall birth and weaning weight of Fogera cattle breed at Andassa livestock research center (ALRC) was 21.4 ± 0.09 kg and 102.2 ± 0.77 kg, respectively. Similar to livestock agriculture in our world, Fogera cattle breed is expected to face climate change and variability in the future. Hence, it is essential to evaluate HS effect on the breed, particularly, as young stock calves are more susceptible to HS than adult ones. The study showed that the thermo neutral zone (TNZ) of a 1-month-old calf is between 13 and 25 °C and the TNZ of a heifer with 0.8 kg daily gain is between 0 and 15 °C [15].Moreover, the mean rectal temperature(RT), respiration, pulse, and heart rate (per minute) were significantly higher in young animals than in adult cows and bulls [16]. However, the response of the Fogera cattle calves to the HS under different seasonal variations has not been well investigated. Therefore, this study was aimed to determine the degree of physiological, hematological, biochemical, and growth responses of Fogera cattle calves during different seasons.2. Materials and Methods2.1. Description of the Study AreaThe study was carried out in Metekel ranch, located in Guangua district of Awi zone in Amhara National Regional State and is situated at about 505 km North-west of Addis Ababa, 200 km from the regional town Bahir Dar. Its altitude ranges from 1500 to 1680 m above sea level (m.a.s.l). Metekel ranch is located at 10°57’6.5232” N latitude and 36°30’45.0864” E longitude.Theranch was established in 1986 for the Fogera cattle conservation and improvement program. The vegetation is mostly composed of perennial and annual grasses and a few scattered trees. According to Melak [17], Cynodondactylon, Digitariaternata, Digitariavelutina, Brachiariadictyoneura, Commelinagenghalensis, and Panicum maximum were the most palatable grass species while Schinusmolle, Casuarina equisetifolia, and Eucaulyptus camaldulensis were the predominant tree species. The annual mean RH is 61.7% and it reaches peak from June to October (76.7–83.8%). The ranch receives an average annual rainfall of 1730 mm and the average temperature ranges from 13.7 to 29.5 °C [18].Rainfall distribution is bimodal. According to Ababa [19], the study area has three seasons classified as the dry season (October–January), short rainy season (February–May), and long rainy season (June–September).2.2. Experimental Design and Animal ManagementA total of 72 calves, 6 months of age were randomly selected and used in this study. The calves were grouped into three of 24 animals with an equal number of males and females. Then, they were subjected to three different seasons such as dry season (January), short rainy season (April), and the long rainy season (July) for 30 days each. These months were selected since they are assumed to be representatives of the three seasons in the study area. The calves were fed as per the National Research Council [20] feeding standards with concentrate supplement in the simple shade. Additionally, they were also allowed to graze from 09:30 to 12:30 hours throughout the study period [21]. About 55, 26, 17, 1, and 1% maize, nougseedcake, wheat bran, salt, and ruminant premixes (minerals and vitamins) mixtures, respectively, were used as concentrate supplements. Before the study’s commencement, calves were treated with Ivermectin (Hebei Hope Harmony Pharmaceutical Co., Ltd., Hebei Province, China) and Albendazole (Ashish life science Pvt., Ltd., Maharashtra, India) for external and internal parasite treatments, respectively. All animal handling practices were followed the international guiding principles listed by the council for international organizations of medical sciences and the international council for laboratory animal science [22]. Meteorological parameters such as AT (°C) and RH (%) were collected two times per day, during morning and afternoon to get maximum and minimum values inall the experimental periods using a Hygrothermometer. Then, average daily THI was calculated from AT and RH using the equation; THI = (1.8 × AT + 32) − [(0.55 − 0.0055 × RH) × (1.8 × AT − 26)] developed by the National Research Council [20] for ruminant animals. Where, THI= temperature–humidity index, AT = ambient temperature (°C), and RH = relative humidity (%).This equation was used for calves in this study because the THI threshold for calves and heifers remains unknown due to minimal information about THI and HS [3].For the same reason, the classification reported by Habeeb et al. [23] was also adopted to quantify the intensity of HS in calves. Thus, calves were assumed to be in comfort zone if (THI < 68), mild discomfort zone if (68 < THI < 72), discomfort if (72 < THI <75), alert (75 < THI < 79), danger (79 < THI < 84), and emergency (THI > 84).2.3. Data Collection2.3.1. Physiological ParametersRectal temperature (RT) was measured using a clinical veterinary thermometer inserted at the animal’s rectum wall at a depth of approximately 3.0 inch for 3 min. Heart rate (HR) expressed in the number of beats per minute and respiratory rate (RR) expressed in breaths per minute were measured using a stethoscope and a stopwatch for the 30 s and multiplying the results by two to obtain the counts per minute.2.3.2. Blood Sample Collection and AnalysisIn each season, three times, blood samples were collected for each hematological and biochemical analysis on days 1, 15, and 30 following the procedure of Nikhil et al. [21]. A total of 432 blood samples were collected from the jugular vein of calves. Out of these, about 216 blood samples were collected in tubes containing ethylene diamine tetraacetic acid (EDTA)as an anticoagulant for hematological analysis while 216 blood samples were collected in serum separator tubes(SST)for biochemical analysis. Hematological study was performed within 24 h after blood collection. Whole blood was collected between 10:00 and 11:00 h. Then, the collected blood sample was analyzed by Hematology analyzer Germany version 2.5 to determine the value of white blood cell (WBC), red blood cell (RBC), hemoglobin (HGB), packed cell volume (PCV), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), and mean corpuscular hemoglobin concentration (MCHC).The plasma was separated from 216 blood samples by centrifugation at 3500 rpm for 10 min at 24 °C and stored at 4 °C for estimation of biochemical parameters until further analysis. Moreover, the blood serum analysis method is shown in Table 1.2.3.3. Growth ParametersThe calves were weighed at the beginning of the study and every 15 days thereafter using the weighing balance in each season for 30 days. All measurements were taken after overnight withdrawal of feed and water. Total gain (TG) was calculated as the difference between final body weight (FBW) and initial body weight (IBW). Then, average daily gain (ADG) was determined by dividing the differences of the FBW and IBW by the number of experimental periods (30 days for each season).2.4. Statistical AnalysisThe average of days 1, 15, and 30 in each season for all parameters were obtained and comparisons were made among seasons [21]. The data obtained on various parameters were statistically analyzed using a two-way analysis of variance. The data were analyzed by the general linear model (PROC GLM) using SAS software version 9.4. Sex of the calves and season were fitted as independent variables while physiological, hematological, biochemical, and growth parameters were fitted as response variables. When the GLM showed the presence of a significant difference between the parameters, the Tukey Kramer test was used for mean comparison. Moreover, the Pearson correlation test was performed to check the strength of the relationship between physiological parameters and THI as well as growth parameters and THI.The general linear model used for the analysis of physiological parameters was (1)Y=μ+Si+Kj+(S × K)ij+eijk where; Yijk = the response variables (HR, RR, RT); µ = overall mean; Si = effect of sex (female, male); kj = effect of season (dry season, short rainy season, long rainy season); (S × K)ij = interaction between sex and season; eijk = random error.The general linear model used for the analysis of hematological parameters was (2)Y=μ+Si+Kj+(S × K)ij+eijk where; Yijk = the response variables (WBC, RBC, HGB, PCV, MCV, MCH, MCHC); µ = overall mean; Si = effect of sex (female, male); kj = effect of season (dry season, short rainy season, long rainy season); (S × K)ij = interaction between sex and season; eijk = random error.The general linear model used for the analysis of biochemical parameters was (3)Y=μ+Si+Kj+(S × K)ij+eijk where; Yijk = the response variables (total protein, urea, creatine, glucose, total cholesterol); µ = overall mean; Si= effect of sex (female, male); kj = effect of season (dry season, short rainy season, long rainy season); (S × K)ij = interaction between sex and season; eijk = random error.The general linear model used for the analysis of growth parameters was (4)Y=μ+Si+Kj+(S × K)ij+eijk where; Yijk = the response variables (IBW, FBW, TG, ADG); µ = overall mean; Si = effect of sex (female, male); kj = effect of season (dry season, short rainy season, long rainy season); (S × K)ij= interaction between sex and season; eijk = random error.3. Results3.1. Meteorological Variables during the Study PeriodsHigher daily mean AT was observed during the dry and short rainy seasons compared to the long rainy season (Table 2) (p < 0.05). Moreover, a higher mean daily THI was noticed during the dry and short rainy seasons (p < 0.05). Whereas, higher mean daily RH was observed during the long rainy season (p < 0.05). The maximum AT in dry and short rainy seasons was greater than the long rainy seasons. However, there was no statistically significant difference among the three seasons in minimum AT during the experimental periods. On the other hand, maximum RH increased by 40.95 and 47% during the long rainy season over the dry and short rainy seasons, respectively. Moreover, the minimum RH increased by 49.18 and 51.71% during the long rainy season over the dry and short rainy seasons, respectively.3.2. Physiological ParametersPhysiological parameters of calves were significantly affected by season except for RT (Table 3). HR was increased by 5.7 and 4.1 beats per minute during the dry season over long and short rainy seasons, respectively. Moreover, RR was increased by 3.8 and 3.4 breaths per minute during the dry season over the long and short rainy seasons, respectively. The calves’ HR was significantly affected by sex (Table 3) (p = 0.007). Moreover, RR of the calves were significantly affected by the sex of the calves (p = 0.046). Male calves had consistently higher HR and RR than female calves. However, RT of the calves were not significantly affected by the sex of the calves (p = 0.165). On the other hand, physiological parameters were affected by season and sex interaction effect except for RT of the calves.3.3. Hematological ParametersHematological parameters were significantly affected by season variation except for PCV (Table 4). A significantly lower WBC value was observed in the dry season compared to short and long rainy seasons (p = 0.011). Moreover, considerably lower values of HGB were observed in the dry and long rainy seasons than short rainy seasons (p = 0.000). Appreciably, the concentration of RBC was higher in the short rainy season followed by dry and long rainy seasons, respectively (p = 0.000).The concentration of MCV was significantly affected by the variation in season (p = 0.000). Lower mean MCV was observed in the short rainy season compared to dry and long rainy seasons. MCH was significantly lower during the short rainy season than dry and long rainy seasons. Considerably, lower MCHC values were observed in the dry season than short and long rainy seasons (p = 0.000). The interaction effect of sex and season did not affect all hematological parameters (Table 4). Almost all of the hematological parameters were not significantly affected by the sex of the calves except PCV (Table 4) (Figure 1).It was noticed that higher PCV was observed in females than male calves (Figure 1) (p = 0.028).3.4. Biochemical ParametersThe least-square mean value of total protein, cholesterol, and glucose concentration was lower in the dry season than short and long rainy seasons (Table 5) (p = 0.000). Conversely, creatine, and urea concentration were higher in dry and short rainy seasons compared to the long rainy season (p = 0.000). Our result indicated that most of the biochemical parameters were not affected by sex and season interaction effect except for creatine, and urea (Table 5). Significantly, female calves had higher creatine during the short rainy season than males. However, there was no significant difference in creatine during the dry and long rainy seasons. Female calves had higher blood urea than males in the short rainy seasons, but male calves had higher blood urea than females during long rainy seasons.3.5. Growth ParametersAll growth parameters considered in our study were significantly affected by the seasons (Table 6). In the long and short rainy seasons, IBW was increased by 15.8 and 14.5 kg compared to the dry season, respectively. Moreover, in the long and short rainy seasons, the FBW was increased by 22 and 10.6 kg compared to the dry season, respectively. Total gain (TG) was increased by 10.3 and 4 kg in the long and dry seasons, respectively, over the short rainy season. Similarly, ADG was increased by 300 and 100 g/day in the long and dry seasons, respectively, over the short rainy season. All growth parameters were not significantly affected by the sex of the calves. However, TG and ADG were significantly affected by sex and season interaction. TG of female calves was higher in dry and short rainy seasons whereas male calves were 9 kg higher than females in the long rainy season. Moreover, significantly higher ADG was observed in females in the dry and short rainy seasons while higher ADG was observed in males in the long rainy season (Table 6).3.6. Relationship between THI and Physiological ParametersThe study revealed that THI was positively correlated with HR and RR. The HR increased by six beats per minute and RR increased by four breaths per minute when the THI increased from 66 to 78 (Figure 2). Moreover, the Pearson correlation test showed that there was a relationship between THI and HR (r = 0.395, p = 0.001) and THI and RR (r = 0.331, p = 0.005). The result also showed a positive relationship between HR and RR as illustrated in Figure 2. Thus, from this study, it can be noticed that the THI value of 66 is a critical value for all physiological parameters considered in this particular breed under the current situation. Up to this critical value, animals could perform well since there was a minimal effect of HS. However, as THI advances calves may lose their performance to maintain their constant body temperature.3.7. Relationship between THI and Growth ParametersGrowth parameters such as IBW, FBW, and TG were negatively associated with the increase in THI (Figure 3). The Pearson correlation test also confirmed that there was a negative relationship between THI and IBW (r = −0.274, p = 0.02); THI and FBW (r = −0.397, p = 0.001) (Figure 3a,b). Furthermore, there was a negative relationship between THI and TG (r = −0.205, p = 0.042) as indicated in Figure 3c. From Figure 3a–c, it can be generalized that all the growth parameters were declining after the THI value of 66. However, up to this value, all the growth parameters were slightly increasing. This result further justifies that THI plays a significant role in determining young calves’ performance, especially during dry and short rainy seasons. Thus, greater attention should be given to environmental factors such as AT and RH when designing any management interventions for improving their performance.4. Discussions4.1. Meteorological Variables during the Study PeriodsThe AT and RH were the commonly used measure of HS in animals until THI was derived [21,24]. THI accounts for AT and RH’s combined effect and is considered one of the best methods to evaluate HS in animals [25]. However, previous studies showed that THI values only serve as a rough measure of HS effect on production [8]; they call for necessary adjustments because the environmental stimulus includes other factors such as wind speed and solar radiation [9]. The result showed that AT values were higher than the critical values reported by Hahn [15] for TNZ for a heifer with 0.8 kg daily gain in all the seasons while AT values were within TNZfor1-month-old calves reported by the same author during the long rainy season. Conversely, RH values in dry and short rainy seasons were lower than the critical values reported by Neuwirth et al. [6] for bull calves whereas higher than the report of Neuwirth et al. [6] during the long rainy season. Based on the assumption by Habeeb et al. [23] calves were in a discomfort zone during the dry and short rainy seasons while in mild discomfort during the long rainy season. However, the THI value was relatively lower during all seasons than reported by Neuwirth et al. [6] for bull calves. This difference might be due to breed, location, and management practices. Therefore, it can be concluded that under the current situation the overall meteorological data indicated that the animals were heat-stressed in all three seasons as confirmed by upper critical AT, RH, and THI. However, the impact of HS was high in the dry season and short rainy seasons compared to the long rainy season evidenced by eminent AT and THI in the dry season and short rainy season compared to the long rainy season. This discomfort zone for calves observed during the dry season and short rainy seasons in this study indicated the need to implement further amelioration strategies in the anticipated climate change in the future.4.2. Physiological ParametersNo difference in RT under different seasons indicates a better adaptive capacity of Fogera cattle calves to their current production environment under different conditions. The HR of the calves was significantly affected by seasons (p = 0.0001). This might be because there was high AT and THI during the dry season compared to both short and long rainy seasons. The mean RR noticed in the dry season was comparable with Hariana breed’s value during the summer [26]. Moreover, the mean RR observed in the short and long rainy seasons was higher than the value reported for Hariana, and Sahiwal breeds, respectively [26]. Thus, it can be suggested that Fogera cattle calves are good at heat tolerance under tropical climate of THI less than or equal to 66.An increase in RR was able to eliminate excess heat and maintain a constant deep body temperature [27]. It is generally true that in matured animals due to physiological differences, males are better at thermoregulation than females. Likewise, our finding also confirmed that higher HR and RR were observed in male calves than females, showing that male calves are relatively better at heat tolerance than females. These differences might be emanated from increased HR and RR in male calves at elevated environmental AT and RH. Neuwirth et al. [6] also reported that bull calves responded with HR, arterial pressure, skin temperature, plasma cortisol, and thyroxine concentrations to acute HS only above 32.2 °C at 60% RH which is consistent with our findings. In contrary to our findings, it has been reported that heifers experience an increased heart rate during HS [28,29].This helps maintain blood pressure as a response to the elevated vasodilatation and increased blood flow caused by HS [28]. Another study also showed that heifers were reported to deposit more fat than steers [30], which might affect their thermoregulation at higher AT, RH and THI than young bulls.4.3. Hematological ParametersSignificantly higher PCV values were noticed in females than male calves. The difference in PCV between females and male calves in this study might be due to sex differences which are further linked with differences in cortisol hormone production in response to HS. Contrary to this result, Kubkomawa et al. [31] reported males having higher mean PCV than female cattle.However, there was no significant difference in WBC, RBC, HGB, MCV, MCH, and MCHC between male and female calves. According to Celik et al. [32], HGB and MCH levels were not statistically different between the sex of the calves, which is concordant with our findings. Moreover, Camargo et al. [33] also did not find a significant difference between males and females in MCHC, which is also in line with our result.The lower WBC counts noticed in the dry season might be due to the body system’s response to stress stimuli compared to the long rainy season. During the study periods, irrespective of the seasons, calves were given similar routine health management such as deworming and continuous follow-up of the health condition. Moreover, they were given feed with good nutrient content in addition to shade from the extreme sunlight. This suggested that more amelioration strategies need to be devised for calves to face severe climate change anticipated in the future. Likewise, Mirzadeh et al. [34] conducted a study and reported lower WBC in summer as compared to spring, autumn, and winter seasons in all age groups of Iranian cattle which is in line with our result.The mean value of total RBC was significantly lower in the long rainy season than short and dry seasons (p = 0.000). Moreover, Naik et al. [16] did a similar study and found significantly higher RBC values in summer compared to winter and monsoon in different age groups in Punganur cattle. Aengwanich et al. [35] did not find any significant changes in RBC among summer, rainy, and winter seasons in beef cattle. The higher number of RBC in dry and short rainy seasons might be associated with relatively higher stress in the body, requiring more oxygen transport throughout the body. Furthermore, the lower RBC during the long rainy season could be due to increased water intake through the lush grasses that were available in that season [36] or high RH observed in the long rainy as compared to other seasons would have compromised evaporative heat loss mechanisms resulting in HS and, therefore, animals would have ingested more water and subsequent hemodilution and hence decreased RBC [37].The lower HGB observed in the dry and long rainy seasons may be because of smaller RBC counts observed in these seasons (Table 4).The lower MCV and MCH in the short rainy season and MCHC in the dry season in this study might be due to a low amount of HGB present per RBC. However, the effect of season was not significantly affected PCV (p > 0.05). Similar findings were reported by Aengwanichet al. [35] who reported that season did not affect hematological parameters of heifers in the northeastern part of Thailand. Contrary to this finding, significantly lower PCV was noticed in summer as compared to winter in all age groups of cattle [34]. Despite upper critical values of AT, RH, and THI in all three seasons and the absence of variation in PCV indicates once again the adaptive potential of Fogera cattle calves to their current production environment.4.4. Biochemical ParametersThe lower total protein observed in this study might be due to the higher AT and THI observed in the dry season. Similarly, Dar et al. [38] reported significantly lower serum protein levels during the summer season in >1 year of age than in the winter season in Badri cattle. Contrary to this result, Nikhil et al. [21] did not find a significant change in total protein concentration among summer, rainy, and winter seasons in crossbred female calves of 6–12 months age in hot and humid tropics. Our study showed that total cholesterol concentration was higher in the long rainy season followed by short rainy and dry seasons (p = 0.000). Farooq et al. [39] also showed significantly lower cholesterol during the hot season compared to cool dry winter in adult Cholistani bulls. Considerably, low plasma cholesterol concentration during dry, and short rainy seasons may be due to HS-induced accelerated fat catabolism [40] or increased lipid mobilization by peripheral tissues [41] or may be due to reduced liver capacity [42] under HS in these seasons.The decreased glucose concentration in dry and short rainy seasons compared to long rainy seasons could be due to more energy requirement to the animal to disperse more heat to maintain the body temperature at a normal level or could be due to heat stress-induced depressed dry matter intake [43]. Similarly, Nikhil et al. [21] also reported significantly lower glucose in pre-monsoon followed by monsoon and post-monsoon in crossbred female calves in hot and humid tropics. The higher serum creatine concentration during dry and short rainy seasons might be because of excess muscular catabolism for energy supply as voluntary feed intake is reduced due to these seasons [39]. The serum creatine concentration was smaller than values reported for Gir, Sahiwal, Kangayam, and Tharparkar but higher than Rathi cattle. However, the result is comparable with DeoniZebu cattle breeds of India during early summer [44]. A higher concentration of creatine during summer stress was also reported by Dar et al. [38].Blood urea concentration was significantly higher during the dry season, followed by short and long rainy seasons (Table 5). Likewise, a higher blood urea concentration was reported during summer compared to winter in >1 year age of Badri cattle [38]. Our results were also in agreement with Rasouliet al. [11], who reported higher blood urea nitrogen during the summer season. This increase may be due to the utilization of amino acids for energy. The other reason may be due to protein mobilization from muscle tissue and stress-related cortisol elevation, which increases catabolism of body proteins [38].4.5. Growth ParametersThe increase in IBW and FBW of calves in the long and short rainy season was attributed to natural pasture availability and decrease in maximum temperature compared to the dry season. The increment in IBW and FBW of calves further attributed to the increase in TG by 10.3 kg and ADG 300 g/day in the long rainy season as compared to dry and short rainy seasons. Despite the change in IBW and FBW in different seasons in the current study, the result was slightly smaller than the value reported by Tesfa et al. [14] in the dry and long rainy seasons, respectively, for the same breed at ALRC. This difference might be due to the difference in availability of forage at these locations and environmental temperature which further inhibits feed intake and growth of calves at Metekel ranch as it is hotter than ALRC.A previous study conducted by Pereira et al. [45] also showed that feed intake was reduced by 11.4% in Limousine calves under thermal stressful conditions, consistent with our finding. However, the ADG in this result was slightly higher than reported for the same breed in dry and long rainy seasons [14], respectively. The difference could be because of the short time experiment and the small number of observations used in this study. The increase in TG and ADG in females over the males during dry and short rainy seasons may be due to female calves’ special care rendered to them as a replacement and distribution to the surrounding farmers. However, all growth parameters were not significantly affected by sex differences (Table 6). The result is in line with Mekuriaw et al. [46] for Ogaden breed and [14,47] for Fogera cattle breed.4.6. Relationship between THI and Physiological ParametersAs THI increased from 66 to 78, the HR of the calves was increasing from 80 to 90 beats per minute (Figure 2) (p < 0.05). Moreover, there was an increase in RR of the calves as THI advances from 66 to 78. This might be due to the increase in THI affecting internal physiological activities such as the calves’ metabolic rate, further increasing the HR and RR to respond against HS. The current result is in line with the findings of Aziz et al. [24] who reported that THI exhibited a positive correlation with the respiratory rate and pulse rate in crossbred cattle calves in Indian. Constant RT and increased HR and RR when calves are exposed to a THI from 66 to 78, indicated that the calves had a high heat tolerance level. They maintained thermal balance by removing excess heat through enhanced respiratory rate and pulse rate [24]. Moreover, the positive relationship between physiological parameters (HR and RR) justifies that calves were well adaptive to their production environment by maintaining their homeostasis through elevated HR and RR. Bianca [48] also found a positive correlation between HR and RR at severe HS in Ayrshire bull calves.4.7. Relationship between THI and Growth ParametersThe decline in the growth parameters might be because calves dissipate heat from their body to adapt to the increase in THI, especially during the dry season. Moreover, the decrease in growth rate with the increase in THI might be because animals used most of the energy to maintain their homeothermy instead of growth [24]. Therefore, to maintain the growth rate and sustain this breed’s adaptability under increased THI, appropriate feeding, shade provision, and frequent watering may be suggested for the coming climate change. A similar study conducted in Japanese black calves indicated that body weight gain with THI of >75 was significantly lower than that with THI ranging from 56 to 60 [49]. Moreover, Broucek et al. [50] showed that calves under HS conditions (74.8 of THI) had reduced starter intake compared with those raised under moderate conditions (59.7 of THI).Colditz and Kellaway [51] also reported that heifers raised under HS condition (38 °C environment) had reduced feed intake and ADG compared to those maintained under cool ambient conditions (17 °C environment) which is corroborated with this result.5. Conclusions and RecommendationsOur study revealed that the existence of upper critical AT, RH, and THI in all three seasons indicated the HS reality in all the animals in the study area. There was a change in growth parameters such as IBW, FBW, TG, ADG, but no change in physiological (RT) and hematological (PCV) parameters with varying seasons. From the finding, it is possible to say animals adapted to the existing stress by similar physiological and hematological responses. There was also a positive relationship between THI and physiological parameters. This implies that as THI gets advanced, Fogera cattle calves adjust their body temperature by increasing their HR and RR. Thus, Fogera cattle calves are good at thermoregulation under the tropical climate of THI less or equal to 66. Therefore, the THI value of 66 can be considered optimum for high weight gain and normal physiological responses to HS in Fogera cattle calves under their current production system. However, some more amelioration strategies such as better nutrition, availability of shade, and implementation of routine health management practice should be considered for the resilience of the breed to HS in the future. Future research should be done in a similar environment by including different indigenous breeds to select the best-performing animals at higher THI values.

animals : an open access journal from mdpi

10.3390/ani13111811

PMC10251964

The quality of in vitro matured oocytes is usually lower than that of in vivo matured oocytes, possibly due to the absence of some important signal regulators in vitro. In this study, we found that progesterone and androstenedione were upregulated in FF during in vivo pig oocyte maturation. The supplementation of progesterone or androstenedione during in vitro maturation significantly improved the pig oocyte maturation rate or subsequent embryo developmental competence.

Oocytes matured in vitro are useful for assisted human and farm animal reproduction. However, the quality of in vitro matured oocytes is usually lower than that of in vivo matured oocytes, possibly due to the absence of some important signal regulators in vitro. In this study, untargeted metabolomics was used to detect the changes in the metabolites in the follicular fluid (FF) during in vivo pig oocyte maturation and in the culture medium during in vitro maturation. Our results showed that the total metabolite changing profile of the in vivo FF was different from that of the in vitro maturation medium, but the levels of 23 differentially expressed metabolites (DEMs) changed by following the same trend during both in vivo and in vitro pig oocyte maturation. These 23 metabolites may be important regulators of porcine oocyte maturation. We found that progesterone and androstenedione, two factors in the ovarian steroidogenesis pathway enriched from the DEMs, were upregulated in the FF during in vivo pig oocyte maturation. The levels of these two factors were 31 and 20 fold, respectively, and they were higher in the FF than in the culture medium at the oocyte mature stage. The supplementation of progesterone and androstenedione during in vitro maturation significantly improved the pig oocyte maturation rate and subsequent embryo developmental competence. Our finding suggests that a metabolic abnormality during in vitro pig oocyte maturation affects the quality of the matured oocytes. This study identified some important metabolites that regulate oocyte maturation and their developmental potential, which will be helpful to improve assisted animal and human reproduction.

1. IntroductionIn vitro matured oocytes play an important role in farm animal reproduction [1], human-assisted reproduction [2], and biological science [3] because they can be used to produce in vitro fertilization, parthenogenetic activation, and cloned embryos. However, the embryos generated with in vitro matured oocytes usually exhibit a lower developmental potential than those derived from in vivo matured oocytes [4,5,6,7]. This suggests that the quality of in vitro matured oocytes is inferior compared to that of their in vivo matured counterparts.The quality difference between in vitro and in vivo matured oocytes is mainly caused by the differences in the maturation environment between these two types of oocytes. Although the in vitro oocyte maturation condition has been improved in the past decades, it still cannot fully mimic the in vivo oocyte maturation condition. For example, the in vitro oocyte maturation medium lacks Stromal cell-derived factor 1 (SDF1) [8], growth differentiation factor 9 (GDF9) [9], bone morphogenetic protein 15 (BMP15) [9], neurotrophic factors [10], and melatonin [11], which all are present in in vivo follicular fluid (FF) and are critical for oocyte maturation quality.FF is an important microenvironment for in vivo matured oocytes. FF contains various regulatory molecules such as RNAs, proteins, metabolites, and hormones, which are transported from the blood or are secreted by follicular cells [12,13]. The level of many molecules in the FF is changed dynamically during in vivo oocyte maturation [14,15,16,17]. The abundance of some factors in the FF has been shown to be correlated with the quality or developmental ability of mature oocytes [18,19,20,21].In this study, we analyzed the FF metabolic profile of porcine in vivo mature and immature follicles and the culture medium metabolic profile of porcine in vitro mature and immature oocytes. After a metabolomics analysis, two metabolites enriched in the ovarian steroidogenesis pathway, including progesterone and androstenedione, were selected to further investigate their roles in pig oocyte maturation quality. We showed that the addition of progesterone and androstenedione into a pig oocyte in vitro maturation medium has positive effects on improving the oocyte maturation rate and subsequent embryo development.2. Materials and Methods2.1. ChemicalsUnless otherwise mentioned, all chemicals used in the experiment were procured from Sigma-Aldrich (Saint Louis, MO, USA).2.2. Collection of In Vivo Follicular Fluid and In Vitro Oocyte Culture MediumThe collection of the in vivo follicular fluid and in vitro oocyte culture medium were performed as previously described [5,8]. Briefly, Duroc sows (about 14 months old) were treated with altrenogest (Yofoto, Ningbo, China) at 20 mg per sow per day for 18 days, and then they were injected with 1000 IU of PMSG (Yofoto, Ningbo, China). Forty hours after the PMSG injection, estrous sows were anesthetized and their ovaries were exposed via surgery. The pig mature follicular fluid (pMFF) was collected from preovulary follicles (diameter > 15 mm) in the ovaries with an obvious ovulation point. The pig immature follicular fluid (pIFF) was collected from immature follicles (3 mm < diameter < 5 mm) of nonestrous 14-month-old Duroc sows via surgery. Eight pMFF samples and nine pIFF samples were collected from 8 estrous and 9 anestrous sows, respectively. Eight pig oocyte mature medium (pMM) samples were collected from the IVM medium of porcine COCs cultured for 44 h, centrifuged, and stored in liquid nitrogen; eight immature pig oocyte medium (pIM) samples were collected from the fresh IVM medium of porcine COCs, centrifuged, and stored in liquid nitrogen.2.3. Ultra-High-Performance Liquid Chromatography-Mass Spectrometry (UHPLC-MS)/MS AnalysisA total of 100 μL of the follicular fluid or mature medium samples was taken out from the liquid nitrogen and thawed slowly at 4 °C. In total, 400 μL of prechilled (−20 °C) 80% methanol was added. The mixed liquor was vortexed, incubated at −20 °C for 60 min, and then centrifuged at 14,000× g at 4 °C for 20 min. The supernatants were subsequently transferred to a fresh centrifuge tube. All the samples were mixed in equal volume as quality control samples to assess the stability of the samples over the entire experiment period before testing. The supernatants of the follicular fluid or medium were analyzed via LC-MS/MS. An LC-MS/MS analysis was completed by Beijing Novogene Technology Co., Ltd. (Beijing, China) using high performance liquid chromatograph (Vanquish UHPLC, Thermo Fisher, Waltham, MA, USA) and mass spectrometer (QE-HF-X, Thermo Fisher, Waltham, MA, USA).The peak alignment, peak selection, and quantification of each metabolite from the raw data file generated by UHPLC-MS/MS were obtained by using compound Discoverer 3.0 (CD 3.0, Thermo Fisher, Waltham, MA, USA). The molecular formula was predicted based on additive ions, molecular ion peaks, and fragment ions by using normalized data. For the metabolites that responded more strongly to positive or negative ions, we chose the mode with a stronger response as the detection result of this metabolite. Peak matching was performed by using mzCloud (“https://www.mzcloud.org/” accessed on 23 February 2019) and ChemSpider (“http://www.chemspider.com/” accessed on 23 February 2019) databases to obtain accurate qualitative and relative quantitative results. The data were imported into EZinfo software (version 2.0; Umetrics AB, Umeå, Sweden) for a principal component analysis (PCA). The first principal component of the PLS-DA model was used to obtain the variable importance in projection (VIP) value. The p values of the metabolites in the two groups were analyzed by using Student’s t-tests. Finally, the differentially expressed metabolites were identified by using the following criteria: p value < 0.05, VIP value >1 and fold change ≥2. The Kyoto encyclopedia of genes and genomes enrichment (KEGG) analysis was performed by using hypergeometric test to identify pathways that were significantly enriched in the differentially expressed metabolites compared with the background of the total identified metabolites.2.4. Preparation of In Vitro Matured OocytesThe collection of COCs was performed as previously described [8]. Briefly, pig COCs were obtained from the ovaries via aspiration by using an 18-gauge needle attached to a 10 mL disposable syringe. The COCs that met the criteria of having at least three layers of compact cumulus cells and intact cytoplasm were selected for IVM. Approximately 50~60 COCs were transferred to each well of a four-well Nunc dish into 500 μL of fresh IVM medium. The IVM medium was prepared by supplementing Medium-199 with 10 ng/mL epidermal growth factor (EGF), 40 ng/mL fibroblast growth factor 2 (FGF2) (PeproTech, Suzhou, China), 20 ng/mL leukemia inhibitory factor (LIF) (PeproTech, Suzhou, China), 20 ng/mL insulin-like growth factor (IGF) (PeproTech, Suzhou, China), 5 μg/mL inositol, an ITS-liquid medium supplement (1×), 10 IU/mL pregnant mare serum gonadotropin (PMSG) (Yofoto, Ningbo, China), 10 IU/mL luteinizing-hormone-releasing hormone A3 (Yofoto, Ningbo, China), and 0.6 mM cysteine, and it was cultured at 38.5 °C with 5% CO2 for 44 h in a humidified atmosphere. The COCs were incubated in Dulbecco’s phosphate buffered saline (DPBS) containing 0.1% polyvinyl alcohol (PVA) and 1 mg/mL hyaluronidase for 5 min, followed by gentle pipetting approximately 200 times to remove the surrounding cumulus cells. The oocytes were examined by using a stereomicroscope and were considered matured when the first polar body was visible in the perivitelline space. These oocytes were utilized for the preparation of cloning, in vitro fertilization (IVF), and parthenogenetic activation (PA) embryos. 2.5. Production of EmbryosThe production of PA, IVF, and cloned embryos were performed as previously described [8]. The production of PA embryos occurred as follows: the matured pig oocytes were electrically activated using two direct current pulses of 85 V/mm for 100 μs in 0.28 mol/L mannitol supplemented with 0.1 mM MgSO4 and 0.1% PVA. The production of IVF embryos occurred as follows: the capacitated sperm (final concentration: 1 × 105 sperm/mL) were added to the well containing the matured oocytes in a humidified incubator for six hours at 38.5 °C under 5% CO2. The production of cloned embryos occurred as follows: the nuclei of the matured oocytes were stained with 1 g/mL Hoechst 33,342, removed using a microinjection needle, and then a round donor cell that was slightly spiculated was selected and injected into the perivitelline space of the enucleated oocytes. Finally, the reconstructed oocytes were activated in 0.28 mol/L mannitol supplemented with 0.1 mM MgSO4 and 0.1% PVA by using two direct current pulses of 150 V/mm for 50 μs.All the embryos were washed three times in a porcine zygote medium-3 (PZM-3) medium and were then cultured in a PZM-3 medium at 38.5 °C under 5% CO2 and saturated humidity, and the cleavage and blastocyst rates were assessed two and six days after embryo production.2.6. Experimental DesignExperiment 1: Porcine COCs were cultured in the IVM medium supplemented with 100 μM progesterone. After a 44 h culture, the number of oocytes that excreted the first polar body was calculated under a stereoscope. The cleavage and blastocyst rates were assessed two and six days after PA and SCNT embryo production.Experiment 2: Porcine COCs were cultured in the IVM medium supplemented with 0 ng/mL, 5 ng/mL, 75 ng/mL, 125 ng/mL, and 250 ng/mL androstenedione. After the 44 h culture, the number of oocytes that excreted the first polar body was calculated under a stereoscope. The cleavage and blastocyst rates were assessed two and six days after embryo production.2.7. Statistical AnalysisThe data on the oocyte maturation rate and embryo development were analyzed with chi-square test. The data on the progesterone, androstenedione, and testosterone in the pIFF, pMFF, pIM, and pMM were compared between different groups by using an ANOVA with LSD by using IBM SPSS Statistics 27 software. Only p < 0.05 was considered statistically significant.3. Results3.1. Untargeted Metabolomics Analysis of Porcine In Vivo Follicular FluidThe principal component analysis (PCA) indicated significant differences in the metabolic profile between the pMFF and pIFF samples regarding their negative (Figure 1A) and positive (Figure 1C) ion models. In addition, permutation test showed that R2Y = 0.58, |Q2|= 0.75 in the negative ion model (Figure 1B) and R2Y = 0.59, |Q2| = 0.79 in the positive ion model (Figure 1D), which suggests that the ion models established in this study were reliable. Differentially expressed metabolites (DEMs) in the FF were screened based on the variable importance in projection (VIP) values > 1, p-value < 0.05 and |log2(fold change)| > 1. In total, 118 metabolites (42 upregulated and 76 downregulated) were identified in the negative ion model (Figure 2A) and 294 metabolites (98 upregulated and 196 downregulated) were identified in the positive ion model (Figure 2B).3.2. Untargeted Metabolomics Analysis of Porcine Oocyte In Vitro Culture MediumThe PCA indicated significant differences in the metabolic profile between the pMM and pIM samples regarding their negative (Figure 3A) and positive (Figure 3C) ion models. A permutation test showed that R2Y = 0.82, |Q2| = 0.69 in the negative ion model (Figure 3B) and R2Y = 0.86, |Q2|= 0.62 in the positive ion model (Figure 3D), which suggests that the models were reliable for subsequent analysis. In the porcine oocyte in vitro culture medium, 62 DEMs (31 upregulated and 32 downregulated) were identified in the negative ion model (Figure 4A) and 152 DEMs (63 upregulated and 89 downregulated) were identified in the positive ion model (Figure 4B).3.3. Metabolites with a Same Change Trend during In Vivo and In Vitro Pig Oocyte MaturationThe metabolites changed and followed the same trend during both in vivo and in vitro pig oocyte maturation, and this probably plays an important role in regulating the oocyte maturation quality. To identify these potential critical metabolites, a Venn analysis was conducted. The results showed that 4 and 19 metabolites were upregulated and downregulated, respectively, during both in vivo and in vitro oocyte maturation (Figure 5). These 23 possible important metabolites are shown in Table 1.3.4. Untargeted Metabolomics Analysis of Porcine In Vivo Follicular FluidTo investigate the functions of the identified DEMs, a Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis was performed. The DEMs in the follicular fluid during oocyte in vivo maturation were enriched in the metabolic pathways during the degradation of compounds, pyrimidine metabolism, and microbial metabolism in diverse environments (negative ion mode; Figure 5A). They were also enriched during ovarian steroidogenesis and pyrimidine metabolism; during alanine, aspartate, and glutamate metabolism; during the biosynthesis of unsaturated fatty acids; when in the prolactin signaling pathway; during endocrine resistance; during steroid degradation; and during galactose metabolism (positive ion model; Figure 5B).The DEMs in the culture medium during oocyte in vitro maturation were enriched in two metabolic pathways, including “stilbenoid, diaryheptanoid and gingerol biosynthesis” and “glucocorticoid and mineralocorticoid receptor agonists/antagonists” (negative ion model; Figure 5C). 3.5. Untargeted Metabolomics Analysis of Porcine In Vivo Follicular FluidAmong the four DEMs that were upregulated during both in vivo and in vitro pig oocyte maturation (Table 1), progesterone attracted our interest because it was also enriched in the KEGG pathway of ovarian steroidogenesis (Figure 6). This suggests that the ovarian steroidogenesis pathway may participate in regulating porcine oocyte maturation. The levels of the three DEMs enriched in the ovarian steroidogenesis pathway, including progesterone, androstenedione, and testosterone, are shown in Figure 7. Progesterone and androstenedione were significantly upregulated while testosterone was significantly downregulated in the FF during in vivo oocyte maturation (Figure 7). Although progesterone was also significantly upregulated in the culture medium during in vitro oocyte maturation, its level was about 31-fold higher in the pMFF than in the pMM (Figure 7A). The level of androstenedione in the pMFF was approximately 20-fold higher than that in the pMM (Figure 7B). This implies that the levels of progesterone and androstenedione in the pig oocyte in vitro culture medium were insufficient compared to those in the FF.Progesterone and androstenedione were added into the in vitro culture medium so we could study their effects on the pig oocyte maturation quality. As shown in Table 2, the supplementation of 100 μM of progesterone during IVM significantly increased the rate of oocytes extruding the first polar body (71.63% vs. 58.70%, p < 0.05). However, the supplementation of progesterone during IVM did not affect the subsequent development of PA embryos (Table 3) and cloned embryos (Table 4).The addition of 5, 25, 125, and 250 ng/mL androstenedione into the IVM medium did not affect the porcine oocyte maturation rate (Table 5). Supplementation with 125 ng/mL androstenedione significantly improved the blastocyst formation rate of PA embryos (50.11% vs. 34.72%, p < 0.05; Table 6). Supplementation with 125 ng/mL androstenedione during pig oocyte IVM significantly enhanced the cleavage rate (57.63% vs. 42.84%, p < 0.05) but had no significant effect on the blastocyst formation rate (8.65% vs. 9.86%, p > 0.05) of IVF embryos (Table 7).4. DiscussionWe demonstrated in this study that porcine ovarian follicular fluid contains elements that are beneficial for the oocyte maturation rate and developmental competence. Similar results were found in our previous studies [5,8]. We also showed that a metabolomics analysis is a powerful tool that can be used to find the molecules regulating the oocyte maturation quality in FF or in a maturation medium.Our data indicated that the levels of 412 metabolites were changed significantly in the FF during pig oocyte in vivo maturation. Nevertheless, during in vitro pig oocyte maturation, only 214 DEMs were detected in the culture medium. These results suggest that more metabolites in the FF are changed dynamically during in vivo oocyte maturation than in the culture medium during in vitro oocyte maturation. Four DEMs that were upregulated and nineteen DEMs that were downregulated during both in vivo and in vitro pig oocyte maturation was identified in the present study. These 23 metabolites may play important roles in pig oocyte maturation. Among them, at least six metabolites, including vitamin C [22,23], progesterone [24,25], 8-hydroxy-deoxyguanosine [26,27], putrescine [28], N-Acetyl-L-cysteine [28,29], and hypotaurine [30,31] have been reported to regulate oocyte maturation rate or quality. In the positive ion model, the KEGG pathways enriched from the DEMs in the FF were different from those enriched from the DEMs in the maturation medium. In addition, the DEMs in the FF detected by the negative ion model were enriched in eight KEGG pathways, yet no KEGG pathway was enriched from the DEMs in the maturation medium identified by the same ion model. These results suggest that the global metabolic profile between the in vivo ovarian FF and in vitro oocyte maturation medium is different. This difference probably mainly resulted from the difference in the maturation environment between the in vivo and the in vitro matured oocytes. During in vivo maturation, granulosa cells, cumulus cells, oocytes, and other somatic cells secrete various metabolites into the FF to modulate oocyte maturation [12,32,33]. However, during in vitro maturation, granulosa cells and other follicular somatic cells are absent, which results in an abnormal metabolite profile in the culture medium and decreased oocyte maturation quality.Our results showed that during in vivo pig oocyte maturation, the DEMs in follicular fluid were enriched in a metabolic pathway called alanine, aspartate, and glutamine metabolism. A previous study also found that the alanine, aspartate, and glutamine metabolism pathway is disturbed and the glutamine level is decreased in follicular fluid from low reproductive sows compared to that from normal reproductive sows [34]. Glutamine is a precursor of glutathione [35], which is significantly reduced in follicular fluid during human in vivo oocyte maturation [36] and during in vivo pig oocyte maturation, as shown in the present study. Gamma-glutamylvaline is regulated by glutathione and its precursor glutamine, and the upregulation of gamma-glutamylvaline in aged follicular fluid is an important factor that impairs the oocyte quality [37]. These results suggest that the alanine, aspartate, and glutamine metabolism pathway is involved in regulating human and porcine oocyte maturation. The progesterone concentration in porcine follicular fluid is positively correlated with the developmental potential of oocytes [38]. The addition of progesterone into the porcine oocyte IVM medium enhances subsequent embryo development, while the supplementation of RU486, a specific progesterone inhibitor, during IVM prevents porcine oocyte maturation and reduces the subsequent embryo developmental rate [24]. In addition, the supplementation of progesterone during IVM could rescue the negative effects of estradiol-17 beta on pig oocyte maturation [39]. All these results indicate that the progesterone level in the FF is a biological indicator of pig oocyte maturation quality. We found that although the progesterone abundance was increased in the FF and in IVM medium during in vivo and in vitro pig oocyte maturation, the concentration of progesterone in the in vivo FF was much higher than that in the in vitro culture medium. This implies that in the pig oocyte IVM medium, progesterone is insufficient. In this study, the supplementation with progesterone during IVM significantly improved the pig oocyte maturation rate, which confirms the positive effects of progesterone on pig oocyte maturation.We showed that androstenedione, another element in the ovarian steroidogenesis pathway, is significantly upregulated in FF during pig oocyte in vivo maturation but not in the IVM medium during in vitro maturation. Previous studies have demonstrated that one’s androstenedione level is associated with follicular development and oocyte capacitation [38,40,41]. In this study, we found that supplementation with 125 ng/mL androstenedione during pig oocyte IVM effectively promoted subsequent embryo development. 5. ConclusionsIn conclusion, the changing patterns of the metabolites in the FF during pig oocyte in vivo maturation was different from that in the culture medium during in vitro maturation. The ovarian steroidogenesis pathway plays an important role in modulating pig oocyte maturation as the supplementation of two factors in this pathway, progesterone and androstenedione, during IVM increased the pig oocyte maturation rate and subsequent embryo developmental ability. This study helped identify the key molecules that regulate the oocyte maturation quality, which is beneficial for animal production, human health, and biological science.

animals : an open access journal from mdpi

10.3390/ani11061742

PMC8230489

No predominant changes between R- vs. Ad-feed intake on leukocyte defense against pathogens were observed in broiler breeder hens despite some differences in inflammatory and respiratory burst responses. Overall, supplemental 25-OH-D3 had more pronounced effects on the innate immunity of Ad-hens. In vitro studies confirmed the differential effects of 25-OH-D3 to rescue immune functions altered by glucose and/or palmitic acid exposure.

Past immunological studies in broilers focused on juveniles within the rapid pre-slaughter growth period and may not reflect adult immune responses, particularly in breeders managed with chronic feed restriction (R). The study aimed to assess innate immune cell functions in respect to R vs. ad libitum (Ad) feed intake in breeder hens with and without dietary 25-hydroxycholecalciferol (25-OH-D3) supplementation. Ad-feed intake consistently suppressed IL-1β secretion, respiratory burst, and cell livability in peripheral heterophils and/or monocytes along the feeding trial from the age of 51 to 68 weeks. Supplemental 25-OH-D3 repressed IL-1β secretion and respiratory burst of both cells mostly in R-hens, but promoted monocyte phagocytosis, chemotaxis, and bacterial killing activity in Ad-hens in accompany with relieved hyperglycemia, hyperlipidemia, and systemic inflammation. Overnight cultures with leukocytes from R-hens confirmed the differential effects of 25-OH-D3 to rescue immune functions altered by glucose and/or palmitic acid exposure. Studies with specific inhibitors further manifested the operative mechanisms via glucolipotoxicity in a cell type- and function-dependent manner. The results concluded no predominant changes between R- vs. Ad-feed intake on leukocyte defense against pathogens despite some differential differences, but supplemental 25-OH-D3 exerts more pronounced effects in Ad-hens.

1. IntroductionModern broilers can reach a market weight of ~2 kg within 35 days, half of the time needed by their ancestors 50 years ago. Genetic selection expedited growth, but also resulted in reproductive inefficacy, and increased susceptibility to sudden death syndrome, fatty liver syndrome, and obesity-related morbidities such as cardiomyopathy [1,2,3,4]. Feed restriction is typically used to achieve target bodyweight gain in broiler breeders and improve reproductive performance and livability by limiting obesity and related dysfunctions [1,2,3,4].Several studies found that immunological response was also a trade-off for rapid growth selection [5,6,7,8]. Genetic selection for improved growth performance has resulted in a decline of adaptive immune responses co-incident with an increase in cellular immunity and inflammatory responses [8]. To date, most of the immunological studies in growth selected lines of chickens were performed in juvenile birds during the 6 weeks prior to slaughter when massive lean tissue gain occurs. However, the effect of chronic feed restriction on the ability to develop optimal immunity in response to vaccination or against pathogen infection in breeder chickens is sparsely studied.Increased fat mass in obese patients can induce metabolic dysregulation. It has also been recognized that obesity development is tightly associated with inflammatory status, particularly in the adipose tissue [9,10]. The abnormal provocation of factors of innate and adaptive immune system by inflammation acts as a strong inducer in the development of type 2 diabetes mellitus (T2DM) [9,11]. Our previous studies showed many metabolic commonalties existing between obese (T2DM) and broiler breeder hens consuming feed to appetite. Indeed, we have shown that prolonged Ad-feed intake by broiler breeder hens causes rapid fat deposition and produces lipotoxicity and an inflammatory response arising from lipid dysregulation and associated changes in gene expression and signaling that results in impaired functions in ovarian follicles, circulating leukocytes, heart, and pancreatic β-cells [3,4,12,13,14,15].Innate immunity functions as the frontline defense against pathogen invasions; moreover, it also directs cellular and humoral responses to eliminate pathogens. Neutrophils and macrophages are key cells in the innate immune response against invading pathogens. Both types of cells function to clear pathogens through their phagocytic capacity and generation of oxidants that kill engulfed microorganisms [12]. The clearance of pathogens and apoptotic cells by phagocytic cells plays an important role in resolving inflammatory responses, and impairments of these processes often result in a chronic inflammatory state [16,17]. In a variety of models, vitamin D and its receptor signaling play an anti-obesity and –inflammatory role in the development of T2DM and cardiac pathogenesis [18,19]. Vitamin D was shown to be a pivotal component of the monocyte/macrophage response to infection [20,21], inducing their antimicrobial activity. Vitamin D3 (VD3) supplementation not only reversed VD3 deficiency–induced inflammatory responses but also alleviated immunological inflammation caused by LPS (lipopolysaccharide) in table-egg type laying hens [22,23]. We found that dietary 25-hydroxycholecalciferol (25-OH-D3) supplementation improved cardiac health and rescued the livability in the breeder hens partially by ameliorating systemic and cardiac inflammation and fibrotic progression [24,25,26,27]. The purpose of the present studies was to assess the innate immune functions of breeder hens consuming feed to appetite (Ad) and hens fed to achieve target bodyweight by commercial feed restriction recommendations (R) as we found no literature on the topic. Besides, whether the beneficial effects of dietary 25-OH-D3 supplementation extended to innate immunological responses was evaluated in both R- and Ad-hens. In vitro studies with isolated heterophils and monocytes in combination with pharmacological inhibition in glucose and fatty acid metabolism were used to delineate the mechanisms of 25-OH-D3 on innate immune functions.2. Materials and Methods2.1. Animal ManagementA flock of broiler breeder hens (Arbor Acres Plus FF) at age 23 weeks were purchased from a local breeder farm. Hens were caged individually and fed a standard soy and corn-based breeder layer mash (11.6 MJ/kg metabolizable energy; 16% crude protein) with weekly adjustment of feed allocation to achieve the targeted bodyweight as per recommendations [3,4]. The study was designed as a 2 × 2 Latin Square. At the age of 51 weeks, 48 birds were continued with restricted rations (R-hens, 150–160 g feed/day/hen) as recommended, while another group of 48 birds were allowed sufficient feed for consumption to appetite (Ad-hens) until 68 weeks of age. Within each feeding level (Ad or R), half of the hens (n = 24) were provided with the standard breeder diet supplemented with an additional 69 μg 25-OH-D3/kg feed (2760 IU/kg feed equivalent, DSM Nutritional Products Ltd., The Netherlands). All birds were maintained at ambient temperatures and humidities of 25–28 °C and 60–75%, respectively. Birds were provided with free access to water throughout the experiment. Feed was placed at 08:30 a.m. in conjunction with a 14L:10D (lights on at 05:00 a.m.) photoschedule. All bird husbandry and tissue collections were conducted in accordance with an approved animal care protocol of the National Chung Hsing University, Taiwan (IACUC Permit No. 102–113). Egg production and feed intake were recorded daily, and body weight was recorded each week.2.2. Determination of Plasma Glucose, Triacylglycerol, NEFA, Insulin, and IL-1β ConcentrationsOvernight fasting blood samples were collected from 6 randomly selected hens in each treatment group, consisting of hens at 58 and 65 weeks of age from a wing vein. On those same days, a second blood sample was collected 30 min after re-feeding 20 g of feed. Commercial kits were used to determine plasma triacylglycerol (TG, Sigma-Aldrich, St. Louis, MO, USA), glucose and NEFA (non-esterified fatty acid, Wako Chemicals, Osaka, Japan) levels. Plasma insulin and interleukin-1β (IL-1β) concentrations were determined by a validated commercial ELISA kit (Cat.# 10-1249-01, Uppsala, Sweden, KMC006; Biosource International, Camarillo, CA, USA, respectively) [28,29].2.3. Isolation of Peripheral LeukocytesHens in the Ad-group that were 59–60 weeks old had consumed unrestricted amounts of feed for 8–9 weeks. In that same Ad-group, hens that were 66–68 weeks old had consumed unrestricted amounts of feed for 15–17 weeks. At each age, venous blood from two randomly selected hens was pooled for leukocyte isolation and functional analysis of freshly prepared cells. A total of six independent preparations were prepared for each diet group. For mechanistic studies, peripheral blood leukocytes were isolated from 61 to 65-week-old hens and used for overnight cultures with various treatments. In those mechanistic studies, blood from four hens was pooled prior to monocyte and heterophil isolation by discontinuous gradient centrifugation using commercial Histopaque® 1077 and 1119, respectively (Sigma-Aldrich, St. Louis, MO, USA) [12,30,31]. A total of four independent preparations were analyzed for mechanistic studies.2.4. Cell CulturesIsolated heterophils (equivalent to neutrophils in mammals) or monocytes were cultured in RPMI-1640 medium (containing 10% fetal bovine serum, 50 units/mL penicillin G sodium and streptomycin, pH 7.4) at 37 °C, 5% CO2, 95% humidity for 3 h. Cells were then treated with 25-OH-D3 (dissolved in anhydrous ethanol, Sigma-Aldrich, St. Louis, MO, USA), glucose (in distilled water, Sigma-Aldrich, St. Louis, MO, USA), and/or palmitic acid (PA, Sigma-Aldrich, St. Louis, MO, USA) at the indicated concentrations within the physiological conditions for overnight [3,4,13,32]. Palmitic acid was supplemented as a complex with 5% fatty acid-free BSA (8:1, M:M) (Research Organics, Cleveland, OH, USA) [33,34]. Fatty acid free BSA-supplemented media was used as a control.To ascertain the mechanisms of glucolipotoxicity, cells were pre-treated with an individual inhibitor of key steps in the metabolic processes identified for glucolipotoxicity. The inhibitory compounds (Sigma-Aldrich, St. Louis, MO, USA) were fumoninsin B1 (FB1), a ceramide synthase inhibitor in ceramide and sphingomyelin synthesis pathway; n-(2-mercaptopropionyl)-glycine (nMPG), a scavenger of reactive oxygen species (ROS); Desipramine (DPS), a functional inhibitor of acid sphingomyelinase that blocks sphingomyelin breakdown into ceramide; and Triacsin-C (TC), a pharmacologic inhibitor of long chain fatty acyl-CoA synthetase that blocks fatty acid activation and downstream metabolism. Inhibitor concentrations and exposure times varied for FB1: final concentration 25 µM for 1 h; nMPG, final concentration 0.3 mM for 30 min; DPS, final concentration 10 µM, for 2 h; and TC, final concentration 5 μM for 2 h. Fumoninsin B1, n-MPG, and DPS were dissolved in distilled water and TC was dissolved in DMSO (dimethyl sulfoxide). Compounds were added to the medium in a 1:1000 ratio (vol/vol). Following pre-treatment, medium was replaced with treatment media comprised of basal medium in the presence or absence of glucose or PA, or both at the indicated levels and the presence or absence of 25-OH-D3 for overnight (16 h) incubation. Media and cells were collected following incubation for further analyses.2.5. Respiratory Burst and Phagocytosis AnalysisCollected cells were incubated with 1 mL of freshly made luminol (4 × 10−4 M in HBSS containing 0.05 M Na2CO3 and 1.5 mM CuSO4) in the presence of phorbol 12-myristate 13-acetate (PMA; final concentration 5 ng/mL; Sigma-Aldrich, St. Louis, MO, USA) stimulation. Generation of oxidative radicals was measured with luminol-dependent chemiluminescence by a Triathler luminometer (Hidex, Turku, Finland) [12,31,35]. Cellular phagocytic activity was measured by a zymosan-dependent method as described previously [12,31].2.6. Cell Chemotaxis AnalysisChemotaxis were analyzed by cell migration across the membrane according to the trans-well procedures developed previously with some modifications [31]. A total of 600 μL Roswell Park Memorial Institute (RPMI)-1640 medium containing 10% BSA was loaded in a 24-well plate. Approximately 150 μL RPMI-1640 medium treated cells (2 × 105) were loaded into the cell culture inserts (5 μm, Millipore, Billerica, MA, USA). The inserts were put on the 24-well plate and incubated at 37 °C for 1 h. Unmigrated cells were removed by wiping the top surface of the membrane. The migrated cells present on the bottom surface of the membrane were fixed with methanol for 30 min, stained by Coomassie blue R250 (0.5% in 50% ethanol) for 30 min and destained in 10% acetic acid. After drying, the membrane was mounted onto a slide and migrated cells were counted (5 random fields per slide) in 400× fields. The number of migrated cells was enumerated by light microscopy.2.7. Bacterial Killing AnalysisBacterial killing activity was evaluated using a chicken nalidixic acid-resistant strain of virulent Salmonella typhimurium (ST, clone number #15721) [36]. In brief, ST was prepared by propagation in Luria Bertani (LB) broth (Acumedia, Lansing, MI, USA) at 37 °C for 12 h and an aliquot of broth was used for viable bacterial count using the LB agar plate. Freshly isolated peripheral heterophils or monocytes and opsonized ST were seeded in a well of microplate at 1:2 ratio (5 × 105 cells: 1 × 106 colony forming units, CFUs of ST) in duplicates. The plates were centrifuged (500× g, 5 min, 4 °C) and incubated at 41 °C with 5% CO2 for 45 min. After centrifugation and (1000× g, 5 min, 4 °C) and removal of the supernatants, cells were then lysed for 3 min after adding 100 μL of deionized water into the wells. After lysis, 10 μL of WST-8 (supplied in Cell Counting Kit-8, Sigma-Aldrich, St. Louis, MO, USA) were added to the wells for color development as instructed by the supplier. Optical density was read at 450 nm using a microtiter plate spectrophotometer. The method relies on the proportionality between viable bacteria conversion of WSTR-8 into water-soluble formazan dye. Wells seeded with opsonized ST only served as a positive control. The Cell Counting Kit-8 assay was validated by counting colony forming units from control and test wells, as described above. The killing activity was calculated as per the following formula; %killing = (OD cell + OD bacterial − OD sample)/(OD cell + OD bacterial) × 100(1)2.8. Cell Viability and Livability AnalysisCells harvested after overnight cultures were used for viability analysis by Fluorescein (FITC)-Annexin-V/propidium iodide (PI) staining and cytometry sorting as described previously [13]. Results of early (defined as FITC-annexin V-positive and PI-negative), late (FITC-annexin V- and PI-positive) stage apoptosis, and necrosis (FITC-annexin V-negative and PI-positive staining) were pooled to calculate total cell death and viability. Freshly prepared leukocytes suspended in RPMI-1640 medium containing 10% BSA for 1 h were used for livability analysis based on the same FITC-Annexin-V/PI staining method.2.9. Western Blot AnalysisCollected media were concentrated (approximately 20×) using a centrifugal concentrator with a molecular weight cut-off of 3 kDa (Millpore Merck, Darmstadt, German). Aliquots of collected cells and concentrated medium were used for total protein extraction using RIPA buffer and then for Western blot analysis, as described previously [12,14]. A rabbit anti-chicken IL-1β primary antibody (Abcam, Cambridge, UK) and antibodies cross-reactive to chicken antigens, including rabbit anti-VDR (vitamin D receptor, Cat.# GR37-100UGCN, Merck Billerica, MA, USA), anti-p65 (a subunit of NFκB, nuclear factor kappa B, Cat.# 8242), and β-actin (Cat.# 4967, Cell Signaling Technology, Beverly, MA, USA), and a horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology, Beverly, MA, USA) were used in Western blot study.2.10. StatisticsData from functional analyses with freshly prepared cells were analyzed by two-way ANOVA, in which feed intake (Ad or R) and 25-OH-D3 treatment were the classifying variables. Differences between group means were tested using Bonferroni corrected t-test when the main effect was significant. If an interaction between feed intake and 25-OH-D3 treatment was found, a mean comparison was performed. Data from mechanistic studies were analyzed using one-way ANOVA by Tukey multiple comparison test or t-test. Results were expressed as means ± SE. Mean differences were considered significant at p < 0.05. All statistical procedures were carried out using SPSS (Chicago, IL, USA) for Windows 13.0.3. Results3.1. Body Weight, Feed Intake, Plasma Glucose, Triglyceride, NEFA, Insulin, and IL-1β LevelsRegardless of 25-OH-D3 supplementation, Ad-feed intake showed an initial burst of feed consumption in the first week from the prescribed 157 to 196 g/day/hen. Feed intake then declined gradually to reach a nadir at 62 weeks (132 g/day/hen), and thereafter increased slowly to 153 g/day/hen at 68 weeks (Supplementary Figure S1A). The BW of Ad-hens increased sharply to reach 4.6 kg/hen at age of 60 weeks, declined slightly to 63–64 weeks, and subsequently, increased to reach 4.6–4.7 kg/hen at 68 weeks (Supplementary Figure S1B). This pattern of feed intake and BW gain was in marked contrast to a slow increase of BW from ~3.6 kg at 51 weeks to ~3.95 kg at 68 weeks with the breeder prescribed feed allotments.Ad-feed intake provoked obesity-associated metabolic dysregulations including increased fasting plasma glucose levels and 30 min re-fed plasma glucose concentrations at 65 weeks. Increases in circulating insulin, TG, and NEFA concentrations were observed in Ad-hens at both 58 and 65 weeks (p < 0.05, Figure 1A–D). Dietary 25-OH-D3 supplementation suppressed plasma glucose, insulin and NEFA levels of Ad-hens at 65 weeks but did not affect concentrations in R-hens. Ad-feed intake also provoked chronic systemic inflammation as shown by increased plasma IL-1β concentrations at 58 and 65 weeks. Supplemental 25-OH-D3 significantly suppressed plasma IL-1β levels in both R- and Ad-hens at 65 weeks (p < 0.05, Figure 1E). A significant interaction between Ad and 25-OH-D3 treatments was found for plasma glucose, NEFA, and IL-1β level at 65 weeks (p < 0.05, Figure 1A–D) as 25-OH-D3 decreased values more in Ad-hens than in R-hens.3.2. IL-1β Secretion, Phagocytosis, and Respiratory Burst of Fresh LeukocytesIn contrast to R-hens, freshly isolated heterophils at 67–68 weeks and monocytes at both 59–60 and 67–68 weeks from Ad-hens had lower IL-1β secretion (p < 0.05, Figure 2A). Supplemental 25-OH-D3 suppressed heterophil IL-1β secretion in both R- and Ad-hens at 59–60 weeks and in Ad- and R-hen monocytes at 59–60 and 67–68 weeks, respectively, but increased heterophil IL-1β secretion in Ad-hens at 67–68 weeks. Ad-hen monocytes had a lower phagocytic activity at 67–68 weeks and supplemental 25-OH-D3 increased the phagocytosis of Ad-hen heterophils and monocytes at 67–68 weeks and at 59–60 weeks, respectively (p < 0.05, Figure 2B). Ad-feed intake also significantly suppressed respiratory burst response in both types of leukocytes at 59–60 and 67–68 weeks (p < 0.05), which 25-OH-D3 did not alter in Ad-hens. This same variable was significantly decreased in R-hen heterophils and monocytes at 59–60 and/or 67–68 weeks (p < 0.05, Figure 2C) with 25-OH-D3 supplementation. A significant interaction between Ad×25-OH-D3 treatments was observed in IL-1β secretion and phagocytosis of both cell types at 59–60 and/or 67–68 weeks, and in the respiratory burst response of heterophils at 59–60 and 67–68 weeks (p < 0.05, Figure 2A–C).3.3. Chemotaxis and Bacterial Killing of Fresh LeukocytesAd-feed intake exerted no significant effects on chemotaxis and bactericidal activity in either type of leukocytes. However, supplemental 25-OH-D3 increased monocyte chemotaxis in both Ad- and R-hens, as well as bactericidal activity in Ad-hens (p < 0.05, Figure 3A,B). Viability of both types of leukocytes was decreased in Ad-hens compared to those of R-hens. Early apoptosis accounted for most of the heterophil death in both R- and Ad-hens, whereas late apoptosis contributed most of the death in monocytes (p < 0.05, Figure 3C). Supplemental 25-OH-D3 had no effects on the viability of either cell type.3.4. VDR Protein Amounts and NFκB Activation of Fresh LeukocytesAd-feed intake exerted no significant effects on VDR expression and p65 activation in both types of cells and supplemental 25-OH-D3 upregulated heterophil and monocyte VDR expression in both R- and Ad-hens, but suppressed heterophil p65 translocation (p < 0.05, Figure 4A–C). The results confirmed the inhibitory action of supplemental 25-OH-D3 on NFκB signaling in the animal model.3.5. Effects of 25-OH-D3 on IL-1β Secretion of LeukocytesIn order to delineate the effects of 25-OH-D3 on innate immune functions of R- or Ad-fed hens, isolated leukocytes from R-hens were used for overnight culture studies. Treatment with 25-OH-D3 increased IL-1β secretion in heterophils in a dose-dependent manner, but decreased IL-1β secretion in monocytes (p < 0.05, Figure 5A). 25-OH-D3 differentially promoted phagocytosis and respiratory burst response in both types of leukocytes (p < 0.05, Figure 5B,C). Based on these results, the dose for 25-OH-D3 was optimized at 50 nM for the following studies.3.6. Effects of Glucose and Fatty Acid on Leukocyte FunctionsTreatment with glucose suppressed IL-1β secretion, phagocytosis, and respiratory burst response in both types of leukocytes in a dose-dependent manner (p < 0.05, Figure 6A–C). Since the glucose treatment at 300 mg/dL completely blocked the respiratory burst response in heterophils, and glucose at 100 mg/dL failed to affect IL-1β secretion, the dose for glucose effects in the mechanistic studies was optimized at 200 mg/dL. A dose-dependent suppression of IL-1β secretion and phagocytic activity by PA treatment was also observed in both cell types (p < 0.05, Figure 6D,E). Interestingly, PA treatment suppressed heterophil respiratory burst in a dose-dependent fashion, but increased the response in monocytes (p < 0.05, Figure 6F). The dose for PA effects on following studies was optimized at 1.5 mM.3.7. Effects of 25-OH-D3 on Leukocyte Functions during GlucolipotoxicityTreatment of glucose or PA alone impaired IL-1β secretion in both heterophils and monocytes (p < 0.05) and combination of glucose and PA did not exacerbate the impairment (Figure 7A). It was shown that 25-OH-D3 had no effects on IL-1β secretion impaired by glucose, PA alone, or by their combination in either cell type. In the presence of 25-OH-D3, but not its absence, PA treatment further reduced IL-1β secretion caused by monocyte treatment with glucose (p < 0.05, Figure 7A). Combination of glucose and PA exacerbated phagocytic activity impaired by glucose or PA alone in both types of leukocytes and treatment of 25-OH-D3 rescued phagocytosis impaired by glucose and PA combination (p < 0.05, Figure 7B). In heterophils, glucose or PA alone impaired respiratory burst, while the addition of PA to the glucose treatment rescued the response impaired by glucose (p < 0.05, Figure 7C). In fact, 25-OH-D3 completely reversed heterophil respiratory burst suppressed by PA but not that caused by glucose alone or the glucose + PA treatment. Further, PA treatment promoted heterophil respiratory burst in the presence of glucose and 25-OH-D3 (p < 0.05, Figure 7C). In monocytes, glucose treatment impaired, but PA promoted respiratory burst response. Treatment with the combination of glucose and PA completely reversed the response impaired by glucose (p < 0.05, Figure 7C). Treatment with 25-OH-D3 significantly rescued the monocyte respiratory burst suppressed by glucose alone, but had no effects in the presence of PA. Interestingly, PA treatment significantly increased respiratory burst in both the glucose alone and glucose + 25-OH-D3 treatments (p < 0.05, Figure 7C).3.8. Mechanisms of Gluco/Lipotoxicity on Leukocyte FunctionsIn both types of cells, treatment with TC, a pharmacologic inhibitor of long chain fatty acyl-CoA synthetase that blocks fatty acid activation and downstream metabolism, FB1, the ceramide synthase inhibitor within the ceramide and sphingomyelin synthesis pathway, DPS, a functional inhibitor of acid sphingomyelinase that blocks sphingomyelin breakdown into ceramide, and nMPG that acts as a ROS scavenger, differentially rescued IL-1β secretion suppressed by PA or glucose (p < 0.05, Figure 8A). TC treatment exacerbated PA mediated suppression of phagocytosis in heterophils but not in monocytes, whereas FB1 partially rescued phagocytosis in both types of cells (p < 0.05, Figure 8B). Treatment with nMPG relieved phagocytosis impairment by PA or glucose in both types of leukocytes (p < 0.05, Figure 8B). In heterophils, DPS partially relieved respiratory burst suppressed by PA, whereas nMPG exacerbated PA or glucose mediated impairments (p < 0.05, Figure 8C). Treatment with TC or DPS reversed the PA-mediated increase in monocyte respiratory burst, while FB1 potentiated the increase by PA (p < 0.05, Figure 8C). nMPG significantly suppressed the promotion of respiratory burst by PA in monocytes to a level even lower than the control, and further exacerbated the suppression of respiratory burst by glucose (p < 0.05, Figure 8C).3.9. Effects of 25-OH-D3 on Leukocyte Viability Following Glucose or Palmitic Acid ChallengeIn both types of leukocytes, treatment of glucose or PA suppressed cell viability in a dose-dependent manner and 25-OH-D3 partially rescued cell survival (p < 0.05, Figure 9A,B). However, 25-OH-D3 alone had no significant effects on cell viability in either cell type.4. DiscussionVery few studies have examined immune responses of broiler breeder hens, particularly with a view to compare hens reared with a restricted feeding regimen compared to those provided with unlimited access to feed. Consistent with our previous reports [14,24,25,26], Ad-feed intake provoked obesity-associated metabolic derangements including hyperglycemia, hyperlipidemia, and systemic inflammation. In humans and mice, the development of obesity and T2DM was suggested to associate with several defects in the innate immune responses, including decreased chemotaxis, phagocytosis and antimicrobial mechanisms in leukocytes [37,38]. Mice with T2DM were more susceptible to subcutaneous infection with B. pseudomallei and the increased severity of infection was associated with a higher expression of proinflammatory cytokines in paralleled with a rapid drop of blood glucose levels, as is frequently observed in sepsis. A failure of early immune responses to limit dissemination of infection and septic progression with fatal outcome in diabetic mice was primarily attributed to decreased phagocytic and antimicrobial activities in macrophages independent of neutrophil and dendritic cell functionality [38]. Ad-hens showed significantly diminished innate immune responses including impaired IL-1β secretion, respiratory burst, and cell viability in peripheral heterophils and monocytes. These results, particularly in heterophils, were confirmed by our previous studies using the same strain of breeder hens at younger ages (26–35 weeks) under Ad-feed intake for 3–8 weeks [12]. Interestingly, however, the present in vitro studies of immune cells isolated from Ad-hens showed no pronounced differences in chemotaxis, bacterial killing, and phagocytic activity compared to those isolated from R-hens. Accordingly, we concluded that only modest changes occurred in Ad-hen leukocyte defense against pathogens despite some differential differences in inflammation and respiratory burst responses. This is the first report to address the innate immune defense of broiler breeder hens managed with chronic feed restriction in comparison to hens allowed to consume feed to appetite. In a longitudinal study from age of 20 to 60 weeks, infectious causes by bacteria accounted for over 50% of the mortality in a broiler breeder flock [39]. While incompletely described, the commercially reared flocks studied would be expected to be managed with restrictive feeding. In vivo studies, therefore, are required to validate the response of innate immune system against infection of pathogens.Excess of glucose and saturated fatty acids provoke a variety of cellular dysfunctions including cell death, ER (endoplasmic reticulum) stress, autophagy, insulin signaling, and inflammation, namely, glucolipotoxicity [40,41,42]. Type 2 diabetes is characterized by hyperglycemia, dyslipidemia, and increased inflammatory tone [43]. The stimulatory effects of hyperglycemia are enhanced by saturated fatty acids by engaging Toll-Like Receptor 2 (TLR2) and TLR4 receptors to induce ROS production, NFκB activation, and proinflammatory factor release in the adipose tissue and other tissues including leukocytes of the innate immune system to contribute to systemic inflammation and exacerbation of metabolic derangements as obesity and T2DM progress [43,44]. Consistent with our earlier reports, Ad-feed intake increased circulating glucose, insulin, NEFA, TG, and IL-1β concentrations. Surprisingly, however, peripheral heterophils and monocytes from Ad-hens had consistently lower IL-1β secretion in culture. Reduced secretion of IL-1β in the obese chicken model used here (i.e., Ad-hens) contrasts sharply with most reports in models of mammalian obesity, T2DM, and cellular models of glucolipotoxocity [40,42,43,44,45]. We further validated the observations with cell culture studies that used leukocytes from R-hens treated with glucose and/or palmitic acid within physiological levels [20] and again observed a suppressed inflammatory response in these leukocytes, including other functions such as respiratory burst in Ad-hens. Therefore, the innate immune cells apparently contribute little to the well documented systemic elevation in proinflammatory cytokines found in the adipose tissue, liver, heart, and ovary of the obese chicken model used here [14,24,26]. The combination of LPS and palmitate led to a synergistic increase of cellular ceramide in macrophages that augmented proinflammatory cytokine production of metabolic diseases characterized by dyslipidemia [46]. In addition to inflammatory response, ceramide and sphingomyelin metabolism play a critical role in modulating leukocyte functions including chemotaxis, respiratory burst, phagocytosis, and apoptosis [47,48,49,50,51]. The present results confirmed the development of glucolipotoxicity in the leukocytes of overfed chickens, in which palmitate and glucose differentially altered leukocyte functions that could subsequently be rescued by specific pharmacological inhibitors related to ceramide and/or ROS production. Since TC treatment only partially rescued the functions altered by PA, PA per se without activation of downstream metabolism via coenzyme A activation also contributed to lipotoxic development, possibly by interacting with a TLR to activate downstream signaling [42,43,44]. Among the assessed leukocyte functions, TC, FB1, DPS, or nMPG treatment rescued IL-β secretion suppressed by PA. Differential effects were observed for phagocytosis and respiratory burst activities depending on the function and leukocyte type. For example, PA treatment promoted respiratory burst in monocytes, whereas it suppressed this function in heterophils. The differential effects by PA alone or in combination with the specific inhibitors suggest an intrinsic difference in PA metabolism in TG storage, β-oxidation, and ceramide/sphingomyelin synthesis, by which heterophils apparently are more susceptible to lipotoxicity than monocytes. The suggestion was confirmed by differential ceramide and sphingomyelin contents of fresh heterophils and monocytes from R- and Ad-hens [12].In innate immunity, vitamin D modulates both leukocyte and barrier cell functions to combat with pathogen infection [52,53]. It enhances antimicrobial peptide production, TLR expression and signaling, chemotaxis, and phagocytic activity. The anti-mycobacterial activity of 1α,25(OH)2D3 (1α,25 dihydroxycholecalciferol) is mediated by NADPH-dependent oxidase of phagocytes through phosphatidylinositol 3-kinase signaling [52,53]. Vitamin D also regulates the inflammatory states in the local microenvironments of infected sites by upregulating MAP kinase signaling and IL-4 expression and suppressing NF-kB activation and production of chemokines, proinflammatory cytokines and prostaglandin in both infiltrated leukocytes and damaged cells to resolve the inflammatory responses [54,55,56,57,58]. Previously, we reported that dietary 25-OH-D3 supplementation improved cardiac health and rescued the livability in breeder hens by ameliorating systemic hypoxia, hypertension, vascular remodeling, systemic and cardiac inflammation and fibrotic progression, and thereby alleviating pathological remodeling and functional compromise [24,25,26,27]. The present results further showed that supplemental 25-OH-D3 repressed IL-1β secretion and respiratory burst of both heterophils and monocytes primarily in R-hens, but promoted monocyte phagocytosis, chemotaxis, and bacterial killing activity in Ad-hens. The differential effects of 25-OH-D3 to rescue innate immune functions, particularly phagocytic activity, were further confirmed in the in vitro model using R-hen leukocytes in response to glucose and/or palmitic acid within physiological levels.5. ConclusionsNo predominant changes between R-hens vs. Ad-hens on leukocyte functions against pathogens in vitro were observed in broiler breeder hens despite some differences in inflammatory and respiratory burst response. Overall, supplemental 25-OH-D3 had more pronounced effects on the innate immunity of Ad-hens. In vitro studies showed detrimental effects of glucose and/or palmitic acid exposure on leukocyte functionality in a cell type- and function-dependent manner and further confirmed the differential effects of 25-OH-D3 to rescue the alterations of innate immune functions due to glucolipotoxicity.

animals : an open access journal from mdpi

10.3390/ani11123575

PMC8697954

Efficient amino acid (AA) utilization in broilers is crucial concerning the accuracy of feed formulation, economy of diet and minimizing nitrogenous pollution in the environment. A range of factors (ingredient type, age, sex, feed form and bird type) can affect the AA digestion in poultry. The first week of broiler life is considered the most critical period while chicks source nutrition from residual yolk. Dynamic changes are noted in the digestive tract development, secretion and activity of protein digestive enzymes during the first few weeks in broiler. Limited data exist on the age effect on AA digestibility in broilers, and the results are paradoxical. The aim of the present study was to investigate the influence of age on the standardized ileal digestibility coefficients (SIDCs) of AAs in corn and barley from hatching to the end of the broiler growth cycle (day 42). Based on the results, the age influence on AA digestibility is grain- and AA-dependent. The pattern of the age effect on the SIDC AA in corn was not consistent. In the case of barley, the SIDC AA increased with advancing age.

The aim of this study was to determine the standardized ileal digestibility coefficients (SIDCs) of nitrogen (N) and amino acids (AAs) in corn and barley at six different ages (days 7, 14, 21, 28, 35 and 42) of broilers using the direct method. The apparent AA digestibility coefficients were corrected using age-appropriate basal endogenous AA losses. No age effect (p > 0.05) was noted for the SIDC of N in corn. The average SIDC of indispensable AAs (IAAs) and total AAs (TAAs) was influenced in a quadratic manner (p < 0.05) with the values being higher at day 7 that decreased at day 14, increased and plateaued between days 21 and 35 and dropped again at day 42. The average SIDC of dispensable AAs (DAAs) was influenced linearly (p < 0.05). In barley, the SIDC of N and average IAAs, DAAs and TAAs was affected (quadratic; p < 0.001) by age. The digestibility increased from day 7 to 21 and then plateaued up to day 42. The present findings confirm that the SIDC of AA in corn and barley are influenced by broiler age and that the age effect on AA digestibility may need to be considered for precise feed formulation.

1. IntroductionTo improve broiler performance and production economics, accurate feed formulation that closely matches nutrient requirements is crucial. Determination of amino acid (AA) digestibility in feed ingredients is an important way to reach the goal of meeting the AA requirements. Poultry feed formulations based on digestible AA are superior to those based on total AAs because they are reflective of the actual amounts utilized for maintenance and production [1]. A better understanding of the factors influencing AA digestibility is vital to supply available AAs at optimum levels for accurate feed formulation and reduction of diet cost and nitrogen pollution into the environment [2].The AA digestibility in poultry was measured at the excreta level in the past [3,4] but is currently measured at the terminal ileal level. The excreta digestibility assay has several shortcomings including possible modifying action of hindgut microflora on excreta AA profile and contamination with nitrogen (N) and AAs from urine [5]. Ileal AA digestibility can be categorized as either apparent or standardized/true. For the calculation of standardized ileal digestibility coefficients (SIDCs), the apparent ileal digestibility coefficients (AIDCs) are corrected for basal endogenous AA (EAA) losses originating from various digestive, pancreatic and enzymatic secretions [1]. The SIDC is more additive than AIDC in broiler feed formulations [2].Grains are the major energy sources in broiler diets. However, they also supply about 40% of the total dietary protein and contribute significantly to the provision of some indispensable AAs (IAAs). Corn (Zea mays L.) is used extensively worldwide in poultry diets because of its high palatability, low fiber, high energy and essential fatty acids. Despite the low protein content in corn, owing to its higher inclusion levels (50–70%), it may contribute approximately up to one-third of the protein requirement of broilers. Barley (Hordeum vulgare L.) is another grain used in the European Union, Western Canada, New Zealand and Australia. Nevertheless, the inclusion of barley in poultry diets remains limited because of relatively low metabolizable energy, high content of fiber (220 g/kg) and high soluble (45 g/kg) and insoluble (122 g/kg) non-starch polysaccharide (NSP) contents [6]. Because of their viscous nature, the water-soluble fractions of barley exert a negative impact on the digestion and absorption of nutrients, including AAs [7].Several datasets are available on the AIDC AA in feed ingredients for broilers [8,9]. However, the AA digestibility varies depending on ingredient type [9], class of bird (rooster, broiler, layer) [10] and feed form (mash vs. pellet) [11,12]. Despite the potential effects of age [3,13,14], only sporadic and inconsistent data exist on the age influence on the AA digestibility of ingredients in broilers [3,4,13,15]. Although a number of studies [1,9,12,16] have reported the SIDC AA in a range of feed ingredients, only a few [15,17,18,19,20] exist on the age-related standardized ileal digestibility (SID) of AAs, and the data are limited to two or three specific broiler ages. To the authors’ knowledge, no studies to date have investigated the SIDC AA in grains from hatching to the end of the growth cycle of broilers. The current study was designed to determine the SIDC AA in corn and barley at six different ages (days 7, 14, 21, 28, 35 and 42 posthatch) of broilers.2. Materials and MethodsThe experimental procedure complied with the New Zealand Revised Code of Ethical Conduct for the use of live animals for research, testing and teaching and was approved by the Massey University Animal Ethics Committee.2.1. Diets and Experimental DesignCorn and barley were obtained from a commercial supplier and ground in a hammer mill to pass through a screen size of 3.0 mm. Two experimental diets were developed with similar inclusions (938 g/kg) of either corn or barley as the only source of AAs in the diet (Table 1). Titanium dioxide (5 g/kg; Merck KGaA, Darmstadt, Germany) was incorporated in both diets as an indigestible marker. The diets were steam-conditioned at 70 °C for 30 s and pelleted using a pellet mill (Model Orbit 15; Richard Size Limited Engineers, Kingston upon Hull, UK) capable of manufacturing 180 kg of feed/h and equipped with a die ring with 3 mm holes and 35 mm thickness. Pellets were crumbled for the feeding of young chicks during the first two weeks of the experiment.Representative grain samples were analyzed, in duplicate, for dry matter (DM), N, starch, crude fat, crude fiber, neutral detergent fiber (NDF), gross energy (GE), AAs, calcium (Ca), phosphorus (P) and ash. The AIDC of N and AAs in each grain was determined using the direct method. The AIDCs were then standardized using the age-dependent (days 7, 14, 21, 28, 35 and 42) basal endogenous N and AA losses measured in a previous experiment [21].2.2. Birds and HousingA total of 696, one-day-old male broilers (Ross 308), obtained from a commercial hatchery, were used in this study. The birds were raised in floor pens and fed a commercial broiler starter diet (12.14 MJ/kg metabolizable energy; 225 g/kg crude protein; CP) from day 1 to 21 and a commercial broiler finisher diet (12.69 MJ/kg metabolizable energy; 190 g/kg CP) from day 22 until day 42 in pelleted form (Table 2).On day 1, 168 chicks were individually weighed and allotted to 12 cages (14 chicks per cage) in such a way that group mean body weight (BW) per replicate was identical. The remaining chicks were allotted to 12 cages at 5 different ages, namely day 7 (12 birds per cage), day 14 (10 birds per cage), day 21 (8 birds per cage), day 28 (8 birds per cage) and day 35 (6 birds per cage). The test diets were offered for 4 days (days 3–7 and 10–14 (crumbled); days 17–21, 24–28, 31–35 and 38–42 (pelleted)) before collecting ileal digesta on days 7, 14, 21, 28, 35 and 42 posthatch, respectively.The birds were offered ad libitum feed, and water was freely available throughout the whole experimental period. The room temperature was 32 ± 1 °C in the first week that was gradually reduced to 23 °C by the end of the third week. The floor pens, battery brooders and grower cages were housed in an environmentally controlled room with 20 h of fluorescent illumination per day.2.3. Growth Performance DataDuring the 4-day study period, feed intake and BW were recorded on a cage basis each week.2.4. Determination of the Coefficient of Apparent Ileal DigestibilityAt the end of each experimental period (days 7, 14, 21, 28, 35 and 42), all birds were euthanized by intravenous injection (0.5 mL per kg BW) of sodium pentobarbitone solution (Provet NZ Pty. Ltd., Auckland, New Zealand). The digesta were collected from the lower half of the ileum and processed as described by Ravindran et al. [8]. The ileum was marked as that portion of the small intestine extending from the Meckel’s diverticulum to a point ~40 mm proximal to the ileocecal junction. In brief, the ileum was excised and divided into halves (proximal and distal ileum), and the digesta samples were collected from the lower half toward the ileocecal junction after gently flushing with distilled water into plastic containers. The ileal digesta from birds within a cage were pooled after collection, frozen immediately and then lyophilized (Model 0610, Cuddon Engineering, Blenheim, New Zealand). Diet and lyophilized digesta samples were ground to pass through a 0.5 mm sieve and stored in airtight plastic containers at 4 °C pending analysis. 2.5. Gizzard pH and Jejunal Digesta ViscosityFrom the birds euthanized for ileal digesta collection, two birds from each replicate cage were used for the measurement of gizzard pH by a digital pH meter (pH spear, Oakton Instruments, Vernon Hill, IL, USA). The glass probe was inserted through an opening made in the gizzard and was placed directly in the digesta. Three values were taken from the proximal, middle and distal regions, and the average value was considered as the final pH value. The jejunal digesta viscosity was also determined from these birds. The digesta was collected from the distal jejunum, followed by centrifugation at 3000 × g at 20 °C for 15 min. A 0.5 mL aliquot of the supernatant was used in a viscometer (Brookfield digital viscometer, Model DV2TLV; Brookfield Engineering Laboratories Inc., Stoughton, MA, USA) fitted with CP-40 cone spindle with shear rates of 5–500/s to measure the digesta viscosity.2.6. Chemical AnalysisDry matter was measured using the standard procedure (Method 930.15) [22]. Titanium was analyzed on a UV spectrophotometer (Berthold Technologies GmbH and Co. KG, Bad Wildbad, Germany) following the method described by Short et al. [23]. Gross energy was determined by an adiabatic bomb calorimeter (Gallenkamp autobomb, Weiss Gallenkamp Ltd., Loughborough, UK) standardized with benzoic acid. Starch was analyzed using the Megazyme Total Starch Assay kit (Megazyme International Ireland Ltd., Wicklow, Ireland) based on thermostable α-amylase and amyloglucosidase [22,24]. Nitrogen was determined by combustion (Method 968.06) [22] using a carbon nanosphere-200 carbon, N and sulfur autoanalyzer (LECO Corporation, St. Joseph, MI, USA). The CP content was calculated as N × 6.25. Fat was determined using the Soxhlet extraction procedure (Method 2003.06) [22]. Neutral detergent fiber was determined (Method 2002.04) [22] using Tecator FibertecTM (FOSS Analytical AB, Höganäs, Sweden). Ash was measured by ashing in a muffle furnace at 550 °C for 16 h (Method 942.05) [22]. Calcium and phosphorus concentrations were measured by Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) using a Thermo Jarrell Ash IRIS instrument (Thermo Jarrell Ash Corporation, Franklin, MA, USA).Amino acids were analyzed following standard procedures (Method 994.12) [25]. Briefly, the samples were hydrolyzed with 6 N HCl containing phenol for 24 h at 110 ± 2 °C in glass tubes in an oven. Amino acids were measured using AA analyzer (ion exchange) with ninhydrin post-column derivatization. The chromatograms detected at 570 and 440 nm were integrated using dedicated software (Agilent Open Lab software, Waldbronn, Baden-Württemberg, Germany). Cys and Met were analyzed as cysteic acid and methionine sulphone, respectively, by oxidation with performic acid–phenol for 16 h at 0 °C prior to hydrolysis. For the measurement of Trp, the samples were saponified under alkaline conditions with barium hydroxide solution in the absence of air at 110 °C for 20 h in an autoclave. The internal standard α-methyl Trp was added to the mixture following hydrolysis. After adjusting the hydrolysate to pH 3.0 and diluting with 30% methanol, Trp and the internal standard were separated by reverse phase chromatography (RP-18) on an HPLC column (CORTECS C18 Column; 2.7 µm, Waters Corporation, Dublin, Ireland). Finally, detection was selectively performed by means of a fluorescence detector to prevent interference by other AAs and constituents.2.7. CalculationsData were expressed on a DM basis. The AIDCs of AAs were calculated from the dietary ratio of AA to Ti relative to the corresponding ratio in the ileal digesta using the following formula. AIDC of AA = [(AA/Ti)d − (AA/Ti)i]/(AA/Ti)d where (AA/Ti)d = ratio of AA to Ti in the diet, and (AA/Ti)i = ratio of AA to Ti in the ileal digesta.Apparent digestibility values for N and AAs were then standardized using the age-appropriate basal endogenous N and AA estimates (EAA; grams per kilogram of DM intake (DMI)) analyzed at different ages (days 7, 14, 21, 28, 35 and 42) in a previous experiment [21]. SIDC = AIDC + [Basal EAA (g/kg DMI)/Ing. AA (g/kg DM)] where SIDC = standardized ileal digestibility coefficient of the AA, AIDC = apparent ileal digestibility coefficient of the AA, Basal EAA = basal endogenous AA loss, and Ing. AA = concentration of the AA in the ingredient.2.8. Data AnalysisCage was considered as the experimental unit. Data were analyzed by the GLM procedure of SAS (version 9.4; 2015; SAS Institute, Cary, NC, USA) for each grain. Differences were considered significant at p < 0.05. Orthogonal polynomial contrasts were performed to determine the linear and quadratic effects of age. The relationships between SIDC AA and other parameters were analyzed by Pearson correlation. 3. Results3.1. Proximate and Nutrient CompositionThe proximate and nutrient composition of the grains is summarized in Table 3. The results are presented on an “as-received” basis.In both grains, starch was the main component followed by NDF in corn and CP in barley. The starch content was higher in corn than in barley, and the opposite was observed for the CP content. The NDF in corn was determined to be 83.1 g/kg which was lower than that in barley (110 g/kg). The contents of Ca in both corn (0.17 g/kg) and barley (0.14 g/kg) were negligible.Among the IAAs, the content of Leu was the highest followed by Val, Arg, Ile and Thr in both grains, whereas lower contents were determined for Trp and Met. The Glu was the major dispensable AA (DAA) followed by Pro in both corn and barley. The variations in CP contents between the two grains were reflected in total AA (TAA) contents with a higher value in barley (92.5 g/kg) compared to corn (61.8 g/kg).3.2. Growth Performance, Gizzard pH and Jejunal Digesta ViscosityWeekly data on the performance, gizzard pH and jejunal digesta viscosity of birds fed corn- or barley-based diets are presented in Table 4.Mortality during the experiment was negligible. Out of the 696 birds, only four died, and the deaths were not related to any specific treatment. The daily feed intake (DFI) and daily weight gain (DWG) increased (quadratic; p < 0.001) in both corn- and barley-based diets as birds grew older. Gizzard pH increased in a quadratic manner (p < 0.001) with advancing age in both grains. A decline in gizzard pH was observed from day 7 to day 14 but increased beyond day 21. The jejunal viscosity in corn was unaffected (p > 0.05) by age. In the case of barley, however, the jejunal digesta viscosity was influenced quadratically (p < 0.05) by age. Higher viscosity was observed on days 7 and 42 (2.94 cP). After day 7, a reduction in viscosity was observed at day 14 that plateaued until day 35. A further increase was observed at day 42.3.3. Ileal Digestibility Coefficients of N and AAs in CornThe influence of broiler age on the AIDC, SIDC and SID content of N and AAs in corn is presented in Table 5, Table 6 and Table 7, respectively. A quadratic increase (p < 0.001) was observed for the AIDC of N and average digestibility of IAAs, DAAs and TAAs of corn with the advancing age of broilers (Table 5). The AIDC of N and average AIDC of IAAs, DAAs and TAAs increased from day 7 to 21 then plateaued up to day 42. The AIDC of all individual IAAs increased in a quadratic manner (p < 0.001) with age. With the exception of Cys (p > 0.05), an increase (quadratic; p < 0.001) was observed for the AIDC of all individual DAAs.The SIDC of N in corn was unaffected (p > 0.05; Table 6) by age. Bird age, however, quadratically influenced the average SIDC of IAAs (p < 0.002) and TAAs (p < 0.05). The higher values were recorded on day 7 than day 14, and the SIDC values increased until day 21 and plateaued until day 35, followed by a decrease on day 42. The SIDC of average DAAs was influenced in a linear (p < 0.05) manner with a higher value on day 7 (0.881) than day 14 (0.788). Afterward, an increase in the SIDC was observed on day 21 which plateaued until day 35, followed by a decrease on day 42. Except for Thr (p > 0.05), the SIDC of all individual IAAs and DAAs was influenced (linear or quadratic; p < 0.05 to < 0.001) by broiler age. No age influence (p > 0.05) was recorded for the SID protein content of corn (p > 0.05; Table 7). The SID contents of total IAAs, DAAs and total AAs was influenced by age (quadratic; p < 0.05 to < 0.001). The SID content of total AAs was higher at day 7 (54.1 g/kg), increased from day 14 to 21 and plateaued until day 42. The SID contents of all individual AAs, except Thr and Ser, were influenced quadratically (p < 0.05 to < 0.001) by bird age. A linear pattern (p < 0.001) was observed for age effect on the SID content of Ser.3.4. Ileal Digestibility Coefficients of N and AAs in BarleyThe impact of broiler age on the AIDC, SIDC and SID content of N and AAs in barley is presented in Table 8, Table 9 and Table 10, respectively. The AIDC of N and average digestibility of IAAs, DAAs and TAAs increased (quadratic; p < 0.001) from days 7 to 21 and then plateaued from days 21 to 42 (Table 8). The AIDC of all individual AAs in barley increased in a quadratic manner (p < 0.001) as the birds grew older.The SIDC of N and average SIDC of IAAs, DAAs and TAAs increased quadratically (p < 0.001) with the advancing age of broilers (Table 9). The values increased from days 7 to 21 and then plateaued until day 42. The lower SIDC of TAAs was recorded at day 7 (0.617), followed by day 14 (0.738). The SIDC of every single AA in barley increased (quadratic; p < 0.05 to 0.001) with bird age with lower values on day 7.The SID protein content of barley increased (quadratic; p < 0.001) from 77.7 g/kg on day 7 to 94.1 g/kg on day 21 and then plateaued from days 21 to 42 (Table 10). The SID contents of total IAAs, DAAs and total AAs increased (quadratic; p < 0.001) from day 7 to 21 and then plateaued until day 42. The SID contents of all individual AAs increased in a quadratic manner (p < 0.05 to 0.001) with lower values on day 7, which increased either at day 14 or 21 and then plateaued until day 42.3.5. Uplift in Digestibility Coefficients Due to Correction for Age-Appropriate Endogenous Amino Acid Losses The percentage increase in the digestibility coefficients of N and AAs after standardization of apparent values for basal endogenous N and AA losses is shown in Table 11.The correction of AIDC for age-appropriate endogenous N and AA losses resulted in an increase in the SIDC regardless of age, though the extent of increase reduced as the birds grew older. After standardization of the AIDC estimates, the average TAA digestibility coefficients increased in corn by 32.5% (day 7), 17% (day 14), 13.9% (day 21), 14.7% (day 28), 14.2% (day 35) and 9.85% (day 42). In the case of barley, the corresponding increases were recorded as 31.6% (day 7), 11.5% (day 14), 10.0% (day 21), 10.8% (day 28), 10.0% (day 35) and 7.17% (day 42).4. DiscussionMost available data on the AIDC and SIDC AA of ingredients have been determined using older broilers (22 to 35 days of age), and the estimates are applied in feed formulations regardless of broiler age. Several reports exist on age-related AA digestibility in ingredients for poultry, but the results are contradictory and inconclusive. Some studies have documented reductions in protein or AA digestibility [3,4] with advancing age, while others reported increases in digestibility estimates [13,26]. The present experiment aimed at identifying whether the broiler age has an impact on the SIDC of AAs in corn and barley.Although it may be intuitively expected that the AA digestibility in broilers will vary depending on broiler age, studies comparing the SID of AAs corrected using age-appropriate EAAs are limited [18,19,20]. A previous study in our laboratory [14] reported the SIDC AA in wheat and sorghum at six different ages (days 7, 14, 21, 28, 35 and 42) of broilers. In the current research, the AIDC and SIDC AA in corn and barley were determined from hatching to the end of the growth cycle of broilers, and the AIDC values were standardized using age-appropriate basal EAA losses.4.1. Nutrient CompositionThe proximate nutrient contents of corn and barley were comparable to the values reported previously [12,16,27]. The higher starch content in corn (590 g/kg) than barley (541 g/kg) was expected. Due to high starch (620 to 720 g/kg) and crude fat (34 to 52 g/kg) contents in corn, it contains higher energy than any other grain. The CP content of corn (67.8 g/kg) was lower than the range (71 to 94 g/kg) reported by Cowieson [28].The CP content (115 g/kg) in barley was marginally higher than the value (101 g/kg) reported by Perera et al. [27] and lower than the range (121 to 180 g/kg) of Bandegan et al. [16]. The higher CP content in barley compared to corn was in agreement with previous studies [9]. The AA contents of corn and barley were identical or close to those reported in previous studies [9,12,16,27].4.2. Performance, Gizzard pH and Jejunal Digesta ViscosityAs anticipated, regardless of the grain type, an increase in both the DFI and DWG was observed with advancing age. The gizzard pH on day 7 of birds fed corn (2.53) and barley diets (2.14) was close to the values of 2.39 and 2.33 observed by David et al. [29] and Morgan et al. [30], respectively, for broilers of similar age. The reduction of gizzard pH at day 14 compared to day 7 was in line with the findings of David et al. [29] when feeding a corn-based diet. Based on the review of 15 published studies, Angel et al. [31] also reported a reduction in gizzard pH in broilers at day 14. According to Rynsburger [32], the secretion of gastric acid in the proventriculus increased from day 2 to 15, causing a decrease in pH.As observed by Nitsan et al. [33], the secretion and activity of digestive enzymes and hydrochloric acid (HCl) secretion from proventriculus increase with broiler age. However, the observed increase in gizzard pH with bird age after day 14 in both diets could be, at least in part, explained by the increasing intake of feed with neutral pH with advancing age [34]. An increased feed load can dilute the HCl secreted and consequently increase the gizzard pH. The secretion of HCl is fundamental to sustain an acidic environment and to convert pepsinogen to pepsin, the first step in protein digestion [32]. Thus, the implication is that a high gizzard pH would compromise the protein digestion. In this experiment, however, gizzard pH was not correlated (r = 0.242; p > 0.05) with the average SIDC of TAAs in corn. On the contrary, in the case of barley, there was a positive correlation (r = 0.644; p < 0.001) between the gizzard pH and the average SIDC of TAAs. Therefore, it could be speculated that, though the gizzard pH was elevated by age in barley, contrary to expectations, AA digestibility also increased. Besides the pH, there are several other factors, to be discussed later, that potentially influence the AA digestibility in broilers.The jejunal digesta viscosity in corn was not influenced by age, and no correlation (p > 0.05) existed between the digesta viscosity and the average SIDC of TAAs. In barley, the range (2.69–2.94) of jejunal digesta viscosity at different ages was notably higher than that of corn (2.03–2.34), with viscosity being higher on days 7 and 42. Yu et al. [35] replaced corn with increasing barley inclusions (0, 125, 250, 500 and 1000 g/kg) in 3-week-old broilers and reported an increase in the duodenal digesta viscosity from 1.46 cP (0 g/kg) to 2.40 (125 g/kg), 2.15 (250 g/kg), 2.71 (500 g/kg) and 2.81 cP (1000 g/kg). A positive relationship exists between the soluble NSP content and digesta viscosity [7]. Corn contains negligible amounts of soluble NSP compared to wheat and barley [28]. The NSP in plant cell walls such as β-glucans, the major NSP in barley, and the pentosans of rye and wheat exhibit significant antinutritive effects in poultry [36]. A portion of NSP of high molecular weight dissolves in the gastrointestinal tract, increasing the viscosity of gut contents that impedes the digestion and absorption of nutrients [37]. In the case of barley, however, the jejunal digesta viscosity tended (r = −0.292; p = 0.084) to be negatively correlated with the average SIDC of TAAs.Besides the soluble NSP content, a myriad of factors such as growing location, storage time, ingredient inclusion level, age of bird, heat processing and pelleting temperature have been shown to influence the digesta viscosity in barley-containing diets [6]. It is difficult to explain the high viscosity at day 42 observed in the current work since previous studies [38] have reported decreased intestinal viscosity in older birds fed barley-based diets. During the evaluation of a high viscosity hull-less barley, Salih et al. [38] recorded a drop in the digesta viscosity in broilers from two weeks (2.59 cP) to eight weeks of age (1.74 cP).4.3. Ileal Digestibility Coefficients of Nitrogen and Amino AcidsThe first week is the most critical period in a bird’s life when they consume only small amounts of feed and depend mostly on the nutrients from residual yolk [39]. Notable changes occur in the morphology and development of the gastrointestinal tract during the first few weeks of life that contribute to improved digestion and absorption of nutrients. During the first few weeks, there is high protein demand for the development of organs and muscle [39]. The secretion and activity of different proteolytic enzymes such as trypsin, chymotrypsin, intestinal peptidase and dipeptidase also generally increase with age [33].From previous studies, it is evident that the AIDC of AAs is variable depending on broiler age [3,13,40,41]. In the current experiment, with advancing broiler age, an increase in the AIDC of N, all individual AAs (except Cys) and average of TAAs was observed for corn. Compared to day 7, the average AIDC of all AAs increased by 0.46, 22.0, 19.4, 18.6 and 20.5% at days 14, 21, 28, 35 and 42, respectively. The increased AIDC AA with age is in agreement with previous findings [13,19,26]. An increase in ileal N digestibility from 78% at day 4 to about 90% at day 21 has been reported in broilers fed diets based on corn–soybean meal by Noy and Sklan [40]. They concluded that the hydrolysis of exogenous and endogenous proteins was not optimum due to insufficient proteolytic activity at the early posthatch period. An increase in apparent AA digestibility from 1 to 10 days of age was reported with a corn–soybean meal diet by Batal and Parsons [41]. Wallis and Balnave [26] recorded an increase in AIDC AA from day 30 to 50 posthatch (0.732 vs. 0.814) feeding a finisher diet containing a wide range of feed ingredients (wheat, sorghum, soybean meal, cottonseed meal, meat and bone meal, poultry tallow, poultry offal meal, feather meal). Huang et al. [13] determined the AIDC AA of eight ingredients (corn, wheat, sorghum, soybean meal, canola meal, meat and bone meal, cottonseed meal and millrun) at three broiler ages (days 14, 28 and 42). Combining all the results, it was concluded that the digestibility increased with advancing age. The trends, however, were variable depending on the AA and ingredient type. In their study, higher AIDC AA was recorded at days 28 and 42 compared to day 14 in the case of corn, soybean meal, canola meal and meat and bone meal. The AIDC in millrun at day 42 was higher than those at 14 and 28 days. Whilst there was no age influence on the AIDC of most AAs in cottonseed meal, Lys and Arg digestibility increased with age. On the contrary, in wheat, the AIDC of most individual AAs was recorded to be higher at day 14 than at days 28 and 42. In sorghum, the AIDC AA was higher at day 42 compared to day 28 but similar to those at day 14 with the exception of His, Lys, Ser and Gly which were higher at day 42.In the present study, with the exception of Thr, linear or quadratic responses to broiler age were observed for the SIDC of all individual AAs in corn. Unlike the AIDC, the pattern of increase in SIDC AA in corn was not gradual with increasing age. Rather, a decline was observed from day 7 to 14 followed by an increase from day 14 to 21, a plateau between days 21 and 35 and a decline at day 42.Differing age-related trends between the AIDC and SIDC estimates have also been observed by Adedokun et al. [19]. These researchers, comparing AA digestibility between days 5 and 21 of broilers for five ingredients (corn, light and dark distiller’s dried grains with solubles, canola meal and soybean meal), reported an increased AID AA with age in all test ingredients. However, increasing broiler age elevated the SID AA only in corn and distiller’s dried grains with solubles and had no influence on the SID AA in soybean meal and canola meal.Apart from the age effect, another notable observation was made on the AA digestibility in corn in the current study. Though maize contained lower CP and TAA contents compared to barley (Table 3), the average SIDC in maize was 39.4% (day 7), 3.4% (day 14), 9.6% (day 21), 10.8% (day 28), 6.9% (day 35) and 8.5% (day 42) higher than barley. Compared to other grains, maize contains low levels of soluble NSP (1 g/kg) and highly digestible nutrients for broilers. Barley contains a high amount of soluble NSP (45 g/kg) compared to maize [28]. According to Andriotis et al. [42], the barley endosperm is composed of β-glucans (70%) and arabinoxylans (20%). The high content of soluble β-glucan is the major antinutritional factor in barley, and, to assess the potential feed value of barley for poultry, determination of the content and properties of β-glucan is crucial [27]. When birds are fed diets with high inclusions of barley, the NSP impedes nutrient digestion and absorption by two mechanisms. First, the soluble NSP forms gel-like viscous matter, which impairs the interaction between nutrient substrates and endogenous enzymes. Second, the insoluble NSP fraction exerts a “cage effect” by encapsulating the nutrients (starch, protein) in endosperm cells, impeding the contact between nutrients and digestive enzymes. High levels of β-glucans result in thicker cell walls than low levels of β-glucans [43,44]. Though barley has a potential to be included in broiler diets without compromising growth performance [45,46], its inclusion is not recommended for young birds (<7 days) in this experiment due to very low SIDC AA at day 7.In the present study, an increase in the AIDC of N and average AIDC of IAAs, DAAs and TAAs with age was observed in barley. The AIDC of all AAs increased from day 7 to 21 and then plateaued until day 42. Similar patterns of increase were observed in the SIDC of N, SIDC of all individual AAs and average SIDC of IAAs, DAAs and TAAs with lower values on day 7, then an increase until day 21, which plateaued from day 21 to 42. The average SIDC of TAAs in barley at day 7 was 16.4% lower than day 14, 25.0% lower than day 21 and 23.4% lower than the average of days 28, 35 and 42. Szczurek et al. [20] determined the AA digestibility in three grain sources (barley, triticale and wheat) in broilers at day 14 vs. 28 posthatch. In barley and triticale, the average AIDC of IAAs was notably higher at day 28 than 14. The average AIDC of IAAs in barley was reported to be 0.730 on day 14 and 0.780 on day 28. The SIDC of all AAs in barley and most AAs in triticale was also higher at day 28. In the case of wheat, however, no difference was recorded for the average AIDC of IAAs between the two ages. The effect of age on the AA digestibility in wheat and sorghum at six different ages (days 7, 14, 21, 28, 35 and 42) of broilers has already been documented [14]. The AIDC of all individual AAs in wheat increased as the birds grew older. Although the average SIDC of TAAs was unaffected by age, the digestibility of some individual AAs (Met, Trp, Asp and Cys) was higher in the older birds. It is possible that the GIT of older birds is developed sufficiently enough to counteract any adverse effects of digesta viscosity induced by β-glucan. Similar observations were made in other studies [38,47], indicating that the digestibility of nutrients in barley diets improves with age and that viscosity is not a limiting factor in older birds.Several factors may explain the increased AIDC AA with advancing age for corn and barley. Lower production and activities of pancreatic enzymes are a constraint for nutrient digestion in young chicks [39]. Nitsan et al. [33] measured the activities (units/kg BW) of trypsin and chymotrypsin in the pancreas and small intestine from hatching to 23 days of age. The activity of trypsin was low at days 3 to 6 posthatch and increased on day 14. A gradual increase in chymotrypsin activity was also observed from hatch to day 14, and this remained constant afterward. Maximum activities of trypsin and chymotrypsin were observed on day 11. In the intestinal contents, the activity of trypsin increased 10-fold from hatching to day 14, and that of chymotrypsin increased 3-fold by day 20. As stated by Tarvid [48], the total intestinal peptidase (aminopeptidase and dipeptidase) activity increases with the advancing age of broilers.Another possible factor contributing to the increasing apparent AA digestibility with age may be the greater absorptive area [39]. According to Nitsan et al. [33], the weight of the small intestine increased 10-fold at day 8 and 20-fold at day 23. Insoluble NSP is known to promote the development of gizzard [49]. As barley contains high insoluble NSP (122–142 g/kg) [6,27], it is likely that a more developed gizzard in barley will facilitate better nutrient digestion by mechanical breakdown of digesta. Well-developed GIT in older birds can overcome the negative viscosity effects of β-glucans [47].Digesta retention time in the GIT plays a vital role in nutrient digestion because it determines the actual contact time among nutrients, digestive enzymes and microbiota. An increase in digesta retention enhances the absorption of nutrients by increasing the exposure time between the digesta and intestinal absorptive surface. Feed passage time through the GIT is reduced with bird age [50]. Increased digesta retention, in combination with a well-developed digestive tract, especially foregut, in older birds may also increase protein and AA digestibility by enhancing intestinal refluxes, subsequently re-exposing the digesta to pepsin [51].Higher AA digestibility in barley with advancing age might be due to better tolerance of older birds to high NSP and viscosity owing to the increased stability of intestinal microbiota with advancing age [7]. Gut microbiota, a major consumer of AAs, may potentially influence digestion by competing for nutrients [52]. The GIT of the newly hatched chick is sterile. The primary source of initial microbiota is the farm environment where the hatchlings are reared. With increasing age, a cascade of changes occurs in microbiota including species diversity, followed by the complexity of population structure and finally maturation and stabilization. This process continues with age in commercial broilers, and the microbiota profile is stabilized by week 3 [53]. According to Choct [54], dietary NSP can promote microbial growth and fermentation. Increased degradation of soluble NSP and the resultant reduction in the viscosity with age may limit bacterial growth and partly mitigate the adverse effects caused by the excessive population of microbiota.In contrast to the present findings, earlier age-related studies measuring the excreta or total tract AA digestibility in poultry [3,15,55] indicated higher protein and AA digestibility in younger birds. Fonolla et al. [3] reported a decline in the excreta digestibility of protein in broilers with advancing age (day 21 vs. 52), which was attributed to an increased excretion of metabolic N. A drop in true digestibilities of protein and AAs from 3 to 6 weeks of age was reported by Zuprizal et al. [15]. According to Carré et al. [55], the apparent protein digestibility in pea was lower in adult roosters than in 3-week-old broilers. Following feeding of a mixed diet with a wide variety of ingredients, ten Doeschate et al. [4] observed a reduction in protein digestibility coefficients with broiler age. The values were reported to be 0.849 (days 13 to 15), 0.830 (days 27 to 29) and 0.840 (days 41 to 43), respectively. However, these studies are based on total tract digestibility and are not comparable with current findings due to possible contamination of AAs from urine, microbial fermentation in the hindgut and, consequently, modifications in the AA constituents [5].With advancing broiler age, the reduction in ileal digestibility of AAs along with other major nutrients in feed ingredients has also been reported. Adedokun et al. [18] reported 18.5% higher AIDC for TAAs in meat and bone meal in 5-day-old broilers (0.705) compared to that of 21-day broilers (0.595). This pattern remained unchanged even after the standardization of AIDC, with SIDC values of 0.760 and 0.632 at days 5 and 21, respectively. In a previous experiment [14], though there was no age influence on the average AIDC of TAAs in sorghum, the average SIDC of TAAs on day 7 (0.903) was 6.9% higher than on day 14 (0.844) and 8.5% higher than the average from day 21 to day 42 (0.819–0.854). Comparisons among published age-related SID digestibility estimates are not straightforward because of differences in ingredient type [1,13], assumed age-appropriate endogenous losses [21], secretion and activities of digestive enzymes [33], development and activity of gastrointestinal microbiota [53], environment, chemical nature of nutrients and methodology.In the SIDC calculations, the AIDCs were corrected for the basal endogenous AA losses from various digestive, pancreatic and enzymatic secretions [56], and unsurprisingly, these two values were different with SIDC being higher. In practical feed formulations, the SIDC AA of ingredients is preferred because it is more additive especially when mixed into a complete diet. Moreover, standardization eliminates the underestimation of some nutritionally critical and less additive AAs in poultry diets [2]. Though there are some constraints in AIDC estimates, the AIDC is reported in this study to better understand the magnitude of age impact on the SIDC due to correction for age-dependent EAA losses. The differences between the AIDC and SIDC illustrated in Figure 1 and Figure 2 were due to the correction of apparent values by age-appropriate EAA losses determined in a previous study in our laboratory [21]. Although data [1,9,16] exist on the SID AA of different poultry feed ingredients, only a few studies [14,15,17,18,19,20] have determined the SID AA at different broiler ages. Except for some [14,18,19,20], all previous studies have used a single EAA flow value, derived from older birds, to standardize the AIDC values at different ages. From our previous findings [21], the basal EAA losses of TAAs at day 7 (12.93 g/kg DMI) were twice that of the average of days 14 to 35 (6.61 g/kg DMI) and almost three times higher than day 42 (4.48 g/kg DMI). These resulted in substantial differences in the percentage of increase in SIDC over AIDC (Table 11) after correction of apparent digestibility data with age-appropriate EAA losses. The current findings suggest that correcting AIDC AA using a single EAA flow value across ages underestimates the SIDC AA of feed ingredients in young birds and overestimates in older birds.In general, the magnitude of increase in the SIDC AA of corn and barley after correction for age-appropriate EAAs decreased with advancing broiler age. After correcting for age-appropriate EAAs, an increase of 32.5% in the average SIDC of TAAs at day 7 was observed in corn that was almost two times higher compared to that from day 14 to 35 (13.9–17.0%) and three times higher than the increase at day 42 (9.85%). Similar to corn, the average SIDC of TAAs in barley at day 7 was 31.5% higher than the average AIDC of TAAs, which decreased to 7.17% at day 42.To summarize, several factors such as the composition of grains, antinutritive factors, digestive tract development, secretion of enzymes, feed intake, nutrient load, digesta retention time and gut microbiota contribute to the age effect on the AA digestibility in broilers. The endogenous AA losses are a key factor with the greatest impact on the SIDC being observed in young broilers.5. ConclusionsThe present study provides information on the SIDC of AAs in corn and barley from hatching to the end of the broiler growth cycle. The findings suggest that the age influence on AA digestibility is dependent on the grain type and AA. The AIDC AA in corn increased with advancing age. The SIDC AA was higher at day 7, decreased at day 14 and increased and plateaued between days 21 and 35. A further decrease was observed at day 42. In the case of barley, both the AIDC and SIDC of AAs increased as the birds grew older. Standardization of AIDC AA with age-appropriate EAA flows resulted in marked differences in the SIDC of both grains. Application of a single EAA value for correction of the AIDC for broilers of different ages can result in the underestimation of the SIDC AA in young birds and overestimation in older birds. It is concluded that the precision of feed formulations can be improved by using age-specific EAA values for the standardization of AIDC AA values.

animals : an open access journal from mdpi

10.3390/ani12040500

PMC8868152

Small ruminants fed on cactus pear exclusively in the diet may present nutritional disorders, especially diarrhea. This is due to the high moisture requirements present in this food, related to microbiological factors inherent to poor hygienic quality, such as handling and prolonged exposure of crushed palm to the air, which can cause greater ingestion by animals of certain bacterial groups with pathogenic potential. The aim of this research was to investigate the levels of cactus associated with buffel grass hay on the microbiological quality of diets and the influence on physiological parameters in sheep. Thus, this research revealed that the participation of buffel grass hay was able to reduce the contamination of bacteria that cause diarrhea and depress animal performance.

This study aimed to evaluate the microbiological composition of cactus pear-based diets with increasing levels of buffel grass hay, and its effect on the blood and physiological parameters and occurrence of diarrhea in feedlot sheep. Four diets containing different percentages of buffel grass hay were tested. Diets were composed of forage cactus, buffel grass hay and concentrate, and the treatments were represented by different levels of hay in the dry matter of the feed: 7.5% buffel grass hay; 15% buffel grass hay; 30% buffel grass hay; and 45% buffel grass hay on a dry matter basis. There was a significant effect (p = 0.0034) of inclusion levels of buffel grass hay on fecal score. Only at the 45% inclusion level diarrhea was not observed, showing that the level of buffel grass affected more the animals than the collection period, although the collection period has affected the microbial counts. Probably there was a physiological adaptation of animals over time. There were significant changes (p < 0.0001) in the blood parameters of sheep. The reduction of the proportion of cactus and the inclusion of greater than 15% buffel grass hay, on a dry matter basis, provides less contamination of the diet and animal feces by enterobacteria, such as E. coli.

1. IntroductionIn semiarid regions that present irregular patterns in the distribution of rainfall at a certain time of the year, the performance of animals may be limited by forage availability. One way to circumvent this limitation is to provide concentrate feed for animals, and/or use forage plants adapted to climatic and soil conditions, such as cactus pear [1,2].In this sense, the cactus pear is a plant with high adaptability and resistance to weather and that is present in several semiarid regions of the world.As a food, the cactus pear is an excellent energy source, due to its high content of non-fiber carbohydrates, and is an excellent source of water, due to its moisture content [2]. Thus, the tendency of farms in the semiarid region is to use as much cactus pear as possible in rations.In turn, the exclusive use of cactus pear in animal diets, can cause nutritional disorders to animals, such as diarrhea, foamy bloat and ruminal acidosis [3]. Several authors relate these nutritional disorders to low levels of physically effective neutral detergent fiber (NDFpe) in cactus pear [4,5]. However, the appearance of nutritional disorders in ruminants fed high levels of cactus pear is also related to microbiological factors regarding poor hygienic quality, such as handling and exposure of crushed cactus pear to air for a long time, which can provide a greater intake by animals of certain bacterial groups with pathogenic potential, for example, the order Enterobacteriales, responsible for diarrhea [4]. Considering that cactus pear is rich in water, when it is chopped and exposed to air for hours during the animal feeding it could allow microorganisms access to moisture and nutrients, leading to a microbial contamination.Thus, it is important to use other sources fiber sources from hay or forage in diets based on cactus pear, as it implies in reducing the incidence of nutritional disorders not only by improving the NDFpe of the diet, but possibly also by reducing both moisture and soluble carbohydrate content [5].Thus, buffel grass (Cenchrus ciliaris L.) can be an alternative, as it is one of the most cultivated grasses in arid and semiarid regions, mainly due to its tolerance and adaptability to low rainfall [6]. Among the ways to use the buffel grass, that the grass has favorable characteristics to be preserved as hay, as it has a high leaf:stem ratio, thin stems and narrow cuticle.However, although there are several studies on the combination of buffel grass hay with cactus pear, aiming to improve animal performance, there are no studies on the relationship between nutritional disorders and the growth of microorganisms with pathogenic potential in diets with cactus pear.In this context, this study aimed to evaluate the microbiological composition of cactus pear-based diets with increasing levels of buffel grass hay, and its effect on the blood and physiological parameters and occurrence of diarrhea in feedlot sheep.2. Materials and Methods2.1. Place and Period of Execution of the ExperimentThe experiment was conducted on a private property located in the municipality of São José dos Cordeiros, between September and November 2020.The municipality of São José dos Cordeiros is located in the central region of the state of Paraíba, Borborema Mesoregion and Western Cariri Microregion, with latitude: 7°23′26″ S, Longitude: 36°48′30″ W and 529 m altitude. The climate of the region is classified as semiarid BSh, according to the Köppen classification, with average rainfall of 551.0 mm per year.2.2. Animal Management and Experimental TreatmentsFor the experimental trial, 40 rams with non-descript breed with an initial body weight of 18 kg ± 2.6 kg were used. In the period before the experiment, animals were weighed, vaccinated against Clostridiosis and treated against endo- and ectoparasites. Animals were housed in individual pens, measuring approximately 3 m2, with a feeding trough to supply the diet and a drinking fountain with unrestricted access to water. The experiment lasted 60 days, divided into ten days for the pre-experimental period for adaptation to the facilities and diets, and 50 days for data collection. Rations were based on cactus pear “Mexican Elephant Ear” (Opuntia stricta Haw) and buffel grass hay (Cenchrus ciliaris L.), both harvested one year after planting (Table 1).Animals were distributed in a completely randomized design, with four treatments and 10 replications. Four diets consisting of different levels of inclusion of buffel grass hay in the dry matter of the feed were evaluated: 7.5% FB = 7.5% buffel grass hay; 15% FB = 15% buffel grass hay; 30% FB = 30% buffel grass hay; and 45% FB = 45% buffel grass hay on a dry matter (DM) basis. Diets were formulated to be isonitrogenous, and meet the requirements, according to the [7], of sheep with an average weight of 18 kg, for a weight gain of 200 g/day (Table 2).2.3. Food Management and ConsumptionFeed was supplied ad libitum in two equal daily portions, at 08:00 and 16:00 h, weighed and adjusted, allowing 10% leftovers. Leftovers were daily weighed to control the intake of dry matter and other nutrients by the animals. Leftovers were collected before supplying the feed, both in the morning and in the afternoon.Dry matter (DM) intake occurred with the daily weighing of the diet offered and the leftovers between days 15 and 49 of the experimental period. Dry matter intake (DMI) was calculated from the difference between the ingested amount of feed and leftovers, both on a DM basis. For this, samples of food, diets and leftovers were collected every seven days of the experimental period, to determine the dry matter content.2.4. Food Sampling and AnalysisChemical analyses were performed on samples of ingredients and feeds in the food analysis laboratory of the Semiarid National Institute (INSA), Campina Grande, state of Paraíba. For this purpose, samples were analyzed for dry matter (DM; method 934.01), mineral matter (MM; method 942.05), crude protein (CP; method 954.01), ether extract (EE; method 920.39) and lignin (method 973.18) were determined, according to [8]. Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were determined according to the methodology proposed by [9], using the ANKOM fiber analyzer (ANKOM200 Fiber Analyzer—ANKOM Technology Corporation, Fairport, NY, USA).The non-fiber carbohydrate (NFC) content of ingredients and feed was calculated using the equations recommended by [10]: = 100 − (%CP + %NDFcp + %EE + %MM).2.5. Blood Collections, Serum and Physiological ParametersBlood was collected from all sheep by puncturing the jugular vein using Vacutainer tubes containing EDTA (ethylenediaminetetraacetic acid) (0.1 mL 10% EDTA) with a volume of 5 mL each, on the 1st, 6th and 26th day of the experimental period. Collections were carried out before the morning meal and sent immediately to the laboratory for complete blood count and identification of microangiopathic hemolytic anemia due to verotoxins, considering as standard values of the experiment those found in the blood analyzed in the 1st day of the experimental period.The blood count evaluated: Hemoglobin (Hb) (cyanomet hemoglobin described by [11], Mean corpuscular volume (MCV), Mean globular hemoglobin concentration (CHCM), Platelets, Total plasma protein (TPP) by refractometry, Fibrinogen (FP), Leukocytes (hemocytometer), Rods, Segmented, eosinophils and monocytes [12]On the second and 22nd of the experimental period, at 9:00 am, the physiological variables of all animals were collected. The respiratory rate was determined by counting the flank movements for 15 s, this result was multiplied by four to obtain the number of respiratory movements per minute. Rectal temperature was measured using a clinical thermometer inserted into the rectum for one minute, with individual readings, followed by notes. Skin surface temperatures on the nape, shoulder, leg and belly were also measured using an infrared thermometer, at a distance of 30 cm from the animal’s surface, directed transversely to the specific location.2.6. Fecal Score and Microbial Population CountThe fecal score of the animals was determined according to consistency, which varies from 0 (normal consistency) to 4 (watery consistency), according to the [13].On the 1st and 21st day of the experimental period, samples were taken from feces (directly from the rectum) and leftovers (removed from the troughs) for microbiological evaluation, in which the growth of enterobacteriaceae and E. coli (fecal contamination indicator). For this, samples were taken and immediately sent to the laboratory in a Styrofoam box with ice.To evaluate the microbial populations, 25 g feces and dietary leftovers were collected in the trough, added in two hundred and twenty-five milliliters (225 mL) of 0.1% sterile buffered peptone water solution. Through the selective technique of culture medias, violet-red bile lactose agar was used for growth, for the cultivation of enterobacteriaceae and E.M.B Levine agar for the cultivation of E. coli. From the first dilution, new serial dilutions were made, the material from each treatment was grown in sterile Petri dishes using the pour-plate method with the respective culture media, in duplicate, and incubated in a bacteriological incubator at 36 °C for 24 h.Plates for counting were those with values between 30 and 300 CFU (colony forming units), considering the dilution. Colonies of enterobacteria showed a circular shape and pink coloration, and E. coli showed a similar shape, but with a metallic green color.2.7. Statistical Analysis2.7.1. Blood, Microbiological Count and Physiological ParametersThe means of the different periods were compared by Tukey’s Test using the SAS® software. A 5% significance level was used in all hypothesis tests. Given the correlation between the values obtained from the same animal in different periods of evaluation, analysis of variance with repeated measures over time was used. The choice of the covariance matrix for this analysis was made from the analysis of different matrices and the matrix that resulted in the lowest BIC (Bayesian Information Criterion) was selected. When the effect of treatment was significant, their degrees of freedom were broken down into linear, quadratic and cubic contrasts. A mixed model was used, which can be described as Yijkl = μ + Ai + Xj + Pk + (XP) + jk + εijkl(1) where Yijkl is the record of the variable of interest; Ai is the random effect of the i-th animal, where i = 38 levels in the analysis of blood count data and i = 20 levels in the analysis of micobiology data; Xj is the fixed effect of the jth level of buffel grass hay, where j = (1,…,4), Pk is the fixed effect of the kth period, where for hemogram variables k = (1,…,3), while for microbiology variables k = (1,2); (XP) jk is the fixed effect of the interaction between the jth buffel grass hay and the kth period; and εijkl is the random effect of the residual.2.7.2. Fecal ScoreFor the analysis of fecal score, as it is a discrete ordinal variable, it was decided to implement a generalized linear model. This model can be described as Yijk = μ + Xi + Pj + (XP) + ij + εijk (2) where Yijk is the fecal score record; Xi is the fixed effect of the i-th level of buffel grass hay; Pj is the fixed effect of the j-th period; (XP) ij is the fixed effect of the interaction between the i-th% buffel grass hay and the j-th period; and εijk is the random effect of the residual. This model was fitted with a multinomial distribution and the CUMLOGIT option as a link. In this analysis, 5% was also assumed as the significance level.3. ResultsAn effect of interaction of collection period and buffel grass hay levels was detected on the counts of Enterobacteriaceae (p < 0.0001) and E. coli (p = 0.0043) (Table 3). Both on the 1st and 21st day, there was a cubic effect of the levels of buffel grass hay on the count of Enterobacteriaceae in the leftovers (Figure 1). There was also a quadratic effect of buffel grass hay levels on E. coli count on day 1 (Figure 2). The lowest E. coli count in the leftovers occurred when the animals consumed 32.7% buffel grass hay (Figure 2). There was no significant difference in the E. coli count in the leftovers on the 21st day, with a mean value of 4.45 log CFU/g.On the first day of collection, leftovers of animals consuming the lowest levels of buffel grass hay (7.5% and 15%) had the highest count values for both Enterobacteriaceae (7.97 log CFU/g and 6.37 log CFU/g) and E. coli (6.64 log CFU/g and 6.21 log CFU/g) compared to the 21st day, with values of 5.46 log CFU/g for Enterobacteriaceae and 4.54 log CFU/g for E. coli (7.5% buffel hay) and 4.41 log CFU/g for Enterobacteriaceae and 4.44 log CFU/g for E. coli (15% buffel hay). The counts in the leftovers of animals consuming 30% and 45% buffel grass hay did not differ between collection periods, with mean values of 5.57 log CFU/g and 3.83 log CFU/g for Enterobacteriaceae and 4.71 log CFU/g and 4.66 log CFU/g for E. coli, respectively.As for feces, there was a significant difference (p < 0.0001) in the counts of Enterobacteriaceae and E. coli (Table 3). The highest counts were found in the 1st collection period for both Enterobacteriaceae (8.01 log CFU/g) and E. coli (7.06 log CFU/g). There was no effect of diets on the growth of Enterobacteriaceae and E. coli in animal feces, with mean values of 6.72 log CFU/g and 5.73 log CFU/g, respectively.There was no effect of interaction (p = 0.9311) on the fecal score of the animals (Table 4). However, there was a significant effect (p = 0.0034) of the levels of buffel grass hay on this variable (Table 4). Animals fed 7.5% and 15% buffel grass hay presented the highest probabilities of score two, with averages of 70.7% and 63.2%, respectively. Animals fed 30% buffel grass hay had higher probability values in score one. Animals fed 45% buffel grass hay scored zero in 100% samples.The quadratic effect of the levels of buffel grass hay on the dry matter intake of the animals is observed, in which the maximum consumption of sheep is estimated at the level of 25.3% of buffel grass hay (Table 4).There was no effect on the physiological variables: rectal temperature (p = 0.953), nape temperature (p = 0.263), shoulder temperature (p = 0.246), leg temperature (p = 0.257), and heart rate (p = 0.496). Probably, there was an adaptation of the animals to a high load of microorganisms with pathogenic potential (Table 5), which can be proven by the non-significance of the levels of buffel grass hay and the period on most of the physiological parameters of the animals (Table 5).There was no effect of buffel grass hay levels in the diet and collection day on mean corpuscular volume (VHCM) and platelets, with mean values of 30.5 fL and 776,070.75 μL, respectively (Table 6).There was an effect of the collection period (p < 0.0001). On the 21st day (9.45 × 106 mm3), there was a higher concentration of erythrocytes than on the 1st day (7.33 × 106 mm3) and on the 7th day (7.90 × 106 mm3). However, there was no effect of buffel grass hay levels on erythrocyte concentrations (p = 0.4139) with a mean value of 8.22 × 106 mm3.Hemoglobin values had a significant effect (p < 0.0001) for the collection periods and there was an effect of interaction of hay levels with collection periods (p < 0.0001). Animals consuming 7.5% buffel grass hay had higher hemoglobin concentrations on the 21st day (9.24 g/dL and 9.43 g/dL) than on the 1st day (7.17 g/dL and 8.10 g/dL). Animals consuming 15% and 30% buffel grass hay had higher hemoglobin concentration on the 21st day. There was no effect of period on the hemoglobin concentration in animals fed 45% buffel grass hay, with a mean value of 8.26 g/dL.Hematocrit values showed a significant difference (p < 0.0001) for the collection periods, with an increase in the amount of hematocrit on the 21st day compared to the 1st and 7th day (Table 6).There was a significant effect (p < 0.0001) for the collection periods and interaction of buffel grass hay levels with collection periods (p = 0.0092) on the mean corpuscular hemoglobin concentration (CHCM). Mean values of CHCM of animals fed 7.5% buffel on the 1st day (30.18 g/dL) were lower than the 7th (33.07 g/dL) and 21st (32.90 g/dL) days. With respect to animals fed the other levels of buffel grass hay, there was no effect of period on this variable.There was a significant effect (p = 0.0106) for the collection periods and interaction of the levels of buffel grass hay with the collection periods (p = 0.0090) for leukocytes (Table 6). Animals receiving 15% buffel grass hay had higher concentrations of leukocytes on the 7th (8011.11 μL) and 21st (7955.56 μL) days when compared to the same animals on the 1st (5277.78 μL) day of collection. There was no effect of period on leukocytes of animals consuming 7.5; 30 and 45% buffel grass hay, with mean values of 7349.73 μL; 7185.19 μL and 8769.58 μL, respectively.The values of segmented showed significant differences for the collection period (p < 0.0001) and interaction (p = 0.0083). Animals consuming 15% buffel grass hay had higher segmented values on the 7th day (8011.11 μL) and 21st day (7955.56 μL) when compared to the 1st day (5277.78 μL). Animals fed 45% buffel grass hay had higher segmented values on the 7th day when compared to the 21st day. Animals receiving 7.5 and 30% buffel grass hay had similar segmented values in the different collection days.There was interaction (p < 0.0001) on eosinophils, and animals receiving 30% on the 21st day had the highest blood eosinophil levels (Table 6). Animals from other treatments had similar eosinophil values on the 1st, 7th and 21st day. There was also a quadratic effect of buffel grass hay levels on eosinophil values on the 1st day and cubic effect on the 7th and 21st days.There was no effect of interaction of buffel grass hay levels with period in relation to lymphocytes. However, there was a quadratic effect of levels of buffel grass hay on this parameter, with a minimum value of 18.7% buffel grass hay (Figure 3). There was also an effect of period on lymphocytes, in which there was a lower concentration of lymphocytes on the 7th day when compared to the 1st and 21st days.There was an interaction (p < 0.0001) for monocytes (Table 6). On the 1st day of collection, animals consuming 7.5% buffel grass hay had lower monocyte values 217.8 μL) when compared to the 21st day (554.40 μL). In relation to the animals consuming 15% buffel grass hay, on the 7th day, the animals had higher values than on the 1st day. Animals fed 30% and 45% buffel grass hay showed no significant difference in monocytes between the collection periods.Values of total plasma proteins showed interaction of buffel grass hay levels with collection periods. The highest concentration was found in animals consuming 15% buffel grass hay, on the 7th day of collection (7.52 g/dL).4. DiscussionIt is possible to infer that the growth of both Enterobacteriaceae and E. coli as reduced with the lowest proportions of cactus in the diets. At the level of 7.5% of buffel grass hay, there was a greater proliferation of these microorganisms, possibly due to the high moisture content in the diet. The moisture content is a factor that improves microbial growth, as it must be considered when the food is chopped and offered to the animals, since it can be exposed to the air for several hours, providing an ideal environment for development. of pathogenic microorganisms such as Enterobacteriaceae, including E. coli. Cactus pear is a food with high moisture content and its composition is rich in substrates for microbial fermentation, such as soluble carbohydrates. When cut into smaller structures, there is a greater surface area of the forage cactus available for access to nutrients by microorganisms, allowing for greater growth and fermentation [4].This may have occurred due to the increase in the pH of the diet, as well as the greater exposure of the material to oxygen, when reducing the amount of cactus in the diet provided. However, the increase in the proportion of buffel grass hay, which has a more alkaline pH [14], combined with cactus pear mucilage that has as a buffer effect [15], in combination, may have created a suitable environment for the growth of these microorganisms. In addition, water in food can form a physical barrier for oxygen to enter material exposed to air, so the drier the food, the more easily air can penetrate it.The interaction of high moisture content of the diet with 7.5% cactus pear, particles smaller than 2 cm2, and time of exposure to air in the diet, proved to be the combination that provides the highest counts of Enterobacteriaceae and E. coli in leftovers, which can lead to deterioration of the material as well as ingestion of foods with a high concentration of agents with the potential to harm the animal health.E. coli is the most common species in the intestines, and when found in water and food, it indicates fecal contamination, which can be pathogenic and cause diseases [16]. Very high counts may indicate the ingestion of contaminated food that culminates in the infection of the animal organism, causing diarrhea, intense fever and dehydration due to watery feces [4]. Probably, the processing of cactus before feeding the animal and stored in an environment with poor hygiene stimulates the growth of enterobacteria, reflecting in a diet with a higher bacterial count, therefore, high amounts of these bacteria can be excreted in the feces. Enterobacteriaceae are commonly found in the most diverse environments and in animals feces, thus poor hygiene environment could result in contamination of the rich cactus diets when feeding the animals, resulting in infections and diarrhea.The lower count of these microorganisms in the leftovers on the 21st day in diets containing 7.5 and 45% buffel grass hay compared to the 1st day of the experimental period is probably due to the adaptation of the animals over the days to this high concentration of microorganisms.The presence of Enterobacteriaceae and E. coli was expected since these microbial groups are present in the animal organism [16]. The microbiological counts of Enterobacteriaceae on the first experimental day are similar to the values reported by [4], who observed a count of 8.48 CFU/g in the diet with crushed cactus pear eight hours before feeding. In a biochemical profile test, the presence of E. coli in the feces of animals was confirmed, reinforcing that these microorganisms inhabit the gastrointestinal tract of sheep.The highest count of Enterobacteriaceae and E. coli in feces of the animals on the first day of collection was possibly because the animals were still adapting to the levels of cactus pear in the diets, and it may be a consequence of the possible ingestion of a food with a higher bacterial load, given the higher count of these microorganisms in food leftovers, also on the first day of collection. However, throughout the experimental period, the animals become adapted to their diets, promoting a balance between the microorganisms present in the gastrointestinal tract of sheep, as the counts of Enterobacteriaceae and E. coli in feces were lower on the 21st day (Table 3).Changes in the fecal score of sheep are related to the amount of cactus pear used in this study. The higher probability of fecal score two in animals fed 7.5% and 15% buffel grass hay is likely due to the lower DM content of these diets (Table 3). The animals had higher counts of Enterobacteriaceae and E. coli only on the first day of collection for these levels (Table 4) which suggest that that it was an effect of moisture on the microbial counts.When cactus pear is offered in too high amounts in the ruminant diet and without the provision of an extra source of forage, changes in the characteristics of animal feces are observed, generally less solid than normal, due to the increased passage rate caused by the high amount of water available and low fiber content in the diet based on cactus pear [17].Therefore, the opposite should also be taken into account, since the higher levels of fiber in the diet for ruminants can reduce the moisture content and regulate the passage rate through the gastrointestinal tract, also affecting the growth of Enterobacteriaceae and E. coli in the feed (Table 4), which interferes with the microbiota of the gastrointestinal tract and, consequently, with the occurrence of diarrhea in the animals (Table 4). As a consequence, the level of 45% buffel grass hay presented a fecal score equal to zero, which is considered the normal value for sheep.There is a quadratic effect of the levels of buffel grass hay on the dry matter intake of the animals, in which the maximum intake of sheep is estimated at the level of 25.3% buffel grass hay (Table 4). The reduction in intake within the range of tested levels (7.5 and 45% FB) may be an adaptation of the animal to food with a high microbial load. Contaminated feed tends to reduce the intake in ruminants as this can cause negative effects on ruminal fermentation affecting the performance and health of animals [18].A reduction in the dry matter intake of animals receiving diets with the highest counts of Enterobacteriaceae and E. coli was also observed [4]. Therefore, it can be assumed that the animals tended to consume a smaller amount of food due to its greater contamination, as a defense mechanism of the animal itself.When ruminants are fed low-fiber diets, their ruminal pH drops, their microbial ecology changes, and they become more prone to metabolic disturbs [16]. Evaluating the effect of dietary fiber level in relation to dry matter (DM) intake of feedlot sheep, [19] observed that DM intake increased linearly as dietary fiber levels increased from 8.67% to a maximum of 34.69%. In the present study, dietary fiber levels were also increased according to the levels of buffel grass hay (Table 2) with DMI maximization with 25.3% buffel grass hay, above that, there was a gradual reduction in the DMI with the inclusion of this forage (Table 4). Considering that the 15% buffel grass hay diet contained 36.34% NDF and the 30% buffel grass hay diet contained 43.92% NDF (Table 3), the estimated dietary NDF value that would maximize the DMI was close to the cited authors. In addition to the high dietary pH, which may have negatively influenced the microbiological quality of the diet, as discussed above, excess fiber can also depress the animal’s DMI due to the filling effect [19].There was an effect of interaction (p < 0.0001) on the belly temperature of the animals. However, this effect does not allow the inference of any pathological change in the animals, since, of the physiological parameters analyzed, rectal temperature and respiratory rate are the parameters that could indicate some type of anomaly in thermal comfort or in the health of the animals [20]. Nevertheless, the mean values of these parameters are in accordance with the reference values used by [21] for Dorper sheep that have been naturalized or crossed with native sheep, under tropical conditions, which are 38.3–39.9 °C and 20–34 moves per minute, respectively. These results reinforce that the animals get adapted to a high load of potentially pathogenic microorganisms.The way of managing diets can positively influence, favoring the animal adaptation to diets with microbiological risk, such as cactus pear. In the present study, cactus pear was crushed and immediately offered to the animals, and the leftovers were removed twice a day, with cleaning of the troughs before supplying a new meal. Studies prove that in diets with high levels of cactus pear, the provision of cactus pear to the animals, immediately after grinding, contributes to a significant drop in Enterobacteriaceae counts in sheep feces, as reported by [4]. Thus, the way of managing and using cactus pear in diets with high levels of this plant in its composition can interfere with the growth of microorganisms with pathogenic potential, which may be correlated with cases of nutritional disorders or reduced performance of sheep confined.In the present study, an effect of interaction (p < 0.0001) was detected for the hemoglobin data, which were lower on the first day and higher in animals fed 45% buffel grass hay. Moreover, hematocrit values showed significant differences between collection periods, which may reinforce some anomaly in animal health (possible anemia) during the first days of the experimental period and a possible adaptation by these animals over the days. Thus, the results of blood parameters allowed inferring that the animals were able to adapt to the diets even with the high levels of cactus pear in its composition.The mean values of erythrocytes, hemoglobin and hematocrit are close to those reported by [4,22]. Reference [4] concluded that animals can adapt to diets with high levels of cactus pear depending on the way of feed management. This study reinforces our findings, as it was carried out under similar conditions and with animals with the same genotype and phenotype patterns.In the present study, animals fed 7.5% buffel grass hay had lower hemoglobin concentration values on the first day of collection, which would be explained by the fact that the animals are still adapting to their diets (Table 6). However, it is noteworthy that despite the lower means presented by the level of 7.5% buffel grass hay on the 1st day of collection, these values were within the reference range suggested in the complete blood count exam (30–36 g/dL) and remained both in the second (33.07 g/dL) and in the third collection (32.90 g/dL).Values of total leukocytes are responsible for indicating a possible infection in the animal organism, whether caused by viruses or bacteria, in suspected blood diseases, parasitic infections or allergic processes [23]. In the present study, values of total leukocytes at 15% buffel grass hay were lower on the first day of collection compared to the 7th and 21st days, which may indicate an immune response of the animal to a possible infection caused by ingestion of foods with higher counts of Enterobacteriaceae and E. coli in the first days of the experimental period (Table 3). The highest counts of Enterobacteriaceae and E. coli in the animals’ feces on the first day of collection coincide with the period of blood collection, in which the animals had the lowest total leukocyte counts, so it is possible to state that the microorganisms affected the health of the animals. Values of total leukocytes below those indicated may represent a condition of leukopenia, where severe cases suggest infections [23]. However, the total leukocyte counts on the 1st day of collection was 5277.78 μL, being within the reference range indicated by the complete blood count test, which is 4000–12,000 μL, indicating that despite the drop, the animals did not progress to more severe conditions and over the days, they managed to recover, through the immune system, because in the 2nd collection period (7th day) the concentration of total leukocytes increased (Table 6).The increase in leukocytes indicates that there was a possible inflammatory or infectious process occurring in the organism. The reduction in the counts of Enterobacteriaceae and E. coli on the 21st day of the experimental period also demonstrates that the animals were able to recover from a possible infection over the experimental days, without major complications. Values of total leukocytes on the 21st day, in animals fed 15% buffel grass hay, were similar to those found for healthy male Dorper sheep [24] which was 7576.0 μL, reinforcing that there was recovery of the animals over the days.Values of segmented at the 15% inclusion level of buffel grass hay, on the first day of collection, may be related to the values of total leukocytes indicating the presence of an infection caused by bacteria. On the 7th day, the highest values indicate the immune response of the organisms and on the 21st day, the values reduced to a value that is in accordance with the reference literature [6,24].Some blood parameters of the animals were affected by the levels of buffel grass hay in the diet with inclusion of cactus pear. Most of them may be related to the highest total leukocyte count, since the lowest count was found at the level of 15% buffel grass hay inclusion on the 1st day of collection, and the highest on the 7th day of collection, with also the lowest (1st day) and highest (7th day) values for segmented, eosinophils, monocytes and plasma proteins. This indicates recovery by the animal organism from a possible bacterial infection caused by lower levels of buffel grass hay in the diet and higher concentrations of cactus pear.Despite the pasty consistency presented by the feces of animals receiving 7.5% and 15% buffel grass hay, no cases of severe digestive disorders, such as totally liquid feces or even with the presence of blood were observed. Probably, defenses of the animal organism managed to prevent the appearance of diseases even at certain levels of food contamination. Thus, it can be stated that sheep become adapted to diets with high content of cactus pear and low fiber content from buffel grass hay, even those with higher counts of Enterobacteriaceae and E. coli.Thus, high levels of cactus pear in the sheep diet alter the microbiological, blood and fecal parameters of confined sheep in the first periods of diet allowance, however, the animals manage to adapt to the diets without major complications.The research results demonstrate that the cause of nutritional disorders in ruminants consuming diets with a high proportion of cactus pear is not exclusively due to the low fiber content of physically effective diets. The high concentration of non-fiber carbohydrates in fresh cactus pear, and the inadequate feeding management of this forage, can predispose the animals to contamination by pathogenic microorganisms that will negatively interfere with the animal health, causing, among other problems, diarrhea. As forages have a low content of NFC, in the form of hay, they contain low moisture, reduce the proliferation of microorganisms in diets with cactus pear, by decreasing the favorable environmental characteristics for the growth of these microorganisms.Nevertheless, it was also observed that sheep were able to adapt to diets with low levels of buffel grass hay. Thus, producers, especially in semiarid regions, can use cactus pear as a basis for the diet of confined sheep, provided it is correctly managed, so that the multiplication and ingestion of undesirable and possibly pathogenic microorganisms is avoided, for favoring secondary fermentations in the feeding trough of the animals.Although physically effective fiber levels help to reduce nutritional disorders in sheep fed diets with high levels of cactus pear, the inadequate management of this food can also cause nutritional disorders. However, studies on the subject are still scarce, suggesting the need for more research to obtain more precise results.5. ConclusionsReducing the proportion of palm and adding more than 15% buffel grass on a dry matter basis provides less contamination of the diet and animal feces by enterobacteria, such as E. coli, without any negative change in blood parameters and reducing the occurrence of diarrhea in confined sheep.

animals : an open access journal from mdpi

10.3390/ani13061117

PMC10044397

Neonatal mortality represents a major cause of production loss in camelids, reaching rates higher than 25%; despite the increasing global importance of this species, most of the physiological features of the newborn dromedary camel remain unknown. Tailored reference values for dromedary calves are needed for a more accurate discrimination of healthy and sick animals. This study aimed at investigating the blood gases, acid-base and electrolyte profiles in healthy dromedary calves during the first 3 weeks of age, assessing possible associations with age. A total of 21 healthy dromedary camel calves aged between 1 and 21 days were sampled on the same day, and venous blood was analyzed through a VETSTAT® analyzer. Calves were divided in two groups; younger calves aged 1–10 d, and older calves aged 11–21 d. Age was associated with differences in K+, Na+, Cl−, sO2 and pCO2 between the two groups. These results suggest an effect of age on some blood parameters, and provide preliminary data regarding the blood gas, acid-base and electrolyte profiles in the healthy dromedary calf during the first 3 weeks of age.

The importance of prompt evaluation and care of the newborn is essential for reducing neonatal mortality, which represents a major cause of loss in camelids. This study investigated the blood gases, acid-base and electrolyte profiles in healthy dromedary calves during the first 3 weeks of life, assessing possible associations with age. Twenty-one dromedary camel calves aged 1 to 21 days were sampled, and venous whole blood analyzed through a VETSTAT® analyzer. The following parameters were measured: sodium (Na+), potassium (K+), chloride (Cl–), hydrogen ion concentration (pH), partial pressure carbon dioxide (pCO2), partial pressure oxygen (pO2), total hemoglobin concentration (tHb), hemoglobin oxygen saturation (sO2), total carbon dioxide (tCO2), bicarbonate (HCO3–), base excess (BE) and anion gap (AG). Calves were divided in two groups; younger calves (1–10 d), and older calves (11–21 d). Statistical analysis showed an effect of age, with lower K+ (p < 0.001) and higher Na+ and Cl− (p < 0.05) mean concentrations in the younger calves compared to the older ones, and higher pCO2 and lower sO2 mean concentrations in the older group. These preliminary results firstly described the blood gas, acid-base and electrolyte profiles in the healthy dromedary calf during the first 3 weeks of age, suggesting an effect of age on some parameters.

1. IntroductionOld World camel species include the Dromedary (Camelus dromedarius) and the Bactrian (Camelus bactrianus), which are native to the Middle East region and Northern Africa, and to central Asia, respectively. According to a recent study [1], the world camel population is growing and should reach 60 million heads in 25 years. This growth is mainly driven by the raising demand for camel meat and milk of distinguished quality [2], and by the increased use of dromedary camels as a production animal worldwide, thanks to the high level of sustainability of camel farming and to the ability of camels to cope with heat stress [3,4,5]. The dromedary represents an important sport and tourism resource in the Arabian Gulf countries, and it is also employed for packing, transport, and riding [6,7]. In Saudi Arabia (KSA) the dromedary camel plays a primary role in the history of the Kingdom, representing a national wealth and a source of income to the majority of citizens, particularly in desert areas.Due to seasonality, the usual calving interval for the camel is 24 months but it might easily extend to 36 months [8]; neonatal mortality, beyond that affecting animal welfare, exerts a great economic impact on farm profitability in the dromedary camel species [9]. Calf losses slow down herd’s reproduction rate, thus reducing the income for meat production. The economic importance of calf mortality is also linked to the milk production, as the presence of the calf is essential for initiation and maintenance of the lactation [8]; consequently, calf mortality threatens the sustainability of the camel farming system, and affects food and milk production in marginal areas. In a survey carried out in eastern Sudan, a 48% mortality rate was reported among calves under 6 months of age [10], while dromedary calf mortality rate between 30 and 50% has also been reported in Kenya [11], Tunisia [12] and Somalia [13]. In some camel herds in the UAE, dromedary calf mortality rates of up to 60% have been reported during the first 3 months of life, but with overall morbidity and mortality rates of 40% and 25%, respectively [14]. Many factors contribute to calf mortality, among which is calf diarrhea, that causes dehydration, together with electrolytes and acid-base imbalances. Inflammation of the intestinal lining impairs the calf’s ability to digest nutrients, leading to weight losses and other life-threatening disorders. A mortality rate of 39.9%, due to diarrhea, was reported in calves aged less than six months in Sudan [15]. Al-Harby et al. [16] evaluated calf scour morbidity and mortality rates in 1200 0–14 day old camel calves in Taif area (KSA); a 20% morbidity rate and 50% mortality rate were observed, respectively, with an equal distribution between winter and summer. Due to the critical issues highlighted above, health management of dromedary calves require care and understanding, particularly during the neonatal life, as they are susceptible to illness until they reach one year of age [8]. Notwithstanding the growing number of camels and the intensification of camel farming systems, there is a dearth in the literature regarding the physiology of newborn dromedary camels. It is essential to understand that young and adults are physiologically different, which justifies the need to obtain age-specific reference intervals. Several studies have indicated significant differences between the physiology of young and adult cattle [17,18,19,20,21], and the same can be found for other species such as horses [22], dogs [23,24] and pigs [25]. Some studies showed significant differences also in the hematological and biochemical profiles of growing dromedary calves from birth to the 4th week of age [26], and between young and adult [27,28,29]. Camel farms are often located far from diagnostic facilities and laboratories, and assessment of sick dromedary calves is still commonly mainly based on physical examination [30,31,32]. However, Old World camelids are legendary in their ability to preserve water, and this adaptation may hamper the accurate identification of early hypovolemia in critically ill camelids by physical examination alone, making the clinical estimation of fluid deficits particularly challenging [33]; for example, their small, ovoid erythrocytes are resistant to osmotic changes, and can continue to circulate in situations of increased blood viscosity. The increasing availability of pen-side blood gas analyzers can facilitate diagnostic work-up and permit evaluation of degree and nature of alterations in acid-base and electrolyte balance, which usually accompany diarrhea and other neonatal diseases. Blood gas and electrolyte analysis combined with appropriate reference ranges for healthy animals represents a prompt and useful tool in the early detection and close monitoring of disturbances; early diagnosis and treatment of neonatal disease, particularly infections, result in greater positive outcome [34]. Mean values for blood gas analysis of newborn dromedary calves immediately after birth have been investigated [26], but no information is available regarding the blood gas and acid-base profile in dromedary calves during the first weeks after birth. Tailored reference values for dromedary camel calves are urgently needed for more accurate differentiation of healthy and sick animals. Therefore, the aim of the present study was to investigate the blood gas, acid-base and electrolyte profiles of healthy dromedary calves up to 21 days of life.2. Materials and Methods2.1. Sample PopulationIn March 2022, twenty-one dromedary camel calves aged between 1 and 21 days (d) were sampled. The animals were reared through a semi-intensive breeding system in a farm located in the Al-Qassim region (Centre of Saudi Arabia). About one hundred dromedary camel females were reared on the farm, distributed into pregnant and calf-rearing dams. The paddocks were provided with shaded areas. Pregnant animals were left free to graze on pastures for three hours/day and received alfa-alfa upon their return. Calf-rearing animals received alfa-alfa and concentrates. For both groups water was provided at libitum. All procedures were performed on the same day, to avoid environmental biases. The health status of each calf was assessed on the day of sampling through complete physical examination performed by veterinarians: body temperature, calf demeanor, mobility and reactivity, suckle reflex, feed intake, fecal consistency and hydration status were evaluated. Only calves that were deemed clinically healthy were enrolled in the study. 2.2. Blood Sampling and AnalysisEach calf was blood sampled by jugular venipuncture. A venous blood sample was collected from the jugular vein into 2.5 mL lithium-heparin syringes, and immediately analyzed by a VETSTAT® analyzer (IDEXX, Westbrook, ME, USA), with individual body temperature set for each calf. Before testing, all visible air bubbles were expelled from the syringe, which was continuously agitated to prevent formation of microclots. Single-use disposable cassettes were employed for the assessment of sodium (Na+, mmol/L), potassium (K+, mmol/L), chloride (Cl–, mmol/L), hydrogen ion concentration (pH), partial pressure carbon dioxide (pCO2, mmHg), partial pressure oxygen (pO2, mmHg), total hemoglobin concentration (tHb, g/dL), hemoglobin oxygen saturation (sO2, %), total carbon dioxide (tCO2, mmol/L), bicarbonate (HCO3–, mmol/L), base excess (BE, mmol/L) and anion gap (AG, mmol/L) in whole blood. Before performing the analysis, a temperature-correction was applied for each sample on the basis of the rectal temperature registered during clinical examination of the subject. To ensure optimal performance, standard reference cassette (SRC) measurements were checked on the day of sampling, by using SRC level 1 and SRC level 3 provided by the manufacturer.2.3. Data Collection and Statistical AnalysisCalves were divided into 2 groups according to their age: between 1 and 10 d of age (n = 10), and between 11 and 21 d of age (n = 11). Mean values (±standard deviation, SD) for each parameter were calculated for both groups, and statistical analysis was applied to determine differences between the two age groups for each parameter (Student t-test for independent samples). Subsequently, the linear regression analysis was performed on the total number of calves to assess the possible effect and magnitude of each day of increasing age on the studied parameters. Statistical analysis was performed using Jamovi® ver. 2.3.21 for Windows. Significance was set for p < 0.05. 3. ResultsAmong the 21 calves, 12 were males and 9 were females. Mean age for the 1–10 d group was 4.3 d, while it was 15.4 d for the 11–21 d group. Statistic for the pooled data is shown in Table 1. Regression analysis showed that the age of the calf was associated with changes in some of the variables evaluated. Specifically, younger neonates (1–10 d) had significantly lower mean K+ concentrations (p < 0.001), and higher mean Na+ (p < 0.05) and Cl− concentrations (p < 0.05) compared to older calves (11–21 d). Regarding blood gas values, pCO2 was lower for younger calves (p < 0.05) while sO2 was higher (p < 0.05) compared to 11–21 d old calves.Linear regression analysis showed that the age of the calf, independently by the groups, was associated with changes in some of the variables evaluated, with diverse magnitude. In fact, K+ was shown to increase of 0.06 mmol/L each day of increasing age (p < 0.001); Cl− decrease of 0.26 mmol/L each day of increasing age (p < 0.05); sO2 decrease of 0.58% each day of increasing age (p < 0.05) and pCO2 showed a tendency to increase of 0.47 mmHg (p = 0.05). Mean concentrations of Na+ decreased significantly from younger to older dromedary calves (p < 0.05) but not linearly.4. DiscussionIn view of the increasing interest for the dromedary species and due to the paucity of knowledge regarding laboratory parameters, particularly when different ages are considered, this study focused the attention on physiological age-specific values of blood gases, electrolytes and acid-base in healthy dromedary calves up to 3 weeks of life. In the assessment of age as an independent variable, a single cut-off point of 10 d was applied, in view of previous findings on newborn bovine calves, which detected most significant changes in blood gases and electrolytes profile before and after 10 days of age [35]. The point-of-care (POC) analyzer employed in the present study was not specifically validated for camelids, but a recent study demonstrated that a similar POC analyzer can be a valid blood gas and biochemistry instrument for use with venous blood from clinically healthy New World camelids [36].The hypothesis that reference ranges for blood gas and electrolyte parameters should be specific for the age of the patient was proven in other species [35,37,38,39], and corroborated also in the dromedary species by the results of the present study. The age of the dromedary calves has been in fact significantly associated with differences in both blood gas analysis and electrolyte. Being the first study reporting blood gas, acid-base and electrolytes concentrations in the young dromedary calf, the comparison with previous results in age-matched dromedaries is unfeasible, while some discussion can be made in comparison with newborn dromedaries immediately after birth [26], with adult dromedaries [40], and with newborns of other species. Electrolytes concentrations were affected by the age of the calves. Mean concentrations of K+, Na+ and Cl− registered in 1–10 days old calves in the current study were similar to the mean concentrations reported by Tharwat et al. [26] in dromedary calves immediately after birth (4.4 ± 1 mmol/L, 156 ± 2.2 mmol/L and 116 ± 1 mmol/L, respectively). The significant rise of blood concentrations of K+ and the decrease of blood Cl− and Na+ concentrations from 1–10 d old calves to 11–21 d old calves in the present study seem to suggest a still on-going maturation process of homeostasis in the newborn. A study on donkey foals also showed an increase in K+ concentrations from samples collected within 1 day after birth to samples collected at day 14 and 21 of life [39], while no significant changes were found in Na+ and Cl− concentrations with age [39]. A study on equine foals showed no differences in the whole electrolyte profile during the first year of life [41], and therefore proposed combined average mean values for the electrolyte profile in foals aged <1 year. On the contrary, Viesselmann et al. [36] found that the age of the calf influenced the electrolyte profile in the bovine species, but with decreasing K+ and increasing Na+ and Cl− concentrations from 1–10 d old calves to 11–30 d old calves, in contrast with the present findings on the dromedary calf. Age-related differences in electrolytes concentrations have been reported for llamas [42]; reference ranges reported for Na+, K+ and Cl− in adult llamas (148–156 mmol/L, 4–5.3 mmol/L and 110–124 mmol/L, respectively) and in adult alpacas (140–156 mmol/L, 4–5.8 mmol/L and 104–119 mmol/L, respectively) were very similar to those found in the present study in the older dromedary calves group (11–21 d). This finding suggests that electrolytes concentrations in healthy calves older than 10 days of age are already comparable to those of adults [40].Blood gas analysis was performed on venous blood, and although some parameters related to the oxygenation ability are affected by the sampling site, some others were reported to be similar in arterial and venous blood in humans [43,44,45]. No correlation has been shown between arterial and venous blood samples for pO2 and sO2 in some animal species, underlining that reliable measurement for pulmonary gas exchange evaluation should only be performed on arterial blood samples [46,47]. Moreover, camelids have small, ellipsoid red blood cells with an unusually high concentration of hemoglobin, and this factor can affect the accuracy of oximetry in camelids [48]. For new world camelids (namely llama and alpaca), an arterial sample is recommended for the evaluation of respiratory parameters, preferably paired with a venous sample [49], as arterial parameters are influenced by of the high hemoglobin affinity for oxygen in these species. This implies that the present findings regarding sO2 on dromedary calves cannot reflect the effective percentage of hemoglobin fully combined with oxygen, lacking in clinical significance. Instead, studies in dogs [50,51] and equine [52,53] suggested that venous pH, pCO2 and actual HCO3- measurements are in sufficient agreement with the arterial value, provided that circulatory status is not impaired. Arterial sampling can be technically difficult [54,55] particularly in hypovolemic subjects, and carries the risk for some complications, especially when dealing with large animals under field conditions. The jugular vein can provide representative samples of whole-body metabolism, since this vessel is of large caliber and has normally relatively high blood flow [56]. Therefore, the choice for sampling venous blood in this research was made in order to obtain an accurate but also easily replicable method to assess blood gas, acid-base and electrolyte values in young dromedary calves, especially to support diagnosis in critical emergency situations. As the influence of body temperature on the values of some blood gas parameters has been previously demonstrated [46], a temperature-correction was applied for each sample analysis on the basis of the rectal temperature registered during the clinical examination; this is particularly important considering that pH, pO2 and pCO2 values are automatically corrected according to the patient temperature by the VETSTAT® analyzer.Mean pCO2 concentrations were affected by age in the present study, with increasing values in older dromedary calves; this could be partly ascribed to differences in respiratory rate according to age. The pCO2 values in 1–10 d old dromedary calves (43–57 mmHg) were similar to those reported for age-matched newborn bovine calves (43–58 mmHg) [35], but higher than those reported for newborn foals (42.2 ±2.1 mmHg) and donkey foals (37.1–38.9, [39]; 42.9–41.2, [47]). In these last studies however, researchers sampled the newborns immediately after birth and until 96 hours after birth, thus involving younger newborns compared to those of the present study. The high pCO2 values registered in the older dromedary calves group (11–21 d) in this study were similar to those from adult dromedaries [40], suggesting that high pCO2 can be a distinctive feature of the old world camelids, maybe an attribute adaptive for life in areas with high environmental temperatures and low water availability, and that this feature is already expressed in calves older than 10 days of age. Concentrations of tCO2 were not different in the two groups in the present study, and they were in agreement with those reported in the adult dromedary [40].No difference in pH values were found in the present study between the different age groups, underlining how lungs, kidneys, and the buffer system interact and respond to physiologic changes in order to tightly regulate blood pH [57]. The mean pH values of both groups in this study (7.34–7.38) were higher than the mean pH registered in the newborn dromedary calf at birth (7.29 ± 0.03) [26], and superimposable to the mean pH reported for adult dromedary (7.35 ± 0.03) [40]. When compared to newborns of other species, the dromedary calves show the lowest pH range values (7.26–7.4); pH ranges were in fact 7.37–7.5 in healthy newborn calves within 30 days after birth [35], 7.35–7.5 in the juvenile llama <1 year [49], and 7.41–7.48 in the donkey foals at 24 hours of age [39]. The fact that a similar pH range was registered in the adult dromedary (7.28–7.44) [40], suggests a distinctive trait of this species, and needs to be taken into account when a clinical evaluation of sick dromedary calves is performed, in order to avoid overestimation of acidemia or underestimation of alkalemia in this species. While a minimum pH cut-off value of 7.36 is suggested for diagnosing acidemia in bovine calves aged 7 to 26 d [58], the present study suggests that a pH value of 7.34–7.38 can be considered as normal for healthy dromedary calves between 1–21 d of age, as these values were found to be the mean pH values in the two groups of healthy dromedary calves enrolled. None of the other acid-base parameters significantly differed among calves of different ages. This may be due to the fact that, within few days after birth, the main metabolic adaptations to the extrauterine life have already taken place; when evaluating hematological and biochemical parameters in the newborn of many species according to age, in fact, most of the significant changes were detected during the very first hours after birth [47,59,60], mostly due to the effect of parturition and to the newborn adaptation to the extrauterine life. Regarding HCO3−, the mean and lower and upper limits registered in the present study were very similar to those reported in the adult dromedary [40] and in the newborn donkey foal [39,47], while lower and higher mean HCO3− concentrations were reported in 7-day-old foals [59] and in the calf [35], respectively. The BE, commonly used in evaluating the metabolic status of patients [61] ranged from −1.1 to 4.3, almost superimposable to the range reported for adult dromedaries (−2 to 4) [40]; when compared to newborns of other species, mean BE concentrations in the present study were lower than those reported in healthy bovine calves of similar ages [35], and more similar to findings on the newborn foals and donkey foals [47,56]. The linear regression analysis showed that a progressive increase of K+, decrease of Cl− and sO2, and borderline increase of pCO2 is seen when all the calves were considered as a whole unique group, highlighting that these changes occur linearly, along age progression.In the present study, some efforts were adopted to avoid possible confounding variables; a single farm was enrolled, and all dromedary calves were sampled on the same day, in order to reduce farm and environmental influences on blood results. All these aspects reduced the availability of calves that could be enrolled, representing a limitation of the study. Due to the special environmental conditions in which dromedaries are reared, and to their unique physiology especially regarding water metabolism, specific blood variable ranges are needed for this species, and according to the present study, age should be considered as an influencing factor. 5. ConclusionsAlthough the number of calves enrolled prevents from establishing definite reference ranges, the present study provides preliminary data for evaluating blood gas, acid-base and electrolyte concentrations in young dromedary calves. This is of fundamental importance in the diagnosis and correction of imbalances resulting from the most common diseases affecting the newborns during the first month of life, such as diarrhea. According to the present study, when compared to newborns of other species, the dromedary calf presents some unique characteristics, such as a lower pH range and higher pCO2 concentrations. Further studies on a greater number of calves are advisable to corroborate these results, and to evaluate possible influences of other variables such as gender, breed and ease of parturition on the blood gas, acid-base and electrolyte profile of the dromedary calf.

animals : an open access journal from mdpi

10.3390/ani13071147

PMC10093310

Bovine mastitis is an inflammation of the mammary gland in response to invasion by opportunistic agents. The objective of this work was to identify and evaluate the antibacterial resistance profile of mastitis milk samples, milking hands and milking equipment from small dairy farms belonging to northwest region of the state of Paraná, Brazil. Fifteen small, non-technical dairy farms in three municipalities, all belonging to the northwest region of the state of Paraná, Brazil, were selected. Of the 199 samples collected from the 15 selected properties in the municipalities of Boa Esperança, Tapejara and Juranda, 36.20% were classified as multiresistant. It is also worth noting the presence of an isolate of Enterobater agglomerans and one of Moellerella wisconsensis in the hands of milkers and milking machines, phenotypically producing extended-spectrum beta-lactamase (ESBL). As for the presence of the mecA gene, 72.72% isolated came from milk, 18.18% from insufflators and 9.1% from milking hands. Mastitis can be spread to the herd through the milking process by the milkers’ instruments and hands and adequate management measures can prevent its transmission and the conscious use of antibiotics decreases the prevalence of multidrug-resistant pathogens. The results of this work directly reflect on the health of the animals, the health of the workers and the health of the respective environment, which can make possible the continuity of the propagation of the etiological agents involved in the mastitis infection.

Bovine mastitis is an inflammation of the mammary gland in response to invasion by opportunistic agents. Due to the high economic importance of dairy production and the complexity related to animal health, the objective of this work was to identify and evaluate the antibacterial resistance profile of samples of mastitis milk, milking hand and milking equipment from small rural dairy farms belonging to the northwest region of the state of Paraná, Brazil. Five small, non-technical dairy farms in the municipalities of Boa Esperança, Juranda and Tapejara, all belonging to the northwest region of the state of Paraná, Brazil, were selected. The properties had Holstein and/or crossbred herds, carried out a bucket-by-foot milking system and all had the presence of animals with subclinical mastitis confirmed by the California Mastitis Test. Samples of sterile swabs from the milking insufflators, the milking hand and milk samples were collected—and later, isolation tests and phenotypic characterization of the samples, sensitivity tests to antimicrobials and phenotypic tests for the detection of beta-producing strains were performed with extended-spectrum beta-lactamase (ESBL), molecular identification of Staphylococcus aureus isolates and mecA gene research. Of the 199 samples collected from the 15 selected properties in the municipalities of Boa Esperança, Tapejara and Juranda, 72 (36.20%) were classified as multiresistant. Isolated from milkers’ hands and milking machines, which phenotypically produce extended-spectrum beta-lactamase (ESBL), the presence of the mecA gene was also observed in 11 isolates of Staphylococcus spp. of milk samples, machines and milking hands. Mastitis can be spread to the herd through the milking process by the milkers’ instruments and hands, and adequate management measures can prevent its transmission and the conscious use of antibiotics decreases the prevalence of multidrug-resistant pathogens. In this work, different pathogenic bacteria were detected in mastitic milk, milking equipment and milking hand with a high percentage (36.20%) of isolates classified as multidrug resistant. In addition, the presence phenotypically (ESBL) and molecularly (mecA gene) of isolates carrying resistance genes was also verified. These results directly reflect on the health of the animals, the health of the workers and the health of the respective environment, which can enable the continuity of the propagation of the etiological agents involved in the mastitis infection. The awareness of producers and workers on these properties about the disease, transmission, sanitary aspects and adequate management and treatment are essential for improving milk production and production efficiency.

1. IntroductionBrazilian cattle raising has great prominence in world dairy production, responsible for the production of 34 billion liters per year, generating high financial turnover and employment, with more than 415,000 people make a living from the activity in the country and about 170,000 considered small producers, who milk up to 250 L per day [1,2].Bovine mastitis generates high economic losses. It is estimated that the economic losses caused by mastitis, from its clinical manifestation to its cure, range from $60.23 to $123.50 per cow [3]. As a result, there is a reduction in milk production, expenses with medicines, veterinary assistance, decrease in milk quality and disposal of contaminated milk and animals with chronic infection [4].Bovine mastitis is an infectious disease characterized by inflammation of the mammary gland in response to the invasion of opportunistic agents, mainly caused by bacteria [5]. This disease can manifest itself in a subclinical, clinical or chronic form, which can severely compromise the general condition of the animal [6].The clinical form causes visible changes in the udder, evident signs of inflammation, associated with changes in milk composition [7]. The subclinical is characterized by an asymptomatic infection or only a decrease in milk production can be observed [8]. In the chronic form, there is fibrosis—with an absence of signs of the inflammatory process and changes in the milk, the result of a persistent process—which may form fistulas in the affected mammary gland [9].The microorganisms involved in the infection can be from contagious or environmental source [9]. The contagious ones live and multiply inside the mammary gland and their transmission occurs horizontally, during the milking process, through milking machines that indicate hygiene failure, through the hands of milkers and multiple-use towels [10]. On the other hand, environmental microorganisms live in the environment where the animals are raised and their transmission occurs between milkings, through contact of the teats with the contaminated environment, favoring the penetration of pathogens into the mammary gland [11].The main factors that contribute to the spread of the disease in the herd are related to the history of the presence of mastitis and treatments used, type of milking, failure to evaluate the first three jets of milk, failures in teat disinfection and drying, incorrect cleaning and maintenance (and use) of equipment and lack of hygiene of the milkers’ hands [12].Due to the high economic importance of dairy production and the complexity related to animal health, the aim of this study was to identify and evaluate the antibacterial resistance profile of mastitis milk samples, milkers’ hands and milking equipment from small rural dairy farms belonging to the northwest region of the state of Paraná, Brazil.2. Material and Methods2.1. Criterion for Inclusion of Rural Properties, Origin and Number of SamplesFive small, non-technical dairy farms were selected, with a bucket milking system, intended for dairy production with Holstein and/or Crossbred herds, with the presence of animals with subclinical mastitis confirmed by the CMT test (California Mastitis Test). Eight to ten animals were included in this research, with one milker and two samples of milking machines before and two after the start of milking on each rural property located in the municipalities of Boa Esperança, Juranda and Tapejara, all belonging to the northwest region of the state of Paraná, Brazil.2.2. Sample CollectionFrom May to June 2022, samples were collected from the milking machines and the milker’s hand using sterile swabs containing AIMES medium + Activated Charcoal (Copan Transystem©, Brescia, Italy), which were collected through rotational movements of the swabs in the insufflators of the milking machine and the hands of the milkers in the respective places.Before proceeding with the collection of milk from the animals, they were submitted to the CMT and the animals diagnosed with two crosses in the respective tests were selected to collect the milk samples using sterile flasks; the extraction was performed manually.The milkers were also invited to complete the epidemiological questionnaires of the property and the milker for a possible explanation of the procedures performed in the milking routine, hygiene and maintenance of equipment, personal hygiene of the milker, along with the questionnaire, and monitoring of milking was carried out to observe the important points for the spread of mastitis in the herd.After collection, the biological materials were kept under refrigeration for a period not exceeding 24 h and sent to the Laboratory of Preventive Veterinary Medicine and Public Health of the Graduate Program in Animal Science with Emphasis on Bioactive Products at Universidade Paranaense (UNIPAR) for processing and the carrying out of diagnostic techniques.In total, 199 biological samples were collected—45, 42 and 35 milk samples, respectively—from the municipalities of Boa Esperança, Juranda and Tapejara. Five samples from the milker’s hand in each municipality and 20, 20 and 22 samples from milking machines, respectively, from the municipalities of Boa Esperança, Juranda and Tapejara were collected, totaling 122 milk samples, 15 hand samples from milkers and 60 samples from milking machines.2.3. Culture, Isolation and Phenotypic Characterization of SamplesSamples were inserted into brain heart infusion (BHI) medium and incubated at 37 °C for 24 h. Then, they were sown using the depletion technique on plates containing blood agar medium and incubated at 37 °C for up to 48 h for bacterial isolation.The predominant isolate was selected (where the bacterial colony was visually in greater quantity on the plate), and it was submitted to the analysis of the macroscopic and microscopic characteristics and catalase and coagulase tests, allowing the classification into positive coagulase Staphylococcus (CoPS) and coagulase negative Staphylococcus (CoNS) [13,14].The biochemical identification of bacteria belonging to the Order Enterobacteriales was performed using the “Kit for Enterobacteria” (NewProv®, Pinhais, PR, Brazil), according to the manufacturer’s recommendations.2.4. Antimicrobial Sensitivity TestsAntimicrobial susceptibility tests were performed according to the criteria of the Clinical and Laboratory Standards Institute. The antibiotics were chosen based on the European Medicines Agency’s “Categorization of antibiotics for use in animals for prudent and responsible use”. Bacterial isolates using the disk diffusion test were evaluated against Amoxicillin + Clavulanate (30 µg), Amikacin (30 µg), Amoxicillin (10 µg), Cefoxitin (30 µg), Clindamycin (2 µg), Chloramphenicol (30 µg), Ceftiofur (30 µg), Doxycycline (30 µg), Erythromycin (15 µg), Meropenem (10 µg), Oxacillin (1 µg) and Rifampicin (5 µg). All strains classified as intermediate were considered non-susceptible as determined by the World Health Organization. Strains that showed resistance to three or more classes of antimicrobials were considered multidrug resistant (MDR) [15].2.5. Phenotypic Test for Detection of Extended Spectrum Beta-Lactamase (ESBL) Producing StrainsThe phenotypic test for the detection of enterobacteria producing extended-spectrum beta-lactamases (ESBL) was performed by the synergic double disc test with cefotaxime (30 μg), ceftazidime (30 μg), ceftriaxone (30 μg) and aztreonam (30 μg). The disks were distributed at a distance of 20 mm from a disk containing amoxicillin + clavulanate (20/10 μg). Any increase or distortion of the inhibition zone of one of the antibiotics toward the amoxicillin + clavulanate disk was considered suggestive of ESBL production [16].2.6. Molecular Identification of Staphylococcus aureus IsolatesPolymerase chain reaction (PCR) was performed on coagulase positive Staphylococcus (CoPS) samples to verify which of these isolates were Staphylococcus aureus. The DNA was extracted with the PurelinkGenomic DNA Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s information and the reactions were performed using the primer Sa442-1 (5′-AAT CTT TGT CGG TAC ACGATA TTC TTC ACG-3′ and the primer Sa442-2 (5′-CGT AAT GAGATT TCA GTA GAT AAT ACA ACA-3′) following the methodology [17]. For DNA amplification, an AppliedBiosystems Veriti™ 96-Well ThermalCycler (Waltham, MA, USA) was used.The amplification of the products was visualized by electrophoresis on a 2% agarose gel stained with Gel Red (Uniscience, Osasco, SP, Brazil) using a molecular marker of 100 pb and the products were visualized as a single band of 241 pb.2.7. mecA Gene ResearchThe DNA of Staphylococcus spp. classified as resistant to oxacillin was extracted using the Purelink Genomic DNA Kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s information and the PCR reactions were performed using the mecA1 primer (AAAATCGATGGTAAAGGTTGG) and mecA2 (AGTTCTGCAGTACCGGATTTG) [18]. For DNA amplification, an AppliedBiosystems Veriti™ 96-Well Thermal Cycler (Waltham, MA, USA) was used.The amplification of the products was visualized by electrophoresis in a 2% agarose gel stained with Gel Red (Uniscience, Osasco, SP, Brazil) using a molecular marker of 100 pb and the products were visualized as a single band of 533 pb.3. ResultsA total of 199 samples were collected from the 15 selected properties, 70 from the municipality of Boa Esperança, 67 from the municipality of Juranda and 62 from Tapejara.Regarding the bacterial growth of the samples, the municipality of Boa Esperança had the highest percentage among the three, with 71.43% (50) of the samples with growth (Table 1), followed by the municipality of Tapejara with 56.45% (35) samples (Table 2), and Juranda with 47.76% (32) of the samples with bacterial growth (Table 3).Regarding microorganisms, in the municipality of Boa Esperança, non-aureus CoPS (12.85%), Staphylococcus aureus (22.85%) and Escherichia coli (5.71%) were detected. In the municipality of Tapejara, non-aureus CoPS (6.45%), Staphylococcus aureus (4.53%) and Escherichia coli (6.45%) were also detected. In the municipality of Tapejara, non-aureus CoPS (6.45%), Staphylococcus aureus (4.53%) and Escherichia coli (6.45%) were also detected. In the municipality of Tapejara, non-aureus CoPS (6.45%), Staphylococcus aureus (4.53%) and Escherichia coli (6.45%) were also detected. In the municipality of Juranda, non-aureus CoPS (5.97%), Staphylococcus aureus (2.98%) and Escherichia coli (19.40%) were detected.In these municipalities, there was also the presence of isolates resistant to the antibiotics tested mainly against oxacillin, with resistance of 78%, 64.86% and 81.25%, respectively, in the municipalities of Boa Esperança, Tapejara and Juranda, which also showed resistance to clindamycin in 44%, 72.97% and 62.5%, respectively, in addition to other antibiotics with significant numbers of resistant isolates, such as erythromycin, cefoxitin and rifampicin (Table 4).Among the 199 (100%) isolates, 72 (36.20%) were classified as multidrug resistant—that is, resistant to at least 1 drug from 3 or more classes of antimicrobials [17]. Of this total, 43 (59.72%) came from equipment used in milking, 20 (27.78%) came from milk and 9 (4.52%) came from the hands of milkers.It is also worth mentioning, within these results, the presence of an isolate of Enterobacter agglomerans and one of Moellerella wisconsensis, having been, respectively, isolated from the hand of a milker and an insufflator, phenotypically producing ESBL (Figure 1).In addition to the presence of phenotypically resistant isolates, the presence of the mecA gene was found in 11 of the Staphylococcus spp. isolates, 8 (72.72%) from milk, 2 (18.18%) from milking machine and 1 (9.1%) from the hand of a milker (Table 5 and Table 6).4. DiscussionMastitis is an endemic and economically important pathology, mainly in family farms, which are farms with low production; milkers do not have technical knowledge and guidance from trained professionals.Related to bacterial growth, the municipality of Boa Esperança had the highest percentage, among the three municipalities studied, with 71.43% of the samples with growth. Some authors, such as [5], described the main points for mastitis control as the correct hygiene and periodic maintenance of milking equipment. Ideally, it should be checked every six months of its operation, in addition to changing the hoses that come into contact with the milk and the milking machines, as well as the hygiene of the milker and the use of gloves.The results demonstrate the lack of hygiene and the failure to carry out periodic maintenance of equipment used in the milking process and possibly have contributed to the spread of mastitis-causing agents in the herd, because the insufflators of the teat cup sets were dirty before and after milking, as well as their maintenance only being carried out when there is a defect in their operation. Furthermore, most properties go years without carrying out this maintenance, in addition to the presence of poor hygiene (presence of mud and feces) in the facilities where the animals remained after milking.In addition to the situations above, this high bacterial growth may also be related to the failure to pre-dip and dry the animals’ teats. The teats were dirty even after carrying out the pre-dipping and the drying of the teats was not carried out completely. In turn, failure in this management can lead to contamination of the insufflators of the teat cup sets, and consequently, contribute to the dissemination of different pathogenic agents, as already described by [10].Another factor that may have contributed to the increase in this bacterial growth was the lack of hygiene of the milkers (due to lack of knowledge). It was observed that when starting the milking process they did not have the habit of washing their hands, and even during the process, they maintained contact with different body parts of the animals and handled the ropes and gates before returning to the milking routine, a situation that could favor cross-contamination (man × animal × environment). A study, conducted by Shin et al. [11], points out that the personal hygiene of the milker associated with adequate handling, especially during milking, can decrease the number of animals affected by clinical and subclinical mastitis, reduce the rate of new infections, improve the somatic cell count (SCC) of the herd and the quality of the milk produced.Algharib et al. [1] point out that recommending the training of milkers on good practices for performing milking, principles of personal hygiene and the correct use of milking equipment are essential for controlling mastitis in the herd. Another factor that can be considered relevant to the result of this study is that most of the milkers in that municipality were over 50 years old and had been exercising this activity for many years; however, they did not seek and evolve in terms of obtaining correct instructions/technical information to improve their work routine, milk production and, consequently, the health of the herd.The municipality of Tapejara was the second municipality with the highest bacterial growth in the collected samples, with 56.45% of the samples, followed by Juranda with 47.76%. This decrease in bacterial growth can be highlighted by better hygiene of the equipment used in the milking process. From the beginning and end of the milking process, the equipment was clean, in addition to the observed efficiency in carrying out pre-dipping and drying from the ceilings. When starting the milking process, milkers wash their hands and avoid touching dirty utensils. In these two municipalities, most properties received private veterinary technical assistance and producers have a basic technical understanding of good practices in milk production.In this study, a common situation in all properties of the three municipalities is the non-adoption of the milking line, which possibly may have favored the spread of mastitis to the herd. This must be instituted to avoid the transmission of pathogenic agents from an infected animal to healthy animals—that is, healthy animals must be milked first, followed by animals with mastitis, a situation that is also reported in studies by [5].In the municipality of Boa Esperança and Tapejara, Non-aureus CoPS, 12.85% and 6.45%, respectively, and Staphylococcus aureus, 22.85% and 4.53%, were isolated in the municipality of Juranda for Non-aureus CoPS 5.97% and Staphylococcus aureus 2.98% was isolated less frequently. These agents are associated with subclinical mastitis of contagious origin—that is, they are easily disseminated by the herd, they are passed unnoticed because they do not present evident clinical signs and in some cases a decrease in milk production [8,11].The Non-aureus CoPS treatment is carried out in drying and has good responses to conventional dry cow therapies, since Staphylococcus aureus is the causative agent of mastitis with the lowest percentage of cure, due to its protective barrier, which is the ability to form an abscess inside the mammary gland, making the action of antibiotics difficult, in addition to being methicillin resistant bacteria, their treatment is carried out in the drying process with the use of an intramammary antibiotic associated with an injectable [6,12].Non-aureus CoPS and Staphylococcus aureus can be controlled or even stopped from herd transmission through the adoption of the milking line, starting milking with animals negative for mastitis in the dark bottom cup test and in the CMT, followed by animals with mastitis or by the method of disinfecting the insufflators with disinfectants. When milking the animal before moving on to the next one, the insufflators can be sprayed or immersed in the disinfectant [3,4]. Another very important method for the control and eradication of this agent is good personal hygiene practices, such as washing hands before starting the milking process and whenever touching utensils and dirty objects or the use of gloves are measures that reduce the spread of this agent by the herd [1].Authors [19] isolated 436 samples of raw milk from animals with clinical and subclinical mastitis from 3 farms and identified 135 samples of Staphylococcus aureus. Additionally, Ref. [20] isolated 23 samples of Staphylococcus aureus from a total of 48 samples of bovine milk diagnosed with subclinical mastitis—which are similar results to this work, which featured isolated Staphylococcus aureus from (16/70) samples from the municipality of Boa Esperança, (3/62) samples from Tapejara and (2/67) samples from Juranda.Authors [21] evaluated in their work 1,549 milk samples from 952 cows, including cows with recurrent mastitis. Non-aureus Staphylococcus (NaS) (27.6%) was isolated, followed by Escherichia coli (18.9%) and Staphylococcus aureus (7.7%) of milk samples. In this study, it was possible to isolate 12.85% of samples with non-aureus CoPS from the municipalities of Boa Esperança, 6.45% from Tapejara and 5.97% from Juranda, followed by 22.85%, 4.53% and 2.98%, respectively, positive for Staphylococcus aureus (7.7%).Researchers [22] collected 400 samples of bovine subclinical mastitis milk and in their analysis of 173 (43.25%) E. coli isolates was detected, which corroborates this work where in the municipality of Juranda it was the one that presented the highest percentage of Escherichia coli, with (13/67) 19.40%, followed by Tapejara with (3/62) 6.45% and Boa Esperança with (4/70) 5.71%.Escherichia coli is one of the causative agents of environmental clinical mastitis, they are bacteria that live in the environment, so the main way to prevent mastitis caused by this agent is to keep the place where the animals are destined after milking clean, free of manure and sludge [5,19].Mastitis caused by Escherichia coli has clinical signs such as a drop in milk production, fever, dehydration, severe depression, loss of appetite and signs of inflammation in the udder, which can lead to severe infections and even death of the animal [7]. Mastitis caused by Escherichia coli is unlikely to cure spontaneously, requiring treatment with antibiotics and cephalosporins which have been widely used in the treatment of mastitis, including cephalexin and ceftiofur. Other widely used antibiotics are amoxicillin, erythromycin, gentamicin and penicillin. Animals that are vaccinated against strains of Escherichia coli have a better cure rate for treatments with antibiotics [4,8].Related to antimicrobial susceptibility, in the properties of the three municipalities of this study, bacterial isolates resistant to different antibiotics were detected, mainly against oxacillin (with resistance of 78%, 64.86% and 81.25%), clindamycin (44%, 72.97% and 62.5%) and erythromycin (9.95%, 15.67% and 11.33%) in the municipalities of Boa Esperança, Tapejara and Juranda, respectively. Furthermore, Dea et al. [3] also showed in their study antimicrobial susceptibility in 202 isolates of S. aureus from samples of mastitic bovine milk and only identified resistance to penicillin (12.4%) and erythromycin (0.5%); however, they did not detect resistance to oxacillin, a situation that differs from the results of this work.It was also identified, in this work, a higher percentage of resistance to clindamycin in 44%, 72.97% and 62.5%, respectively, in the municipality of Boa Esperança, Tapejara and Juranda. Another important antibiotic with important numbers of resistant isolates was erythromycin.Molineri et al. [12] aimed to determine the prevalence of phenotypic resistance of S. aureus to antimicrobial agents collected worldwide, in the context of bovine intramammary infections between the years 1969 and 2020, whereby the highest global prevalence of resistant S. aureus was for the penicillin class, followed by clindamycin, erythromycin and gentamicin. The ceftiofur and cephalothin class showed the lowest global prevalence of antimicrobial resistance, highlighting the results obtained in our work.Different classes of antibiotics are used to treat animals with mastitis. The most commonly used first-line treatments are penicillins (oxacillin, amoxicillin, methicillin), alone or associated with aminoglycosides (amikacin, gentamicin, streptomycin, neomycin), macrolides (erythromycin, azithromycin), lincosamides (clindamycin), fluoroquinolones (norfloxacin), tetracyclines (doxycycline, oxytetracycline) and cephalosporins (ceftiofur) [4,10,11].Antimicrobial resistance is the ability of a microorganism to grow or survive in the presence of an antimicrobial at a concentration that is generally sufficient to inhibit or kill microorganisms of the same species [20]. Nader et al. [6] reported in their work that the indiscriminate use of antimicrobials, with the wrong concentration, dosage or application interval, has increased and accelerated this resistance process, making it a serious threat to public health. One of the main causes of antimicrobial resistance in dairy cattle is the fact that it is associated with the use of antibiotics for the treatment of mastitis, whether in the lactation phase or in the drying process of the animal, associated with its erroneous use in the production [12].Dea et al. [3] justify that the increase in antimicrobial resistance in mastitis-causing agents is caused due to the attempt to treat mastitis conditions without knowing which agent caused the infection, thus the treatment in the dark develops new agents with genetic mutation making them more resistant the drugs of choice used. The conscientious use of antibiotics in the production line, such as the use of the active ingredient of antibiotics that is effective and recommended for the causative agent, in addition to dosage, concentration and correct period, are methods of prevention that mitigate the emergence of new causative agents of genetically resistant mastitis [7].Molineri et al. [12] also analyzed all antimicrobials which showed a growing pattern over time, being more evident from 2009 that the antimicrobials with the highest prevalence of resistance over the years were clindamycin, gentamicin and oxacillin, which is similar to the results of this study where oxacillin and clindamycin were the most prevalent.In this study, a result that draws attention is the prevalence of 72 (36.20%) bacterial isolates classified as multi-drug resistant. Of this total, 45 (62.50%) come from the equipment used in milking, 18 (9.04%) from the milk and 9 (4.52%) from milk workers’ hands.In our study, 18 (9.04%) multiresistant samples isolated from milk were identified and a similar result was detected [20], where they isolated 48 Staphylococcus aureus from samples of mastitis bovine milk and all of them were multi-drug resistant. Related to the presence of resistant multidrug isolates [4], they also detected it in different species of Streptococcus.Taniguchi et al. [21] isolated Listeria spp. of 79 (26.3%) of the 300 samples, including 29 (36.7%), 32 (40.5%) and 18 (22.8%) isolates found in raw milk, milking equipment and hand swabs of milkers, respectively, with 88% of the total multidrug resistant isolates, a result similar to our work, where 72 (36.20%) samples were classified as multidrug resistant from milk sample isolates, swabs from milking equipment and milkers’ hands.It is also worth mentioning the presence of an isolate of Enterobacter agglomerans and one of Moellerella wisconsensis, having been isolated from milking hand and insufflator hand, respectively, phenotypically producing ESBL [21], isolated from 1549 milk samples from mastitis cows and 952 of the samples with bacterial growth. The incidence of ESBL-producing extended-spectrum β-lactamase (ESBL) was 1.4% of all samples and 1.4% of Klebsiella pneumoniae enterobacteria were identified. Additionally [23] evidenced in their study of mastitis milk samples the occurrence of Escherichia coli (118/372, 31.7%) and Klebsiella pneumoniae (77/372, 20.7%), two environmental pathogens known to cause bovine mastitis. When searching for extended-spectrum beta-lactamases (ESBLs) agents were detected in selective medium in (3/118, 1.59%) Escherichia coli and (6/77, 7.79%) Klebsiella pneumoniae, emphasizing the results obtained in our work. Enterobacteriaceae are more frequently present in properties with precarious hygienic-sanitary conditions, with a lot of fecal contamination, accumulation of manure and sludge [22].The inappropriate use of antimicrobials can lead to the emergence of resistant strains. Enterobacteria are among the main etiological agents of environmental bovine mastitis and are often resistant to antimicrobials, especially to β-lactams due to the production of beta-lactamases and some enterobacteria produce broad-spectrum beta-lactamases (ESBL), constituting the bacteria the ability to degrade β-lactams, in addition to presenting broad spectrum on various antimicrobials such as ceftazidime, cefotaxime and aztreonam, which are strongly inhibited by clavulanic acid [4,10].Haenni et al. [10] suggest that cows should have access to feed immediately after milking, in order to keep them standing until the teat dries and the striated canal closes completely, preventing possible contamination after milking by agents from the environment. This method is carried out in all three municipalities—the animals are kept standing if fed for more than an hour after milking.mecA gene expression is constituted or induced by beta-lactam antibiotics, such as oxacillin and cefoxitin. The mecA gene is inserted into the staphylococcal chromosome through a mobile genetic element called the chromosomal staphylococcal cassette. The mecA gene sequence is highly conserved in strains of Staphylococcus aureus and Staphylococcus spp. negative coagulase. The Staphylococcus spp. are pathogens that cause bovine mastitis and may present multiple resistance to different antimicrobial groups.Carvalho et al. [24] aimed in their study to phenotypically identify isolates of Staphylococcus spp. obtained from bovine milk and to characterize its antimicrobial resistance profile. Of the 101 strains isolated, the mecA gene was detected in 27% of the milk samples. In our studies, similar results were obtained, which noted the presence of the mecA gene in 11 of the isolates of Staphylococcus spp., 8 (72.72%) isolated from milk, 2 (18.18%) from insufflators and 1 (9.1%) from the hand of a milker. Other studies also detected the presence of the mecA gene in milk, which showed the importance of this gene for human, animal and environmental health [25,26,27].A study conducted by Dea et al. [3] recommended the treatment of dry cows, with the aim of reducing subclinical infections and preventing new infections in the dry period that occurs between two lactations, which is the phase in which the animal was in productive rest, an important situation for the health of the animal’s mammary gland to regenerate for the next lactation. This can last an average of 60 days, counting from drying to delivery. In the first post-drying weeks, the risk rate for new infections is very high and the treatment of subclinical mastitis has higher cure rates compared to treatment during lactation. According to the study carried out in this work, all properties in the three municipalities adopt drying therapy, which helps in the control and treatment of possible subclinical infections in animals.One of the main mastitis control methods is dry cow therapy, adopting the milking line, keeping cows standing after milking, washing the milk pool correctly and the application of a treatment protocol, effectiveness in the pre- and post-dipping procedure and the use of gloves by the milkers. An important role of veterinarians is to advise on the importance of mastitis in the production line, the economic losses it entails, how to prevent it and the correct treatment. Furthermore, much of this assistance is oriented around management planning to maintain or improve the health status of the family properties.5. ConclusionsIn this study, different pathogenic bacteria were detected in mastitis milk, milking equipment and milker’s hand with a high percentage (36.20%) of isolates classified as multidrug resistant. In addition, the presence phenotypically (ESBL) and molecularly (mecA gene) of isolates carrying resistance genes were also verified. These results directly reflect on the health of the animals, the health of the workers and the health of the respective environment, which can make possible the continuity of the propagation of the etiological agents involved in the mastitis infection. The awareness of producers and workers on these properties about the disease, transmission, sanitary aspects and adequate management and treatment are essential for improving milk production and production efficiency.

animals : an open access journal from mdpi

10.3390/ani11071935

PMC8300113

Poultry can be classified as broilers for meat production and layers for egg production. Modern poultry farming improved economically important traits of broilers and layers by breeding and genetic selection. Myostatin (MSTN) has gained attention as a potential selection marker for higher meat production in the poultry industry, because MSTN mutant chickens and quail showed increased muscle mass. In this study, the effect of MSTN mutation on egg production was investigated to evaluate potential use of MSTN for higher egg production in the layer industry. MSTN homozygous mutant quail showed a significantly delayed onset of egg laying, a higher egg weight, and a lower number of eggs produced during the active laying period compared to wild-type quail. However, there were no significant differences in total egg production for 20 days, percentage proportion of egg white and yolk in egg weight, and egg fertility, and hatchability between MSTN mutant and WT quail. Although a clear benefit on egg production by MSTN mutation in quail was not revealed, this study provided useful information to understand the productive performance of MSTN mutant hens.

Increased body weight and muscle mass, along with improved feed efficiency, by myostatin (MSTN) mutation in quail, supports the potential use of MSTN as a selection marker for higher meat yield in the poultry industry. Although economically important traits of broilers have been studied using recently generated MSTN mutant quail, the effect of MSTN mutation on egg production has not yet been investigated. In this study, several economically important traits of layers, including egg production, reproduction, and body composition of hens, were compared between MSTN homozygous mutant, heterozygous mutant, and wild-type (WT) quail. In terms of egg production, MSTN homozygous mutant quail, showing significantly delayed onset of egg laying, laid significantly heavier eggs, but a significantly lower number of eggs compared to WT quail for 20 days after 3 months of age, resulting in similar total egg production among groups. In addition, the percentage proportion of egg white and yolk in egg weight were similar among groups. Furthermore, similar fertility and hatchability of eggs from MSTN homozygous mutant breeding pairs and WT breeding pairs indicated normal reproductive function of MSTN mutant quail. These findings will provide scientific rationales for the consideration of MSTN as a potential selection marker for layers in the poultry industry.

1. IntroductionSince the discovery of the anti-myogenic function of myostatin (MSTN) in mammals, MSTN was considered as a potential selection marker for a higher meat yield in the livestock industry. Unlike other MSTN mutant mammals reported and studied for a relatively long time [1,2,3], genome-editing of the MSTN gene in avian species including quail and chickens has been recently generated and studied mainly regarding growth rate and development of muscle and adipose tissue [4,5]. Therefore, other economically important traits of poultry need to be further investigated to consider the usage of MSTN as a potential selection marker in the poultry industry. In addition to increased body and muscle weight, improved feed efficiency in MSTN homozygous mutant quail was also reported in our recent study [6]. Although previous studies demonstrated positive effects of MSTN mutation on economically important traits of broilers, the effect of MSTN mutation on egg production traits of layers has not been reported. In addition to the meat production of broilers, commercial egg production of layers is an important sector in the poultry industry. Therefore, an increase in egg production has been achieved by improving layer genetics and a farm management system [7]. In modern commercial poultry farming, optimal management systems have been practiced to control factors affecting egg production such as environmental factors of poultry farms and physical characteristics of egg laying hens. Especially, feed is one of the main concerns in poultry farming because not only does feed comprise a major portion of the total production cost in the poultry industry, but also body weight and fatness affect the onset of egg production [8]. Body weight has been known to affect egg size positively and fat accumulation is increased prior to egg laying [9,10]. However, Overweight with excessive fat accumulation could negatively affect egg laying performance [11]. In our previous study, body and fat weight of MSTN homozygous mutant females were approximately 15% heavier and 30% lower than wild-type (WT) females, respectively, at 6 weeks old, when egg laying is about to be initiated in quail [4]. Because MSTN mutation affects both body weight and fatness of hens, the effect of MSTN mutation on egg production of MSTN homozygous mutant quail needs to be investigated.In addition to the productive traits, functional traits of layers, such as fertility and hatchability of eggs, are also important for higher profitability in poultry industries. Although the direct effect of MSTN mutation on reproductive performance in birds has not been reported, fertility and hatchability were significantly higher in the small egg size group compared to the large egg size group in broiler breeder chickens [12]. Likewise, the egg weight of MSTN mutant quail and resultant fertility and hatchability can be affected by MSTN mutation. These unknown effects of MSTN mutation on reproductive traits of avian species can now be studied using recently generated MSTN mutant quail. Therefore, the aim of this study was to investigate the effect of MSTN mutation on egg production traits of female quail by examination of age of first egg, egg weight, total number of eggs produced during the active laying period, weights of egg white and yolk, and egg fertility and hatchability.2. Materials and Methods2.1. Animal CareAll animal care protocol and experimental procedures were approved by the Institutional Animal Care and Use Committee at The Ohio State University (Protocol 2019A00000024). Japanese quail (Coturnix japonica) with MSTN mutation were generated from our previous study [4] and maintained at The Ohio State University Poultry Facility in Columbus. After raising quail as groups in brooder cages, 5 weeks old quail were transferred to small individual cages (12″ × 16″ × 16″) to record first egg laying dates and collect eggs from each quail. All quail were fed the same standard chicken diet produced by The Ohio State University’s Research Feed Mill in Wooster and euthanized via CO2 inhalation. 2.2. Egg Collection and Tissue SamplingAfter the initiation of laying approximately at 6 weeks old, quail were raised until 3 months of age to give enough time for stabilization of egg weight and number during the active laying period. Then, eggs were collected daily for 20 days from 10 homozygous and 9 heterozygous mutants, and 9 WT female quail at 3 months old to count the numbers of eggs and measure the weights of whole eggs, egg whites, and egg yolks. Total egg amount was calculated by multiplying average egg weights and average number of eggs produced for 20 days. After collection of eggs, pectoralis major, leg fat, and abdominal fat were sampled from the same quail at 4 months old. To analyze the onset of egg laying, additional female quail were raised and dates of first egg laying were recorded from 25 MSTN homozygous and 40 heterozygous mutants, and 22 WT female quail.2.3. Egg Fertility Another batch of MSTN homozygous mutant and WT quail groups were prepared. At 3 months old, 9 pairs of MSTN homozygous mutant male and female quail and 7 pairs of WT male and female were mated in separate small cages, respectively. Then, eggs were collected daily from each breeding pair and incubated weekly. Egg fertility and hatchability were measured by counting total incubated eggs, hatched quail, and developed embryos from unhatched eggs.2.4. Statistical AnalysesAll data were expressed as means ± SEM. The Student t-test was used for statistical analysis to compare fertility and hatchability of eggs. One-way ANOVA followed by Tukey’s multiple comparisons test were used for comparisons of onset of egg laying, egg weights, numbers, and egg white and yolk among homozygous and heterozygous mutants, and WT groups. All statistical analyses were performed by the GraphPad PRISM (version. 6.02, Graphpad Software, La Jolla, CA, USA) and detailed numbers of samples were described in each of the Figure Legends. 3. Results3.1. Egg Production Traits of MSTN Mutant QuailAverage ages of hens reaching onset of first egg laying were very similar between WT (48.45 days) and MSTN heterozygous mutant hens (48.73 days). However, MSTN homozygous mutant quail had approximately 5 days of significant delay in onset of egg laying (53.08 days) compared to heterozygous mutant and WT quail (Figure 1). In terms of egg weight, homozygous mutant quail laid significantly heavier (7%) eggs compared to WT quail, and the weights of eggs from heterozygous mutant quail were intermediate (Figure 2A). On the contrary, total number of eggs produced for 20 days during the active laying period were significantly lower in homozygous mutant quail compared to heterozygous mutant and WT quail (Figure 2B), resulted in similar total amount of eggs among groups (Figure 2C). Although the egg yolk weights of homozygous mutant quail were significantly heavier than the one from WT quail (Figure 2D), the percentage proportion of egg white and yolk in egg weight were similar among groups (Figure 2E), showing an increase of egg white and yolk weights in proportion to the egg weight of different genotypes. 3.2. Fat and Muscle of MSTN Mutant QuailBody and pectoralis major muscle weights of MSTN homozygous mutant quail were significantly heavier than those of heterozygous mutant and WT quail at 4 months old (Table 1). On the contrary, leg and abdominal fat weights and percentages of homozygous mutant quail were significantly lower than those of WT quail. However, leg and abdominal fat weights and percentages of MSTN heterozygous mutant quail were intermediate between those of homozygous mutant and WT quail.3.3. Fertility and Hatchability of Eggs from MSTN Mutant QuailThere was no significant difference in fertility and hatchability of eggs from MSTN homozygous mutant breeding pairs and WT breeding pairs (Table 2). 4. DiscussionThe positive correlation between egg weight and body weight of hens has been reported in chickens and turkeys [9,13]. Furthermore, the quail line selected for increased body weight laid heavier eggs compared to quail line with selection for decreased body weight [14]. Likewise, the average egg weight during the active laying period of MSTN homozygous mutant quail was significantly greater than that of WT quail (Figure 2A) and the body weight of homozygous mutant females was approximately 15% and 10% higher than that of WT females at 6 weeks old [4], right before the onset of egg laying, and at 4 months old, active laying period, respectively.On the contrary, total number of eggs produced for 20 days during the active laying period were significantly lower in MSTN homozygous mutant quail compared to WT quail (Figure 2B). Similarly, egg production of the quail line selected for higher body weight was decreased [14]. However, similar total egg production (Figure 2C), with similar percentage proportion of egg white and yolk (Figure 2E), for 20 days among groups might indicate the capacity of egg production might not be affected by MSTN mutation, but it took more time to produce bigger eggs from heavier quail.Similar percentage proportions of egg white and yolk among groups indicated that the capacity of egg white and yolk production was not compromised by MSTN mutation. Notably, the egg yolk weight of MSTN homozygous mutant eggs were significantly greater than those of WT eggs (Figure 2D). Indeed, positive correlation between body weight and yolk weight were also shown in the quail line selected for higher body weight [14]. This can be partially explained by more body fat of the quail line selected for higher body weight compared to the quail line selected for lower body weight [15], because moderate correlation between the body fat contents and egg yolk percentage were observed in white and brown egg layer chickens [10]. However, MSTN homozygous mutant quail had significantly lower leg and abdominal fat compared to those of WT quail (Table 1), indicating lipid mobilization for yolk formation was not compromised by lower fat accumulation from MSTN homozygous mutation. Although lower leg and abdominal fats of MSTN homozygous mutant quail did not negatively affect the egg production during the active laying period, significantly delayed onset of egg laying of MSTN homozygous mutant quail compared to WT quail might be associated with low fat accumulation, because negative correlation between body fat content and onset of egg laying were observed in broiler breeder chickens [8]. Furthermore, increased fat accumulation during the onset of the first egg laying period have been reported in quail [16]. It has been generally accepted that there are multiple threshold traits, including age, body weight, and body composition, needs to be reached for the onset of egg laying in quail and chickens [17,18]. According to the delayed egg laying of MSTN homozygous quail in our study, however, body composition changed by weight gain, mainly body fat accumulation, would be a more critical factor for the onset of egg laying than the body weight itself. In broiler breeder male and female chickens, Overweight negatively affects their reproductive traits [11], and thus intensive management is required to prevent overconsumption of feed to attain optimal body weight for high reproductive performance. Although improved feed efficiency and low leg and abdominal fats of MSTN homozygous mutant female indicated increased body weight is not caused by overconsumption of feed [4,6], the effect of MSTN mutation on egg fertility and hatchability was examined in this study. There was no significant difference on egg fertility and hatchability among groups, preventing potential problems of reproductive traits on breeder line with MSTN mutation. In mammals, double-muscled cows with MSTN mutation have increased risk of dystocia due to heavier weight of their offspring, but the quality of semen has not been affected by MSTN mutant in pigs. Therefore, fetal size of MSTN mutant animals can influence the reproductive traits of MSTN animals indirectly. However, birds lay eggs and the size is determined by the capacity of reproductive organs and, thus the indirect effect of egg size on the reproductive traits of MSTN birds can be negligible in avian species. In addition, when chicken lines are selected for abdominal fat content, the lean chicken male and female showed better reproductive traits with higher fertility and hatchability [19]. Likewise, low fat accumulation by MSTN homozygous mutation might contribute to the normal fertility and hatchability in MSTN homozygous mutant quail. 5. ConclusionsIn this study, we investigated the effect of MSTN mutation on egg production traits to provide information about potential usage of MSTN as a selective marker in the layer industry. Although MSTN homozygous mutant quail showed significantly heavier eggs with normal fertility and hatchability compared to WT quail, there was a negative effect on the onset of egg laying and egg production during the active laying period in MSTN homozygous mutant quail. While the positive effects of MSTN mutation on egg production cannot be concluded in this study due to various implications, consideration of MSTN as a potential selection marker for layers will remain a decision of the layer breeding companies depending on their prior selection criteria.

animals : an open access journal from mdpi

10.3390/ani12030327

PMC8833429

The gill tanscriptomes of greater amberjack (Seriola dumerili) reared under different salinity stress were analyzed. The regulatory networks of salinity-related pathways were explored through Kyoto Encyclopedia of Gene and Genome (KEGG) pathway enrichment and bioinformatics analyses. This will be of great value in understanding the molecular basis of salinity adaptation in greater amberjack.

Salinity significantly affects physiological and metabolic activities, breeding, development, survival, and growth of marine fish. The greater amberjack (Seriola dumerili) is a fast-growing species that has immensely contributed to global aquaculture diversification. However, the tolerance, adaptation, and molecular responses of greater amberjack to salinity are unclear. This study reared greater amberjack juveniles under different salinity stresses (40, 30, 20, and 10 ppt) for 30 days to assess their tolerance, adaptation, and molecular responses to salinity. RNA sequencing analysis of gill tissue was used to identify genes and biological processes involved in greater amberjack response to salinity stress at 40, 30, and 20 ppt. Eighteen differentially expressed genes (DEGs) (nine upregulated and nine downregulated) were identified in the 40 vs. 30 ppt group. Moreover, 417 DEGs (205 up-regulated and 212 down-regulated) were identified in the 20 vs. 30 ppt group. qPCR and transcriptomic analysis indicated that salinity stress affected the expression of genes involved in steroid biosynthesis (ebp, sqle, lss, dhcr7, dhcr24, and cyp51a1), lipid metabolism (msmo1, nsdhl, ogdh, and edar), ion transporters (slc25a48, slc37a4, slc44a4, and apq4), and immune response (wnt4 and tlr5). Furthermore, KEGG pathway enrichment analysis showed that the DEGs were enriched in steroid biosynthesis, lipids metabolism, cytokine–cytokine receptor interaction, tryptophan metabolism, and insulin signaling pathway. Therefore, this study provides insights into the molecular mechanisms of marine fish adaptation to salinity.

1. IntroductionMarine environmental factors, such as salinity, low O2 concentration, temperature, and pH value, influence the physiological and biological status of marine animals [1]. For instance, environmental stresses activate the sympathetic nervous system [2], the release of adrenaline and noradrenaline [3], and the hypothalamic–pituitary–interrenal axis in fish [4], thus causing the release of the steroid glucocorticoid hormones and other hormones for adaptation [5]. Although salinity enhances optimum fish growth, it can influence growth rate, immunity, antioxidant capacity, and lipid metabolism in fish [6,7]. For instance, salinity stress induces more active energy metabolism, including lipid metabolism and glycogen metabolism [8]. Salt stress environments trigger various metabolic changes in fish, thus enhancing fish adaptation to salinity [9,10]. Moreover, osmolality and water balance are energy-demanding processes maintained through osmoregulatory mechanisms, and they often alter survival, growth, and other physiological processes [11,12,13].Osmotic stress responses associated with salinity changes activate molecular and physiological adaptations, such as variations in cell proliferation and differentiation of osmoregulatory organs [12], modulation of the expression and activity of ATPases [14], secondary activation of ion transporters [15], and structural proteins [16]. The osmoregulatory organs of fish include the gill, kidney, and digestive tracts. However, the gill plays the most crucial role because it has a large surface area and direct contact with the external environment [17,18]. Blood and water in gills are separated by just a few micrometers, thus facilitating the exchange of gases and allowing gill tissue to be exposed to environmental variation and pollutants [19]. Moreover, gills can balance the ion concentration in blood and acid base in freshwater and seawater-adapted fish [20,21,22]. Salinity changes can cause many lesions in gills, such as vascular congestion [8], lamellar fusion, mucosal cell and gill filament epithelium hyperplasia, loss of the structural integrity of pillar cells, and an increased number of chloride cells [23,24]. These lesions can change the crucial functions of gills and alter their morphological structure [19].Some researchers have focused on the influence of salinity on the osmotic regulation of gills to investigate ionic and osmotic regulation in fish [9]. Furthermore, many studies have assessed the transcript expression profiles of teleosts, such as marbled eel (Anguilla marmorata) [25,26], spotted sea bass (Lateolabrax maculatus) [27], half-smooth tongue sole (Cynoglossus semilaevis) [28], Nile tilapia (Oreochromis niloticus) [15,29], silvery pomfret (Pampus argenteus) [30], Mozambique tilapia (Oreochromis mossambicus) [6], and Atlantic salmon (Salmo salar), after exposure to salinity changes [17]. As a result, some significantly differentially expressed genes and pathways related to salinity changes have been identified [31]. For instance, many classical ion transporters, including channels for amino acids [32], water [33], small solutes [34], calcium ions [35], sodium [36], chloride [37], and potassium [38], such as SLC (solute transport protein) families [39], AQP (aquaporin) families [40], NPY (neuropeptide Y receptor) families [41], and TRP (transient receptor potential) families are differentially expressed in the gills of teleosts under salinity changes [42]. Furthermore, researchers have also focused on pathways, including steroid biosynthesis, immune response, energy metabolism, apoptosis, cytokine–cytokine receptor interaction, and toll-like receptor signaling pathways [43].The greater amberjack Seriola dumerili is a large, fast-growing species in the aquaculture industry worldwide with high commercial value [27,28]. Many researchers have assessed the factors limiting reproduction [44], ectoparasites [45], and weaning diets of greater amberjack [46]. However, no research has identified the optimum and limiting environmental factors for greater amberjack juveniles in captivity. This study explored the influence of different salinities on the gill transcriptome and gene expression of greater amberjack juveniles to identify and assess the genes with potential roles in salinity adaptation using RNA sequencing. Therefore, this study can provide a basis for understanding the physiology of greater amberjack and practical guidance for its commercial aquaculture production.2. Materials and Methods2.1. Ethics StatementThe experiments were conducted following the guidelines and regulations of the Animal Research and Ethics Committee of Guangdong Ocean University (NIH Pub. No. 85–23, revised 1996) and China’s laws and regulations on biological research. This study did not include any endangered or protected species.2.2. Experimental Fish, Salinity Development, and Tissue CollectionA total of 80 greater amberjack juveniles (body length, 8.33 ± 0.45 cm and body weight, 6.38 ± 1.33 g) were used in this study. Before experiments, the greater amberjack juveniles were reared in tanks at 22 ± 1.0 °C in Donghai Island (Zhanjiang, Guangdong, China). They were randomly divided into four cylindrical 1000 L tanks (20 individuals per tank) at various salinities: 40, 30, 20, and 10 ppt (parts per thousand) groups. The salinity levels in the experiment were selected based on the previous salinity adaption study of greater amberjack larvals [47] and our unpublished data of juveniles. The fish in the control group were reared in natural seawater with a salinity of 30 ppt. The 40, 20, and 10 ppt groups were regulated using a commercial seawater salty crystal and aerated tap water. The fish were fed on commercial float bait twice a day at 9:00 and 19:00 for 30 days. Six fish were randomly selected from each group on 0, 15, and 30 days in our study, which was based on studies of other fishes, such as Asian seabass (Lates calcarifer, Bloch, 1790) [48], catfish (Lophiosilurus alexandri) [24], and cobia (Rachycentron canadum), under different salinity stress levels [31]. Unfortunately, the fish in the 10 ppt group were all dead within 10 days. The fish were then anesthetized using 100 mg/L tricaine methane sulfonate (MS 222; Sigma-Aldrich, St. Louis, MO, USA) and dissected. The gill tissues were immediately collected in centrifuge tubes containing 1 mL RNA stabilization reagent overnight, then stored at −80 °C for RNA extraction, sequencing, and gene expression analysis. The RNA of gill samples after 30 days were used for transcriptomic analysis, all gill samples after 15 and 30 days were used for qPCR verification.2.3. Total RNA Extraction, Library Construction, and Illumina SequencingTrizol reagent (Invitrogen, Carlsbad, CA, USA) was used to extract total RNA from the gill tissues, following the manufacturer’s instructions. As previously described, the cleavage of gill tissue samples, RNA extraction, RNA purity, degradation, and contamination examinations were performed [49]. An Agilent 2100 bioanalyzer (Agilent Technologies, Palo Alto, CA, USA) was used to detect RNA integrity. Total RNA with an RNA integrity number (RIN) score >7 was used for sequencing.TrueSeq RNA Sample Prep Kit (Illumina, San Diego, CA, USA) was used to obtain complementary DNA (cDNA) libraries from gill tissue, following the manufacturer’s instructions. Then, 3 µL of USER Enzyme (NEB, New England Biolabs, Palo Alto, CA, USA) was used with size-selected, adaptor-ligated cDNA at 37 °C for 15 min, followed by 5 min at 95 °C before PCR. PCR was performed using Phusion High-Fidelity DNA polymerase, universal PCR primers, and Index (X) Primer. The AMPure XP system was used to purify the PCR products. An Agilent Bioanalyzer 2100 system was used to assess the library quality. A cBot Cluster Generation System was used to cluster the index-coded samples via the TruSeq PE Cluster Kit v4-cBot-HS (Illumia, San Diego, CA, USA), following the manufacturer’s instructions. The library preparations were then sequenced on a HiSeq X-ten platform to generate the paired-end reads (150 bp).2.4. Transcriptome Assembly and Functional Gene AnnotationThe assembled Seriola dumerili genome (deposited in the DNA Data Bank of Japan (DDBJ) under accession numbers BDQW01000001-BDQW01034655 (Biosample ID: SAMD00083043_sdu_WGS.acclist.zip)) was used as a reference database for mapping reads [50]. The Illumina high-throughput sequencing platform was used to sequence the cDNA library, generating raw reads/data. The raw reads/data were filtered, then the adapter sequence and ploy-N (unable to determine base information) and low-quality (reads with <50% bases of quality value) reads were removed to obtain high-quality clean reads/data. Raw data and clean data were saved in FASTQ format. The Q20, Q30, GC-content, and sequence duplication levels of the clean data were measured. All the subsequent analyses were based on high-quality clean data.BLASTx (version 2.2.26) (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 29 December 2021) (E-value threshold, 1 × e−5) was used to annotate the gene function via the NR (https://www.ncbi.nlm.nih.gov/refseq/, accessed on 29 December 2021), Swiss-Prot (protein) (https://www.uniprot.org/;%20version%2020140317, accessed on 29 December 2021), KEGG (https://www.genome.jp/kegg/, accessed on 29 December 2021), COG (https://www.ncbi.nlm.nih.gov/research/cog/, accessed on 29 December 2021), and KOG (https://www.ncbi.nlm.nih.gov/research/cog-project/, accessed on 29 December 2021) databases, using the default setting. The clean library sequencing data were submitted to the National Center for Biotechnology Information (NCBI) Sequence Read Archive (SRA) (Bioproject number: PRJNA783747 (SRR17041711-SRR17041719)).2.5. Analysis of Differential Expressed Genes (DEGs) AnalysisThe greater amberjack gene expression levels were detected using the fragments per kilobase per million (FPKM) method. The DEGs in 20 vs. 30 ppt and 30 vs. 40 ppt groups were identified using the DESeq2 R package (version 1.16.1). DESeq2 was used to determine differential expression in digital gene expression data based on the negative binomial distribution model. Genes with a fold change ≥2 and a false discovery rate (FDR) <0.05 were considered DEGs. A KEGG pathway analysis was conducted, and DEGs with p < 0.05 were considered statistically significant [31]. In Donghai island, 30 ppt was the natural salinity of seawater. The DEGs in both 20 vs. 30 ppt and 30 vs. 40 ppt were identified and screened in order to clarify the effects of hypo-salinity and hyper-salinity seawater environments on gills of greater amberjack compared with the natural salinity.2.6. Quantitative Real-Time PCR (QPCR) ValidationThe expression patterns of DEGs in the RNA sequencing analysis were validated using qPCR. Sample collection, RNA extraction, and reverse transcription were performed as previously described [51]. RNA samples were obtained from the 20, 30, and 40 ppt groups. Each group had three replicates. The DEGs primers were designed using Primer5 software based on the assembled transcripts. A light CyclerTM96 (Roche, Indianapolis, IN, USA) was used for qPCR analysis, following the protocol of SYBR Green Real-Time PCR Master Mix (Takara, Tokyo, Japan). β-actin was used as a reference gene to normalize the expression levels [52,53,54]. The relative abundance of DEG mRNA transcripts was evaluated using the 2−△△Ct method. The primer sequences for qPCR are shown in Table S1.2.7. Statistical AnalysisThe relative mRNA expression levels and FPKM values are expressed as mean ± standard error (SE). The one-way ANOVA with Tukey’s post hoc test was used to evaluate the significant differences between 20 vs. 30 ppt and 30 vs. 40 ppt groups. The significance level was set at α = 0.05. The Statistical Package for the Social Sciences (SPSS) 16.0 (SPSS, Chicago, IL, USA) was used for all statistical analyses.3. Results3.1. Illumina SequencingThe gill transcriptome can provide a valuable RNA resource for future analysis of greater amberjack adaptability to salinity and artificial culture. The HiSeq X Ten platform was used for RNA sequencing of gill samples at 20, 30, and 40 ppt. A total of 94.79, 93.66, and 78.83 million clean reads were obtained at 20 (G20), 30 (G30), and 40 ppt (G40), respectively, after quality control. The Q30 values and GC content of the clean reads were more than 95% and 49%, respectively. The high-quality reads were used for further analysis. The reads’ Q20 and Q30 values, GC content, and transcript numbers for each cDNA library are shown in Table 1.3.2. Annotation and Functional Analysis of the Gill TranscriptomeIn total, 23,972 genes were annotated. A total of 23,950 (99.91%); 22,838 (95.27%); 21,503 (89.70%); 16,505 (68.85%); 15,991 (66.71%); 14,525 (60.59%); 14,970 (62.45%); and 7382 (30.79%) were annotated in the NR, eggNOG, Pfam, Swiss-Prot, KOG, GO, KEGG, and COG databases, respectively (Table 2). The annotated genes provided the basis for further analysis of the specific molecular processes in greater amberjack.The best BLAST results of reads were enriched for closely related fish species, including Seriola lalandi (6.24%), Lates calcarifer (2.09%), Larimichthys crocea (1.49%), Stegastes partitus (0.63%), Oreochromis niloticus (0.55%), and other species (4.98%) (Figure 1).3.3. Identification and Analysis of Differentially Expressed Genes (DEGs)A total of 417 (205 up-regulated and 212 down-regulated) and 18 (nine up-regulated and nine down-regulated) DEGs were identified in the G30 vs. G20 and G30 vs. G40 groups, respectively, using DESeq2 software, FDR-adjusted p-value < 0.05 and |Log2(fold change)| ≥ 1 (Figure 2). Heat maps of the clustered DEGs under hypo- and hyper-salinity stresses are shown in Figure 3. The top 20 DEGs in the G30 vs. G20 group and the top 18 DEGs in the G30 vs. G40 group are shown in Table 3.3.4. DEG Annotation and Pathway AnalysisSome enriched KEGG pathways related to Seriola dumerili metabolism and molecular signaling pathway were identified, including steroid biosynthesis, cytokine–cytokine receptor interaction, porphyrin and chlorophyll metabolism, cytosolic DNA-sensing pathway, the intestinal immune network for IgA production, and tryptophan metabolism (Figure 4 and Figure 5). The top 10 enriched KEGG pathways of the DEGs under hypo- and hyper-salinity stresses are shown in Table 4.3.5. Validation of RNA Sequencing Data by QPCRConsidering the top 20 most important salinity stress responses of DEGs identified in the transcriptomic analysis, which had been reported in the previous studies, nine genes were selected for RNA sequencing data validation using qPCR (Figure 6). The mRNA expression of ebp, msmo1, nsdhl, sqle, lss, and ogdh was significantly increased in the G20 group, while that of ebp, msmo1, and sqle was significantly decreased, compared with the G30 group. However, the mRNA expression of nsdhl, lss, and ogdh showed no difference between the G40 and G30 groups. The results were consistent with the RNA sequencing analysis results conducted using qPCR. Moreover, the mRNA expression of edar, wnt4, slc25a48 were decreased in the G20 group. However, the mRNA expression of edar and slc25a48 showed no difference between the G40 and G30 groups. Interestingly, the expression of these genes showed no differences at 15 days (Figure S1).4. Discussion4.1. Transcriptomic Analysis of Differentially Expressed GenesSalinity alterations can lead to various physiological reactions to maintain homeostasis, including osmotic regulation, ion transports, and respiratory metabolism [55]. For instance, low (12 ppt) and high salinity water (32 ppt) can significantly induce some specific pathways in gills of silvery pomfret (Pampus argenteus), including calcium transport, neuroactive ligand–receptor interaction, NOD-like receptor signaling, Toll-like receptor signaling, and cytokine–cytokine receptor interaction pathway, indicating that salinity stress affects the immune system and osmotic pressure-regulated pathways [30]. Moreover, low salinity stress (6 ppt) affects ion transport, immune response, energy metabolism, and protein synthesis in marbled flounder (Pseudopleuronectes yokohamae) [9]. A total of 417 DEGs (205 up-regulated and 212 down-regulated) in the 30 ppt vs. 20 ppt group and 18 DEGs (nine up-regulated and nine down-regulated) in the 30 ppt vs. 40 ppt group were identified. KEGG analysis showed that the DEGs were mainly enriched in the cytokine–cytokine receptor interaction, apoptosis, steroid biosynthesis, and mTOR signaling pathways, similar to cobia (Rachycentron canadum) [31]. As discussed below, the DEGs were involved in several potential complex molecular biological processes in the gill.4.2. DEGs Involved in Steroid Biosynthesis and Lipid MetabolismSteroid hormones, such as epinephrine and cortisol, can influence the metabolic capacity of the gill [56]. Epinephrine induces glycogenolysis after exposure to stress, thus increasing the plasma glucose level, which provides energy for the target tissue, including gill, to transfer ions like Na+ [57]. In addition, cortisol can stimulate active Ca2+ uptake under asymmetrical conditions, thus regulating the tight junction morphology between pavement cells of euryhaline fish [56,58]. Herein, KEGG pathway analysis showed that some DEGs (ebp, lss, sqle, nsdhl, msmo1, sc5d, dhcr7, and dhcr24) were involved in the steroid biosynthesis pathways under the long-term hyper-salinity (40 ppt) and hypo-salinity (20 ppt) stresses. Moreover, salinity stress (from 20 to 40 ppt) decreased the mRNA expression levels of sqle, msmo1, and ebp. Ebp (Emopamil binding protein), also known as EBP cholestenol delta-isomerase, is essential in the sterol biosynthesis pathway [59]. Sterols are essential cell membrane components and transporters in many biofilms [60]. SQLE (squalene epoxidase) is rate limiting and the first oxygenation enzyme in cholesterol synthesis [61]. Lanosterol, especially cholesterol, is the upstream precursor of sterol biosynthesis in fungal steroids and animals [62]. Herein, hypo-salinity up-regulated sqle, ebp, and lss in the gills, indicating the stimulation of cholesterol synthesis in the gills. Studies have shown that lss and dhcr24 are up-regulated in O. niloticus and Rachycentron canadum under hypo-salinity [31,63], consistent with this study.Msmo1 is a key cholesterol biosynthetic enzyme. Nsdhl regulates adipogenesis via a synergized expression pattern with Msmo1 [64,65,66,67]. Herein, hypo-salinity up-regulated both nsdhl and msmo1, indicating that hypo-salinity can affect adipogenesis in greater amberjack. Moreover, hypo-salinity down-regulated edar and tnfsf12 mRNA in greater amberjack. Studies have shown that EDAr (ectodysplasin A receptor), belonging to the tumor necrosis factor receptor (TNFr) superfamily, can regulate cell activities, such as differentiation, proliferation, maturation, and lipid metabolism, by binding to the ectodysplasin A1 (EDA1) [68,69,70,71]. A previous study showed that tnfsf12 (TNF superfamily member 12) can inhibit lipid deposition in a dose-dependent manner without any cytotoxic effects. However, an agonistic antibody of the tnfsf12 receptor can alleviate the repression [72]. Herein, hypo-salinity decreased the mRNA expression of tnfsf12 in gill, indicating that lipid deposition is essential under low salinity stress. Previous reports have suggested that lipids are the energy source for euryhaline fish under osmotic stress [49,73]. Therefore, more lipids may be produced in greater amberjack under salinity changes by regulating the steroid biosynthesis related to lipid metabolism, adipogenesis, and lipid deposition, as revealed in other fish species [18]. The Ogdh enzyme is a key entry point for carbon into the Krebs cycle. Moreover, it affects all redox signals in mitochondria and cells [74,75]. Herein, hypo-salinity significantly increased the mRNA expression of ogdh mR, implying that adequate energy is needed in gills under salinity changes.4.3. DEGs Involved in Ion TransportSolute transport protein (solute carrier SLC) is the largest class of intracellular transport proteins, with over 300 members, mostly located in cell membranes. They mainly facilitate the transport of various substrates, including amino acids, nucleotides, glucose, and inorganic ions, across biological membranes [39]. Some of these proteins (slc5a6a, slc4a1a, slc4a4a) had different expression patterns between the hypo-salinity and control groups. The SLC4 gene family mediates HCO3− extrusion and Cl− uptake across cellular plasma membranes, thus regulating the cell volume and intracellular pH and stabilizing resting membrane potential through the regulation of cytoplasmic Cl− [76]. Moreover, slc4a1a and slc4a1b are positively correlated with NKA (Na+/K+-ATPase) [77]. Herein, slc4a1a was down-regulated in the hypo-salinity group, while it was up-regulated in the hyper-salinity group, indicating the ion exchange in the gills.Cystic fibrosis transmembrane conductance regulator (CFTR), an ABC transporter, acts as a channel across the cell membrane for transporting Cl− into and out of cells. Hyper-salinity increases cftr expression in striped bass (Morone saxatilis) [78]. Moreover, cftr mRNA levels are significantly increased in gills of killifish (Fundulus heteroclitus) and Atlantic salmon (Salmo salar) under seawater conditions [79,80]. Similarly, this study showed that cftr was up-regulated in the hyper-salinity group, indicating the Cl− transport changes.Aquaporins (AQPs) are a family of integral membrane proteins that facilitate water transport across biological membranes along an osmotic gradient [81]. Thirteen AQP isoforms (AQP0-AQP12) have been identified in humans and rodents [40]. However, AQP4 has the most potential for high water permeability [82]. Herein, the expression of aqp4 was significantly decreased in the hypo-salinity group compared with the control group, indicating the possible water permeability changes.4.4. DEGs Involved in Immune ResponsePrevious studies have shown that the changes in fish immune status depend on the intensity and duration of the environmental stresses [83]. Herein, several DEGs (wnt4, slc25a48, slc6a8, tlr5, etc.) related to immune responses were found in the gill transcriptome of greater amberjack. Studies have reported that wnt genes enhance defense against pathogenic virus infection in the innate immune of Litopenaeus vannamei [84]. Moreover, Wnt4 protein can stimulate white blood cells and thymopoiesis in mice [85]. Herein, wnt4 was significantly down-regulated in the hypo-salinity group, indicating that low salinity can affect the immune system of the greater amberjack.Toll-like receptors (TLRs) are key pathogen pattern recognition receptors that control the host immune responses against pathogens by recognizing molecular patterns specific to microorganisms [86]. Previous studies have reported that tlr5 plays a crucial role in the immune responses of turbot (Scophthalmus maximus L.) to the infections of various pathogens [87]. For instance, TLR5 stimulates the expression of proinflammatory, antibacterial, and stress-related genes by binding to bacterial flagellin, thus enhancing host defense against bacterial pathogens [88]. Moreover, flagellin increases Tlr5 activity and the release of its downstream factor, IL−8 in mice [88,89]. Herein, the expression level of tlr5 was decreased in the hypo-salinity group, indicating that low salinity inhibits tlr5 activity on key adaptive functions, thus lowering efficient immune responses.5. ConclusionsThis work used transcriptome analysis to investigate the molecular changes in gills of greater amberjack (Seriola dumerili) under three different salinity concentrations (20, 30, and 40 ppt). The results provided large transcriptome data, abundant DEGs, and signaling pathways related to salinity adaptation. The signaling pathways analysis indicated that a complex molecular regulatory network is involved in metabolism, including steroid synthesis, lipid metabolism, tryptophan metabolism, ion transporters, and immune response for adaptation to salinity stress. Therefore, this study can provide insights into the molecular mechanisms of greater amberjack adaptation to salinity.

animals : an open access journal from mdpi

10.3390/ani11072131

PMC8300340

The maintenance of cows in good health and physical condition is an important component of dairy cattle management. One of the major metabolic disorders in dairy cows is subclinical ketosis. Due to financial and organizational reasons it is often impossible to test all cows in a herd for ketosis using standard blood examination method. Using milk data from test-day records, obtained without additional costs for breeders, we found diagnostic models identifying cows-at-risk of subclinical ketosis. In addition, to select the best models, we present a general scoring approach for various machine learning models. With our models, breeders can identify dairy cows-at-risk of subclinical ketosis and implement appropriate management strategies and prevent losses in milk production.

The diagnosis of subclinical ketosis in dairy cows based on blood ketone bodies is a challenging and costly procedure. Scientists are searching for tools based on results of milk performance assessment that would allow monitoring the risk of subclinical ketosis. The objective of the study was (1) to design a scoring system that would allow choosing the best machine learning models for the identification of cows-at-risk of subclinical ketosis, (2) to select the best performing models, and (3) to validate them using a testing dataset containing unseen data. The scoring system was developed using two machine learning modeling pipelines, one for regression and one for classification. As part of the system, different feature selections, outlier detection, data scaling and oversampling methods were used. Various linear and non-linear models were fit using training datasets and evaluated on holdout, testing the datasets. For the assessment of suitability of individual models for predicting subclinical ketosis, three β-hydroxybutyrate concentration in blood (bBHB) thresholds were defined: 1.0, 1.2 and 1.4 mmol/L. Considering the thresholds of 1.2 and 1.4, the logistic regression model was found to be the best fitted model, which included independent variables such as fat-to-protein ratio, acetone and β-hydroxybutyrate concentrations in milk, lactose percentage, lactation number and days in milk. In the cross-validation, this model showed an average sensitivity of 0.74 or 0.75 and specificity of 0.76 or 0.78, at the pre-defined bBHB threshold 1.2 or 1.4 mmol/L, respectively. The values of these metrics were also similar in the external validation on the testing dataset (0.72 or 0.74 for sensitivity and 0.80 or 0.81 for specificity). For the bBHB threshold at 1.0 mmol/L, the best classification model was the model based on the SVC (Support Vector Classification) machine learning method, for which the sensitivity in the cross-validation was 0.74 and the specificity was 0.73. These metrics had lower values for the testing dataset (0.57 and 0.72 respectively). Regression models were characterized by poor fitness to data (R2 < 0.4). The study results suggest that the prediction of subclinical ketosis based on data from test-day records using classification methods and machine learning algorithms can be a useful tool for monitoring the incidence of this metabolic disorder in dairy cattle herds.

1. IntroductionSubclinical ketosis is one of the major metabolic disorders in dairy cows [1,2,3]. Subclinical ketosis increases the risk of clinical ketosis [4] as well as other disorders, e.g., displaced abomasum, metritis and lameness [5,6,7], which can lead to an increased culling rate [4] and higher costs at herd level [8,9]. It has been determined that subclinical ketosis is also associated with reduced milk production [6] as well as with reduced reproductive performance of cows [10,11]. Subclinical ketosis is mostly observed during early lactation and can be diagnosed based on an elevated ketone bodies in body fluids (blood, milk, urine) in the absence of clinical signs [12]. The β-hydroxybutyrate concentration in blood (bBHB) is an indicator used for diagnosing subclinical ketosis in dairy cows. The review paper by Benedet et al. [13] indicates various bBHB thresholds used in the literature for distinguishing between healthy cows and those with subclinical ketosis. In the majority of publications, the threshold is defined as 1.2 mmol/L [7,14,15,16,17,18], and only rarely it is defined as 1.0 mmol/L [19,20] or 1.4 mmol/L [6,21]. The detection of ketone bodies in blood is not a standard procedure used in the management of dairy cattle herds. Due to practical (financial and organizational) limitations, it is impossible to test all cows in a herd at regular intervals. A search for indicators in milk samples during the assessment of milk performance, which would allow identifying cows-at-risk of subclinical ketosis during early lactation is ongoing. The strong correlation between the ketone bodies in blood and milk [22] can be an indication for the use of acetone (ACE) and β-hydroxybutyrate concentrations in milk (mBHB) for diagnosing subclinical ketosis. van Knegsel et al. [23] found that the inclusion of ACE and mBHB helps to detect subclinical ketosis with greater accuracy as compared to the inclusion of fat-to-protein ratio in milk.The incidence of ketosis varies greatly between individual farms. Clinical ketosis is observed in about 4–10% of cows per herd and subclinical ketosis—in about 10–50% of cows [4,15,24]. In Poland, about 10% of cows per herd are at risk of ketosis on average. This percentage is even higher, up to 30%, during the early period of the first lactation [25]. According to Oetzel [1], the identification of subclinical ketosis in 10% of cows in a herd should be considered an alarming level.The objective of the study was (1) to design a scoring system that would allow choosing the best machine learning models for the identification of cows-at-risk of subclinical ketosis, (2) to evaluate various machine learning methods and to choose the best performing models, and (3) to validate the best performing models using a testing dataset containing unseen data.The advantages of machine learning methods include the possibility of generating models without any previous knowledge of relationships between variables [26], the smaller number of assumptions concerning data (e.g., normal distribution is often not required), as compared to linear methods [27].Machine learning has been used in the field of dairy science for early detection of subclinical mastitis [28,29,30,31]. Much attention has been paid to the development of machine learning expert systems for detection of subclinical mastitis from milking parameters. Such parameters as milk yield, fat, protein and lactose concentration, milking time and peak flow are easily accessible due to widely used in dairy farms automatic milking systems, which provide breeders a large amount of information about each cow. Using machine learning techniques and information from non-invasive sensors allow prediction of time-to-calving in beef and dairy cows [32], modeling of milk yield of dairy cows under heat stress condition [33], and identification of heat-stressed cows [34].In the traditional approach, models are often built as a result of a good understanding of the application domain which helps to create and select variables that can be included in models. Model validation is based mainly on the goodness-of-fit evaluation and hypothesis testing. In machine learning, the effort is shifted from a deep understanding of the application domain towards computationally constructed and tested models [35].2. Materials and Methods2.1. Initial DatasetThe original dataset consisted of 882 test-day (TD) records for Polish Holstein–Friesian cows. Some records were excluded from further analysis if the lactation number was unknown (n = 5), the sample collection day was incorrect (<6 or >60 days in milk) (n = 37) and the test-day results were missing (n = 7). Following the removal, the initial dataset consisted of 833 unique TD records from the first eight lactations, grouped into four categories of lactation (1, 2, 3, 4–8). The cows calved in 37 herds in 2013 and 2014. The data were provided by the Polish Federation of Cattle Breeders and Dairy Farmers. The records included nine milk traits: TD milk yield, fat, protein and lactose percentages, fat-to-protein ratio (FPR), milk urea concentration, somatic cell count (SCC), ACE and mBHB. The daily FPR was calculated as the ratio of TD fat percentage to protein percentage. To normalize the distribution, the SCC in milk was common log-transformed to the somatic cell score (SCS). All milk variables were recorded as continuous traits and were not assigned to categories. The number of lactation was used as a categorical variable. Table 1 shows the descriptive statistics of the initial dataset.There was only one sample per cow in the dataset. Milk samples were analyzed using a MilkoScan FT6000 analyzer (Foss, Hillerod, Denmark). The acetone and β-hydroxybutyrate concentrations in milk were determined by Fourier-transform infrared method (FTIR) using a CombiFoss analyzer (Foss, Hillerod, Denmark). Sampling of individual cows comprised single milk and blood samples collected on the same test-day. The β-hydroxybutyrate concentrations in blood were measured using an OptiumXido glucometer (Abbott, Winey, UK). The data were collected between September 2013 and June 2014. For further analysis, three bBHB thresholds were used as the diagnostic reference for subclinical ketosis: 1.0, 1.2 and 1.4 mmol/L. Cows with circulating bBHB lower than the pre-defined threshold were considered to be healthy.2.2. ApproachThe scoring system for the identification of subclinical ketosis was developed using two machine learning (ML) modeling pipelines, one for regression and one for classification.The analyses were performed with Python version 3.8, using pandas (1.1.2), numpy (1.19.2), scipy (1.5.2), imbalanced-learn (0.7.0), scikit-learn (0.23.2), lightgbm (3.0.0), xgboost (1.2.0) and catboost (0.24.2) libraries. Figure 1 presents an overview of the regression and classification modeling pipelines.2.3. Data Pre-Processing for Machine LearningFor the best performance of ML algorithms, 12 versions of the initial dataset were prepared using different feature selection and outlier detection methods.2.3.1. Feature SelectionTo select features for modeling, two feature selection methods were used: one based on Pearson’s and Spearman’s correlation coefficients, and another one based on ML recursive feature elimination method (RFE).Table 2 shows Pearson’s correlation coefficients for continuous features in the initial dataset. The correlation coefficients between independent features ranged between 0.41 and 0.86. To eliminate multicollinearity between independent features, the ones with correlation coefficient above 0.80 were examined. Fat percentage was eliminated from the modeling dataset because it was highly correlated with FPR (0.86). The acetone and β-hydroxybutyrate concentrations in milk were also highly correlated (0.76), however, below the pre-defined threshold of 0.80. Finally, features correlated with the target variable (bBHB), having an absolute value of correlation coefficient equal to or greater than 0.20, were selected for further processing: ACE (0.63), mBHB (0.62), FPR (0.44), lactose percentage (0.24) and days in milk (DIM) (0.21). In addition, the only categorical feature in the initial dataset (parity) was selected as having Spearman’s correlation coefficient with the target variable of 0.20.Based on scatter plots of all combinations of features, no non-linear relationships were observed, neither between independent features nor between independent features and the target feature.The recursive feature elimination machine learning method with scikit-learn DecisionTreeRegressor estimator was used for selecting the three best-suited groups of features. The goal of the recursive feature elimination is to select features by recursively considering smaller and smaller sets of features using an external estimator that assigns weights to features. First, the estimator is trained using the initial set of features to determine the importance of each feature. Then, the least important features are pruned one by one out of the current set of features. That procedure is recursively repeated until the desired number of features is achieved.The best three groups of features selected using the RFE selection method were termed as RFE1, RFE2, and RFE3. The RFE1 group included ACE only. The RFE2 group contained milk yield, fat percentage, protein percentage, FPR and ACE. The RFE3 group contained protein percentage and ACE.2.3.2. OutliersTwo approaches: analytical and numerical were used for the identification of outliers. In the analytical approach, for features with non-Gaussian distribution (ACE, mBHB and bBHB), observations with values higher than 1.5 of interquartile range (IQR) were removed. For features with Gaussian distribution (milk yield, FPR, fat, protein and lactose percentages), observations with values higher than three standard deviations (SD) from the mean were removed. In the numerical approach, outliers were detected using unsupervised one-class classification (OOC) approach based on the scikit-learn local outlier factor (LOF) machine learning method. The unsupervised anomaly detection LOF algorithm is a method which computes the local density deviation of a given data point with respect to its neighbors. It considers as outliers the samples that have a substantially lower density than their neighbors. Table 3 summarizes the differences between the 12 datasets generated using different feature selection and outlier detection methods.2.4. Modeling Pipelines—Description and Validation of ModelsIn order to predict subclinical ketosis based on a numerical (continuous) target feature (bBHB), two ML modeling pipelines were designed and used to score regression and classification models. All the 12 datasets prepared during the data pre-processing stage were used as input for both pipelines.Each input dataset was split into training (with 70% of observations) and testing (30% of observations) subsets using the scikit-learn train_test_split method with the same random_state parameter for reproducibility and comparability. Using the same random state guarantees the same split into training and testing datasets at all times. In addition, the stratified sampling method was used in the classification pipeline. The use of such sampling leads to the generation of training and testing subsets that have the same proportions of class labels as in the initial dataset. The same random state and stratified sampling defined while splitting data into the training and testing subsets made it possible to compare different ML algorithms based on the same input data.Each training dataset was scaled using four scikit-learn feature scaling methods: StandardScaler, RobustScaler, Normalizer and MinMaxScaler. Non-scaled version was also used for comparison. Some algorithms perform better if features are in the same scale or are scaled using a different feature scaling method.In both pipelines, dummy estimators were used to establish the performance baseline (point of reference) for all other modeling techniques. If a model achieves performance at or below the baseline, the technique should be improved or abandoned.2.4.1. Regression PipelineIn the regression pipeline, 14 ML algorithms were used. In the scikit-learn package: DummyRegressor (always returns mean), LinearRegression, ElasticNet, SupportVectorRegressor (SVR) with linear and squared exponential (rbf) kernels, DecisionTreeRegressor, AdaBoostRegressor, BaggingRegressor, RandomForestRegressor, ExtraTreesRegressor, and BayesianRidge; in the xgboost package: XGBRegresor; and in the lightgbm package: LGBMRegressor. All the methods (except SVR) were used with their default hyperparameters. The mathematical details and the conceptual underpinnings of the methods used in the pipelines can be found in Hastie et al. [36].For each feature scaling method, performance of the fitted models was evaluated using the training dataset by repeating ten times the 10-folds cross validation (CV) with the mean coefficient of determination (R2), mean absolute error (MAE), root mean square error (RMSE) and their standard deviations as model performance metrics. Next, the best performing models were fitted to the entire training dataset for making predictions at a later stage (using unseen data represented by the testing dataset).The testing dataset was used for 12 best regression models (one per dataset) to compare their performance with classification models. bBHB predicted by regression models were split into binary classes based on three cut-off points (1.0, 1.2 and 1.4) and their classification power was evaluated based on sensitivity, specificity, balanced accuracy, Matthews correlation coefficient and F2 score classification metrics. bBHB values lower than the cut-off point were classified into ketosis negative class (class label = 0). bBHB values equal to or greater than the cut-off point were classified into ketosis positive class (class label = 1).2.4.2. Classification PipelineThe first step was to create three binary target features based on bBHB original continuous values according to three cut-off points at 1.0, 1.2 and 1.4. bBHB values lower than the cut-off point were classified into ketosis negative class (class label = 0). bBHB values equal to or greater than the cut-off point were classified into ketosis positive class (class label = 1).In the classification pipeline, 12 machine learning algorithms were used. In the scikit-learn package: DummyClassifier, LogisticRegression, SGDClassifier, DecisionTreeClassifier, KNeighborsClassifier, AdaBoostClassifier, BaggingClassifier, RandomForestClassifier, ExtraTreesClassifier, SupportVectorClassification (SVC), and GaussianNB; in the catboost package: CatboostClassifier. For all the methods, their default hyperparameters were used.Table 4 shows the number of observations in ketosis positive and ketosis negative classes, including the prevalence of subclinical ketosis for each cut-off point of bBHB (1.0, 1.2 and 1.4).As observed, positive and negative classes were imbalanced in all cases (Table 4). To balance the target binary classes, 5 oversampling methods (from the scikit-learn package) were used: SMOTE, BorderlineSMOTE, RandomOverSampler, ADASYN and SVMSMOTE. The oversampling was performed during the cross-validation iterations after the training dataset was split into folds to eliminate potential data leakage.For each of the cut-off points, feature scaling and oversampling algorithm classification models were evaluated on the training datasets using the scikit-learn RepeatedStratifiedKFold cross-validation method (repeated ten times with 10-folds). The mean cross-validation sensitivity, specificity, balanced accuracy, Matthews correlation coefficient and F2 score, and their standard deviations were used as model performance metrics. Next, the 12 best performing models (one per dataset) were fitted to the entire training dataset for making predictions at a later stage (using unseen data represented by the testing dataset).Using the testing datasets, for each of the class cut-off points, the 12 best classification models were evaluated based on sensitivity, specificity, balanced accuracy, Matthews correlation coefficient and F2 score classification metrics.The application of different outlier detection methods (during data pre-processing) resulted in a varying number of observations and features in each of the 12 input datasets. The ML algorithms were scored separately for each of the twelve input datasets to ensure that the scoring was performed on the same training and testing datasets for each algorithm in the regression and classification pipeline. As a result, 72 best performing models were selected (one per each input dataset, pipeline and cut-off point).2.4.3. Evaluation MetricsTo compare and assess the final performance of each regression and classification model on the testing dataset, the continuous values of the target feature (bBHB), as predicted by a regression model, were translated into classes using three cut-off points (1.0, 1.2 and 1.4) and the same logic of positive class assignment as in case of classification. Next, the same set of classification metrics was used consistently across all models, regardless of the initially used type of machine learning method.Five metrics were used for the evaluation of classification models: sensitivity (recall, true positive rate, TPR), specificity (true negative rate, TNR), balanced accuracy (bACC), Matthews correlation coefficient (MCC) and F2 score.Sensitivity indicated the proportion of cows with subclinical ketosis that were correctly predicted as cows with subclinical ketosis and specificity indicated the proportion of healthy cows that were correctly predicted as healthy.Instead of accuracy which indicates the percentage of correctly predicted cows in the dataset but can be misleading in the case of an imbalanced dataset, balanced accuracy [37] was calculated using the following formula:bACC=sensitivity+specificity2The value of this metric can range from 0 to 1, where 1 means perfect performance of a model and 0 means random scoring.Additionally, the Matthews correlation coefficient [38] was calculated. This metric has values in the range of −1 to 1, where −1 represents the total disagreement between predicted and actual value, and 1 indicates that the prediction generated by the model entirely agrees with the actual value. The MCC was calculated according to the following formula:MCC=TP·TN−FP·FN(TP+FP)(TP+FN)(TN+FP)(TN+FN) where TP, TN, FP and FN are true positive, true negative, false positive, and false negative, respectively.The Fβ score can be interpreted as a weighted harmonic mean of precision and recall reaching the optimal value at 1 (indicating perfect precision and recall) and the least optimal value at 0. Precision is the percent of correct positive predictions. Recall (also referred to as sensitivity) is the percentage of correctly classified positive values. Fβ=(1+β2)·precision·recallβ2·precision+recallThe two values commonly used for β are 2, which means that greater weight is attributed to recall than to precision, and 0.5, which means that greater weight is attributed to precision than to recall. For ketosis predictions, it is more important to identify as many cows with subclinical ketosis as possible, and therefore the F2 score was chosen as a metric. F2=5·precision·recall4·precision+recall3. Results3.1. Number of ModelsFor each of the 12 input datasets (Table 3), characterized by the different compositions of features and the use of different methods for their selection and the elimination of outliers, five datasets were generated, for which different methods were used for the scaling of independent variables (no scaling was performed for one of these five datasets). As a result, a total of 60 sets were generated.The classification approach included the generation of three sets for each of these 60 sets, where the continuous values of the dependent variable (bBHB) were assigned to the following classes: 0 and 1, based on one of the bBHB thresholds (1.0, 1.2, 1.4). Next, using five oversampling methods, for each of the resulting 180 datasets and 12 machine learning methods, a total of 10800 models were trained.In the regression-based approach for each of 60 datasets characterized by the different composition of features, the use of different methods for the elimination of outliers and scaling of variables, a total of 840 models were trained using 14 machine learning algorithms.3.2. Performance of Classification ModelsIn the classification-based approach, out of the models trained for each threshold (1.0, 1.2, 1.4) for the dependent variable (bBHB), a total of 12 models were chosen (one per each dataset shown in Table 3), for which the greatest mean sensitivity, specificity, bACC and MCC were obtained in the cross-validation, including the lowest respective standard deviations. The values of the selected metrics for the best classification models are shown in (Table 5, Table 6 and Table 7).For the threshold of 1.0 mmol/L for the dependent variable (bBHB), the mean sensitivity in the cross-validation ranged between 0.63 and 0.90 (with standard deviation in the range of 0.09 and 0.20) and the specificity was in the range of 0.14 and 0.73 (with standard deviation in the range of 0.05 and 0.19) (Table 5). The model with both high average sensitivity (0.74) and specificity (0.73) was a model based on the SVC machine learning algorithm. Variables for the model were selected based on the calculated coefficients of correlation between variables and outliers were eliminated using the LOF machine learning method (Table 5, dataset 3). For scaling of variables, RobustScaler was used. Oversampling was performed using the ADASYN method. For this model, bACC obtained in the cross-validation was equal to 0.74. This model was characterized by the highest MCC (0.40) among the models selected for each of the 12 datasets (Table 5). Taking also into account the average F2 score (0.63) as determined during the cross-validation, this model should be considered to be superior as compared to the other. For the testing dataset containing unseen data, the sensitivity, specificity and bACC were lower than those obtained in the cross-validation at 0.57, 0.72 and 0.65, respectively.If the bBHB threshold was defined as 1.2 mmol/L, then the best classification model was the logistic regression (Table 6). Oversampling was performed using the ADASYN method. As previously, variables to be included in the model were selected based on correlation coefficients and they were scaled using the MinMaxScaler (Table 6, datasets 1 and 3). In principle, it was not important whether outliers were eliminated using a machine learning method or were not eliminated because mean values of metrics selected in the cross-validation were similar in both cases: 0.74 and 0.73 for sensitivity, 0.76 and 0.77 for specificity, and bACC was 0.75 for both models. MCC was 0.38 and 0.39 respectively for the model used for a dataset from which outliers were eliminated and for a model used for a dataset where outliers were not eliminated. For the bBHB threshold defined as 1.2 mmol/L, the F2 score was highest for the two models indicated by us and it was 0.6 in both cases. For both models, the values of the selected metrics for the testing dataset were close to the mean values of these metrics as obtained in the cross-validation. Taking into account the bBHB threshold of 1.2 mmol/L out of the best performing models (with the greatest sensitivity), each of the 12 datasets also included models for which the average sensitivity obtained in the cross-validation was above 0.8, however, the specificity for these models ranged between 0.14 and 0.30 (Table 6).For the bBHB threshold of 1.4 mmol/L, the best performing model with high average sensitivity (0.75) and specificity (0.78) obtained in the cross-validation was again a model based on logistic regression (Table 7). As previously, variables to be included in the model were selected based on correlation coefficients and they were scaled using the StandardScaler, and oversampling was performed using the ADASYN method (Table 7, dataset 3). Outliers were removed using the LOF machine learning method. The model based on the SVC machine learning algorithm seems to be an equally good model (cross-validation sensitivity of 0.74 and specificity of 0.79). Calculations for this model were performed based on data with outliers removed using the LOF machine learning method and variables selected using the RFE machine learning method (Table 7, dataset 12). This model included such independent variables as protein percentage and acetone concentration. The variables were scaled using RobustScaler and oversampling was also performed using the ADASYN method. The mean MCC obtained in the cross-validation for the two discussed models were the same (0.38) and highest among the considered models for the bBHB threshold of 1.4 mmol/L (Table 7). The F2 score was 0.58 in both cases, which demonstrates the superiority of the models referred to above as compared to the others. The values of the analyzed metrics for the testing dataset were no different from those obtained during the cross-validation (Table 7).For the best performing models, independent variables were selected based on the calculated coefficients of correlation between variables. The features that were taken into account included FPR, ACE, mBHB, lactose percentage, lactation number and DIM. Oversampling using the ADASYN method was performed for the sets used for fitting of these models. The most desired values of metrics (sensitivity, specificity, bACC, MCC, F2 score) were obtained for a logistic regression model (bBHB cut-offs 1.2 and 1.4) as well as for a model based on the SVC (SupportVectorClassification) machine learning algorithm (bBHB cut-off 1.0).3.3. Performance of Regression ModelsIn the regression-based approach, a total of 12 models were selected out of the trained models (one per each dataset specified in Table 3), for which the greatest R2, the lowest MAE and the lowest RMSE were obtained in the cross-validation, having the lowest respective standard deviations. The values of the metrics selected in the cross-validation for the best performing prediction models, for each of the 12 datasets, are shown in Table 8.The highest R2 (0.39) in the cross-validation was obtained for the model based on the SVR machine learning algorithm. For this model, the related features were selected taking into account the values of correlation coefficients (Table 8, dataset 1). No outliers were removed from the set. Variables were scaled using StandardScaler. The MAE for the model in question was 0.34 and the RMSE was 0.55 while for the other models, the MAE was in the range of 0.30 and 0.35, and the RMSE—in the range of 0.44 and 0.58. The low R2 score obtained in the cross-validation can be indicative of the limited possibilities for using the regression model for predicting bBHB.To compare the classification and regression models, the estimated continuous values of the dependent variable bBHB were assigned to two classes (0 and 1) using the same rules that were used for the classification models, taking into account three bBHB thresholds (1.0, 1.2, 1.4). Subsequently, the same metrics were calculated for the testing dataset as those calculated for classification models (Table 9). For the regression-based model characterized by the greatest coefficient R2 (0.39) in the cross-validation, the sensitivity for the testing dataset was in the range of 0.32 and 0.40 according to the pre-defined bBHB threshold and the specificity was in the range of 0.94 and 0.97 (Table 9, dataset 1). The sensitivity was lower as compared to the recommended classification models, for which the sensitivity was in the range of 0.57 and 0.74 for the testing dataset.4. Discussion4.1. The Use of Classification Models for Diagnosing Subclinical KetosisThe classification models most commonly used for diagnosing of cows-at-risk of subclinical ketosis are those based on logistic regression [15,18,39]. In our study, the logistic regression model also proved to be the best, both when the bBHB threshold was defined as 1.2 or 1.4 mmol/L (Table 6 and Table 7, dataset 3). Taking into account the threshold of 1.0 mmol/L, the best performing classification model was the model based on the SVC machine learning algorithm (Table 5, dataset 3). The average sensitivity achieved in our study in the cross-validation for the best performing models ranged between 0.74 and 0.75 (Table 5, Table 6 and Table 7). Chandler et al. [18], who also used a logistic regression model, obtained lower sensitivity of 0.56 and 0.32 for primiparous and multiparous Holstein cows, respectively, and 0.40 and 0.42 for primiparous and multiparous Jersey cows, respectively. On the other hand, van der Drift et al. [15] obtained higher sensitivity (0.82) with equally high specificity (0.84), however, they did not perform cross-validation or external validation for the final model. The specificity obtained in this study (0.73–0.79) for the best models is not as high as that shown in the study of van der Drift et al. [15] and Chandler et al. [18] (0.83–0.99), however, given that the models proposed in this study are characterized by sensitivity, which is quite high, they can be considered for practical use. Denis-Robichaud et al. [40] achieved very high sensitivity and specificity (>0.90) for their model which included only ACE and mBHB, however, the level of ketone bodies in milk was determined using flow-injection analysis and not the FTIR method. The ketone bodies in milk as determined using flow-injection analysis are more strongly correlated with the ketone bodies in blood than the ketone bodies determined using the FTIR method [40].It should be highlighted that the values of metrics for the best models in the external validation on a testing dataset (about 0.70 for sensitivity and about 0.80 for specificity) were similar to those obtained in the cross-validation, which may be indicative of their good suitability for correct classification of new data.The most desirable values of metrics were obtained for logistic regression models (or models based on the SVC algorithm) when they were validated using a dataset containing features selected based on correlation coefficients. These features included milk yield parameters such as FPR, ACE, mBHB, and lactose percentage. Additional features accounted for in the models included DIM and lactation number. Less optimal results were obtained for the three datasets containing features selected using the RFE machine learning method. One of these datasets included only ACE, the other one also included protein percentage, and the third one also included—in addition to ACE and protein percentage—milk yield, fat percentage and FPR.Other authors also included ACE and mBHB in their logistic regression models. Chandler et al. [18], for example, used ACE in all the models studied by them (for primiparous and multiparous Holstein and Jersey cows), however, mBHB was not included in models oriented towards primiparous cows. On the other hand, Denis-Robichaud et al. [40] who took into account only ACE and mBHB, generated a model that allowed predicting subclinical ketosis with sensitivity and specificity greater than 0.90, however, as it was mentioned above, ketone bodies in milk were determined based on flow-injection analysis and not the FTIR method.The fat-to-protein ratio, in addition to ketone bodies level in milk, was a traditional tool used for screening for ketosis [15,41,42]. Hyperketonemia is associated with an increase in fat percentage and a decrease in protein percentage in milk, which increases the FPR. However, the inclusion by some authors of fat-to-protein ratio as the only feature in a model for predicting subclinical ketosis was not sufficient because the sensitivity of such models was in the range of 0.58 and 0.69 and the specificity—in the range of 0.66 and 0.71 [23,40,41], and these values were lower than those presented in this study.In future, it would be advisable to extend models for the prediction of subclinical ketosis to include other features, e.g., fatty acids in milk [18]. Fatty acids, mobilized from the fatty tissue, are characterized by a high concentration of long-chain fatty acids [14] which are taken up by the mammary gland and secreted in milk fat. Chandler et al. [18] indicated that Jersey cows, which had subclinical ketosis, produced milk with a higher concentration of monounsaturated fatty acids (MUFA) and trans fatty acids, and a lower concentration of short-chain fatty acids as compared to healthy cows.In our study, the recommended models also included such features as lactation number or DIM. A number of studies demonstrated that the risk of subclinical ketosis increases with lactation number [18,43,44], and therefore it is reasonable to continue using that feature in models. The logistic regression model generated by Chandler et al. [18] for predicting subclinical ketosis in primiparous cows also included the gestation length and the dry period length. The authors highlighted that primiparous cows with hyperketonemia remained pregnant seven days longer as compared to healthy cows. No such correlation was identified for older cows with subclinical ketosis. The authors suggested that the relationship of the features referred to above and the risk of subclinical ketosis should be studied further.4.2. The Use of Regression Models for Diagnosing Subclinical KetosisThe study also attempted to use a linear regression model for the prediction of bBHB and subsequently, based on the estimated bBHB, for the classification of cows as healthy or ketosis-affected in accordance with the pre-defined bBHB thresholds. However, even for the best model out of the selected ones, the coefficient R2 was relatively low (0.39) (Table 8). This model included the same features as the logistic regression model recommended in our study (FPR, ACE, mBHB, lactose percentage, lactation number and DIM) and it was based on the SVR machine learning algorithm (Table 8, dataset 1). Chandler et al. [18] also tested the suitability of linear regression models for predicting bBHB in primiparous and multiparous Holstein and Jersey cows. Regression models were fitted to data covering two periods: 5–11 DIM and 12–20 DIM. The R2 coefficient obtained by those authors in the cross-validation was in the range of 0.20 to 0.71 according to period and breed, and the highest values of the coefficient were obtained for primiparous Holsteins. The RMSE of prediction ranged between 0.29 and 0.92, and it was 0.55 for our best model. The R2 coefficients obtained in our study demonstrate that there are limited possibilities of using regression models for predicting bBHB and their application for the identification of cows-at-risk of ketosis. In contrary to our study, Chandler et al. [18] obtained higher sensitivity in the cross-validation for linear regression (0.53–0.74) than for logistic regression (0.31–0.55). In our study, the sensitivity obtained in the external validation using a testing dataset for the best regression model based on the SVR machine learning algorithm ranged between 0.32 and 0.40 according to the pre-defined bBHB (Table 9, dataset 1). To compare, the sensitivity for a testing dataset for the recommended classification models ranged between 0.57 and 0.74, and it was higher in the cross-validation (0.73–0.75) (Table 5, Table 6 and Table 7).5. ConclusionsThe study evaluated various machine learning algorithms designed for predicting if a cow is at risk of subclinical ketosis. The logistic regression model was found to be the best fitted model, which included features such as fat-to-protein ratio, acetone and β-hydroxybutyrate concentrations in milk, lactose percentage, lactation number and days in milk. Regression models were characterized by poor fitness to data. In the event that it is possible to acquire additional features as determined during the assessment of milk performance (e.g., milk fatty acids), it should be considered including such features in the model and validating the model with the new features. A greater number of observations, including repeated test-day records, could also help to achieve better results using the model. Using machine learning models and milk data, breeders can efficiently identify dairy cows-at-risk of subclinical ketosis and implement appropriate management strategies to optimize or prevent losses in milk production.

animals : an open access journal from mdpi

10.3390/ani13091570

PMC10177098

The stable microecological system in animal intestines is beneficial for the healthy growth of calves, and as zinc is an important trace element in the body, the question of how to better promote its absorption by the body has become an important research hotspot. Research has shown that an appropriate amount of zinc addition can promote the growth and proliferation of beneficial bacteria in the body’s intestines and inhibit the proliferation of harmful bacteria. This study showed that adding 80 mg/kg zinc amino acids to the diet can improve the growth performance, immune performance, antioxidant capacity, intestinal microbiota enrichment, and the intestinal microbial environment, as well as promote the healthy growth, of calves.

The aim of this study was to investigate the effect of dietary supplementation with zinc (Zn) amino acids at different concentrations on immunity, antioxidant capacity, and gut microbiota composition in calves. Twenty-four one-month-old healthy Angus calves of comparable body weight were randomly divided into three groups (four males and four females in each group) based on the amount of Zn supplementation added to the feed the animals received: group A, 40 mg/kg DM; group B, 80 mg/kg DM; and group C, 120 mg/kg DM. The experiment ended when calves reached three months of age (weaning period). The increase in dietary Zn amino acid content promoted the growth of calves, and the average daily weight gain increased by 36.58% (p < 0.05) in group C compared with group A. With the increase in the content of dietary Zn amino acids, the indexes of serum immune functions initially increased and then decreased; in particular, the content of immunoglobulin M in group A and group B was higher than that in group C (p < 0.05), whereas the content of interleukin-2 in group B was higher than that in the other two groups (p < 0.05). In addition, the content of superoxide dismutase and total antioxidant capacity in the serum of calves in group B was higher than that in group C (p < 0.05), and the MDA level was lower than in group C (p < 0.05). Moreover, alpha diversity in the gut microbiota of calves in group B was higher than that in group A and group C (p < 0.05); the dominant phyla were Firmicutes and Bacteroidota, whereas the dominant genera were Unclassified-Lachnospiraceae and Ruminococcus. Linear discriminant analysis showed that the relative abundance of Bacteroides in the gut microbiota of calves in group B was higher than that in group A, and the relative abundance of Prevotellaceae-UCG-003 was higher compared to that in experimental group C. Thus, dietary supplementation of 80 mg/kg of Zn amino acids to calves could improve the immune function and antioxidant capacity, as well as enrich and regulate the equilibrium of gut microbiota, thus promoting the healthy growth of calves.

1. IntroductionThe alternance between dynamism and stability in the gut microbial ecosystem impacts the animal host, and the type of microorganisms in the gut is closely associated with the host’s metabolism, immunity, as well as tissue and intracellular homeostasis [1]. A stable gut microbiota can regulate animal health by promoting nutrient digestion and absorption, maintaining intestinal barrier integrity, and enhancing organism immune function [2]. On the contrary, disruptions to the gut microbiota can cause metabolic disorders, such as diabetes and non-alcoholic fatty liver, which hinders animal growth and development [3,4]. Therefore, maintaining a stable gut microbiota, being beneficial to animal growth and development, is one of the focuses of animal husbandry practices.Zinc (Zn) is a component of more than forty different enzymes and two-hundred enzyme activators, and thus plays a key role in animal growth and development [5]. Inorganic zinc (IOZ) is the main form of Zn added to feed. However, due to its low absorption rate, unstable chemical properties, and environmental pollution capacity, IOZ has been gradually replaced by organic zinc (OZ), which has good palatability, high chemical stability, lipophilicity, and bioavailability [6,7]. Previous studies have shown that the addition of OZ to feed (protein zinc, amino acid chelated/complexed zinc, and polysaccharide zinc [8,9]) can alleviate diarrhea [10], improve immune function [11] and antioxidant capacity [12], promote intestinal health [13], and enhance reproductive performance [14], thus supporting the healthy growth of animals.Different sources and doses of Zn affect the gut microbiota of animals. The appropriate amount of Zn can promote the growth and proliferation of beneficial bacteria while inhibiting harmful bacteria in the gut. Yu et al. found that the addition of 120 mg/kg of OZ to the diet increased the relative abundance of Firmicutes, while decreasing the relative abundance of Proteobacteria in the intestinal tract of fattening pigs [15]. Xie et al. found that the addition of 100 mg/kg of zinc chitosan to the diet of weaned piglets increased the number of Lactobacillus in the cecum and colon, while decreasing the number of Escherichia coli and Salmonella [16]. Broom et al. and Højberg et al. also showed that a zinc-rich diet decreased the number of anaerobic bacteria and Lactobacillus in the ileum of piglets, but increased the number of Escherichia coli [17,18]. Few studies have been conducted on the application of OZ on calves, especially considering its effects on gut microorganisms.Therefore, the aim of this study was to analyze the effects of supplementation with Zn amino acid complexes at different concentrations on the immunity, antioxidant capacity, and gut microbiota composition of calves. High-throughput second-generation 16S rRNA sequencing was used for assessing bacterial community composition in the guts of calves. The findings discussed herein provide a theoretical basis for investigating the appropriate level of Zn amino acid complex dietary supplementation to calves.2. Materials and Methods2.1. Animals and Experimental DesignAll animal procedures used in this study were previously approved by the Animal Care and Use Committee (IACUC) of the Institute of the Ningxia Academy of Agriculture and Forestry Sciences, China (DK2019060). The experiment was conducted at the institutional experimental farm.Twenty-four Angus calves (30 ± 5.10 days of age; 69.34 ± 7.67 kg in body weight) were included in the study. Calves were housed from birth with their mothers in pens equipped with a separate feeding trough, which enabled the obtainment of supplementary pellets from ten days of age. Animals were equally divided into three experimental groups based on the level of Zn supplementation (eight animals per group): (i) group A, 40 mg/kg DM; (ii) group B, 80 mg/kg DM; and (iii) group C, 120 mg/kg DM. Zn was added in the form of Zn amino acid complexes (Zn content, 120 g/kg; total amino acid content, ≥21%), replacing the zinc in the basic feed premix (zinc content, 80 mg/kg; added in the form of zinc sulfate) and mixed with other raw materials to make pellet feed. The entire experiment lasted for 60 days.The animals were numbered and fed with pellets containing different amounts of Zn every day. The cow pen was equipped with a calf feeding pen, allowing only calves to freely enter and exit for feeding. Feeding of the calves with pellets occurred twice daily, at 9:00 and 16:00. Pens were equipped with a sink, cleaned and disinfected weekly, and the calves had access to hay and water ad libitum. The feed dietary composition and nutrient levels of the basal diet are shown in Table 1.2.2. Measurements of Growth Performance and Immunity Indexes2.2.1. Growth PerformanceCalves were weighed on an empty stomach in the morning at initial body weight (one month old) and final body weight (three months old), and the average daily gain (ADG) was calculated. Calf pellet input and the remaining amount were weighed for three consecutive days weekly to determine calf feed intake. The average daily dry matter intake and feed-to-weight ratio were also calculated.2.2.2. Serum Indicators of Immune and Antioxidant PerformanceAt 7 a.m. the next day after the experiment, 5 mL of blood was collected from the calf tail vein using a vacuum anticoagulant device, left to stand, centrifuged at 3500× g for 5 min, and divided into two samples in 1 mL centrifuge tubes, followed by storage at −20 °C in 1.5 mL Eppendorf tubes until further analysis. Serum concentrations of immunoglobulin A (IgA), immunoglobulin G (IgG), immunoglobulin M (IgM), and interleukin-2 (IL-2) were determined using an Enzyme-Linked Immunosorbent Assay in a Multiskan FC Microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). Colorimetric determination of serum glutathione peroxidase (GSH-Px), superoxide dismutase (SOD), total antioxidant capacity (T-AOC), and malondialdehyde (MDA) was conducted using an ELISA analyzer Rayto RT-6100 (Shenzhen Raydu Life Science Co., Ltd., Shenzhen, China).2.3. Metagenomic Analysis of the Gut Microbiome2.3.1. DNA Extraction and 16S rRNA Gene SequencingAt the end of the experiment, fecal samples of the calves supplemented with Zn amino acid complexes at different concentrations were collected by rectal extraction. Fecal samples were placed in 2 mL lyophilization tubes, which were immediately placed into liquid nitrogen and then stored at −80 °C until further analysis. Total DNA extraction and PCR amplifications were performed by QingKe Bio, Ltd. (Beijing, China). 2.3.2. Sequencing Data ProcessingThe quality control of raw sequencing reads was conducted using Trimmomatic software v.0.33. Primer sequences were identified and removed using cutadapt software 1.9.1 to obtain clean sequencing reads without adapters. Clean reads were then spliced by overlapping using Usearch software v.10. Spliced sequencing data were length-filtered according to the length ranges of different genomic regions, and chimeric sequences were identified and removed to obtain the final effective sequencing reads using UCHIME software v.4.2. Reads were clustered at a 97.0% similarity level using Usearch software to obtain operational taxonomic units (OTUs) [20]. Classification of obtained OTUs was conducted using QIIME2 software. Alpha-diversity indexes were calculated [21], and beta-diversity analysis was performed using QIIME software. The taxonomic annotation of species in each feature was performed using a plain Bayesian classifier with the SILVA database. The community composition of each sample was determined at each taxonomic level (phylum, class, order, family, genus, and species), and species at different taxonomic levels were determined using QIIME software. Abundance tables were generated for different taxonomic levels using QIIME software, while the microbial community composition of samples at each taxonomic level was constructed using R language. Finally, microbial biomarkers were identified in RStudio based on linear effect size (Lefse) analysis using the MicroBiotaProcess program package [22].2.4. Data AnalysisData were expressed as means ± standard deviations. Body weights, immune function indexes, and other data were initially processed in Microsoft Excel software. One-way ANOVA was performed in SPSS software (version 22.0) (IBM Corp., Armonk, NY, USA), and Duncan’s test was used for multiple comparisons, in which p < 0.05 indicated significant differences. The significance of the alpha-diversity of gut microbiota composition in different experimental groups was determined using t-test in QIIME2 software (version 2.0). Principal coordinate analysis (PCoA) was used for analyzing beta-diversity measurements. Linear discriminant analysis (LDA) effect size (LEfSe) analysis was performed using the MicroBiotaProcess program package in RStudio to determine microbial biomarkers, including the abundance of bacterial taxa from phylum to species level [21]. The thresholds for the false-discovery rate (FDR) and the log LDA score were 0.01 and 4, respectively.3. Results3.1. Growth PerformanceAs shown in Table 2, the final weight of calves showed an increasing trend with the increase in the concentration of Zn amino acid complex supplementation; the final weight of calves in group C was higher than in group A (p < 0.05); moreover, the average daily weight gain was 36.58% higher in group C compared to group A (p < 0.05).3.2. Immune FunctionAs shown in Table 3, IgM content in group A and group B was higher than in group C (p < 0.05). Moreover, IL-2 content in group B was higher than in the other two groups (p < 0.05). 3.3. Antioxidant CapacityThere was no significant difference in serum GSH-Px content among treatments, SOD and T-AOC levels were higher in group B compared to group C (p < 0.05), and the MDA level was lower in group A and group B compared to group C (p < 0.05) (Table 4).3.4. Gut Microbiota Composition3.4.1. Sequencing Data Quality ControlAs shown in Table 5, 1,200,040 reads were obtained in total, and 1,159,596 valid reads were obtained after quality control and the splicing of paired-end reads was conducted to remove low-quality reads and chimeras. The ratio of valid reads in all experimental groups of samples was more than 94%, and Q20 and Q30 values were above 99% and 96%, respectively, which indicated that sequencing results could adequately reflect the microbiota in the guts of calves receiving Zn supplementation.3.4.2. Number of OTUs As shown in Figure 1, 818 OTUs were identified in the three sample groups, and the OTUs uniquely assigned to group A, group B, and group C were 13, 3, and 5, respectively. OTUs shared by samples in the three groups corresponded to 89.40% of the total OTUs; in contrast, OTUs uniquely found in each sample group accounted for only 2.30% of the total OTUs. OTU compositions in fecal samples showed high similarity, indicating that the gut microbiota structure was relatively stable among calves receiving Zn supplementation at different concentrations.3.4.3. Sequencing DepthAs shown in Figure 2, the dilution curves of each sample group stabilized as sequencing depth increased, indicating that sequencing depth covered the bacterial species in the samples and that sequencing volume and depth were reasonable, such that they could be used for the analysis of bacterial population diversity.3.4.4. Alpha-Diversity AnalysisTable 6 shows that both group A and group B outperformed group C in terms of Simpson and Shannon indexes (p < 0.05).3.4.5. Beta-Diversity AnalysisPCoA results are shown in Figure 3A. Principal component 1 (PC1) contributed 28.78% to explanation of the variance of the samples, whereas PC2 contributed 21.25% to explanation of the variance of the samples. In addition, the samples between the three groups were separated, indicating large differences in the composition and structure of the gut microbiota of calves. In contrast, samples from the same groups clustered together, indicating small differences in intra-group diversity. Analysis of similarities (ANOSIM) results are shown in Figure 3B. The ANOSIM value between sample groups was R = 0.351 > 0, indicating that inter-group differences were greater than intra-group differences (p < 0.05). Collectively, the results indicated a high degree of reliability and reasonable grouping.3.4.6. Relative Abundance and Structure of Gut Microbiota at Phylum and Genus LevelsAs shown in Figure 4 and Table 7, the dominant microbial phyla in each sample group were Firmicutes, Bacteroidota, and Proteobacteria; the dominant genera were unclassified-Lachnospiraceae, Alloprevotella, and Ruminococcaceae-UCG-005. Differences in species abundance between groups were analyzed using non-parametric statistics at different taxonomic levels, and the results are shown in Table 8. The relative abundance of Lactobacillus and Faecalibaculum was higher in group B compared to group A at the genus level (p < 0.05); the relative abundance of Prevotellaceae-UCG-003, Faecalibaculum, and Bifidobacterium was higher compared to group C (p < 0.05); and the relative abundance of Bacteroides was higher in group C than in group A (p < 0.05).3.4.7. LEfSe AnalysisHerein, the LEfSe score threshold was set to 4, and species with LDA scores ≥ 4 were considered the dominant species with very significant high abundance. As shown in Figure 5, species with significant differences in group B were Spirochaetacene, Bacteroidaceae, unclassified Prevotellaceae, Prevotellaceae-NK3B31-group, Treponema, Bacteroides, and Prevotellaceae-UCG-003. The relative abundance of Bacteroidaceae increased (p < 0.05) and the relative abundance of Spirochaetaceae and Treponema decreased (p < 0.05), whereas the relative abundance of Prevotellaceae-UCG-003 increased compared with group C (p < 0.05).4. DiscussionZinc is a crucial component for animal growth and development, being also crucial for preserving proper metabolism [23]. Recent studies have demonstrated the value of dietary supplementation with adequate levels of zinc to support animal development performance [24]. Pei et al. demonstrated that weaned pigs fed with 450 mg/kg of ZnO nanoparticles had increased growth performance [25]. In addition, Graget et al. reported that adding 20 mg/kg of ZnO and Zn methionine to feed promoted the growth of lambs, the Zn-meth group having considerably greater daily weight gain than the ZnO group [26]. Chang et al. showed that neonatal calves receiving 80 mg of Zn methionine daily had increased daily weight gain after 14 days of supplementation [27]. Moreover, the average daily weight gain of bull calves was greatly increased by daily supplementation with 0.45 g of Zn methionine complexes (80 mg of Zn) [28]. Thus, the results of the current study are consistent with those of previous studies, in that the average daily weight gain of calves increased as the level of dietary zinc amino acid supplementation increased; in particular, group C outperformed group A in terms of weight gain. Subsequently, we investigated the effect of Zn amino acids at different levels on the immune function and antioxidant capacity of calves to better determine the impact of Zn supplementation on calves’ health.Animal health and metabolic state are reflected in blood markers. The levels of IgA, IgM, and IgG are significant markers of immune response, and the trace element Zn is intimately involved in the immune response [29,30,31,32,33]. The levels of IgM and IgG in the blood of calves increased after dietary supplementation with 80 mg/day of ZnO [27]. Wei et al. described that the levels of IgA, IgM, and IgG in the blood of calves showed an upward tendency with the increase in ZnO concentration (40–120 mg/kg) in the diet [34]. In the present study, no discernible effect was observed as a result of dietary supplementation with Zn amino acids on the levels of IgA and IgG in blood; in contrast, the level of IgM in group B was higher than that in groups A and C, which was inconsistent with the results of the above studies. This could be likely due to different sources of Zn used in this study, which led to differences in actual concentrations of Zn supplemented. IL-2 is a T-lymphocyte growth factor that plays an important role in the immune response [35]. Chen et al. revealed that feeding dairy cows with 60 mg/kg of Zn methionine dramatically raised the levels of IL-2 in the blood [36]. In the present study, the levels of IL-2 in the blood of calves increased initially and then decreased as the amount of Zn amino acids in the feed increased, thus suggesting that increasing dietary supplementation with 80 mg/kg DM of Zn amino acids will benefit the immune system of calves.Antioxidative compounds are produced by the body in response to harmful external stimuli to withstand challenges [37]. The body’s potential for antioxidant defense is indicated by levels of T-AOC, which is composed of numerous antioxidative compounds and enzymes. SOD can neutralize a great number of free radicals generated in the body as a result of stressors and scavenge reactive oxygen species (ROS) [38]. Moreover, the level of cell damage is indicated by MDA concentration, which is a byproduct of lipid peroxidation in cell membranes. Several studies have demonstrated that Zn can increase the body’s capacity for antioxidant response, which is directly linked to the equilibrium of the body’s redox system [39,40]. Yu et al. revealed that serum T-AOC levels during the fattening stage were considerably greater in pigs receiving 90 mg/kg, 120 mg/kg, and 150 mg/kg of cysteamine chelated Zn compared to pigs receiving 60 mg/kg [15]. The addition of protein Zn to the diet at concentrations greater than 80 mg/kg dramatically raised serum SOD levels, as demonstrated by Kannan et al. [41]. In the present study, the levels of T-AOC and SOD in the blood of calves tended to increase and then decrease as the concentration of dietary Zn amino acids increased, whereas the serum MDA levels increased. Thus, these findings suggest that, under the conditions employed in the current experiment, moderate addition of Zn amino acids to the diet improved the antioxidant capacity of calves, while an excess of Zn decreased antioxidant capacity. Subsequently, we carried out correlation analysis of the gut microbiome of calves following the addition of Zn amino acids at different concentrations to further investigate the ideal amount of Zn amino acids to be supplemented in calf diets.In order for calves to digest and absorb nutrients, the gut microbiota is essential, as it directly impacts calves’ health. When the gastrointestinal microbiota is unbalanced and prone to invasion by pathogenic bacteria, diarrhea and other intestinal disorders can occur [42]. Herein, the alpha-diversity of the gut microbiota of calves in group B was substantially greater than in groups A and C, showing that the addition of 80 mg/kg DM of Zn amino acids to the diet improved the variety of the gut microbial composition. Jensen et al. showed that the variety of the rectal microbiota was not affected by the addition of 2500 mg/kg of ZnO to the diet of weaned pigs [43] which was not in accordance with the results of the present study, possibly due to the differences in feed composition, nutrition level, feeding environment, and Zn source.Firmicutes play a significant role in how well proteins and carbohydrates are absorbed [44]. In contrast, Bacteroides primarily target non-fibrous materials for digestion, but species within this phylum can also colonize the gut to reduce the adherence of invasive pathogens [45,46]. It is known that obesity level is reflected in the ratio of Firmicutes and Bacteroides in the gut microbiota [47]. Tian et al. found that the dominant phyla in the gut microbiota of calves were Firmicutes and Bacteroidetes [48]. In the present study, the dominant microbial phyla and genera in the fecal samples of calves receiving Zn supplementation were not impacted by dietary levels of Zn amino acids. In addition, Ruminococcus and Bacteroides were the dominant genera in the gut microbiota of calves in all sample groups. The phylum Bacteroides is primarily involved in the hydrolysis of proteins and the degradation of carbohydrates [46]. According to Chang et al., the relative abundance of Bacteroides in the gut microbiota of calves increased considerably after receiving daily supplementation with 457 mg (80 mg/day) of Zn methionine [27]. In the present study, the relative abundance of Bacteroides in calves of group C was considerably higher than in group A. Moreover, supplementation with 120 mg/kg DM of Zn amino acids increased the utilization of protein and carbohydrate in the feed, thus promoting the growth of calves, based on average daily weight gain data.The immune system and Lactobacillus are closely associated [49]. The production of lactic acid prevents the growth and multiplication of pathogens such as Colidextribacter, and it has been demonstrated that Lactobacillus can lower local pH in the gastrointestinal tract [50]. To maintain the environmental acid and prevent the growth of infective microbial species in the gut, Lactobacillus can also convert lactic acid to butyrate. Hou et al. found that feeding weaned piglets with chitosan-chelated zinc (100 mg/kg zinc) increased the relative abundance of Lactobacillus in the cecum [13]. In addition, Faecalibaculum is a probiotic species that produces a natural antibiotic that can improve health in the colon by generating metabolites, such as butyric acid and short-chain fatty acids, among others [51]. In the present study, the relative abundances of Lactobacillus and Faecalibaculum were considerably higher in group B than in group A, suggesting that feeding calves with 80 mg/kg of Zn amino acids was more effective in improving intestinal health.Prevotella degrade starch into monosaccharides and other non-cellulose polysaccharides to provide energy to the body [52] as well as to boost nutritional digestion and absorption [53,54,55]. However, this degradation might promote inflammation [56]. In addition, Bifidobacterium is another physiologically significant intestinal bacterium that can interact with immune cells and control the immune system [57,58]. In another study, compared to the group receiving no dietary addition of Zn methionine, it was revealed that the addition of 70 mg/kg and 140 mg/kg of Zn methionine considerably increased the abundance of Bifidobacterium in the cecum of laying hens [59]. In the present study, a considerably higher relative abundance of Prevotella-UCG-003 and Bifidobacterium was found in group B compared to group C. Thus, the addition of Zn amino acids to the feed boosted the immune system of calves, while it also promoted nutrient absorption. Therefore, it was established that the optimal amount of Zn amino acids to be added to the feed given to calves was 80 mg/kg DM.Furthermore, using LEfSe analysis, significant differences in the abundance of Spirochaetacene, Treponema, Bacteroides, Prevotellaceae-UCG-003, and other species were found in group B. Among these, Spirochaetacene and Treponema are Gram-positive intestinal pathogenic bacteria [60,61]. By interacting with the immune system, Bacteroides stimulate T-cell-mediated responses and prevent potentially harmful bacteria from colonizing the gut [62,63]. Chang et al. demonstrated that 104 mg/day of ZnO supplementation decreased the incidence of diarrhea in calves and raised the relative abundance of Bacteroides in seven-day-old calves [27]. In the present study, compared to group A, a higher relative abundance of Bacteroides and a lower relative abundance of Spirochaetaceae and Treponema were found in group B, which suggests that including Zn amino acids in the diet of calves will improve the body’s immune system.Finally, Prevotellaceae-UCG-003 was first described by Koh et al. as having the capacity to control intestinal inflammation by triggering dendritic cells through succinate synthesis [64]. In the present study, Prevotellaceae-UCG-003 was found in greater relative abundance in group B than in group C, showing that dietary addition of the Zn amino acids at the levels used in group B was more effective in enhancing host immune function. Collectively, it can be stated that the addition of 80 mg/kg DM of Zn amino acids to feed improved intestinal health and led to an increase in the relative abundance of beneficial microorganisms in the guts of calves.5. ConclusionsIn the present study, the levels of IgM, IL-2, T-AOC, and SOD in the blood of calves in Experiment B were considerably greater than those in Experiments A and C, as well as the variety of fecal microbiota; Group B’s average daily weight gain rose by 13.41% compared to Group A.Additionally, LEfSe scores revealed a much higher relative abundance of Prevotellaceae-UCG-003 in group B than in group C. Moreover, the relative abundance of Bacteroides in group B was higher than in group A. Thus, the findings discussed herein highlighted that the addition of 80 mg/kg DM of Zn amino acids to the diet of calves could improve immune response and antioxidant capacity, enrich the gut microbiota, regulate the equilibrium of the gut microbiota, and promote the healthy growth of calves.

animals : an open access journal from mdpi

10.3390/ani11030915

PMC8004997

As the longissimus dorsi muscle is the largest muscle in the equine back, it has great influence on the stability of the spine and facilitates proper locomotion. In general, muscle function is determined by its specific intramuscular architecture. However, only limited three-dimensional metrical data are available for the inner organisation of the equine longissimus dorsi muscle. The thoracic and lumbar longissimus muscles of five formalin-fixed cadaveric horse backs of different ages and body types were dissected layerwise from cranial to caudal. Three-dimensional coordinates along individual muscle fibre bundles were digitised from the origin to the insertion and 3D models were created using imaging software and computed tomography. The muscle was divided into functional compartments and morphometric parameters (muscle fascicle length, pennation angles, muscle volume and the physiological cross-sectional area (PCSA)) were determined. Fascicle length showed the highest values in the thoracic region and decreased from cranial to caudal, while in most caudal compartments, fascicle length was less than 50% of the fascicle length in thoracic compartments. The pennation angles differ between compartments. In the cranial compartment, fascicles almost run parallel to the horizontal plane (mean angle 0°), while in the caudal compartment, the angles increase up to a mean angle of 38°. In the sagittal plane, the pennation angles varied from parallel (0°) in cranial compartments to 0–22° in the caudal compartments. The muscle volume ranged from 1350 cm3 to 4700 cm3 and PCSA from 219 cm2 to 700 cm2. This study lays the anatomical basis for a biomechanical model to simulate muscle function.

As the longissimus dorsi muscle is the largest muscle in the equine back, it has great influence on the stability of the spine and facilitates proper locomotion. The longissimus muscle provides support to the saddle and rider and thereby influences performance in the horse. Muscular dysfunction has been associated with back disorders and decline of performance. In general, muscle function is determined by its specific intramuscular architecture. However, only limited three-dimensional metrical data are available for the inner organisation of the equine longissimus dorsi muscle. Therefore, we aimed at investigating the inner architecure of the equine longissimus. The thoracic and lumbar longissimus muscles of five formalin-fixed cadaveric horse backs of different ages and body types were dissected layerwise from cranial to caudal. Three-dimensional coordinates along individual muscle fibre bundles were recorded using a digitisation tool (MicroScribe®), to capture their origin, insertion and general orientation. Together with skeletal data from computed tomography (CT) scans, 3D models were created using imaging software (Amira). For further analysis, the muscle was divided into functional compartments during preparation and morphometric parameters, such as the muscle fascicle length, pennation angles to the sagittal and horizontal planes, muscle volume and the physiological cross-sectional area (PCSA), were determined. Fascicle length showed the highest values in the thoracic region and decreased from cranial to caudal, with the cranial lumbar compartment showing about 75% of cranial fascicle length, while in most caudal compartments, fascicle length was less than 50% of the fascicle length in thoracic compartments. The pennation angles to the horizontal plane show that there are differences between compartments. In most cranial compartments, fascicles almost run parallel to the horizontal plane (mean angle 0°), while in the caudal compartment, the angles increase up to a mean angle of 38°. Pennation angles to the sagittal plane varied not only between compartments but also within compartments. While in the thoracic compartments, the fascicles run nearly parallel to the spine, in the caudal compartments, the mean angles range from 0–22°. The muscle volume ranged from 1350 cm3 to 4700 cm3 depending on body size. The PCSA ranged from 219 cm2 to 700 cm2 depending on the muscle volume and mean fascicle length. In addition to predictable individual differences in size parameters, there are obvious systemic differences within the muscle architecture along the longissimus muscle which may affect its contraction behaviour. The obtained muscle data lay the anatomical basis for a specific biomechanical model of the longissimus muscle, to simulate muscle function under varying conditions and in comparison to other species.

1. IntroductionBack pain is a very common problem in horses and very hard to detect. It is often discovered too late by poor performance or gait abnormalities [1]. Due to the variety of possible clinical signs, equine back problems are often challenging to diagnose [2,3,4]. To avoid long-term effects, such as secondary skeletal modifications, early detection of pain and its cause is essential and poses a challenge to horse owners and veterinarians [5,6].The longissimus dorsi muscle (LG) is the largest muscle of the back and is part of the active epaxial musculoskeletal system, providing stability of the spine [7,8]. A decrease in spinal stability and muscle strain of the epaxial back muscles are considered the main causes of back pain in humans [9]. Because of the important role of the thoracolumbar longissimus dorsi muscle, especially in the ridden horse [8], it is plausible that strain in this muscle might lead to back pain.Skeletal muscle architecture describes the internal arrangement of the muscle fibres in a muscle and partly determines the contraction behaviour [10]. Bundles of muscle fibres are surrounded by perimysium—a sheath of connective tissue. Those muscle fibre bundles or fascicles are distinguishable in formalin-fixed tissue and can therefore be used to characterise muscle architecture. Architectural parameters, such as fascicle length and pennation angle, have a major impact on the contractibility and muscle function [11,12,13,14,15]. Therefore, muscle morphology clearly determines muscle function.Besides architectural parameters, metabolic parameters, such as muscle fibre type distribution, also influence muscle function [12,14]. Previous anatomical studies of the equine back mainly focus on the general skeletal and ligamentous structures and their function [16,17,18], whereas detailed data and knowledge about the inner muscle architecture are limited. Von Scheven (2001) and Ritruechai et al. (2008) gave important insight into anatomical characteristics of the equine longissimus muscle. Ritruechai et al. (2008) showed segmental variations in the cross-sectional area and moment arm length, as well as differences in direction cosines measured in a multitude of segments. This work unravelled regional specialisation of the longissimus dorsi muscle, also suggesting functional consequences of the anatomical characteristics. However, the segmental methodological approach and evaluation of mean direction cosines left scope for interpretation. We therefore undertook the challenge to reduce intervals of measurement, increase resolution and quantify fascicle length and three-dimensional orientation. In addition, we aimed at producing a 3D model of the completely thoracolumbar LG in order to visualise inner muscle structure of the largest muscle of the horse back and lay a necessary quantitative basis for future biomechanical models that could be used to mimic or simulate muscle function.Additional information, like PCSA, total length and volume, should give insight into the general longissimus muscle architecture and support already available literature. With the available literature, the following hypotheses are formulated: fascicle length will decrease and PCSA will increase from cranial to caudal aspects of the LG. In addition, the fascicle orientation to the spine will change along the length of the LG. In addition to the description and measurements of inner muscle architecture of the LG as one of the most important back muscles, a three-dimensional visualisation will help understand muscle structure more thoroughly and also be of great didactic value for veterinary anatomy training.2. Material and Methods2.1. Specimen PreparationIn this study, cadaveric backs obtained from five horses from the anatomical collection of the University of Veterinary Medicine Vienna were used to examine the thoracic and lumbar LG. None of the horses had any known musculoskeletal problems. All animals were euthanised or died at the equine hospital of the University of Veterinary Medicine Vienna, for reasons unrelated to this study. As the study was performed on dead horses, it did not require ethical approval. However, the animals owners’ consent to dissect and publish resulting data was obtained according to the standard procedure, which was approved by the ethics and animal welfare committee of the University of Veterinary Medicine Vienna. The horses were dissected according to the “Good Scientific Practice and Ethics in Science and Research” regulation emplemented at the University of Veterinary Medicine Vienna. Only small breeds, ponies of different breeds and a Standardbred foal (one Shetland pony, one Shetland–Haflinger crossbreed, one Dutch Riding Pony, one Icelandic Horse and one Standardbred) were used, as the storage space as well as computed tomography (CT) gantry dimensions were limited. The age ranged between two months and 29 years. The body mass ranged from 119 kg to 280 kg with mean ± SD of 177.2 kg ± 58.02. The horses were for recreational use.Prior to LG preparation, skin, cutaneous muscle, fat and muscles of the shoulder girdle, such as the latissimus dorsi, dorsal and ventral serratus muscle and trapezius muscle as well as other epaxial back muscles like the spinalis muscle and the gluteus medius muscle, as an intrinsic muscle of the hindlimb, were removed, in order to expose the region of interest. The backs were trimmed to a minimal size, ranging from the 6th cervical vertebra to the pelvis. Subsequently, they were fixed in 4% formalin solution to preserve the tissue.To guarantee an adequate fixation of the muscle, formalin was also injected into the multifidi muscles and into the iliocostalis muscle along the thoracolumbar LG. The LG thoracis et lumborum itself was not injected with fluid to avoid local muscle fascicle compression. Then the whole backs were stored in a formalin tank until further processing.2.2. Dissection, Digitisation, Three-Dimensional ModellingScrews were drilled into the spinous processes (from the 6th cervical vertebra to the sacrum), the coxal tuber and the body of the ribs to serve as reference coordinates during the digitisation process. Helical CT scans of the formalin-fixed backs were performed with a 16-slice helical CT scanner (Somatom Emotion, Siemens AG Medical Solutions, Erlangen, Germany) from the last cervical vertebrae to the pelvis including the os sacrum. Technical settings used for the scans were 130 kV tube voltage, 150–200 mAs tube current, collimation of 1.2 or 0.6 in one case, respectively, rotation time of 1–1.5 s and slice thickness of 0.75–1.5 mm. The CT DICOM image sequences were imported into the 3D software package Amira (version 5.3.3, Visage Imaging, Berlin, Germany) and a surface model of each skeleton was created. All skeleton models were oriented into a standardised coordinate system to make the obtained data comparable. For establishing this standardised coordinate system, we used a set of reference points dorsal to the spinous processes and ventral on the ventral crest of the body of each thoracic and lumbar vertebra. From this point cloud, the three principal axes were calculated. The x-axis was defined as the cranio-caudal axis, the y-axis as the left–right axis and the z-axis as the dorsoventral axis (see Figure 1).The thoracic and lumbar part of the LG was dissected in layers from cranial to caudal and superficial to deep. Layer thickness was adjusted to the total muscle thickness and degree of variance of muscle fascicle orientation (approx. 2–20 mm). The thinner the muscle and the more changes were seen in fascicle orientation, the thinner the layers were designed. For each layer, three-dimensional coordinates of representative muscle fascicle were captured, using a Microscribe G2X digitiser (Revware Systems, Inc., San Jose, CA, USA). At least three fascicles were recorded per layer. Origin, insertion and at least three points in between along the course of the fascicles were recorded. Muscle fascicles are bundles of muscle fibres that are surrounded by a sheath of connective tissue—the perimysium. Those muscle fibre bundles or fascicles are distinguishable in formalin-fixed tissue and could therefore be used to characterise muscle architecture.This procedure was repeated layer by layer for the whole muscle length and for both body sides, in order to display and describe the inner fascicle structure for the entire thoracolumbar LG.For each fascicle digitisation session, reference points in neighbouring spinous processes, ribs and coxal tuber were also digitised and with their help the digitised muscle data could be imported to Amira and aligned to the surface model of the same skeleton within the same reference system, as described before.2.3. Data Analysis For further description and analyses, the muscle was divided macroscopically into four compartments (A–D, see Figure 2), with compartment A being the most cranial part inserted at the last cervical vertebra, followed caudally by compartment B and C. Architectural parameters such as fascicle length, muscle volume, physiological cross-sectional area (PCSA) and pennation angles were calculated from digitised data or obtained from the axial CT scans of the backs. The fascicle length was calculated with MATLAB R2015b (The MathWorks, Inc., Natick, MA, USA) as the sum of the distances between the consecutive digitised points between origin and insertion of each digitised fascicle. Mean fascicle length was calculated for each compartment of the LG (see Figures 3–5). To compare between animals of different sizes, relative fascicle lengths were used. One hundred percent was set for the compartments with the longest mean fascicle lengths. As in compartment A and B, due to the thin shape of the muscle, fewer data for fascicle length could be collected, and data for these two compartments, A and B, were pooled for fascicle length calculations. Muscle volume was determined for the right longissimus muscle of all animals with the help of the CT image volumes based on manual segmentation in Amira (version 5.3.3, Visage Imaging, Berlin, Germany). As no specific side differences were assumed, we only used the right side data for determining muscle volume. Due to tissue shrinkage that occurs during the fixing process in formalin, the overall measurements are slightly shorter than the in vivo situation [19]. Therefore, we do not discuss quantitative but relative data that allow for seeing the biologically relevant differences across intra- and inter-individual longissimus muscle.Finally, PCSA was calculated for the LG by dividing the muscle volume by the mean fascicle length of the muscle.To describe the orientation of the fibre bundles, pennation angles were measured as 2D projected angles between the fascicles (assumed as straight lines from origin (first point of digitalisation) to insertion (last point of digitalisation)) and anatomical planes of the reference coordinate system. Two angles were determined for fibres of each compartment: (i) the horizontal projection of the angle between fascicle and sagittal plane (xz plane) and (ii) the sagittal projection of the angle between fascicle and horizontal plane (xy plane). Data are given as boxplots (see Figures 6 and 8) and mean + standard deviation (SD) angles. Digitised fascicle orientation enabled us to generate 3D angles (visualised in the 3D model)—however, we decided to present 2D values for the pennation angles in two different planes as they are much easier to understand and more suitable in the anatomical description. The collected data were processed in Microsoft Excel 2010 (Microsoft Corporation, Redmond, WA) and calculated using MatLab R2015b (The MathWorks, Inc., Natick, MA, USA). Data are given as mean values with standard deviation. A linear regression (goodness of fit (R2)) was calculated to show that the fascicle length as well as the orientation (pennation angles) of the fascicle change from cranial to caudal.3. Results3.1. Anatomical Description The thoracolumbar LG extends in its total expansion from the sacrum and iliac wing all along the vertebral column to the last cervical (C7) vertebra. Along its length, the LG inserts into the mammillary, transversal and dorsal spinous processes and to the ribs. It has its greatest dimension in the lumbar region while it narrows from caudal to cranial. The muscle is covered by the thoracolumbar fascia that releases an inner fascia at the height of the 13th rib, which partly divides the muscle. The inner muscle architecture varies widely along the length of the LG. Therefore, different compartments were defined and described from cranial to caudal, shown in detail below.Compartment A is the most cranial part of the muscle. Here, the muscle inserts into the transversal process of the last cervical vertebra via flat tendons. The muscle strand is thin and fascicles run almost parallel to the spinal column. Most fascicles are long and span over several vertebral segments. In addition, shorter fascicles can be found, particularly in the deeper layers of the muscle. The caudal part of compartment A is overlapped by the muscle belly of the following compartment B and only fully visible after removing cranial parts of compartment B, see Figure 2.At the height of the 5th rib, the cranial part of compartment B follows and partly overlaps compartment A. The muscle becomes bulkier and muscle diameter increases. The fascicles run almost parallel to each other in a caudodorsal to cranioventral direction. Macroscopically, fascicles are shorter than in compartment A. The aponeurosis of the muscle provides origin for the fascicles that run in a uniform course parallel to each other towards their insertion at the articular, mammillary and transversal processes of the thoracic vertebrae.The diameter of the muscle continues to increase from the height of 13th to 18th rib, where the muscle has its largest expansion. Compartment A and B were joined as one functional unit. They were only split during the preparation process, because of some overlapping muscle parts, and then rejoined in the analysis.The separation between compartment B and C is not apparent from the outside of the muscle, because of the thoracolumbar fascia that covers the muscle. In the lumbar region, the fascia gets thicker and joins the supraspinous ligament. The classification is based on the internal structure of the muscle, as at the height of the 13th rib, the thoracolumbar fascia releases a fascia sheet, which divides approximately half of the muscle into a dorsomedial and a ventrolateral portion. The thoracolumbar fascia provides the origin for fascicles that span between the fascia and the ribs, as well as between the fascia and the aponeurosis of the multifidus muscle. Compartment C shows the structure of a multipennate muscle, as the inner fascia not only divides the muscle into two parts, but also provides the origin for fascicles that span between the inner fascia and the deep layer of the fascia spinocostotransversalis. In compartment C, fascicles are no longer parallel to each other, as the inner fascia as the origin influences the orientation of the fascicles. To describe and discuss the results, the fascicles are grouped according their orientation and therefore also their origin. Fascicles lying dorsal to the inner fascia and arising at the thoracolumbar fascia are grouped in compartment C1. Fascicles lying dorsally to the inner fascia and originate at the inner fascia are assigned to C2. In compartments C1 and C2, fascicles are orientated from caudolateral to craniomedial. Fibres that are situated ventrally to the inner fascia are divided into compartment C3 and C4. Muscle fascicles that originate at the inner fascia are grouped in C3 and fascicles that span between the thoracolumbar fascia and the ribs, having no contact with the inner fascia, are assigned to C4. In compartments C3 and C4, fascicles are orientated from caudodorsal to cranioventral, see Figure 3.After removal of the gluteus medius muscle that attaches to the aponeurosis of the LG, as far forward as the 18th thoracic vertebra, compartment D becomes visible. In this region, the LG can be divided macroscopically into two compartments. It has a dorsal thin muscle belly, which reaches the os sacrum and a ventral wide and flat portion that attaches to the iliac crest. Fascicles from the dorsal portion are grouped in compartment D1, fascicles from the flat ventral portion are assigned to compartment D2. Fascicles in compartment D1 originate at the thoracolumbar fascia and insert at the mammillary, transversal and dorsal spinous processes. They also attach to the fascia that covers the multifidi muscles and are orientated from caudolateral to craniomedial, while in compartment D2, fascicles run from caudodorsal to cranioventral, as they originate at the thoracolumbar fascia and attach to the ribs, see Figure 4.3.2. Inner Muscle ArchitectureAll data are presented as mean ± SD. Mean fascicle length of all animals was 65.09 ± 24.71 mm with a minimum length of 12.47 mm and a maximum length of 202.19 mm. Fascicle length showed the highest values in the cranial compartments (A + B) and decreased from cranial to caudal. Compartment A + B with the longest fascicle length was set to 100%. The middle part of muscle compartment C showed 74.93 ± 9.6% compared to the fascicle length in compartment A + B. In compartment D, fascicle length was less than half of the length (44.53 ± 10.86%) of the fibre bundle length in compartment A + B (see Figure 5).The muscle volume of the LG ranged between 1350 cm3 and 4700 cm3 with a mean (±SD) volume of 2180 ± 1269.88 cm3. The PCSA was calculated for each examined longissimus muscle and ranged, depending on the muscle volume and mean fascicle length, from 219.42 cm3 to 700.40 cm3 with a mean PCSA of 365.53 ± 174.07 cm3.Horizontally projected pennation angles of the fascicles relative to the sagittal plane (describing the orientation of the fascicles in the xy relative to the xz plane—see Figure 1) ranged from 0.5 degrees to 34 degrees and varied between compartments. An angulation of 0 degrees means that the fascicle runs parallel to the sagittal plane (spine). For the left body side, negative angles indicate that the fascicles are running from medial (origin) to more lateral (insertion), while positive angles indicate that fascicles are running from lateral (origin) to more medial (insertion). In compartment A, the mean angle is 8.8 ± 6.0°, with the fascicles running almost parallel to the spine and slightly in the lateral direction. Fascicles in compartment B almost run parallel to the spine with a mean angle of 8.8 ± 5.9°. In compartment C1, the direction of the fascicle is influenced by the inner fascia and therefore the pennation angles varied widely within this muscle part. For the fascicles lying dorsally to the fascia, the mean angle was 20.1 ± 4.9° and for the fascicles that arise at the fascia C2, the mean angle was 24.3 ± 6.6°, while the mean angle for the ventrally lying fascicles in C3 was 3.6 ± 2.7° for the fascicles arising at the fascia C4 and 6.7 ± 3.8° for the ventral ones. In compartment D, the mean angle was 16.6 ± 9.1° and varied between 4.9° and 32.3° for fascicles of the dorsal muscle part (D1) and 19.1 ± 7.7° for the ventral fascicles (D2), with sub-compartment D2 showing most mean variation, see Figure 6 and Figure 7.Sagittaly projected pennation angles of the fascicles relative to the horizontal plane describe the orientation of the fascicles in the xz relative to the xy plane (see Figure 1). An angle of zero degrees means that a fibre is parallel to the horizontal plane, while positive angles indicate a fibre course from caudodorsal (origin) to cranioventral (insertion). Negative angle values would indicate fibre course from ventral (origin) to dorsal (insertion), but did not occur in the present study. In the cranial part of the muscle (compartment A), the fascicles run almost parallel relative to the horizontal plane (xy plane), and the mean angle was 2.5 ± 10.3°. The more caudal the fascicles were situated, the steeper the angles. While in compartment B the mean angle was −20.4 ± 3.6°, in compartment C, the mean angle varied from fascicles dorsally to the inner fascia between C1 (−16.3 ± 8.9°), and C2 (−7.2 ± 6.7°). In C3 and C4, the angle was reduced and lay between −25.1 ± 7.0° for C3 and −23.8 ± 7.4° for C4. In the most caudal part of the muscle compartment D, the angles to the horizontal plane showed a mean angle of −29.1 ± 12.2° in the dorsal part (D1) and −38.0 ± 8.6° in the ventral part (D2), see Figure 8 and Figure 9.4. DiscussionMuscle performance highly depends on architectural parameters such as fascicle length and pennation angle. Thus, a detailed knowledge of inner muscle architecture is necessary to understand muscle function during movement [12,14,20,21,22]. Using dissection and digitisation in five horse backs, we were able to capture architectural data of the thoracolumbar LG. With the help of three-dimensional reconstructions, regional differences in fascicle architecture of the thoracolumbar LG of the horse could be quantified and documented for the first time.Muscle Architecture The LG muscle is considered an epaxial muscle with a segmental structure. Here, we give a detailed anatomical description of the anatomical subdivisions of the equine thoracolumbar LG and compare them in terms of muscle architecture parameters. While compartment A and B are both unipennate, A inserts at the last cervical vertebra and compartment B at the first thoracic vertebrae and ribs. In compartment C, an internal fascia is apparent leading to a completely different arrangement of fascicles and bipennate inner structure. The most caudal portion of the lumbar LG, compartment D, can be divided into an upper belly reaching the os sacrum, while another rather flat portion is attached to the iliac crest. Whether a muscle is unipennate or mulitpennate also influences fascicle length and contractile forces. Fascicle length showed the highest values in the cranial compartment (A + B) and decreased from cranial to caudal, with compartment C showing 75% of the cranial fascicle length, while in compartment D, fascicle length was 45% of the fascicle length in compartment A + B. Therefore, we can conclude that fascicle length decreases from cranial to caudal.Our results show that the fascicle length decreased from cranial to caudal, with most caudal fibres in compartment D being about 40% of the length of the fascicles in the cranial–thoracic longissimus muscle, therefore, this part of the hypothesis is supported. This is in accordance with previous studies [8,23] that also reported that the muscle fascicles get shorter from cranial to caudal. The high standard deviation is a result of the different horses used for this study. Fascicle length depends on how many sarcomeres are arranged in series. Longer fascicles have a greater contraction velocity and a greater excursion over which a muscle can exert its force [24,25,26]. The longer fascicles of the thoracic part of the LG indicate that the LG has a greater muscle excursion and contraction velocity in that region compared to the lumbar region, where fascicles are shorter. Studies about the mobility of the spine in the horse showed that the movement between the vertebrae decreases from cranial to caudal due to differences in intervertebral articulation. Between the second and the 17th thoracic vertebra, most lateral flexion occurs, while there is almost no dorsoventral flexion. The spine between the 17th thoracic and the sixth lumbar vertebra shows the least mobility, while between the sixth lumbar vertebra and the os sacrum, a lot of dorsoventral flexion occurs [27]. The greater range of thoracolumbar mobility between the 14th and 18th thoracic vertebra [28] predisposes the thoracic LG with its long muscle fascicles to play an important role in movement and lateral bending of the spine.The pennation angles to the horizontal and sagittal plane become bigger in the lumbar region, compared to the thoracic region. In the thoracic part of the longissimus, fascicles run almost parallel to the spine (sagittal plane) and to the horizontal plane, while in the lumbar region (compartment C + D), the angles get steeper and the fascicles no longer run parallel to each other, as the inner fascia influences the orientation of the fascicles The last part of the hypothesis (the orientation of the fascicle to the spine will change along the length of the LG) can be supported.Other studies on the equine LG [23] calculate the pennation angle as the angle between the aponeurosis or internal fascia and the muscle fascicle. We calculated the pennation angles to the horizontal and sagittal plane to describe the orientation of the fascicles [15]. The pennation angles become larger from cranial to caudal. The pennation angles of the cranial thoracic LG (compartment A + B) indicate a high contraction velocity and potential for high lateral excursions. Together with the fact that the fascicles almost run parallel to the spine, this suggests that the thoracic longissimus muscle rather plays a role in motion rather than in stabilisation, especially as a lateral flexor of the spine [23,29].In contrast, in pennate muscles, more muscle fibres can be packed into the muscle volume. Therefore, higher forces can be generated [14,30]. In the lumbar region of the LG, the inner muscle architecture changes compared to the more cranial compartments. While the fascicle length decreases from cranial to caudal, the fascicles no longer run parallel to the horizontal and sagittal plane. The lumbar LG is partly divided by the inner fascia into a dorsomedial and a ventrolateral portion. The fascia not only divides the muscle, it also provides an insertion origin for fascicles that are attached to it. This makes the muscle be built up like a multipennate muscle. The inner structure of the longissimus in the lumbar region (compartment C, D) indicates that in this region, the muscle can exert large forces and plays an important role in stabilisation of the spine. Our results are in line with previous studies [8,23], which found that the lumbar LG has the potential to excert greater forces than the thoracic LG, due to its inner architecture.These architectural findings match with kinematic studies [28,31,32] that showed a higher proportion of excursion in the cranial thoracic vertebral column. The thoracic part of the LG has long fascicles that almost run parallel to the vertebral column, which results in a high contraction velocity and the prevalence of a larger excursion, as mentioned before.Stability of the thoracic vertebral column is provided by the rib cage and other trunk muscles. In the lumbar region, the LG has the largest PCSA. In addition to the large PCSA in the lumbar region, short fascicle lengths and big pennation angles underline the important role in stabilisation of the spine in this area. In the caudal part, the LG stays via its aponeurosis in close contact with the middle part of the gluteal muscle that is one of the major propulsion muscles of the hindlimb. In this area, the muscle seems to have an important role in stabilising the lumbar spine against dynamic forces during push-offs of the hind limbs [16]. In pennate muscles, as the caudal compartments of the LG, fascicles rotate to greater angles of pennation as they shorten [33]. For that reason, the velocity of the contracting fibres can be slower compared to the muscle velocity and the caudal compartments of the LG might still be capable of fast contraction.Using muscle volume and fascicle length, the PCSA can be estimated. It describes the number of parallel sarcomeres and is directly proportional to the force a muscle can exert [34,35]. It is therefore a good parameter to estimate the force potential of a muscle. Von Scheven [23] calculated the PCSA for the cranial middle and caudal LG and detected that the caudal part has significantly higher PCSA. Volumes, and therefore PCSA, vary widely in our study because of the different races, ages and sizes of the horses. In addition, using formalin fixation leads to tissue shrinkage and confounds the interpretation of total measures [19]. Therefore, we decided to leave measures of volume and calculation of PCSA as important parameters of muscle architecture in the results section, but do not use them for discussion to avoid overinterpretation of heterogenous samples and methodological influence on total measures and, consequently, this part of the hypothesis could neither be supported nor rejected.Muscle function is not only determined by its inner architecture, but partly also by its muscle fibre type composition. Muscles can be more postural or locomotory, depending on if they have more slow type I muscle fibres or more fast type II muscle fibres [36,37,38,39].Previous studies showed that most muscles in horses have a combination of different muscle fibre types, but fast type 2 muscle fibres are usually more common [37,40]. We exemplarily also took fresh muscle samples of superficial and deep aspects of six different positions along the thoracic and lumbar longissimus muscle to underline fibre type results for horses that have already been published [37,40]. Together with the fascicle data, the information on fibre type and diameter allows a broader view on the physiological parameters of the muscle. All of our samples showed relatively more fast type II (72.81% ± 6.07) fibres and, therefore, our results match with the literature about the longissimus muscle [37,40] concerning muscle fibre type proportion. Studies on the gluteus medius muscle in horses [41] reported differences in fibre type composition in different sampling depths. An increasing number of slow type I fibres was found from superficial to deep. In the LG, we could not find differences concerning fibre type proportions between deep and superficial probes. Kawai et al. [40] found regional differences in fibre type proportion between the thoracic und lumbar LG. Similarly to others [8], we could not confirm differences in fibre type proportion at the sampling sites along the longissimus muscle. In our muscle samples, the measured mean muscle fibre diameter of the fast type II fibres seemed to be slightly larger than the slow type I fibres, although we did not perform any statistical analysis due to small sample size. These results match with other studies about the equine gluteus muscle [41] and the equine lumbar longissimus muscle [42].For both fibre type proportion and muscle fibre diameter, it has been demonstrated that the age and training status of the horse has an impact [43]. Further studies will be needed to confirm these findings as we only had samples from one muscle to examine fibre size and fibre type proportion.5. ConclusionsAnatomical observations and three-dimensional measurements reveal a very complex inner structure of the LG in the horse. Based on our three-dimensional model, the thoracolumbar LG of the horse with its multiple origins and areas of insertion and its complex inner structure due to the inner fascia could be described, and are also shown in a three-dimensional model for the first time. Muscle fascicle architecture changes within the LG from cranial to caudal. These regional anatomical differences indicate that the LG performs different functions along its length. Data for fascicle architecture such as fascicle length and fascicle orientation using the digitisation tools make all the measurements visible and represent the muscle fascicle architecture in its entirety. With the obtained detailed data, a base for enhanced understanding and further biomechanical studies, such as simulating the muscle function, is laid (see Figure 10).

animals : an open access journal from mdpi

10.3390/ani11113272

PMC8614362

It is common practice for animal shelters to evaluate the behavior of dogs a few days after admission. These evaluations typically consist of a series of tests and subtests that expose dogs to diverse stimuli and situations they might encounter postadoption. Limited information exists on whether behaviors displayed during an evaluation predict a dog’s length of stay at the shelter. We examined records from 975 dogs behaviorally evaluated and released for adoption at a New York shelter. Proportions of the study population evaluated as displaying concerning or especially dangerous behavior on tests and subtests were generally low. Nevertheless, dogs’ scores on some tests or subtests (food guarding and meeting another dog) predicted length of stay at the shelter. Dogs evaluated as showing dangerous behavior had longer lengths of stay than dogs evaluated as showing either concerning behavior or no concerning behavior; the latter two groups did not differ from one another in length of stay. We suggest that dogs with challenging behaviors have smaller pools of potential adopters, which leads to longer lengths of stay. Our findings may aid shelter management of dog populations and help highlight dogs needing special adoption efforts to avoid long stays at shelters.

Most animal shelters conduct behavioral evaluations before making dogs available for adoption. However, little information exists on whether behaviors displayed during these assessments predict a dog’s length of stay at the shelter. We reviewed nearly 5 years of records from 975 dogs released for adoption at a New York shelter to see whether behaviors shown during their evaluation predicted length of stay. For most tests and subtests, the prevalence of concerning and especially dangerous behaviors was low. Nevertheless, dogs’ scores on some tests or subtests—food guarding and meeting another dog—predicted length of stay at the shelter. Dogs evaluated as showing dangerous behavior had longer lengths of stay than dogs evaluated as showing either concerning behavior or no concerning behavior; the latter two groups did not differ from one another in length of stay. The most likely explanation for the relationships found between behavior during the evaluation and length of stay at the shelter is that dogs with challenging behaviors had smaller pools of potential adopters. Understanding the relationships between performance on behavioral evaluations and length of stay may inform shelter management of canine populations and also help identify dogs requiring special adoption efforts to avoid long shelter stays.

1. IntroductionThere is much debate over the usefulness of canine behavioral evaluations conducted at animal shelters. These evaluations typically occur several days after a dog has been admitted to the shelter and include a series of tests and subtests meant to assess a dog’s reaction to diverse stimuli and situations. Proponents of behavioral evaluations see them as useful, when combined with other sources of information, in promoting better informed matches between dogs and adopters and helping to prevent dangerous dogs from entering communities [1,2]. Critics describe behavioral evaluations as provocative and unlikely to reliably predict behavior postadoption, especially because obviously dangerous dogs are usually screened out of the population at or shortly after intake (i.e., before the behavioral evaluation), resulting in a low prevalence of warning or biting behaviors in the evaluated dog population [3,4]. Additionally, critics suggest time and resources would be better spent engaging with dogs in positive activities they will experience in adoptive homes, such as training, taking walks on a leash, and playing with other dogs [3]. Several studies have surveyed adopters of shelter dogs to determine if results from shelter behavioral evaluations predict the behavior of dogs in adoptive homes. In general, these studies have found poor predictability for behaviors such as food guarding and separation-related behaviors and somewhat better predictability for behaviors related to sociability and fearfulness [5,6,7,8,9,10,11]. Another approach taken to assess the effectiveness of shelter behavioral evaluations uses privately-owned dogs: owners complete the Canine Behavioral Assessment and Research Questionnaire (C-BARQ), a validated questionnaire [12], and then dogs are evaluated by researchers using different shelter assessments. Correlations between behavior reported by owners and behavior scored during shelter behavioral evaluations are typically weak to moderate at best [13,14].Less explored than their predictability regarding behavior postadoption, is whether results from behavioral evaluations predict other metrics of interest to shelters, such as the percentage of adopted dogs that are returned (often called return rate) and length of stay (time from intake to adoption). Some information is available for the specific tests that assess food guarding because most shelters test for this behavior and about half of those surveyed in one study did not make food guarding dogs available for adoption [5], making it essential to understand the predictive abilities of food guarding tests. Of the dogs assessed as food guarding in shelters, most (83%) exhibit behaviors consistent with mild to moderate guarding (e.g., stiffening and growling); the remaining dogs (17%) exhibit behaviors classified as severe guarding (e.g., lunging, snapping, and biting; [15,16]). In one study, dogs that showed severe food guarding during the shelter behavioral evaluation were more likely to be returned than those that showed either mild to moderate guarding or no guarding behavior, and the latter two groups did not differ from one another in the likelihood of return [16]. Another study found slightly lower return rates for dogs that displayed food guarding during the behavioral evaluation (5%; not differentiated by level of severity) when compared with dogs that did not display food guarding (9%; [5]). However, food guarding dogs in the study by Mohan-Gibbons et al. [5] were placed on a free-feeding program in the shelter as well as on a specific feeding program in the adoptive home (although adopters did not always follow the program). Regarding length of stay, dogs that showed food guarding at the time of evaluation stayed at one shelter four days longer, on average, than dogs in the general shelter population [15]. Behavior around food, however, represents only one of the many situations and behavioral tendencies typically examined in a shelter evaluation (e.g., fearfulness, sociability, arousal during play, and responses to an unfamiliar person, unfamiliar dog, and handling). The scarcity of information concerning the relationship between behaviors displayed during shelter evaluations and length of stay is surprising given that other factors potentially influencing length of stay have been extensively studied, including canine demographic and phenotypic characteristics [17,18,19,20,21,22,23], as well as in-kennel behavior [17,21,24,25] and whether a dog is housed at the shelter or in a foster home [26].We reviewed nearly five years of data from a New York SPCA to determine whether behavior displayed during separate tests and subtests of the shelter behavioral evaluation, together with the demographic factors sex and age, predicted length of stay for dogs. Based on previous research at this shelter regarding whether food guarding predicted the likelihood of return to the shelter [16], we predicted that dogs evaluated as showing dangerous behavior on specific tests and subtests would have longer lengths of stay than dogs evaluated as showing either concerning behavior or no concerning behavior. We expected dogs showing either concerning behavior or no concerning behavior to have similar lengths of stay. Understanding the relationship between length of stay and behavior during specific tests and subtests of canine behavioral evaluations could inform shelter management of dog populations. A second goal was to determine the prevalence of concerning or dangerous behaviors during specific tests and subtests of the behavioral evaluation because few measures of prevalence for these assessments have been reported in the literature (exceptions include [8,9]).2. Materials and Methods2.1. Study Shelter and RecordsWe examined canine records from the Tompkins County SPCA in Ithaca, NY, USA, which is a no-kill, open-admission shelter with scheduled intake. The shelter has several programs to promote dog adoptability including a small set of foster homes, playgroups for suitable pairs of dogs, and volunteer dog walking, in-kennel socialization, and taking dogs on day trips or overnight stays. Adoptions are promoted in local print and social media, at off-site events, and by a volunteer group independently advertising hard-to-place dogs.We analyzed data from 1 September 2014 through to 31 May 2019 entered by shelter staff into the PetPoint data management system (Supplementary Material). We extracted demographic data on dogs (sex, age, and body mass), as well as information on behavioral evaluations, adoptions, and lengths of stay (date adoption paperwork signed minus intake date, in days; for dogs returned one or more times to the shelter, we used their first length of stay). We excluded records from the following groups: all puppies (behavioral evaluations of puppies differed from those of older dogs and puppy results were not entered into the PetPoint database); five dogs with serious medical conditions; and dogs returned to owners, transferred to rescue organizations, or euthanized for either medical or behavioral reasons. We also excluded records from 34 dogs released for adoption and kept in foster homes during the study period because length of stay differed between dogs housed at the shelter and in foster homes (mean ± SD: shelter, 19.7 ± 19.7 days; foster home, 48.4 ± 47.5 days; t = 3.52, d.f. = 33.40, p < 0.01). Our final sample consisted of 975 dogs released for adoption following behavioral evaluation and housed at the shelter until adoption (note that this sample essentially reflects that of McGuire [16], but without the dogs in foster homes or with serious health issues).2.2. Dogs, Care, and HousingWe provide a brief description of housing and care of dogs because details have been presented by McGuire [27]. Upon admission to the shelter, dogs were housed in chain link cages in the Rescue building. Each cage had an indoor space (2.2 m2) and an outdoor run (3.5 m2). Veterinary staff examined dogs on the day of intake and performed routine procedures such as vaccinations, flea control, fecal exam, deworming, and a heartworm test. Each dog was scheduled for behavioral evaluation (Section 2.3) and after completing the behavioral evaluation, dogs were moved within a few days to the Pet Adoption Center, adjacent to the Rescue building. Thirteen cubicles on the adoption floor ranged in size from 5.2 to 7.3 m2 and almost all dogs were housed individually; only dogs surrendered from the same household and assessed by staff as needing to stay together shared the same cubicle. Each cubicle contained a water bowl, raised bed, blanket, and toys. Staff fed dogs each day between 08:00 and 09:00 h and between 15:00 and 16:00 h. Dogs were exercised several times a day when volunteers or staff either walked them or brought them to a large outdoor enclosure. Intact dogs were spayed or neutered before adoption.Demographic data on the 975 dogs are summarized in Table 1. Most dogs at the Tompkins shelter were mixed breeds; due to a lack of pedigrees and DNA analyses, the number of purebred dogs in the shelter population during the study period was unknown.2.3. Behavioral EvaluationsApproximately 3 days after intake, shelter staff evaluated each dog’s behavior using a series of tests based on Sternberg’s Assess-a-Pet [28], with modifications described by Bollen and Horowitz [1]. Present at each test was an evaluator from the shelter’s Behavior Program and a scribe. Over the nearly 5-year study period there were four evaluators (all female); beginning in June of 2015 and running through to the end of our study, evaluations were almost always conducted by one of the four evaluators. Behavioral evaluations included nine tests in the following sequence, with the Cage presentation subtests conducted while the dog was in its kennel in the Rescue building and all other tests and subtests conducted in a conference room in the Pet Adoption Center with the dog leashed.Cage presentation (two subtests): confrontational, evaluator faces the dog, bends at the waist, and makes direct eye contact; friendly, evaluator faces sideways, bends down, and talks to the dog in friendly manner.Sociability (four subtests): evaluator stands and ignores the dog for 60 s; strokes the dog three times; sits and ignores the dog for 5 s; sits and talks to the dog for 20 s.Teeth exam: evaluator makes five attempts to lift the dog’s upper lip and hold for 5 s.Handling (eight subtests): evaluator strokes the dog’s far side; lifts hind foot; runs hand down tail and tugs slightly; checks ears; presses on shoulders; leads with collar; wipes with towel; hugs.Arousal: evaluator initiates play for 30 s using balls, rope toys, etc., and then stops.Food bowl: evaluator gives the dog a mix of kibble and canned food in a bowl, and using the Assess-a-Hand, strokes the dog’s back and attempts to pull the bowl away.Possession: evaluator gives the dog a valued possession such as a raw hide chew or pig’s ear, and using the Assess-a-Hand, attempts to take the item away.Stranger: an unfamiliar person knocks on the door to the conference room, enters when prompted by the evaluator, makes eye contact with the dog, steps forward and reaches toward the dog, then squats and talks to the dog in a friendly manner.Dog-to-dog (two subtests): with the evaluator holding the leash of the dog being tested, the scribe brings from the adoption floor a leashed dog that did not show aggression to other dogs during its own evaluation and reactions of the test dog are recorded first upon seeing and then upon meeting the previously tested dog.Most dogs were tested with one dog, although sometimes a test dog’s reaction prompted testing with a second previously tested dog, and early in the study period, dogs with a history of dog aggression were tested with a fake dog. Use of a fake dog was largely discontinued after 2017 due to questions about the predictive value of a dog’s reaction to the fake dog compared to real dogs.Dogs evaluated as showing concerning or dangerous responses were placed on behavior modification plans and handling plans; housing strategies were also employed (e.g., dogs evaluated as reactive to other dogs were housed in cubicles on the adoption floor that minimized exposure to other dogs). Dogs on behavior plans were not retested.2.4. Scoring Methods and Statistical AnalysesEvaluators attempted to conduct all nine parts of the evaluation to their full extent, but exceptions were made for safety reasons for two tests requiring close contact. If a dog snapped at the evaluator during one of the early attempts of the Teeth exam, then subsequent attempts to check the teeth were skipped. Similarly, if a dog reacted poorly to one of the handling subtests, such as being led by the collar, then the evaluator might skip the hug subtest and proceed to other components of the evaluation. To address the reality that the Teeth exam was sometimes cut short and certain subtests of the Handling test may have been skipped, we scored whether a dog showed concerning or dangerous behavior during at least one attempt of the Teeth exam and during at least one subtest of the Handling test.Table A1 lists by test the behaviors considered concerning or dangerous by evaluators. For most tests in which a sufficient number of dogs showed concerning or dangerous behavior, we considered the level of behavior in our analyses using the following categories: no concerning behavior; concerning behavior; and dangerous behavior (Teeth exam; Handling; Stranger; Dog-to-dog meeting; an exception was made for the Food bowl test and Possession test, see below). Dangerous behaviors were not listed in the shelter scoring system for the four subtests of the Sociability test and the seeing subtest of the Dog-to-dog test because none of these subtests created situations in which dangerous behaviors were displayed. Thus, for Sociability subtests and the Dog-to-dog seeing subtest, we present data for concerning behaviors only. Finally, for the Cage presentation confrontational and friendly subtests, we present data for dangerous behaviors only, due to differences between the options available on the shelter behavioral evaluation form for concerning behaviors and entry options in PetPoint.To be consistent with previously published research, we categorized data from the Food bowl test and Possession test somewhat differently than the above descriptions for other tests. First, studies of resource guarding in shelter dogs typically combine results from these two tests [6,7,16], so we combined them in our analyses as well. Second, previous studies classified the level of guarding based on behaviors shown during either the food bowl test, possession test, or both tests as follows: dogs that stiffened, exhibited whale eye, snarled, froze, or growled were classified as showing mild to moderate guarding, and dogs that lunged, snapped, or bit the Assess-a-Hand were classified as showing severe guarding [15,16]. Note that this system of scoring differs somewhat from the shelter scoring system in that only the last three behaviors considered dangerous at the shelter are classified as severe guarding (lunged, snapped, bit the Assess-a-Hand; Table A1), and the behaviors froze and growled, classified by the shelter as dangerous, are classified as mild to moderate guarding, along with stiffened, exhibited whale eye, and snarled.For each test on the behavioral evaluation, we first assessed the prevalence of dogs in the study population showing concerning or dangerous behavior (number of dogs showing concerning or dangerous behavior/number of dogs tested and released for adoption). For tests with a sufficient number of dogs showing concerning or dangerous behavior, we used least squares models to determine significant predictors of length of stay at the shelter. Fixed factors in the models for length of stay were sex, age class, and behavior during the specific test or subtest (e.g., no concerning behavior, concerning behavior, or dangerous behavior). For all models, we examined the main effects and two-way interactions. From the final models, we dropped two-way interactions that were not significant at the p < 0.05 level; none of our two-way interactions were significant, thus all were dropped. We used Tukey’s HSD to correct for multiple comparisons within models. To control for multiple testing across models and decrease the likelihood of false positives, we set the p value threshold at p ≤ 0.01. Statistical analyses were completed in JMP Pro (version 15.0.0). For subtests or tests with a very low prevalence of concerning or dangerous behaviors (Cage presentation friendly subtest and Arousal test), we provide descriptive statistics only because models were unstable.3. Results3.1. PrevalencePrevalence of dogs evaluated as showing concerning or dangerous behavior on each test or subtest of the behavioral evaluation is shown in Table 2. Of all the tests and subtests, we found the highest prevalence for showing concerning behavior during the stand and ignore portion of the Sociability test (34.7%; only concerning behaviors scored, no dangerous behaviors), meaning that it was fairly common for dogs to make very brief nonsocial contact with the evaluator or to completely ignore the evaluator during this subtest. The prevalence of dogs evaluated as showing dangerous behavior during tests ranged from 0.0% (Arousal test) to 4.1% (Cage presentation confrontational subtest; Table 2). For the Food bowl and Possession tests, 12.0% of dogs showed mild to moderate guarding and 2.6% of dogs showed severe guarding (these results are reported here rather than in Table 2 because of slight differences in how we categorized behaviors).3.2. Length of StayFor ease of presentation, we first provide descriptive statistics for length of stay at the shelter in relation to demographic variables (Table 3) and then in relation to the behavior displayed during specific tests and subtests of the behavioral evaluation (Table 4, tests for which the level of behavior could be analyzed using the categories: no concerning behavior, concerning behavior, or dangerous behavior; Table 5, subtests for which only concerning behaviors were scored). Because we used a slightly different scoring system for food guarding, we describe these results in the text rather than including them in Table 4. The specific results of statistical models incorporating both demographic variables and behavior during the test or subtest are summarized in Table 6.Sex did not predict length of stay of dogs at the shelter (Table 3 and Table 6). In contrast, age class did predict length of shelter stay: seniors stayed longer than adults, which in turn stayed longer than juveniles (Table 3 and Table 6).animals-11-03272-t003_Table 3Table 3Length of stay (mean ± SD, in days) for shelter dogs placed up for adoption in relation to sex and age class, with sample sizes in parentheses. Within specific variables, values with different superscript letters are significantly different.Demographic VariablesLength of Stay (Days)Sex Female18.7 ± 20.6 (479)Male20.6 ± 18.9 (496)Age class Juvenile14.8 ± 11.6 (183) aAdult19.8 ± 21.1 (674) bSenior26.4 ± 19.9 (118) cLevel of behavior displayed by dogs on some tests/subtests predicted length of stay at the shelter. For the Dog-to-dog meeting subtest, dogs evaluated as showing dangerous behavior had longer lengths of stay than dogs evaluated as showing either concerning behavior or no concerning behavior; the latter two groups did not differ from one another in length of stay (Table 4 and Table 6). Although results for the Teeth exam, Handling test, and Stranger test showed a similar pattern, the effect of level of behavior on length of stay failed to reach statistical significance at the p ≤ 0.01 level (Table 4 and Table 6). Level of food guarding behavior displayed during either the Food bowl test, Possession test, or both tests predicted length of stay at the shelter (Table 6). Dogs evaluated as showing severe guarding had longer lengths of stay (mean ± SD, in days, number of dogs; 34.2 ± 20.2, 25) than dogs evaluated as showing either mild to moderate guarding (20.6 ± 13.8, 117) or no guarding behavior (19.1 ± 20.3, 833); the latter two groups did not differ from one another in length of stay.animals-11-03272-t004_Table 4Table 4Length of stay (mean ± SD, in days) for dogs in relation to behavior shown during tests or subtests on the behavioral evaluation at the shelter. Number of dogs shown in parentheses. Within rows, values with different superscript letters are significantly different.TestNo Concerning BehaviorConcerning BehaviorDangerous BehaviorTeeth exam19.4 ± 19.9 (940)21.8 ± 11.1 (16)31.1 ± 15.4 (19)Handling19.2 ± 20.1 (836)21.5 ± 17.3 (113)26.9 ± 16.4 (26)Stranger19.6 ± 19.6 (919)16.4 ± 13.5 (31)29.0 ± 27.2 (25)Dog-to-dog: meeting19.0 ± 19.3 (829) a22.2 ± 17.8 (130) a36.7 ± 41.3 (16) bFor the Sociability test in which only concerning behaviors were scored by shelter staff, behavior during the four subtests did not predict length of stay at the shelter: length of stay did not differ between dogs that showed concerning behavior and dogs that did not (Table 5 and Table 6). In contrast, there was a trend (p = 0.012) for behavior during the seeing component of the Dog-to-dog test to predict length of stay at the shelter. The post-hoc comparison revealed that dogs evaluated as showing concerning behavior had longer lengths of stay than dogs evaluated as showing no concerning behavior (Table 5 and Table 6).animals-11-03272-t005_Table 5Table 5Length of stay (mean ± SD, in days) for dogs in relation to behavior shown during the four subtests of the Sociability test and the seeing subtest of the Dog-to-dog test. Number of dogs shown in parentheses. Within rows, values with different superscript letters are significantly different 1.TestNo Concerning BehaviorConcerning BehaviorSociability: stand and ignore19.0 ± 20.8 (637)21.0 ± 17.6 (338)Sociability: three strokes19.6 ± 20.0 (910)20.3 ± 16.1 (65)Sociability: sit and ignore19.1 ± 19.9 (802)22.3 ± 18.9 (173)Sociability: pet and talk19.7 ± 19.8 (948)21.0 ± 19.8 (27)Dog-to-dog: seeing19.4 ± 19.0 (942) a28.4 ± 34.6 (33) b1 Only concerning behaviors were scored because these subtests do not create situations in which dangerous behaviors occur.Behavior during the Cage presentation confrontational subtest (only dangerous behaviors scored) did not predict length of stay at the shelter: dogs that showed dangerous behavior did not differ in length of stay from those that did not show dangerous behavior (mean ± SD, in days, number of dogs; dangerous behavior, 20.2 ± 15.3, 40; no dangerous behavior, 19.7 ± 19.9, 935; Table 6). We present descriptive statistics only for length of stay for the Cage presentation friendly subtest (only dangerous behaviors scored) and the Arousal test because models were unstable due to the small number of dogs showing dangerous or concerning behavior during these two tests (Cage presentation friendly subtest: dangerous behavior, 16.3 ± 10.3, 16; no dangerous behavior, 19.8 ± 19.9, 959; Arousal test: concerning behavior, 24.7 ± 14.8, 16; no concerning behavior, 19.6 ± 19.8, 959; even though dangerous behaviors were possible during the Arousal test, none of the dogs tested displayed them).animals-11-03272-t006_Table 6Table 6Results of least squares models to determine predictors of length of stay of dogs at the shelter. Fixed factors are sex, age class, and behavior displayed during the specific test or subtest. Threshold for significance set at p ≤ 0.01 1.TestSource d.f. F p Cage presentation: confrontationalSex12.120.15Age class212.650.0001Behavior during subtest 210.030.87Sociability: stand and ignoreSex12.250.14Age class211.910.0001Behavior during subtest 310.960.33Sociability: three strokesSex12.160.15Age class212.640.0001Behavior during subtest 310.100.76Sociability: sit and ignoreSex12.380.13Age class211.840.0001Behavior during subtest 312.330.13Sociability: pet and talkSex12.120.15Age class212.630.0001Behavior during subtest 310.070.79Teeth examSex12.350.13Age class212.710.0001Behavior during test 423.550.03HandlingSex11.730.19Age class213.240.0001Behavior during test 422.820.06Food bowl and PossessionSex11.850.17Age class211.520.0001Behavior during tests 425.980.003StrangerSex12.360.13Age class212.670.0001Behavior during test 423.370.04Dog-to-dog: seeingSex11.970.16Age class212.560.0001Behavior during subtest 316.380.012Dog-to-dog: meetingSex12.860.09Age class211.740.0001Behavior during subtest 427.230.0011 Results are not shown for the Cage presentation friendly subtest and the Arousal test because few dogs showed dangerous or concerning behaviors during testing and least squares models were unstable. 2 Only dangerous behaviors were analyzed for the Cage presentation confrontational subtest because options available on the shelter behavioral evaluation form for concerning behaviors differed from entry options in PetPoint. 3 Only concerning behaviors were analyzed for the four Sociability subtests and the seeing component of the Dog-to-dog test because these subtests do not create situations in which dangerous behaviors occur. 4 For three tests (Teeth exam, Handling, and Stranger) and one subtest (Dog-to-dog: meeting) we classified behavior by the following levels: no concerning behavior, concerning behavior, or dangerous behavior. For the Food Bowl and Possession tests, we used the following levels to be consistent with existing literature: no guarding behavior, mild to moderate guarding, or severe guarding.4. DiscussionA relatively low prevalence of concerning and especially dangerous behaviors characterized our study population of 975 dogs evaluated and released for adoption. The Sociability stand and ignore subtest had the highest prevalence at 34.7%, indicating that about one third of dogs tested and made available for adoption showed concerning behavior during this subtest, defined as ignoring or making only brief social contact with the evaluator. The prevalence of dangerous behaviors in our study population never exceeded 4.1% on any of the tests and subtests in which such behaviors might be displayed.It is challenging to compare the prevalence of behaviors during shelter behavioral evaluations across different studies because of variation in the tests and subtests conducted and scoring systems used. For example, whereas the evaluation at our study shelter classified behaviors displayed by dogs as dangerous, concerning, or not concerning, other studies classified behaviors into categories such as fearful, anxious, or aggressive [8,9]. A few direct comparisons are possible, with the caveat, that even tests with the same or similar names may be conducted and scored differently at different shelters. The prevalence of food guarding among dogs placed up for adoption at our study shelter (overall, 14.6%) is similar to values reported for other shelters (14% [5]; 20.6% [6]; 17% [15]). Clay et al. [8] reported 9.9% of dogs guarded a bone and 7.4% guarded a pig’s ear, but specific data on guarding a food bowl were not included. Despite testing for food guarding, van der Borg et al. [9] did not report the prevalence of this behavior. These authors did, however, report that 32.1% of dogs tested showed aggressive behavior (described as growling, baring teeth, snapping, biting, and piloerection) when meeting another dog. In contrast, we found the prevalence of concerning behavior (13.3%) and dangerous behavior (1.6%) during the Dog-to dog meeting subtest to be lower than that reported by van der Borg et al. [9]. This might reflect different testing conditions: whereas van der Borg et al. [9] matched dogs for sex and size, and described the dogs used for testing as dominant, these conditions did not apply to the Dog-to-dog subtests at our study shelter. Finally, the prevalence of concerning behavior (1.6%) and dangerous behavior (0.0%) during the Arousal test at our study shelter was also lower than the 12.3% of dogs reported by van der Borg et al. [9] to show either aggressive responses or play escalating into aggression during the Play with handler test.Level of behavior displayed by dogs on some tests or subtests predicted length of stay at the shelter. When using the levels, dangerous behavior, concerning behavior, and no concerning behavior, the dogs evaluated as showing dangerous behavior during the Dog-to-dog meeting subtest had longer lengths of stay (on average, about 15 days longer) than dogs that showed either concerning behavior or no concerning behavior. We found no difference in length of stay between dogs evaluated as showing concerning behavior and those evaluated as showing no concerning behavior during the Dog-to-dog meeting subtest. The level of food guarding behavior, assessed as severe, mild to moderate, or no guarding during either the Food bowl test, Possession test, or both tests, also predicted length of stay at the shelter. Dogs evaluated as showing severe guarding had longer lengths of stay (on average, about 14 days longer) than dogs evaluated as showing either mild to moderate guarding or no guarding; there was no difference in length of stay between dogs that showed either mild to moderate guarding or no guarding. These length of stay results for the Dog-to-dog meeting subtest and food guarding tests mirror those described by McGuire [16]. For likelihood of return of food guarding dogs to this shelter, dogs evaluated as showing severe guarding were more likely to be returned than those evaluated as showing either mild to moderate guarding or no guarding, and the latter two groups did not differ in likelihood of return. It is worth noting, however, that even though scores on food guarding tests predicted length of stay (present study) and likelihood of return [16] at our study shelter, food guarding during shelter testing did not consistently signal such guarding would occur in adoptive homes. Surveys of adopters revealed that more than half of the dogs evaluated as food guarding at our study shelter did not show guarding postadoption [7]; similar findings have been reported for dogs at other shelters [5,6].When levels of behavior were limited to concerning behavior or no concerning behavior for subtests that did not create situations in which dangerous behavior might be displayed, there was a trend (p < 0.012) for the level of behavior during the Dog-to-dog seeing subtest to predict length of stay. More specifically, dogs evaluated as showing concerning behavior remained at the shelter longer (on average, about nine days longer) than those evaluated as not showing concerning behavior. The remaining tests or subtests either did not predict length of stay (Cage presentation confrontational subtest, four Sociability subtests, Teeth exam, Handling test, and Stranger test) or too few dogs displayed concerning or dangerous behaviors to allow a formal statistical analysis (Cage presentation friendly subtest and Arousal test). Protopopova et al. [25] identified several in-kennel behaviors that predicted longer lengths of stay at the shelter; these included leaning on kennel walls, facing away from the front of the kennel, and standing. Given this connection between a dog’s in-kennel presentation and length of stay, we were surprised to find that the Cage presentation confrontational subtest, a subtest for which only dangerous behaviors were analyzed, did not predict length of stay at our study shelter. One possible explanation for our failure to find a relationship between behavior during this subtest and length of stay concerns the cage design. Cages in the Rescue building, where this test is conducted, are chain link, allowing the evaluator to bend and look directly into the eyes of the dog. In contrast, in the Pet Adoption Center, where potential adopters typically view dogs, the lower parts of cubicle doors and walls are covered with opaque material, making it impossible for visitors to bend and look directly into the eyes of dogs. Thus, the cage design in the Pet Adoption Center may reduce the occurrence of strong responses by dogs to visitors, and could explain our failure to find a relationship between behavior during the Cage presentation confrontational subtest and length of stay.The longer lengths of stay found for dogs evaluated as displaying dangerous or concerning behavior during specific tests and subtests of the behavioral evaluation may reflect a smaller pool of potential adopters for dogs with challenging behaviors. A reduction in the size of the pool of potential adopters could happen in several ways. It is possible that potential adopters might skip over dogs with signage on their cubicle indicating issues such as food guarding or intolerance of other dogs. Another possibility is that potential adopters decline to adopt based on their direct interactions with a dog when introduced by shelter staff or volunteers. Visitors to one animal shelter stopped to look at less than one third of available dogs in their kennels, and when they did stop, spent 70 s, on average, in front of a cage [29]. Removal of breed labels from kennel cards was associated with reduced lengths of stay for dogs at one shelter, but other changes during the study period, such as increased advertising and longer operating hours, likely also played a role [18]. Another study found that most potential adopters asked to interact with only one dog and interactions with the dog were typically of short duration, about 8 min [30]. Nevertheless, dogs that were adopted spent more time lying in proximity to the adopter and less time ignoring play initiation by the adopter than dogs that were not adopted [30].Based on these findings regarding visitor behavior at other shelters, we suggest that visitor response to signage or direct interactions with dogs probably play some role in limiting the pool of potential adopters for dogs with behavioral challenges. Our study shelter does not conduct policy-based adoptions, described in Weiss et al. [31] as screening potential adopters based on factors such as daily time spent away from home and veterinary care provided to pets already in the household. Instead, Tompkins County SPCA uses a conversation-based approach during which adoption counselors disclose all available information about the dog, including results from behavioral evaluations, to help potential adopters decide whether the dog would be a good fit for their household. Some potential adopters may decide that a dog with challenging behaviors would not be a good match for them, which might result in longer shelter stays for these dogs compared with dogs without such challenges. Tompkins County SPCA does require a meeting between any dog(s) living in the household of potential adopters and the dog they are considering bringing home. These dog meets might be another point at which potential adopters decide not to pursue a dog with behavioral challenges and could explain the relationships we found between the behavior shown during Dog-to-dog tests and length of stay. Based on the data available to us, however, we cannot assess the relative importance of these various factors—signage, direct interactions with dogs, meetings with adoption counselors, and dog meets—in the decision made by visitors to adopt or not adopt.For the two demographic variables examined, we found that age class predicted length of stay at the shelter, but sex did not. More specifically, seniors stayed longer than adults (on average, by about 7 days), which in turn, stayed longer than juveniles (on average, by about 5 days). Our finding of increasing length of stay with age agrees with results from several studies [19,22,23]. Protopopova et al. [17] found that age did not predict length of stay at their study shelter and Luescher and Medlock [32] found age did not predict the likelihood of adoption; these findings might reflect the absence from both studies of dogs more than 7 years of age. For comparison, senior dogs, defined as those at least 8 years of age, made up about 12% of our study population. No consistent pattern has emerged from studies examining the effects of sex on length of stay: some, like ours, report no effect of sex, whereas others find an effect, often with longer stays for males [17,19,21,22,23].Our study focused on a single shelter rather than multiple shelters in different regions, which may reduce the generalizability of our results to other shelters [20,33]. Other limitations of our study derive largely from the necessary safety measures taken by shelter staff conducting the behavioral evaluations. For example, we could not examine individual subtests of the Handling exam because reactions of a dog during one subtest may cause the evaluator to skip another subtest for safety reasons. As a result, we could not assess the prevalence of dogs showing concerning or dangerous behavior for each of the eight subtests, which ranged from touching the back foot to hugging the dog, nor could we determine whether individual subtests predicted length of stay. Instead, our more general analysis considered whether showing concerning or dangerous behavior on at least one subtest of the Handling exam affected length of stay. A similar situation existed for the Teeth exam where all five attempts to check the teeth might not have been completed for safety reasons. Another limitation concerns variation in how the Dog-to-dog seeing and meeting subtests were conducted, because early in the study period, dogs with a history of intolerance of other dogs were sometimes tested with a fake dog rather than a dog from the adoption floor. However, over the nearly five-year study period with results from 975 dogs tested and released for adoption, evaluators used a fake dog with only four dogs during the seeing subtest and two dogs during the meeting subtest, so we expect use of the fake dog had little effect on our findings for the two Dog-to-dog subtests.5. ConclusionsA low prevalence of concerning and especially dangerous behaviors during the multiple tests and subtests of the behavioral evaluation characterized our study population of 975 dogs evaluated and released for adoption. Nevertheless, for some tests and subtests —food guarding and meeting another dog—display of behavior evaluated as dangerous predicted longer lengths of stay at the shelter. Display of concerning behavior on one test that did not create a situation in which dangerous behavior would occur (seeing another dog at a distance) fell just short of being a significant predictor of longer lengths of stay at the shelter. We suspect the longer lengths of stay for dogs that displayed challenging behaviors at the time of evaluation reflect a smaller pool of potential adopters, perhaps resulting from one or more of the following: signage on dog cubicles, direct interactions between potential adopters and dogs, conversation-based discussions between shelter staff and potential adopters when all available information about a dog is disclosed, or dog meets. Understanding the relationships between performance on canine behavioral evaluations and length of stay may aid shelter management of dog populations and help highlight individual dogs requiring special adoption efforts to avoid long shelter stays.

animals : an open access journal from mdpi

10.3390/ani11030854

PMC8002757

Defining animal welfare definition is still controversial, and a definition is difficult to establish. Furthermore, welfare detection is often complicated and subject to different interpretations. This work aimed to provide valid indicators to evaluate the welfare of dairy cows. The Animal Welfare and Biosecurity Evaluation form (AWB-EF) checklist developed and validated by the Italian National Centre of Reference for Animal Welfare (CReNBA) was submitted to 16 Sardinian dairy cattle farms. Blood samples from the 230 Holstein breed dairy cattle housed in these farms were analyzed for hematological parameters. Correlation analysis revealed a strong association between AWB-EF (considered as gold standard) and laboratory parameters, indicating correspondence between the health and welfare status of the animals. Our study clearly indicates that the use of a validated checklist in combination with the identification of well-known laboratory parameters can be a fundamental tool for veterinarians to detect stress conditions early.

The need for animal welfare definition and assessment is increasing worldwide, and several studies have been conducted to help fill the knowledge gaps regarding the welfare of cattle. However, further studies are needed to provide valid synthetized measures for welfare evaluation. The aim of this study was to assess the welfare status of 16 Sardinian dairy cattle farms, based on the developed Animal Welfare and Biosecurity Evaluation checklist (AWB-EF) and the corresponding hematological, biochemical, and electrophoretic profiles of these animals. Considering the AWB-EF as gold standard, blood samples were collected from 230 Holstein breed dairy cattle, aged between 3 and 8 years, out of the periparturient period, and with no clinical signs of specific pathologies. Principal Component (PC) and correlation analyses were performed to simplify phenomena interpretation and assess positive/negative associations. Four PCs were able to explain 76% of the total variability, and six laboratory parameters were strongly associated with the AWB-EF score (Spearman’s correlation coefficient ≥ 0.40, p-Value < 0.05), reflecting the real health status of the animals. Given the complexity of animal welfare as a multidimensional concept and the need to include both animal-based and non-based measures in welfare evaluation, the present work represents a sound basis for future evaluation and veterinary health planning.

1. IntroductionThe original concept of animal welfare, developed over 50 years ago by Brambell (1965) [1], was updated by Lawrence and Stott [2], who defined animal welfare as an ethical concern for the mental and physical health of animals over which we have a degree of control, in 2010. However, defining and measuring animal welfare remains controversial [3]. As underlined by Devitt et al. in 2018, if, on the one hand, there is a need to take the relationship between farmers and animals into account when considering farm animal welfare standards, then on the other hand, there is limited understanding of how the nature of this relationship influences welfare outcomes, and thus welfare assessment would require a multidisciplinary approach [4].The need for animal welfare has been underlined by the World Organization for Animal Health (OIE), who recommended that “veterinarians should be the leading advocates for the welfare of all animals, recognizing the key contribution that animals make to human society through food production, companionship, biomedical research, and education” [5]. Increased attention has been given to farm animal welfare in developed countries, especially given the expansion of intensive animal production systems that improve profit and efficiency but challenge the conscience of many consumers [6].Furthermore, the need for measurable outputs relevant to animal health and welfare, and that are able to determine whether the welfare program is effective, efficient, and transparent, is increasing worldwide [7].Poor environmental conditions can affect several homeostatic functions and reduce the productive and reproductive performances of livestock. Stress factors and poor welfare can also compromise the host immune system and lead to increased susceptibility to diseases among animals [8]. Farm animal welfare should be viewed as a global condition, where the effects of infectious and non-infectious stressors cannot be easily discriminated and can overlap, challenging the host’s immune system [9]. In fact, the innate immune system can rapidly respond to both infectious and non-infectious stressors, such as metabolic stress conditions, psychological stress, high/low temperatures, oxidative stress, and hypoxia [8].To fill the knowledge gaps regarding cattle welfare, several studies have attempted to develop scientifically valid methods for assessing welfare [10,11,12,13,14,15,16,17,18,19,20,21,22]. Some of these studies primarily focused on improving external factors, management-based measures, or non-animal-based measures (N-ABMs), which affect welfare without considering the reactions and consequences for the animals [17,18]. Indeed, most of the methods used have not been refined or validated, possibly because they aimed to detect illness rather than welfare [19,20]. Thus, resource-based assessment cannot answer questions about animal welfare. For all these reasons, attempts have been made worldwide to develop animal-based measures (ABMs) to estimate the actual welfare of animals [21,22]. Conversely, the European Food Safety Association (EFSA) has provided evidence that in some cases N-ABMs may be more efficient than ABMs; thus, it follows that both ABMs and N-ABMs are necessary to obtain a holistic approach and achieve an effective overall classification of animal welfare at the farm level [23,24,25]. Currently, about two million dairy cows are reared in the Italian national territory, with more than half of these in the north of the country [26]. Despite this large number of animals, there is no official protocol for evaluating the welfare of dairy cows that uses ABMs, N-ABMs, and risk assessment. In 2017, the Italian National Centre of Reference for Animal Welfare (CReNBA) developed a simple and easy-to-use on-farm protocol for assessing the welfare of dairy cows in loose housing systems. They then carried out expert opinion elicitation to characterize a list of management and housing factors potentially associated with negative or positive welfare outcomes in dairy cows [25].Welfare assessment requires a multidisciplinary approach, and the CReNBA Animal Welfare and Biosecurity Evaluation form (AWB-EF) monitor both ABMs and N-ABMs in dairy cattle farms. Health is a key component of welfare and health status (e.g., presence/absence of disease, organ function, metabolic processes, and internal body condition) is primarily monitored using hematological and biochemical tests [27,28,29]. In this study, we aimed to assess whether the level of welfare estimated by the AWB-EF positively correlated with laboratory parameters (metabolic profile, electrophoresis, and blood count), which are regarded as indicators of health status. To achieve this goal, an in-depth investigation on livestock welfare and biosecurity status was conducted in a broad sample of Sardinian dairy cattle herds using the CReNBA AWB-EF as gold standard, and its correlation with laboratory parameters was assessed. Considering that there is no specific official indication about how laboratory parameters relate specifically to animal welfare (except for hemoglobin levels in calves, Italian Health Ministry D.L. 331, 1st September 1998) [30], this study could represent an important starting point for identifying simple and easily detectable parameters that can help monitor animal welfare.2. Materials and Methods 2.1. Study Context Sardinia is an Italian island in the center of the Mediterranean Sea (40°03′ N 9°05′ E) with a total surface area of 24,100 km2 and a population density of 69 inh/km2 [31]. Administratively, the Sardinian region comprises five provinces, as established by the Regional Law of 4 February 2016: Nuoro, Sassari, Oristano, South of Sardinia, and Cagliari (metropolitan city), with 377 municipalities in total [32]. Given the low population density and the unpolluted environment, the Sardinian economy is mostly based on agropastoral activities [33]. Across the whole island, around 9,200 dairy cow farms are regularly recorded in the Italian Veterinary National database, with a dairy cow population of about 260,000 animals. Sardinian cattle breeding is mainly intended for meat production (85%; 7700 farms). The main area dedicated to beef production is located in the north east of the region, (Sassari and Nuoro provinces), while most of the bovine livestock for milk production is located in the center of the island (Oristano province), where the territory mainly comprises large lowland areas [26]. 2.2. The Animal Welfare and Biosecurity Evaluation Form For the purpose of this work, the AWB-EF checklist was implemented by two trained veterinarians, during the farms’ official controls, in 16 dairy cow farms in Sardinia (Italy) between 2013–2015. The checklist was provided by the CReNBA, aimed to monitor the welfare conditions of dairy cows, and consisted of 90 items divided into 5 sections (A1, A2, A3, B, C) (Table S1). In detail, section (A) is divided into (A1) farm management (22 items, i.e., number of stockpeople and their training, animal grouping, inspection, type of handling, feeding strategy, water provision, cleanliness, bedding material and calving pen management), (A2) structures and equipment (29 items, i.e., features about space availability, calving pen, bedding material, floor, feeding space, water points, facilities for sick animals, milking machine, temperature, humidity, ventilation, gas concentration, and artificial lighting), (A3) animal-based measures (18 items, i.e., details about lameness, mortality rate, body condition score, treatments for mastitis, integument alterations, and mutilations; cleanliness of flank, leg, and udder).The other sections are (B) biosecurity (15 items, i.e., general measures against rodents, insect, precautions for entry of strangers, disinfection, quarantine, carcasses management, animal movement management, and disease prevention) and (C) risks and alarm system (5 items, i.e., noise, fire and ventilation alarm, and electricity generator). Each item contributed equally to generating a percentage score for each section, constituting five different outcomes. Furthermore, general information on the number of animal breeds, as well as the animals’ average age, breed, the number of lactating cows and milk production (kg of milk by animal per day), was collected by the AWB-EF checklist for each farm. Animals in the periparturient period were excluded to avoid confounding bias in laboratory parameters detection [34]. The AWB-EF was provided in paper format, and all data were subsequently collected and stored in an ad hoc password-protected electronic database using a closed response data collection instrument (Microsoft Excel, Microsoft Corporation, Redmond, WA, USA).2.3. Laboratory AnalysisLaboratory parameters that were able to verify whether the farm conformed to Italian (D.Lgs.n.146/2001; D.Lgs.n.126/2011) and European legislation (Directive 98/58/EC; Directive 2008/119/EC) were chosen based on EFSA reports [23,24,35,36,37,38,39,40,41], to evaluate whether the AWB score corresponded with the cows’ welfare conditions on the basis of the five outcomes. During the veterinarian’s official visits, blood sampling was performed on Holstein cows between 3 and 8 years old at different stages of lactation. Individual blood samples were collected from the animals’ coccygeal vein before feeding. They were transferred into vacuum tubes containing EDTA anticoagulant for hematological profiling and then into serum gel separator tubes without anticoagulant for both biochemical profiling and analysis of the electrophoretic pattern of serum proteins. The samples were placed in a container with ice and forwarded to the laboratory within 2 h. The serum was separated by laboratory centrifugation at 3500× g for 10 min at 4 °C, placed in 1.5-mL tubes, and stored at −20 °C until analysis, while the EDTA tubes were analyzed within 3 h. Hematological determinations were made using an ADVIA 2120 automatic hematology analyzer (Bayer Healthcare, Siemens, Monaco, Germany) with software that allows blood determination in cattle. Biochemical parameters were analyzed using a Dimension RXL chemistry analyzer (Siemens), and serum protein electrophoresis was carried out using an INTERLAB G26 Automated Agarose Gel Electrophoresis Analyzer (Interlab, Rome, Italy). The set of parameters used for biochemical profiling are provided in Table S2. 2.4. Sample Size and Inclusion/Exclusion CriteriaDairy cow livestock breeding (not mixed species farming system) business’s start and end dates were available, and business activities for the entire study period (i.e., business start date before on or before 1 January 2013 and business end date after on or after 31 December 2015) and the available animal census data were essential farm characteristics to be included in this study. If the farmer had more than one farm with a unique fiscal code, or the farms were inactive for the entire study period, or the farm had missing business start data and/or end data or animal census data for one or more of the study years, the farm was excluded. The minimum number of animals to be observed is defined by the Welfare and Biosecurity Manual for dairy cows, performed by CReNBA and specific for the AWB-EF checklist (Table S2). Otherwise, given that programmed statistical analysis (factorial analysis) is strongly influenced by the ratio of sample size (N) to the number of variables being analyzed (p) [42], a specific sample size calculation was carried out before proceedings, following MacCallum et al., 1999 [43]. Given the large number of variables collected (31) and the number of hypothesized factors (5), a total sample size of 200 animals (plus 15% of these animals to account for drop out) was considered appropriate. A total of 16 randomly selected dairy cow farms were included in the study. Animals raised on the 16 farms were included in the study, except for animals in the periparturient period and those outside of the age range (3–8 years). 230 samples from 230 randomly selected cows were collected and analyzed.For each farm, the following baseline characteristics were collected: location (latitude and longitude), province, municipality, opening date, number of animals on farm, age, race, number of lactating cows, milk production, details of previous disease (i.e., skin lesions, lameness, mastitis), mortality rate (overall mortality year rate), and number of animals with laboratory parameters within the normal range. 2.5. Statistical AnalysisData quality and completeness were tested based on an extensive data check. Descriptive analyses based on mean (SD), median (I–III quartile) and frequency (percentages) were performed to evaluate the distribution of the parameters at baseline compared with the laboratory reference values. A correlation matrix was performed to evaluate association between variables. The linear or nonlinear nature of the relationship between the dependent variable (i.e., percentage score for each section) and each of the continuous independent variables was assessed graphically. When a linear relationship was assumed, its strength was initially evaluated through bivariate analyses by means of the Spearman non-parametric correlation coefficient. When relationships were assumed to be other than linear, mathematical transformations (i.e., log-normal) were applied. The homogeneity of the samples between farms was graphically evaluated to exclude bias generated by the within group correlation or sub-group populations.The main hypothesis to be tested by the statistical analyses was the correlation between the percentage score for each section of the AWB-EF checklist with the laboratory results (electrophoresis, metabolic profile, and blood count; Table S3). A principal component analysis (PCA) with a varimax rotation was used to make the factors orthogonal and more interpretable, and to extract the laboratory patterns and confirm the number of dimensions underlying the set of variables [44]. The number of retained components was determined according to eigenvalues (≥2.0), scree plot examination and interpretability [45].Given the non-normal distribution of all the variables included, to evaluate the strength of the relationship (association) between the AWB-EF scores of each checklist section and the laboratory parameters, the nonparametric Spearman’s correlation coefficient was estimated. A correlation matrix and correlation graphs were produced. Any outlying observation that appeared to deviate markedly from other observations in the sample was checked. Only 8 values significantly deviated from the mean, and 3 of these were related to reporting errors (and consequently corrected). The other five were included, considering that, in such cases, outliers may be due to random variation or may indicate something scientifically interesting [46]. A high statistically significant correlation (coeff. ≥ 0.7, p-Value < 0.05) means that two variables have a strong relationship, while a weak correlation means that the variables are hardly related. All statistical tests were two-sided, and a p-Value less than 0.05 was considered statistically significant. Statistical analyses were performed using Stata13 (StataCorp, Stata statistical software, Release 13; StataCorp LP, College Station, TX, USA) and R-software (Version 3.6.2; R-Foundation for Statistical Computing, Vienna, Austria).3. Results3.1. Descriptive and Principal Component Analysis ResultsThe baseline characteristics of each farm and thirty-four laboratory parameters were collected from 230 dairy cows raised on 16 extensive farms. The farm baseline characteristics are reported in Table 1, while the laboratory parameters and comparisons with the reference ranges are reported in Table 2. Farms included in this study were well-distributed around the Sardinian island. Totally, 5110 Holstein cows were bred on these 16 farms, with an average of 216 animals (SD = 131) and a median age of 4 years (I–III quartile = 3–9), as established by the inclusion criteria. On each farm, an average of 109 (SD = 63) animals were in the lactating period, with an associated average milk production of 31 (SD = 2.44) kilograms of milk per cow per day. About 15% of the animals on each farm reported skin lesions, 6% lameness, and 30% has undergone treatment for mastitis. The overall mortality rate was about the 2% of the overall population. In total, 217 out of 230 cows (94%) showed laboratory parameters within the normal range.The five AWB-EF sections were evaluated and summarized in Figure 1. The average value of section A1 (farm management) was 76.9% (SD = 8.8), while section A2 (structures and equipment) showed an average percentage of 64.5% (SD = 12.3) and section A3 (animal-based measures) had a lower average value of 62.9% (SD = 10.1). Finally, sections B (biosecurity) and C (risks and alarm system) of the AWB-EF checklist showed lower averages, with values of 52.3% (SD = 11.3) and 54% (SD = 11.3), respectively. Based on PCA analysis, according to the eigenvalues and the scree plot (Figure 2), four components are able to explain why 76% of the total variance was maintained.The first component (fact1) was able to explain 43% of the total variance and was positively characterized by alkaline phosphatase (ALP), glutamic-pyruvic transaminase (GPT), white blood cells (WBC), red blood cells (RBC), hemoglobin (HGB), hematocrit (HCT) and the total number of lymphocytes (LYMPH) and basophils (BAS). The second component (fact2) explained 18% of the total variance and was positively characterized by α1-globulin (A1GB), α2-globulin (A2GB), β-globulin (BGB), γ-globulin (GGB), WBC count, platelets (PLT), average platelet volume (MPV), and the total number of monocytes (MONO) and basophils. The third PCA component (fact3) was positively characterized by WBC, RBC and HGB count, average corpuscular volume (MCV), average hemoglobin content (MCH), medium corpuscular hemoglobin concentration (MCHC), MPV, and the total number of neutrophils (NEUT) and eosinophils (EOS), explaining 9% of the total variance. The last component (fact4) explains 6% of the total variance and was positively characterized by albumin (ALB), BGB, gamma-glutamyl transpeptidase (GGT), GPT, HGB, HCT, and MCV.Considering the main parameters that characterized each PCA factor, the four dimensions highlighted by the PCA were named as “Overall Welfare”, “Electrophoresis”, “Blood count” and “Metabolic profile”, respectively, rather than dimension 1, 2, 3 and 4 as conventionally used [44].3.2. Results of Correlation AnalysisFive correlation matrices were computed to evaluate the association between A1, A2, A3, B, and C scores with the laboratory parameters of Pearson’s correlation coefficient and correspondent p-Values (Table 3).The correlation matrix was interpreted by considering the positive relationships between variables as ‘very strong’ when the Spearman’s correlation coefficient was higher than 0.70, ‘strong’ when it was between 0.40–0.69, ‘moderate’ when it was between 0.30–0.39, ‘weak’ when it was between 0.20–0.29, ‘negligible’ when it was < 0.20, and likewise for the negative relationships between negative correlation values [47]. In particular, 16 parameters were statistically correlated with the score A1–management factor, and the Spearman’s correlation coefficient of 0.54 and −0.48 indicates a strong positive relationship with amplitude of hemoglobin distribution (HDW) and strong negative relationship with GPT (Table 3, Score A1), respectively. All 12 variables statistically correlated with the score A2–housing factors and showed a weak or negligible positive or negative coefficient, except for GPT and LYMPH, which showed a negative and positive moderate correlation, respectively (Table 3, Score A2), as well as neutrophil-lymphocyte ratio (NRL). A strong negative correlation was described by a 0.52 coefficient between the score A3–animal-based measures and A1GB, while the A2GB was moderately positively correlated with the score (coefficient = 0.37) (Table 3, Score A3), and the other five variables showed moderate (BGB, MCHC, BAS) or weak correlation (GGB, MPV). Between the 10 variables statistically correlated with score B–biosecurity, a very strong negative relationship was identified with A1GB (Table 3, Score B), a very strong positive relationship with A2GB (Table 3, Score B), and a strong negative relationship with GPT, with Spearman’s correlation coefficients of −0.71, 0.72 and −0.51, respectively. The score C–risk and alarm system were statistically significantly correlated with seven variables, but only blood urea nitrogen (BUN) and HDW showed strong positive relationships (coefficients = 0.66 and 0.40, respectively) (Table 3, Score C). 4. DiscussionThe present study evaluated the welfare status of 230 dairy cows on 16 dairy cattle farms in Sardinia, testing the association between the welfare score detected by the AWB-EF checklist and the individual hematological and biochemical parameters observed. Considering the welfare protocol score as gold standard, and by correlating it with laboratory results, we were able to estimate whether the data recorded by the protocol reflected the health status of the animals. The five different AWB-EF sections reflecting the laboratory patterns individuated with PCA analysis, and each score showed strong and statistically significant association with one or more of the laboratory parameters. Considering the PCA results, welfare status in farm management could be well described by the electrophoresis, blood count and metabolic profile. Besides, the correlation matrix results described a strong association between AWB-EF checklist scores and laboratory parameters such as A1GB, A2GB, GPT, HDW, LYMPH and BUN. Serum protein levels, including α1-globulin and α2-globulin, are correlated with important functions of organic synthesis (mainly in the liver), and their levels depict the animals’ capacity to cope with growth and production demands [48]. Electrophoresis has been used to monitor the ability of young bulls to adapt to a different environment, and researchers observed statistically significant variations in levels of either albumin, α-globulin, β-globulin or γ-globulin several times after arrival at the new farming center [49]. Another study described abnormality in electrophoretic parameters in calves after transportation stress, and particularly α-globulin peak was the most predictive of diseases in Holstein Friesian cattle among the clinical immunological parameters evaluated [50]. In the same study, hematological parameters were monitored in animals after transportation stress, and researchers observed that WBC was the most predictive of disease among the hematological parameters evaluated. Nevertheless, the differences between leucocyte subsets were not monitored in that study [50]. Overall, WBC and the variation of different leukocyte populations can be monitored to evaluate the sanitary status of the herds [51]. In our study, a strong association of AWB-EF was also observed with two biochemical parameters: GPT and BUN. Evaluation of the serum activities of hepatic enzymes, including GPT, is routinely used to monitor liver health status, whereas BUN levels are inversely correlated with the decline of kidney function [51,52,53,54]. Correct functionality of these two vital organs is strongly correlated with animal welfare. The health status of animals is one criterion of welfare assessment, and it is important to monitor management and housing factors, which are strongly related to a lower incidence of disease and mortality, mental comfort, absence of stress, good appetite, body homeostasis, and maintenance of proper animal welfare levels [55]. Assessing welfare requires detailed knowledge of the available scientific information, since the definition involves describing how well the animals experience their environment based on the best possible judgement of their situation. Such information is necessary to avoid errors in interpreting a given measure and cannot be based solely on science or on data collected from experiments or laboratory analysis [56]. Proper management conditions are essential for organisms to function normally [57], and hematological and biochemical tests may help veterinarians to understand animals’ welfare status based on factors other than the presence/absence of disease [27,58,59].Several studies used hematological and biochemical parameters to assess animal health and welfare status, both in livestock and small animals [16,22,25,48,52,60,61,62,63,64]. Nevertheless, the use of laboratory parameters requires the collection of samples on the farm and, depending on the parameters used, can entail quite high costs in the purchase of reagents and require time for sample collection and analysis. Thus, welfare protocols are a more immediate and less expensive method for judging the living conditions of animals. Few reports have focused on welfare scores in cows, and more in-depth studies are needed to detect the causes of possible deviations from normal reference ranges. Animal-based measures alone are not enough to ensure a complete evaluation of animal welfare [41]. As a result, applying welfare protocols requires a more general farm evaluation carried out by well-trained veterinarians, which could be difficult to organize. The combination of both welfare protocol and laboratory parameters could be a valid tool to assess a robust estimation of both animal health and welfare status.In this study, the score assigned by the welfare protocol mainly corresponded with the animals’ health status, and the analysis showed that some laboratory patterns may be particularly useful indicators of welfare. Otherwise, the results of correlation analysis must be carefully considered, as their extrapolation could be dangerous. Correlation analysis only considers the linear relationship between two variables (i.e., other variables that could influence the response variable are not studied) and could be affected by outliers. Furthermore, correlation analysis does not establish if one variable is dependent and the other is independent. Therefore, correlation analysis provides information about the strength and the direction (positive or negative) of a relationship between two continuous variables, but a strong correlation does not imply a cause-and-effect relationship. Furthermore, a limit of this study is that information (i.e., day in milk, amount of milk production, age, lactating period) is collected overall by farm and is not available for each animal. Thus, considering that the single animal was the epidemiological unit for the correlation analysis, this information was not included in the final analysis, generating a possible bias considering that these conditions could significantly influence blood and milk parameters [65].We can conclude that the average values of the hematological and biochemical parameters fell within the range of reference values. Mainly strong or moderate associations have been highlighted between the average welfare score of the dairy farms and laboratory analysis, suggesting that the welfare protocol score mainly reflected the real health status of dairy cattle. The use of a validated checklist in combination with the identification of few well-defined parameters, able to synthetize the health and welfare status of the animals, can be a fundamental tool for veterinarians to detect stress conditions early. Although more in-depth analysis is needed to provide not only an association between measures and the checklist but to quantify and model this association, the results obtained in this study are a strong starting point for future research. Finally, it is necessary to underline that health is a key component of welfare, and welfare assessment requires a multidisciplinary approach, so a laboratory evaluation of the health status of animals cannot be considered the sole criteria for determining animal welfare.

animals : an open access journal from mdpi

10.3390/ani11082377

PMC8388692

Neogobius melanostomus is a highly invasive fish that has colonized most major European rivers and is dispersing into their tributaries. Its foraging behaviour does not show particular prey preferences, which makes predicting its interactions with endangered members of the macrozoobenthic community in tributaries a challenge. We observed the interaction of N. melanostomus and crayfish juvenile or A. aquaticus in single- and multiple-prey systems to better predict its ecological impact. The results suggest an impact of N. melanostomus on crayfish similar to that on A. aquaticus, potentially making it a threat to crayfish population stability. Destabilization of a keystone species such as crayfish in river tributaries may lead to a trophic cascade in the ecosystem with irreversible consequences.

Despite the spread of round goby Neogobius melanostomus into freshwater streams, there is a lack of information with respect to its effect on macroinvertebrate communities, especially crustaceans. We studied foraging efficiency of N. melanostomus on Procambarus virginalis and Asellus aquaticus, using a functional response (FR) approach. Stocking density of the prey species was manipulated to determine its effect on consumer utilization, with prey offered separately or combined at 1:1, 3:1, and 1:3 at each tested density. For both prey species, N. melanostomus exhibited type II FR, occasionally with a high proportion of non-consumptive mortality. Procambarus virginalis suffered a significantly higher attack rate compared to A. aquaticus. Neogobius melanostomus killed significantly more of the most prevalent prey, regardless of species. In trials with prey species of equal proportions, a difference in the number of each species killed was observed only at the highest density, at which P. virginalis was preferred. Neogobius melanostomus may be an important driver of population dynamics of prey species in the wild. The non-selective prey consumption makes N. melanostomus a potential threat to macrozoobenthic communities of river tributaries.

1. IntroductionCrayfish have an impact at multiple trophic levels through predation, shedding, burrowing, and competition [1,2,3] and are considered keystone species influencing stability and functionality of ecosystems, particularly in tributaries to major streams [4,5,6]. Crayfish populations worldwide are threatened by multiple stressors: Climate change, water pollution, habitat modification, invasive species, and disease [5,7]. Nearly one third of crayfish species worldwide are threatened with extinction [7]. Although interventions in the EU [8] and throughout the world [9,10] aim to improve the ecological status of freshwater lotic ecosystems, the threat presented by non-indigenous species is ever-increasing [11]. In addition to interactions with non-indigenous crayfish, native crayfish interact with small benthic fishes, including non-native species [1].The round goby Neogobius melanostomus (Pallas 1814), among the most invasive of freshwater fish species [12], has expanded substantially beyond its native range the Ponto-Caspian region. It poses a serious threat to freshwater and brackish ecosystems [13] causing critical food web disruptions, shifts in trophic levels, extermination of native species through direct predation and/or competition for resources and habitat, and spread of disease [14,15,16,17]. In major rivers, after establishing a viable population, N. melanostomus spreads both down- and up-stream [18,19]. It is increasingly found in tributaries of major rivers [20,21,22] that are often used as refugia for native species [23] and contain unique highly diverse macrozoobenthic communities including endangered species such as crayfish [24]. These communities may be seriously threated by N. melanostomus invasion and dispersion [25,26].Macrozoobenthos represent a predominant proportion of the N. melanostomus diet [27,28], reflecting the community structure in a given locality [29,30]. In contrast to major rivers and lakes, which often harbour several non-native macrozoobenthos species, in small streams with highly diverse macrozoobenthic communities, N. melanostomus remains a generalist omnivore [30]. This can lead to a significant transformation of the community structure with severe consequences to endangered species, since even partial depletion of a single prey population can alter the predator food selectivity [31]. Nevertheless, crayfish are rarely reported in N. melanostomus diet in invaded regions [32,33], possibly the result of a unique flip-tail escape strategy, as observed for dragonfly nymph predation on early-stage crayfish [34].With respect to the coexistence of small benthic fish and crayfish, due to similar body size, the primary focus has been on competition for food and shelter and on behaviour interactions in the presence of a common predator, as opposed to their mutual predation relationship [1]. However, crayfish juveniles that have become independent after leaving the female are threatened by fish predation due to their small size [35,36] and limited antipredator defences, usually restricted to the tail-flip escape movement [35,36,37,38]. The impact of small voracious benthic fish such as N. melanostomus on early crayfish stages may be intensified when sharing a common habitat. The ecological impact of N. melanostomus on crayfish populations has not been quantified.Understanding and predicting novel predator-prey interaction dynamics and their consequences for invaded freshwater communities is a critical issue in invasion management [39]. Invasive predators, often possessing better foraging efficiency and/or resource utilization, may have higher maximum feeding rates than the analogous native predators and therefore greater ecological impact [40,41] with especially pronounced consequences in aquatic environments [42].Resource availability represents a crucial determinant of feeding rate as illustrated by a functional response (FR) curve [43,44]. The shape and asymptote of the curve depict important parameters of consumer-resource interactions and population community dynamics [45,46]. Invasive species often display elevated FRs compared to native or low-impact non-native ecologically analogous species [47,48,49] making comparative FR a valuable tool for invasion biologists [48,49,50]. Functional response has been calculated for comparison of N. melanostomus foraging efficiency with native [51] as well as non-native analogous species [52] and can be employed for comparison of predator impact on prey components, since predator response to prey may be prey species–dependent [53,54,55,56,57]. A higher FR asymptote denotes more effective prey exploitation, possibly due to greater prey attractiveness or palatability and/or greater predator adaptation to prey antipredation behaviour. Currently, knowledge of the relationship between N. melanostomus and crayfishes is lacking, especially in tributaries serving as refuges for native aquatic biota and sources of genetic diversity for main stream ecosystems.The aim of our study was to characterize N. melanostomus foraging efficiency on early juvenile crayfish. While natural ecosystems generally consist of multiple prey species per predator, the majority of research experiments address interaction between a single predator and prey species. We observed the predation behaviour of N. melanostomus in the presence of two prey species differing in escape behaviour at several densities and stocking proportions. We hypothesized that prey defence, as well as the presence of an alternative prey in various proportions, may significantly influence predator foraging efficiency.2. Materials and Methods2.1. Predator and Prey Acquisition and AcclimatizationNeogobius melanostomus were collected with a backpack pulsed-DC electrofishing unit (FEG 1500, EFKO, Leutkirch, Germany) in early October 2018 from a recently colonized locality in the Elbe River (50.6524583 N, 14.0441314 E). Specimens (TL = 55.9 ± 2.6 mm; W = 2.1 ± 0.3 g) were transported to the Institute of Aquaculture and Protection of Water and acclimated in a 1600 L recirculating aquaculture system for 7 days. They were fed frozen chironomid larvae to satiation twice daily. Water temperature (20.3 ± 0.3 °C), dissolved oxygen (100.6 ± 2.9%), and pH (7.7 ± 0.2) were measured twice daily with an HQ40d digital multimeter (Hach Lange GmbH, Düsseldorf, Germany).We used two hard-bodied benthic invertebrate prey species of similar body mass differing in escape strategy: The native water louse Asellus aquaticus (L.) (W = 5.56 ± 1.94 mg) is representative of isopods that form a component of the N. melanostomus diet [58,59]. Isopod locomotion is restricted to slow crawling with no escape strategy [60]. The second species was the juvenile non-native marbled crayfish Procambarus virginalis (Lyko 2017) (W = 5.45 ± 0.66 mg), a common crayfish model species for laboratory research [61], which exhibits a flip-tail escape strategy as the native crayfish species [34]. Both native crayfish species in the Czech Republic (i.e., Astacus astacus and Austropotamobius torretium) are classified as critically endangered species in the Red list of threatened species of the Czech Republic with a continual populations decline [62]. Therefore, their use for experiments performance is strongly forbidden and dispensation from law is impossible. Asellus aquaticus was collected with hand nets in late September 2018 in the Kyselá voda stream (49.0195475 N, 14.4640344 E). The P. virginalis were obtained from the Laboratory of Ethology of Fish and Crayfish, FFPW USB. Both prey species were housed in 200 L glass aquaria equipped with PVC trickling filter media (Hewitech GmbH, Ochtrup, Germany) that served as shelter and filter. Half the water volume was exchanged daily with dechlorinated tap water.2.2. Experiment DesignTransparent plastic boxes (295 × 185 × 155 mm; total volume = 6000 mL) filled with 5000 mL dechlorinated tap water and 200 mL fine aquarium sand (particle size < 0.3 mm) were used as experimental arenas. Five prey exposures were tested: A. aquaticus and P. virginalis separately and combined at respective ratios of 1:1, 1:3, and 3:1. Each exposure included prey densities of 4, 8, 20, 36, 60, and 100 individuals/box with six replicates per density. Overall, 180 N. melanostomus specimens were used in the experiment, whereas each predator was used only once. Baseline prey mortality was assessed with control groups of the same combinations, ratios, and densities in six replications without predators. Neogobius melanostomus were starved for 24 h before each trial to standardize hunger level and placed individually into the experimental arenas 1 h after prey insertion. A light regime of 500 lux m2 was maintained in a 12 L:12 D photoperiod. The predator was removed from the arena after 24 h, and the number and species of surviving prey and non-consumptive mortality (NCM) were determined. Non-consumptive mortality was calculated as in [63] including dead prey not ingested by the predator. Each predator was used once to avoid experience bias. 2.3. Data AnalysisThe FR of N. melanostomus was fitted separately for each prey organism and ratio and calculated as a total number of killed prey (sum of NCM and eaten prey). Hence, FR quantified the overall impact of N. melanostomus on prey. The FRs of N. melanostomus on prey were compared between species and among stocking ratios. The type of FR was determined by fitting of logistic regression on the basis of the relationship between the killed prey (Ne) and the initial prey density (N0):(1)NeN0=exp(P0+P1N0+P2N02+P3N03)1+exp(P0+P1N0+P2N02+P3N03) where P0, P1, P2, and P3 represent intercept, linear, quadratic, and cubic coefficients, respectively, estimated using the method of maximum likelihood. If P1 reaches a positive value with P2 negative, the proportion of prey killed is positively density-dependent, which is peculiar to type III FR. However, if P1 is a negative value, the proportion of prey killed declines monotonically from initial prey density, indicating type II FR [46]. Based on logistic regression, we used Rogers’s random predator equation [64] for type II FR in all prey types and ratios, which is suitable for non-replacement design:(2)Ne=N0−(1−exp(a(Neh−T))) where T is time of prey exposure to predator (24 h), a is predator attack rate (predator relative consumption rate corresponds to search efficiency in low prey density manifested in an initial slope steepness on FR curve; L day-1), and h is predator handling time (time pursuing, subduing, and eating of prey combined with time spent prey searching and digestive pause; days prey-1) [65]. For bordering of the Rogers’s random-predator equation by Ne on both sides of the equation, we used the Lambert W function for solving Equation [66]:(3)Ne=N0−W{ahN0 exp[−a(T−hN0)]}ahWe estimated parameters a and h using non-linear least-squares regression and Lamber W function included in the EMDBOOK package [66]. Differences in parameters among prey species and ratios were evaluated based on an overlap of 95% confidence intervals. If no overlap was observed, the parameters significantly differed among the treatments [67].The effects of prey species, ratio, density, and their interaction upon the number of prey eaten, NCM, and killed prey were tested using a generalized linear model (GLM) with Gaussian distribution. Tukey’s HSD post-hoc test was subsequently used for determination of significant differences among exposures. Since the survival rate in all control treatments exceeded 97% (97.2–100.0%), the mortality of predator-exposed prey was attributed exclusively to the presence of N. melanostomus, and datasets were not adjusted for natural mortality. All analyses were conducted in R version 4.0.3 (R Development Core Team 2018). 3. Results3.1. Functional Response TypeIn all exposures, N. melanostomus exhibited the type II functional response (Figure 1): Significant negative linear coefficients in logistic regressions (Table 1).3.2. Attack Rate and Handling TimeSignificantly higher values of attack rate were observed in the trial with P. virginalis offered separately as well as in both 3:1 prey combinations compared with the 1:1 combination and A. aquaticus offered separately. Neogobius melanostomus displayed the highest handling time in the 3:1 trials, with no significant differences among groups in which prey species were offered separately or at 1:1 (Table 2 and Figure 2).3.3. Number of Killed and Eaten Prey and Non-Consumptive MortalityThe number of prey eaten by N. melanostomus was significantly affected by the interaction of species and ratio (F2,102 = 4.71, p = 0.011). This was reflected in a significantly higher number of P. virginalis consumed than A. aquaticus in the group with 3:1 for P. virginalis. There were no other significant differences among trials in the number of prey eaten (Figure 3). The NCM was affected by prey density (F1,103 = 7.33, p = 0.008) and the interaction between prey species and ratio (F2,101 = 5.87, p = 0.004). The NCM at 1:1 was significantly higher at the highest density (100 ind/box) than at densities < 60 ind/box at the same ratio (Figure 4B). There was no difference in NCM of A. aquaticus among the three ratios. In contrast, the NCM of P. virginalis at 3:1 for P. virginalis was significantly higher than 3:1 for A. aquaticus. The NCM was always significantly higher in the prevalent prey species than in the less abundant (Figure 4A). At 1:1, no significant species differences were observed in NCM (Figure 4B). In all exposures, NCM ranged from 0 to 100% of killed prey.The number of killed prey was significantly affected by prey density (F1,103 = 29.82, p < 0.001), species (F1,106 = 4.21, p = 0.042), and interaction of prey species with prey ratio (F2,101 = 40.07, p < 0.001) and density (F1,100 = 6.13, p = 0.015). In both 3:1 trials, N. melanostomus killed a significantly higher number of the prevalent prey species. The number of killed A. aquaticus differed significantly with the proportion and reflected the number offered. In contrast, the number of killed P. virginalis reached similar values at 1:1 and 3:1 for P. virginalis only at 3:1 for A. aquaticus and was significantly lower than at other ratios (Figure 5A). At 1:1, there was no significant difference between species in the number of killed prey at densities <60 individuals/box. With 100 individuals/box at 1:1, N. melanostomus killed significantly more P. virginalis than A. aquaticus (Figure 5B).4. DiscussionThe ability to utilize different prey sources and to switch among prey species as required is an attribute of successful invasive predators that can negatively affect not only prey species populations but also coenoses stability [31,68]. Neogobius melanostomus significantly changes composition of the macrozoobenthic communities in the invaded freshwater ecosystems [25,69]. Tributaries of major rivers serve as refuges for native aquatic biota and as sources of genetic diversity for the main streams [23] that are currently heavily affected by biological invasions [70].Neogobius melanostomus exhibited type II FR toward prey organisms differing in escape strategy regardless of presentation. This type of functional response is typical of carnivorous predators [63,71] and is usually associated with destabilization of prey organism populations [72]. Type II FR was previously observed in N. melanostomus towards amphipods [49,51,73], A. aquaticus [49], and common carp Cyprinus carpio L. larvae [52] under experimental conditions. With increasing habitat complexity [74], switching among prey types [45] and consumption of less preferred prey [75] or prey with a well-developed antipredator defence [76] commonly involves a shift from type II FR to type III FR. However, this expected phenomenon was not observed in our two prey–species system, although prey organisms displayed different escape abilities. This is consistent with Gebauer et al. [77] who found no shift in N. melanostomus FR with increased habitat complexity, suggesting that N. melanostomus is a highly effective predator irrespective of habitat conditions [76] and prey behaviour (this study). Handling time, as the ability to find and process prey, determines the predator maximum feeding rate [50]. This parameter closely correlates with habitat complexity [57,77,78] and, especially, with prey morphology and behaviour [53,79]. The typical crayfish flip-tail escape is generally considered a successful antipredation strategy [34,80] that reduces predator success or at least requires higher predator energy [81,82]. Contrary to expectations, we observed no significant differences in handling time of A. aquaticus and P. virginalis, suggesting that the crayfish escape strategy is ineffective against N. melanostomus predation, at least in early crayfish ontogenetic stages and in sandy substrates. In the trials with a single prey species at low density, N. melanostomus exploited P. virginalis more effectively than A. aquaticus, reflected in its significantly higher attack rate on P. virginalis. Based on these results, we can conclude that crayfish populations, including native species (e.g., genus Astacus and Austropotamobius for European regions), in freshwater ecosystems may be exposed to predation stress by N. melanostomus similar to that on A. aquaticus. Lawton et al. [83] reported that the predator attack rate decreases and handling time is elevated when alternative prey items are available [83], and Colton [53] demonstrated that, in a multi-prey system, both handling time and attack rate vary with quantity and characteristics of the second most available prey item [53]. However, our experimental design did not allow analysis of those parameters with respect to prey species separately in the multiple-prey exposures. Handling time and attack rate in our multi-prey trials reached values different from those that would be expected in single-prey exposures. Regardless of the proportion of P. virginalis, the prey item considered to be a driver of the N. melanostomus attack rate, on the overall offered prey amount, N. melanostomus showed a higher attack rate in both 3:1 ratios compared to 1:1 or A. aquaticus offered separately. At 1:1, the attack rate was similar to that in the system with only A. aquaticus. In both 3:1 trials at lower densities, the attack rate was positively affected, while, at higher densities, N. melanostomus handling time was prolonged compared to expectations based on results gained in the single-prey systems, implying ongoing predator switch to the alternative prey. The prey alternation could be more challenging when prey species occur in unequal quantities. This is in agreement with Colton [53], who stated that the addition of a prey species to a system leads to additional interactions and behaviour changes, and the food system becomes unpredictable. Lawton et al. [83] reported reduced predator pressure on individual prey in such conditions due to the increased handling time and depressed attack rate. However, our data clearly showed that addition of a second prey item led to an increase in N. melanostomus attack rate as well as elevated impact on the prey community. In addition, our study confirms the value of multi-species experimental design in ecological studies to gain a more realistic assessment of predator impact upon prey communities. Several studies have documented N. melanostomus feed selectivity [30,84,85] that differs with locality. The optimal foraging theory states that a predator will maximize energy profit to cost with respect to prey acquisition and processing [86]. Prey selectivity in aquatic ecosystems is affected by multiple factors including prey availability [87]; predator experience [56]; prey size, morphology, and colour [87,88]; and water turbidity [89,90]. The latter is demonstrated by N. melanostomus diet shift to easily available prey under experimental conditions of high turbidity [89]. Therefore, it can be assumed that prey exhibiting an effective escape response and/or high mobility will be less preferred by predators [86]. However, studies of N. melanostomus feed selectivity have often shown contradictory results, with respect to preferences for native [91,92] or non-native [93] species. In addition, overexploitation of certain benthic species regardless of abundance has been observed [30,84,85] and confirmed by our findings of no species-differences in the number of prey killed when presented in equal numbers, while at 3:1, N. melanostomus killed significantly more specimens of the prevalent species. These findings support the hypothesis that N. melanostomus often shows indiscriminate foraging, taking the most readily available prey and easily switching to another source [30,94]. The ineffectiveness of crayfish tail-flip escape strategy against N. melanostomus predation was also shown. An exception was 1:1 presentation of prey at density of 100 ind/box, when N. melanostomus killed significantly more P. virginalis than A. aquaticus, possibly showing predator food preference after satiation [45]. Although the focus is generally on the direct consumption of prey, this is not the only means by which predator ecological impact may occur [95]. We observed that non-consumptive mortality (NCM) may have an even higher effect on prey populations than direct predation [96]. This component of predator behaviour, also known as waste or surplus killing, has been observed in invertebrates [63,97,98] and mammals [99,100,101]. Ignoring NCM may cause a significant underestimation of predator ecological impact [102] as well as energy transfer among trophic levels [97]. In our experimental exposures, N. melanostomus exhibited a high rate of NCM, indicating its potential role in the effect of this predator on prey population abundance and ecosystem function. Our observed NCM is in contrast with previous studies of N. melanostomus FR with fish larvae as prey [52,77] in which no NCM was observed. In mammals, NCM is usually connected either with an ineffective anti-predator response due to lack of co-evolution with the predator [101] or to lack of prey escape response as a consequence of isolated short-term events [102]. In invertebrates, it seems that the satiation level determines whether the prey is consumed. However, hunting and killing of prey are probably directed by mechanisms [97] in invertebrates that differ from that of vertebrates [103]. Johnson et al. [97] assumed that an empty midgut may stimulate predatory damselfly nymphs to capture more prey than can be processed due to filled foregut. It seems that an effect of satiation was not confirmed in our experiment, since N. melanostomus killed both prey species without their consumption after 24 h starvation, even at the lowest densities. Although the NCM has usually been reported to increase as prey density rises [63,97,98], we did not find a correlation of NCM rate and prey density in N. melanostomus, and the proportion of NCM in total prey mortality ranged from 0–100% (33.4 ± 39.2%). Fantinou et al. [63] described NCM elevation at temperatures outside the predator thermal optimum, i.e., in stressful conditions. Similarly, Veselý et al. [98] in a study of Aeschna cyanea nymphs, and Jedrzejewska and Jederzejewski [100] in Mustela nivalis, described higher NCM at lower temperature. However, it is unclear whether the low temperature directly caused change of predator behaviour or influenced prey occurrence and/or behaviour and subsequently predator response. In our study, the temperature ranged within the optimum range reported for N. melanostomus [104]. We can assume that a potential reason for observed high NCM values might be the absence of shelter as a possible trigger of stress, although we have no evidence supporting this assumption or quantifying its importance in the wild in N. melanostomus. Neogobius melanostomus successfully exploited both hard-bodied prey species differing in escape strategy without showing a distinct preference. The simultaneous effects of high N. melanostomus foraging efficiency on P. virginalis and previously documented successful competition of N. melanostomus for shelter with crayfish [105] may demonstrate a potential to regulate P. virginalis populations in the wild. Bovy et al. [106] pointed out that a destabilization effect of predator presence on prey populations is negatively correlated with prey reproduction and dispersal abilities. Therefore, despite a strong interaction between P. virginalis and N. melanostomus as invasive non-native species, an eradication effect is less likely in established P. virginalis populations due to its high fertility rate and overall reproduction ability [107]. However, for native crustaceans, including indigenous European crayfishes that are threatened for many reasons [7,62] and exhibit lower fecundity [108], N. melanostomus may pose a serious risk. Particularly with regards to increasing records of N. melanostomus in smaller tributaries [30,109,110] inhabited by native crayfish, this can be crucial for continuing crayfish existence. More attention should be focused on identifying and clarifying non-consumptive mortality in the wild as a potential element of N. melanostomus foraging behaviour. The reason for ineffective predation in N. melanostomus is unclear, and this is one of the first laboratory foraging studies to report non-consumptive predation in fish. Both indiscriminate foraging behaviour and non-consumptive mortality are important factors that should be taken into consideration for quantification of N. melanostomus impact on native crustaceans in freshwater ecosystems.5. ConclusionsAlthough N. melanostomus shows comparable predation pressure on both preys, it can be a threat to the population stability of already endangered crustaceans such as crayfish. Effective control to limit further spreading of N. melanostomus to tributaries should be a priority. There is a need for more multiple-prey studies, as quantification of N. melanostomus impact on the macrozoobenthic community based on the single prey model may be insufficient. In addition to prey species, their density and relative proportions can significantly influence the N. melanostomus foraging efficiency.

animals : an open access journal from mdpi

10.3390/ani11113197

PMC8614579

Antibiotic resistance happens when bacteria develop the ability to defeat the mechanism of action of chemical compounds designed to kill them. This has become one of the major global concerns in the food chain since it has an effect in diverse steps such as livestock. Poultry products are one of the most consumed type of meat in Spain. In farms, antibiotics are normally used for therapeutic treatments although in the past they were utilized as growth-promoting agents which provoked a high selection pressure in the natural microbiota of fowl. Escherichia coli is a gram negative Enterobacteriaceae that is commonly found in chicken microbiota and can be use as interesting indicator of antibiotic resistance in poultry products.

The prevalence of Escherichia coli was analysed in poultry products from different Spanish retailers and determined its antibiotic resistance capability by phenotypic (ampicillin, amoxicillin, chloramphenicol, gentamicin, imipenem, cefotaxime, tetracycline, ciprofloxacin, trimethoprim, and colistin) and genotypic assays. A total of 30 samples (hindquarters or livers) were collected from supermarkets and butchers. Enterobacteriaceae counts ranged between 3.2 and 6.5 log colony-forming units (CFU)/g, and the highest values were found in livers and in samples from supermarkets. E. coli was detected in 83% of the samples tested, and the highest prevalence was observed in livers (100%) and supermarkets (91%). Regarding the antibiotic sensitivity test, 100% of the E. coli showed resistance to at least one antibiotic. The highest resistance rates were detected for colistin (87%) and gentamicin (79%), while only two antibiotics (chloramphenicol and cefotaxime) showed a resistance lower than 10%. Furthermore, the resistance genes of tetracycline and beta-lactams were analysed by multiplex PCR, revealing that tet(A) and blaTEM were the majority genes, respectively.

1. IntroductionSince 2010, European Union (EU) chicken meat production has grown by 29.8% owing to it being a less-expensive alternative protein and it is perceived as a healthier product. Now, Poland is leading the production market, but the 2nd EU poultry meat producer is Spain (13% in 2021) [1]. Moreover, in the 2019 report issued by the Spanish Ministry of Agriculture (MAPA), it was shown that the consumption of poultry meat has increased more (37.5%) than other types of meat, such as pork (29.8%) and beef (14.6%) [2].As reported, in the past years the consumer demand for poultry products has increased, which has resulted in continuous pressure on the sector to obtain maximum productivity in minimum time. For this reason, diverse strategies have been developed to increase production, such as genetic selection, innovation in feed products, suitable animal welfare, and biosecurity on farms, as well as a proper supply of antibiotics for the treatment and prevention of infectious diseases [3].Antibiotics in poultry farms have been employed as therapeutic treatments, prophylactic measures, and growth-promoting agents [3,4,5], among others, although the third use was prohibited on the 1st of January 2006 by the EU owing to the emergence of bacterial resistance [6]. As a consequence of their wrong utilisation, antibiotics have caused a selective pressure for the appearance and spread of antibiotic-resistant bacteria in the chicken biota, allowing the creation of reservoirs of resistance genes [7].It is important to highlight that diverse zoonotic bacteria are naturally present in the chicken microbiota, such as Campylobacter (principally Campylobacter jejuni and Campylobacter coli), Salmonella spp., and Escherichia coli [8]. Therefore, when antibiotic resistance occurs in these bacteria, the problem acquires a greater dimension since it can endanger the effect of antibiotics in human medicine [9].The presence of antibiotic-resistant strains of E. coli has been evaluated by diverse authors, which has improved the knowledge of the inadequate use of antibiotics in poultry farms [10]. Both in developing countries and in the main poultry producers, the highest rates of resistance have been detected for beta-lactam antibiotics, tetracyclines, and fluoroquinolones. On the other hand, the lowest rates of resistance have been identified for carbapenems or third-generation cephalosporins. Regarding multi-resistance antibiotics, Spain, Brazil, and China, among others, have a prevalence greater than 80% [11,12,13].Therefore, the emergence of antibiotic resistance has become a global setback with a great impact on the economy and public health. Therefore, the actions to be taken must be intersectoral and coordinated by the different organisations involved [14].Accordingly, the main objective of this work was to study the prevalence of phenotype and genotype of antibiotic-resistant E. coli isolates from chicken hindquarters and livers acquired from different supermarkets and butchers in Spain.2. Materials and Methods2.1. SamplingA total of 30 samples of chicken meat were acquired between February and March 2020 from different supermarkets and butchers located in Ciudad Real province (39° N 4° W), a central region of Spain. Of these, 80% were chicken hindquarters, as they are one of the most consumed pieces. The remaining samples (20%) were livers, which have viscera high metabolic importance.2.2. Enterobacteriaceae Counts: Isolation and Identification of Escherichia coliAll the samples were sliced, and 25 g of each sample was homogenised with peptone water (10 g/L) and tween 80 (0.1%) using a masticator (IUL Instruments, Barcelona, Spain). The samples or their decimal dilutions were inoculated in duplicate onto Tryptone Bile X-glucuronide (TBX) agar (Condalab) employing an automatic spiral plater (Eddy Jet 2W, IUL Instruments, Barcelona, Spain). The plates were incubated at 37 °C for 24 h. After that period, colonies were counted, in colony-forming units (CFU), with the automatic colony counter Flash & Go (IUL Instruments, Barcelona, Spain).Presumptive E. coli isolates in TBX agar developed a blue-green colour. Therefore, all these colonies were selected and purified in the same media and subsequently identified by the API 20E test system (Biomeriux) for the elimination of false positives. All E. coli isolates were kept at −80 °C with glycerol until they were studied.2.3. Phenotypic Sensitivity Test of E. coli IsolatesWith the aim of identifying the sensitivity or resistance of the E. coli isolates to diverse antibiotics with clinical interest, a phenotypic test was carried out. A total of ten antibiotic solutions at their minimum inhibitory concentration (MIC) were prepared in Tryptone Soya Broth (TSB) broth (Bioser): ampicillin (32 µg/mL), amoxicillin (32 µg/mL), chloramphenicol (32 µg/mL), gentamicin (16 µg/mL), imipenem (4 µg/mL), cefotaxime (4 µg/mL), tetracycline (16 µg/mL), ciprofloxacin (4 µg/mL), trimethoprim (4 µg/mL), and colistin (2 µg/mL). These doses were selected following the indications given by the Clinical and Laboratory Standards Institute (CLSI, Annapolis, MD, USA).Cell concentrations from overnight cultures were adjusted to an optical density of 1.2 (OD600). In a 96p microplate, 20 µL of the standardised cultures was inoculated in 180 µL of TSB supplied with each of the antibiotic compounds independently.The test was carried out in duplicate, and two negative controls were added (TSB broth + E. coli isolate without antibiotic and TSB broth + antibiotic without cells). Microplates were incubated at 37 °C/24 h, and absorbance measurements at Time 0 (T0) and Time 24 h (T24) were taken at 600 nm using a plate reader (HiPo MPP-96, Biosan, Riga, Latvia).Growth Curves of Beta-Lactam- and Tetracycline-Resistant E. coliWith the obtained results, five E. coli isolates were selected randomly among all the isolates previously classified thanks to the phenotypic assay: one sensitive (code 7.1), two with medium resistance (code 6.8 and 7.15), and the other two with high resistance (code 3.10 and 19.27). The assay was carried out as described above, but only the beta-lactam and tetracycline antibiotics were used. In brief, 20 µL adjusted overnight cultures (OD = 1.2) was added to 180 µL of TSB with the selected antibiotics. As negative controls, TSB + antibiotics and TSB + E. coli isolates were used and all samples were evaluated by triplicate.The OD600 of the growth curves was monitored for 24 h at 37 °C using a plate reader (HiPo MPP-96, Biosan, Riga, Latvia) and taking measurements every 30 min which entailed a total of 48 measurements. Before the samples were read, they were agitated for 5 s at 150 rpm. Growth curves were obtained by plotting OD versus time (h). Kinetic parameters were calculated using the model described by Warringer and Blomberg [15]: lag phase (λ), generation time (G), maximum OD (ODmax) and specific growth rate constant (μmax).2.4. Multiplex Polymerase Chain Reaction (PCR) for the Detection of Antibiotic Resistance GenesThe phenotypically confirmed beta-lactam- and tetracycline-resistant E. coli were analysed for the presence of resistance genes owing to their importance in the poultry industry.2.4.1. Beta-Lactam Resistance GenesA multiplex PCR was carried out to detect the blaTEM, blaSHV, and blaCMY-2 genes. For this assay, the primer sequences were selected from Kozak et al. [16] study and were synthesised by Metabion (Germany). The oligonucleotides sequences, their band size and studies in which oligonucleotides were first reported are shown in Table 1. Each 25 µL reaction mix was prepared with a reaction buffer (10×; Biotools), a MgCl2 solution (2.5 mM; Biotools), dNTPs (200 µM; Biotools), a blaTEM (0.2 µM) primer, a blaSHV (0.4 µM) primer, a blaCMY-2 (0.2 µM) primer, Taq polymerase (1.25 U/µL; Biotools), and 2.5 µL of extracted DNA. Two controls were included in the reaction, a negative control where the extracted DNA was substituted by Milli-Q water, and a positive control which was E. coli with different genes. The amplification process was carried out in a Life Touch thermocycler (Bioer, Hangzhou, China), as follows: 1 initial denaturation cycle at 94 °C/15 min; 30 cycles with the subsequent conditions 94 °C/1 min (denaturation), 55 °C/1 min (hybridisation), and 72 °C/1 min (extension); and a final cycle at 72 °C/10 min (final extension).PCR products and a 100 bp DNA length standard (Biotools) were loaded in an agarose gel of 1.2% and were subjected to 90 V for 1 h. Fragments were visualised with the gel Green (6×) in a gel documentation system and discrimination was carried out based on the band size.2.4.2. Tetracycline Resistance GenesA similar protocol with slight changes was conducted for detecting the tet(A) and tet(B) genes (Table 1). Reaction mixes were elaborated as described before, although in this case tet(A) (0.1 µM) and tet(B) (0.2 µM) primers were added, as well as 3.5 µL of extracted DNA. The multiplex PCR was performed with the same equipment and under the same conditions as before except that the hybridisation reaction was at 63 °C/1 min. The PCR products were visualised as described in the previous section and the same discrimination process was carried out.2.5. Latex Agglutination Test for E. coli O157:H7In order to check if any of the antibiotic-resistant E. coli isolates were serotype O157:H7, a latex agglutination test was carried out following the manufacture’s indications (Microgen Bioproducts, Camberley, UK).2.6. Statistical AnalysisAn analysis of variance (ANOVA) followed by a Duncan test (p < 0.05) was performed to determine if significant differences existed in the Enterobacteriaceae mean log CFU/g values of all samples. In that case, no comparaison between groups was carried out—neither anatomical parts nor retailers. The same procedure was carried out with the kinetics parameters. On the other hand, Student’s t-test was carried out for comparing Enterobacteriaceae mean log CFU/g values between retailers (supermarkets vs. butchers) and types of samples (hindquarters vs. livers). It was established that significant differences existed when p < 0.05.All the statistical analyses were carried out using the IBM SPSS Statistics program version 24.3. Results3.1. Enterobacteriaceae Counts: Isolation and Identification of Escherichia coliCounts were ranged between 3.2 and 6.5 log CFU/g (Table 2). The results revealed that the concentration of Enterobacteriaceae was not homogeneous among all chicken samples analysed. This was supported by the ANOVA, which showed that there were significant differences between all the Enterobacteriaceae counts, and by the Duncan test, which grouped the samples into 15 different groups.Only focusing on the different anatomical parts of the chicken, the liver samples presented counts between 3.6 and 5.4 log CFU/g, while the counts from hindquarters varied between 3.2 and 6.5 log CFU/g. The Student’s t-test determined that no significant differences were detected among the counts of the two groups compared (hindquarters and livers).In the case of retailer’s samples, the Enterobacteriaceae distribution in supermarket samples was heterogenous (3.2–6.5 log CFU/g). A different trend was observed in butchers’ samples, with lower but more homogeneous counts (4.0 and 4.9 log CFU/g). Nevertheless, the Student’s t-test indicated that no significant differences existed between supermarkets and butcher’s counts.All the presumptive E. coli isolates were confirmed using the API 20E test system, which corroborated that all isolates belong to this species. A total of 240 E. coli isolates were identified in poultry pieces, and they were found in 25 of the 30 samples (Table 2). Samples without E. coli were both from supermarkets or butchers, as well as from hindquarters and livers. It was observed that the Enterobacteriaceae counts were not related to the presence or higher proportion of E. coli. In fact, the sample with the highest Enterobacteriaceae/E. coli proportion (74%) did not present the highest Enterobacteriaceae counts (3.4 Log[CFU/g]). The most usual proportion was ≤10% among all the sample types, except six hindquarters samples and two from livers.Regarding the E. coli distribution, 83% of the samples were positive, indicating an extensive prevalence in poultry products (Figure 1). It was noticeable that E. coli was detected in 91% of the supermarket samples while it was found in only 62% of the butcher samples. On the other hand, all liver pieces and 79% of the hindquarter samples presented E. coli.3.2. Phenotypic Sensitivity Test of E. coli IsolatesThe optical density difference (∆OD) between the final (ODT24) and initial (ODT0) values allowed the E. coli isolates to be classified as sensitive (0.00–0.200), medium resistant (0.201–0.400), or highly resistant (>0.400) to antibiotics [20]. All microorganisms presented some antibiotic resistance, and one was even capable of resisting all 10 antibiotics. Isolates were grouped by their susceptibility or resistance to one or more antibiotics in Table 3 where it can be observed various resistance patterns in the anatomical samples.The E. coli isolates that presented resistance to a smaller number of antibiotics were those sampled from livers bought from supermarkets. It can be also observed that isolates from hindquarters sampled in supermarkets showed resistance to a greater number of antibiotics than those which came from hindquarters sampled in butchers. In fact, for hindquarters sampled in supermarkets, the maximum of E. coli isolates documented were able to resist at least six antibiotics while the maximum for those from butchers was detected for three antibiotics. Regarding livers, different resistance trend was documented depending on the retailer of origin. E. coli isolates from butchers resisted more antibiotics than isolates from supermarkets.The distribution of the E. coli isolates (sensitive, medium and high resistance) grown with different antibiotics is shown in Figure 2.Firstly, the highest resistance value was observed for colistin (87% of the strains showed high resistance). In addition, the resistance rate determined for four antibiotics (colistin, gentamicin, imipenem, and tetracycline) was >60% which are considered high and medium resistance isolates. On the other hand, gentamicin and tetracycline showed more medium-resistant E. coli isolates than high-resistant isolates. It was also noticed that the beta-lactam antibiotics tested (ampicillin and amoxicillin) had a similar resistance rate (≈60%) and exhibited a remarkable resistance. In contrast, chloramphenicol and cefotaxime showed the weakest resistance (90% of the strains presented values between 0.00 and 0.200).The distribution of E. coli isolates with medium and high antibiotic resistance is shown in Figure 3. In general, the resistance rates were higher in hindquarters than in liver isolates (Figure 3A), except in the case of cefotaxime. Additionally, it was observed that for all antibiotics, the resistance ratio was much higher in supermarket E. coli isolates than in other isolates (Figure 3B).Regarding multi-resistance (resistance against at least three antibiotics), almost 93% of the 240 E. coli isolates showed multi-resistance.Growth Curves of Beta-Lactam- and Tetracycline-Resistant E. coliThe absorbance measurements were plotted versus time in order to monitor the growth of the E. coli isolates against the three antibiotics that are important in poultry farms.As expected, the sensitive (rate between 0.00 and 0.200) E. coli isolate (7.1) did not show growth (Figure 4). Regarding beta-lactam, the isolates with the highest resistance (3.10 and 19.27) showed better growth curves than those with medium resistance (6.8 and 7.15) in the presence of ampicillin (Figure 4A). The amoxicillin curves (Figure 4B) showed a similar tendency for the four resistant E. coli isolates. In the case of tetracycline (Figure 4C), the best results were observed for the 3.10 isolate, followed by the 19.27 and 7.15 isolates, while the 6.8 isolate presented the slowest growth.Additionally, four kinetic parameters of all curves were calculated which are collected in Table 4. The ANOVA analysis showed that existed significant differences between the kinetic parameters of the sensitive E. coli isolate and those with antibiotic resistance. Regarding medium and high resistance isolates, rate values were ranged between 0.09 h−1 (6.8 in tetracycline curves) and 0.19 h−1 (3.10 in ampicillin curves) while generation times were between 1.61 h and 3.23 h for the same isolates. These isolates were not separated into groups with significant differences by Duncan’s test, except isolate 3.10 in ampicillin and tetracycline curves. On the other hand, the highest ODmax values and the longest latency phase times were detected both in high resistance isolates (3.10 and 19.27) in each antibiotic tested (Not taking into consideration the sensitive isolate). Moreover, the statistical analysis of these parameters catalogued high and medium resistance isolates in two groups with significant differences between them in the three scenarios.3.3. Multiplex PCR for Detection of Beta-Lactam and Tetracycline Resistance GenesWith the aim of detecting genes related to E. coli antibiotic resistance against ampicillin, amoxicillin, and tetracycline, a multiplex PCR with specific primers was carried out. In total, 175 E. coli isolates with medium or high tetracycline resistance and 156 isolates with medium or high beta-lactam (ampicillin and amoxicillin) resistance were analysed. The results of the amplification products are shown in Table 5.Of the three primers tested for beta-lactam antibiotics, only one gene (blaTEM) was amplified in all the samples, and multiple amplification or no amplification was not observed. In contrast, 117 of the isolates that were tetracycline resistant (medium and high) presented the tet(A) gene, while only 16 isolates had the tet(B) gene, which was the minority gene. Additionally, no amplification was observed in 42 of the samples.3.4. Latex Agglutination Test for E. coli O157:H7A total of 241 E. coli isolates with phenotypic resistance to any of the antibiotics assayed were subjected to this analysis. The latex agglutination test revealed that none of the samples were E. coli O157:H7.4. DiscussionAn analysis of the prevalence of antibiotic-resistant E. coli in poultry products was carried out in this study. For this aim and based on the MAPA information [2], chicken was selected as a raw material since is the main type of meat consumed in Spanish homes (12.4 kg per capita/year). The majority of poultry products consumed in Spain come from lean meat and the hindquarters are the most popular anatomical part. Regarding livers, these foods were selected as representing poultry offal products due to their metabolic importance and ease to be found in stores, although offal consumption is much lower than other product types (0.9 kg per capita/year). Other important factors including the sampling location, supermarkets and butchers were selected because they are the main channels of selling poultry meat. The highest volume of purchases during 2019 in Spain was observed in supermarkets (50.0%) while butchers represented 22.5% of the total. Due to this different consumption importance and distribution, the number of samples from each type of meat and retail was different. More samples were selected from those with greater relevance in the transmission of resistant bacteria (hindquarters and supermarkets). Finally, E. coli was chosen among all the diverse Enterobacteriaceae species since is part of the intestinal microbiota of chickens and is considered as an indicator species of antibiotic resistance [21].The data presented in this study revealed a high prevalence of antibiotic-resistant E. coli isolates from poultry products (hindquarters and livers) either from supermarkets or butchers.The high variability in counts in the poultry pieces could be explained by several factors related to Enterobacteriaceae development, such as the initial cell concentration, intrinsic (i.e., pH, water activity and redox potential) and external (i.e., storage temperature) factors of the food, and the hygienical conditions in the food chain [22]. In fact, it is also usual to observe changes in enteric bacteria concentration in different parts and during various stages of poultry products, which could justify why counts are variable among samples studied [22]. A similar Enterobacteriaceae prevalence was found by Blanco Guarner [23] in different poultry pieces, such as livers, sweetbread, and carcasses. Nevertheless, other authors have reported smaller Enterobacteriaceae populations, although comparable counts were observed in organic chicken meat [24].Regarding the presence of E. coli, similar values were observed by other authors, showing a prevalence of E. coli in poultry products between 77% and 100% [12,23,25]. This high prevalence is possible because E. coli is part of the chicken microbiota and poultry meat is a nutritive substrate, with a pH and Aw suitable for the development of E. coli; moreover, this bacterium can survive long refrigeration times [22]. Therefore, E. coli prevalence in poultry products may not be as influenced by geographic origin as other factors such as poultry products that are a favourable matrix for the growth of Enterobacteriaceae.Strict regulations on the use of antibiotics in meat production have contributed to a drastic decline in the use of these compounds in farms. Fortunately, since the first plan against antimicrobial resistance was established in the EU in 2011, the overall sales of veterinary antibiotics in European countries have decreased by more than 34% [26]. Nevertheless, in the past decades, antibiotics have been used indiscriminately as growth promoters or therapeutical agents, leading to an emergency in the health field [27,28]. This fact was observed in the phenotypic sensitivity test of E. coli that indicated the existence of diverse resistance percentages for the antibiotics tested. Our results are similar to those shown in the EFSA and European Centre for Disease Prevention and Control (ECDPC) report on antimicrobial resistance in zoonotic and indicator bacteria from food in 2017 [29].The resistance percentages of E. coli isolates against ampicillin, tetracycline, cefotaxime, ciprofloxacin, trimethoprim, and chloramphenicol were comparable to the values observed in European countries. In the case of cefotaxime and chloramphenicol, which presented the smallest resistance, the same trend has been observed in Belgium (cefotaxime), France, and the United Kingdom (chloramphenicol) [29].Based on the information collected by the EMA (European Medicines Agency) in the 10th ESVAC (European Surveillance of Veterinary Antimicrobial Consumption) report [16], tetracyclines and penicillins (amoxicillin, ampicillin, and metampicillin) were the most sold group of antibiotics between 2011 and 2018 for animal uses in 31 European countries. This could explain the similar resistance rates observed in diverse EU countries and the present study for these types of antimicrobial compounds [29], except for the amoxicillin results (60%), which were closer to those identified in studies carried out in China and India. Furthermore, the resistance rates of trimethoprim and ciprofloxacin in EU were smaller than those of the previously named antibiotics, probably because fewer doses of trimethoprim and fluoroquinolones were sold for use in slaughtered and livestock animals [26,30].Progress towards the development of national action plans against antimicrobial resistance appears high in those countries with large livestock sectors. Brazil, Mexico, Argentina, India, Indonesia, Iran, Russia, China, Japan, and the USA are included in the top 10 chicken-producing countries. Among all of them, nine have at least developed a national action plan [31]. However, some of these countries have, for some antibiotics, higher resistance rates in E. coli than the rates presented in this study such as India, China and Mexico, owing to their plans to monitor the overuse of antimicrobial compounds are minimal or have been recently established compared to EU countries [32]. Bezerra et al. [13] reported a high prevalence of E. coli isolates in broiler chickens, showing resistance to ampicillin (87%), tetracycline (95%), ciprofloxacin (91%), and chloramphenicol (51%). Poultry products from China presented similar resistance values for tetracycline (93%) and chloramphenicol (50%) but higher rates for ampicillin (99%) [11]. Furthermore, other countries, such as Mexico, have reported E. coli strains isolated from retail meats with a marked resistance rate against ampicillin (92%), cefotaxime (78%), and tetracycline (75%) [33]. The percentages showed above are higher than those documented in this study, although, as it was commented before, these results were expected since the monitorisation of antibiotics is much lower in these countries and these compounds are extensively used in the food chain in developing countries [32].Regarding the antibiotics that presented the highest resistance percentage in our study, the E. coli isolates showed a notable resistance to gentamicin (80%), which is commonly injected outside the EU in combination with in vivo vaccines to prevent cross-contamination between eggs. It has been documented that this gentamicin-supplemented vaccine could be a critical driver for antimicrobial-resistant bacterial contamination in poultry since it has permitted the adaptation of some bacteria populations to this antibiotic [34,35], so the use of this vaccine may explain the high resistance we found in isolates from livers or hindquarters. In the case of colistin, an increment in resistance has been monitored in recent years in E. coli isolates from livestock and slaughtered animals owing to its extensive use which could have resulted in a high selective pressure [36]. Finally, imipenem-resistant bacteria in poultry have been studied in India and Nigeria, where resistance rates of, respectively, 31% and 73% were found in multidrug-resistant beta-lactamase-producing E. coli; the latter percentage is in agreement with the data collected in the present study [37,38].Tetracyclines, together with penicillins in which beta-lactams are included, are currently the most sold drugs for livestock use in Spain [26], so it was interesting to monitor the growth curves of sensitive, medium and high resistance isolates and to carry out a genetic analysis of the principal genes which encoded the resistance to these antibiotics in E. coli.The utilization of growth curves in measuring the effect of antibiotics in bacteria has previously helped in establishing the correlation between bacterial morphology and growth as well as to prove their effect in kinetic parameters [39]. The ODmax and latency phase were the two parameters statistically different between medium and high resistance isolates. This is consistent with other antibiotic tolerance assays where it has been documented that bacteria with longer lag phases are more tolerant to these compounds owing to their prolonged exposition to them at the beginning of the division stage which could lead to a greater final culture density [40]. Therefore, the latency phase seems to be a key kinetic parameter to identify antibiotic resistance isolates as well for knowing their tolerance intensity.In the case of specific resistance mechanisms developed by bacteria, tetracycline’s resistance mechanism is generally mediated by efflux pumps, which are codified in E. coli by tet genes, such as tet(A) and tet(B) [41]. Similar results were observed by other authors in poultry products, in which the principal gene amplified was tet(A), while tet(B) (7%) or both genes (2%) were the least amplified genes [42]. Regarding beta-lactams, they are not efficacious when Gram-negative bacteria are able to produce beta-lactamase enzymes, which can hydrolyse penicillins, among other antibiotics, and which are encoded by the plasmid-mediated blaTEM, blaSHV, and blaCTX-M genes [43]. Other authors have documented that the principal gene detected in beta-lactam-resistant E. coli strains was blaTEM [11,44]. In fact, Blanco Guarner [23] found similar results to those reported in the present study, and blaTEM was amplified in 100% of the ampicillin-resistant E. coli strains.5. ConclusionsIn conclusion, the Enterobacteriaceae counts were significantly different among the analysed samples. A high prevalence of E. coli has been observed both in different poultry products and in poultry meat from different types of retailers. Moreover, the phenotypic test carried out with antibiotics used in the poultry industry has revealed high resistance rates, especially against colistin, gentamicin, imipenem, tetracycline, and beta-lactams.These circumstances pose a complex global problem that may have a great impact on the economy. Therefore, the emergence of antimicrobial resistance goes beyond the consequences for human and animal health and is becoming a global public health concern, with the need for intersectoral measures coordinated by the different international organisations involved.

animals : an open access journal from mdpi

10.3390/ani11102793

PMC8532635

Anaplasma phagocytophilum and Anaplasma ovis, tick-borne pathogens with zoonotic potential, have been detected in small ruminants in Europe and North America in the past. These intracellular bacteria cause tick borne fever and ovine anaplasmosis, respectively. The most common clinical signs of infection are fever, lethargy and anaemia. To date, little is known about the distribution of these pathogens in sheep and goats from Germany. Therefore, 3178 serum samples of small ruminants from 71 farms distributed in five German federal states (Schleswig-Holstein, Lower Saxony, North Rhine-Westphalia, Baden-Wuerttemberg and Bavaria) were examined for IgG antibodies to Anaplasma species by a cELISA based on the MSP5 antigen. In 70 flocks, antibodies to Anaplasma spp. were detected in both sheep and goats. Furthermore, a risk factor analysis was carried out by means of a questionnaire answered by the farmers. Older animals and females were more likely to have antibodies to Anaplasma spp. Moreover, sheep had a higher probability of becoming seropositive than goats. Using flocks for landscape conservation and the presence of cats and dogs on the farm increased the risk of having more than 20% seropositive animals within the flock significantly. Since antibodies to Anaplasma spp. have been detected in almost all flocks (70/71), it can be assumed that Anaplasma spp. might be underdiagnosed in small ruminants from Germany.

Knowledge about the distribution of Anaplasma spp. in small ruminants from Germany is limited. Therefore, serum samples were examined from 71 small ruminant flocks (2731 sheep, 447 goats) located in the five German federal states: Schleswig-Holstein (SH), Lower Saxony (LS), North Rhine-Westphalia (NRW), Baden-Wuerttemberg (BW) and Bavaria (BAV). Antibodies to Anaplasma spp. were determined by a cELISA based on the MSP5 antigen. A risk factor analysis at animal and flock level was also performed. Antibodies to Anaplasma spp. were detected in 70/71 flocks without significant difference in the intra-flock prevalence (IFP) between the federal states. The mean antibody levels from sheep were significantly lower in northern Germany (LS, SH) compared to west (NRW) and south Germany (BW, BAV). Sheep had a 2.5-fold higher risk of being seropositive than goats. Females and older animals (>2 years) were more likely to have antibodies to Anaplasma spp. in one third and one quarter of cases, respectively. Flocks used for landscape conservation had a five times higher risk of acquiring an IFP greater than 20%. Cats and dogs on the farms increased the probability for small ruminant flocks to have an IFP of above 20% 10-fold and 166-fold, respectively. Further studies are necessary to assess the impact of Anaplasma species on the health of small ruminants in Germany.

1. IntroductionAcross Europe and North America, sheep and goats can become infected with obligate intracellular bacteria of the genus Anaplasma. Whereas Anaplasma phagocytophilum is widespread in many European countries, an infection with Anaplasma ovis mainly occurs in the Mediterranean Basin [1]. However, reports about the focal occurrence of A. ovis in Central European countries like Hungary, Slovakia and Germany are increasing [2,3,4]. Both pathogens are also present in the US sheep population, but detailed information about the dissemination is lacking [5,6].Wild ruminants may act as a reservoir for both pathogens in Europe and North America [3,7,8,9,10,11]. The transmission of Anaplasma spp. usually happens through tick bites [1]. The main vectors of A. phagocytophilum are Ixodes ricinus in Europe, as well as Ixodes scapularis, Ixodes pacificus and Ixodes spinipalpis in North America [1,11,12,13]. Different tick species belonging to the genera Dermacentor, Rhipicephalus and Hyalomma are considered to transmit A. ovis [13,14]. In recent years, A. ovis was also found in sheep keds (Melophagus ovinus) but their vector competence remains doubtful [15,16].The replication of A. phagocytophilum takes place within the vacuoles of neutrophil granulocytes and sometimes also lymphocytes [17]. This causes granulocytic anaplasmosis in many domestic animals, such as horses [18,19], cattle [20,21], dogs [22,23] and cats [24,25], and also in humans [12]. In small ruminants, A. phagocytophilum results in tick-borne fever (TBF) and affected animals suffer from high fever, anorexia and dullness [26,27,28]. Neutropenia and thrombocytopenia are the haematological key findings in affected sheep and goats [26,29]. Immunosuppression causes a high susceptibility to secondary infections like Mannheima haemolytica and Bibersteina trehalosi and leads to respiratory distress in lambs [30,31]. Furthermore, A. phagocytophilum favours co-infections with staphylococcal bacteria which cause tick-pyaemia with polyarthritis [29]. TBF and co-infections can be fatal for lambs [29,30,31]. However, mild courses of A. phagocytophilum were reported but affected lambs had reduced growth rates [28]. Goats show similar clinical signs to sheep after an infection with A. phagocytophilum, but to a lesser extent [26,32,33].Anaplasma ovis mainly affects the ovine and caprine erythrocytes [34] but can also be found in wild ungulates like roe deer (Capreolus capreolus) and red deer (Cervus elaphus) [7,15,35]. Humans rarely become infected [36]. The pathogen causes ovine anaplasmosis especially in sheep in poor health [37]. Main clinical signs in sheep are fever, severe anaemia, extreme weakness, anorexia, and weight loss [34,37,38,39]. Moreover, haemoglobinuria and icteric carcasses were also described in sheep infected with A. ovis [2,39]. An acute infection results in decreased values of red blood cells, haemoglobin and packed cell volume [40]. Although the same signs are described for goats as for sheep, A. ovis appears to be more pathogenic for goats [41].In Germany, A. phagocytophilum was identified in I. ricinius across the country with detection rates between 1.9% and 5.4% [42,43,44]. Although A. phagocytophilum has been well described in domestic animals [19,20,25,45] and wild ungulates [9,46], knowledge of the occurrence of the pathogen in German sheep and goat flocks is still limited. A molecular investigation revealed an infection rate of 4% (n = 255) in sheep from Northern Germany [8] and a clinical case of TBF was described in a goat from western Germany [47]. Recently, A. phagocytophilum was detected by PCR in five sheep flocks located in the southern part of the country [4]. In the same study, A. ovis was identified for the first time in a German sheep flock.Due to the lack of information about the occurrence of Anaplasma spp. in the German small ruminant population, the present study aimed to determine the seroprevalence of Anaplasma spp. in sheep and goat flocks across Germany by using a cELISA to receive further information about the dissemination of Anaplasma species. Moreover, a risk factor analysis was performed to identify potential threats for sheep and goats in Germany to be exposed to Anaplasma species. This risk factor analysis was based on data from a standardised questionnaire which was performed with the sheep farmers [48].2. Materials and Methods2.1. AnimalsSerum samples from sheep and goats were available from a Q fever study conducted from winter 2017 to spring 2018, and details were described elsewhere [49]. In brief, the specimens were collected from 3178 small ruminants (2731 sheep, 447 goats) within 71 flocks located in five federal states: Schleswig-Holstein (SH), Lower Saxony (LS), North Rhine-Westphalia (NRW), Baden-Wuerttemberg (BW) and Bavaria (BAV) (Figure 1). These states have the largest sheep populations within Germany and the farms were selected based on the owners’ willingness to participate in the study. The number of required samples to determine the intra-flock prevalence (IFP) was calculated on the assumption of 3% expected prevalence, 95% confidence interval, 80% power and 5% precision. A maximum of 44 animals per flock were sampled. In sheep flocks with goats, the sample size for the goats was calculated under the same assumptions as for sheep. Individual ear tag number, species (sheep or goat), sex and age of every sampled animal were recorded.2.2. Detection MethodThe serum samples were tested for IgG antibodies against the MSP5 protein of Anaplasma spp. [50] by using a cELISA (Anaplasma Antibody Test Kit, cELISA v2, VMRD, Inc., Pullman, WA, USA). The assay was performed in accordance with the manufacturer’s instructions and is approved for the detection of antibodies against A. marginale, A. ovis and A. centrale in cattle. Results with an inhibition of ≥30% were specified as positive. This cELISA has already been successfully applied to ovine and caprine sera from areas where A. ovis and A. phagocytophilum are present [37,51].2.3. Risk AnalysisInformation about farm management and flock health were available from a recently performed Q fever study [48]. The standardised questionnaire consisted of questions concerning: (1) general farm indicators, (2) information on livestock kept on the farm, (3) the husbandry system, (4) flock history, and (5) last lambing season [48].2.4. Statistical Analysis2.4.1. SeroprevalenceThe IFP of small ruminant flocks among the five federal states were tested for normal distribution followed by a Kruskal-Wallis test. To examine the distribution of the values of the individual animals between the federal states, the differences were calculated with an ANOVA. Subsequently, a test for least significant differences (Fisher’s Least Significant Difference (LSD) test) was carried out. In the southern federal states (BW, BAV) proportionally more goats were sampled than in the northern and western ones (SH, LS, NRW). Therefore, goats were excluded to avoid a distortion of the analysis and only the antibodies to Anaplasma spp. from the sheep are presented in Figure 2 (PROC LOGISTIC, SAS Institute, Inc., Cary, NC, USA).2.4.2. Risk AnalysisCorrelation AnalysisFirst, all variables were verified in terms of their distinguishable content due to the large number of possible risk factors. A correlation analysis was carried out to support this step.The following measures were determined and confirmed for correlation: Cramer’s V > 0.5 for qualitative variables, ANOVA (equal variances) or Kruskal-Wallis test (unequal variances) (p > 0.05) and coefficient of determination (R2 > 0.1) for qualitative and quantitative variables, and the Pearson correlation coefficient (>0.7) for quantitative variables. Either correlated variables were summarised, one of them removed for further analysis but considered in the subsequent interpretation of the results, or, if there was a moderate correlation, both variables were included in the model selection using an interaction term.Risk Factor AnalysisA risk factor analysis was conducted to identify risk factors for exposure to Anaplasma spp. at animal and flock level. As almost all flocks tested Anaplasma spp. positive, the probability of acquiring an Anaplasma spp. seroprevalence of >20% was determined at flock level. The target variable was dichotomised (positive/negative). Furthermore, the geographical location was dichotomised (North = SH, LS, NRW; South = BAV, BW) to reduce the results’ distortion. Due to the different farm management systems in these two regions, the geographical location of the examined farms was considered as a confounder and therefore the model was stratified for the two regions.For risk factor analysis at animal level, the farm was considered as a cluster variable. Therefore, an extended generalised linear model approach was chosen to take the hierarchical structure of the data into account. The parameters were estimated by using generalised estimating equations [52,53] (PROC GENMODE, SAS, Institute Inc.).For the risk factors at flock level, univariable and multivariable logistic regression models were provided (PROC LOGISTIC, SAS, Institute Inc.). The model assumes the predictors (i.e., risk variables) x1,…, xn and a binary response variable Y (i.e., anaplasmosis <=/> 20%), with p = P (Y = 1). A linear relationship is assumed between the predictors and the logit function of Y = 1. Where ℓ is the logit function and b is the base of the logarithm:ℓ=logbp1−p=β0+β1x1+…+βnxnOdds ratio (OR), a 95% confidence interval (CI), Akaike Information Criterion (AIC) at flock level, or Quasi likelihood under the Independence Model Criterion (QIC) at animal level and p values were calculated for categorical and continuous variables. The Odds ratio was determined using:p1−p=bβ0+β1x1+…+βnxnA variable was used for further analysis if it had a p value lower than 20% (p < 0.20) of the model [54]. Additionally, a distinctive OR < 0.75 or OR > 1.33 and a reasonable corresponding 95% CI (lCI > 0.001; uCI < 999.99) led to the variable being taken further into account. In rare cases, a variable took on the same value on all observed farms. As a result, it was impossible to calculate meaningful ORs and CIs and the corresponding variables were not considered for the multivariable models. These criteria allowed variables to be considered for further analysis if they did not have a p value lower than 20% but still had a distinctive OR. Hence, the multivariable model could be selected from the largest number of possible risk factors and the probability of wrongly removing influencing factors was minimised.A forward selection was carried out with the variables that met the above-mentioned criteria. The variables which most improved the model fittings and whose addition achieved the best p values of the models were selected. The addition of the variables to the models was terminated either if all variables were included or if the addition of variables led to no further improvement of the model fittings and the p values. In the ultimate step, the final models were each examined for collinearity using the variance inflation factor.3. Results3.1. Occurrence of Anaplasma spp. in German Small Ruminant FlocksIn total, antibodies against Anaplasma spp. were detected in 70 out of 71 small ruminant flocks. The IFP ranged between 0% and 97.7%. However, the level of IFP did not differ in principle between the federal states (p > 0.05). There was only one sheep flock in BAV without antibodies to Anaplasma spp. (Figure 2).Overall, the mean antibody levels against Anaplasma spp. in sheep from the northern federal states, SH and LS, were significantly lower compared to NRW, BW and BAV (Figure 3). Among the northern federal states, there was no difference in the sheep’s antibody response (p > 0.05). Moreover, the mean antibody levels were also not statistically different between BW vs. NRW and BAV vs. NRW (p > 0.05), but sheep from BAV had significantly lower Anaplasma spp. values than animals from BW.3.2. Univariable Analysis3.2.1. Risk Factors at Animal Level for Exposure to Anaplasma spp.Sheep had a 2.5 times higher risk acquiring an Anaplasma spp. infection than goats. Females had a 37% increased chance of being seropositive, but the likelihood of antibody detection in young animals (≤2 years) was reduced by one quarter (Supplementary Table S1).3.2.2. Risk Factors at Flock Level for Exposure to Anaplasma spp.At flock level, only landscape conservation and cats had a significant p-value (Supplementary Table S2). Small ruminant flocks used for landscape conservation were four times more likely to have an IFP of above 20%. Farms with cats had an almost four-fold higher risk of having more than 20% IFP than farms without cats.3.3. Multivariable Analysis3.3.1. Risk Factors at Animal Level for Exposure to Anaplasma spp.The results of the multivariable analysis were in line with the univariable analysis and revealed no additional information (Table 1).3.3.2. Risk Factors at Flock Level for Exposure to Anaplasma spp.The resulting multivariate model at flock level included four risk factors (Table 2) which were all significant except for contact to deer.Small ruminant flocks which performed landscape conservation had about a five times higher risk of having more than 20% seropositive animals compared to flocks without this farming purpose. Observations of deer near the flock reduced the risk to less than one-sixth, while the presence of cats and dogs on the farms increased the probability of having an IFP above 20% 10-fold and 166-fold, respectively.4. DiscussionInformation about Anaplasma spp. in the German sheep and goat population is still scarce [4,8] despite A. phagocytophilum having been well described in horses [18,19], cattle [20,21], dogs [22,23], cats [24,25] and wild ruminants [9,55] from Germany.4.1. Occurrence of Anaplasma spp. in German Small Ruminant FlocksIn the present study, antibodies to Anaplasma spp. were determined in almost all small ruminant flocks, indicating a high distribution of Anaplasma species across the country which is consistent with molecular findings in small ruminants in different parts of Germany [4,8,47]. The comparison of our serological results with findings from other studies is hampered due to the different study designs and methods to identify antibodies to Anaplasma spp. in small ruminants.However, our results are similar to serological studies from Italy, Hungary and the US, which also detected a high seroprevalence in small ruminants by using the same cELISA. In Italy, antibodies against Anaplasma spp. were detected in 69.59% (n = 217) and 45.45% (n = 22) of the examined sheep and goats, respectively [56]. In a subsequent study, almost all sheep (98%) in a single flock (n = 200) from Italy had antibodies to Anaplasma species [37]. A high detection rate was also reported from five sheep flocks (n = 156) in Hungary, with 99.4% seropositive animals [2] and from a sheep flock (n = 357) from Idaho (USA) with an antibody response to A. ovis of 94.8% [5]. In contrast, the presence of antibodies to Anaplasma species varied from 0% to 43.3% among eight sheep flocks in California and Oregon (USA) determined by cELISA and indirect immunofluorescence assay [6]. The mean antibody levels in sheep from the northern federal states (SH, LS) were significantly lower compared to the levels from sheep in the West (NRW) and the South (BW, BAV) of Germany. This might indicate a different exposure to ticks and to Anaplasma species of sheep and goats in different parts of the country. Despite I. ricinus being widely distributed across Germany [57], sheep from LS and SH may have less contact to ticks. In these federal states, sheep play an essential role for coastal protection by grazing on dykes. These areas usually have a low vegetation which is less preferred by ticks [58,59]. In contrast, sheep flocks in the South are widely used for landscape conservation and grazing on forest edges and scrubs representing the natural habitat of ticks [58,59]. This hypothesis is supported by our findings from the multivariable risk factor analysis. Small ruminant flocks which performed landscape conservation had a five times higher likelihood of having an IFP above 20%. Moreover, the focal dissemination of Dermacentor spp. in NRW, BW and BAV [57,60] as a potential vector for A. ovis and the recent detection of this pathogen in BAV [4] might also influence the higher antibody response in small ruminants from these areas compared to animals from Northern Germany.4.2. Risk Factor Analysis on Animal LevelThe likelihood of sheep testing positive for antibodies to Anaplasma spp. was 2.5 times higher than for goats. This corresponds to observations from other serological studies in which the percentage of seropositive animals was higher in sheep than in goats [56,61]. Sheep and goats responded differently to A. phagocytophilum regarding the severity of clinical disease and haematological changes [26]. These might also result in differences in production and longevity of antibodies against Anaplasma species. Additionally, it might be possible that ticks are attracted to a varying extent by sheep and goats due to their different odour. Some tick species might prefer individual mammal species and are attracted by the specific host’s odour, but I. ricinus seems not to have any preferences [62,63]. However, the conducted studies did not include goats as targeted species.Due to the lack of Anaplasma risk factor analyses for sheep or goats in Europe and North America, we will discuss our outcomes in a broader context. Older animals (>2 years) and female small ruminants had a higher risk of becoming exposed to Anaplasma spp. in the current study. In general, older animals might be exposed to ticks for a longer period of time, which increases the chance of being exposed to Anaplasma species. This is supported by findings from A. ovis DNA positive goats in France [64]. Animals older than three years tested positive significantly more often. However, age was not related to Anaplasma spp. antibody levels in sheep from Hungary [2]. Furthermore, antibodies against A. phagocytophilum were detected age-independent in horses from the Czech Republic [65]. These conflicting results might occur due to differences in Anaplasma species, environmental conditions and animal husbandry. Similar to our findings, mares were more likely to be A. phagocytophilum seropositive than stallions [65]. Generally, male breeding animals are managed differently compared to females; for instance, males graze on a separate pasture or are kept indoors. The latter especially reduces the likelihood of tick infestation and thus the risk of Anaplasma spp. exposure.4.3. Risk Factor Analysis on Flock LevelThe presence of cats on farms significantly increased the risk of Anaplasma spp. exposure in small ruminants. In Germany, there is evidence of A. phagocytophilum infections in cats. A seroprevalence of 16.7% (n = 326) was detected in cats from LS and BAV [24] and a proportion of 23% A. phagocytophilum seropositive cats (n = 956) were reported from a nationwide study [25]. However, both studies examined sera from pet cats which might have less access to tick habitats compared to barn cats. Therefore, the infection rate with A. phagocytophilum might be higher in the barn cat population. This must be validated in the future. Like cats, dogs were also identified as a risk factor for small ruminant flocks having an IFP greater than 20%. Their presence on farms increased the risk 166-fold. In total, only three farms (n = 71) had no dogs and from this group, one flock had an IFP of >20%. This small sample size might have distorted our findings and probably overestimates the risks of dogs. Nevertheless, canine granulocytic anaplasmosis has been regularly diagnosed in Germany and an A. phagocytophilum seroprevalence between 43.2% and 50.1% was reported within the dog population [22,66]. Finally, dogs and cats are hosts for A. phagocytophilum but their epidemiological role for Anaplasma spp. infections in small ruminants remains unclear.The deer population is considered to be a reservoir for A. phagocytophilum and A. ovis in several countries [3,7,8,9,10,11,67,68]. Molecular investigations determined high infection rates of A. phagocytophilum in roe deer (Capreolus capreolus, 98.9%), red deer (Cervus elaphus, 100%), sika deer (Cervus nippon, 86.4%) and fallow deer (Dama dama, 72.1%) in Germany [9,46,69]. In the present study, the farmers were not asked to distinguish between deer species when these animals were in the vicinity of their small ruminant flocks. Our findings indicate that deer did not increase the likelihood of exposure to Anaplasma spp. in small ruminants. The roe deer population seems to have its specific A. phagocytophilum variants and gene clusters, which are rarely found in sheep [4,8,70]. In contrast, red deer and sheep might share some A. phagocytophilum variants/gene clusters [4,8,70,71]. Therefore, the interaction of Anaplasma spp. between sheep and different deer species needs further clarification.The observed tick infestation by the farmers and the ectoparasitic treatment had no effect on the seropositivity of the sheep and goats, which is in line with findings from a previous study on tick-borne encephalitis virus infections in small ruminants [72]. Although a significant reduction in tick infestation was achieved after the prophylactic treatment of lambs with flumethrin, the seroprevalence was not reduced through the treatment [73]. Consequently, other preventive measures are necessary to protect small ruminants from exposure to Anaplasma species.Species differentiation by cELISA is not possible and is a limitation of our study. Hence, molecular investigations are crucial to distinguish between A. phagocytophilum, A. ovis and possible other Anaplasma species. The recent detection of A. ovis for the first time in a German sheep flock underlines the necessity of species identification [4]. Furthermore, A. phagocytophilum is widely distributed within the German small ruminant population, knowledge on different genetic variants is scarce but the limited numbers of studies indicate a large diversity of A. phagocytophilum variants [4,8,47]. Therefore, molecular analyses of Anaplasma spp. isolates from sheep and goats may reveal new insights into the epidemiological situation and the possible clinical impact of different Anaplasma variants on sheep and goats in Germany.5. ConclusionsThe present study generated important information about the dissemination of Anaplasma spp. infection within the small ruminant population in Germany and contributes to a better understanding of the complex epidemiology of anaplasmosis in sheep and goats. Almost all small ruminant flocks tested seropositive for Anaplasma spp. antibodies. Moreover, due to the wide dissemination of Anaplasma spp. in the German small ruminant population, an infection with these pathogens is probably completely underdiagnosed. Thus, in cases of unspecific clinical signs like high fever, anorexia, anaemia, dullness and poor growth rate in sheep and goats, veterinarians should rule out an infection with Anaplasma species.

animals : an open access journal from mdpi

10.3390/ani13050835

PMC10000100

Optimal nutrition is important for Norwegian-farmed Atlantic salmon in the challenging early seawater phase, which shows a higher mortality leading to significant economic losses. Phospholipids are reported to enhance growth, survival, and health in the early stages of the fish life. Atlantic salmon (74 to 158 g) were fed six test diets to evaluate alternative phospholipid (PL) sources in freshwater and were transferred to a common seawater tank with crowding stress after being fed the same commercial diet up to 787 g. Krill meal (KM) was evaluated using dose response with the highest 12% KM diet compared against 2.7% fluid soy lecithin and 4.2% marine PL (from fishmeal) diets, which were formulated to provide the same level of added 1.3% PL in the diet similar to base diets with 10% fishmeal in the freshwater period. A trend showing increased weight gain with high variability was associated with an increased KM dose in the freshwater period but not during the whole trial, whereas the 2.7% soy lecithin diet tended to decrease growth during the whole trial. No major differences were observed in liver histology between the salmon that were fed different PL sources during transfer. However, a minor positive trend in gill health based on two gill histology parameters was associated with the 12% KM and control diets versus the soy lecithin and marine PL diets during transfer.

Growth and histological parameters were evaluated in Atlantic salmon (74 g) that were fed alternative phospholipid (PL) sources in freshwater (FW) up to 158 g and were transferred to a common seawater (SW) tank with crowding stress after being fed the same commercial diet up to 787 g. There were six test diets in the FW phase: three diets with different doses of krill meal (4%, 8%, and 12%), a diet with soy lecithin, a diet with marine PL (from fishmeal), and a control diet. The fish were fed a common commercial feed in the SW phase. The 12% KM diet was compared against the 2.7% fluid soy lecithin and 4.2% marine PL diets, which were formulated to provide the same level of added 1.3% PL in the diet similar to base diets with 10% fishmeal in the FW period. A trend for increased weight gain with high variability was associated with an increased KM dose in the FW period but not during the whole trial, whereas the 2.7% soy lecithin diet tended to decrease growth during the whole trial. A trend for decreased hepatosomatic index (HSI) was associated with an increased KM dose during transfer but not during the whole trial. The soy lecithin and marine PL diets showed similar HSI in relation to the control diet during the whole trial. No major differences were observed in liver histology between the control, 12% KM, soy lecithin, and marine PL diets during transfer. However, a minor positive trend in gill health (lamella inflammation and hyperplasia histology scores) was associated with the 12% KM and control diets versus the soy lecithin and marine PL diets during transfer.

1. IntroductionFarmed salmon are typically transferred from early phase production in tanks on land to seawater cages that constitutes a challenging environment, where fish can experience significant mortality before reaching harvest size. For example, mortality in Atlantic salmon ranged from 15 to 16% from 2017 to 2021 in Norway, with approximately 35% of sea cage mortality occurring in the first 0–3 months at sea for the 2010–11 salmon generations in the Norwegian-farmed Atlantic salmon [1]. This mortality in the early sea cage phase leads to significant economic loss [2]. Thus, research on optimal nutrition to produce robust smolts for improved survival and growth after transfer to the sea cage is of interest to the aquaculture industry [3]. Fish meal (FM) and fish oil (FO) dominated early commercial salmon feed formulations and provided essential nutrients, but usage of these marine ingredients has declined over time as they are limited resources at generally higher prices compared to alternative ingredients where sustainability measures are also considered [4]. Antarctic krill meal (KM; Euphausia superba) is a commercially known ingredient in salmon feeds, with potential benefits toward enhancing growth and health in salmonids [5]. The krill fishery in the Antarctic Southern Ocean is considered highly regulated and sustainable [6,7]. KM provides a range of nutrients including proteins (similar amino acid profile to FM); water soluble nitrogenous components (free amino acids, peptides, nucleotides, and trimethylamine N-oxide), which can act as potential feed attractants; astaxanthin; marine omega-3 fatty acids (eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)); and phospholipids (PLs) [5]. Substantial evidence exists showing that dietary PL can improve growth, survival, and health (reduced intestinal steatosis and deformities) in the larval and early juvenile stages of the fish [8,9,10,11]. In addition, KM and krill oil (KO) reduced fat accumulation in the hepatocytes in comparison to soybean lecithin as the PL sources in the diet of seabream larvae [10,12,13]. In addition, there was an indication that seabream juveniles that were fed a diet with 9% KM had lower hepatocyte vacuolization (fat storage) versus a control diet without KM that was higher in fishmeal [12,13], and a non-significant trend for lower hepatocyte vacuolization was indicated for seabream larvae that were fed a diet with krill oil versus soybean lecithin as the PL source [10]. PLs from different sources can have different properties. KM has approximately 40% PL consisting of the total lipid with phosphatidylcholine (PC) at >80% of the total PL and ca. 18% EPA + DHA of the total lipid [14]. In comparison, fluid soy lecithin can have approximately 46% PL of product (does not include glycolipids and complex sugars) with ca. 35% PC of the total PL and ca. 55% 18:2n-6 of the total FA as the major FA with no EPA + DHA [15]. KM has been documented in the diet of seawater salmon [16,17,18], however, only KO has been documented in the diet of freshwater salmon during the pre-transfer to the seawater phase [19]. The objective of the present study was to document the effect of the KM dose as a source of PL and compare it against other PL sources in the feed of freshwater Atlantic salmon during the pre-transfer phase followed by the early seawater phase by evaluating the growth and histological health parameters. A four-level graded dose response for KM up to 12% of the diet along with a comparison of alternative PL sources (soy lecithin and marine PL from fishmeal) formulated to provide the same level of added 1.3% PL in the diet as 12% KM was evaluated in freshwater diets for salmon during the pre-transfer phase. Fish identified by pit tag with this pre-transfer freshwater feeding history were then transferred to a common seawater tank with crowding stress after transfer and a drop in water temperature at transfer (crowding and water temperature drop can be experienced at transfer commercially) and then were fed the same commercial feed. Gill and liver histology were also compared for salmon that were fed the alternative PL source diets at the end of the freshwater pre-transfer period.2. Materials and Methods2.1. Feed Formulation and CompositionThree different sources of PL were tested in pre-transfer freshwater feeds: (i) krill meal (QrillTM Aqua; Aker BioMarine Antarctic ASA) at four levels for dose response (4%, 8%, and 12% of diet), (ii) fluid soy lecithin as a vegetable PL source, and (iii) marine phospholipid-rich oil sourced from North Atlantic fish species from Triple 9 (TripleNine, Trafikhavnskaj 9, DK-6700 Esbjerg, Denmark)). , and a control diet. The trial diets are referred to as Control, KM4, KM8, KM12, VegPL, and MarPL, respectively. Trial feeds were formulated using a commercial formulation program with external oil mix calculations and produced by extrusion at Cargill Innovation Center (Dirdal, Norway) for ca. 74 g fish with lipid nutrients and then adjusted for purposes of the trial. The 4-mm pre-transfer freshwater trial feeds were formulated and analyzed to have similar digestible energy (22.1–23.6 MJ/kg gross energy), protein (46–49% range), and fat (22–24% range) (Table 1) and with similar calculated 1.1% EPA + DHA in diet, 15–16% saturated in total FA, and 1.3 n-6/n-3 fatty acid (FA) ratio across trial feeds. Protein was analyzed by the Dumas principle using the Elementar Rapid Max N system. Fat was analyzed by low-field nuclear magnetic resonance scan using the NMR Analyzer Bruker minispec mq10 system (Cargill Innovation Center, Dirdal, Norway). Gross energy was analyzed by the Leco gross energy bomb calorimetry system (Cargill Innovation Center, Dirdal, Norway). Moisture was predicted by the NIR FOSS DS2500 system (Cargill Innovation Center, Dirdal, Norway) by using the feed model at Cargill. A similar 1.3% PL in diet across pre-transfer freshwater diets was calculated from the addition of 12% krill meal, fluid soy lecithin, and marine PL test ingredients to base formulations with the same 10% fishmeal level across the diets. There was variation in the other ingredients (added oil, plant ingredients, and micronutrients) needed for balancing or reaching nutrient targets. The choline level was formulated to be the same for control and VegPL diet with MarPL and KM12 providing additional choline to these diets in the form of phosphatidylcholine (PC). However, formulated choline levels for control diet and fluid soy lecithin diets were in excess of the NRC 2011 requirements for salmonids and in excess of the lowest choline level used by Hansen and coworkers [20] with no growth differences observed (1340 to 4020 mg choline/kg diet dose response trial for 456 g initial weight salmon). Lipid accumulation in the gut was reduced for salmon (456 g initial weight) at increased choline levels [20]. The formulation and composition of feeds are given in Table 1.2.2. Fish Trial ConditionsThe experiment was performed according to the guidelines and protocols approved by the European Union (EU Council 86/609; D.L. 27.01.1992, no. 116) and by the National Guidelines for Animal Care and Welfare published by the Norwegian Ministry of Education and Research.Atlantic salmon (Salmo salar) with an initial weight of ca. 67 g were used for the trial. The fish were pit-tagged and randomly distributed into 24 freshwater flow-through tanks (1 m diameter and 0.45 m3 volume) to have 40 fish per tank at the start of trial diet feeding. These fish after 15 days of tank acclimation were 74 ± 12 g (average ± SD for all 960 fish in 24 tanks at the start of trial feeding) and then were fed the freshwater pre-transfer trial diets (Table 1) over a 53-day period. Water temperature averaged 14.3 °C (13.3–15.3 °C range) with 107% average oxygen saturation at the inlet and 90% oxygen saturation at the outlet during the freshwater acclimation and trial diet feeding period. Fish were fed the six trial diets to four replicate tanks during the 53-day freshwater pre-transfer period using an automatic belt feeder with continuous feeding for 20 h per day in excess of satiation level. Feed intake was calculated on a weekly basis by collecting and weighing uneaten pellets as well as by weighing the amount fed. There was a 12 h light: 12 h dark photoperiod regime from Day 0 at freshwater tank acclimation to Day 33 after which a 24 h light regime was used to initiate smoltification. After this freshwater pre-transfer feeding period, fish from all the tanks (17–20 fish per tank from the 24 freshwater tanks) were transferred to a larger common seawater flow-through tank (5 m diameter and 21.6 m3 volume with 28.5 ppt salinity, and no acclimation time from 0 ppt freshwater to 28.5 ppt seawater) with a water temperature drop at transfer (ca. 14 to 9 °C) and crowding stress after transfer (lowered water level to ca. 20 cm for one hour with supplemental oxygen for all 459 fish of ca. 167 g within a ca. 0 to 30 h period after transfer) in the common seawater tank after all 17–20 fish per tank from the 24 freshwater tanks fish were transferred over and then were fed a common commercial extruded salmon diet (EWOS AS) for a further 98 days. Daily water temperature was lower during the seawater phase averaging 9.4 °C (8.5–11.1 °C range).2.3. Fish GrowthThe 40 fish per tank were weighed individually with pit-tag identification on acclimation to the freshwater tanks (Day 0), at the start of trial diet feeding (Day 15), at intermediate weighing (Day 33), and after 53 days of trial feeding in the freshwater (Day 68). The fish weight gain in the freshwater pre-transfer period from Day 15 (start of freshwater trial diet feeding) to 68 were compared statistically between diets. A total of 17–20 fish from each of the 24 freshwater tanks were transferred to the common seawater tank on Day 68 with fish weighing performed on Days 35, 73, and 98 after transfer to seawater. There were 9 to 17 fish representing the original tanks in the freshwater period with 50 to 58 fish representing each of the test diets from the freshwater period at final weighing in seawater at 98 days after transfer to the common seawater tank. The fish weight gain over the whole trial period in freshwater and seawater from Day 15 to 166 days were statistically compared between diets.2.4. Hepatosomatic IndexHepatosomatic index (HSI) is the liver weight percent of the whole body weight. HSI was measured on 10 fish randomly sampled per tank (four tank replicates per diet) to study 40 fish per diet at the end of the freshwater pre-transfer period when fed test diets and 40 fish per diet (identified by pit-tag) at the end of the seawater phase when fed the common commercial diet.2.5. HistologyGill and liver histology were performed on the fish involved in the dietary phospholipid source comparison (KM12, VegPL, and MarPL) and on fish fed the Control diet at the end of the freshwater pre-transfer period. Liver (half tissue section) and gill (left gill arch 2) tissues were randomly sampled from five fish per tank to give a total of 20 liver and 20 gill tissues per diet group for histological analysis. The tissues were fixed in formalin (4% formaldehyde) and stored at room temperature until sent to Pharmaq Analytiq AS (Harbitzallée 2A, 0275 Oslo, Norway) for histological analysis.2.6. Statistical AnalysisThe weight gain for the different periods was modelled by computing the weight gain of each tagged individual and then using a hierarchical generalized additive model (GAM) with the spline function to describe the possibly non-linear dose-response. A random effect of tank was added to the model to account for the multiple individual observations per experimental unit. The total feed intake over the periods of interest was modelled with a single level GAM with a spline function describing the dose-response function. Hepatosomatic index (HSI) was modelled by a hierarchical GAM model using a spline function to describe the dose-response function, mean-centered round weight of the fish as a covariate, and a random effect of tank to account for the multiple individual observations per tank. From this model the expected liver weight was solved for an average-sized sampled fish and expressed as HSI by dividing the expected liver weight with the mean round weight of the sample. Gill and liver histology scores are ordinal variables for which common arithmetic operations, such as sum or mean, are not defined and therefore scores require an ordinal model returning the score probability for evaluation. A hierarchical GAM for ordinal data was set up by using a spline function to describe the dose-response function, and a random effect of tank was included to account for multiple individuals observed per tank. The models for weight gain, feed intake, and HSI assumed the error distribution is the normal distribution, and the model for gill and liver scores assumed the model is ordinal and the errors followed the ordered categorical family. All data processing and statistical modelling was conducted with the R language [21]. The GAMs were estimated with the “gam” function of the R language add-on package “mgcv” [22].The outcomes from the fitted statistical models are presented graphically by showing the mean response and the 95% credible intervals. The mean (median) response and the 95% credible intervals were computed with the help of a parametric bootstrap (with 10,000 random draws per parameter) by taking the 25%, 50%, and 97.5% quantiles of the computed response vector. In the case of a categorical predictor variable (for comparing the different PL sources), the graphs show the mean and an error bar of the 95% credible interval. In the case of a continuous predictor (for the dose-response of krill meal inclusion), the mean response is shown as a median dose-response curve and the 95% credible interval is shown as a confidence band around the mean curve. This way both the magnitude of any potential effect (biological significance) and the uncertainty of any effect estimate (statistical significance) can be shown in the same graph for all the results independent of the response following the normal, binomial, or ordered categorical distribution.3. Results3.1. Growth PerformanceAtlantic salmon of 74 g (overall tank average) were fed the six test diets up to 158 g (overall tank average), growing 2.1-times the initial fish weight to the end of the freshwater pre-transfer period. There was no clear trend for increased feed intake with KM dose in the FW pre-transfer phase (Figure 1). A trend for increased feed intake was indicated for the Control and KM12 diets compared to the MarPL and VegPL diets in the PL source comparison for the FW pre-transfer phase (Figure 2). There was overall high variability for the feed intake comparisons. A trend for increased fish weight gain with high variability was indicated with increased KM dose in the FW phase (Figure 3). There was similar weight gain during the whole trial with feeding the KM dose in the FW pre-transfer phase followed by feeding the same commercial diet in a common tank for the SW phase (Figure 4). Fish fed the KM12 diet had increased weight gain compared to the VegPL diet with the MarPL and Control diets having intermediate weight gains in the PL source comparison for the FW pre-transfer phase (Figure 5). Weight gain was similar for the fish that were fed KM12, MarPL, and Control diets, with a trend for higher indicated weight gain than the VegPL group during the whole trial, with feeding the KM dose in the FW pre-transfer phase followed by feeding the same commercial diet in a common tank for the SW phase (Figure 6, Tables S1 and S2).3.2. Hepatosomatic IndexA trend for decreased hepatosomatic index (HSI; liver% of fish weight) was indicated for the fish that were fed increased KM dose from 0 to 12% of diet at the end of the freshwater pre-transfer feeding phase (Figure 7). There was no decrease in HSI with feeding KM dose at the end of the whole trial after the FW pre-transfer phase followed by feeding the same commercial diet in a common tank for the SW phase (Figure 8). A lower HSI was indicated for the fish that were fed the KM12 diet compared with the fish that were fed the MarPL, VegPL, and Control diets at the end of the freshwater pre-transfer feeding phase (Figure 9) with a similar minor HSI trend observed over the whole trial (Figure 10).3.3. Histology3.3.1. Gill HistologyAn increased probability for very mild to mild gill lamella inflammation and hyperplasia score was indicated for the salmon that were fed the VegPL and MarPL diets compared to the Control and 12% KM diets at the end of the freshwater pre-transfer phase after 53 d of feeding the trial diets (Figure 11a,b). Other following gill histology responses were evaluated with no major differences between the diets: vascular lesions, filament inflammation, necrosis of respiratory epithelium, necrosis affecting deeper tissues, fusion of lamella ,and other lesions noted as present or absent.3.3.2. Liver HistologyNo major differences were observed in liver histology between the control, 12% KM, soy lecithin, and marine PL diets at the end of the FW pre-transfer phase after 53 d of feeding the trial diets (data not shown). The following liver histology responses were evaluated: total amount of abnormal tissue, inflammation, necrosis, inflammation in liver tissue or capsule (peritonitis), peribiliary or perivascular inflammation, neoplasia, fibrosis, lipid deposition, other degenerative changes, vascular lesions, and other lesions noted as absent or present.4. DiscussionThe present study evaluated the effect of different phospholipid sources fed over 53 d in the freshwater pre-transfer phase followed by feeding the same commercial diet over 98 d in a common seawater tank on growth performance and health parameters of Atlantic salmon. KM was evaluated in dose response (4%, 8%, and 12.0% of diet), and diets with 2.7% fluid soy lecithin (VegPL) and 4.2% MarPL as alternative PL sources were formulated to provide the same level of added 1.3% PL in diet as 12% KM. All the test diets contained 10% fishmeal in the FW phase. A trend was indicated for increased fish weight gain (high variability) with increased KM dose in the FW pre-transfer phase but a carry-over effect on growth was not observed for the same salmon fed the same commercial diet after seawater transfer. Salmon (104 g initial weight) that were fed krill meal at 7.5 and 15% of diet for higher fishmeal diets (40–52% of diet range) than the current trial had increased growth after transfer to sea cage [16]. Fishmeal provides PL, so higher fishmeal diets may reduce the need for KM as a PL source [23]. However, KM also provides amino acids (protein), water-soluble nitrogenous components (potential feed attractants), astaxanthin, and EPA + DHA, hence, it is more than a PL source. KM feeding may need to continue after sea water transfer to have a positive effect on growth at the end of the trial, noting the positive effects of KM on salmon growth observed in other but not all trials, which can depend on life stage and challenges, diet composition, KM refining (de-shelling etc.), and inclusion level [5].A trend for decreased fish weight gain was indicated for the VegPL diet in the FW phase and over the whole trial compared with the control diet, whereas the MarPL diet showed more similar growth to the control diet over the whole trial, noting that only one PL level tested for MarPL and fluid soy lecithin matched that provided by KM12, so optimal dose was not evaluated. The choline level was formulated to be the same for the control and VegPL diets with KM12 and MarPL providing additional choline to these diets in the form of phosphatidylcholine (PC). Formulated choline levels for the control diet and fluid VegPL diets were in excess of the NRC 2011 requirements for salmonids and in excess of the lowest choline level used by Hansen et al. in 2020 with no growth differences observed (1340 to 4020 mg choline/kg diet dose response trial for 456g initial weight salmon) [20]. Lipid accumulation in the gut was reduced for these salmon (456 g initial weight) at increased choline levels [20]. Effects of increased choline with KM inclusion cannot be ruled out and further research would be needed to separate choline from PL effects for these smaller pre-transfer salmon (74 to 158 g fish weight) that were fed lower fat pre-transfer diets (22–24% fat) than during the seawater growth with choline requirements for reducing the lipid accumulation in the intestine, potentially dependent on dietary fat level [20]. Higher growth was generally observed for PL provided by KO over soy lecithin at various PL doses for the first-feeding stage of salmon, but this growth trend was not consistent at various PL doses over the whole trial from the first-feeding to smolt [19]. PL from KO was indicated to be more effective than fluid soy lecithin for reducing intestinal steatosis in smaller salmon (2.5 g salmon, but no steatosis observed across diets for 10–20 g salmon) and low level of vertebral deformities [19]. Marine PL sources (FM and KO) were also compared against soy lecithin at a similar ca. 3.5% PL of diet level for the first-feeding Atlantic salmon (0.14 g initial weight) with these PL sources, giving similar growth to ca. 2.4 g final fish weight with no conclusive mortality or intestinal histology differences between PL sources but these parameters were generally improved for the PL source diets with higher PL compared to the control diets with lower PL. An uncertain observation of higher average growth was indicated for the marine PL sources over soy lecithin at intermediate weighing for salmon at ca. 0.6 g [24]. Effects of PL cannot be isolated from KM but the increased growth for KM12 over the VegPL diet in the pre-transfer phase may be due to PL, choline, water soluble nitrogenous components, etc., noting that there was also an indicated trend for decreased growth of VegPL versus the control diet in the pre-transfer phase.Addition of KM did not give a clear increase in feed intake compared to the control diet and there was an indicated trend of decreased feed intake for the MarPL and VegPL diets, but strong conclusions cannot be made due to the high variability. Feed intake can only be measured on a tank basis, so it was not possible to estimate feed intake of fish with different pre-transfer freshwater feeding histories in a common tank that were fed the same diet in the seawater phase.A trend for decreased hepatosomatic index (HSI) was indicated with increased KM inclusion and for the 12% KM diet versus the other PL sources added to provide the same PL level in the pre-transfer phase, but the effect of KM on decreasing HSI was not carried over into the seawater phase with fish that were fed the same diet in a common tank (Figure 7, Figure 8, Figure 9 and Figure 10). There was no difference in the liver lipid droplet accumulation based on histology (normal scores only) for salmon that were fed the diets containing different PL sources at the end of the freshwater pre-transfer period. The lower HSI in KM12 could be due to the positive effects from krill PL (and choline) on the lipid transport and deposition in organs, with this effect of feeding 12% KM to Atlantic salmon documented by [17] with less pale livers and reduced liver fat. The authors further supported this observation with a significantly higher expression of the cadherin 13 (Chd) gene in the 12% KM group associated with circulating levels of the adipocyte-secreted protein adiponectin that has potential anti-inflammatory effects and plays an important role in metabolic regulation and is associated with the fatty liver index in humans [25]. However, Chd expression was not studied in the current study, and hence, further studies are warranted to explore the association between Chd expression, his, and absolute fat accumulation in the liver in salmon. Increased choline, which KM provided in this trial, was shown to reduce fat accumulation in the intestine of Atlantic salmon [20]. Choline supplementation was also indicated to reduce HSI in Atlantic salmon, but this was not reflected in lower liver fat or histological vacuolization, noting that there are variable trends of dietary choline deficiency on the liver fat level of fish reported in the literature [26]. PL from KO was indicated to be more effective than fluid soy lecithin for reducing intestinal steatosis in smaller salmon (2.5 g salmon but no steatosis observed across diets for 10–20 g salmon). Further studies are required to associate higher liver fat with welfare in salmon.Gills are one of the most vital organs of fish, due to their function in respiration, osmoregulation, excretion of nitrogenous waste, pH regulation, and hormone production [27]. Gill health has become one of the most significant health and welfare challenges in the salmon aquaculture industry in Norway, Scotland, and Ireland [28,29,30]. The gill disorders are generally complex and multifactorial and are related to both biological factors, such as parasites and pathogens, handling stress, treatments, or due to the environmental factors, such as temperature, salinity, algal blooms, etc. Hence, the gill diseases are challenging to prevent and control and lead to high mortality, reduced production performance, and impaired fish welfare, cumulating in huge economic losses [31]. There were no differences reported for histological parameters investigated except in the presence of ectopic epithelial cells containing mucus in the lamina propria in the hindgut (potential inflammatory marker) of salmon (grown from 2.3 to 3.9 kg in sea cages) that were fed 15% fishmeal diet but not for 12% KM of diet in a 5% fishmeal diet, which may suggest anti-inflammatory effects of KM [17]. KM provides astaxanthin (166 mg/kg in the KM used for the present study) to the diet as a natural antioxidant with potential anti-inflammatory properties [32]. KM and MarPL also provide EPA + DHA attached to PL, which may affect bioavailability of EPA + DHA for use in cell membranes and inflammatory response [33] but this is not documented in fish. In the current study, there was decreased probability for very mild to mild gill lamella inflammation and hyperplasia scores indicated in salmon that were fed 12% KM compared to the soy lecithin and marine PL diets but gill histology for salmon that were fed the 12% KM diet was similar to the control diet without KM (Figure 5).5. ConclusionsOverall, increased KM tended to increase growth (high variability), whereas the VegPL diet tended to decrease growth compared to the control diet in the FW pre-transfer phase. The positive growth trend indicated for KM fed pre-transfer was not carried over into the seawater phase for fish fed the same diet. A minor positive trend in gill health (lamella inflammation and hyperplasia histology scores) was indicated for the 12% KM and Control diets compared with the VegPL and MarPL diets in the FW pre-transfer phase. Hepatosomatic index tended to decrease with KM fed in the pre-transfer phase, noting that all livers evaluated by histology were considered normal for lipid droplet accumulation. Only one VegPL and MarPL dose was tested, so dose effect of these PL sources and comparison with krill oil to better isolate the PL effect from other nutrients in KM as well as a post-transfer feeding comparison of these PL sources could be areas to research further in transfer diets for salmon.

animals : an open access journal from mdpi

10.3390/ani11102813

PMC8532665

Nutritional manipulation of livestock diets has long been considered asa tool for the control of infectious diseases. The objective of this literature review is to provide a comprehensive and critical view of the studies that have investigated the interactions between synthetic amino acids supplementation and immune response against infectious diseases in livestock. We performed a literature search in two main databases, PubMed and Web of Science. Based on our criteria for eligibility of the research articles, we selected 58 studies. Most of the studies selected focus on poultry and three amino acids that are all associated with a significant improvement in host response: methionine, threonine and arginine. The most described immune mechanisms associated with synthetic amino acids supplementation were directed against intracellular pathogens. We highlight the need for more analytical studies using these three amino acids, particularly in livestock other than poultry, and their use with other types of pathogens.

Infectious diseases represent one of the most critical threats to animal production worldwide. Due to the rise of pathogen resistance and consumer concern about chemical-free and environmentally friendly productions, the use of antimicrobials drugs is no longer desirable. The close relationship between nutrition and infection has led to numerous studies about livestock. The impact of feeding strategies, including synthetic amino acid supplementation, on host response to various infections has been investigated in different livestock animals. This systematic review provides a synthesis of the experimental studies on the interactions between synthetic amino acid supplementation and immune response to infectious diseases in livestock. Following PRISMA guidelines, quantitative research was conducted using two literature databases, PubMed and Web of Science. The eligibility criteria for the research articles were: (1) the host is a livestock animal; (2) the supplementation with at least one synthetic amino acid; (3) at least one mediator of immunity is measured; (4) at least one production trait is measured. Data were extracted from 58 selected studies. Articles on poultry were the most numerous; few contained experiments using ruminants and pigs. Most of the authors hypothesized that synthetic amino acid supplementation would particularly improve the animals’ immune response against intracellular pathogens. An increase in T and natural killer lymphocytes and macrophages activation, intracellular redox state, lymphocytes proliferation and antibodies production were the most described immune mechanisms associated with synthetic amino acid supplementation. Most of the selected studies focused on three amino acids (methionine, threonine and arginine), all of which are associated with a significant improvement of the host immune response. The use of synthetic amino acid supplementation appears as an encouraging perspective for livestock infectious disease management, and research must concentrate on more analytical studies using these three amino acids.

1. IntroductionInfectious diseases are one of the most significant threats to livestock farming worldwide. These diseases have important economic impacts and can lead to food insecurity by increasing morbidity and mortality, reducing market value and productivity [1,2]. The primary strategy for reducing the spread of pathogens on farms relies on the use of chemical drugs (e.g., antibiotics and anthelmintics). However, this strategy involving pharmaceutical treatments of livestock as the sole method of infectious disease control is no longer desirable for sustainable production due to the rise in drug-resistant pathogens worldwide, together with concerns about the presence of drug residues in edible animal products and as contaminants in the environment [3,4]. Moreover, antimicrobial-resistant pathogens represent a major concern impacting public health, particularly with the transfer of multi-resistant bacteria from animals to humans [5]. This issue also raises questions about animal welfare, given the close relationship between animal health and welfare. Consequently, the control of infectious diseases in livestock has to be included in a global scheme of management in which alternative control strategies have to be developed complementary to a parsimonious use of classical practices in accordance with agroecological concepts [6,7]. Indeed, in the context of global change, the aim is to sustain the economic and environmental viability of agricultural systems closely linked to their local ecosystem [8].A vaccination strategy to improve livestock immunization against specific pathogens is a vital tool for the efficient prevention of infectious diseases. However, vaccination is not available for all infectious diseases of importance. Some vaccines have a number of shortcomings with regard to safety, efficacy and/or user-friendliness that limit their effectiveness [9]. The genetic selection to reduce host susceptibility is a promising strategy, but it is feasible only in the long term and requires operational extension services and is consequently costly [10,11]. Therefore, short-term control strategies, such as tailored management of animal nutrition, are also necessary.In recent decades, knowledge regarding livestock nutrition has considerably evolved in terms of improvements in feed efficiency, whatever the production considered. Most of the time, these nutritional strategies to maximize production performances were developed in healthy, un-challenged animals. However, it has long been shown that optimal immune function is dependent on an adequate supply of protein, energy and micronutrients (e.g., vitamins and minerals), thus highlighting the close relationship between host nutrition and the immune system [12].In animal production, the key role of dietary protein and amino acids, which are the building blocks of proteins, on different functions of the immune system, has been investigated [13,14,15]. Indeed, numerous studies were conducted to find an optimal requirement of amino acids by different livestock animals, including birds [16], pigs [17] and ruminants [18], under numerous developmental, environmental, pathological and nutritional conditions. Amino acids play an important role in regulating immune responses, including the activation of lymphocytes, NK cells and macrophages; proliferation of lymphocytes; regulation of intracellular redox balance; gene expression; and production of cytokines [19]. Thus, the availability of specific dietary amino acids is essential for the control of infectious diseases, including viral infections. This review aimed to provide insight into the current knowledge regarding the effect of dietary supplementation with synthetic amino acids on the immune response to infectious diseases in livestock animals.2. Materials and MethodsThe study methodology was based on the guideline of “Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement” [20]. The literature search was conducted using the Web of Science and PubMed electronic databases. The search strategy is described in Figure 1 and the search terms used are in Table 1.*The use of truncation (*) allows a query to be made only on the root of the word. Furthermore, the reference lists from selected articles provide additional articles. Studies were closely evaluated and selected for inclusion if they were performed using livestock to measure the effect of dietary supplementation (above the National Research Council, NRC requirement). They had to have at least one particular synthetic amino acid on the expression of at least one mediator of the immune response following an immune challenge, and measure at least one production trait. Information extracted from each study included livestock types, immune challenge (live pathogens, microbe-associated-molecular-pattern, non-pathogenic antigens, non-specific immune stimulators), immune mediators and the impact on their expression (e.g., humoral and cellular mediators, lymphoid organ weight), dietary synthetic amino acid supplementation (quantity), basal diets, number of animals per group and production traits (Table 2). A total of 30 studies out of 58 showed a significant effect on the expression of at least one immune effector. The methodological quality of the included studies was evaluated qualitatively with a 9-point scoring system based on SYRCLE’s risk of bias tool (Figure 2) [21].3. Results3.1. Study Selection and CharacteristicsA total of 2322 results were provided by the Web of Science and PubMed databases (Figure 1 and Table 1). After excluding duplicates and screening titles and abstracts, 77 full-text research articles were further evaluated for relevance. Finally, 58 research articles were included in the qualitative synthesis. All included research articles were in English and published between 1969 and 2020. Six papers were not found using the bibliographical research strategy described but were located in the citations of the selected articles.3.2. Synthesized FindingsThe main objective of this systematic review was to synthesize the knowledge on the effect of dietary supplementation with synthetic amino acids on the immune response to infectious diseases in livestock. The bibliographical research was focused on all livestock, but only three types were found: poultry, ruminant and pig. The research articles on poultry were the most numerous compared to pigs and ruminants (48 out of 58 articles), and chicken was the main poultry studied. One of the major inclusion criteria was monitoring an immune challenge, including live bacteria, virus, parasite and their associated molecular pattern. The studies using non-specific immune stimulators (e.g., phytohemagglutinin, concanavalin A) were excluded. Among the 20 proteinogenic amino acids, only 9 were found in the selected research articles. The most studied amino acids were methionine (24 articles), followed by arginine (19 articles) and threonine (12). Figure 3 represents the amino acids used as supplementation, the type of livestock and the pathogens with their associated molecular pattern used in the 58 studies.3.3. Assessment of Risk of BiasThe results of the qualitative evaluation of the risk of bias are presented in Figure 2. An ethical statement about animal treatment in the research was not included in 25% of the studies. Most of these research articles were published before this statement was mandatory. An adequate allocation sequence generation was stated in 50% of the studies, while 24% did not use random housing and 94% reported using similar groups at baseline. However, allocation concealment was not performed in 87% of the studies. Random outcome assessment was stated in 80% of the studies, but the outcome assessment was not blind in 57% of the studies. An incomplete outcome was adequately addressed in 92% of the studies, and other sources of bias were described in 98% of the studies.4. DiscussionThe fine-tuning of nutritional strategies to support optimal function of the immune system in livestock is of great interest for the development of sustainable management in animal health. The immune response against invading pathogens is expensive in terms of proteins and calories because of the energy and protein requirements of immune cells for the synthesis of immune mediators and repairing damaged tissue [52]. Moreover, several vitamins (e.g., A, B6, B12, C, D, E and folate) and trace elements (including zinc, iron, selenium, magnesium and copper) play important and complementary roles in the development of both the innate and the adaptive immune systems. A short-term and easy-to-implement strategy consists of improving immune functions through animal nutrition. In the available literature, dietary supplementation with synthetic amino acids appears as a pertinent lever for preserving and regulating the immune response against pathogens.4.1. Role of Amino Acids in Immune Response4.1.1. Sulphur-Containing Amino Acid (SAA)Methionine and cysteine are the two principal sulphur-containing amino acids used as a substrate for proteins biosynthesis [53]. These two amino acids are also precursors of numerous metabolites implicated in various physiological functions including the immune system [16,54,55]. Methionine is a methyl group donor participating in both transmethylation and remethylation pathways, including the methylation of DNA and proteins and the regulation of gene expression [22]. The implication of methionine in the cellular immune response and the humoral immune response has been suggested in numerous studies [56].In turkeys and broilers, it has been shown that higher dietary methionine levels (from 0.60% and 0.90% in turkeys and broilers, respectively) increased the level of methionine in the peripheral blood, and was correlated with higher levels of peripheral leucocytes and IgG after vaccine challenges [22,31,32]. In addition, increased levels of CD4+ and CD8+ T cell subpopulations and IgM+ B subpopulation in immune organs, such as the thymus and the bursa of Fabricius, were observed in turkeys [31,32]. Higher antibody levels were also observed in broilers supplemented with a combination of 0.65% methionine and 0.13% of choline [24]. Indeed, methionine is a substrate for choline, and thus acetylcholine and phosphatidylcholine synthesis play a central role in numerous metabolic pathways including leucocyte metabolism. In contrast, despite the use of rumen-protected dietary methionine, no effect on the immune response after a vaccine challenge against the herpesvirus was observed in beef heifers [57]. Cysteine is involved in the formation of interchain and intrachain disulphide bonds of proteins. In vitro studies showed that cysteine and other cysteine derivatives can modulate lymphocyte and macrophage functions [58]. In broiler chickens supplemented with 0.65% of cysteine, an increase in immune and the inflammatory responses was shown after an Escherichia coli LPS (lipopolysaccharide) injection, notably by the production of several cytokines (IL-1, IL-6 and TNF-α) by macrophages [28]. The supplementation with a combination of methionine and cysteine (from 0.8%) was also associated with an increase in the level of anti-Eimeria IgA in broilers [30].Few studies have investigated the role of cysteine in the improvement in immune responses. The bibliographic analysis performed here showed only one article in which the impact of cysteine supplementation on the immune response of sheep against Haemonchus contortus and Trichostrongylus colubriformis was studied. Cysteine was reported to increase blood eosinophil counts and globule leucocytes in the abomasal mucosa following experimental infection [47]. Further studies need to be conducted in other types of livestock to complete investigations on metabolic cysteine pathways in response to infectious diseases. Despite the great advances in our knowledge of sulfur-containing amino acids, there are important areas where further work is required. Cysteine appears to influence certain aspects of immunocompetency in sheep, although the exact role of cysteine in the relationship between wool production and parasite susceptibility requires further elucidation.4.1.2. Amino Acid AmideThe interaction between dietary glutamine and immunity has been well studied in mice [59]. Glutamine is implicated in purine and pyrimidine biosynthesis in tissues of the immune system [60], in lymphocyte proliferation [61] and in cytokine production [62]. In dairy livestock, glutamine is usually considered a good feed additive to improve the amino acids profile of milk due to the presence of casein in glutamine residues [19]. Dietary supplementation of broiler chickens with glutamine (3%), arginine (2%) and threonine (2%) was associated notably with an increased number of goblet cells, a lower level of mucosal IgA and a reduction in the thymus, suggesting an improved immune response against a co-infection with Eimeria and E. coli [41]. Similar to the results of Bartell and Bartal (2007) [63], this higher level of dietary glutamine in broiler was associated with lower levels of mucosal IgA, probably due to the rapid and efficient immune response against the parasitic infection.4.1.3. Basic Amino AcidThe basic amino acids, lysine and arginine, play important roles in membrane protein activity and the actions of antimicrobial, toxin and cell-penetrating peptides. In poultry, due to the absence of a functional urea cycle, arginine is an essential amino acid [64]. Many studies demonstrated the importance of arginine supplementation above the NRC recommendations to support growth performance and improve the host immune response against avian infectious diseases [23,39]. In these studies, poultry challenged with viruses such as infectious bronchitis virus (IBV), Newcastle disease virus (NDV) or infectious bursal disease (IBD) were able to establish an immune response with no impact on the growth rate.Arginine metabolism has been described as a key regulator of innate and adaptive immunity. Cells of the myeloid lineage, such as macrophages and dendritic cells, could modulate the immune response by regulating the expression of two enzymes: NOS (nitric oxide synthase) and arginase [65]. NOS uses arginine as a substrate to produce NO (nitric oxide), a key molecule involved both directly and indirectly in host immune response [66]. In broiler chickens, it has been shown that dietary arginine (from 0.45% above the NRC recommendation) may modulate macrophage phagocytosis by induction of the expression of key cytokines, particularly IL-1, IL-2, INF-γ and TNFα [36,38]. In piglets challenged with Escherichia coli LPS (lipopolysaccharide), arginine supplementation (0.5% above the NRC recommendation) increased the infiltration of ileal mucosa by IgA-secreting cells, CD8+ and CD4+ T cells [43]. Many pieces of evidence demonstrated that a dietary deficiency in lysine decreased the synthesis rate of proteins (including cytokines) and lymphocytes proliferation. The impact of dietary supplementation with lysine on the immune response is always studied in association with another amino acid. In broiler chickens, dietary supplementation with lysine and methionine improved antibody production against Newcastle and the Gumboro diseases [49]. This positive effect on the immune response, together with performance production, was also observed in piglets challenged with Escherichia coli K88 [50]. An average optimal standardized ileal digestible tryptophan: lysine ratio of 21% optimized the performance of the challenged piglets and increased the expression of IL-10 in the ileum.4.1.4. Branched-Chain Amino Acid (BCAA)Branched-chain amino acids (BCAA), which include leucine, isoleucine and valine, contribute notably to the synthesis of glutamine in the skeletal muscle and are essential for lymphocytes’ proliferation in response to immune stimulation [67]. Thus, dietary BCAA restriction impairs several aspects of the immune function and increases the susceptibility to different pathogens [68]. In broiler chickens, it has been shown that valine, which constitutes approximately 18% of muscle myofibrillar protein, could be associated with a decrease in the weight of lymphoid organs [69,70]. Furthermore, studies have demonstrated that birds subjected to either killed or live viruses and supplemented with a high level of dietary valine showed increased antibody production. In weaned piglets, isoleucine and valine (0.19 and 0.27% respectively) supplementation to protein-restricted diets protected villous morphology and increased levels of intestinal immunoglobulins [71]. Piglets are particularly sensitive to rotavirus infection that leads to severe gastroenteritis [72]. Mao et al. [51] showed that adding isoleucine in pig diet increases the production of IgA, IgG, several cytokines (IL-1β, IFNβ, IFNγ, TNF-α and IL-10) and β-defensins in serum, ileum and/or mesenteric lymph [51].4.1.5. Other Essential Amino Acids (OAA)Threonine is implicated in many biological functions, such as, body protein synthesis, collagen, antibody and uric production or pancreatic enzymes [73]. In broilers, threonine is considered an essential amino acid because of its impact on growth performance, gut health, carcass traits and immune functions [74]. Dietary threonine supplementation, above the requirements, promotes the growth of immune organs and stimulates the synthesis of immunoglobulins and antibodies in broiler chickens under immune stress [45,46,75]. Indeed, it has been suggested that, in broiler chickens, threonine would be especially implicated in the regulation of the immune system response by controlling the gut microbial population and the production of IgA and IgG [44]. In pigs, during the post-weaning period, due to the high susceptibly to intestinal bacterial infection, a higher requirement of threonine is recommended. Interestingly, threonine is the most abundant amino acid in mucin proteins produced by the intestinal mucosa, [76]. In addition, in a recent study, a high level of dietary threonine (0.9%) increased the level of IgA production against an enterotoxigenic E. coli K88 (ETEC) [77]. In broiler chickens, following avian tuberculin injection, dietary supplementation with tryptophan (0.3 and 0.5%) improved total oxidant status and overall antibody response, together with cellular immunity, [48]. The antibody response and the IgG increased when the tryptophan level exceeded the NRC requirements. The combination of tryptophan and arginine from two times the NRC levels in starter, grower and finisher diets of broiler chickens has an efficient immunomodulatory impact on virus infections [33]. Indeed, this treatment improved both the innate (IFNα) and humoral (IgG) immune responses against the infectious bursal disease virus (IBDV).Most of the selected studies focused on three amino acids: methionine, threonine and arginine, which were all associated with a positive impact on the host immune response. The mechanisms of immunity improved by dietary amino acids in livestock animals involved both Th1 and Th2 responses, and Th17 to a lesser extent. We have summarized the effects of the different amino acids studied in the selected research articles in a schematic representation of the immune response (Figure 4). The most important effects of amino acids on immunity were mostly observed at a high level of supplementation. The majority of the studies were on chicken, and few studies investigated the effect of amino acid supplementation on ruminant immune response. This may be due to the fact that ruminant physiology limits amino acid availability. To overcome this limitation, the utilization of rumen-protected amino acids is of interest. Therefore, there is a need for more studies in pigs and ruminants with live pathogens to increase our knowledge on the effect of dietary supplementation with synthetic amino acids on the immune response against infection.

animals : an open access journal from mdpi

10.3390/ani12030353

PMC8833611

Horses commonly develop low blood oxygen levels during anesthesia, especially when they are placed on their backs. This study investigated whether a 15° head-up tilt, in a homogenous group of anesthetized horses positioned on their backs, would result in better blood oxygen levels as compared to no tilt. The results showed significantly greater blood oxygen levels with tilt compared to no tilt in five out of six horses tested. In one horse the effect was the opposite. The concurrent effect on cardiovascular function remains to be tested in detail. Further studies are needed to confirm these findings in a larger group of horses and to determine the effects on blood pressure and treatment options.

Lower than expected arterial oxygen tension (PaO2) continues to be an unresolved problem in equine anesthesia. The aim of this randomized, crossover, and prospective study using six adult horses is to determine if a 15° reverse Trendelenburg position (RTP) increases PaO2 during inhalation anesthesia. Under constant-dose isoflurane anesthesia, dorsally recumbent horses were positioned either horizontally (HP) or in a 15° RTP for 2 h. Lungs were mechanically ventilated (15 mL/kg, 6 breaths/min). Arterial carbon dioxide tension (PaCO2), PaO2, inspired oxygen fraction (FiO2), and end-tidal carbon dioxide tension (EtCO2) were determined every 30 min during anesthesia. Indices of dead-space ventilation (Vd/Vt), oxygenation (P–F ratio), and perfusion (F–shunt) were calculated. Dobutamine and phenylephrine were used to support mean arterial pressure (MAP). Data are presented as median and range. In one horse, which was deemed an outlier due to its thoracic dimensions and body conformation, indices of oxygenation worsened in RTP compared to HP (median PaO2 438 vs. 568 mmHg; P–F ratio 454 vs. 586 mmHg, and F–shunt 13.0 vs. 5.7 mmHg). This horse was excluded from calculations. In the remaining five horses they were significantly better with RTP compared to HP. Results in remaining five horses showed that PaO2 (502, 467–575 vs. 437, 395–445 mmHg), P-F ratio (518, 484–598 vs. 455, 407–458 mmHg), and F-shunt (10.1, 4.2–11.7 vs. 14.2, 13.8–16.0 mmHg) were significantly different between RTP and HP (p = 0.03). Other variables were not significantly different. In conclusion, the 15° RTP resulted in better oxygenation than HP in dorsally recumbent, isoflurane-anesthetized horses, although worsening of oxygenation may occur in individual horses. A study detailing the cardiovascular consequences of RTP is necessary before it can be recommended for clinical practice.

1. IntroductionRespiratory function impairment, or less than expected arterial partial pressure of oxygen (PaO2), has long been recognized and continues to be very common in equine anesthesia [1,2,3,4,5,6,7,8]. Horses in dorsal recumbency appear to be particularly prone to impaired oxygenation, although it can also occur in lateral recumbency [5,9]. The problem arises primarily from increased intrapulmonary right-to-left shunting, secondary to the development of alveolar atelectasis in dependent lung areas [5,10]. Compression atelectasis appears to be the most common [1,5,10], although absorption atelectasis might also occur in some situations [6]. Physical features such as body mass, chest dimensions, and abdominal shape all have been shown to positively correlate with the degree of impairment [1,10,11,12]. Several physical and pharmacological interventions aimed at mitigating impaired oxygenation have been studied in anesthetized horses. These include reverse Trendelenburg position (RTP) [13,14], positive end-expired pressure (PEEP) combined with repeated alveolar recruitment maneuvers (RM) [15,16], continuous positive airway pressure [17], intravenous and inhaled β-adrenergic receptor agonists [18,19], inhaled nitric oxide [10,20,21], acepromazine [7], and reducing FiO2 [22,23,24,25]. Overall, physical methods appear to improve PaO2 more consistently than pharmacological methods, which seems logical given the prominent contribution of compression atelectasis to impaired oxygenation [5,10]. In morbidly obese human patients, the reverse Trendelenburg position (RTP; i.e., head-up tilt) was initially advocated to facilitate upper abdominal surgery [26]. A 30° RTP was subsequently shown to be efficacious in improving pulmonary gas exchange and total respiratory system compliance [27,28,29]. In isoflurane-anesthetized horses [13,14] and sevoflurane-anesthetized steers [30] positioned in dorsal recumbency, RTP at angles between 5° and 10° did not improve arterial oxygenation compared to the horizontal position (HP). Another study comparing 7° RTP with a 7° head-down (i.e., Trendelenburg) position for 90 min in dorsally recumbent isoflurane-anesthetized horses breathing 0.85 FiO2 found that the RTP did not improve arterial oxygenation compared to baseline. However, gas exchange was better maintained compared to the head-down position [31]. Finally, a recent clinical study employing 5° RTP did not show any significant improvement in arterial oxygen tension in isoflurane-anesthetized horses positioned in dorsal or lateral recumbency [13]. Thus, these veterinary studies using relatively conservative RTP angles compared to that commonly applied in humans do not appear to support the use of RTP to improve gas exchange compared to the HP during inhalation anesthesia in the horse. However, it remains unclear if steeper RTP angles would result in improved arterial oxygen tension. The study aimed to determine the effect of a 15° RTP on arterial oxygen tension in dorsally recumbent, mechanically ventilated, and isoflurane-anesthetized horses. It was hypothesized that arterial oxygen tension would be significantly better with a 15° RTP as compared to the HP. Effects on the alveolar dead-space ventilation (Vd/Vt) and the oxygen content-based index (F–shunt) were evaluated as secondary goals. Results indicated that a 15° RTP resulted in better PaO2 than HP in dorsally recumbent, isoflurane-anesthetized horses, but may lead to negative cardiovascular consequences that warrant further investigations.2. Materials and Methods2.1. Animals and Study DesignA total of 6 horses (1 mare and 5 geldings) aged 8.5 ± 3 years (range 5–13 years) of several breeds (1 Arabian, 3 Quarter Horses, and 2 mixed breeds) and weighing 462 ± 50 kg (range 397–536 kg) were studied in a randomized (randomization.com; accessed on 3 July 2017) crossover design with two treatments, separated by at least one-week washout in between. Horses were deemed healthy based on physical exam, complete blood cell counts, and serum biochemistry analyses. 2.2. AnesthesiaFood was withheld for approximately 12 h and water was available ad libitum prior to anesthesia. Thoracic dimensions were obtained as previously described [12] and horses were classified subjectively as round-bellied or flat-bellied, according to their abdominal contour as described by Moens and co-workers [11]. A 14-gauge intravenous catheter (Angiocath)a was aseptically placed in one jugular vein for drug and fluid administration. Horses were sedated with 5 μg/kg dexmedetomidine (Dexdomitor; Orion Pharma, Orion Corporation, Espoo, Finland) or 1 mg/kg xylazine (AnaSed; Akorn, Inc., Lake Forest, IL, USA) as part of a separate study [32]. After 5–7 min, anesthesia was induced with 0.06 mg/kg midazolam (Midazolam Injection USP; Hospira Inc., IL, USA) and 2.2 mg/kg ketamine (KetaVed; Vedco Inc., Saint Joseph, MO, USA) behind a swing gate and inside of a padded stall. The horse’s trachea was orally intubated with a 26 mm internal diameter cuffed endotracheal tube. The horse was subsequently hoisted and positioned in dorsal recumbency on a padded table (see positioning description below) and connected to a large animal circle breathing system and anesthesia machine (Mallard; Mallard Medical Inc., Redding, CA, USA). General anesthesia was maintained with an end-tidal isoflurane (IsoFlo; Abbott Animal Health, Abbot Park, IL, USA) concentration of 1.2 times its minimum alveolar concentration (MAC), considered to be 1.31% [33], with an oxygen flow rate of at least 6l/min. The target end-tidal isoflurane concentration was achieved in all horses within 10 min after connecting the endotracheal tube to the breathing circuit of the anesthesia machine. Horses received fluid therapy with Hartmann’s solution (Vetivex; Dechra Veterinary Products, Overland Park, KS, USA) at 3 mL/kg/h.2.3. Body Positioning Protocols during AnesthesiaThe anesthetized horse was hoisted onto a surgical operating table capable of hydraulic tilting, which was previously adjusted to either the HP or with a 15° RTP. The angle of the table was measured using a 25.4 cm multi-function standard digital level (Husky; The Home Depot, Atlanta, GA, USA) that was positioned on a fixed solid part of the table. To prevent the possibility of horses sliding off the table, a rope padded with towels was positioned around the horse’s hindquarters, approximately at the level of the ischiatic tuberosity, with each end secured to the head of the table (Figure 1). Horses were maintained in the designated position for 2 h.2.4. Monitoring and Support during AnesthesiaHorses were monitored with base-apex electrocardiography. A 20-gauge, 4.4 cm catheter (Insight)a was aseptically placed in the facial artery and connected via a fluid-filled regular 83.3 cm extension set to a calibrated disposable pressure transducer (BD DTXTM disposable pressure transducers; Becton Dickinson Infusion Therapy Systems, Sandy, UT, USA) that was zeroed and positioned at the level of the heart (i.e., scapulohumeral joint) for invasive blood pressure measurement. Dobutamine (DOBUTamine injection USP; Hospira Inc., Lake Forest, IL, USA) was administered as a continuous infusion as needed to maintain mean arterial pressure (MAP) between 70 and 80 mmHg. A dobutamine dose was started at 0.5 μg/kg/min and then doubled every 2–3 min until MAP was within the desired range. If signs of dobutamine-induced side effects occurred with a given dose (tachycardia or cardiac arrhythmias), its dose was reduced to the previous infusion rate and phenylephrine (Neo-Synephrine® HCl; Hospira Inc., Lake Forest, IL, USA) was administered for additional cardiovascular support as 1–2 μg/kg boluses to maintain MAP within the desired range. Approximately 2–3 min was allowed after each bolus before the next one was administered if needed. The amounts of dobutamine and phenylephrine used in both treatment groups were recorded.The horses’ lungs were mechanically ventilated (Mallard; Mallard Medical Inc., Redding, CA, USA) at a rate of 6 breaths/minute and tidal volume 15 mL/kg, rounded to the nearest liter, with an inspiratory time of 2 s. Peak inspiratory pressures, as measured by the pressure gauge located in the breathing circuit of the anesthesia machine, were recorded. Airway gas samples were obtained continuously throughout anesthesia using a calibrated multi-parameter monitor (DPM 7; Mindray DS USA Inc., Mahway, NJ, USA) for side stream monitoring of inspired and end-tidal concentrations of oxygen, carbon dioxide, and isoflurane. Arterial blood samples (3 mL) were collected at 30, 60, 90, and 120 min of anesthesia for the determination of blood gas tensions, pH, and lactate concentrations using a calibrated portable blood gas analyzer (i-STAT CG4 test cartridges; Abbott Point of Care Inc, NJ, USA). Rectal temperature was monitored with a digital thermometer.2.5. Recovery from AnesthesiaAt the end of the experiment, the urinary bladders were emptied via catheterization, monitoring equipment was removed, the orotracheal tube was disconnected from the breathing circuit, and the horses were transferred to padded recovery stalls for unassisted recovery as part of a separate study [32].2.6. CalculationsThe PaO2 and the ratio between arterial oxygen tension and a fraction of inspired oxygen (P–F ratio) were used to assess oxygenation. The P-F ratio was determined by dividing the arterial oxygen tension (PaO2) by the inspired oxygen fraction (FiO2). The Vd/Vt was calculated via the Enghoff modification of the Bohr equation where Vd/Vt = ([PaCO2 − PETCO2]/PaCO2) × 100 [34]. The F-shunt was calculated using the equation ([Cc′O2 − CaO2]/[Cc′O2 − CaO2 + 3.5 mL/dl]) × 100 [13,14,35,36], where Cc′O2 and CaO2 were, respectively, the capillary and arterial oxygen contents. The value 3.5 is the arteriovenous oxygen content difference in mechanically ventilated humans [37], which closely approximates that of isoflurane-anesthetized, dorsally recumbent horses [23,38]. The CaO2 was calculated with the equation (1.38 × Hb × SaO2) + (PaO2 × 0.003), where 1.38 is the oxyphoric capacity of equine hemoglobin [39], Hb is the hemoglobin concentration measured in the awake horse before the experiment, SaO2 is the arterial hemoglobin saturation with oxygen determined via blood gas analysis, and 0.003 is the oxygen solubility in blood. The Cc′O2 was calculated similarly, except that the pulmonary end-capillary partial pressure of oxygen was assumed to be equal to the alveolar oxygen partial pressure (PAO2), which was estimated using the alveolar gas equation in which PAO2 = (PB × FiO2-PH2O) − (PaCO2/0.8). The PB is the local barometric pressure on the day and time of the experiment (range 761–766 mmHg), PH2O is the partial pressure of water (47 mmHg), and 0.8 is the carbon dioxide production to oxygen consumption ratio. Total respiratory system plus breathing system compliance (Ctot) was calculated as peak volume/pressure and was corrected for body weight (mL/cmH2O/kg). 2.7. Statistical AnalysisStatistical analyses were performed using commercial software (GraphPad Prism Version 7.0d for MAC OS, GraphPad Software, Inc., San Diego, CA, USA). The data were assessed for normality via visual inspection of QQ plots and the Shapiro–Wilk normality test. Indices of oxygenation (PaO2 and P–F ratio) and indices of ventilation and perfusion (PaCO2, PETCO2, Vd/Vt, F–shunt) over time were analyzed by means of two-way repeated measures ANOVA followed by Holm–Sidak post-test. The median PaO2, P–F ratio, F-shunt, and Ctot over the 2 h of anesthesia were compared between groups using Wilcoxon matched–pairs signed–rank test. Spearman correlation was used to determine the relationship between changes in PaO2 and Ctot. Statistical significance was set at p < 0.05. Data are shown as mean ± SD or median (range).3. ResultsAll horses completed the study and recovered uneventfully. Demographic information for each horse is presented in Table 1. One horse (#5) was excluded from statistical analysis due to abdominal contour and thoracic length that was markedly different from the remaining five horses.3.1. Respiratory VariablesIndices of oxygenation, ventilation, and perfusion are summarized in Table 2. Statistically significant differences were detected between HP and RTP for PaO2, P-F ratio, and F-shunt at several time points during anesthesia. The treatment effect accounted for approximately half of the total variance (after adjusting for matching) with <1% chance of randomly observing the measured effect in an experiment of this size. The PaCO2 and Vd/Vt were not significantly different between RTP and HP. The Vd/Vt increased over time during RTP such that it was significantly greater at 90 and 120 min compared to 30 min of anesthesia. The 2 h median PaO2 of each horse is shown in Figure 2a, and the difference between group medians (437 mmHg and 502 mmHg during HP and RTP, respectively) was statistically significant (p = 0.03). The median P-F ratio was significantly (p = 0.03) greater during RTP (518 mmHg, range 484–598) than HP (455 mmHg, range 407–458). F-shunt was significantly lower during RTP (10.1%, range 4.2–11.7) than HP (14.2%, range 13.8–16.0). Individual Ctot values for each horse during HP (median 0.69 mL/cmH2O/kg) and RTP (median 0.75 mL/cmH2O/kg) are shown in Figure 2b, and the group medians were not significantly different. Spearman correlation between changes in PaO2 and Ctot for all horses (n = 6) was 0.64 (p = 0.09), and the results are shown in Figure 2c.3.2. Clinical Cardiovascular and Other VariablesMedian rectal temperature, dobutamine, and phenylephrine use during HP and RTP, as well as the body position during the first anesthesia for each horse, are listed in Table 3. Clinical cardiovascular parameters (HR, MAP), rectal temperature, and requirement of dobutamine and phenylephrine were not compared statistically between groups. The heart rates and mean arterial pressures for individual horses are shown in Figure 3. Cardiac dysrhythmias were observed during both HP and RTP. During HP, dysrhythmias included tachycardia (heart rate > 50 beats/min) in three horses (horses # 1, 3 and 4), atrial premature complexes in one horse (horse #2), and sinus bradycardia (heart rate < 26 beats/min) followed by junctional escape rhythm, junctional escape-capture bigeminy, and atrial premature complexes in one horse (horse #6, shown in Figure 4). During RTP, dysrhythmias included tachycardia in four horses (horses # 2, 3, 4, and 5) and atrial premature complexes in one horse (horse #6).4. DiscussionThe results of the current study suggest that a 15° RTP can significantly improve arterial oxygen tension compared to the HP in dorsally recumbent, mechanically ventilated, isoflurane-anesthetized horses. However, worsening of arterial oxygenation can also happen, so its use in clinical cases should be very carefully monitored. The 15° RTP better preserved the V/Q relationship compared to HP by decreasing venous admixture, as indicated by a significantly lower F–shunt. These findings are in agreement with reports in mechanically ventilated morbidly obese humans at 30° RTP [27,28,29] but are in contrast to those obtained in mechanically ventilated horses at 5° RTP [13] or 7° RTP [14] and in spontaneously breathing steers at 5° and 10° RTP [30]. It is important to note that our results were obtained in a homogeneous group of round-bellied horses, a group of horses known to have lower PaO2, and a higher alveolar-to-arterial oxygen tension difference than flat-bellied horses in both dorsal and lateral recumbency [11]. Further studies are needed to confirm or refute our results in round-bellied horses given the relatively small sample size, as well as to determine the response of flat-bellied horses to RTP. Interestingly, the PaO2 decreased during RTP in relation to HP in the single horse classified as flat-bellied. It is likely that a proportion of the general equine population will not respond satisfactorily to RTP as is the case with approximately a third of morbidly obese human patients [29].The actual body mass may also affect the efficacy of RTP in improving PaO2. In humans, RTP resulted in significantly better PaO2 during anesthesia compared to HP in morbidly obese individuals [27,28] but not in non-obese individuals with similar body mass [40]. In horses, a non-significant trend was reported for PaO2 and P-F ratio, which were higher in 7° RTP than in HP in horses weighing between 300 and 599 kg but lower in those weighing 600 kg or more [14]. In morbidly obese human patients, 30° RTP improves PaO2 at least in part by increasing total respiratory system compliance [27,28,29]. In the current study, a similar trend was observed in four out of six horses, and a strong positive correlation between changes in PaO2, and Ctot was observed, albeit these were outside the pre-set statistical significance level. The markedly greater Ctot in the single flat-bellied (and long-chested) horse during HP, as compared to the remaining five round-bellied horses, was also notable. The Ctot values obtained in the study are likely not true physiologic values because they also include compliance of the breathing system. However, a between-group comparison is still relevant since the same exact equipment was used in all instances. In the current study, the 15° RTP was associated with an apparent greater negative cardiovascular effect compared to HP. With the caveat that the study was not designed to test the cardiovascular effects of 15° RTP and these data were not compared statistically, our findings appear to be in contrast with reports using smaller RTP angles in horses [13,14] and steers [30] as well greater angles in humans [27,28,40]. While positive pressure ventilation and the constant-dose isoflurane likely had some contribution to the hypotension observed in the current study, the difference between RTP and HP can be attributed to the 15° RTP given the study design with each horse receiving both treatments and thus serving as their own control. Dobutamine and phenylephrine were used as pharmacologic blood pressure support. Dobutamine is the preferred drug to treat hypotension in anesthetized horses [41] although it can cause sinus tachycardia if administered at rates greater than 3 to 5 μg/kg/min or in patients with inadequate intravascular volume [41,42,43]. Accordingly, sinus tachycardia associated with increasing infusion rates of dobutamine occurred in several horses of the current study, especially during RTP, which became a limiting factor for the administration of dobutamine. It is unlikely that the healthy horses of the study had inadequate intravascular volume. Instead, increases in regional vascular capacitance associated with the 15° RTP, as shown in humans [44,45], likely decreased preload and emulated a decreased intravascular volume. Thus, phenylephrine was employed for vasoconstriction [46] and its requirement was markedly greater during RTP than HP. The study suggests that the effect of different RTP angles on cardiovascular function in anesthetized horses, including the physiological basis and optimum therapeutic strategies, merits further investigation since hypotension is a significant risk factor for anesthetic morbidity and mortality in horses [8].Isoflurane causes dose-dependent decreases in MAP and skeletal muscle blood flow [43]. In the current study, the isoflurane dose was standardized at 1.2 × MAC to ensure the same intra-individual anesthetic depth for HP and RTP and to minimize the confounding effects of additional drugs possibly required to correct insufficient anesthetic depth. However, this approach could have resulted in different inter-individual anesthetic depths, possibly excessive in some cases given the absence of invasive procedures. The cardiovascular support with dobutamine and phenylephrine was effective in correcting hypotension. Inadequate blood flow to skeletal muscle, if it occurred, was not sufficient to result in clinical signs of post-anesthetic myopathy although subclinical myopathy cannot be ruled out since muscle enzymes were not determined. Phenylephrine does not change skeletal muscle blood flow in anesthetized horses, but dobutamine can be effective in mitigating anesthetic-induced skeletal muscle hypoperfusion and post-anesthetic myopathy [42,43,47,48]. In horizontally positioned isoflurane-anesthetized horses, phenylephrine significantly decreased intestinal blood flow, most profoundly at the colon [49]. Studies are needed to characterize the intestinal regional arterial and venous responses to the 15° RTP and the effects of phenylephrine in anesthetized horses.The junctional escape-capture bigeminy seen in one of the horses in the current study has not been previously reported in horses. Only two reports could be found in the veterinary literature, one in a domestic shorthair cat [50] and another in a Yorkshire Terrier dog [51]. In this horse, the junctional escape-capture bigeminy was not related to RTP since it occurred during HP. It coincided with starting dobutamine to correct hypotension in the presence of sinus bradycardia (heart rate of 20–25 beats/min). Upon starting dobutamine, occasional junctional escape beats were observed, followed by a junctional rhythm and then junctional escape-capture bigeminy. Bradycardia was the common finding among the two reported veterinary cases [50,51] and in this horse. In this horse, the escape-capture bigeminy subsided as the heart rate increased during anesthesia. This did not happen in the second anesthetic episode (RTP) in which bradycardia did not occur. In escape-capture bigeminy, a conducted complex can only follow an escape complex if the R-P interval of the retrograde P wave is sufficiently long (i.e., via the slow atrioventricular nodal pathway) to be conducted back to the ventricles via the fast atrioventricular nodal pathway. When the P wave is caused by retrograde conduction via the fast atrioventricular nodal pathway, anterograde atrioventricular nodal conduction is blocked, resulting in a blocked P wave [52], which was occasionally evident in the horse of the study. The study has several limitations. First, the exclusion of one horse from statistical comparisons resulted in a relatively small sample size. However, it also resulted in a very homogeneous group of horses regarding body shape (all round-bellied) and thoracic dimensions, which may have increased the chances of finding statistically significant differences. Despite the relatively small sample size, statistical analysis indicated that the treatment effect explained approximately half of the total variance, with a very small chance (<1%) of randomly obtaining the observed effect. It should be recognized nonetheless that our results are conditioned to a small homogeneous group of horses, which likely does not fully represent the wider equine population. Larger studies are needed to confirm or refute our findings. The use of xylazine and dexmedetomidine for pre-anesthetic medication could be considered another limitation. However, these α-2 adrenergic receptor agonists were used at approximately equipotent sedative doses, and the available literature suggests that there are no significant cardiopulmonary differences between them during isoflurane anesthesia [53] or total intravenous anesthesia with ketamine and midazolam [54] in horses. Another potential concern was the estimation of venous admixture via calculation of F-shunt instead of deriving it from a calculation based on mixed venous, capillary, and arterial oxygen contents, using the traditional Berggren shunt equation [55]. However, venous admixture can be accurately estimated by calculating F-shunt, as demonstrated during one- and two-lung ventilation in anesthetized sheep [35]. This approach has been utilized in anesthetized horses in several publications [13,14,36]. The arteriovenous oxygen content difference used in the F-shunt calculations originated from mechanically ventilated humans [37] but is similar to that of isoflurane-anesthetized, dorsally recumbent horses [23,38]. Further, the F-shunt strongly correlated with shunt fraction estimated by the Berggren shunt equation during different infusion rates of dobutamine in isoflurane-anesthetized horses [56]. The markedly higher phenylephrine requirement during RTP than HP could be considered another concern but research in isoflurane-anesthetized horses showed that phenylephrine did not affect PaO2 in horses positioned in HP [49]. Finally, the use of a point-of-care (POC) instrument to determine arterial blood gas tensions could be criticized. However, the instrument was validated against our laboratory’s reference values, is routinely verified for quality control, and all samples were analyzed immediately after collection using the same POC instrument. In addition, the POC instrument used in our study was previously found to have good agreement with a central analyzer [57]. 5. ConclusionsUnder the conditions of this study, a 15° RTP resulted in better oxygenation than HP in dorsally recumbent, isoflurane-anesthetized horses, although worsening of oxygenation may occur in individual horses. Further studies are warranted to determine the effect of RTP in horses with different body conformation (i.e., round-bellied vs. flat-bellied horses) and to characterize the cardiovascular responses to different RTP angles before it can be recommended for clinical practice.

animals : an open access journal from mdpi

10.3390/ani11113061

PMC8614451

Total dissolved gas (TDG) supersaturation severely threatens the survival of fish downstream of a dam in the Yangtze River due to flood discharge. However, most studies have been executed in the laboratory. Few works have evaluated the effects of TDG supersaturation on fish in natural rivers during periods of flood discharge. In the present study, we investigated the survival of two rare species (Procypris rabaudi and Myxocyprinus asiaticus) when subjected to TDG-supersaturated water at varied water depths in the natural river during periods of flood discharge. The results of this study showed that deeper water depths can increase the tolerance of juvenile Procypris rabaudi to TDG supersaturation in natural rivers during periods of flood discharge while it cannot improve the survival of juvenile Myxocyprinus asiaticus. Juvenile Procypris rabaudi were more vulnerable to TDG supersaturation than juvenile Myxocyprinus asiaticus. The study results can promote the protection of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus (or other rare species) and contribute to the improvement of reservoir operation practices in the Yangtze River.

Total dissolved gas (TDG) supersaturation, which can be caused by flood discharge, results in gas bubble disease (GBD) in fish and threatens their survival downstream of dams. TDG supersaturation has become a serious environmental problem in the Yangtze River. Few studies have evaluated the effect of TDG supersaturation on fish in natural rivers during periods of flood discharge. To estimate fish tolerance to TDG supersaturation under natural conditions, juvenile Myxocyprinus asiaticus and juvenile Procypris rabaudi were exposed to TDG-supersaturated water for 96 h at various depths (0–0.3 m, 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m) during periods of flood discharge of Dagangshan hydropower station. The results showed that juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus exhibited obvious GBD signs. An increase in exposure time decreased survival probability of the two species. Deeper water depths can increase the tolerance of juvenile Procypris rabaudi to TDG supersaturation in natural rivers during periods of flood discharge while it cannot improve the survival of juvenile Myxocyprinus asiaticus. Compared with juvenile Myxocyprinus asiaticus, juvenile Procypris rabaudi showed weaker tolerance of TDG supersaturation in shallow water, and juvenile Procypris rabaudi were more vulnerable to TDG supersaturation than juvenile Myxocyprinus asiaticus even if the TDG level (116%) was low.

1. IntroductionDissolved gas supersaturation is a condition that results from natural processes (e.g., water cascades, changes in water temperature, photosynthesis) and human activity (flood discharge) [1,2,3,4,5]. The excessive dissolved gas in water results in gas bubble disease (GBD) in fish and threatens their survival [6,7,8]. The signs of GBD and its mortality rate in fish, such as Esox lucius, Pomoxis nigromaculatus and Oncorhynchus tshawytscha, have been reported due to dissolved gas supersaturation [9,10,11]. In the last ten years in China, a large number of high dams (>200 m) (e.g., Shuangjiangkou Dam, Xiluodu Dam, Xiangjiaba Dam and Lianghekou Dam) have been built or have begun construction on the Yangtze River. When floodwaters are released from these high dams, total dissolved gas (TDG) supersaturation occurs in the water downstream of the dams. TDG supersaturation maintains a high saturation level (130–143%) downstream of Three Gorges Dam and Ertan Dam during periods of flood discharge [5,12]. This may cause fatal effects in fish due to GBD. In 2014, the flood discharge at Xiluodu Hydropower Station caused the death of 40 tons of fish. Recently, many studies have been performed on the impact of TDG levels on the supersaturation tolerance of fish in the Yangtze River, for example, Acipenser dabryanus, Procypris rabaudi, Hypophthalmichthys nobilis, Myxocyprinus asiaticus and Schizothorax prenanti [13,14,15,16,17,18,19,20]. However, most of these studies have been executed in the laboratory. Few works have evaluated the effects of TDG supersaturation on fish in natural rivers during periods of flood discharge.Myxocyprinus asiaticus and Procypris rabaudi inhabit in the upper Yangtze River and are endemic and rare species of Yangtze River. With the development of high dams in the Yangtze River, the populations of these two species have declined greatly owing to habitat destruction [21,22] and have been listed as key protected animals in China. In order to estimate their tolerance of supersaturation under natural conditions, it is necessary to determine the effect of water depth on their survival in TDG-supersaturated waters. Therefore, the aim of this study was to investigate the survival of these two species when subjected to TDG-supersaturated water at varied water depths in the natural river during periods of flood discharge. The study results can promote the protection of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus (or other rare species) and contribute to the improvement of reservoir operation practices in the Yangtze River.2. Materials and Methods2.1. Ethics StatementThe experimental proposal was approved by Xihua University (202150-51). All experiments were executed in accordance with the animal management regulations of Sichuan Province in China.2.2. Experimental Site and Experimental FishThe experimental site (29°21′43.1″ N, 102°14′23.5″ E) is located in Shimian county (Sichuan province) and lies between Dagangshan Dam and Longtoushi Dam in the Dadu River (a tributary of the Yangtze River) (Figure 1). The test facilities were set up on 10 × 10 m covered barges that moored in the backwater region of the Dadu River approximately 12 km downstream from Dagangshan Dam.In this experiment, juvenile Myxocyprinus asiaticus and juvenile Procypris rabaudi were obtained from the Sichuan Fisheries Research Institute (Chengdu, China) with authorization from the government. The weight and length of juvenile Myxocyprinus asiaticus were 16 ± 0.5 g and 17.6 ± 0.3 cm, respectively. The weight and length of juvenile Procypris rabaudi were 4.3 ± 0.4 g and 10.7 ± 1.1 cm, respectively.2.3. Experimental DesignTo investigate the survival of test fish exposed to TDG-supersaturated water at various depths, four different water depths (0–0.3 m, 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m) were used for the tests (Table 1). Eight cages were used in the experiment. Two cages (length × width × height: 0.5 m × 0.5 m × 0.5 m) were placed at water depths from 0–0.3 m. Two cages of the same size were placed at the 0.3–1.3 m depths. Two cages of the same size were placed at the 1.3–2.3 m depths. Two cages (length × width × height: 0.5 m × 0.6 m × 2.3 m) were placed at water depths from 0–2.3 m. Furthermore, the control group (TDG 100%) was established in a tank (720 L) filled with aerated river water in the river bank. Two replicates were also set up for the control group.Before the experiment, the tested fish were held in aerated river water (dissolved oxygen (DO) level: 5.81–8.02 mg L−1; water temperature: 16–17 °C; TDG level: 98–100%) for 24 h to allow them to acclimate to the new environment. After acclimation, 20 juvenile Myxocyprinus asiaticus and 20 juvenile Procypris rabaudi were placed into each cage (Table 1). The same number of test fish was placed in the control group. During the experiment, the tested fish were observed at 8.00 a.m. in the morning and at 1:00 and 7:00 p.m. in the afternoon by raising the cages to the surface. The signs of GBD and the number of dead fish were recorded. The TDG level was continually measured by a Point Four Tracker (Pentair Aquatic Eco-Systems Inc., Coquitlam, Canada). A multiparameter water quality sonde (6600, YSI Inc., Yellow Springs, OH, USA) was used to measure the pH value, temperature and DO level. The weight and length of dead fish were measured with an electronic scale and a ruler. The whole experiment lasted for 96 h from 4 to 8 July 2017.2.4. Data AnalysisIn this study, the survival process of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus in the TDG supersaturated water with varied depth was investigated with the survival probability. The survival probability is calculated based on the following mathematical formula:(1)p=(1−nN)×100% where p, n, and N are the survival probability, number of deaths and total number of the test fish in each case, respectively.Furthermore, the survival analysis of the test fish was performed with an accelerated failure time (AFT) model in R version 3.6.1 [23,24]. The application of the AFT model to detect the effect of TDG supersaturation on fish has been described in our previous studies [20,21,25]. Briefly, the AFT model used in this study is depicted as follows:(2)T=exp(β⋅d)⋅T0 where T is the time-to-death of juvenile Myxocyprinus asiaticus (or juvenile Procypris rabaudi) experienced the exposure of TDG-supersaturated water at various depths, and T0 stands for the time-to-death of these two species in the control group. The water depth of 0–0.3 m was considered the control group in the analysis. The variable d represents the experimental water depth, and the unknown coefficient β represents the influence of the water depth on the time-to-death of juvenile Myxocyprinus asiaticus (or juvenile Procypris rabaudi) exposed to TDG supersaturation. Hypothetically, T0 should have a loglogistic distribution, and the water depth was treated as the categorical variable to fit the survival data. The parameters involved in the AFT model were evaluated by the maximum likelihood method. Thus, the survival S(t) can be determined at a given time (t) by using the following equation:(3)S(t)=(1+(t⋅exp(ε)−1⋅exp(β⋅d)−1)1/ϕ)−1 where the unknown parameters ε (intercept) and φ (scale) can be determined in the AFT model. Furthermore, the survival curve determined from the AFT model can completely describe the survival process of the test fish. The survfit function was applied to determine the estimated coefficients and their standard errors from the survival curve related to each experimental condition. In addition, the function survdiff was selected to analyze and compare the survival curves among the experimental groups. The Holm method was used to examine the p-value correlation [26]. The average survival time (AST) calculated from the AFT model was used to estimate the effect of water depth on the survival of the test fish experienced the TDG exposure by one-way analysis of variance (ANOVA). The least significant difference test was used for a post hoc multiple comparison with the one-way ANOVA. Tamhane’s T2 test was applied when homogeneity of variance was not confirmed. The significance level was set at p < 0.05.3. Results3.1. TDG Supersaturation Analysis and GBD SignsDuring the observation period, the water temperature was 14.9–17.4 °C. The TDG and DO values are shown in Figure 2. The TDG level varied around 115% during the first 36 h. A high value (127% TDG) and a low level (105% TDG) appeared in 36–48 h. The range of the TDG level was 105–115% in 48–60 h, and the TDG level fluctuated obviously at approximately 84 h. The average values of TDG and DO were 116% and 8.89 mg L−1, respectively. During the experiment, the test fish showed obvious GBD signs, such as gas bubbles in their fins and gills.3.2. Survival of the Test Fish under TDG Supersaturation at Various Water DepthsThe survival probability of the test fish increased with increased water depth while increased TDG supersaturation exposure time led to decreased survival probability (Figure 3). Figure 3a shows that there was a 95% survival probability for juvenile Procypris rabaudi in the 12–24 h period at depths of 0–0.3 m and 0.3–1.3 m. After exposure for 48 h, 65% of juvenile Procypris rabaudi survived at 0–0.3 m, and no test fish died at 1.3–2.3 m. At the end of 96 h of exposure, the survival probability of juvenile Procypris rabaudi had decreased to 40%, 80%, 95% and 75% at the 0–0.3 m, 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m depths, respectively. Additionally, as shown in Figure 3b, the survival probability of juvenile Myxocyprinus asiaticus was 95% in the 12–24 h period at 0–0.3 m and 1.3–2.3 m, respectively. No dead juvenile Myxocyprinus asiaticus appeared at 0.3–1.3 m or 0–2.3 m during the same exposure period. After exposure for 48 h, the survival probability of juvenile Myxocyprinus asiaticus decreased to 85% at 1.3–2.3 m, while there was a 95% survival probability at other water depths (0–0.3 m, 0.3–1.3 m and 0–2.3 m). Until the exposure time of 96 h, the survival probability of juvenile Myxocyprinus asiaticus was 70%, 80%, 75% and 65% at 0–0.3 m, 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m, respectively. Throughout the experiment, all juvenile Procypris rabaudi and Myxocyprinus asiaticus survived at 100% TDG.Table 2 lists the estimated parameters obtained from the AFT model fitting. The effect of water depth on the survival of juvenile Procypris rabaudi and Myxocyprinus asiaticus was estimated using the absolute values of the coefficients. The results showed that water depth had a strong effect on the survival of juvenile Procypris rabaudi exposed to TDG supersaturation. Compared with shallow water (0–0.3 m), deep water significantly increased the survival of juvenile Procypris rabaudi (|Coeff. β0.3−1.3 m| = 0.63, p = 0.05 and |Coeff. β1.3 − 2.3 m| = 1.18, p < 0.01). However, there was no significant difference in the survival of juvenile Procypris rabaudi between 0–0.3 m and 0–2.3 m (Coeff. β0 − 2.3 m| = 0.58, p=0.06 > 0.05). Furthermore, as shown in Table 2, no significant difference in juvenile Myxocyprinus asiaticus survival was found between the water depth of 0–0.3 m and the other water depths (|Coeff. β0.3 − 1.3 m| = 0.48, p = 0.23 > 0.05, |Coeff. β1.3 − 2.3 m| = 0.26, p=0.32 > 0.05, and |Coeff. β0−2.3 m| = 0.19, p=0.6 > 0.05).In this study, the average survival time (AST) was also used to assess the survival of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus experienced the exposure of TDG-supersaturated water at varied water depths (Figure 4 and Table 3). As shown in Figure 4a, the AST was significantly lower at water depths of 0–0.3 m than at other water depths (0.3–1.3 m: F = 4.67; df = 1,40; p=0.03 < 0.05; 1.3–2.3 m: F = 14.54; df = 1,40; p < 0.01; 0–2.3 m: F = 7.99; df = 1,40; p < 0.01). There were no significant differences in the AST of Procypris rabaudi at water depths of 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m. For juvenile Myxocyprinus asiaticus (Figure 4b), no significant differences were observed in the AST values at 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m compared with that at 0–0.3 m (0.3–1.3 m: F = 0.31; df = 1,40; p=0.58 > 0.05; 1.3–2.3 m: F = 0.04; df = 1,40; p=0.83 > 0.05; 0–2.3 m: F = 0.10; df = 1,40; p=0.75 > 0.05).In addition, the results in Table 3 show that the AST values of juvenile Procypris rabaudi were 72.61 h, 85.85 h, 92.88 h and 89.36 h at 0–0.3 m, 0.3–1.3 m, 1.3–2.3 m and 0–2.3 m, respectively. The AST values of juvenile Myxocyprinus asiaticus were 80.01 h, 90.36 h, 84.72 h and 86.26 h in the same water depths, respectively. The AST value of juvenile Procypris rabaudi was lower than that of juvenile Myxocyprinus asiaticus at 0–0.3 m and 0.3–1.3 m. However, juvenile Procypris rabaudi had higher AST values than juvenile Myxocyprinus asiaticus at 1.3–2.3 m and 0–2.3 m.4. DiscussionIn the mid-1960s, it gradually became evident that a serious dissolved gas problem existed in the Columbia River system [7]. Westgard (1964) observed adult Oncorhynchus tshawytscha suffering gas bubble disease at the McNary Spawning Channel [9]. Subsequently, the effect of gas supersaturated water on Oncorhynchus tshawytscha, Oncorhynchus and Salmo gairdneri were investigated in the Nechako River and the Columbia River [2,11]. In China, some studies have been carried to estimate the effect of TDG on endemic species of Yangtze River (e.g., Acipenser dabryanus, Schizothorax prenanti and Procypris rabaudi) in the laboratory [17,18,19]. However, relatively few studies have investigated the effect of TDG supersaturation on fish in natural rivers. Xue et al. [27] indicated that all juvenile Ctenopharyngodon idella died at 0–1 m water depth (average TDG: 125%; exposure time: 140 h) when the Xiangjiaba Dam discharged floodwaters in the upper Yangtze River in 2014. At depths of 1–2 m, the survival rate of the species was approximately 60%. In the next year, they also found that approximately 30% Procypris rabaudi survived at 0–0.5 m water depth in 100 h (average TDG: 123%; average temperature: 22.3 °C). Cao et al. [28] indicated that no juvenile Myxocyprinus asiaticus survived at 0–0.7m water depth in 60 h (average TDG: 123%; average temperature: 22.3 °C) downstream of Xiangjiaba Dam in August 2014 and September 2015, but the mortality decreased with increasing water depth. In our results, the survival probability of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus was 40% and 70% at 0–0.3 m water depth (average TDG: 116%; exposure time: 96 h; average temperature: 16.1 °C), respectively. This may indicate that increasing TDG level has great effects on the survival of these species in natural rivers. However, the effect of temperature should not be ignored because increasing water temperature can decrease the survival of fish exposed to TDG based on Yuan et al. [29]. They found that an increase in water temperature (from 12 to 20 °C) decreased median survival time (from 18.62 to 5.22 h) of juvenile Schizothorax prenanti exposed to 130% TDG for 96 h.Furthermore, in the study of Yuan et al. [30], 23% and 43% of adult Schizothorax prenanti survived at water depths of 0–1 m and 1–2 m in TDG-supersaturated water (average TDG: 117%), respectively. At the same water depths, adult Schizothorax davidi showed a higher survival rate (53%) after 6 d of exposure. Their results also showed that the survival rates of juvenile Schizothorax prenanti, juvenile Schizothorax davidi and juvenile Leptobotia eloungata were 86%, 70% and 96% at water depth of 0–1 m, respectively. The juveniles exhibited greater tolerance than the adults under the same conditions. Juvenile Leptobotia elongata showed the highest tolerance of all fish in their study. In the present study, the survival probability of juvenile Procypris rabaudi was 40%, 80% and 95% at 0–0.3 m, 0.3–1.3 m and 1.3–2.3 m, respectively, while that of juvenile Myxocyprinus asiaticus was 70%, 80% and 75% at the same water depths (Figure 3). Compared with the surface water (0–0.3 m), deeper depths increased the survival of the test fish. This was consistent with the studies of Xue et al. [27], Cao et al. [28] and Yuan et al. [30]. Furthermore, the water depth made a greater contribution to the survival of juvenile Procypris rabaudi than to the survival of juvenile Myxocyprinus asiaticus. Juvenile Myxocyprinus asiaticus showed a higher tolerance to TDG supersaturation in shallow water (0–0.3 m) than juvenile Procypris rabaudi. When the water depth was 0–2.3 m, the survival probability of juvenile Procypris rabaudi rose from 40% (0–0.3 m) to 75%. We speculate that juvenile Procypris rabaudi can escape from TDG supersaturation by swimming to deeper depths. Cao et al. [31] found that Myxocyprinus asiaticus can avoid highly TDG-supersaturated water (140%–150% TDG). At present, no observations have been made to investigate TDG supersaturation avoidance by Myxocyprinus asiaticus at other TDG levels (<140%). In our results, there was only a small change in the survival probability of juvenile Myxocyprinus asiaticus between 0–0.3 m and 0–2.3 m (from 70% to 65%) after 96 h of exposure. We speculate that juvenile Myxocyprinus asiaticus did not exhibit obvious avoidance ability at a low TDG level (116%).It has been shown that GBD in fish can lead to fish death owing to hypoxemia, which is caused by embolism in the blood vessels [25,32,33]. In the present study, juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus subjected to TDG-supersaturated water at different water depths exhibited signs of GBD, such as bubbles in their gills and fins (caudal fin, ventral fin and dorsal fin). In addition, our results also demonstrated that increased water depth extended the survival time of the test fish and increased their survival. The possible explanation for this finding is that the compensation depth of water decreased the development of GBD [34]. Previous research demonstrated that higher TDG level (>125%) led to high mortality (nearly 100%) in fish after 96 h of exposure [13,19,22,31]. Relatively high survival probabilities for juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus were observed in the study. This may be related to the low TDG level (116%) in addition to the effect of compensation depth. In addition, one possible reason for this finding is that the exposure time of 96 h at the low TDG level was not long enough to cause high mortality in the fish [28,35,36].Based on the results of the AFT model (Table 2 and Figure 4), there were significant differences in the survival of juvenile Procypris rabaudi between the surface water (0–0.3 m) and the other water depths (0.3–1.3 m,1.3–2.3 m and 0–2.3 m). No significant differences were observed for juvenile Myxocyprinus asiaticus survival under the same conditions because of high survival probability. This indicated that water depth has a greater effect on the survival of juvenile Procypris rabaudi than on that of juvenile Myxocyprinus asiaticus exposed to TDG supersaturation (116% TDG). Additionally, our results also showed that the AST of juvenile Procypris rabaudi was lower at 0–0.3 m and 0.3–1.3 m than that of juvenile Myxocyprinus asiaticus (Table 3). In comparison to juvenile Myxocyprinus asiaticus, juvenile Procypris rabaudi exhibited a weaker tolerance to TDG supersaturation in the relatively shallow water layer, and juvenile Procypris rabaudi was more vulnerable to TDG supersaturation even if the TDG level was low. However, when the test fish were provided with more space (0–2.3 m) to swim freely, the AST of juvenile Procypris rabaudi was higher than that of juvenile Myxocyprinus asiaticus. The reasons for this result may be that juvenile Procypris rabaudi can exhibit their ability to avoid TDG.According to the results of a previous laboratory study, the mortality of juvenile Procypris rabaudi (weight: 2.41–2.48 g; length: 6.6–7.2 cm; water depth: 0.2 m; temperature: 25 °C) was minimal in TDG-supersaturated water (115% TDG) after 60 h of exposure [13]. No juvenile Myxocyprinus asiaticus (average weight: 39.54 g; average length: 11.54 cm; water depth: 0.2–0.5 m; temperature: 21 °C) died at 120% TDG after exposure for 48 h [15]. In our study, the survival probabilities of juvenile Procypris rabaudi (average weight: 4.3 g; average length: 10.7 cm) and juvenile Myxocyprinus asiaticus (average weight: 6 g, average length: 7.6 cm) were 40% and 70% respectively, at 0–0.3 m depth in the Dadu River (average temperature: 16.1 °C) after 96 h of TDG supersaturation exposure. This may be related to the peak TDG levels that occurred during the test (Figure 2). In addition, existing results showed that suspended sediment might accelerate fish death in TDG-supersaturated water [19,21,22]. For example, Li et al. [22] found that an increase in suspended sediment concentration (from 200 to 1000 mg L−1) increased mortality (from 10 to 70%) of juvenile Myxocyprinus asiaticus exposed to 125% TDG for 4 h, while increasing suspended sediment concentration (from 200 to 1000 mg L−1) caused 20 to 81% mortality in juvenile Procypris rabaudi exposed to 125% TDG for 10 h [21]. The water temperature and the size of the fish affect fish tolerance to TDG supersaturation [20,29,33,37,38]. The differences in TDG supersaturation tolerance between our results and the results of previous laboratory studies might be explained by the above factors. To provide accurate predictions of what will really happen, the results of laboratory and field bioassay experiments must be interpreted in terms of all discernible natural conditions. Thus, due to the lack of field results in China, more field bioassay experiments should be performed on the effects of TDG supersaturation in fish, such as examining the avoidance ability of different species (especially rare fish), examining lethal and sub-lethal effects for all life stages of same species and examining effects caused by repeated exposure in natural rivers during the discharge of flood. Furthermore, although some useful results were obtained in the present study, we only set up two replicates for each water depth. The results may have limitations in assessing the effect of TDG supersaturation on the survival of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus at varying water depth.5. ConclusionsOur study indicates that deeper water depths can increase the tolerance of juvenile Procypris rabaudi to TDG supersaturation in natural rivers during periods of flood discharge while it cannot improve the survival of juvenile Myxocyprinus asiaticus. Compared with juvenile Myxocyprinus asiaticus, juvenile Procypris rabaudi show a weaker tolerance to TDG supersaturation in shallow water even if the TDG level is low. According to our results, we speculate that juvenile Procypris rabaudi may have an ability to avoid TDG supersaturation (average value: 116%), but that juvenile Myxocyprinus asiaticus did not show this ability. In this study, 40–95% juvenile Procypris rabaudi and Myxocyprinus asiaticus can survived at 116% TDG for 96 h at 0–0.3 m, 0.3–1.3 m and 1.3–2.3 m depths. If the exposure time is sufficiently long or the TDG level is higher (>120%), the survival of juvenile Procypris rabaudi and juvenile Myxocyprinus asiaticus may be seriously threatened even at water depths of 2.3 m. Maintaining a low TDG level (<116%) and deep water depths (>3 m) downstream of dams is necessary for the protection of Procypris rabaudi and Myxocyprinus asiaticus (or other rare species). In addition, we find that fluctuating TDG levels with a mean of 116% TDG cause greater mortality in both species compared to static exposures in the lab. To provide an effective measurement for managing GBD effects in fish, more field studies should be performed. The thresholds of effect of TDG should be determined in future research for the formulation of reservoir operation systems during the flood season. Moreover, a long exposure time to high TDG levels should be avoided. Therefore, some measures should be taken to eliminate and reduce harm to fish from TDG supersaturation, for example, shortening discharge periods, increasing water depth and restricting TDG levels.

animals : an open access journal from mdpi

10.3390/ani11041052

PMC8068381

The motility of the smooth muscle of the rumen and the reticulum is tightly synchronized since they are responsible for degrading the ingested fibrous diet. It involves mixing contractions of partially digested material (cycle A), cyclic eructation of fermented gases (cycle B), and contractions associated with regurgitation and rumination (cycle C). Disorders of cycles A, B, and C in reticulo-rumen contraction occurred in cases of many metabolic diseases. Since the smooth muscle electromyography (EMG) is closely associated with the motility pattern, our study aimed to implement the long-term EMG telemetry recording method to analyze the basic reticulo-rumen motility pattern in conscious unrestrained sheep. The study was conducted on nine ewes chronically fitted with bipolar electrodes in the rumen and reticulum. Cycles A and B occurred constantly in a similar share, whereas cycles C appeared every 30 min and lasted for the most part 15–30 min. EMG signal analysis showed more differences at the level of single bursts than on burst bundles. Recording reticulo-rumen EMG signal with a long-term telemetry approach offers a reliable tool to assess the forestomach motility in conscious, unrestrained sheep, allowing for easy recognition of reticulo-rumen cycles.

The reticulum and rumen are considered a single functional unit (the reticulo-rumen) with regards to myoelectrical and contractile activities. The specialized contractions of the reticulo-rumen provide constant mixing of partially digested material (cycle A), its flow into the omasum during eructation (cycle B), and regurgitation-rumination (cycle C). This study aimed to investigate the feasibility of electromyography (EMG) registered by a long-term telemetry method for assessment of the basic reticulo-rumen myoelectrical activity in sheep, to develop the effective recognition of the reticulo-rumen cycles at rest with no food stimulation, and to investigate the relationship between cycles A, B, and C in such basic conditions. The experiment was carried out on nine ewes. Myoelectric activity of the rumen, reticulum, and abomasum was recorded by the combination of three silver bipolar electrodes and a 3-channel transmitter implant. The myoelectrical activity registered successfully in the reticulum and rumen was determined as three characteristic patterns of cycles A, B, and C. The percentage of each type of cycle changed at different intervals from equally cycles A (43–50%) and B (50–56%), occurring when cycle C was not observed to the domination of cycle C (57–73%) with a decrease of cycles A (6–14%) and B (20–28%). The long-term EMG telemetry registration is feasible in the assessment of the reticulo-rumen myoelectrical activity in sheep.

1. IntroductionIn ruminants, the myoelectric activity of the smooth muscle of the rumen and the reticulum is characterized by contractility that allows intake, degradation, and transport of various fibrous components and their metabolic products. This is followed by evacuation of fermented gases. The reticulum and rumen are considered a single functional unit when it comes to myoelectrical and contractile activities. This is because their contractions and their control are primarily vagal-dependent, they are initiated in the reticulum and then their myoelectrical signal is transferred to the rumen [1,2,3]. This activity, known as the reticulo-rumen cycles pattern, allows mixing and degradation of ingested plant material, eructation of fermented gases, separation, and transport of plant particles. The latter is done either backwards for rumination purposes or continuously for further digestive processes in the omasum, abomasum, and intestines [3,4]. Due to the alternating contractions of the rumen and the reticulum, the insoluble heavier, crumbled nutrients (e.g., cereal grains) fall to the bottom of the ventral rumen sac, while the lighter, finely divided components form a spongy weave of fibers on the surface of the rumen, over which fermented gases accumulate [5]. In the reticulo-rumen cycles, regular contraction sequences within individual parts of the stomach chambers are organized in three schemes [5,6]. The specialized contractions of the reticulo-rumen provide constant mixing of partially digested material (cycle A) and its flow into the omasum as well as cyclic eructation (cycle B) and regurgitation-rumination (cycle C). Braun et al. [7,8] used the ultrasonographic examination to describe A and B cycles, differentiating between primary and secondary rumen contractions. Additionally, Plaza et al. [4] used the terms primary and secondary contractions to describe mixing and eructation in the reticulo-rumen using electromyography (EMG). However, the authors often reported additional contractions of the reticulum during rumination. Until now, the involvement of the rumen and of the reticulum in contraction cycles as well as the correlation between mixing/eructation and rumination contractions was unclear. Therefore, this preliminary study aimed to record EMG signals in the reticulum and the rumen of conscious unrestrained sheep with telemetry method. This investigation will enable further long-term study on the changes of the reticulo-rumen cycles in response to different diets and/or pharmaceutical treatments.The main advantage of the present approach is the long-term recording of EMG signals, which allows recognizing cycles A, B, and C at rest. The forestomach motility in ruminants can be assessed physiologically using direct and indirect methods. Indirect methods, such as ultrasonography [7,8] and bolus wireless sensors with an accelerometer [9,10], have some limitations. Bolus sensors have numerous artifacts whereas ultrasonography approaches do not provide easy access to all components. This is turn skews the actual contraction of the examined part of the gastrointestinal tract. On the other hand, the direct methods, such as the force transducer method [11,12] or electromyography approach [4,13], require surgery. However, this can record continuous measurements over a long time and allow to obtain a signal with the least number of artifacts. The EMG method is characterized by a high correlation with the motor activity of the examined structures and high repeatability of results [14]. The EMG method, when the signal is collected directly from the muscle layer of the forestomach via electrodes, has been recognized as effective and reliable in sheep [15,16]. In later studies, the usefulness of EMG as a method that reflects the electrical activities of myocytes in the muscular layer of the gastrointestinal wall was confirmed as the most reliable source describing motor activity [4,13]. The most important limitation concerns signal transmission. Kania et al. [17], Sarna et al. [18], and Plaza et al. [4] used cable signal transmission, which caused long-term immobilization of the animals and limited the duration of registration from 6 [5,18] to 12 h [4]. In contrast, considering all available methods, ultrasonography allowed us to exam the few minutes of forestomach activity [7,8], whereas, the telemetry bolus-type sensor enable a few hours of observation [9,10,12]. Therefore, this study presented EMG registration using long-term telemetry method of signal transmission. The long-term telemetry method allows to significantly extend the time of registration from several days to around 2 months [19,20,21]. The long-term telemetry method enabled EMG measurements which maintained animal welfare and reduced the animal’s stress associated with immobilization during registration [13,22]. However, it should be kept in mind that the advantage of telemetry may not even compensate for the surgery necessary to implant the electrodes. The long-term telemetry method has been successfully used to study the myoelectric activity of the gastrointestinal tract in pigs [13,23] and the reproductive tract in pigs [22,24,25,26]. Therefore, the implementation of the long-term telemetry method to evaluate the reticulo-rumen myoelectrical activity in sheep was proposed in this paper.There is a need for introducing an accurate registration and transmission method that allows proper identification and description of cycles A, B, and C of the rumination process, due to the importance of forestomach diseases. Motility disorders of the forestomach which accompany many diseases of bacterial and metabolic etiology cause severe pain symptoms and are an important clinical problem in ruminants [27]. Gas accumulation in the rumen results in free gas bloat observed in hypocalcemia [28] or grain overload [29] and are consequences of inhibition of cycle B, which are vital for eructation. On the other hand, improper mixing of the digest in the rumen (cycle A) and rumination (cycle C) could result in the rumen overload and impaction on the rumen and omasum [30]. In many metabolic diseases, such as ruminal acidosis [31], hypocalcemia [11], and numerous diseases causing symptoms of severe pain [32], disturbances of the frequency and amplitude of contractions of the reticulo-rumen were reported. Therefore, scientific research on the myoelectrical activity of the gastrointestinal tract is needed to prevent clinical cases of gastrointestinal motility disorders, mainly of the forestomach in ruminants [27,33,34]. However, the EMG evaluation of the reticulo-rumen motility cannot be considered as a practical solution in clinics or agriculture.The primary objective of this study is to investigate the feasibility of electromyography registered by a telemetry method for the assessment of the reticulo-rumen myoelectrical activity in sheep. Ancillary goals were to develop the effective recognition of the reticulo-rumen cycles at rest with no food stimulation, and investigate the relationship between cycles A, B, and C in such basic conditions. To achieve this goal, the reticulo-rumen contractile pattern at rest was characterized in detail in nine ewes based on the amplitude, root mean square (RMS), duration, and frequency of myoelectrical activities.2. Materials and Methods2.1. AnimalsThe experimental protocol was approved by the II Local Ethical Committee on Animal Testing in Warsaw (Permit Number: 14/2013, from 27 March 2013) on behalf of the National Ethical Committees on Animal Testing. The experiment was carried out on nine mature crossbred ewes (n = 9), each weighing 35–40 kg body weight. The ewes were adapted to the animal facilities for 7 days before the experiment. During the whole experiment, the ewes were kept in individual cages and had unlimited tap water and hay. During the entire study, the ewes were clinically healthy. Myoelectric activity of the rumen, reticulum, and abomasum was recorded by the combination of three silver bipolar electrodes [13,18] connected to a 3-channel transmitter implant (TL10M3-D70-EEE, DSI, St. Paul, MN, USA). In brief, the bipolar needle electrodes were made from a silver wire (0.4 mm, Ag 99.98% purity, Mennica Polska, Warsaw, Poland) and connected to a stranded type steel wire immersed in Teflon (A-M Systems, Inc., Sequim, WA, USA). The electrodes were then fixed between two silicone foil pieces (Σ Sigma medical, No. 2079, Nanterre, France) with silicone glue (Silastic® No. 891, Dow Corning, Midland, MI, USA). A set of three electrodes was connected to the 3-channel transmitter implant. Before the surgery, the electrodes and the implant were sterilized in a 2% solution of glutaraldehyde (Aldesan E, Septoma, Warsaw, Poland). They were then washed with aseptic physiological saline (0.9% NaCl, Polfa S.A., Lublin, Poland).Implantation of electrodes and transmitter was performed under general isoflurane anesthesia [35] in animals premedicated with midazolam (Midanium, 0.25 mg/kg b.wt., i.m., Polfa S.A., Lublin, Poland). Following midline laparotomy, the reticulum, abomasum, and rumen were extracted from the abdominal cavity, and electrodes were sutured, on the central section of the cranial part of the reticulum, on ventral sac of rumen, and abomasum (4 cm to pylorus), respectively. The electrode on the abomasum was used to follow the passage of ingesta from the reticulo-rumen to the abomasum. The electrode on the abomasum did not affect the myoelectrical activity of the reticulo-rumen, nor the classification of signals as mixing, eructation, or rumination patterns. In the next step, the telemetry transmitter TL10M3-D70-EEE (DSI, St. Paul, MN, USA) was fixed in the pocked between the abdominal muscles on the right flank, a ground electrode was sutured to the abdominal muscles near the transmitter, and the laparotomy was routinely closed. In the recovery period, ewes were treated with tolfenamic acid (Tolfedine, 2 mg/kg b.wt., i.m., Vetoquinol, Lure, France) and procaine penicillin with dihydrostreptomycin (Pen-strept, 10 mg/kg b.wt., i.m., Scanvet, Dublin, Ireland) for 5 days. The healing was fast and without complications.2.2. Electromyography Recording and AnalysisControl EMG recordings started the next day after the surgery, however, the EMG signal free of injury potentials was observed just a few days after surgery, in general, from the 3rd to 5th day after the surgery. Regular EMG recordings lasted usually for 5–6 weeks. After that time the increasing number of artifacts made the EMG recordings useless for analysis. Before each EMG recording session, the access to feed was limited for 12 h before the beginning of recording with free access to water. Registration started at 7 a.m., and it was conducted for 3 h. To spare the battery, after each recording session, the telemetry transmitter implant was turned off. Recorded with 100 Hz sampling frequency EMG signal was digitized and modulated into radiowave and received by a receiver (RMC-1, DSI, New Brighton, MN, USA) located nearby the animal’s cage. The radiowave signal was then demodulated, amplified (Matrix, DSI, St. Paul, MN, USA), and archived (Neuroscore, DSI, St. Paul, MN, USA) for further off-line analysis. The EMG signals were digitally filtered with a band-pass filter (5–50 Hz) [36,37,38]. The myoelectrical activity of the rumen and reticulum was defined as a series of spike potentials with an amplitude exceeding 5 μV and duration longer than 0.5 s [24]. Spiking activities were organized into bursts and bundles forming A, B, and C cycles. Bursts were defined as series of spiking activities with amplitude exceeding 5 μV. In bursts, spike activities occurred one after the another with pause between spikes shorter than 0.5 s. The next burst began when the pause between subsequent spikes was longer than 0.5 s [4]. Bundles were defined as series of bursts with amplitude exceeding 5 μV and pauses between bursts longer than 5 s. The myoelectrical activity of the rumen and the reticulum was analyzed using the following parameters: burst mean amplitude (mV), burst mean RMS (root mean square) (mV), burst duration (s), bundle mean amplitude (mV), bundle mean RMS (mV), bundle duration (s), and a number of bursts forming a bundle. Parameters characterizing individual A, B, and C cycles were described in 12·15-min intervals (T1–T12) for each 3 h recording session.2.3. Electromyography Recognition of A, B, and C CyclesThe classification of signals as mixing, eructation, and rumination patterns was proposed basing on Plaza et al. [4] and Braun et al. [7,8] works. Braun et al. [7,8] described A and B cycles, as primary and secondary rumen contraction. Each primary contraction cycle (cycle A—mixing) starts with biphasic reticular contraction, which is followed by contraction of the anterior blind sac of the rumen and then contraction of the dorsal sac of the rumen in a cranio-caudal direction. Each secondary contraction cycle (cycle B—eructation of gas) does not involve the reticulum and ruminal atrium and consists of contraction of the dorsal sac and then the ventral sac of the rumen. Therefore, using electrode arrangement presented in this study, cycle A was identified as the reticulum monophasic or biphasic contraction that occurred before the single rumen contraction, whereas cycle B was identified as the reticulum monophasic or biphasic contraction that occurs before the double rumen contraction.Additionally, Plaza et al. [4] used the terms primary and secondary contractions to describe mixing and eructation in the reticulo-rumen using electromyography. However, the authors proved during rumination, the reticular biphasic contraction became triphasic. It was the first borderline criterion for the identification of the C cycle. Braun et al. [7] confirmed Plaza’s et al. [4] findings that a triphasic contraction of the reticulum occurs when a biphasic contraction is immediately preceded by a so-called rejection contraction, which function is to transport the cud into the esophagus, and it represents the rumination. It was the second borderline criterion for the identification of the C cycle in our work. Therefore, using electrode arrangement presented in this study, cycle C was identified as the reticulum triple contraction that occurs before the rumen contraction, whether the contraction of the rumen was single or double.2.4. Data AnalysisAll statistical evaluation was performed by GraphPad Prism6 software (Graph-Pad Software, San Diego, CA, USA), and the level of statistical significance was set to p < 0.05. The non-parametric Kolmogorov–Smirnov test, one-way analysis of variance (ANOVA) supported by the Holm–Sidak multiple comparison test or Kruskal–Wallis rank ANOVA supported by the Dunn multiple comparison test were used to compare the tested features between intervals (T1–T12) and to compare more than two features at individual intervals and Student’s t-test with Welch correction or Mann Whitney’s test to compare two features at individual intervals. The results in Table 1 were reported as mean ± SD. Additionally, the variability across animals and variability over the 5–6 week period of observations were calculated using the coefficient of variation (CV, %). First, the CVs were calculated across animals for each day, and mean values are presented in Table 2. Then, the CVs were calculated throughout observations for each animal, and mean values are presented in Table 3.3. ResultsThe myoelectrical activity registered in the reticulum and rumen was determined as a three-cycle pattern using burst and bundle parameters. Cycle A (mixing) was defined as single or double bursts forming bundles in the reticulum (limit values: A > 5 μV; T > 0.5 s), before occurrence of single bundle in the ventral rumen sac (limit values: A > 5 μV; T > 2 s) (Figure 1). Cycle B (eructation) was defined as single or double bursts forming bundles in the reticulum (limit values: A > 5 μV; T > 0.5 s), then double bursts forming bundles in the ventral rumen sac (limit values: A > 5 μV; T > 2 s) (Figure 2). Cycle C (regurgitation-rumination) was determined as triple bursts forming bundlez in the reticulum (limit values: A > 5 μV; T > 0.5 s), followed by single or double bursts forming bundles in the ventral rumen sac (limit values: A > 5 μV; T > 2 s) (Figure 3).During the preprandial period, the number of ruminal cycles remained constant 12 ± 0.2 to 16 ± 0.1 in 15 min intervals, which means that they appeared with frequency 0.013–0.018 Hz. Cycles A and B were observed in all 12 intervals, whereas cycle C was observed only in T3–4, T6–8, T11–12. We found that the percentage of each type of cycle changed at different intervals during the rest period. In the rest period, cycles A and B occurred equally with the percentages of 43.75–50.00% and 50.00–56.25% respectively, when cycle C was not observed. Cycle C appeared every 30 min for at least two intervals and in that time were dominating cycles (57.14–73.33% of all cycles recorded). At this time only intermittent cycles A (6.67–14.29%) and B (20.00–28.57%) were observed. In the summary of observations, the repeated occurrence of the rumination phase during 30 min intervals was repeatable in all animals from the start point of synchronization (Figure 4).The EMG parameters of each cycle were characterized in detail in the reticulum and the ventral sac of the rumen (Table 1), and the variability across animals and variability throughout observations are summarized in Table 2 and Table 3, respectively.In cycle A, single or double bursts forming bundles in the reticulum, and a single bundle in the ventral sac of rumen were described using amplitude, RMS, and duration parameters. There were no differences in amplitude between the first and the second burst in the reticulum (1.93 ± 0.29 versus 1.74 ± 0.30 mV, p = 0.511), whereas for both the amplitude was always higher (p < 0.0001) than in the rumen (0.72 ± 0.14 mV). The coefficient of variation for amplitude was higher in the rumen than in the reticulum across animals (22.16% vs. 11.13–14.30%) and throughout observations (20.38% vs. 13.19–16.87%). The RMS of the second burst in the reticulum was higher than in the first (0.48 ± 0.06 versus 0.35 ± 0.05 mV, p = 0.0041), and again for both the RMS was always higher (p < 0.0001) than in the rumen (0.10 ± 0.02 mV). The coefficient of variation for RMS was lower in the rumen than in the reticulum across animals (19.29% vs. 26.05–47.62%) and throughout observations (20.38% vs. 24.57–45.69%). On the other hand, the burst lasted longer (p < 0.0001) in the rumen (7.71 ± 1.13 s) in comparison to the reticulum, where the first burst lasted significantly longer than the second (1.95 ± 0.26 versus 0.82 ± 0.49 s, p < 0.0001). The coefficient of variation for the duration was similar in the rumen and reticulum across animals (18.53% vs. 12.26–18.96%) and throughout observations (14.96% vs. 11.86–13.74%).In cycle B, single or double bursts forming bundles in the reticulum, and double bursts forming bundles in the ventral sac of rumen were similarly described. There were no differences (p > 0.05) in amplitude and RMS between the first (A: 0.75 ± 0.15 mV; RMS: 0.11 ± 0.02 mV) and the second (A: 0.71 ± 0.19 mV; RMS: 0.13 ± 0.02) burst in the rumen, however, both parameters were always lower (p < 0.0001) than in the reticulum. The RMS of the second burst in the reticulum was higher than of the first (0.37 ± 0.05 versus 0.51 ± 0.07 mV, p = 0.002) with no differences in amplitude (1.89 ± 0.31 versus 1.73 ± 0.36 mV, p = 0.149). The coefficient of variation for amplitude was similar in the rumen and in the reticulum across animals (14.34–20.16 vs. 19.66–24.55%) and throughout observations (18.74–26.53% vs. 15.33–20.66%). The coefficient of variation for RMS was lower in the rumen than in the reticulum across animals (10.44–15.09% vs. 27.60–51.57%) and throughout observations (11.24–16.33% vs. 24.77–50.54%). The duration of the first burst lasted always longer than in the second burst, in both the rumen (7.85 ± 1.23 versus 5.27 ± 0.75 s, p < 0.01) and the reticulum (2.01 ± 0.19 versus 0.73 ± 0.12 s, p < 0.01). Again, the burst lasted longer (p < 0.0001) in the rumen than in the reticulum. The coefficient of variation for the duration was similar in the rumen in reticulum across animals (13.34–17.54% vs. 13.40–16.51%) and throughout observations (10.91–15.66% vs. 15.62–19.36%).In cycle C, triple bursts forming bundles in the reticulum, and single or double bursts forming bundles in the ventral sac of rumen were described, likewise. The burst parameters in the rumen were similar to cycle B, with no differences in amplitude and RMS, and longer duration of the first burst. Additionally, the burst parameters in the reticulum corresponded to those in cycle B in the case of the first and the second burst. However, the third burst (0.78 ± 0.08 s) was shorter (p < 0.0001) than the first burst (2.77 ± 0.14), with no differences (p = 0.113) with the second one (1.06 ± 0.27 s). The coefficient of variation for the duration was higher in the rumen than in the reticulum across animals (12.80–52.69% vs. 12.47–18.26%) and throughout observations (13.70–53.22% vs. 13.09–17.17%). There were also no differences in amplitude (p = 0.263) and RMS (p = 0.132) between the third burst (A: 1.60 ± 0.28 mV; RMS: 0.45 ± 0.16 mV) and the previous two in the reticulum. The coefficient of variation for amplitude was similar in the rumen and reticulum across animals (10.66–29.00% vs. 10.83–20.82%) and throughout observations (19.79–24.11% vs. 15.91–22.86%). The coefficient of variation for RMS was slightly lower in the rumen than in the reticulum across animals (13.68–19.23% vs. 17.73–36.22%) and throughout observations (16.97–17.54% vs. 21.69–35.61%).Summarizing the values of the general parameters of bundles in all three types of the reticulo-rumen cycles, there were no differences (p > 0.05) in amplitude and RMS of bundles recorded both in the rumen and in the reticulum, between cycle A, cycle B, and cycle C. At the bundles level, the major differences between cycles concerned duration.In the rumen, bundles lasted the longest (p < 0.0001) in cycle B (26.14 ± 2.68 s) in comparison to shorter bundles in cycle C (16.79 ± 8.94 s), and the shortest in cycle A (7.71 ± 1.13 s). The observed difference resulted from the recorded number of bursts forming bundles, and the length of pauses between subsequent bursts. In the reticulum, the durations of the bundles were comparable (p = 0.095) in cycles A (3.24 ± 0.37 s) and B (3.25 ± 0.22 s), whereas in cycle C (6.48 ± 0.39 s) the bundle lasted longer (p < 0.05), due to the presence of an additional, third burst in the reticulum. Generally, the amplitude and RMS of myoelectrical activity were higher in the reticulum than in the rumen, however, were shorter, regardless of the type of the reticulo-rumen cycles.4. DiscussionThe analysis of the physiological motor activity of the gastrointestinal tract of ruminants is an important element of both basic and applied research in the context of the clinical application [4,17,39,40,41,42,43]. The latest studies characterizing the motor activity of the forestomachs carried out using ultrasound techniques [44,45] give way to EMG studies of the gastrointestinal tract in sheep [4] and pigs [13,23], which were considered crucial for understanding the mechanisms regulating the motor activity of the gastrointestinal tract and for clinical practice.The obtained EMG recordings were subjected to a detailed off-line analysis taking into account the assessment of the parameters of bundles and, carried out for the first time, an analysis of the parameters of the bursts forming the bundles. To perform a comprehensive characterization of the physiological myoelectric activity of the smooth muscle of the rumen and the reticulum in sheep, a division of the reticulo-ruminal was proposed. The presented characteristic of cycles A, B, and C connects the available theoretical knowledge [5,15,30] with the novel, detailed pattern of EMG activity obtained during long-term registration in the reticulo-rumen.The general pattern of EMG signals presented in this paper is comparable to the characteristic of analog records obtained by Ruckebusch [5], Plaza, [4], and Kania et al. [43] in consecutive segments of the gastrointestinal tract. The average frequency of occurrence of the reticulo-rumen cycles at rest according to Ruckebusch and Bardon [5] is about 1/minute (0.8–1.2 cycles/minute), and according to Forbes and Barrio [30] 1–3 cycles/minute. On the other hand, Kania et al. [42] estimated that, physiologically, sheep are having around three reticulo-ruminal cycles within 2 min (1.5 cycles/minute). In our studies, the number of ruminal cycles remained constant 12 ± 0.2 to 16 ± 0.1 in 15 min intervals, which is in the range of 0.8–1.1 cycles/minute recorded at rest with no food stimulation. Therefore, it can be concluded that the detectability of the reticulo-rumen cycles based on myoelectrical activity registration by a telemetry method is comparable with the results obtained with other recording techniques at rest. It can be concluded that the facility and compatibility of obtained data allow their subject a detailed analysis of interrelationships. However, it should be kept in mind, that the pressure of the ingesta inside the reticulo-rumen and the chemical composition of the ingesta are triggers for the reticulo-rumen motility [1,2,3,4]. Therefore, the data presented here reflected only the basic reticulo-rumen motility pattern at rest, and could be the reason for the relatively low number of ruminal cycles observed over time. When the physiological motility without feed restriction was taken into account, the rumination time (total time spent to rumination per day) ranged from 6.53 to 8.21 h depending on the amount of fiber in the feed [3,46]. In the study presented here, at rest with no food stimulation, the rumination time appears quite short as estimated as 4.50 h.One of the most important parts of this study was to measure the reticulo-rumen myoelectrical activity in conscious, unrestrained sheep. Measuring methods recently used in ruminants, such as force transducer method [11,12], restrained electromyography [4], or ultrasonographic examination [7,8] required an animal restriction. On the other hand, an application of accelerometers in bolus wireless sensors [9,10] does not allow long-term registration. Therefore, since the myoelectric activity of smooth muscle of the reticulo-rumen reflects its contractile activity [1,2,3], telemetry EMG recording appears to be currently the only method that allows measures of long-term gastrointestinal motility on the unrestrained animal. Moreover, the use of the long-term telemetry method eliminates the greatest limitation in experimenting on the in vivo model in the form of registration and signal transmission to the data processing and archiving system. The long-term telemetry method provides continuous recording and real-time radio signal transmission from animals moving freely in their cages, in contrast to the cable method, which requires training of experimental animals and severe immobilization of them during periodic recording [4,22,24].The comprehensive characteristics of the myoelectric activity of the reticulo-rumen cycle were based on the analysis of the parameters of the EMG signal: duration of periods of activity (T) and inactivity (P), amplitude (A) and root mean square (RMS), not yet used for describing gastrointestinal tract EMG recording. The action potentials described in this way, recorded in the time domain, were considered as burst forming bundles [23,47]. This detailed description made it possible to unambiguously identify closely synchronized sequences of activities described by Ruckebusch and Tomov [15] including mixing (cycle A), gas eructation (cycle B), and rumination (cycle C) [7].The classification of the reticulo-rumen cycles described in this paper, supports the existing Ruckebusch and Bordon [5], Braun et al. [7,8], and Plaza et al. [4] classification and extends them with the possibility of evaluating individual bursts forming bundles. Braun [7] et al. analyzed the motor activity of a goat’s reticulum and confirmed the occurrence of 1-, 2-, or 3-phase contractions in the goat. However, he could not determine, using only an ultrasound examination, which contraction of the reticulum corresponds to the motor activity of the rumen, and thus which reticulo-ruminal cycle should be classified. The presented studies also confirmed the occurrence of bundles consisting of 1, 2, or 3 bursts in the reticulum, and the methodology used allowed them to be associated with the rumen contractions, and then classifying the bundles registered in the reticulum consisting of 1 and 2 bursts of activity to cycle A or B, while consisting of three bundles of activities up to cycle C. Available literature lacks publications describing in detail the myoelectric activity of bundles in each of the reticulo-rumen cycles. Julia and Latour [48] limited the automatic analysis of the myoelectrical activity of the gastrointestinal tract to the contraction of the reticulum, describing the duration of periods of activity and inactivity for bundles. Only Plaza et al. [4] conducted a more complex computerized analysis of the myoelectrical activity in the gastrointestinal tract of sheep. The authors implanted five triple nickel-chromium electrodes into the muscular membrane of the reticulum, dorsal rumen sac, omasum, abomasum, and duodenum for recording an analog signal using a 10-channel electroencephalograph. In the available literature, no reports are describing all discussed parameters characterizing the burst in each of the bundles, which significantly hinders the discussion. The results obtained for bundles in subsequent cycles A, B, and C correspond with the results obtained by other authors, while the characteristics of burst registered in the reticulum and the rumen were presented for the first time.Our studies showed the same duration of the bundles registered in the reticulum, in the A and B cycles, and significantly longer in the C cycle. Plaza et al. [4] obtained similar results for the bundles, however, they did not consider bursts’ parameters, due to method limitations. Additionally, Braun et al. [7], in the ultrasound examination, showed a significantly longer duration of the three-phase contraction of the reticulum (in the C cycle) compared to the single and two-phase contraction (in the A and B cycles). We demonstrated also, lower duration of bundles in the reticulum compared to the rumen in each reticulo-ruminal cycle, which corresponds with the result obtained by Plaza et al. [4].Analysis of signal amplitude and RMS confirmed that those parameters are relatively constant and did not fluctuate significantly inhomogeneous bundles and burst in the reticulum and rumen. In the study of Plaza et al. [4], the amplitude of bundles was averaged over the entire record and was not analyzed for individual bursts. In our study, the amplitude of the bundles in the reticulum was always higher than in the rumen, moreover, the amplitude of the burst forming bundles did not differ between the cycles for the reticulum and rumen, respectively. Similarly, the RMS of the bundles in the reticulum was always higher than in the rumen. The analysis of individual bursts showed an interesting difference between the first and second reticular contractions and the first and second rumen contractions is RMS, but not in amplitude. In recent studies, RMS was not used for the characterization of the motor activity of the forestomach [4,5,16,40]. Plaza et al. [4] describing the second contraction of the reticulum as stronger could not calculate the EMG signal strength in an analog data recording system on paper with a speed of 3–4 cm/min. Only the analysis of the digital signal carried out in the presented work allowed the separation of the amplitude and RMS (as the mean square after the amplitude in the frequency domain) as two dependent, but separately analyzed, parameters. Therefore, it can be concluded that the frequency of electrical events in a burst, as a variable affecting signal strength, may be an important regulator of the reticulum myoelectrical activity, in contrast to the rumen, in which no signal strength differences between burst in cycles B and C were found.The main limitation of this study is the lack of electrodes placed on the dorsal sac of the reticulum or the atrium. The lack of additional biopotential measurements results from the technical limitation of implantable telemetry for large animals to 3-channel transmitter implant. In the study presented here, three electrodes were positioned on the central section of the cranial part of the reticulum, on the ventral sac of the rumen, and the abomasum. Future research with modified arrangements of electrodes are required to achieve useful information about the contribution of the dorsal sac or the atrium to the reticulo-rumen motility. One of the solutions to this limitation may be an application of multimodal telemetry recordings combining simultaneous measurements of EMG and intraruminal pressure. With the development of technological advances, the availability of specific sensor-based technologies that can monitor these physiological changes increases [49]. Recently, multimodal telemetry was successfully applied to monitoring brain function and intracranial pressure in macaques [50], and may support a new approach in the minimally invasive monitoring of the reticulo-rumen activity in sheep.5. ConclusionsIt may be concluded that the long-term EMG telemetry registration is feasible in the assessment of the reticulo-rumen myoelectrical activity in sheep. The obtained results suggest that the EMG signal registered in the wall of the nonstimulated gastrointestinal tract in sheep is not random, but it occurs in an orderly array and percentage. Those EMG signals of each rumination cycle formed a specific pattern of the reticulo-ruminal motility during the rest period. The recognition of the A, B, and C reticulo-rumen cycles ineffective based on the EMG signal structure. The use of bursts parameters allows very detailed characteristics of the reticulo-rumen contractile pattern. These results may be the basis for researching the regulation of gastrointestinal motility using the long-term telemetry method.

animals : an open access journal from mdpi

10.3390/ani11092500

PMC8468924

Widespread media reports suggest that unusually high numbers of the public purchased, or sought to purchase, puppies following the first ‘lockdown’ phase of the COVID-19 pandemic in the UK. This study aimed to explore this phenomenon by comparing the reasons why, the methods how, and by whom “Pandemic Puppies” were purchased during this period (from 23 March 2020–31 December 2020), and compare these responses with owners who purchased their dog during the same date-period in 2019. Valid responses were analysed from owners of 1148 puppies from 2019 and 4369 Pandemic Puppies. Key differences included Pandemic Puppy owners being more likely to be first-time dog owners, have children in their household, pay a deposit without seeing their puppy, collect their puppy from outside their breeders’ property, see their puppy without their littermates, and pay > £2000 for their puppy, compared with 2019 puppies. Over 1 in 10 Pandemic Puppy owners had not considered purchasing a puppy before the pandemic, while 2 in 5 felt their decision to purchase a puppy had been influenced by the pandemic, most commonly due to having more time to care for a dog. Changes in puppy purchasing during the pandemic raise a range of welfare concerns including relinquishment, behavioural problems and poor health.

Widespread media reports suggest that unusually high numbers of the public purchased, or sought to purchase, puppies following the first ‘lockdown’ phase of the COVID-19 pandemic in the UK, dubbed “Pandemic Puppies”. This study aimed to explore this phenomenon by comparing pre-purchase motivations and behaviours, and purchase behaviours of UK owners purchasing puppies aged < 16 weeks from 23 March 2020–31 December 2020 with those of owners who purchased puppies during the same date-period in 2019. An online survey was conducted during November-December 2020, from which 5517 valid responses were analysed (2019 puppies: n = 1148; 2020 ‘Pandemic Puppies’: n = 4369). Over 1 in 10 Pandemic Puppy owners had not considered purchasing a puppy before the pandemic, and 2 in 5 felt their decision to purchase a puppy had been influenced by the pandemic, most commonly by having more time to care for a dog (86.7%). Multivariable logistic regression models revealed that Pandemic Puppy owners were more likely to be first-time dog owners and have children in their household, were less likely to seek out a breeder that performed health testing on their breeding dog(s) or view their puppy in-person, and were more likely to pay a deposit without seeing their puppy. At purchase, Pandemic Puppies were more likely to be younger, delivered or collected from outside their breeders’ property, seen without their littermates, and cost > £2000 compared with 2019 puppies. Changes in puppy purchasing during the pandemic raise welfare concerns for this unique population, including relinquishment, behavioural problems and poor health.

1. IntroductionEfforts to suppress the spread of the SARS-CoV-2 coronavirus in the UK during 2020 involved periods of national and local lockdown along with other restrictions that changed the lifestyles of UK households (see Figure 1). These periods of variable confinement, whereby a large proportion of the public spent several months restricted at home with limited external social contact, commenced on 23 March 2020 with a ‘stay at home’ order across the whole of the UK [1]. During April to May 2020, media and Non-Governmental Organisation (NGO) reports, began to emerge suggesting that unusually high numbers of the UK population had sought to purchase a puppy, dubbed the “Pandemic Puppy” phenomenon [2]. The true extent of puppy purchasing is challenging to quantify in the UK owing to the largely ‘cottage industry’ nature of dog breeding and selling, and lack of a national dog licensing scheme. However, as a proxy, data on activity on puppy-finding websites supported a marked increase in puppy purchasing intention. The Kennel Club (KC) reported that searches using its “Find a Puppy” online tool increased by 168% between 23 March and 31 May 2020 compared to the same period in 2019, with May 2020 alone accounting for a 237% increase during a period when the majority of UK residents continued to work from home [3,4]. Similarly, the Dogs Trust reported that online searches regarding puppy acquisition increased by 120% in April 2020 [5] whilst statistics from the People’s Dispensary for Sick Animals (PDSA) and the Brachycephalic Working Group (BWG) demonstrated that internet searches via Google for ‘buying a puppy’ increased by 175% within one month of the UK lockdown [4,6]. Rising demand for dog ownership in April and May 2020 has been reported internationally [7,8].This ‘Pandemic Puppy’ phenomenon has given rise to concerns that high demand for puppies during this period may have exhausted supply from good welfare sources and thereby pushed prospective puppy buyers towards purchasing from unscrupulous breeders and sellers hoping to ‘cash in’ on the phenomenon by producing large numbers of puppies under low welfare conditions, including puppy farms and via illegal importation at inflated prices [4]. In addition, concerns were raised over a rising risk of impulse purchasing of puppies, particularly by households unsuited to maintaining an appropriate human lifestyle for dog ownership after lockdown restrictions were lifted [2]. Impulse buying, defined as unplanned behaviour involving quick decision-making and a tendency toward immediate acquisition of a product [9], was observed for a variety of products perceived to be scarce during the COVID-19 pandemic [10]. Impulse purchasing of puppies pre-dates the pandemic, with the power of the desire to ‘own a puppy right now’ eclipsing rational human thought, with the inherent “cuteness” of puppies at weaning age [11] considered a facilitator of such purchases [12]. Impulsive purchasing of puppies during the pandemic has been reported [2], but has not been evidenced by scientific study and comparison with impulsivity of puppy purchases prior to the pandemic.Research has demonstrated that many aspects of the puppy-purchasing process, from decision-making through the actual purchasing behaviour, were already suboptimal in many respects in the UK prior to the pandemic [13]. In one UK study (n = 1844), one quarter (461/1844) of puppies were acquired under eight weeks of age and 8.1% were obtained without viewing their mother (dam) [14]. In a further study of UK dog owners (n = 1427), one sixth (15.7%) of owners admitted that they had carried out no pre-purchase research, one quarter of owners purchased their puppy on first visit (24.7%), and nearly half (46.1%) of owners did not ask to see health records for either the dam or sire of their puppy [15]. Such sub-optimal purchasing behaviours leave owners vulnerable to purchases from unscrupulous breeders and suppliers, including illegal imports and puppy farms [16], who breed to meet demand rather than good health or welfare, and which can have long-term negative consequences for behaviour and welfare in these dogs [17,18,19,20].A plethora of studies on the impact of COVID-19 on canine welfare have been published during 2020 and 2021; however, these papers have largely focused upon dogs that were purchased and owned pre-pandemic. Specific topics have included dog-owner relationships [8,21,22,23,24,25], the social support provided by dogs to owners [22,24,25,26,27,28,29], owner concerns over caring for their dog and accessing veterinary services [21,22,23,28,29,30,31], adoption from and relinquishment to rescue centres [8], along with changes in dogs routines and behaviour [23,29,32,33]. Thus, despite widespread interest and concern over the Pandemic Puppy phenomenon, there is a critical data gap on puppy purchasing during the pandemic.The present study aimed to explore the impact of the COVID-19 pandemic upon puppy purchasing in the UK. Using a cross-sectional analysis of a national survey, we sought to:(i)describe the pre-purchase motivations and behaviours and purchase behaviours of owners of puppies purchased 23 March–31 December 2020;(ii)compare data from (i) with the motivations and behaviours of owners of puppies purchased during the same date frame in 2019.2. Materials and Methods2.1. Survey Design and ContentAn online questionnaire was designed to explore the pre-purchase and purchase motivations and behaviours of UK puppy purchasers. Questions were designed iteratively amongst the authors and piloted on a small number of respondents to ensure ease of understanding and comprehensiveness of scope. The survey included five broad sections: (1) General owner demographics e.g., gender, age, household members, prior dog ownership; (2) General puppy demographics e.g., breed, sex; (3) Pre-purchase motivations e.g., factors influencing choice of breed; (4) Pre-purchase behaviours e.g., research conducted; (5) Purchase behaviours e.g., requests for health records, cost of puppy, which parents of their puppy were seen (if any). Owners of Pandemic Puppies were additionally directed to a set of questions exploring COVID-specific impacts. This study received ethical approval from the Social Science Research Ethical Review Board at the Royal Veterinary College (URN: SR2020-0259). To reduce missing data for core study questions, ‘survey logic’ was used to offer every owner the option to complete just core questions, or core questions plus an extended set of questions (Figure 2). The full survey is included in Supplementary Material S1 File.2.2. Participant RecruitmentThe questionnaire was hosted on SurveyMonkey and was open from 10 November to 31 December 2020. The survey was distributed by snowball sampling via a wide range of sources, including social media, the veterinary, canine, and general press (including radio interviews and printed articles), and through key stakeholders including the commercial sector (e.g., insurance bodies, canine registration organisations and puppy-selling websites) and charity sector (e.g., animal welfare and rehoming charities).Respondents were required to give informed consent for their data to be held on a secure server in accordance with UK GDPR legislation. IP addresses were used to eliminate duplicate responses prior to permanent deletion.Inclusion CriteriaRespondents were required to:Be over 18 years of ageResident in the UKHave brought home a puppy aged under 16 weeks at any date during 2019 or 2020 (N.B. although all puppy purchases in 2019–2020 were eligible for the broader project, responses analysed in this study were limited to owners who brought home their puppy between 23 March–31 December during 2019 or 2020).Have purchased their puppy rather than rehomed or bred the puppy themselves.Where participants had purchased more than one puppy during the period, they were asked to answer for the youngest at the time of the survey and, in the case of co-purchased littermates, owners were asked to answer for the dog whose name came first alphabetically.2.3. Data CleaningThe raw survey data were exported from SurveyMonkey into Microsoft Excel for manual data cleaning prior to analysis, including removing responses from duplicated IP addresses (where the more complete response was retained), responses with no data beyond the consent and inclusion criteria stage, and respondents that did not meet inclusion criteria.2.4. Spatial AnalysisParticipants were asked to provide the first half of their postcode to allow assessment of the representativeness of the study sample to the UK population. These partial postcode data were checked for validity against the Office for National Statistics (ONS) National Statistics Postcode Lookup (NSPL) May 2021 data [34] and were allocated to one of the 12 UK regions. Regional response rates were calculated as (total responses from a region in a year/ONS region population [35]) ×100,000. Choropleth maps were produced using ArcGIS 10.2 (Environmental Systems Research Institute, Redlands, California, CA, USA) to show the regional response rates per 100,000 population for 2019 and 2020, together with the percentage increase in response rate between the two years.2.5. Qualitative Content Analysis of Free-Text OptionsIn addition to fixed-choice responses to multiple-choice questions, many questions included an ‘Other, please specify’ answer to capture novel owner insights. To allow these responses to be quantified alongside existing fixed-choice responses, free-text data were coded using qualitative content analysis [36]. Three authors (R.M.A.P., Z.B. and C.L.B.) familiarised themselves with the data by reading all free-text responses to the multiple-choice option ‘Other, please specify’. Two of these authors (Z.B. and C.L.B.) independently used an inductive approach similar to that explained elsewhere to develop a coding framework for each question [37]. Based on the two sets of independently derived codes, Z.B. and C.L.B. agreed an overall set of categories that were finalised by R.M.A.P. and applied to the free-text responses by C.L.B., C.L.P. and R.M.A.P. Where free-text responses were deemed to fit within the scope of existing fixed-choice responses, considered to be deductive categories, data were back allocated into that category if not already selected. The inductive coding framework for each question including free-text responses can be found in Supplementary Material S2 File.2.6. Quantitative AnalysisFollowing cleaning in Excel, data were imported into IBM SPSS Statistics v27 (SPSS Inc, Chicago, IL, USA). Initial data analysis included calculation of descriptive statistics (frequency and percentage) for all variables. Data from puppies purchased between 23 March–31 December 2019 (“2019 puppies”) and puppies purchased between 23 March–31 December 2020 (“Pandemic Puppies”) were compared at the univariable level using chi-squared (X2) analysis for categorical variables and Mann-Whitney U tests for non-normally distributed continuous data (with data distribution ascertained by visual inspection of histograms). Variables liberally associated with acquisition year in the univariable analyses (p < 0.2) were included in four multivariable binary logistic regression models describing (i) demographics, (ii) pre-purchase motivations, (iii) pre-purchase behaviour, and (iv) purchase behaviour, with year of acquisition (2019 vs. 2020) as the binary outcome. Model development used manual backwards stepwise elimination. Confounding was assessed for all variables retained in the final models through addition of each independent variable in a stepwise manner to the model and assessing for substantial (>20%) change in OR of any other variable when each new variable was added to the model [38]. Collinearity of variables was assessed through evaluation of the correlation matrices, the variance inflation factor (VIF) and tolerance [39]. Interactions between all independent variables in the final models were assessed for significance. The Hosmer-Lemeshow test was used to evaluate the quality of the model fit. Statistical significance was set at the 5% level.3. ResultsIn total, n = 7545 responses were returned to the survey. Following cleaning, n = 96 responses were removed due to duplicated IP addresses, n = 799 responses were removed that held no data beyond the consent and inclusion criteria, n = 123 were removed due to not meeting inclusion criteria (of which n = 53 had brought their puppy home after the age of 16 weeks, n = 30 had purchased their puppy from another owner choosing to relinquish their puppy, n = 25 had adopted their puppy rather than purchased it, n = 8 who were yet to bring their puppy home, n = 5 had bred the puppy themselves, n = 1 who had purchased their puppy from a third-party seller, and n = 1 whose puppy was part of a ‘puppy walking scheme’ for an assistance dog charity). Of the remaining valid sample (n = 5517), there were 1148 puppies (20.8%) purchased between 23 March–31 December 2019 (“2019 puppies”) and 4369 puppies (79.2%) purchased between 23 March–31 December 2020 (“Pandemic Puppies”).3.1. Spatial AnalysisAll UK regions were represented in the sample, with geographical distribution of respondents by year and changes between years described in Figure 3. There was no significant difference in geographical distribution of the sample compared with Office for National Statistics population data for mid-2020 (t = 0.013, df = 11, p = 0.990). Owners of Pandemic Puppies were significantly less likely to live in Scotland (2019: 15.8% vs. 2020: 9.2%) and more likely to live in London (2019: 6.8% vs. 2020: 20.7%; X2 = 64.41, p < 0.001) than 2019 puppies.3.2. Owner Demographics and LifestyleOwners were predominantly female, with no difference in gender distribution between 2019 puppy and Pandemic Puppy owners (female, 2019: 92.0% vs. 2020: 90.0%; X2 = 6.61, p = 0.202). The most common age group of owners was 45–54 years old (2019: 25.2% vs. 2020: 24.6%) followed by 25–34 years (2019: 23.4% vs. 2020: 24.1%), with no difference in age between Pandemic Puppy and 2019 puppy owners (X2 = 11.77, p = 0.067). The majority of owners were the primary carer for their puppy (i.e., providing the majority of care such as feeding and walking; 2019: 62.9% vs. 2020: 57.8%); however, Pandemic Puppy owners were significantly more likely to share the caring role with someone else in their household (2019: 34.1% vs. 2020: 39.7%; X2 = 11.28, p = 0.024).3.2.1. Experience with DogsPandemic Puppy owners were less likely to have previously owned a dog, with 59.7% of Pandemic Puppy owners having previous dog ownership experience compared with 66.7% of 2019 puppy owners (X2 = 16.90, df = 2, p < 0.001). There was no difference in proportion between owners of Pandemic Puppies and 2019 puppies that had grown up with a dog (2019: 68.1% vs. 2020: 69.3%, respectively; X2 = 0.48, df = 1, p = 0.487). Owners of Pandemic Puppies were less likely to be employed in the canine and/or animal care sector (e.g., veterinary surgeons, veterinary nurses or dog trainers) compared with 2019 puppy owners (2019: 17.9% vs. 2020: 10.0%; X2 = 49.02, df = 2, p < 0.001).3.2.2. Household DemographicsDemographics of Pandemic Puppy households differed in several ways to those of 2019 puppy households at the univariable level. Pandemic Puppies were more likely to live in households with both adults and children compared to 2019 puppy owners (2019: 26.8% vs. 2020: 33.2%) (X2 = 15.64, df = 3, p < 0.001). Of those households with children, children in Pandemic Puppy households were more likely to be younger, with half aged 5–10 years (2019: 37.8% vs. 2020: 50.3%; X2 = 14.69, p < 0.001), and less likely to be older teenagers aged 16–18 years (2019: 35.4% vs. 2020: 24.8%; X2 = 13.69, p < 0.001). Pandemic Puppies were more likely to be the only dog in the household than 2019 puppies, with 70.1% of Pandemic Puppies living in single-dog households vs. 58.7% of 2019 puppies (X2 = 54.62, df = 3, p < 0.001). The majority of Pandemic Puppies lived in households with access to a private garden or yard (95.1%), which did not differ from 2019 puppies (96.9%; X2 = 7.78, df = 3, p = 0.051).3.2.3. Impact of COVID on Household LifestyleAt the univariable level, during the 2020 phase of the pandemic, Pandemic Puppy owners were more likely to have worked from home (2019: 51.8% vs. 2020: 57.3%; X2 = 10.01, df = 1, p = 0.002) and to have been home-schooling children (2019: 22.7% vs. 2020: 27.5%; X2 = 9.11, df = 1, p = 0.003) compared with owners’ of 2019 puppies. In contrast, during the 2020 phase of the pandemic, significantly fewer owners of Pandemic Puppies had been furloughed from their job compared with owners’ of 2019 puppies (2019: 30.7% vs. 2020: 24.6%, respectively; X2 = 15.63, df = 1, p < 0.001). There was no difference in the proportion of owners who became unemployed due to COVID-19 between Pandemic Puppy owners and 2019 puppy owners (2019: 5.8% vs. 2020: 6.0%; X2 = 0.07, df = 1, p = 0.794), in the proportion considered to be keyworkers (2019: 34.8% vs. 2020: 33.7%; X2 = 0.43, df = 1, p = 0.513), or to have another household member considered to be a keyworker (2019: 24.8% vs. 2020: 24.3%; X2 = 0.10, df = 1, p = 0.748).3.3. Puppy DemographicsDetailed breakdown of puppy demographics will be provided in a follow-on publication. For context, the majority of puppies in the overall study were purebred (72.0%) (i.e., of a recognised breed). The proportion of purebred puppies in the 2020 Pandemic Puppies population was lower than for the 2019 puppies population (2019: 78.7% vs. 2020: 70.3%; X2 = 32.484, df = 1, p < 0.001), with a corresponding increase in designer crossbred puppies in the Pandemic Puppies population (2019: 18.8% vs. 2020: 26.1%; X2 = 27.668, df = 1, p < 0.001). There was a significant reduction in puppies registered with The Kennel Club in the Pandemic Puppy population compared to the 2019 puppies population (2019: 58.2% vs. 2020: 46.2%; X2 = 54.1, p < 0.001). There was no difference in sex distribution between the 2019 puppies and Pandemic Puppies (male, 2019: 51.7% vs. 2020: 53.4%, X2 = 1.12, df = 1, p = 0.290).3.4. Owner Demographics and Lifestyle: Multivariable AnalysisMultivariable logistic regression modelling identified four variables related to owner demographics and lifestyle with significant association with Pandemic Puppies (Table 1). The Hosmer-Lemeshow test indicated acceptable model fit (p = 0.904).3.5. Pre-Purchase MotivationsCompanionship for the owner was the most common reason cited as to why prospective owners wanted to purchase a dog in both 2019 Puppy and Pandemic Puppy owners (Table 2). Pandemic Puppy owners were significantly more likely to cite exercise encouragement, improving their/their family’s mental health and companionship for their children as reasons their household wanted to acquire a dog compared with 2019 owners (Table 2). Conversely, a significantly lower proportion of Pandemic Puppy owners cited companionship for their other dog(s), a specific working role (e.g., gundog) or a non-working role (e.g., dog sports) compared with 2019 owners. Puppy acquisition was less likely to be driven solely by the responding owner in the Pandemic Puppy population (2019: 46.6% vs. 2020: 41.4%, X2 = 9.26, p = 0.002), with a household-wide equal desire to acquire a puppy more common in the Pandemic Puppy population than in the 2019 puppy population (2019: 41.3% vs. 2020: 45.6%, X2 = 5.36, p = 0.021).When deciding which breed or crossbreed to purchase, the most commonly cited characteristics sought by owners were being a good companion (most common reason in 2019 and 2020) and size being suited to owner lifestyle (second most common reason in 2019 and 2020). At the univariable level, Pandemic Puppy owners were significantly more likely to seek a breed/crossbreed whose size suited to their lifestyle, that they believed was good with children, was easy to train, that friends or family members currently owned and that was hypoallergenic, compared with 2019 puppy owners (Table 3). Conversely, 2019 puppy owners were significantly more likely to seek a puppy from a breed/crossbreed they had owned before or had working abilities or, conversely, had no specific characteristics in mind when looking for a breed/crossbreed, compared with Pandemic Puppy owners (Table 3).When considering the characteristics that owners sought out in a breeder, the most common characteristics cited were consistent between 2019 and 2020, namely a breeder who would allow them to see the puppies’ mother (dam), who they felt cared for their dog, and that they felt was trustworthy. At the univariable level, Pandemic Puppy owners were significantly more likely to seek out a breeder they felt cared for their dogs and who communicated well with them, but were significantly less likely to seek out a breeder that performed health tests for the breed/crossbreed they wanted, who was a member of The Kennel Club Assured Breeders Scheme, or bred from dogs that had been awarded prizes at dog shows, compared with 2019 puppy owners (Table 4).Pre-Purchase Motivations: Multivariable AnalysisMultivariable logistic regression modelling identified eleven variables related to pre-purchase motivations with significant association with Pandemic Puppies (Table 5). The Hosmer-Lemeshow test indicated acceptable model fit (p = 0.939).3.6. Pre-Purchase BehavioursPre-purchase research was more commonly carried out by Pandemic Puppy owners (+11.4%) and was reported by 58.1% of Pandemic Puppy owners compared with 46.7% of 2019 puppy owners. However, this difference was in part explained by the proportion of owners in each cohort who considered themselves to be an experienced dog owner who did not need to conduct pre-purchase research, which was 11.4% higher in 2019 puppy owners (50.3%) than Pandemic Puppy owners (38.9%) (X2 = 49.76, df = 2, p < 0.001). Only 3.0% of owners in each year reported that they conducted no pre-purchase research. Of those owners who conducted pre-purchase research, the most common sources were friends or family who own or had owned a dog (most common source in 2020 and joint most common source in 2019), alongside breed/crossbreed specific online resources (e.g., website/forum) and The Kennel Club website (second most common source in 2020). At the univariable level, Pandemic Puppy owners were significantly more likely to conduct pre-purchase research by talking to friends or family who own or had owned a dog before, using breed/crossbreed-specific online resources (e.g., forums, websites), using animal charity websites and reading books than 2019 puppy owners (Table 6).Owners found their puppy most commonly via animal selling websites in 2019 and 2020, followed by their puppy’s breeder being somebody they already knew, such as a colleague, friend/family or someone they had bought a puppy from before. At the univariable level, Pandemic Puppy owners were more likely to find their puppy via animal selling websites or via breeder recommendations from friends, and significantly less likely to find their puppy by already knowing their puppy’s breeder, via The Kennel Club website, or from a physical advertisement compared to 2019 owners (Table 7).The interval from prospective owners’ initial decision to look for a puppy to when their puppy was brought home was most commonly between 1–6 months for both 2019 and 2020. At the univariable level, Pandemic Puppy owners were more likely to take longer from their decision to look for a puppy to acquiring their puppy than 2019 puppy owners, although this difference was small (X2 = 36.28, df = 4, p < 0.001; Figure 4). Around three-quarters of owners did not join a waiting list for their puppy in either year (2019: 73.0% vs. 2020: 71.5%) with no difference in levels between Pandemic and 2019 puppy owners (X2 = 0.76, df = 3, p = 0.683).At the univariable level, Pandemic Puppy owners were more likely to put down a deposit before they saw their puppy (2019: 8.9% vs. 2020: 17.2%; X2 = 81.87, df = 5, p < 0.001), and less likely to never be asked to put down a deposit for their puppy at any point during the puppy-purchasing process (2019: 37.5% vs. 2020: 28.6%) (Figure 5).The majority of owners visited what was reported to be their puppies’ breeders house in-person prior to the day they brought the puppy home; however, this was significantly lower (−21.0%) in Pandemic Puppy owners compared to 2019 puppy owners (2019: 80.6% vs. 2020: 59.6%; X2 = 165.46, p < 0.001), with a corresponding significant increase (+22.4%) in owners using live video calls with their breeder (2019: 6.5% vs. 2020: 28.9%; X2 = 233.73, p < 0.001) or in breeders using photos/pre-recorded videos of what was claimed to be their puppy (+20.2% increase) (2019: 31.3% vs. 2020: 51.5%; X2 = 141.34, p < 0.001) (Table 8). The median (25th–75th centile) number of owner visits to see their puppy in-person prior to being brought home was 2 (1–3) visits for 2019 puppies, but was significantly lower for Pandemic Puppies at 1 (1–2) visit (U = −14.88, p < 0.001). Correspondingly, the median (25th–75th centile) number of live video calls with the breeder of Pandemic Puppies was significantly higher at 0 (0–3) calls compared to 0 (0–0) calls for 2019 puppies (U = 14.94, p < 0.001). One quarter of owners were not questioned on their suitability as a dog owner before their breeder agreed to sell them their puppy (2019: 75.6% vs. 2020: 75.2%); however, 2019 owners were less likely to remember whether the breeder asked them this question (“I don’t remember”, 2019: 4.3% vs. 2020: 2.7%; X2 = 9.24, df = 2, p = 0.011).The majority of owners purchased their first choice of breed; however, owners of Pandemic Puppies were less likely to purchase their first-choice breed (2019: 91.6% vs. 2020: 86.6%; X2 = 42.09, df = 10, p < 0.001), with the most common reasons being that their first-choice breed was too expensive or they were unable to find a seller that had puppies available at the desired time of purchase (Table 9).Pre-Purchase Behaviours: Multivariable AnalysisMultivariable logistic regression modelling identified eight variables related to pre-purchase behaviour with significant association with Pandemic Puppies (Table 10). The Hosmer-Lemeshow test indicated acceptable model fit (p = 0.689).3.7. Purchase BehavioursThe most common location for owners to collect their puppy was the inside of what was reported to be the breeders home; however, at the univariable level there was a significant decrease (−33.7%) in this behaviour between 2019 and 2020 (2019: 84.7% vs. 2020: 51.0%; X2 = 407.7, p < 0.001), with a corresponding 24.3% increase in puppies being collected from outside their breeder’s home, for example their doorstep or garden (2019: 5.5% vs. 2020: 29.8%; X2 = 277.0, p < 0.001). Although rare, at the univariable level, Pandemic Puppies were significantly more likely to be delivered to their owners property, or collected from a service station or car park, compared with 2019 owners (Table 11). The vast majority of 2019 puppy (96.6%) and Pandemic Puppy (96.5%) owners were happy with the location they collected their puppy from, which did not significantly differ between years (X2 = 0.035, df = 1, p = 0.851).The proportion of owners not asking their breeder to see any information related to the health testing related to their puppy’s parents increased significantly between 2019 and 2020, for both DNA (genetic) tests (+5.7% not asking; X2 = 20.90, df = 3, p < 0.001) and veterinary screening tests (e.g., hips, elbows, knees, eyes, respiratory testing; +10.4% not asking X2 = 42.19, df = 3, p < 0.001) at the univariable level (Table 12).On the day that the puppies were finally acquired, puppies were most likely to be seen with their mother (dam) followed by their littermates); however, this was significantly lower for Pandemic Puppies, with over a quarter seen without their littermates (27.9%, −12.8% less than 2019 puppies) (Table 13), and one quarter seen without their dam (24.9%, −10.6% less than 2019 puppies). Pandemic Puppies were also significantly less likely to be seen with another dog(s) they were not related to, their sire, other adult dog(s) they were related to, or other puppies (that the owner was unsure whether they were littermates with), with a corresponding +7.8% increase in puppies seen on their own between 2019 and 2020 (2019: 4.9% vs. 2020: 12.7%; X2 = 52.42, p < 0.001).Around one in seven puppies came from a breeder that had another litter for sale at the same time as the puppy that was purchased (of the same breed, 2019: 8.7% vs. 2020: 10.1%; of a different breed, 2019: 4.6% vs. 2020: 4.4%); however, this did not differ between Pandemic Puppies and 2019 puppies (X2 = 2.50, df = 4, p = 0.645). Pandemic Puppies were more likely to be purchased at a younger age, with two thirds (67.3%) purchased aged 7–8 weeks compared to 52.5% of 2019 Puppies (X2 = 113.01, p < 0.001) (Figure 6). A small minority of owners felt pressured by their puppy’s breeder to commit to purchasing their puppy (2019: 1.5% vs. 2020: 2.3%), which did not differ between 2019 puppies and Pandemic Puppies (X2 = 5.16, df = 3, p = 0.161).Owners of Pandemic Puppies paid significantly more for their puppy than 2019 owners, with one quarter of Pandemic Puppy owners (24.3%) paying £2000–2999 for their puppy compared to just 1.8% of 2019 puppy owners (X2 = 636.8, df = 6, p < 0.001, Figure 7).Around two fifths (40.2%) of owners claimed to have heard of The Puppy Contract before purchasing their puppy; levels of awareness did not differ between 2019 puppy owners and Pandemic Puppy owners (2019: 41.5% vs. 2020: 40.2%; X2 = 0.53, df = 1, p = 0.466). Of those owners who were aware of The Puppy Contract (n = 1858), two thirds used the contract during their purchase, with levels of use the same across 2019 puppy and Pandemic Puppy purchasers (63.0% vs. 62.8%; X2 = 0.01, df = 1, p = 0.952). The most common reason for not using (either choosing not to use or being unable to use) The Puppy Contract was a belief that it was not needed as the owners were confident in their own purchasing decision, followed by the breeder not offering to use The Puppy Contract. Pandemic Puppy owners were significantly more likely to find their breeder did not agree to use The Puppy Contract when asked (Table 14).Purchasing Behaviours: Multivariable AnalysisMultivariable logistic regression modelling identified seven variables related to purchase behaviour with significant association with Pandemic Puppies (Table 15). The Hosmer-Lemeshow test indicated acceptable model fit (p = 0.972).3.8. Influence of COVID on Puppy-Purchasing Decisions in 2020Of those households that purchased puppies during the 2020 Pandemic, 11.0% had not considered purchasing a puppy before the COVID-19 pandemic (n = 424), and a further 0.5% (n = 21) were unsure whether they had. More than two in five Pandemic Puppy purchasers felt their decision to purchase a puppy had been influenced by the COVID-19 Pandemic (41.5%, n = 1596), with a further 2.4% unsure as to whether it had influenced their decision (n = 106). Of those 2809 owners who felt the pandemic had influenced their decision, reasons are listed in Table 16, the most common of which was having more time to care for a dog (86.7%, n = 1378).4. DiscussionThis study characterises the nature of the Pandemic Puppy phenomenon in the UK and reveals some remarkable insights into how puppy purchasing in the UK has changed following the start of the first lockdown on 23 March 2020. These findings indicate that the surge in puppy purchasing reported during the 2020 phase of the pandemic was not simply pre-pandemic puppy purchasing at a greater scale, but that the COVID-19 pandemic was associated with several critical changes in why and how puppies were purchased, and by whom. Although many elements of the puppy purchasing process remained constant from 2019 to 2020, some of the pandemic-related changes documented in this study pose increased risks to the long-term welfare of puppies purchased during this period and are of practical importance. The pandemic prompted many households that had never previously considered owning a dog to purchase a puppy during this unprecedented period, with 2 in 5 Pandemic Puppy purchasers feeling that the pandemic had influenced their purchase in some way, predominantly because they had more time to care for a dog. Given that public health restrictions related to COVID-19 relaxed during July 2021, including a reduced emphasis on working from home as restrictions relax, and the Government furlough is scheme is due to end in September 2021, temporarily increased time to care for a dog acting as a major driver of puppy acquisition may result in increased relinquishment risk of this unique Pandemic Puppy population. Lack of time is a consistently documented risk factor for relinquishment of dogs [40,41,42,43,44,45,46], commonly in dogs 2 years of age or younger (70%) [41], and often precipitated by changes in the household [45]. It is possible that some societal changes seen during the pandemic in the UK, particularly around working from home, may continue permanently in some form for many employees [47], which may facilitate increased levels of dog ownership in the UK population. However, many roles including those in retail, manufacturing, leisure and hospitality will still require employees to travel to their place of work [48] and thus there are likely to be individual and regional differences in time-related relinquishments of puppies or dogs. Media reports suggest that increased relinquishment of Pandemic Puppies may have already begun [49,50,51,52,53]. Enhanced support mechanisms for Pandemic Puppy owners to avoid relinquishment of their dog (where this is in the dog’s best interests) are urgently needed, including implementation of training to help dogs to be left alone without distress, and awareness (and potentially increased provision) of services such as day care and dog walkers to avoid dogs being left alone for long periods of time, an existing problem for an estimated one in five of the UK dog population prior to the pandemic [54].Mental health in the wider UK human population has deteriorated during the pandemic, with life satisfaction and happiness decreasing and anxiety levels increasing in adults [55]. Several findings from the current study indicate that Pandemic Puppies were often purchased in an attempt to mitigate some of these mental health challenges: Pandemic Puppy owners commonly sought out a dog to improve their own or their family’s mental health, significantly more than by 2019 puppy owners. When reporting on whether and why the pandemic influenced their puppy purchase, three of the top five reported reasons had a mental health component: to have ‘something happy’ to focus on, a reason to go outside to exercise more and a desire for more company due to being at home more. These findings are in agreement with many of the mental health benefits reported by existing dog owners during the pandemic, including making isolation easier to tolerate via offering companionship, having someone to talk to, physical connection and protection against loneliness, improving wellbeing via offering routine, purpose and motivation for life, exercise encouragement, and increased opportunity to socialise [34,56]. Some of these benefits were promoted to prospective puppy owners via media reports during the pandemic [57,58].Although dogs play a variety of roles in 21st century society, restrictions on activity and social contact during the pandemic may have limited some of these functions. In support of this view, Pandemic Puppy purchasers were less likely to seek out puppies for specific non-working roles such as dog sports and conformational showing, which may have changed the demographics of the puppy population during 2020. Instead Pandemic Puppy owners increasingly selected their puppy’s breed/crossbreed based on anthropocentric lifestyle factors including being a size suited to the owners’ lifestyle, being easy to train and being good with children. more commonly than 2019 owners. International evidence suggests that shorter and smaller breeds are becoming more popular over time, supporting the concept that people attempt to purchase dogs that can fit their lifestyle niche [59]. The desired features of Pandemic Puppies could reflect the relative inexperience of this owner population. Puppy purchasing during the pandemic appears to have been driven partially by individuals and families seeking to purchase a dog for the first time, with two in five Pandemic Puppy owners having no previous dog ownership experience compared with one in three of the 2019 owners. The relative inexperience of Pandemic Puppy owners may pose risks to the future behaviour, welfare and homing status of this puppy population. First-time dog ownership is associated with an increased risk of relinquishment [60,61], and it has been argued that owners’ knowledge and attitudes, including ignorance of species-specific behaviours, unrealistic expectations for the roles pets play in children’s lives, and the expense and time required for dog ownership and caretaking contribute to this high relinquishment [61]. First-time dog ownership has been associated with increased perceived ‘costs’ (or burden) of dog ownership [62], which may reflect a lack of understanding of the day-to-day responsibility of dog ownership prior to acquisition. In addition, first-time dog owners are more likely to report problem behaviours in their dog, including aggression [63], separation-related problems [63], noise phobias [63,64,65], nervousness [66] and overexcitement [63,66]. Whether these perceived problems represent true behavioural pathologies or are within the realm of ‘normal’, non-pathological dog behaviour, with new owners more inclined to consider them undesirable due to their inexperience, has been speculated as a reason for this association [63,66]. Previous data indicate that greater animal care knowledge is associated with an increased expectation for effort required in owning a dog [67], and those with arguably the most experience of dogs—professionals from the canine and/or animal care sector—were significantly less likely to be Pandemic Puppy owners. This may also reflect concerns over the Pandemic Puppy phenomenon from informed parties, who may have greater awareness of welfare risks to puppies and breeding dogs during this period of high demand. Indeed, similar avoidance of other companion animal acquisition phenomena that put animal welfare at greater risk have been reported in this professional group, for example, a reduced preference towards brachycephalic cats and rabbits [68,69]. Given the relatively inexperienced demographic of Pandemic Puppy owners, and the consequent potential negative outcomes for dog and owner, provision of increased support and education for Pandemic Puppy owners is likely to be required by canine behaviour and welfare professionals and organisations to maintain the welfare of this vulnerable puppy population.Owners of Pandemic Puppies were more likely to live in households with children. Despite this, only 4.1% of this group explicitly stated that they purchased a puppy to keep their children busy during the pandemic. However, compared with 2019 owners, Pandemic Puppy owners were more likely to want to own a dog to improve their/their family’s mental health. In addition, Pandemic Puppy owners were more likely to desire a dog breed they perceived to be good with children compared with 2019 owners. Although dogs may have offered positive benefits to families during lockdown [56], public health measures to limit the spread of COVID in the UK were associated with an increase in paediatric dog bite emergency department attendances, potentially due to children spending more time at home, with greater exposure to dogs [70]. Bites peaked during May to July 2020, with the authors of that study speculating that increased exposure to dogs due to the acquisition of Pandemic Puppies as 2020 progressed may have been one reason for this increase [70]. The desire for a dog that is safe with children has been cited in several studies characterising the ‘ideal companion dog’ [71,72] and perceived suitability for living with children has driven the recent surge in brachycephalic dogs [73]. However, with no robust evidence that breed is a risk or protective factor for dog bites [74], urgent owner education is needed to raise awareness of the dog bite risk to children and promote safe interactions with dogs within and outside of the household.Although one in ten Pandemic Puppy owners had not considered owning a dog before the pandemic, the proportion of owners who reported that they carried out no pre-purchase research did not differ between 2019 and 2020, and was consistently low (3.0%). Significantly more Pandemic Puppy owners performed pre-purchase research than 2019 owners, albeit this difference can be explained by a greater number of 2019 owners feeling sufficiently experienced as dog owners not to undergo this process. Commonly used pre-purchase information sources both before and during the pandemic were breed/crossbreed-specific online resources, The Kennel Club website and talking to dog breeders. Pandemic Puppy owners were more likely to conduct pre-purchase research on animal charity websites (+16.2%) or by talking to friends and family who had owned a dog (+10.5%), highlighting the importance of reliable online resources but also the power of anecdote over evidence, with peer-peer and family communication more common, which although trusted may be of variable quality. With social distancing in place throughout the pandemic, owners were less likely to talk to strangers that owned dogs of the breed they eventually purchased, albeit this may also be an unreliable source of information. Although The Kennel Club reported that searches using its “Find a Puppy” online tool increased during the pandemic compared to the same period in 2019 [3,4], Pandemic Puppy owners were significantly less likely to find their puppy using this tool, which may reflect exhaustion of the section of the market that advertised via this website.Both before and during the pandemic, puppy owners valued dog breeders who would allow them to see their puppies’ mother (dam), they felt cared for their dogs, and whom they felt were trustworthy and offered good communication, with Pandemic Puppy owners more likely to seek out breeders who offered good communication. The increased desire for a communicative breeder in 2020 may reflect owners becoming increasingly cautious of unscrupulous breeders cashing in on demand, given media reports and public campaigns ran by DEFRA (‘Petfishing’ [75]), Dogs Trust (‘DogFishing’ [5]) and The Kennel Club (#BePuppywise [76]) to educate prospective owners on this issue. Despite seeking a breeder who was perceived to care for their dogs, Pandemic Puppy owners were less likely to seek out a breeder who was a member of the Kennel Club Assured Breeders Scheme (ABS) or performed health tests for the breed/crossbreed they wanted. It is possible that prospective owners were not aware of these schemes prior to their purchase, or that scarcity of ABS breeders and/or those advertising health testing during the pandemic influenced prospective owners’ priorities when looking for a breeder. Understanding how owners define a ‘caring breeder’ is an important future research area given that this was an important characteristic for owners in both years, yet many puppies were purchased from breeders who were not following recommended or even legal standards.The common desire by prospective owners for a breeder who allowed prospective owners to see their puppy’s mother (dam) may reflect the success of the campaign that led to Lucy’s Law being passed into English legislation in 2019, which made it illegal to sell a puppy without showing its mother in the place it was born [77]. Despite this desire, this outcome was not fully achieved in many purchases during both 2019 or 2020, with one in seven puppies seen without their mother (dam) in 2019 and 1 in 4 in 2020. Understanding why owners make compromises during the purchasing process requires further research; however, previous data has demonstrated that actions differing from plans when acquiring a dog are common, particularly regarding the source of the dog [78,79]. It is possible that the ‘sellers’ market’ created in 2020, where demand outstripped supply, may have pushed some owners to make suboptimal decisions in their desperation to purchase a puppy. In May 2020, the media reported a UK ‘puppy shortage’ [80]. Scarcity messaging of this type during the pandemic led to panic buying of many household products—most infamously toilet rolls—with such messaging potentially heightening the urgency consumers feel to buy impulsively [10]. It is possible these messages had a similar effect on puppy purchasing; owners of Pandemic Puppies were willing to pay significantly more for their puppy than 2019 owners, with one in four owners paying over £2000 for their puppy, which was previously extremely rare with fewer than one in fifty owners paying this price in 2019. This was complemented by Pandemic Puppy owners being more likely to place a deposit on their puppy before they had seen them, leaving owners open to ‘PetFishing’ [75] and potentially limiting their freedom to change their minds if unhappy with the puppy or its environment when they met in person. Taken together with population statistics, it is likely that the unprecedented scale of owners strongly driven to purchase a puppy in 2020 enhanced the market for unscrupulous breeding of puppies, and, indeed, illegal imports to meet this demand, as reported by Dogs Trust [81].Public health measures to control the pandemic introduced restrictions that altered the pre-purchase and purchase behaviour of puppy purchasers. Pandemic Puppy owners were less likely to visit their breeder in person to view their puppy, reducing from four in five during 2019 to three in five owners during 2020 physically seeing their puppy prior to the day of purchase. Instead, during the pandemic, prospective owners were significantly more likely to view their puppy via live video calls or pre-recorded videos or photos. For those who did visit in person prior to bringing their puppy home, the number of visits was significantly lower. Limits to contact and viewing of the breeder’s property continued during the purchase, with Pandemic Puppy owners less likely to collect their puppy from inside the breeders’ property, replaced by doorstep/garden collections or deliveries to the new owner. Although these behaviours were unavoidable if adhering to ‘stay at home’ orders during some periods of 2020, these changes to the pre-purchase process increase the likelihood of puppy purchasers purchasing from breeders or dealers who were trying to conceal the poor environment and conditions their puppies were raised in, or indeed hide that the puppies they were selling weren’t raised at that location at all [75]. During the purchase visit, Pandemic Puppy owners were less likely to see their puppy with its littermates compared to 2019 owners. This may further indicate an increased number of unscrupulous puppy dealers in 2020 compared to 2019, who are reported to often separate puppies from their litter [75]. Increased educational efforts to re-emphasise the importance of seeing a puppy in person prior to purchasing them, collecting them from within the breeder’s property where they are able to see and interact with the puppies’ mother and littermates is needed to promote canine welfare.Health testing of breeding dogs, both phenotypic screening (e.g., hips, elbows, knees, eyes, respiratory testing) and DNA (genetic) tests are a critical tool in improving the health and welfare of future generations of dogs, including both pedigree and intentional ‘designer’ crossbreeds. Fewer Pandemic Puppy owners asked to see screening test results during the purchase process, a concerning trend previously documented in the purchase of brachycephalic breeds [15]. The lack of requests to see health records may reduce the priority placed on health by breeders, by reducing demand for healthy, tested dogs. While market forces of supply and demand are not being effectively applied to canine health, breed health is unlikely to improve substantially unless a sufficient proportion of breeders are intrinsically motivated to carry out health testing themselves.5. ConclusionsThe COVID-19 pandemic impacted all stages of puppy purchasing in the UK, from the reasons why prospective owners wanted to purchase a puppy, through to the actual puppy purchase. Given the nature of some of these changes, including an increase in inexperienced owners purchasing a puppy for the first-time, fewer puppies being seen in the environment where they were raised, and a reduced emphasis on canine health, combined with the reportedly large scale of purchasing during this period of 2020, this unique Pandemic Puppy population may face significant welfare problems in the future. In the short-term, whether owners are able to continue owning these now adolescent to young adult dogs as COVID-19 restrictions change in the UK is a major concern, given the initial motivation to purchase a dog during the pandemic being primarily based on the increased time this unique period offered. In the medium-long term, threats to the health and behaviour of this population include the potentially unsuitable environments many of these puppies might have been bred and raised in negatively impacting their behaviour, and the low regard for genetic health during many purchases potentially increasing the risk of breed-related disorders. It is clear that greater support from animal welfare organisations, veterinary professionals and animal behavioural professionals will be required to avoid, where possible, and address, where present, negative welfare outcomes likely to arise in this population. This study has further highlighted many of the deficiencies in UK puppy-purchasing culture both before and during the pandemic, and the need for ongoing education of prospective owners, as well as legislation to improve breeding practices, where consumer demand is not appropriately directed to effect positive welfare changes.

animals : an open access journal from mdpi