Datasets:

disi-unibo-nlp

/

SciLay

like 0

Follow

DISI UniBo NLP 11

10.3390/ani11113294

PMC8614317

Mulberry leaf is widely used in ruminants feeding, such as sheep, beef cattle, and dairy calves. Due to the high content of crude fiber in mature mulberry leaves and branches and the presence of anti-nutritional factors such as tannin, excessive addition will affect the production performance and health of livestock and poultry, and limit its large-scale application in animal production to a certain extent. The disadvantages of woody plants can be improved by microbial fermentation, which can reduce the content of anti-nutritional factors, and increase the content of peptides and amino acids, probiotics, and bioactive components. In this study, Lactobacillus, Saccharomycetes, and Bacillus subtilis were used to make mixed strains to ferment mulberry leaf powder, and different proportions were added to the diet of yellow feathered chicken broilers. The results showed that the addition of fermented mulberry leaf in the diet could improve the digestion and absorption of nutrients, and then improve its growth performance, and increase the contents of inosine monophosphate (IMP), total amino acids, essential amino acids, and delicious amino acids in breast and thigh muscle, and improved polyunsaturated fatty acids and essential fatty acids in breast muscle; this also has a positive effect on improving meat quality.

This study was conducted to investigate the effects of feeding fermented mulberry leaf powder (FMLP) on growth performance, slaughter performance, and meat quality of broilers. A total of 360 1-day-old chickens were randomly divided into 5 groups. The control group was fed basal diet (CON), 3% FMLP, 6% FMLP, 9% FMLP, and 3% unfermented mulberry leaf powder. The (MLP) group was fed basal diet supplemented with 3%, 6%, 9% fermented mulberry leaf powder, and 3% MLP, respectively. The experiment lasted for 56 days, with 1–28 days as the starter phase and 29–56 days as the grower phase. The results on the growth performance showed that diets supplemented with 3% FMLP significantly increased the ratio of villus height to crypt depth in the duodenum, jejunum, and ileum of broilers, enhanced the activity of intestinal amylase and digestibility of dry matter and crude protein, improved the average daily gain (ADG), and decreased the feed to gain ratio (F/G) (p < 0.05). Compared with the control group diet, the 3% FMLP group diet significantly increased the breast muscle yield (p < 0.05), reduced the abdominal fat ratio (0.1 < p < 0.05), and improved the slaughter performance of broilers. The 3% MLP group diet increased the shear force of breast muscle (p < 0.05) and thigh muscle of broilers compared to the control group, and adding FMLP could reverse the above results. Additionally, relative to the control group, FMLP supplementation improved the contents of inosine monophosphate (IMP), total amino acids (TAA), essential amino acids (EAA), and delicious amino acids (DAA) in breast and thigh muscle, and improved polyunsaturated fatty acids (PUFA) and essential fatty acids (EFA) in breast muscle; the 6% and 9% FMLP groups showed preferably such effects (p < 0.05). In conclusion, dietary supplementation of FMLP can improve the digestion and absorption of nutrients, and then improve the growth performance of broilers; it also has a positive effect on improving slaughter performance and meat quality.

1. IntroductionWith the development of population and the improvement of people’s living standards, the demand for livestock and poultry production and the conventional feed resources is increasing. The shortage of feedstuff has become increasingly prominent and the price of conventional feedstuff with large consumption has gradually risen. Therefore, finding cheap and reasonable feed resources to replace conventional feedstuff has become a research hotspot and strategic future after the COVID-19 outbreak [1].Mulberry is a deciduous perennial woody plant, belonging to Morus of Moraceae. Its leaves are considered as a high-quality forage plant resource because of its rich crude protein content (22~29.8%), balanced amino acid composition, rich in vitamins, trace elements, phytosterols, flavonoids, alkaloids, polysaccharides and other bioactive substances [2,3], and so on. However, due to the high content of crude fiber in mulberry leaves and branches and the presence of anti-nutritional factors such as tannin, the excessive addition of mulberry leaves and branches would affect the production performance and health of livestock and poultry, which, to a certain extent, limits its large-scale use in animal production [4]. The related disadvantages of woody plants could be improved by a microbial fermentation treatment, which reduced the content of anti-nutritional factors, increased the content of polypeptides and amino acids, and contains a variety of beneficial products such as probiotics bioactive ingredients [5]. Studies have found that adding fermented mulberry leaves into feeds could enhance immunity [6], regulated lipid metabolism [7], and improved the quality of animal products [8]. In conclusion, fermented mulberry leaf, as a new protein feed resource, has a broad application prospect in animal husbandry production.Mulberry leaves as unconventional feed resources are mainly used in ruminants, such as sheep [9], beef cattle [10], and dairy calves [11]. Fermented feed can be used to improve the intestinal health of broilers [12,13] and growth performance [14,15] has been reported. However, there are few studies on the application of mulberry leaves in poultry production after fermentation.In recent years, probiotic fermentation technology has become a powerful tool to reduce anti-nutritional factors in feed, and improve nutritional quality and the bioavailability of nutrients [16,17]. Therefore, in this study, the mixed strains of Lactobacillus, Saccharomycetes, and Bacillus subtilis were used to ferment mulberry leaf powder to investigate the effects of fermented mulberry leaf powder on production performance, slaughter performance, and meat quality of broilers, so as to provide a theoretical basis for the application of fermented mulberry leaf in livestock production, especially in areas where mulberry leaves are widely planted.2. Materials and Methods2.1. Preparation of FMLP SampleMulberry leaf powder (MLP) and Fermented mulberry leaf powder (FMLP), which were made from the leaves of hybrid feed mulberry, also known as Yajin protein mulberry, were provided by Hunan Institute of Sericulture Science. Fermentation strains (Lactobacillus, Saccharomycetes and Bacillus subtilis = 1:2:1, viable count ≥ 3 × 109 cfu/g), provided by Shandong Kangdien Biotechnology Co., Ltd. (Linyi, China). FMLP was prepared by solid-state fermentation for one week. After laboratory testing, the routine nutrients of MLP and FMLP were obtained and are shown in Table 1.2.2. Experimental Birds and FeedingAll of the experimental procedures were approved by the Animal Care and Use Committee of Hunan Agricultural University. In total, 360 one-day-old male yellow-feathered broilers provided by Hunan Xiangjia Animal Husbandry Co., Ltd. (Hunan, China) were randomly divided into five groups, consisting of 6 replicates of 12 birds each, which was then denoted as CON group (basal diet), 3% MLP group (basal diet supplemented with 3% mulberry leaf powder), 3%, 6%, and 9% FMLP group (basal diet supplemented with 3%, 6%, and 9% fermented mulberry leaf powder). The addition dosage of FMLP was adjusted accordingly, on the basis of the study of Has et al. [18]. All birds were raised in wire cages with 3-level battery following standard temperature regimens, which gradually decreased from 32 to 25 °C. The lighting scheme was all day lighting, throughout the test. Meanwhile, birds were offered basal diet and diet supplemented with mulberry leaf powder and different doses of fermented mulberry leaf powder and provided ad libitum access to water and diet in crumbled (1–28 d) and pelleted form (29–56 d). The experiment lasted for 56 days. The basal diets of the starter (1–28 d) and grower phase (29–56 d) formulated according to the feeding standard of chicken (NY/T 33-2004) are shown in Table 2.2.3. Sample CollectionAt 56 d of age, after 8 h of starvation, 6 birds (1 bird per replicate) were randomly selected from each treatment group. The weight of broilers after plucking and bloodletting was taken as dressed weight (DW) and after removal of head, foot, and viscera was taken as eviscerated weight (EW). Dressing percentages was calculated by DW/BW. Eviscerated yield was calculated as the percentages of BW. Breast muscle, thigh muscle, and abdominal fat pad including leaf fat surrounding the cloaca and gizzard were separated and weighed. Breast and thigh muscle yields were calculated as the percentages of EW. Abdominal fat percentage was calculated by abdominal fat weight/(abdominal fat weight + EW). Subsequently, within 10 min postmortem, all the right entire pectoralis majors and thigh muscle were collected for the determination of meat quality. Parts of the pectoralis major and thigh muscle samples were cut from the same location, quickly frozen in liquid nitrogen, and then kept at −80 °C for further analysis.2.4. Growth PerformanceFeed intake was recorded weekly, and total feed consumption in each replicate were recorded at 1, 28, and 56 d to determine average daily feed intake (ADFI), average daily gain (ADG), and feed to gain ratio (F/G).2.5. Apparent Nutrient DigestibilityDuring the experiment, 0.5% titanium dioxide (TiO2) was added to the diet as an exogenous indicator. The basic diet and four experimental diets were fed to the different treatment groups respectively. The first 3 days were used to adapt the birds while in the last 3 days, about 300 g of representative fecal samples were selected from each replicate every day, pooled, weighed, oven-dried (55 °C), milled, and stored pending chemical analyses. Apparent digestibility values for crude fibre, crude protein were calculated according to the following formula: AD (%) = 100 − [(G1 × F2)/(G2 × F1)] × 100 AD: apparent digestibility of dietary nutrients, G1: titanium content in diet, F1: nutrient content in the diet, G2: titanium content in feces, F2: nutrient content in feces.2.6. Intestinal Digestive Enzyme ActivityAfter slaughtering, the intestines of the experimental chickens were taken out, and the middle part of the jejunum about 10 cm was separated with a scalpel. The contents of the jejunum were put into a centrifuge tube, frozen in liquid nitrogen and stored at −80 °C. Amylase (Amy), lipase (LIP), and protease (PT) in jejunum contents were determined by commercial kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China), according to the manufacturer’s recommendations.2.7. Intestinal HistomorphologyBriefly, the intestinal samples were dehydrated with increasing concentrations of ethanol, cleared with xylene (Surgipath Medical Industries, Richmond, IL, USA), and embedded with paraffin wax (Thermo fisher scientific, Kalamazoo, MC, USA), and cut into 4-μm thick histological sections for hematoxylin and eosin staining. The tissue sections were measured under a microscope using a 40 × combined magnification, and an image processing and analysis system (Version 1, Leica Imaging Systems Ltd., Cambridge, UK). Villus height (VH); villus width (VW); crypt depth (CD); and VH/CD ratio (VH:CD) of the small intestine were determined by Program Image-pro Plus 6.0.2.8. Meat QualityThe meat color was measured at 60 min postmortem from a mean of three random readings made with a portable chromameter (CR-300, Minolta, Japan), which was calibrated with a white tile according to the manufacturer’s manual. At 45 min and 24 h after slaughtering, the pH of breast and thigh muscles were measured with a pH meter (pH-STAR, SFK technology, Denmark), previously calibrated with pH 4.6 and 7.0 buffers. The drip loss of breast and thigh meat was determined as described by Zhang et al. [19]. In brief, take a 3 × 2 × 1 cm piece from position of each sample of breast and thigh meat to determine drip loss. This sample was weighed and the mass was recorded as W1, and then suspended from a hook and placed in an inflatable zip–lock bag with the direction of the muscle fiber parallel to the gravity direction and hung for 24 h at 4 °C. After 24 h, the sample was removed and cleaned of moisture using filter paper, then weighed to obtain W2. Drip loss was then calculated as a percentage, where drip loss (%) = (W1 − W2)/W1 × 100%. L* (lightness), a* (redness), and b* (yellowness) of five random locations surface of the chicken breast and thigh meat were measured using a colorimeter (Konica Minolta Sensing Inc., Osaka, Japan) 1 h postmortem [20]. Cooked breast and thigh meat were cooled to room temperature and then rectangular-shaped samples (1 × 1 × 2 cm) at the same location were removed to measure tenderness using a TA-XT2 texture analyzer (Stable Micro Systems, Godalming, UK) with a Warner-Bratzler blade (code HDP/BS, Stable Micro Systems). Shear force was measured perpendicular to the axis of muscle fibers in 6 replicates for each treatment.2.9. Muscle Chemical AnalysisAbout 50 g breast and thigh muscles samples were sliced up, weighed, placed in a weighing bottle, and reweighed. The weighing bottle was placed into a freeze dryer at 50 °C for 48 h, and then reweighed. The weight difference between the initial sample and the dried sample was used to calculate the moisture percentage. Then, the dried samples were powdered with Muller CS-700 (Wuyi Haina Electric Appliance Co., Ltd., Zhejiang, China) and used for the analysis of crude protein (CP), amino acid and fatty acid composition. The crude protein, ether extract (EE), and crude fiber (CF) content were analyzed according to the method of the Association of Official Analytical Chemists.2.10. Inosine Monophosphate Content MeasurementAbout 5 g fresh muscle samples were weighed into a 15 mL centrifuge tube and homogenized in ice bath at 10,000 rpm for 30 s with t-25 ultra turrax homogenizer (IKA, Staufen, Germany). Then, we weighed 2.5 g of homogenate into a 50 mL centrifuge tube and added 25 mL of 5% perchloric acid. After shaking, it was centrifuged at 3500 rpm for 10 min in 4 °C refrigerated centrifuge, and then filtered into a 100 mL beaker. Then, 15 mL of 5% perchloric acid was added into the centrifuge tube. This was shaken well for 5 min, then centrifuged again and the two filtrates were mixed. After adjusting the pH to 6.5 with 5 mol/L and 0.5 mol/L NaOH, the filtrate was transferred into a 100 mL volumetric flask and diluted to the calibration tail with ultrapure water. The samples were filtered into the automatic vial and then used for HPLC.2.11. Amino Acid Composition of MuscleAbout 150 mg dried breast and thigh muscle were weighed into a glass bottle and 15 mL of 6 mol HCl were added. After nitrogen filling, the mixture was hydrolyzed for 22–24 h at 110 °C. Next, the hydrolysate was transferred to a 50 mL volumetric flask and diluted to calibration tail with ultrapure water. The solution was filtered using a 0.45 μm membrane filter into an autosampler vial, and then analyzed by L-8900 amino acid analyzer (HITACHI, Japan).2.12. Fatty Acid Composition of MuscleLipid extraction from breast muscle samples was performed by the Folch et al. method [17]. The extracted lipid was hydrolyzed in 2 mL KOH–methanol (C = 0.5 mol/L). After shaking for 1 min, the mixture was reacted in 95 °C water for 10 min to obtain a mixture of free fatty acids. The free fatty acid mixture was esterified in 2 mL BF3–methanol solution (W = 10%). After shaking for 10 s, the mixture was reacted in 80 °C water for 20 min. Subsequently, adding 1 mL n-hexane and 5 mL saturated NaCl solution, mixed for 1 min, then centrifuged for 15 min at 3000 rpm. Next, a volume of 800 μL fatty acid methyl esters was separated and analyzed with a GC-2010 plus gas chromatograph (Shimadzu, Japan). The injector and detector temperatures were maintained at 250 °C and 260 °C, respectively. Nitrogen was used as carrier gas, and the flow rate was 2.5 mL/min. The column temperature profile was as follows: maintained at 100 °C for 5 min, increased to 180 °C at 8 °C/min, increased to 210 °C at 4 °C/min, and maintained at 210 °C for 5 min. Next, the temperature was raised to 230 °C at 10 °C/min and then kept unchanged for 10 min. Fatty acids could be identified by comparing the retention time of the peaks with known standards (Sigma, St. Louis, MO, USA).2.13. Statistical AnalysesData are expressed as the mean ± standard deviation. Statistical analysis of the index was carried out according to the replicate of each group. The differences among the groups were analyzed by One-Way Analysis of Variance (ANOVA) followed by Tukey’s test using the SPSS 22.0 software (SPSS, Chicago, IL, USA). Significance was set at p < 0.05.3. Results3.1. Growth PerformanceThe effects of dietary supplementation of FMLP on growth performance are presented in Table 3. In the starter phase, ADG increased by 11.44%, 10.46% (p < 0.01); F/G decreased by 15.88%, 10.59% (p < 0.05), respectively, in chicken receiving 3% mulberry leaf powder and fermented mulberry leaf powder meal compared to those given basal diet group. Moreover, the ADFI in 3% fermented mulberry leaf powder group was significantly increased by 7.52% (p < 0.05) compared to the control group. In the grower phase, compared with the 3% mulberry leaf powder group, the ADG of the 3% fermented mulberry leaf powder group was significantly increased by 14.24% (p < 0.05); the ADFI and F/G of broilers among all groups had no significant differences (p > 0.05); adding low dose of fermented mulberry leaf powder had a trend to increase the ADFI of broilers (0.05 < p < 0.10). In the entire experimental phase, compared to the control group, the ADG of broilers in the 3% fermented mulberry leaf powder group was dramatically increased by 18.39% (p < 0.05), and the F/G of broilers in the 3% fermented mulberry leaf powder group was sharply decreased by 10.88% (p < 0.05).3.2. Apparent Nutrient DigestibilityThe dry matter (DM), CP digestibility of broilers in FMLP group were improved, and the 3% FMLP group were markedly increased by 6.98%, 10.36%, respectively (p < 0.05, Table 4) compared to the control group. Ether extract (EE) digestibility improved by 5.98%, 3.84%, and 9.89% (p < 0.05) in chicken receiving 3%, 6%, 9% FMLP meal compared to those given basal diet group. The digestibility of EE and ASH of broilers in all experimental groups were increased compared to the control group, but there was no significant difference between them (p > 0.05).3.3. Intestinal Digestive Enzyme ActivityThe digestive enzyme activity of amylase, lipase, protease in jejunum of broilers are shown in Table 5. Compared with the control group, the activities of amylase in jejunum of broilers in the 3%, 6%, 9% FMLP groups and 3% MLP group were significantly increased by 15.33%, 13.63%, 11.93%, and 19.46%, respectively (p < 0.01, Table 5). The lipase activity in jejunum of broilers in the 9% FMLP group was significantly increased by 24.44% (p < 0.05), and in other experimental groups it was increased, but there was no significant difference (p > 0.05) compared to the control group.3.4. Intestinal HistomorphologyThe normal function and structure of the intestinal tract were indicated by the villus height, crypt depth, and villus length/crypt depth (V/C), as shown in Table 6. The duodenal villus height of broilers in each experimental group was increased (p > 0.05), and the duodenal crypt depth was lower than that in the control group (p > 0.05). Compared with the control group, the duodenal V/C ratio of broilers in each dose of FMLP group was markedly increased by 25.69%, 19.71%, and 33.72% (p < 0.05). Compared to the control group, the villus height of jejunum in the 9% FMLP group was sharply decreased by 31.07% (p < 0.05), and the V/C ratio of jejunum in the 3% FMLP group was significantly increased (p < 0.05). The ileal crypt depth of broilers in each group decreased (p > 0.05), and the ileal V/C value of broilers in the 3% and 9% FMLP groups and in the 3% MLP group increased by 23.06%, 16.71%, and 18.27%, respectively (p < 0.05) compared to the control group.3.5. Slaughter PerformanceAs shown in Table 7, the breast muscle yield of broilers in the 3% FMLP group was markedly increased by 8.38% (p < 0.05), compared with the control group. In addition, the abdominal fat percentage of broilers in the 3% MLP group was sharply decreased by 29.68% (p < 0.05), and the abdominal fat percentage of broilers in other treatment groups had a decreasing trend (0.1 < p < 0.05) compared to the control group. There were no significant differences in dressing percentage, eviscerated carcass yield, thigh muscle yield among the treatment groups (p > 0.05).3.6. Meat QualityThe meat quality of the breast and thigh muscle fed with diets containing different doses of FMLP and MLP are summarized in Table 8. The L* value of thigh muscle was decreased (p > 0.05) and that of breast muscle was increased by adding FMLP to broiler diet, and in the 9% FMLP group, it was significantly increased by 16.45% (p < 0.05). Adding FMLP to broiler diet could reduce the a* value of thigh muscle (p > 0.05), but increased the a* value of breast muscle. Compared with the control group, the b* value of breast muscle in broiler diets supplemented with FMLP had no significant effect (p > 0.05), but it could reduce the b* value of thigh muscle, and in the 6% and 9% FMLP groups, it decreased by 27.80% and 25.65%, respectively (p < 0.05). Compared to the control group, the 3% MLP group increased in muscle shear force of breast muscle (p < 0.05) and thigh muscle (p > 0.05). However, the shear force of breast muscle and thigh muscle in the FMLP group decreased, especially in the 6% and 9% FMLP groups, in which it markedly (p < 0.05) decreased, compared with the 3% MLP group. Adding FMLP and MLP to the broiler diet could reduce the drip loss of breast and thigh muscles (p > 0.05). In addition, there were no significant effect on the ph45min value and ph24h value of broiler muscles (p > 0.05).3.7. Muscle Chemical Composition and Inosine Monophosphate ContentThe chemical composition including the moisture, EE, CP, and IMP content in breast and thigh muscle is presented in Table 9. The IMP of breast muscle increased (p < 0.05) by 31.11, 35.55, and 32.92%, the IMP of thigh muscle increased by 16.77, 19.88 (p < 0.05), and 24.84% (p < 0.05) in chickens fed 3, 6, and 9% FMLP diet compared to those given basal diet. There were no significant differences in the contents of moisture, EE, and CP in the muscle of broilers among the groups (p > 0.05).3.8. Amino Acid ProfileDietary supplementation of FMLP increased the contents of total TAA, EAA, and DAA in breast muscle and thigh muscle (Table 10). Concentrations of TAA in the 6% and 9% FMLP groups were significantly increased by 2.84% and 3.23%, and in the thigh muscle were remarkably increased by 4.16% and 4.39% compared with those in the control group, respectively (p < 0.05, Table 10). Compared with the control group, seven EAA (Try was not detected) in breast and thigh muscles of the diet treatment group were increased (Table 10 and Table S1). Lys and Val in breast muscles of the 3%, 6%, and 9% FMLP groups were significantly increased by 4.46%, 5.94%, 7.92% and 13.27%, and 14.16% and 15.04%, compared to the control group, respectively (p < 0.05, Table 10 and Table S1). Lys, Met, Phe, and Val in the 6% and 9% FMLP groups were significantly increased by 4.5% and 7.66%, 16.33% and 30.61%, 5.98% and 10.26%, an 11.38% and 13.82%, respectively (p < 0.05, Table 10 and Table S1). Compared to the control group, supplementing FMLP in the diet could significantly increase the content of DAA in the muscle of experimental chickens (p < 0.05, Table 10). Except Gly, Tyr, and Phe, the DAA concentrations in breast muscle of each diet treatment group were remarkably higher than those of the control group (p < 0.05, Table 10 and Table S1). Concentrations of DAA (except Val and Gly) in thigh muscles of the 6% and 9% FMLP groups were remarkably higher than those of the control group, in which Asp, Glu, Ala, and Phe were markedly increased by 6.78% and 9.60%, 3.00% and 11.97%, 16.52% and 25%, and 5.98% and 10.26%, respectively (p < 0.05, Table 10 and Table S1).3.9. Fatty Acid ProfileThe fatty acids detected in the breast and thigh muscles of broilers were basically the same, mainly composed of C16:0, C18:0, C18:1n-9t, and C18: 2n-6*, as shown in Table 11 and Table S2. Compared to the control group, the C18:1n-9t in the breast muscle of MLP group and each dose FMLP group increased 5.22%, 5.93%, 7.36%, and 12.54% (p < 0.05, Table S2). Concentrations of C22: 6n-3*, C18: 2n-6*, and C20: 4n-6* in the breast muscle of the control group were remarkably increased in each group (p < 0.01, Table S2), among which the C22: 6n-3* increased by 13.15%, 15.79%, 18.42%, and 23.68%, and C18: 2n-6* increased by 15.32%, 21.07%, 13.79%, and 24.77%, and C20: 4n-6* increased by 28.67%, 21.67%, 20.97%, and 23.08%, respectively. In addition, the concentrations of PUFA and EFA in the dietary treatment group was significantly higher than that in the control group (p < 0.05, Table 11), and the content of MUFA in the breast muscle had a tendency to increase (0.05 < p < 0.1, Table 11). Compared with the control group, the C16:1n-7 content in broiler thigh muscle was increased by adding FMLP in the diet, and the C16:1n-7 content in the 6% and 9% FMLP groups was significantly increased by 22.71% and 18.98% (p < 0.05, Table S2). Supplementation of 3%, 6%, and 9% FMLP significantly improved the content of C18:3n-6* in thigh muscle of broilers compared to the control group (p < 0.05, Table S2).4. DiscussionMulberry leaf, as a new type of feed resource, has a great potential for development and utilization in animal production due to its characteristics of large yield and balanced nutrition. However, its large-scale use is limited due to the fact that its mature leaves and stems contain anti-nutritional factors such as tannin. Fermentation, especially probiotic fermentation, has attracted more and more attention, because of its potential to reduce dietary anti-nutritional factors, improve feed nutritional quality, and promote animal growth performance [13,14]. Our results showed that, relative to the control group, dietary supplementation of low-dose FMLP markedly increased the ADG and sharply decreased the F/G in the starter phase, remarkably increased the ADFI, and significantly decreased the F/G in the whole experiment period, but the effect was weakened with the increase of the supplemental dose. Nutrient digestion and absorption may play an important role in improving growth performance. Previous studies have also found that the addition of 10% and 20% fermented and unfermented mulberry leaves in the diet of broilers significantly reduced the final body weight and dietary dry matter and crude protein digestibility of broilers with the increase of mulberry leaves supplemental dose [18]. This study showed that dietary supplemented with 3% FMLP could improve the digestibility of dry matter and crude protein nutrients of broilers, and the digestibility decreased with the increase of supplemental dose compared to the control group, which was consistent with the findings of Has et al. [18]. The digestibility decreased with the increase of supplemental dose, which may be attributed to the accelerated digestion rate caused by the increase of dietary fiber content, thus reducing the digestion time of nutrients and the digestion and absorption of nutrients by the gastrointestinal tract [21], and affecting the retention of nutrients (dry matter, organic matter, nitrogen) [22], because the fiber content of MLP and FMLP is higher, with 12.30% and 8.67%, respectively. In addition, in this study, relative to 3% MLP group, the F/G of broilers in the 3% FMLP group markedly decreased from 1 to 56 days. The reason for this result was related to the reduction of antinutritional factors in the diet and the degradation of macromolecular organic matter into small molecular by fermentation which is easy to be absorbed and utilized and the presence of probiotics in the diet, so as to improve the nutrient absorption and animal growth performance [23,24]. Other factors affecting nutrient absorption, such as digestive enzyme activity and intestinal morphology, were also examined. The results showed that dietary supplementation of FMLP can remarkably increase the activity of the intestinal amylase. Relevant studies have demonstrated that adding fermented feed and probiotics to broiler diet can improve the activity of intestinal digestive enzymes, which may be connected to the metabolism of probiotics in the intestine to produce part of digestive enzymes and improve the activity of related digestive enzymes [24,25]. Normal intestinal function and structure are the biological basis for growth and nutrient digestion and absorption of animals [26]. Villus height, crypt depth, and ratio of villus height to crypt depth (V/C) are important indexes to evaluate intestinal digestion and absorption in animals. The higher the villi height, the better the intestinal digestion and absorption function [27]. In this study, dietary supplementation MLP markedly increased ileum V/C value, and supplementation of FMLP significantly increased duodenum, jejunum, and ileum V/C value of broilers, with the 3% FMLP group having the most significant effect. These results were consistent with the study by Feng et al. [28], which revealed improved intestinal tissue morphological structure and increased intestinal digestive enzyme activities by adding fermented feed to broilers’ diets. According to the present results, it is suggested that dietary supplementation of FMLP promotes the growth performance of broilers by improving intestinal tissue structure, digestive enzyme activity, and nutrient digestibility.Slaughter performance is an important index to measure the carcass quality of meat livestock and poultry. It can not only directly reflect the percentage of the mass of different tissue parts in the total mass, but also reflect the difference of the deposition amount of nutrients in different tissue parts. High abdominal fat in broiler chickens will directly affect the processing of meat products, reduce slaughter rate and consumers’ purchase desire, and affect economic benefits [8]. In this study, it was found that the supplementation of MLP and FMLP in the diet of broilers can reduce the abdominal fat percentage of broilers, which may be relevant to the regulation of active substances in mulberry leaves on fat metabolism of broilers. Previous studies have demonstrated that 1-deoxynojirimycin (DNJ), the main alkaloid in mulberry leaves, had the effect of lowering blood glucose and blood lipid [29]. In our study, diets supplementing with FMLP had no effect on dressing percentage, eviscerated carcass yield, and thigh muscle yield of broilers, which was consistent with the finding of Semjon et al. [30]. In addition, diets supplementing with 3% FMLP may significantly increase the breast muscle yield of broilers, which may be related to the improvement of the digestibility of crude protein.Meat color is an important appearance index of meat quality, which directly affects consumers’ purchasing desire [31]. Indicators reflecting meat color are L*, a*, and b* [32]. Within a certain range, higher the a* value, the better the quality and freshness of the meat; the higher the L* value of meat color, the higher the gloss of the meat and the paler the color. The a* value is directly proportional to meat quality, while the b* and L* values are inversely proportional to meat quality [33]. Our results showed that the addition of 9% FMLP significantly reduced the b* of thigh muscle and increased the L* value of breast muscle, which indicated that the different types of muscle fibers might be the reason for the different effects of FMLP on different muscle tissues of broilers [34]. Probiotics and active substances may play an important role in the effect of dietary FMLP on meat color of broilers. This conjecture is consistent with the findings of Yu et al. [35], who reported that dietary supplementation of high concentrations had the most significant effect on meat color [35,36,37], and the findings of Shen et al. [38], who revealed that the effect of adding bamboo leaf extract in the diets with different concentrations on meat color was linearly increased [38]. Tenderness (shear force) may be the most important edible quality parameter that determines consumer acceptance [39]. Shear force is an intuitive indicator of muscle tenderness [40]. In the present study, the addition of FMLP could reduce muscle shear force, especially at medium and high doses. Probiotics in fermented mulberry leaf powder may play an important role in it. Previous studies similarly have found that dietary supplementation probiotic can reduce muscle shear force [35,41]. Relative to the control group, the muscle shear force of MLP group was significantly increased, which may be related to the increase of ADG, resulting in the increase of muscle fiber diameter, which in turn led to the increase of shear force, because the smaller the muscle fiber diameter is, the more tender the muscle is [42].The composition and content of amino acids, fatty acids, and nucleotides flavor substance in animal and poultry meat are important factors that affect the nutritional value and flavor. In the present study, compared to the control group, diets supplementing with FMLP remarkably improved the content of IMP, increased the contents of EAA, TAA, and DAA, and the effect of medium and high dosage FMLP were most significant, but dietary supplementation MLP had little effect on the IMP content in breast and thigh muscle of broilers, suggesting that probiotics may play an important role in the FMLP group. Previous studies similarly have demonstrated that diet supplementing with 5% alfalfa (similar to MLP, it can be used as unconventional protein feed) meal did not affect IMP of breast and thigh muscles; dietary supplementation probiotics increased the content of IMP, DAA, EAA, and DAA in breast muscle [35]. In addition, in the present study, fatty acids in breast muscle and thigh muscle of broilers were mainly C16:0, C18:0, C18:1n-9t, and C18:2n-6. Their total contents accounted for a significant proportion in the total fatty acid composition, and they were the main component of muscle fatty acids in broilers, and unsaturated fatty acids were the main component, which was consistent with the findings of Semjon et al. [30]. Farmer et al. [43] illuminated PUFA are more likely to form volatile flavor substances during lipid oxidation, which makes meat more delicious. C18:2n-6, C18:3n-6, and C20:4n-6 are EFA indispensable to the human body in PUFA, which play a very important role in maintaining normal development and health, and can effectively prevent atherosclerosis and myocardial infarction [44]. C18:3n-6 and C22:6n-3 are important raw materials for the formation of biofilms, which promote the development of the nervous system and brain [45]. Our result showed that dietary supplementation of MLP and FMLP markedly increased the content of PUFA (C18:1n-9t, C18:2n-6, C20:4n-6, C22:6n-3) and EAA in breast muscle of broilers, indicating that active substances and probiotics may play an important role in MLP and FMLP groups. Previous studies similarly have reported that diets supplementing fermented ginkgo biloba leaves increased the contents of flavonoids and polysaccharides in diets, and then increased the content of total PUFA in breast muscles [46]; diets supplementing with probiotics improved the content of PUFA and SFA in breast muscles [47]. Additionally, dietary supplementation with MLP and FMLP had little effect on thigh muscle, only improved the content of C18:3n-6 in muscle, which indicated that the effects of MLP and FMLP on fatty acid content of muscle in different parts of broilers were different, which might be caused by potential differences in nutrient absorption and distribution among different tissues [48]. According to the present results, it is suggested that dietary supplementation with MLP and FMLP can improve the nutritional value and flavor of meat by affecting the composition and content of PUFA in muscle of broilers.5. ConclusionsDiet supplementing with FMLP at a dosage of 3% could improve the digestion and absorption of nutrients, such as the digestibility of dry matter, CP and EE, amylase and V/C ratio, and then improve growth performance. Adding 6% and 9% FMLP could improve the meat quality of breast and thigh muscles without affecting the growth performance, such as increasing the concentration of IMP, TAA, EAA, DAA, PUFA, and EFA, reducing the shear force, and the 9% FMLP group showed preferably such effects.

animals : an open access journal from mdpi

10.3390/ani13050785

PMC10000113

This study was conducted to investigate the effects of dietary supplementation with Bacillus licheniformis and a combination of probiotics and enzymes on the growth and blood parameters of grazing yak calves. The body weight, body size, serum biochemical parameters, and growth hormone levels of grazing yaks were assessed. We found that supplementation with probiotics alone or with a combination of probiotics and enzymes significantly increased the average daily gain, compared to the controls, and the combination of probiotics and enzymes showed a better performance. Supplementation with the complex of probiotics and enzymes significantly increased the concentration of serum growth hormone, insulin-like growth factor-1, and epidermal growth factor, which may be the main reason for the higher daily weight gain. The findings of this study may help improve the growth efficiency of yak calves on the Qinghai–Tibetan Plateau.

Early weaning is an effective strategy to improve cow feed utilization and shorten postpartum intervals in cows; however, this may lead to poor performance of the weaned calves. This study was conducted to test the effects of supplementing milk replacer with Bacillus licheniformis and a complex of probiotics and enzyme preparations on body weight (BW), size, and serum biochemical parameters and hormones in early-weaned grazing yak calves. Thirty two-month-old male grazing yaks (38.89 ± 1.45 kg body weight) were fed milk replacer at 3% of their BW and were randomly assigned to three treatments (n = 10, each): T1 (supplementation with 0.15 g/kg Bacillus licheniformis), T2 (supplementation with a 2.4 g/kg combination of probiotics and enzymes), and a control (without supplementation). Compared to the controls, the average daily gain (ADG) from 0 to 60 d was significantly higher in calves administered the T1 and T2 treatments, and that from 30 to 60 d was significantly higher in calves administered the T2 treatment. The ADG from 0 to 60 d was significantly higher in the T2- than in the T1-treated yaks. The concentration of serum growth hormone, insulin growth factor-1, and epidermal growth factor was significantly higher in the T2-treated calves than in the controls. The concentration of serum cortisol was significantly lower in the T1 treatment than in the controls. We concluded that supplementation with probiotics alone or a combination of probiotics and enzymes can improve the ADG of early-weaned grazing yak calves. Supplementation with the combination of probiotics and enzymes had a stronger positive effect on growth and serum hormone levels, compared to the single-probiotic treatment with Bacillus licheniformis, providing a basis for the application of a combination of probiotics and enzymes.

1. IntroductionYaks (Bos grunniens) occur on the Qinghai–Tibet Plateau at high altitudes and with long cold seasons and limited pasture resources. This species is a unique product of long-term natural selection, providing local herders with the most basic living materials and livelihood resources, such as meat, milk, shelter (hides and furs), and fuel (dung), and is an indispensable part of the ecology and economy of the Qinghai–Tibetan Plateau [1]. However, the low reproductive rate of yaks seriously restricts their production and utilization. The cold season on the Tibetan Plateau lasts for eight months (October to the following May), during which time the quantity and quality of pasture decrease below the nutritional requirements of lactating yaks [2]. The deficiency of feed intake results in a negative body energy balance and metabolic stress [3]. On the other hand, under traditional grazing management, plateau-grazing yak calves are weaned naturally or artificially under various conditions at an age of 18–24 months [4], rather than the weaning age of domestic beef cattle (<6 months). The slow recovery itself and the late weaning of yak calves, which result in a poor postnatal physical condition, severely affect the onset of the next estrous cycle in the cow. Most yaks exhibit a long postpartum anestrous period and calve twice every 3 years or once every 2 years [5]. Therefore, the early weaning of yak calves may help mitigate these adverse effects.Early weaning has become more popular in recent years for various reasons, including the better use of limited feed resources and alleviating grazing pressure on pastures by reducing the nutritional needs of cows [6]. Weaning calves before the start of the breeding season improves the reproductive performance of cows [7,8] because the cows can regain their weight faster, thus accelerating the onset of postpartum estrus. The use of milk replacer in early weaning is common in livestock production [9,10]. The milk replacer has demonstrated positive benefits in animal experiments, such as improved immunity and relieved weaning stress response [11]. Increasing evidence suggests that enhanced milk replacer feeding is beneficial for improving gut microbial development and growth performance in early-weaned lambs [12,13].Over the past few decades, probiotics have been widely used in livestock and poultry production for their ability to enhance animal disease resistance, improve feed utilization, and improve growth performance [14]. In ruminants, yeasts and bacteria, including Lactobacillus, Bifidobacterium, Bacillus, Propionibacterium, and Enterococcus, alone or in combination, are used as additives in diets [15,16]. Probiotics can decrease diarrhea, improve production and feed utilization efficiency, and strengthen the immunity system in young ruminants [17,18,19]. Moreover, supplementation with probiotics improves the rumen and intestinal epithelial cell growth, which enhances the gastrointestinal tract development and health status of calves [17,20,21]. Oral administration of Bacillus licheniformis can increase ruminal digestibility and total volatile fatty acid concentrations in Holstein cows [22] and growth performance in Holstein calves [23]. In vitro inoculation with Bacillus licheniformis also improves ruminal fermentation efficiency of forage of various qualities [24]. However, no information is currently available on the effect of Bacillus licheniformis on the growth performance of yak calves.Compound enzyme preparations are produced from one or more preparations containing a single enzyme as the main entity, which is mixed or fermented with other single enzyme preparations to form one or more microbial products [25], including saccharylases, amylases, cellulases, proteases, phytases, hemicellulases, and pectinases. Depending on the differences in digestive characteristics and diet composition, specific enzyme preparations can be used for livestock [26]. Specific enzyme complex preparations can degrade multiple feed substrates (antinutrients or nutrients), and different types of enzymes can work synergistically to maximize the nutritional value of feed [27]. In buffalo calves, cellulase and xylanase are more effective with regard to average daily weight gain (ADG) and feed efficiency [28]. Further, the addition of exogenous fibrolytic enzymes to wheat straw has no effect on starter feed intake and increases nutrient digestibility and recumbency, but decreases the ADG of weaned Holstein dairy calves [29].The effects of probiotics or compound enzyme preparations on the production performance and biochemical blood indexes of calves are not consistent [29,30,31,32,33]. The respective discrepancies may be due to differences in the amounts of added probiotics and exogenous enzymes, the strains of probiotics, diets, and animal management strategies. Therefore, this study was conducted to compare the effects of Bacillus licheniformis and a combination of probiotics and enzymes on the growth performance and serum parameters in yak calves, so as to provide a theoretical basis for the application of probiotics in grazing yak calves.2. Materials and Methods2.1. Animals and TreatmentThis study was performed in accordance with the Chinese Animal Welfare Guidelines, and the experimental protocols were approved by the Animal Care and Ethics Committee of the Institute of Animal Husbandry and Veterinary Medicine, Tibet Academyof Agriculture and Animal Husbandry Science (No. #TAAAHS-2016–27).The feeding trial was conducted at Damxung Co., (Lhasa, China; 30.5° N, 91.1° E) from July to October. The average altitude was 4200 m, the average annual temperature was 1.3 °C, and the average annual precipitation was 456.8 mm. Thirty two-month-old male yaks (38.89 ± 1.45 kg body weight (BW)) were fed milk replacer solution at 3% of their BW every day and were randomly assigned to three dietary supplementation treatments (n = 10, each), according to BW and age, as follows: T1, supplemented with 0.15 g/kg Bacillus licheniformis (2 × 1010 CFU/g); T2, supplemented with a 2.4 g/kg combination of probiotics and enzymes (containing 0.4 g/kg Bacillus licheniformis, 2 × 1010 CFU/g; 1.0 g/kg yeast, 1 × 1010 CFU/g; 1.0 g/kg mixture of xylanase, cellulase, and glucanase in a 1:1:1 ratio, xylanase, 20,000 U/g, cellulase, 1500 U/g, glucanase, 6000 U/g); and a control treatment. The milk replacer, probiotics, and enzyme preparations were provided by the Chinese Academy of Agricultural Sciences (Beijing, China). All yak calves were allowed to graze on an alpine meadow during daytime for the 60-day trial, and they were individually fed milk replacer before and after grazing (0800 and 2000 h, respectively). The forage of the alpine meadow was mainly composed of Kobresia tibetica, and the nutrient composition (dry matter basis) was analyzed in our previous study [34], i.e., 10.4% crude protein, 2.1% ether extract, 67.8% neutral detergent fiber, 34.2% acid detergent fiber, and 4.6% ash. The powdered milk replacer was weighed and mixed with warm water (approximately 40 °C) at a ratio of 1:7 (w/v) to obtain milk replacer solution, according to our previous study [35]. Based on preliminary assessments, the feeding amount of milk replacer was calculated so that all yak calves were able to feed without surplus [35]. The nutrient composition of the milk replacer is shown in Table 1.2.2. Sample Collection and AnalysisThe BW of each yak calf was recorded before morning feeding on d 0, 30, and 60 using a platform scale, and the ADG was calculated accordingly. The body size indexes of all yak calves were determined using a linen tape at the beginning (d 0) and end (d 60) of the experiment, as previously described [36].Blood samples (approximately 10 mL) were collected from the jugular vein of the yak calves using a vacuum tube before morning feeding on d 0 and 60. The blood samples were centrifuged at 1100× g for 10 min to obtain serum, which was then aliquoted in 1.5 mL centrifuge tubes and stored at −20 °C.The serum biochemical parameters, including blood urea nitrogen (BUN), globulin (GLB), blood glucose (GLU), and non-esterified fatty acids (NEFAs), were analyzed using an automatic biochemical analyzer 7020 (Hitachi, Tokyo, Japan). Metabolic hormones in the serum, including insulin-like growth factor-1 (IGF-1), epidermal growth factor (EGF), cortisol, insulin (INS), and growth hormone (GH), were determined using commercial ELISA kits (Jiahong Technology Co., Ltd., Beijing, China) according to the manufacturer’s instructions. Briefly, 50 μL of each five-fold diluted serum sample was added to each well of a 96-well ELISA plate. After 30 min of incubation at 37 °C, the plate was washed five times using PBS (Servicebio, Wuhan, China) to remove unbound proteins. Then, 50 μL of HRP-conjugated antibodies was added to allow them to bind with their corresponding antigens. The 3,3′,5,5′-tetramethylbenzidine working solution was added to each well, followed by stop solution. Absorbance was measured using a multi-plate reader (Varioskan LUX, Thermo Fisher Scientific, Waltham, MA, USA) at a wavelength of 450 nm.2.3. Statistical AnalysisAll experimental data of this study were statistically analyzed using a one-way analysis of variance followed by Duncan’s post hoc test with SPSS 26.0 software (SPSS Inc., Chicago, IL, USA). Each yak calf was considered an experimental unit. Data are expressed as means ± standard error. p < 0.05 was considered statistically significant.3. Results3.1. Body WeightThe three treatments did not differ significantly in terms of BW on d 0, 30, and 60 (Table 2). The ADG was higher (p < 0.05) in the calves under T2 treatment than those under the control treatment, from d 0 to 30, d 30 to 60, and d 0 to 60, and higher (p < 0.05) than that of those calves under the T1 treatment from d 0 to 60, indicating that the supplementation of Bacillus licheniformis and the combination of probiotics and enzymes could improve the growth performance of early-weaned grazing yak calves. The ADG of calves under T1 treatment was higher (p < 0.05) than that of those under the control treatment from d 0 to 60.3.2. Body SizeThe body size parameters did not differ significantly among the three treatments on d 0 and 60 (Table 3), indicating that the supplementation of Bacillus licheniformis and the combination of probiotics and enzymes did not affect the body size of yak calves within 60 d.3.3. Serum Biochemical ParametersThe concentrations of serum GLB, BUN, GLU, and NEFAs did not differ significantly among the three treatments on d 0 and 60 (Table 4).3.4. Serum HormoneAs shown in Table 5, the concentrations of serum IGF-1 on d 60 were higher in T2-treated calves than in the T1- and control-treated calves (p < 0.05, each). The concentrations of serum EGF and GH on d 60 were higher in the T2-treated calves than in the controls (p < 0.05). The concentration of serum COR on d 60 was higher in the control calves than those under the T1 treatment (p < 0.05).4. DiscussionEarly weaning may have various benefits for cows; however, early weaned calves generally perform poorly compared to naturally weaned calves [37]. Early weaned calves without breastfeeding grew at a lower rate and subsequently took longer to reach their target weight than breastfed calves [38]. To improve the growth performance of early-weaned calves, several improvements were made to the composition of milk replacer or additional feeds were added [39,40,41]. Moreover, the addition of probiotics to the diets of calves significantly improved the ADG [29,30,33]. Dietary supplementation with compound enzyme preparations also improved growth performance in weaned piglets [42,43] and growing-finishing pigs [44]. However, previous studies also reported that supplementation with probiotics, yeast cultures or enzymes had no effect on the growth performance of calves [31,32,45]. In the current study, the addition of Bacillus licheniformis alone or a complex of probiotics and compound enzyme preparations to the milk replacer significantly improved the performance of grazing yaks and calves compared with milk replacer alone. Further, the addition of probiotics is beneficial for the regulation of the intestinal microbiota community structure, improving intestinal health and fecal consistency, and reducing diarrhea prevalence [19,31,46,47,48]. The supplementation of fibrolytic enzyme to the diet of crossbred calves improved their nutrient digestibility with a positive effect on daily gain [49]. Calves typically exhibit high metabolism and fast growth; however, their growth performance is susceptible to environmental stress and nutrient absorption and digestive problems, especially in the period after weaning [50]. Under natural grazing conditions on the Qinghai–Tibet Plateau, due to the long-term lack of pasture and harsh environmental conditions, the normal growth of yak calves is severely restricted [48]. In the present study, none of the study animals died, which may be attributed to the supplementation with milk replacer. Therefore, the addition of probiotics and compound enzyme preparations was beneficial for the growth of grazing yak calves.In most cases, calf weight is positively correlated with body length, and body length can be used to predict calf live weight [51,52]. Supplementation with Bacillus subtilis results in an increased body length and BW in Barki lambs at the third and fourth week, as observed in a four-week continuous feeding trial [53]. In the present study, neither body size nor BW differed among the treatments, which may be due to insufficient trial duration and individual differences in animals. Therefore, more time may be required to elucidate whether the probiotic and compound enzyme preparations affected the calves’ body size.To a certain extent, blood biochemical parameters reflect the metabolism and the acid–base balance of the animal body, and they vary within a certain range [54,55]. The results of the current study revealed that supplementation with Bacillus licheniformis and the complex of probiotics and enzyme preparations had no effect on the blood biochemical parameters of grazing yak calves, which is consistent with previously reported results in crossbred and Holstein calves [56,57]. The blood biochemical values of calves vary with the growing stage and are strongly influenced by weaning [58,59], and these possible factors may be stronger than the influence of diet on blood biochemical indicators.Insulin-like growth factors (IGFs) are small polypeptide hormones mainly synthesized and secreted from the liver, and they are structural homologs of insulin, with similar activities. These consist in binding to specific carrier proteins in the blood to form a composite factor that stimulates systemic body growth and has growth-promoting effects on almost every cell in the body [60,61]. As mediators of GH action, the synthesis of IGFs is also affected by the blood level of GH [62]. EGF is a member of the growth factor family, a single polypeptide of 53 amino acid residues that is involved in regulating cell proliferation [63]. We found that the addition of probiotics and a combination of probiotics and enzymes significantly increased the concentration of serum IGF-1, EGF, and GH, whereas supplementation with Bacillus licheniformis alone did not achieve this effect. These results are consistent with the ADG results. GH and IGF-1 are important controllers in regulating amino acid metabolism in calves, where GH promotes the entry of amino acids in muscle tissue into cells and increases protein synthesis, and IGF-1 increases protein deposition by promoting protein synthesis [63,64]. Cortisol is commonly used as a marker of stress responses (such as weanling stress) in animals, and it occurs at high serum levels for a period of time after calves are weaned [65]. In line with our results, oral supplementation with probiotics markedly decreases the concentrations of serum cortisol in neonatal and weaned calves [66,67]. Interestingly, we found that the concentrations of serum cortisol were lower in the T1 than in the T2 group, which was, however, not statistically significant. This suggested that the addition of Bacillus licheniformis alone may better alleviate weaning stress in grazing yak calves. However, the respective mechanisms remain to be resolved in more detail.A limitation of this study is that the T2 group did not strictly control a single variable compared to the T1 group, and the factors (yeast or xylanase, cellulase and glucanase) that contributed to the difference were unclear. This was due to the initial intention of this study to improve the milk replacer by adding probiotics or compound enzyme preparations, and ultimately promote the growth performance of yak calves on the Qinghai–Tibet Plateau. Further, we were unable to collect data on diarrhea and determine nutrient digestibility in grazing calves, which would have further improved our understanding of the weight gain of yaks under the various treatments.5. ConclusionsOur results suggest that supplementation with Bacillus licheniformis alone or with a complex of probiotics (Bacillus licheniformis and yeast) and compound enzyme preparations (xylanase, cellulase, and glucanase) can improve the ADG of grazing yak calves, and the complex had a better effect on the ADG. The addition of the complexes of probiotics and complex enzyme preparations also increased the concentrations of serum GH, IGF-1, and EGF, which may have led to a higher ADG. Thus, the addition of a combination of probiotics and enzymes to milk replacer may serve as an effective strategy to improve the production of yak calves.

animals : an open access journal from mdpi

10.3390/ani11051397

PMC8156027

Automated training devices are commonly used for investigating learning, memory, and other cognitive functions in warm-blood vertebrates, whereas manual training procedures are the standard in fish and other lower vertebrates, thus limiting comparison among species. Here, we directly compared the two different approaches to training in guppies (Poecilia reticulata) by administering numerical discrimination tasks of increasing difficulty. The automated device group showed a much lower performance compared to the traditionally-trained group. We modified some features of the automated device in order to improve its efficiency. Increasing the decision time or inter-trial interval was ineffective, while reducing the cognitive load and allowing subjects to reside in the test tank improved numerical performance. Yet, in no case did subjects match the performance of traditionally-trained subjects, suggesting that small teleosts may be limited in their capacity to cope with operant conditioning devices.

The growing use of teleosts in comparative cognition and in neurobiological research has prompted many researchers to develop automated conditioning devices for fish. These techniques can make research less expensive and fully comparable with research on warm-blooded species, in which automated devices have been used for more than a century. Tested with a recently developed automated device, guppies (Poecilia reticulata) easily performed 80 reinforced trials per session, exceeding 80% accuracy in color or shape discrimination tasks after only 3–4 training session, though they exhibit unexpectedly poor performance in numerical discrimination tasks. As several pieces of evidence indicate, guppies possess excellent numerical abilities. In the first part of this study, we benchmarked the automated training device with a standard manual training procedure by administering the same set of tasks, which consisted of numerical discriminations of increasing difficulty. All manually-trained guppies quickly learned the easiest discriminations and a substantial percentage learned the more difficult ones, such as 4 vs. 5 items. No fish trained with the automated conditioning device reached the learning criterion for even the easiest discriminations. In the second part of the study, we introduced a series of modifications to the conditioning chamber and to the procedure in an attempt to improve its efficiency. Increasing the decision time, inter-trial interval, or visibility of the stimuli did not produce an appreciable improvement. Reducing the cognitive load of the task by training subjects first to use the device with shape and color discriminations, significantly improved their numerical performance. Allowing the subjects to reside in the test chamber, which likely reduced the amount of attentional resources subtracted to task execution, also led to an improvement, although in no case did subjects match the performance of fish trained with the standard procedure. Our results highlight limitations in the capacity of small laboratory teleosts to cope with operant conditioning automation that was not observed in laboratory mammals and birds and that currently prevent an easy and straightforward comparison with other vertebrates.

1. IntroductionThe study of learning, memory and perception in animals has, since its inception, benefited from the use of automated training equipment. The use of these methods offers a two-fold advantage. First, they reduce the time needed for training and the related human labor required. Some experiments, especially those in discrimination learning, may require thousands of training trials [1,2]. With manual execution, many months and hundreds of hours of work are required to train each subject [3,4,5]. The second advantage is that automated equipment allows for the control of every detail of the experiment, standardizing procedures across different studies and laboratories, while minimizing the need for human intervention and reducing the possible influence from researchers’ expectations [6,7].While automated training devices are frequently employed to study mammals and birds, they are rarely used outside these two vertebrate classes [8,9,10]. In the last two decades, there has been an increasing interest in studying cognition in other animal taxa, such as fish, reptiles, and arthropods. Teleost fish in particular have been thoroughly studied, and in some cognitive domains, they show abilities surprisingly close to those seen in mammals and birds [11,12,13] as well as high degree of genetic homology to humans [14]. Therefore, some small tropical fish, in particular zebrafish (Danio rerio), medaka (Oryzias latipes), and guppy (Poecilia reticulata), have become important models in neurobiological research [15,16,17], and several laboratories have tried to develop Skinner-box apparatuses for these species [18,19,20,21].Numerical cognition is one of the cognitive domains most widely investigated. On average, fish show capabilities comparable to those of mammals and birds [22]. Guppies (Poecilia reticulata), in particular, can be quickly trained to discriminate up to four from five objects, even if continuous perceptual cues (i.e., cumulative surface area or density) are made irrelevant [23]. As for primates, task difficulty increases with decreasing distance between the quantities to be discriminated [23,24].The numerical capacities of guppies exceed those observed in various warm-blooded vertebrates such as dogs, horses, and domestic chicken, though they appear lower than those exhibited by humans, apes and some other mammalian and by avian species [22]. However, these discrepancies are difficult to interpret, due to marked differences in methods. Monkeys, pigeons, and rats were studied with automated conditioning devices, whereas guppies were studied using traditional, manually operated, training methods or with spontaneous preference procedures. Studies on the former species commonly involve hundreds of training trials per day and a few thousand trials per experiment [1,2]. In contrast, guppies usually receive 12–15 daily trials, and the number of trials per experiment rarely exceeds 100 [23,25,26].In recent years there have been some attempts to use automated operant conditioning devices for studying cognitive abilities in fish [18,19,20,21,27]. Our laboratory has developed a device based on the Skinner tank built by Manabe and colleagues [18]. The device is controlled by a microcomputer, which displays stimuli on a monitor, tracks the movements of the fish, and delivers a very small amount of food when the subject makes a correct response. Guppies tested with this device can easily perform 80 trials in each daily session. If required to discriminate between two stimuli with different colors, guppies showed excellent performance, reaching 90–95% accuracy in two–three sessions. The performance in shape discrimination tasks was only slightly lower, with subjects reaching a maximum 80–85% accuracy in approximately one dozen sessions [28]. This is comparable or even superior to the performance obtained by guppies and other fish in color or shape discrimination in other studies using more traditional non-automated approaches to training [29]. Unexpectedly, when asked to discriminate between two sets of items with different numbers of elements (3 vs. 5 and 3 vs. 4), the guppies showed very poor performance [30]. This clearly conflicts with capacities demonstrated by guppy studies that used different training procedures [23,24,31]. It is worth noting that other studies using automated devices have reported unexpectedly low performance in other species and with different tasks, which calls into question the reliability of this approach of operant conditioning in fish [27,32,33]. For example, in one study zebrafish demonstrated good accuracy in color discrimination, but no evidence of learning shape discrimination [32]. No reference data are available in this case, though zebrafish have shown significant capability for recognizing familiar from unfamiliar shapes [34,35].There are two critical aspects concerning this research area. The first is that it is presently difficult to have a precise estimation of the effectiveness of the automated approach compared with the traditional ones. None of the above-mentioned studies included control groups tested in the same task with a reference standard procedure, and many differences with prior studies (i.e., type of subjects, stimuli, operant conditioning protocol) could potentially explain discrepancy in results. For example, the numerical stimuli in Gatto and colleagues [30] differed in the format and in the numerical ratios from those used in previous studies [23,26]. Furthermore, Gatto and colleagues [30] presented a mixture of two numerical discriminations simultaneously, whereas the best performing test [23] used a particular protocol in which subjects were first trained on a very easy numerical discrimination and then, upon achieving the criterion, were given progressively more difficult tasks until they reach the discrimination threshold.The second critical aspect is that even if the difference in certain tasks would be confirmed, it will not be easy to identify the causes of the efficiency gap between automated devices and traditional methods to train fish. Automated operant conditioning represents a fundamentally different approach to training which largely differs from traditional training methods in its objectives, and which is subjected to distinct constraints. The two approaches differ in a larger number of features, many of which are not directly related to automation itself. Historically, Skinner boxes have been developed to reduce human intervention and to speed up training of laboratory animals, such as rats or pigeons. The inner space of a Skinner box is extremely reduced so that stimuli, operandums, and the devices for delivering reinforcers are fitted in a small area and close to the subject. This feature permits a subject to run trials in rapid succession, usually many trials per minute, which, combined with the possibility to administer infinitesimal rewards, allows an animal to perform hundreds of trials in a brief session. Thanks to these characteristics, a single Skinner box is used to train many subjects in rotation.In the traditional, manually-operated, operant conditioning methods, the need to visually assess the subject’s choice, has made necessary testing subjects in large apparatuses. In this way, the two stimuli are kept far apart, and the subject makes its choice by moving in the direction of one of the two targets or by choosing an arm in a T-maze. The need for the experimenter to prepare the stimuli and the reward for the subsequent trial usually determine long inter-trial intervals. This, in turn, implies that the number of daily trials cannot usually exceed one or two dozen. The way stimuli are presented is also very different in the two approaches. The fully-computerized control of a Skinner box requires that stimuli are displayed in some digital format. On the other hand, since all operations of manual training are usually performed by a single experimenter, in most situations adding the extra task of operating a computer can be impractical. For this reason, solid objects or printed stimuli are usually preferred with traditional training approaches.The presence of a large number of differences could make the search for causal factors very long and complex. However, some factors are better candidates than others to account for the low efficiency of automated training approach in fish. To explain the low performance of guppies in numerical tasks, Gatto and coworkers [30,32] argued that automated devices may force fish to perform additional cognitive tasks (related to device functioning), increasing the overall cognitive load (i.e., the amount of cognitive resources that are devoted for dealing with one specific task). This factor is expected to be especially important when a subject is performing complex tasks [36,37]. Indeed, it was argued that numerical discriminations are more cognitively demanding than other types of discrimination and certainly they imply a greater number of steps [38]. A numerical discrimination requires that, in each trial, subjects estimate the first quantity, then the second, and then compare the two amounts. In contrast, color discrimination requires only learning to avoid one color or to choose the other, with no comparison needed after the first trials. A second hypothesis regards the time allowed by the two approaches to subjects for collecting information and issue the correct response. In traditional manual training experiments the subject slowly approaches the stimuli, with plenty of time to make a decision; in Skinner tanks, subjects performed trials in quick succession, with less time to process information, take a decision and, if necessary, to correct a wrong decision [30]. Finally, the automated device for fish was designed, as the original Skinner box, to allow testing several subjects, which need consequently to be moved daily from their home tank to the apparatus. While rats and pigeons adapt very quickly to being manipulated and introduced into the Skinner box periodically, fish may perceive a threat and, therefore, reduce the attention paid to the task. In addition, in fish frequent manipulation may produce chronic stress with a consequent negative impact on cognitive performance [39,40,41].The first experiment of this study aimed to evaluate the influence of the training approach on numerical tasks performance under controlled conditions. Training done with our automated device was benchmarked against a standardized manual training procedure [31,42,43] using the same strain, the same stimuli, the same numerical ratios and adopting in both cases the protocol used by Bisazza and colleagues [23].The fact that automated devices appear very effective in certain types of tasks but unpredictably inefficient in others, makes this approach of operant conditioning unreliable for fish cognition research. In the second part of the study, we tested five variants of the automated procedure, in the attempt to improve its effectiveness (experiments 2–5). We modified some features of this approach to training that we hypothesized to be responsible for its reduced effectiveness in numerical tasks. In particular, we elongated the decision time, increased inter-trial interval, removed internal partitions to increase visibility of stimuli, allowed subjects to reside in the conditioning chamber, and attempted to reduce the cognitive load by uncoupling learning how to use the conditioning chamber from learning the numerical discrimination.2. Materials and Methods2.1. Animal HousingWe used adult female guppies derived from the ornamental strain known as ‘snakeskin cobra green’ that is regularly bred in our laboratory at the Department of General Psychology (University of Padova). We tested subjects of the same sex to avoid the additional confounding effect of sex differences and to have a more homogeneous sample. The guppies were maintained in mixed-sex groups of 30 individuals in 150 L tanks. Each tank was enriched with natural vegetation and a gravel bottom. The water temperature was kept at 27 ± 1 °C and was aerated via a biomechanical filter. A 30 W fluorescent lamp provided illumination for 12 h per day. The guppies were fed live nauplii brine shrimp (Artemia salina) and commercial flakes (AquaTropical, Isola Vicentina, Italy) twice per day. All fish used in the experiments were naïve to the experimental protocol. After completion of the experiment, all fish were released in other maintenance tanks and kept for reproductive purposes only.2.2. Determination of Sample SizeTo determine the sample size for the experiments, we calculated the number of subjects necessary to achieve 90% power at a two-tail significance level of p = 0.05. For the manual conditioning experiments, we fitted parameters derived from a previous study that investigated numerical discrimination using a very similar protocol [31] and for automated conditioning Experiment the parameters from a similar recent experiment [30]. The estimated sample size was n = 6 for the former and n = 9 for the latter. We conservatively used slightly larger sample size (8 and 11 respectively). In Experiment 2 we dropped sample size at six subjects as a very large number of fish were discarded in the pre-training phase, due to failure to learn to detour the transparent barrier. In Experiment 3, the number of subjects was increased to 10 since we had two slightly different methods of presenting stimuli as a between factor.2.3. Experiment 1: Comparison of Manual and Automated TrainingIn this experiment, we directly compared the recently developed automated operant conditioning procedure with the manual training procedure routinely used in our laboratory [42,43]. In both cases, the task consisted of numerical discriminations of progressively increasing difficulty, following the protocol of Bisazza and colleagues [23].The subjects of this experiment were 19 adult females, 11 trained with manual conditioning and eight with the automated conditioning procedure. Stimuli were made with Adobe Illustrator CC 2019 and consisted of sets of black dots of differing numerosity on a white background: 3 vs. 12, 2 vs. 3, 3 vs. 4, 4 vs. 5, and 5 vs. 6. Like other vertebrates, guppies can discriminate two quantities of objects using non-numerical attributes of the stimuli, such as cumulative surface area, density, and convex hull (the convex polygon that circumscribes all items) [44,45]. To prevent fish from using this alternative strategy, we controlled the stimuli for all the above-mentioned non-numerical variables. To control for cumulative surface area, we used dots of different diameters (range 0.75–0.95 cm): in one-third of the stimuli used, the ratio between the cumulative surface area of the smaller over the larger set was between 76 and 85%; one-third, between 86 and 95%; and in one-third between 96 and 105%. Furthermore, we varied the position of the dots in order to equate the convex hull in half of the trials and the density of the dots in the other half of the trials [23]. We used 18 pairs of stimuli for the 3 vs. 12 numerical discrimination and 24 for all others numerical discrimination. The stimuli used for 3 vs. 12 were controlled for density and convex hull, but not for the cumulative surface area.Stimuli were printed on laminated white cards (3 × 3 cm) for the manual conditioning procedure and were displayed on a monitor for the automated conditioning procedure.2.3.1. Apparatus and ProcedureManual ConditioningWe used an apparatus and procedure recently developed for other investigations on guppy cognition [42,43]. Subjects were tested individually in a 20 × 50 × 32 cm glass tank filled with 28 cm of water (Figure 1A).Eleven identical apparatuses were used at the same time and placed in a dark room. Each tank was in olfactory communication with an aquarium placed beneath the apparatuses that housed approximately 20 conspecifics; the water in all tanks was filtered by a silent filtering system. Two trapezoidal lateral compartments (10 × 6 × 32 cm) internally shaped each tank into an hourglass. These lateral compartments, made of transparent plastic, housed natural plants to provide an enriched environment for the subjects. Externally each wall of each tank was covered with green plastic to avoid any possible external influence during the experiment. One 15 W fluorescent lamp illuminated two adjacent apparatuses on a 12:12 h light:dark photoperiod schedule. All trials were recorded with video cameras placed above each tank. To present the stimuli to the subjects, each card was affixed to the end of a transparent panel (3.5 × 15 cm) with an L-shaped blocker that allowed us to fix the panel to the wall of the tank. During each trial of the training phase (see below for details), two transparent panels were simultaneously inserted on the same short wall of the tank.Following the protocol of previous studies [42,43], each subject underwent a habituation phase, a pre-training phase and a training phase. The habituation phase lasted two days, during which the guppies could familiarize themselves with the experimental apparatus. During this phase, fish were fed four times per day using a Pasteur pipette placed in alternating positions near the two short sides of the tank.The pre-training phase also lasted two days, during which guppies could familiarize themselves with the experimental procedure. During the first day of this phase, a white card was presented to the subjects eight times. Guppies performed four trials in the morning session and four trials in the afternoon session; the two sessions were divided by a 90-min interval, and consecutive trials were divided by 15-min intervals. When a subject approached the card, a Pasteur pipette was used to deliver a small quantity of food reward (i.e., reinforcement) consisting of a drop of live brine shrimp (A. salina). During the second day of this phase, the subjects performed a total of 12 trials (six in the morning session and six in the afternoon session) identical to the trials of the first day. To maintain a high level of motivation during the sessions, no other food was provided during the course of the experiment. The number of times the card was presented on each of the two short sides of the tank was evenly distributed in the trials and counterbalanced across each session on both days. We admitted to the training phase only those subjects that approached the card all 12 times in the second day of the pre-training phase. One subject that failed to achieve this criterion was not admitted to the experimental phase and was replaced with a new subject.In the training phase, each subject performed 12 trials per day for a maximum of 12 consecutive days for the 3 vs. 12 numerical discrimination and 10 days for all other discriminations. Specifically, the subjects performed six trials in the morning session and six in the afternoon session, with a 90-min interval between the sessions and a 15-min interval between each consecutive trial. The guppies were presented up to five different discriminations that corresponded to five different difficulty levels. All subjects were trained to select the larger numerosity. A choice was considered when the subject approached (swam at less than one body length) one of the two stimuli. To assess reliability of this measure, a subsample of the video-recorded trials of each subject was analyzed by a second observer who was blind to the experimental hypotheses. In all discriminations, when the subjects approached the correct stimulus first, they were given a food reward. If they approached the smaller numerosity first, no food reward was given and both stimuli were removed simultaneously.Each subject started with the 3 vs. 12 discrimination. Subjects were presented a more difficult discrimination (i.e., 2 vs. 3) if they met one of the two learning criteria. The primary learning criterion was defined as a rate of at least 75% correct choices (18/24) of all trials over two consecutive days (statistically significant using a binomial test). The secondary learning criterion was defined as a frequency of at least 60% correct choices (87/144 for the 3 vs. 12 discrimination and 72/120 for all others) over all trials (statistically significant using a binomial test). If they failed to reach one of the two criteria within 120 trials, the experiment ended. In the case of success, the same procedure with the same criteria was used to present the discrimination tasks for 3 vs. 4, 4 vs. 5 and finally, 5 vs. 6. The left/right position of the larger numerosity line and the short side of the tank on which the cards were presented over the trials were counterbalanced.Automated ConditioningFish were tested in a 12 × 16 × 10 cm conditioning chamber made of semi-transparent white plastic (0.3 cm in thickness). The internal compartment was uniformly illuminated by room light. The bottom of the chamber was made of transparent plastic and a camera was positioned 12 cm below the chamber to track the subjects during the experiment. The chamber (Figure 1B) was internally divided in a starting area (12 × 5.5 cm) connected through a corridor (4 × 2.5 cm) to a V-shaped decision area (12 × 3.5 cm), and two choice areas (6 × 4.5 cm). The distance between the exit of the starting area and the entrance to each choice area was 6. The two choice areas were separated by a suspended semi-transparent white plastic sheet. Each choice area presented a 6 × 5 cm window that allowed projection of stimuli via an LCD computer monitor (Samsung S19C450, Suwon, South Korea). An automatic feeder was placed above and between the two choices areas. The feeder consisted of a servomotor (Futaba S3305) connected to a transparent cylinder filled with A. salina decapsulated eggs (Shg Srl, Alessandria, Italy) held by two metallic rods. When appropriate (see details below), activation of the servomotor caused the vibration of metallic rods, and the release of 3–4 eggs as a positive food reward (i.e., reinforcement). The presentation of the stimuli, the tracking of the fish using the camera, and the activation of the feeder were simultaneously controlled through a Raspberry Pi (Raspberry Pi 3 Model B V1.2, 2015, Sony UK Technology Centre, Pencoed, UK) running custom-made Python software.One week before the experiment started, two subjects were randomly selected and moved into a 30-L aquarium provided with the same conditions as the maintenance tank. Each aquarium housed three immature guppies as social companions. The individual subjects were placed in the conditioning chamber only during the experimental sessions. Fish were identified based on their individual characteristics, such as coloration and tail shape. A pump connected to the housing tank and to the conditioning chamber served to ensure water exchange and to provide social odors to the subjects during the daily training sessions.Pre-training phase. During the pre-training phase, subjects were habituated to the conditioning chamber and learned how operate the automated system. We conducted two daily pre-training sessions between 10:00 h and 14:00 h, lasting 30-min each. Subjects were individually inserted in the conditioning chamber while the monitor projected a white background. When the subject spontaneously moved into the starting area, the monitor background changed from white to grey and the trial started. Once the subject entered one of the two choice areas, the automatic feeder delivered a small quantity of food reward and the monitor’s background changed from grey to white. A new trial did not start until the subject entered again in the starting area. There was no interval between trials. Each session ended when the subject obtained a maximum of 80 reinforcements or when 30 min had elapsed. The second daily pre-training session was conducted after a 2-h interval.During each pre-training session, the experimenter monitored the subject’s behavior using an LCD monitor connected to the camera. We admitted to the training phase only those subjects that consumed at least 30 food rewards in two pre-training sessions. Thirteen subjects failed to achieve this criterion within 12 pre-training sessions. These subjects were discarded and substituted with new subjects.Training phase. Daily training sessions lasted one hour and were administered between 10:00 h and 14:00 h for a maximum of 12 days. Subjects were individually inserted in the conditioning chamber while the monitor projected a white background. A pair of stimuli that differed in numerosity appeared once the subject entered the starting area. All subjects were trained to select the larger numerosity. The system randomly alternated the position of the correct stimulus between trials and was set to prevent three consecutive presentations of the correct stimulus in the same choice area. Once the subject chose the correct stimulus by entering the corresponding choice area, the system activated the feeder to release the food reward above the two choice areas. The stimuli then disappeared, and a new trial started when the fish spontaneously moved back into the starting area. If the subject entered the choice area corresponding to the incorrect stimulus, the monitor immediately displayed a black background, and no food reward was released. In this case, the same pair of stimuli was presented in the same left–right position (correction trials) and the correction trial was repeated until the subject chose the correct stimulus and received the reward. The interval between an incorrect choice and the subsequent correction trial was set to 10 s. These correction trials were not considered in the analysis, whether correct or incorrect, and only served to prevent the fish from developing left–right biases.The maximum number of reinforcements given per day was 80, and the training session ended automatically when the subject reached the maximum number or when 60 min had elapsed. The subject was then gently moved back in its maintenance aquarium. When a fish obtained fewer than 40 reinforcement, an additional training session was performed on the same day 2 h after the end of the first session. The additional training session was conducted using the same schedule as the first; however, the number of possible reinforcements changed based on the number of rewards the fish had already obtained, up to the 80 daily maximum reinforcements.Each day, we quantified the subjects’ accuracy as the proportion of correct choices by considering the first choice of each trial (correction trials were not considered). We used two learning criteria. The primary learning criterion was defined as 75% correct choices in two consecutive days. As in the manual procedure, a secondary learning criterion was defined as the frequency of at least 60% correct choices over a total of 12 days (statistically significant at the binomial test). Subjects that fulfilled one of the two learning criteria were presented with a series of more difficult numerical discriminations (each one lasted for a maximum of 10 days following the same training protocol).2.3.2. Statistical AnalysisWe performed the statistical analysis in RStudio version 1.2.5019 (RStudio Team, 2019; RStudio: Integrated Development for R. RStudio, Inc., Boston, MA, USA). We used a two-tailed statistical test; the significance threshold was set at p = 0.05, and descriptive analyses were reported as mean ± standard deviation.We analyzed subjects’ accuracy in four steps. The first step concerned the assessment of overall accuracy using evidence of learning the task for each numerical discrimination task. We analyzed the performance at the group level by comparing the percentage of correct choices to one expected at chance level (i.e., percentage of 50%) using a one-sample t-test. Since the subjects’ accuracy could be random in the first training session due to the lack of previous experience with automated conditioning, we analyzed the overall performance in the second half of the training session with the same approach. The second step concerned the assessment of individual accuracy using a binomial test on the number of correct choices and incorrect choices. The third step was focused on accuracy improvement over the training session. We analyzed subject performance as a repeated observation of binomial choices (i.e., correct and incorrect choices) in each training session using a generalized linear mixed-effects model with logit link function and binomial error distribution (GLMM, “glmer” function from the “lme4” R package). We fitted the model with the training session as a fixed effect, and subject (i.e., individual ID) as a random effect. The effect of the parameters was evaluated using the ‘Anova’ function of the ‘car’ R package. In Experiment 1, manual conditioning, the model was also fitted with numerical discrimination (i.e., 3 vs. 12, 2 vs. 3, 3 vs. 4, 4 vs. 5) as a fixed factor to evaluate difference in subjects’ accuracy among tasks.The fourth step concerned the comparison of subjects’ accuracy showed in the manual and automated conditioning procedures. We compared the overall performance by using a two-sample t-test. The learning performance was then evaluated using a GLMM, fitted with the training session and type of procedures as fixed factors, and the subject as a random factor.2.3.3. ResultsManual conditioning. Interrater reliability calculated for 936 trials was found to be very high (Cohen’s kappa coefficient κ = 1, p < 0.001). All 11 subjects reached the primary learning criterion of 18/24 correct choices in two consecutive days on the 3 vs. 12 discrimination (Figure 2). Table 1 shows the individual performance on each discrimination task with statistics. On average, the fish needed 55.64 ± 35.30 trials to meet the criterion (average number of days: 4.63 ± 2.94). Overall, subjects’ accuracy was greater than that expected at chance level (70.25 ± 6.99%; one-sample t-test: t10 = 9.604, p < 0.001).Ten out of 11 subjects reached the primary learning criterion in the 2 vs. 3 discrimination. On average, these fish needed 38.40 ± 21.01 trials to meet criterion (average number of days: 3.20 ± 1.75). The remaining subject (N10) stopped participating after 72 trials when its performance was statistically significant (Table 1). Overall, subjects’ accuracy was greater than that expected at chance level (73.96 ± 7.67%; t10 = 10.367, p < 0.001).Eight out of 10 subjects reached the primary learning criterion in the 3 vs. 4 discrimination. On average, these eight fish needed 60.00 ± 37.09 trials to meet the criterion (average number of days: 3.75 ± 1.83). Overall, subjects’ accuracy was greater than that expected at chance level (68.07 ± 9.66%; t9 = 5.918, p < 0.001).Two out of eight subjects reached the primary learning criterion and they needed on average 30.00 ± 8.49 trials (average number of days: 2.50 ± 0.71). One additional subject reached the secondary learning criterion. Overall, subjects’ accuracy was greater than that expected at chance level (60.76 ± 7.71%; one-sample t-test: t7 = 3.947, p = 0.006).None of these three subjects achieved the 5 vs. 6 discrimination according to the learning criteria. Overall, subjects’ accuracy in the 5 vs. 6 numerical discrimination was not greater than that expected at chance level (51.67 ± 0.83%; t3 = 3.464, p = 0.742).The overall analysis revealed a significant improvement in subjects’ accuracy over training session (χ21 = 5.371, p = 0.020), and a significant difference in accuracy among the four numerical discriminations (χ24 = 26.912, p < 0.001). The interaction was significant (χ24 = 10.512, p = 0.033), suggesting a different learning curve among tasks.Automated conditioning. No subjects achieved the 3 vs. 12 numerical discrimination according to the learning criteria. Overall, subjects’ accuracy was not greater than that expected at chance level performing on average 532.88 ± 180.91 trials (51.55 ± 4.38%; t7 = 0.998, p = 0.351). Even when the last six training sessions were considered, the subjects’ accuracy was no greater than that expected at chance level (258.00 ± 114.80 trials; 53.65 ± 5.01%; t7 = 2.061, p = 0.078). Table 2 shows the individual performance with statistics for all subjects.The GLMM revealed a significant improvement in the subjects’ accuracy over training sessions (χ21 = 8.448, p = 0.004).Manual vs. Automated conditioning. Overall, subjects’ accuracy was significantly different between Manual and Automated conditioning (two-sample t-test: t17 = 6.653, p < 0.001). The GLMM revealed a statistically significant improvement in the subject’s accuracy over training sessions (χ21 = 7.516, p = 0.006), a significant effect of the procedure (χ21 = 44.408, p < 0.001), and no session × procedure interaction (χ21 = 1.049, p = 0.306).2.4. Experiment 2: Elongation of Decision TimeIn this experiment, we increased the size of the conditioning chamber, increasing the distance between the corridor and the choice areas. Additionally, a transparent barrier was placed between the corridor and the choice areas so that the animal could not rush into them. These changes aimed to force the subject to take more time before the decision.2.4.1. Subjects, Apparatus and ProcedureFor this experiment, we used six adult females. The conditioning chamber was increased in size to 16 × 32 cm (Figure 1C). As previously described, it was internally divided in a starting area (16 × 16 cm) connected through a corridor (4 × 3 cm) to a V-shaped decision area (16 × 8.5 cm), and two choice areas (8 × 4.5 cm). As a consequence, the distance between the exit of the starting area and the entrance to each choice area was increased from 6 to 11.5 cm. To further increase the travel time necessary to reach the choice areas, a transparent barrier was placed between the V-shaped decision area and the choice areas so that the subject had to detour it in order to enter one of the two choice areas. The two choice areas were separated by a white plastic divider to prevent the fish from seeing the stimulus projected in the other choice area. Two automatic feeders were placed in correspondence with both areas. When the fish entered the choice area associated with the correct stimulus, the system activated the corresponding feeder to release a food reward. In this experiment, we modified the method to provide water exchange and social odors to the subjects. The conditioning chamber was placed inside a larger tank (20 × 50 × 32 cm) filled with 28 cm of water. The tank was provided with a gravel bottom, natural vegetation, a bio-mechanical filter, and it housed three immature guppies. As in Experiment 1, subjects were housed in a 30 L aquarium provided with the same condition as the maintenance tank and placed in the conditioning chamber only during the experimental sessions.During the pre-training phase, we habituated the subjects to detour a 7 × 10 cm transparent barrier. When the subject obtained at least 30 reinforcements in a pre-training session, we changed the transparent barrier for a wider one (9 × 10 cm) that was used during the training phase. Other details of the procedure were identical to those of Experiment 1 (automated conditioning). Stimuli were the same as those used in Experiment 1. Thirty-one fish did not pass the pre-training phase and were replaced with new subjects. The exceedingly large number of fish discarded in this experiment was due to a difficulty of most individuals to learn to efficiently detour the transparent barrier.Subjects’ performance was analyzed using the same statistical approach described in Experiment 1. We finally compared subjects’ performance between Experiment 1 (both manual and automated conditioning procedures) using a GLMM. The model was fitted with training session and type of procedure as fixed factors, and subject as a random factor.2.4.2. ResultsOne subject reached the primary learning criterion in the 3 vs. 12 numerical discrimination after 339 trials in 8 days. The remaining subjects did not achieve this discrimination after performing on average 502 ± 204.62 trials. Overall, subjects’ accuracy was not greater than that expected at chance level (51.47 ± 4.38%; t5 = 0.820, p = 0.449). Even when the last six training sessions were considered, subjects’ accuracy was not greater than that expected at chance level (52.87 ± 12.66%; t5 = 0.555, p = 0.603). Table 3 shows the individual performance with statistics for all subjects.The GLMM did not reveal an improvement in the subjects’ accuracy over the training sessions (χ21 = 0.608, p = 0.435). The performance of this discrimination did not significantly differ from Experiment 1 with an automated device (GLMM; procedure: χ21 = 0.673, p = 0.412; training session: χ21 = 3.092, p = 0.079; training session × procedure: χ21 = 5.909, p = 0.015), but was significantly lower than in the manual training experiment (procedure: χ21 = 20.239, p < 0.001; training session: χ21 = 0.638, p = 0.425; training session × procedure: χ21 = 0.005, p = 0.945).The subject tested in 2 vs. 3 discrimination, failed to achieve the task (overall: 51.94 ± 6.40%; t9 = 0.960, p = 0.362).2.5. Experiment 3: Removal of Internal PartitionsIn Experiment 3, we built a new conditioning chamber from which all internal dividers had been removed, so to increase the visibility of the stimuli. (Figure 1D). In this experiment we also compared two different spatial arrangement of the stimuli presented to the subjects.2.5.1. Subjects, Apparatus, and ProcedureFor this experiment, we used 10 adult females. Stimuli were the same as those of Experiment 1 (automated conditioning) but were presented in two different ways. To half of the subjects, we presented stimuli as in previous experiments, always arranged in the same spatial position except for left-right alternation (Condition A). To the remaining subjects (Condition B), we varied the horizontal and vertical position of the two stimuli and the distance between them, obtaining 18 different positions, the same used in two previous studies [23,46]. Stimuli were presented in these positions in rotation. This procedure is thought to minimize the development of left-right side bias and to favor discrimination learning [23,47].We used a modified version of the apparatus adopted in Experiment 2, consisting of a 16 × 32 × 10 cm conditioning chamber (Figure 1D). Internally, two white plastic walls divided a 16 × 10 cm starting area from the rest of the chamber. A feeder was positioned in correspondence with each choice area (8 × 4.5 cm) and a window allowed the presentation of the stimuli. The distance between the exit to the starting area and the entrance to each choice areas was increased to 17.5 cm. When the subject entered the choice area associated with the correct stimulus, the system activated the corresponding feeder to release a food reward. Other details of the procedure were identical to those of Experiment 1 (automated conditioning). Seventeen subjects did not pass the pre-training phase and were replaced with new subjects.Statistical analysis was the same as that for to previous experiments. Learning performance was evaluated using a GLMM fitted with training session and experimental condition (i.e., stimuli presentation) as fixed effects, and subject as a random effect. We compared subjects’ performance between Experiment 1 and Experiment 3 using a GLMM. In this analysis the subjects of the two conditions in Experiment 3 were considered as a single group. The model was fitted with training session and type of procedure as fixed factors, and subject as a random factor.2.5.2. ResultsNo subjects achieved the 3 vs. 12 numerical discrimination according to the criteria (average number of trials: 390.00 ± 74.48; Table 4). There was no difference between the two conditions (t8 = 0.755, p = 0.472), and the two conditions were pooled for further analysis. Accuracy was significantly greater than that expected at chance level (51.707 ± 2.18%; t-test: t9 = 2.478, p = 0.035). Table 4 shows the individual performance with statistics for all subjects.The GLMM did not reveal an improvement in subjects’ accuracy over training sessions (χ21 = 0.807, p = 0.369), nor effect of condition (χ21 = 0.456, p = 0.499), nor interaction (χ21 = 0.938, p = 0.333).The performance of this experiment did not significantly differ from that of Experiment 1 with automated device (GLMM; procedure: χ21 = 0.010, p = 0.921; training session: χ21 = 1.921, p = 0.166; training session × procedure: χ21 = 7.574, p = 0.006), but was significantly inferior to the performance in the manual training experiment (procedure: χ21 = 43.374, p < 0.001; training session: χ21 = 1.067, p = 0.302; training session × procedure: χ21 = 0.039, p = 0.844).2.6. Experiment 4: Subjects Resident in the Conditioning Chamber and Manipulation of the Inter-Trial IntervalIn all previous experiments, fish were housed in a maintenance tank and individually transferred to the conditioning chamber for the training session. The experimental manipulation may have stressed the subjects and, consequently, affected their learning abilities. In this experiment, guppies were tested in their home tank. The conditioning chamber was built inside a large aquarium and each subject was maintained in the conditioning chamber for the whole experiment.2.6.1. Subjects, Apparatus, and ProcedureFor this experiment, we used 16 adult females. We used the same set of stimuli adopted for Experiment 3 arranged in different vertical and horizontal positions. For this experiment, we built a 20 × 32 × 32 cm conditioning chamber inside a 20 × 50 × 32 cm glass tank, the same type of tank used for the manual training (Figure 1E). The chamber was internally divided into a starting area (20 × 10 cm), a corridor (5 × 2 cm), a V-shaped decision area (20 × 16.5 cm), and two choice areas (4.5 × 10 cm) so that, in this version the shape and the size of the various areas were very similar to the corresponding areas in the manual training apparatus. A feeder was positioned in correspondence with each choice area, and a window allowed the presentation of the stimuli. The distance between the exit of the starting area and the entrance to each choice area was increased to 18.5 cm. The remaining part of the glass tank was supplied with natural vegetation and a gravel bottom, a bio-mechanical filter, and it housed, for the whole experiment, three immature guppies as social companions. A transparent plastic partition furnished with a series of holes separated the conditioning chamber from the area containing social companions. At the end of each experimental session, a male was inserted as a social companion into the conditioning chamber and removed the following day, one hour before the beginning of the daily session.Experiment 4 was, in fact, composed of two distinct experiments which shared the same apparatus. In Experiment 4A the temporal schedule of the trials was the same as that described for Experiment 1 and each fish daily underwent a one-hour session with a maximum of 80 trials. In Experiment 4B we modified inter-trial interval and the overall temporal schedule of the experiment. Each daily training session lasted 8 h and the interval between the two correction trials was increased from 10 s to 10 min. The session terminated if the subject reached 80 reinforced trials. This modification aimed to prevent subjects from developing the strategy of producing a rapid series of random responses to get a 50% on rewarded trials. The change also had the effect of spacing out trials over time, making the temporal schedule more similar to that of manual training procedure. To habituate the fish to an 8-h training session, one day of pre-training was added, in which subjects underwent four pre-training sessions of increasing duration, 15, 30, 45, and then 60 m, separated by 90-min intervals. Other details of the procedure were identical to that of Experiment 1 (automated conditioning). Five subjects in Experiment 4A and two subjects in Experiment 4B did not pass the pre-training phase and were replaced with new subjects.2.6.2. ResultsIn Experiment 4A, six out of eight subjects achieved the 3 vs. 12 numerical discrimination according to the primary learning criterion (average number of trials: 291.00 ± 82.34; average number of days: 5.17 ± 1.47) and one additional subject reached the secondary learning criterion (Table 5). A subject did not achieve this discrimination after performing 597 trials. No subject achieved the 2 vs. 3 discrimination after performing on average 495.14 ± 135.10 trials.In Experiment 4B, five out of eight subjects achieved the 3 vs. 12 numerical discrimination according to the primary learning criterion (average number of trials 184.00 ± 33.96; average number of days: 5.20 ± 1.64; Table 6). Three subjects did not achieve this discrimination after performing on average 329.67 ± 5.69 trials. Only one out of five subjects reached the primary learning criterion in the 2 vs. 3 discrimination, after 160 trials in two days. This subject failed to achieve the 3 vs. 4 discrimination (47.59 ± 6.64%; t9 = 1.147, p = 0.281). The remaining four subjects did not achieve the 2 vs. 3 discrimination after performing, on average, 336.25 ± 110.07 trials.To examine the influence of inter-trial interval (10 s vs. 10 min) we compared accuracy between Experiment 4A and Experiment 4B. In the 3 vs. 12 discrimination, we found no statistically evidence of an effect of time elongation (t10 = 0.285, p = 0.780) and the subjects of the two experiments were pooled for further analysis. Subjects’ accuracy was significantly greater than that expected at chance level (59.88 ± 6.95%; t15 = 5.689, p < 0.001). The GLMM revealed a significant improvement in the subjects’ accuracy over training sessions (χ21 = 20.948, p < 0.001), no effect of condition (χ21 = 0.030, p = 0.863). The session × condition interaction was significant (χ21 = 6.844, p = 0.009), due to the fact that subject with the standard procedure (Experiment 4A) had a steeper learning curve than subjects which had a 10 min delay after each incorrect trial (Experiment 4B). The performance was significantly higher than that of Experiment 1 with automated device (GLMM; procedure: χ21 = 18.474, p < 0.001; training session χ21 = 29.835, p < 0.001; training session × procedure χ22 = 8.775, p = 0.003), but did not reach the performance of manual training (χ21 = 12.190, p < 0.001; session χ21 = 25.220, p < 0.001; session × procedure χ21 = 4.817, p = 0.028).In the 2 vs. 3 discrimination, we found no statistically evidence of an effect of time elongation (t4.455 = 0.747, p = 0.493) and the subjects of the two experiments were pooled for further analysis. Accuracy was not greater than that expected at chance level (54.48 ± 10.03%; t11 = 1.546, p = 0.150). When the last six training sessions were considered, subjects’ accuracy was not greater than that expected at chance level (53.28 ± 9.62%; t11 = 1.179, p = 0.263).2.7. Experiment 5: Reducing Cognitive LoadThe automated procedure may increase the cognitive load of the task, thereby reducing fish learning performance, by requiring the fish to first attain a set of other skills (i.e., learn how to swim through the tank sectors, to approach the stimuli to cause food delivery, swim to the back chamber to prepare a new trial, etc.). In the present experiment, before subjecting guppies to the numerical task, we allowed them to become acquainted with these additional skills by giving them a series of three different non-numerical (shape, color, and size) discrimination tasks.2.7.1. Subjects, Apparatus, and ProcedureFor this experiment, we used eight adult females. The conditioning chamber was identical to those used in Experiment 1. Subjects initially underwent a series of three visual discriminations: the first task required fish to discriminate between a circle (diameter 3 cm) and a 3 × 3 cm triangle on a white background; the second task was a color discrimination between red (red-green-blue color model: 255, 0, 0) and green (RGB: 0, 255, 0); and the third task was a size discrimination between a 3 × 3 cm square and a 1.5 × 1.5 cm square on a white background (ratio: 0.25). Fish that successfully completed this sequence underwent a fourth discrimination that was the same numerical discrimination 3 vs. 12 presented first in Experiment 1. We originally planned to continue the experiment by increasing the difficulty of the numerical task, as in previous experiments. However, as subjects appeared less motivated to complete the task, we decided to skip the last part the experiment and perform instead a second shape discrimination (different from the first one) to compare learning at the beginning of the experiment and after two months. The stimuli were a cross (3 cm in length and 0.75 cm in width) and a 3 × 1.5 cm horizontal bar. The learning criteria and other details of the procedure were identical to those adopted in Experiment 1. Thirteen subjects did not pass the pre-training phase and were replaced with new subjects.We followed the same statistical approach described in previous experiments to analyze subjects’ accuracy. We additionally compared subjects’ performance between the first shape discrimination and the second shape discrimination using a GLMM. The model was fitted with training session and type of discrimination as a fixed factor, and the subject as a random factor.2.7.2. ResultsFive out of eight subjects achieved the shape discrimination according to the primary criterion after performing on average 225.60 ± 156.19 trials (average number of days: 3.20 ± 2.17), whereas the remaining subject did not achieve the first discrimination (average number of trials: 435.67 ± 94.94; Table 7). Overall, the group of subjects did not show an accuracy greater than that expected at chance level (62.43 ± 15.88%; t7 =2.215, p = 0.062). Table 7 shows the individual performance with statistics for all subjects. The GLMM did not reveal a significant improvement in the subjects’ accuracy over training sessions (χ21 = 0.370, p = 0.543).All five subjects achieved the color discrimination according to the primary criterion after performing on average 416.00 ± 266.64 trials (average number of days: 5.60 ± 3.78). Overall, subjects’ accuracy was significantly greater than that expected at chance level (68.52 ± 13.29%; t7 = 3.115, p = 0.004). The GLMM revealed a significant improvement in the subjects’ accuracy over training sessions (χ21 = 43.416, p < 0.001).These 5 subjects also achieved the size discrimination according to the primary criterion after performing on average 373.20 ± 251.21 trials (average number of days: 4.60 ± 3.21). Overall subjects’ accuracy was significantly greater than expected by chance level (75.32 ± 10.15%; t4 = 5.580, p = 0.005, Table 7). The GLMM revealed a significant improvement in the subjects’ accuracy over training sessions (χ21 = 10.594, p < 0.001).Two out of five subjects achieved the 3 vs. 12 numerical discrimination according to the primary criterion after performing on average 564.00 ± 137.18 trials (average number of days: 7.5 ± 2.12), and two additional subjects reached the second learning criterion after performing on average 795.00 ± 7.07 trials. The remaining subject did not achieve it after 596 trials. Overall, subjects’ accuracy was significantly greater than that expected at chance level (60.29 ± 4.80%; t4 = 4.790, p = 0.009, Table 7). The GLMM revealed a significant improvement in the subjects’ accuracy over training sessions (χ21 = 30.408, p < 0.001). The performance on this discrimination was significantly higher than that found in Experiment 1 with automated device (GLMM; procedure: χ21 = 18.474, p < 0.001; training session: χ21 = 29.835, p < 0.001; training session × procedure: χ21 = 8.775, p = 0.003), but did not reach the performance of manual training (procedure: χ21 = 12.190, p < 0.001; session: χ21 = 25.220, p < 0.001; session × procedure χ21 = 4.817, p = 0.028).At this point in the experiment, one subject died, and the others did not appear very motivated to complete the test. We decided to subject the four remaining guppies to a new shape discrimination to determine whether, after nearly two months of experiments and more than two thousand trials, the performance of the subjects had declined. Only one out of four subjects achieved the second shape discrimination according to the primary criterion after 640 trials in 8 days, whereas the remaining subject did not (average number of trials: 746.67 ± 94.94). Subjects’ accuracy was not greater than that expected at chance level (60.20 ± 6.89%; t3 = 2.961, p = 0.060, Table 7). The GLMM did not reveal a significant improvement in the subjects’ accuracy over training sessions (χ21 = 0.800, p = 0.371). A comparison of this task and the shape discrimination at the beginning of this experiment revealed a significant effect of the treatment (GLMM, task: χ21= 10.187, p = 0.001), indicated that subjects showed a higher performance in the first shape discrimination task (Figure 2). There was no significant effect of the session (χ21 = 0.016, p = 0.899), nor the interaction (χ21 = 1.624, p = 0.203).3. DiscussionIn recent years, there have been numerous attempts to produce automated operant conditioning devices for small teleosts. A comparison of studies employing traditional training methods and studies employing automated devices seems to suggest that the latter enhance performance on some tasks [21,28] but worsen it on others [27,30,32]. Using a direct comparison of the two methods, we demonstrated that, when all the parameters of the tasks are controlled for, automated methods reduce the subjects’ performance on a numerical discrimination task. However, we showed that residency and reduction of cognitive load moderately improved the subjects’ performance.3.1. Comparison of Manual and Automated TrainingIn Experiment 1, we compared in the same numerical task subjects trained with a recently developed automated device [28,30] and subjects trained with a traditional approach, a manual operant conditioning protocol that was uses in many previous studies [31,42,43]. We used the protocol that was previously found to be the most efficient in training fish on a numerical discrimination [23]. The subjects were first tested on an easy numerical contrast and the difficulty of the task progressively increased until the fish failed to reach the criterion. This protocol is ideal for making comparisons between different conditions because it allows comparing both learning performance within a single numerical task and the numerical acuity estimated by administering task of increasing difficulty. All of the subjects manually trained achieved the easiest discriminations, 3 vs. 12 and 2 vs. 3 items. Additionally, the two following, more difficult, discriminations were achieved by several subjects (3 vs. 4: 9 out of 11 subjects; 4 vs. 5: 3 out of 8 subjects). The 5 vs. 6 discrimination was not achieved by any subject. This result completely overlaps that obtained in the previous study with a different manual training protocol [23] and could not be attributed to an observer bias effect, as showed by a very high interrater reliability of scoring method. Conversely, no subject trained with automated device reached the criterion in 3 vs. 12 discrimination and only 3 out of 8 fish had a percentage of correct responses significantly above chance. It is worth noting that, with traditional training procedure, guppies were allowed a maximum of 120 trials for each discrimination, and, in most cases, they reached the criterion after a few dozen trials, whereas with the automated device they were allowed to perform up to a thousand trials. The poor numerical performance with automated training procedures observed in a previous study [30] is, thus, confirmed even with the adoption of a most efficient protocol, which starts from simple discriminations to gradually increase the difficulty. We can exclude that the low performance in the Skinner box is merely due to the inadequacy of the apparatus or the procedure since with the same equipment and procedure guppies rapidly reach up to 90% correct response in other types of discrimination, such as color and shape discrimination [28]. This performance is actually higher than previously reported in fish and also superior to that of many warm-blooded vertebrates. Since the proposed numerical tasks are easily discriminated by guppies trained with various non-automated protocols ([23,31], the present study), the cause of the poor performance must be sought in one or more characteristics that differentiate the two approaches to training.One important difference regards the time available to the subjects to gather information about the stimuli, analyze them and make a decision. Indeed, a trade-off between decision speed and accuracy was shown for a variety of species [48,49,50]. Skinner box are typically very reduced in size compared with manual training apparatuses [21,27,51,52]. Specifically, in our study, the distance between the exit from the corridor and the line of choice was 22 cm for the manual training and only 6 cm (approx. one body length) in the automated training. This resulted in a choice time on average three times longer for the manual procedure. The short time interval (~1 s) of the Skinner box may be enough for perceiving the color or the shape of two objects but not sufficient to make an accurate numerical discrimination. In a numerical discrimination, stimuli change for every trial and the subject must independently estimate the number of items in the two stimuli, compare these two quantities and thereby decide based on a learned rule. In addition, if number discrimination requires a longer processing time, the possibility of inhibiting and correcting initially wrong choices may be very reduced in the Skinner box due to the much shorter time allowed to the subject before it reaches the choice areas [30].An additional difference between color or shape discriminations and numerical discriminations is that the latter requires the fish to observe both stimuli in their entirety. A subject, for example, can perform a correct color discrimination even if one or both stimuli are partly hidden from view. Conversely, a larger of two numerosities may become the smaller one if some of its items are not visible. The tiny space of the automated conditioning chamber may cause a problem in visibility of the stimuli that was not present in the large tank used for manual training.Another possibility for explaining the different performances is that automated procedures increase the cognitive load of the task by simultaneously requiring that the fish learn to solve the discrimination, to swim fluidly through the sectors, to associate its approach to the stimuli with food delivery, and to swim back to the starting chamber to launch a new trial. This hypothesis is corroborated by the observation that in experiments with automated training procedures many subjects are discarded before the training phase starts due to their inability to learn how to operate the device ([28,32,53], the present study). In many vertebrates, including fish, the addition of concurrent tasks increases cognitive load and decreases the performance on the primary task [54,55]. This is expected to occur, particularly when subjects are performing complex tasks [36,37]. Although it is not clear whether numerical discriminations are more cognitively demanding for fish than color or shape discriminations [56], they certainly imply a larger number of cognitive operations. In a shape or color discrimination, a subject sees the same pair of stimuli throughout the experiment and can solve the task simply by learning to approach the positive stimulus or move away from the negative one without necessarily making a comparison of the two. Conversely, in a numerical discrimination task, the two stimuli change at each trial because position, size and density of the items vary systematically. In each trial, the subject must make a new estimate of the two quantities, compare the two numerosities and, hence, apply the learned rule (e.g., choose the larger one). It is plausible that the additional cognitive load related to the use of the automated device is particularly detrimental when a fish is required to perform such a task.Finally, the two procedures differ in the amount of manipulation received by the subjects during the experiment. As usual for conditioning protocols of fish (e.g., [43,57,58]), in the manual training experiment, guppies resided for the whole experiment in their experimental tank. In the automated training experiment, by contrast, they resided for most of the time in their home tank, and once or twice a day, they were netted and transferred to the conditioning chamber for the duration of the test. Netting and change in the environment, even for a short period, may cause severe and long-term stress in teleosts [59,60,61,62], and even chronic stress in case of repeated events [63]. Given the well-known negative impact of stress on cognitive performance [39,40,41], the manipulation required for the automated procedure may cause fish reduced discrimination learning success. Being in a new environment is also expected to cause short term increase in vigilance against predators, which typically reduces the attention devoted to foraging activities and decreases food-finding efficiency [54,55,64].Other differences between the two methods of training could potentially be involved. For example, the number of trials per sessions or the presence of correction trials have been sometimes found to affect learning rate in other organisms [65,66]. However, they do not normally have dramatic effects on task achievement. In addition, an automated procedure with all these features led to performance equal or even superior to the traditional methods with color or shape discrimination [28] and with the discrimination of the size of stimulus (see Experiment 5) and it is difficult to devise hypotheses that easily explain why these features should selectively compromise learning in numerical tasks.3.2. Can Automated Training Devices Be Improved?Following the above hypotheses, in the second part of this study we tested four modified versions of the setup to see whether we could detect the key factors and reduce the gap in numerical discrimination performance between the automated training approach and the other methods.Elongation of decision time. The first variant was intended to force guppies to take a longer time interval before making their choice (Experiment 2). The size of the conditioning chamber was increased, almost doubling the distance between the corridor and choice areas, and a transparent barrier was placed between the corridor and the choice areas to further lengthen the distance to reach the stimuli. Only one out of the six subjects learned to discriminate the 3 vs. 12 numerical discrimination, but this subject failed the subsequent 2 vs. 3 discrimination. These results suggest that the lengthening of the decision time, at least as it was induced here, is not enough to improve performance.Removal of internal partitions. In Experiment 3, we further modified the conditioning chamber of the previous experiment by removing all internal dividers so that the animal had a complete view of the two stimuli from afar, as happened in the manual training tanks. None of the ten subjects tested in this experiment learned the 3 vs. 12 discrimination, suggesting that the visibility of the stimuli was probably not a crucial factor in determining the difference between the two training approaches of Experiment 1.Subjects resident in the conditioning chamber. In Experiment 4, we modified the chamber so that we could keep the experimental subject there for the entire duration of the experiment, as was the case in manual training experiment. With this modification, 12 out of the 16 subjects achieved the 3 vs. 12 numerical discrimination, a considerable improvement on the setup of Experiment 1. Fish likely adapt less well to being moved daily between the home tank and the experimental chamber, relative to domestic species of mammals and birds, such as rats and pigeons for which the original Skinner box was developed. In the rainbowfish, Melanotaenia duboulayi, familiarity with the testing apparatus was found to increases task performance probably because familiarity decreases stress, allowing subjects to pay more attention to the task [67]. This hypothesis may also explain the unexpectedly low performance in numerical tasks found in a study in which guppies undergo a traditional manual training but were moved in turn to the conditioning chamber for the trial [44].In order to allow the subject to reside in the apparatus, in this experiment we had to increase the size of the conditioning chamber. Therefore, in this experiment the size of the chamber is a potential confounding factor. One might for example argue that guppies are not at ease in tiny spaces and would be stressed by being tested in small apparatuses. However, in their natural habitat, guppies avoid open spaces and are generally found in the proximity of the river’s margins, in small pools that contain a fraction of the volume of our Skinner box and they often forage in small gaps of thick vegetation that can host just one fish [68,69,70]. As a further confirmation of the fact that the size of the chambers per se is unlikely to represents a key factor, excellent performance was obtained with the smallest version of the conditioning chamber in other tasks [28], and some studies have found good numerical performances using very small manual training apparatuses (e.g., [71]). It is worth noting the global performance of fish resident in the conditioning chamber was still inferior to the manually trained group of Experiment 1. Being resident in the test apparatus is not the only factor that determines the difference between the two approaches of training used here.One of the advantages of the Skinner box was that a single small-sized equipment could be used to train several rats or pigeons in rotation. Maintaining the fish in the conditioning chamber improves performance but suffers the major drawback that it implies the use of a large tank for each experimental subject and therefore experiments require much more space and equipment. A compromise may be the one adopted in a previous study [30], in which the subjects are semi-residential, i.e., they live in a tank that houses the conditioning chamber to which they are moved in turn at the time of the test. In this way, the relocation is rapid and the subject stays in identical physical, chemical and olfactory conditions for the duration of the experiment. A rough comparison between the results of the two studies suggests that this solution is sufficient to achieve some improvement. In our Experiment 1, only 3 of the 8 subjects had a performance greater than that expected by chance; this happened for all subjects but one in the previous study, though in the latter, the ratios to discriminate were more difficult. Indeed, the performance significantly differed between the two studies (Mann–Whitney U test; p = 0.045). It is also possible to devise a conditioning chamber in which the subjects are accustomed to leave their group in turn, and swim to the testing area at the time of test, a common practice in primate studies (e.g., [72,73,74]).The results of Experiment 4 have potential implications for the welfare of the laboratory studies in fish. Moving repeatedly the subject forth and back between home cage and the testing apparatus is a common practice of cognition studied in fish, as it is in mammals and birds (e.g., [4,18,75]). However, this practice could induce stress in fish more than in other vertebrates, especially if the experiment is prolonged, as also suggested by the outcome of Experiment 5 (see below).In this experiment we also manipulated inter-trial interval, another factor that frequently differentiates automated conditioning approach from traditional ones and that also differed considerably in Experiment 1. We increased the inter-trial interval in half of the subjects from 10 s to 10 min, making it very similar to the inter-trial interval adopted in the manual training experiment, but we found no evidence that this treatment affected discrimination learning or numerical acuity of the subjects.Reducing cognitive load. The aim of Experiment 5 was to reduce the cognitive load during the execution of the task by uncoupling learning the use of the conditioning chamber from learning the numerical discrimination. Before being admitted to the numerical discriminations, guppies underwent a shape, a color, and a size discrimination. The performance on these three tasks was substantially similar to that observed in previous studies employing different training procedures [23,29]. When subjected to the 3 vs. 12 numerical discrimination, guppies showed an evident improvement compared to the automated training of Experiment 1, although their performance was significantly lower than in the experiment with manual training.At the end of this series of tasks, after nearly two months of experiments and more than two thousand trials, one animal died, and others showed signs of stress. Therefore, we decided to stop the experiment and we do not know the quality of the performance of the fish of this treatment on the more challenging numerical discriminations. This outcome tells us that, unlike monkeys and mice, fish cannot be tested for unlimited time durations, at least not if it is necessary to move them frequently from their home tank to an experimental chamber. A solution to the problem of moving fish may be to have the subject reside in the experimental chamber as in Experiment 4, in which this solution seems to have brought benefits in terms of performance. It is necessary to verify whether guppies that are resident or semi-resident in the conditioning chamber can do a thousand trials without suffering or diminishing their performance.The first part of Experiment 5 confirms previous finding about the efficiency of our automated training device for discrimination learning [28,32]. Considering the first days of training in each discrimination, before the best performing subjects were admitted to the next discrimination, guppies evidenced an excellent performance in color, shape and size discriminations reaching 75–80% correct choices in few sessions. It is interesting to note that the first part of this experiment also shows that guppies, like mammals and birds [76,77], can learn a series of different discriminations in rapid sequence without apparent interference of one task on the subsequent.3.3. A Comparison with Other Studies Using Automated Conditioning DevicesAs mentioned in the Introduction, learning, memory, and many other cognitive functions are investigated in warm-blooded vertebrates using primarily automated procedures whereas manual training is the standard procedure used with teleost fish, with other lower vertebrates and with invertebrates. Some laboratories have attempted overcome this limitation by developing a Skinner-box apparatus for fish. In the 1960s and 1970s, many studies investigated learning and memory in the goldfish using fully automated operant conditioning devices with the aim to compare a fish species with the classical mammalian and avian model species [78,79,80]. When discrimination learning experiments were performed, goldfish appeared to master well color [79,80,81] as well as shape discrimination [82,83].In recent years, some small tropical fish, in particular zebrafish, medaka and guppy, have become important models in neurobiological research. Taking advantage of the emerging digital technologies, there have been half a dozen independent attempts to implement automated devices, obtaining mixed results. Manabe and colleagues [18] set up a computer-controlled operant apparatus for small fish, in which stimuli were presented by means of LEDs and the approach to response key was sensed through an optical fiber sensor, which triggered an automated feeder delivering small amount of food. With the same apparatus, Kuroda and colleagues [19], using a color discrimination, demonstrated that zebrafish can learn a reversal learning task.Color discrimination is, by far, the most frequently investigated discrimination task in the remaining studies that employed automated training approach. Another device allowed for the simultaneous training of six subjects, each resident in an adjacent chamber. A computer tracked the position of each fish, controlled stimulus presentation on a LCD monitor and commanded the delivery of the food reward [21]. The apparatus was tested by training zebrafish on a blue-green color discrimination. Subjects slowly improved their performance over 30 daily sessions of 20 trials each, reaching 80% of correct choices at the end of experiment. Conversely, Miletto Petrazzini and colleagues [27], who tested zebrafish with a commercially available operant conditioning apparatus, found a much lower performance on two color discriminations. With the exception of the latter study, the others were in agreement with the results obtained with our device in the present study or in previous ones [28,32].Two studies conducted in our laboratory investigated shape discrimination. Guppies showed good learning performance, fully comparable with the results of manual training experiments [28]. Conversely, with zebrafish, no subject reached the criterion on the same shape discrimination task, even after extended training [32], despite this species showing 75% accuracy in color discrimination tasks with the same device. There is scarce information about the discriminative abilities of this species except for color discrimination. Zebrafish proved able to discriminate novel shapes from familiar ones [34,35]. However, the sole study that examined shape discrimination in zebrafish using appetitive conditioning was an interspecific study in which zebrafish showed a rather poor performance compared to other teleosts [25].To date, only another laboratory investigated numerical discrimination in a fish, Gambusia affinis, using an automated system [33]. The device and the procedure differed from those used in the other studies. During the training phase, the quantities to discriminate were presented on two monitors placed at the opposite ends of a rectangular tank and a computer-operated system delivered a food reward when the subject approached the correct stimulus. Learning was measured in a probe trial in which the subject was exposed to the stimuli without reinforcement, and the percentage of time spent near the positive stimulus was calculated from video-recordings. This experiment replicated a study on the same species with an identical procedure, but in which the training phase was done manually [84]. A rough comparison of the two studies shows that, even in this case, the performance on numerical discrimination tasks was better when fish were trained manually (p = 0.016; Welch t-test).4. ConclusionsIn summary, the results of this study confirm that the automated training device we developed modelling the classical Skinner boxes can satisfactorily be used to train guppies in some tasks (i.e., color, shape, and size discriminations) but are totally inadequate for other tasks, such as a numerical discrimination. Similar inconstancy in performance have been reported with other automated devices in different fish species and with different tasks which suggests that small laboratory teleosts may be limited in their capacity to cope with some undetermined aspects of the automated approach to training. These limitations need to be overcome, as they presently prevent an easy and straightforward comparison of teleosts with the other vertebrates.In a series of experiments, we introduced modifications to the automated training apparatus and procedure in an attempt to fill the performance gap with the best performing manual training procedure available for guppies. The difference did not appear to be related to a longer decision time in manual procedure nor to the different visibility of the stimuli in the two procedures. Being resident in the test tank improved the performance of the subjects but this factor alone did not lead to the same performance of the manual training experiment. A similar effect was obtained by reducing the cognitive load through the temporal dissociation of the familiarization with the functioning of the automatic equipment from the numerical discrimination task.It, therefore, seems that there is no single factor that explains the different efficiency of the two procedures tested in this study, but rather several factors acting synergistically to determine the different performance. There are many other differences between the two methods compared in this study, and in future, it will be necessary to investigate other factors that may be important. One interesting difference that was not addressed concerns the way the stimuli are presented in the two approaches, generated on a computer screen or stuck onto real objects introduced into the water. Computer-generated stimuli have already been used successfully in fish for visual discrimination tasks [28,85,86] and a vast literature shows that fish react to stimuli presented on the monitor as they react to the real objects [87,88,89,90,91]. However, it is possible that stimuli introduced into the tank are much more salient and focus the subject’s attention on the task. Indeed, in several species the salience of the stimulus was shown to influence the performance on discrimination tasks (capuchin monkeys [92]; pigeons [93]; keas [94]) and there is also partial support for this effect in the guppy [46]. It will be interesting to verify whether an automated procedure that commands the presentation of solid stimuli introduced into the subject’s tank allows for a performance comparable to that we obtained in our study with the reference manual training procedure.

animals : an open access journal from mdpi

10.3390/ani13101602

PMC10215146

The aim of this study was to characterize the RHU competencies according to the distance (short vs. long), causes of deaths, and associated risk factors. The studied population comprised 16,856 horses that participated in RHU rides from 2007 to 2018. During the entire period, there were 99 fatalities. The percentage of inexperienced horses and those who completed the ride was greater in short races than in long races. In both types of rides, more horses died during than after the ride, and inexperienced horses were more likely to be dead than horses with prior experience in the sport. Short rides were associated with increased risk of sudden death, while long rides were associated with increased risk of death due to metabolic alterations.

RHU is the oldest endurance sport in Uruguay. However, despite 80 years of racing, there are no studies to characterize this type of competition, explore rates and causes of death, and identify the associated risk factors. The aim was to characterize the Raid Hipico Uruguayo (RHU) competencies according to the distance (short (SR, 60 km) vs. long (LR, 80–115 km)), the causes of deaths, and the associated risk factors. The study population comprised horses (n = 16,856) that participated in RHU rides from 2007 to 2018. LR were more frequent than SR (p < 0.001). The average speed of winners was higher in SR (32.12 km/h) than in LR (28.14 km/h) (p < 0.001). There were 99 fatalities (5.9 per 1000 starts). SR had greater frequency of high comfort index (CI = Temp [°F] + Humidity [%]) than LR, and LR had greater frequency of low CI than SR (p < 0.001). The percentage of inexperienced horses and those who completed the ride was greater in SR than in LR (p < 0.001). In both types of rides, more horses died during than after the ride, and inexperienced horses were more likely to suffer fatalities than horses with prior experience in the sport (p < 0.05). SR were associated with increased risk of sudden death, while LR were associated with increased risk of death due to metabolic alterations. The high fatality index shown in this work warrants urgent investigation in this sport to minimize mortality associated with RHU-specific diseases.

1. IntroductionEndurance equestrian sports have a long history, but it has experienced great growth in recent decades, mainly FEI endurance. As described by the FEI, “Horsemanship and Horse welfare are the core of endurance riding. Endurance is a test of the Athlete’s ability to manage the Horse safely over an Endurance course. It is designed to test the stamina and fitness of the Athlete and Horse against the track, distance, terrain, climate, and clock, without compromising the welfare of the Horse” [1]. Because metabolic disturbances and deaths occur more frequently than in any other type of equestrian discipline, all equine resistance sports have a strict veterinary control that ensures the health of equine competitors [1,2]. A series of veterinary inspections and examinations are established in the interest of the health, safety, and welfare of the horse in the competition. Only competitors whose horses have passed all the inspections and examinations are entitled to be classified in the final list of results [1,3].Elimination rates appear to have increased over recent years, which is a source of concern for the sport’s ethics and image. Main reasons for elimination are lameness and metabolic disturbances, associated with dehydration and electrolyte disturbances, and with substrate depletion in active muscle fibers. Moreover, there are severe consequences of this metabolic derangement, including heat stroke, rhabdomyolysis, colic, kidney and liver insufficiency, laminitis, and disseminated intravascular coagulation [2,4].The Raid Hipico Uruguayo (RHU) was the first endurance sport in Uruguay. This sport dates from 1944 and since then, rides have been regulated by the Uruguayan Equestrian Federation (FEU). It is the most popular and typical equestrian discipline of Uruguay, with some unique characteristics. RHU is considered as an endurance discipline. Research into veterinary problems in endurance horses is increasing, but there is almost no information available on endurance races not regulated by FEI. In theory, most of the information generated from FEI endurance races could be applicable to RHU horses. However, many of these practices were not useful in RHU horses, and most management, training and nutrition techniques come from empiric experience without any scientific basis. Something similar occurs with many other equestrian sports that are scientifically overshadowed by disciplines with common characteristics. However, small differences in the sports can produce large differences in metabolic and locomotor behavior, with strong impact on the athletic horse.Ride distances vary from 60 to 115 km and according to this, FEU has determined 2 categories: short rides (60 km) and long rides (80–115 km), all of which are divided in only 2 phases, being the first phase of long duration (2/3 of the total distance). Winning horses average speeds from 25 to 37 km/h, reaching top speeds of 50 km/h. The winning horse is the first to cross the finish line and meets subsequent veterinary requirements. For the rest of riders, the cut-off time of crossing the finish line is 45 min after the winning horse arrives [3].The breed of the horses competing in RHU was not officially registered until 2019. In an unpublished study from Brito et al., of 305 horses that raced in 1 season of RHU, 237 were crossbreeds, 39 thoroughbreds, 15 Anglo-Arabians, and 14 Arabians horses. Horses called crossbreeds mostly had more than 75% thoroughbred blood.The competitions take place on flat terrain with mostly hard surfaces, and the minimum weight of the riders should be 85 kg. Horses and riders can compete in any ride, regardless of their previous experience. Horses are examined by official veterinarians before the ride, after the first phase, and at the end. Veterinary control after the first phase is performed 20 min after the arrival of the horse, and once passed, horses must wait a 40-min compulsory rest period before starting the last phase. Horses are eliminated from the ride if veterinarians consider their metabolic status or orthopedic condition not to be adequate to enable them to continue the ride [3]. Sixty-one percent of the participants do not finish the ride due to lameness or metabolic reasons [5], which can sometimes lead to fatalities.From the point of view of health and welfare, the death of equines during or after competition is a major concern for vets, riders, organizers, and for spectators. Studies on the causes of death in sports horses are scarce and mostly refer to racehorses, being the main causes: sudden death [6,7] and catastrophic musculoskeletal injury [8,9].Fatalities during endurance exercise are recognized as a consequence of prolonged exercise, but data documenting incidence and causes are very limited. Balch et al. (2019) reported 127 fatalities out of 335,456 starts (0.28 fatalities per 1000 starts) during the period 2002 to 2018 [10]. According to Balch et al. (2019), 77% of deaths were attributed to the high demands of endurance exercise (leading to severe muscle cramping and exhaustion, mostly attributable to acute abdominal pain) and 33% due to injuries not associated with the metabolic demands of endurance exercise (such as falling off a cliff or the trail, kick injury) [10]. In addition, the risk of death increased with the distance traveled (0.12, 0.35, and 1.48 fatalities per 1000 starts in rides of 48, 80, and 160 km, respectively) [10].Although this sport (RHU) has been carried out in Uruguay for several decades, there are no reports characterizing the ride conditions, as well as the causes of death and their risk factors. We hypothesize that although this sport has similarities with other endurance disciplines, it has very different characteristics that affect the causes of death during races. Therefore, the aim of this study was to characterize the RHU competencies according to the distance (SR vs. LR), the causes of deaths, and the associated risk factors.2. Materials and MethodsThe study was performed with the endorsement of the FEU, guaranteeing confidentiality regarding the names of the horses and owners.2.1. Data Collection and Studied VariablesThis was a retrospective cohort study. Data from all RHU rides from 2007 to 2018 were collected from FEU archives. All rides were contested annually between the months of March to December. The information obtained was entered into a computerized database, and each RHU ride was assigned a unique identification number. The database included the temperature and humidity, ride type and length, number of horses that started, number of eliminated and retired horses (total, by phases and reason), number of inexperienced horses (no RHU racing experience), number of horses that completed the ride and average speed (km/h) of each phase (phases 1 and 2), and average speed of the winning horse. If a variable was not recorded, the variable was assigned a value of not available. Reports with incomplete data for more than two variables were not included in the study. When fatalities occurred, the cause of death was recorded. To establish the cause of death, necropsy examinations by official veterinarians were performed on all dead equines. All data were obtained from FEU reports.2.2. Variable CategorizationComfort index (CI) was calculated by the sum of the temperature in Fahrenheit degrees and the relative humidity as a percentage [11]. CI was classified into three categories: low (CI < 130), medium (CI 130–150), and high (CI > 150). Ride types were classified by the length into two categories according to the FEU designation: short ride (SR: 60 km) and long ride (LR: 80–115 km). Causes of death were classified in four categories [9]: metabolic conditions (colic, exhausted horse syndrome, disseminated intravascular coagulation), catastrophic musculoskeletal injuries (defined as horses that died or were euthanized due to severe acute bone fractures that carry a poor clinical prognosis), sudden death (defined as acute death in a closely observed and previously apparently healthy animal), and accidental.Additionally, the part of the race where the death occurred was classified as during the ride (phase 1 or phase 2) or after the ride (24 h after the finish of the ride) to know if the horses that suffered fatalities completed the course or not.2.3. Statistical AnalysisDescriptive analyses (mean and standard deviations or percentage values) were calculated for the variables CI, horse experience, completed ride, speed (phase 1, phase 2, and average), overall fatalities, cause of death, and ride type. Statistical differences by ride type were calculated using a Chi-squared test, Fisher’s test, or Wilcoxon rank sum test. Screening of all exposure variables for overall fatalities (Live/Death) were performed separately using univariable Logistic Mixed Model analysis. Only variables with p < 0.2 were considered for inclusion in the multivariate Logistic Mixed Model analysis. In both analyses, all variables were considered fixed effects and the ride was considered as a random effect. The multivariate models were built using a forward selection procedure whereby variables with a Wald-test p < 0.05 were retained in the model. p-values of less than 0.05 were considered statistically significant. All analyses were performed in R (Version 4.2.2, 2022) and RStudio (version 2022.12.0 Build 353) [12].3. ResultsFrom a total of 702 RHU competitions taking place between 2007 and 2018, there was a 3-fold greater frequency (p < 0.0001) of LR (509, 42 rides per year) than SR (193, 16 rides per year). There were 16,856 horse starts, of which the number of horses was also greater in LR than in SR (Table 1, p < 0.0001). The average of horses competing per ride was 22 in SR and 29 in LR. Overall, 40.5% of the horses completed the ride, 43.1% were not able to complete the course due to metabolic reasons, 12.3% did not complete because of lameness, while 4.1% of the horses were retired from the ride due to rider-decision. The average speed of the winning horses was 28.1 km/h and 32.1 km/h (Table 1), with maximum and minimum speeds of 32.6 km/h and 20.3 km/h for LR and 35.9 km/h and 25.5 km/h for SR. The highest average speed in each phase and the average speed of the winning horse were recorded in SR (Table 1).Over the 12-year study period, there were 99 fatalities, and 68 of these horses were euthanized. All the euthanasias were performed with a rigorous criterion evaluating the future life and the suffering of the horse. The risk of fatality over the entire period was 5.87 per 1000 starts. The average number of deaths per year were 8.25 and did not differ over the years studied. The risk of fatality was significantly greater (p = 0.05, odd ratio = 1.52) for participation in SR (7.9 fatalities per 1000 start) than in LR (5.2 fatalities per 1000 start) (Table 2). There were significant differences in causes of death by ride type (Table 1). Short rides were associated with a greater proportion of sudden death, and LR were associated with a greater proportion of deaths due to metabolic alterations (Table 1). Catastrophic injuries occurred in a high proportion in both ride types (Table 1). Of the total deaths for metabolic reasons, 24 (69%) of the fatalities developed acute abdominal pain, 7 (20%) equine exhausted syndrome, and 4 (11%) disseminated intravascular coagulation.CI varied with the type of ride, SR had a greater frequency of high CI in comparison to LR; and LR had a greater frequency of low CI than SR, while there were no differences between ride type for the medium CI (Table 1). However, when CI was evaluated separately for each ride type, CI did not represent a significant risk factor for death in either ride type (Table 2).Inexperienced horses were more likely to suffer fatalities than experienced horses in both ride types (Table 2 and Table 3).There was no association between experience and the probability of dying in LR (Table 2), but under a multivariate analysis (Experience and Completed ride), a significant association was found between these variables in LR (Table 3).Horses that participated in SR completed the ride in a greater proportion than those that participated in LR (Table 1). Besides, regardless of ride type, more horses died during the ride than after them, so most of them did not complete the ride (Table 2 and Table 3).4. DiscussionThis is the first study to characterize RHU rides according to ride-type (short vs. long), as well as the causes of death and their risk factors. This was a retrospective cohort study, where many variables, such as sample population and variability associated with horse background and the environment of the races, could not be recorded. The extent of the limitation should be considered for the interpretation of the results.Equestrian endurance sports require the greatest physiological demand for the athlete horse FEI [2,4,13,14]. The FEI endurance is the most widely described [2,4,15,16,17]. The racing distances were classified as short and long according to the FEU regulations. Long FEI endurance races would be longer than 140 km. Compared to FEI endurance races, RHU has higher speeds, greater weight load, fewer stages, a lower proportion of horses that finish the race, and a higher fatality rate. According to our data, RHU is most likely one of the most demanding events for horses [5].During the period 2007–2018, LR were more frequent than SR, which follows the same pattern as FEI endurance 14. However, unlike endurance, where the speeds of the races were below 25 km/h [14,18,19,20], RHU races were faster, with speed averages between 28 km/h and 32 km/h, reaching maximum speeds close to 36 km/h. In addition, horses completing RHU races were 10 to 40% less than those reported for endurance [20,21,22]. Another clear difference with endurance is that while in RHU the main cause of elimination was due to metabolic condition (43%), in endurance the highest percentage of elimination (25 to 40%) was due to lameness [14,21,22,23].Among the main RHU characteristics, the following are briefly highlighted: LR are more frequent than SR; SR are faster than LR ones; SR had a higher frequency of high CI than LR, and LR had a higher frequency of low CI than SR; and the percentage of inexperienced horses and those who completed the ride was greater in SR than in LR.Regarding fatalities, there were 99 deaths, of which 68 were euthanasia. In both types of rides, more horses died during than after the ride. Most causes of fatalities are incompatible with the endurance exercise. The horses that suffered fatalities and finished the course were due to metabolic causes, except for one animal that presented sudden death. The higher rate of occurrence of metabolic fatalities in LR most likely induced the higher course completion in horses that died during LR (Table 1 and Table 2).The probability of suffering fatalities was higher in inexperienced horses. In addition, SR were associated with increased risk of sudden death, while LR were associated with increased risk of death due to metabolic alterations. It is interesting to note that although equine resistance sports are very popular throughout the world and that their popularity has been growing [24], according to our knowledge, except for one study [10], there are no reports on equine fatalities in endurance equestrian sports. During the 12-year study period, the fatality rate in RHU was 5.87 per 1000 starts.This result is much greater than that reported by Balch et al. (2019) in endurance horses under AERC rules (0.28 fatalities per 1000 starts) [10]. Horse deaths in RHU competitions attract the attention of public opinion and negatively affect the sensitivity of the public towards these sports; they also generate controversial opinions regarding the intensity of the exercise carried out by the animals that participate.The higher fatality rate of the RHU compared to the endurance competition once again highlights the high metabolic and locomotor demand that this sport represents for the horse. The metabolic requirements demanded by high speeds over long distances, with only two stages, create highly challenging conditions for the RHU horse. Additionally, horses run mostly on hard ground, with a high rider weight. Another characteristic of the sport is the important prize money, which adds excitement to the already exciting competition, and can distract competitors from the horse’s state of health and well-being. Since the death of equines during competitions represents a major welfare concern, it should be a priority to know the frequencies and causes of death in all equine sports worldwide. It is a great challenge for veterinarians to try to minimize the frequency of equine deaths during sports. In this sense, the fatalities in the RHU show the need for greater controls and strictness in the limits to which equines are exposed.In this study, the type of the races (SR vs. LR) in RHU was influenced by the comfort index, the percentage of inexperienced horses, whether they completed the race, the speeds, and the number and cause of fatalities. The SR had higher speed, higher frequency of high comfort index, greater percentage of inexperienced horses, and those who completed the ride in comparison with LR. The comfort index was used as an indicator of thermal stress in this work. It is widely used due to its simplicity and low cost, but there are significant weaknesses due to misleading for many combinations of temperature and humidity. Many indices have been evaluated to assess heat stress, even in horses, showing a better predictive capacity. This limitation should be considered for the interpretation of the results.The risk of death of horses was higher in SR than in LR, unlike what was reported in endurance by Balch et al. (2019), in which the risk of death increased with the distance traveled (0.12, 0.35, and 1.48 fatalities per 1000 stars in rides of 48, 80, and 160 km, respectively) [10]. It is possible to speculate that, although less distance is covered, the big locomotor and metabolic demand imposed by a greater speed in SR than in LR determined a greater proportion of equine deaths—mainly catastrophic and sudden deaths. According to the type of death, in relation to the type of race, SR had some similarities with racehorses, since deaths due to catastrophic injuries and sudden death predominate [2]. On the other hand, LR agreed with what has been reported for endurance events where the highest proportion of causes of death were metabolic alterations and catastrophic injuries [6,7]. In previous studies [24,25] no associations were found between the speed of individual horses and elimination for lameness or metabolic reasons. Although in our study, the speed of the races in which horses died did not differ statistically from those in which they survived, we cannot conclude that speed does not influence the deaths of horses. Recorded speed was the average speed of the stage in which they died, or the average speed of the race if they managed to finish it. Therefore, the individual speed of the horse was not considered, nor was the accumulated distance—which requires another study design for its analysis.The highest proportion of deaths in SR was due to sudden death, defined as acute, exercise-associated death in a closely observed and previously apparently healthy animal [26]. Cardiovascular disease is often confirmed or suspected [27]. Most of the studies on sudden death are in thoroughbred racehorses and eventing horses [7,28]. According to Lyle et al. (2011), the prevalence of sudden death in racehorses in the UK between 2000 and 2007 was 0.28 deaths per 1000 starts [6]. Comyn et al. (2017), found a prevalence of 0.14 death per 1000 starts in FEI eventing horses between 2008 and 2014 [29]. In the present study, the mortality rate from this cause was 1.02 deaths per 1000 starts, which is higher than any previous report. Intense exercise requires high oxygen consumption such as horse racing or three-day events, producing large increases in cardiac output and blood pressure, increasing propensity for major cardiovascular events [6,29].Navas de Solis et al. (2018) studied cases of sudden death in many types of equestrian sports and riding horses, and reported that 71.9% of the horses that suffered sudden death collapsed during exercise [27]. These results are in contrast to the work of Lyle et al. (2011) on racehorses, where most sudden deaths occurred after the race [6]. Exercise time during competitions in racehorses is much less than in three-day eventing or RHU. The intensity and duration of the exercise most likely play an important role in the incidence and moment of presentation of sudden deaths in sport horses. Horses collapsing during exercise may suffer catastrophic lesions, and riders may fall and be injured during these episodes and while riding a horse that presented sudden death. Therefore, the study of the causes of sudden death during RHU, and its prevention may be imperative, not only for increasing the horse’s welfare, but also for human safety concerns.The most frequent cause of death in LR was metabolic condition. This cause of death occurs mainly in horses after long duration submaximal exercises [2,4]. The marked increase in metabolism for such a long period of time is accompanied by intense energy consumption and loss of body water and electrolyte stores, as a consequence of thermoregulation [2,4,10,14,15]. Mild dehydration is compatible with competitive performance, however, serious medical problems may develop in horses that compete in endurance events—even death [2,15]. The distance, due to the longer time in exercise, most likely has an influence on the presentation of metabolic alterations, since it represents a risk factor for its presentation in endurance horses [14]. In our work, the proportion of deaths due to metabolic alterations was higher in LR.The main cause of death due to metabolic disorders was Acute Abdomen Syndrome (AAS). Balch (2019) reported AAS as being the most common clinical presentation (85%) as a prelude to death in AERC endurance races [10]. The management of the horse before and during the race can lead to episodes of AAS, such as gastric dilatation due to the effect of fasting before the race and peristalsis problems due to dehydration and electrolyte imbalances [2,15].Catastrophic injuries occurred in a high proportion in both ride types. In this study, the animals with catastrophic injuries did not complete the ride (43/44), being euthanized when an irreparable injury was diagnosed. Of the total catastrophic injuries, 11 were in SR and 33 in LR, although no differences were found. In the present work, the mortality rate due to catastrophic injuries was 2.6 dead horses per 1000 starters, which is even higher than those reported in racehorses or in three-day eventing horses.According to Misheff et al. (2010), the risk of suffering from bone pathologies during endurance races increases with the distance covered and the increasing speeds [30]. Most epidemiological studies show a positive association between the risk of catastrophic injuries and distance covered, although not all findings have been consistent [31].In our study, we also show that experience is a risk factor for fatalities. Therefore, another element that could have contributed to the higher risk of death in SR was the greater proportion of inexperienced horses in comparison to LR. Previous studies in sporting horses as endurance [25] or thoroughbred racehorses [32,33] have found experienced horses to be associated with reduced odds of deleterious outcomes compared to horses that were inexperienced (in endurance for the distance) or less experienced thoroughbred racehorses. However, the concept of inexperience is not the same, since in FEI races there is a qualification system that does not allow horses to compete if they have not completed shorter distance races. Therefore, they are only inexperienced at that distance. In our case, although many animals had competed in other sports, none of the horses considered inexperienced had competed in RHU races. The fatality rate for inexperienced horses was 13.2 and 8 fatalities per 1000 starts in SR and LR, respectively.Based on this set of results, it is clear that although LR are more frequent than SR in RHU, the risk of death is higher in SR than in LR. Therefore, the need to control racing conditions is evident, especially in SR, where speeds are higher.The differences and characteristics of the RHU as an equestrian endurance sport shown in this study, with notable consequences for equine athletes, warn about the importance of independent evaluation of each discipline to ensure the well-being of our equine and human athletes in all equestrian sports. Further studies are needed to assess more specific risk factors, in order to provide objective data that can help to plan racing schedules and serve as a basis for regulations, ultimately improving both the welfare of RHU horses and the public perception of this discipline.5. ConclusionsSR and LR have important differences that are manifested in the characteristics of the race, the causes of death, and the risk of fatality. This must be considered to understand the physical impact of equine participation in this type of event.The high fatality index shown in this work, especially in inexperienced horses, warrants urgent investigation in this sport to minimize the mortality associated with RHU-specific diseases and to improve the welfare of the RHU horses.

animals : an open access journal from mdpi

10.3390/ani11092614

PMC8465983

The red deer (Cervus elaphus) de novo genome assembly (CerEla1.0) has provided a great resource for genetic studies in various deer species. In this study, we used gene order comparisons between C. elaphus CerEla1.0 and B. taurus ARS-UCD1.2 genome assemblies and fluorescence in situ hybridization (FISH) with bovine BAC probes to verify the red deer-bovine chromosome relationships and anchor the CerEla1.0 C-scaffolds to karyotypes of both species. We showed the homology between bovine and deer chromosomes and determined the centromere-telomere orientation of the CerEla1.0 C-scaffolds. Using a set of BAC probes, we were able to narrow the positions of evolutionary chromosome breakpoints defining the family Cervidae. In addition, we revealed several errors in the current CerEla1.0 genome assembly. Finally, we expanded our analysis to other Cervidae and confirmed the locations of the cervid evolutionary fissions and orientation of the fused chromosomes in eight cervid species. Our results can serve as a basis for necessary improvements of the red deer genome assembly and provide support to other genetic studies in Cervidae.

The family Cervidae groups a range of species with an increasing economic significance. Their karyotypes share 35 evolutionary conserved chromosomal segments with cattle (Bos taurus). Recent publication of the annotated red deer (Cervus elaphus) whole genome assembly (CerEla1.0) has provided a basis for advanced genetic studies. In this study, we compared the red deer CerEla1.0 and bovine ARS-UCD1.2 genome assembly and used fluorescence in situ hybridization with bovine BAC probes to verify the homology between bovine and deer chromosomes, determined the centromere-telomere orientation of the CerEla1.0 C-scaffolds and specified positions of the cervid evolutionary chromosome breakpoints. In addition, we revealed several incongruences between the current deer and bovine genome assemblies that were shown to be caused by errors in the CerEla1.0 assembly. Finally, we verified the centromere-to-centromere orientation of evolutionarily fused chromosomes in seven additional deer species, giving a support to previous studies on their chromosome evolution.

1. IntroductionThe family Cervidae (Ruminantia) groups more than fifty extant deer species, including species with growing economic importance. Deer species can be divided into three subfamilies: Cervinae, Capreolinae and Hydropotinae [1] and show a great karyotype diversity reflecting chromosome evolution of the taxon. The diploid chromosome numbers range from 2n = 6 in the female Indian muntjac (Muntiacus vaginalis) to 2n = 70 in several species of Capreolinae [2,3,4]. The 2n = 70 karyotypes of Hydropotes inermis and Mazama gouzoubira, involving 68 acrocentric autosomes, an acrocentric X and a small submetacentric Y, most probably represent an ancestral cervid karyotype [4] which evolved from the hypothetical ancestral pecoran karyotype (2n = 58) by six chromosome fissions [5]. Comparative cytogenetic studies revealing interspecies chromosome homologies and tracking of evolutionary karyotype rearrangements have been still scarce in Cervidae, with the exception of Muntjacini. The published studies were based mostly on standard banding methods [6,7,8] or on fluorescence in situ hybridisation (FISH) using whole chromosome painting probes [5,9,10,11,12,13]. The known data show that the most common mechanism of karyotype evolution in Cervidae is represented by Robertsonian (centric) fusions [4,7], whereas tandem fusions were described as the major evolutionary karyotype shaping factor in Muntiacini [9,11]. On the other hand, fissions of several ancestral pecoran chromosomes conserved in Bos taurus (BTA, 2n = 60) as BTA1, 2, 5, 6, 8, 9 and intrachromosomal rearrangements of the BTA1 orthologue and the X chromosome were also detected in Cervidae using bovine BAC (Bacterial Artificial Chromosome) probes [13,14]. However, the recent rapid development of high throughput molecular methods, namely whole genome sequencing, has brought new resources for comparative phylogenetic studies. At the level of chromosomes and their parts, an analysis of the next generation sequencing data can enable a precise determination of evolutionary chromosome breakpoints and allow a detection of small or intrachromosomal rearrangements that cannot be visualized by conventional cytogenetics or FISH with whole chromosome painting probes. In ruminants, cryptic interspecies chromosome differences as small as 3.3 Mb were identified in cattle and sheep using an in silico comparative bioinformatic approach [15]. This indicates that the use of sensitive methods can bring interesting discoveries even in seemingly well-described taxa.Unfortunately, this approach is only limited to species with completely sequenced and well-assembled genomes. Regarding Cervidae, whole genome assembly divided to chromosome-scale scaffolds (C-scaffolds) and including basic gene annotation is available only for the red deer (Cervus elaphus, CEL, 2n = 68) [16]. The CerEla 1.0 assembly available in the NCBI database has a total length of 3438.62 Mb and a total ungapped length 1960.83 Mb. It includes 406,637 contigs, 11,479 scaffolds and 35 chromosome-scale scaffolds (C-scaffolds) (https://www.ncbi.nlm.nih.gov/assembly/GCA_002197005.1/#/st accessed on 5 November 2020). The C-scaffolds in the CerEla1.0 genome assembly currently available in the NCBI database are arranged in accordance with the red deer genetic linkage map [17]. As a result, their order does not comply with the physical chromosome length and the chromosome order and centromere-telomere orientation in the red-deer karyotype. Generally, the use of other methods, i.e. BAC FISH mapping, is recommended to verify the newly established genome assemblies and physically anchor them to to chromosomes, thus upgrading them to a chromosome level [18,19,20,21]. In this study, we used comparisons with cattle (B. taurus), a closely related species used as model for comparative studies among Cetartiodactyla, with a range of available BACs and, above all, a well established whole genome sequence that served as a reference sequence for the CerEla1.0 assembly establishment [16]. We paired the 34 deer chromosome-scale scaffolds of the C. elaphus (CerEla1.0) genome assembly with bovine chromosomes by comparison of the gene annotation of the C. elaphus (CerEla1.0) and B. taurus (ARS-UCD1.2) assemblies available in the NCBI database. We selected bovine BACs for a construction of FISH probes that we used to anchor the CerEla1.0 C-scaffolds to C. elaphus karyotype, to compare the centromere-telomere orientation of the deer and bovine chromosomes and to analyse cervid evolutionary chromosome rearrangements. Using this approach, we revealed and corrected several incongruences between the CerEla1.0 and ARS-UCD1.2 genome assemblies, specified the orientation of the C. elaphus C-scaffolds and adjusted the predicted positions of evolutionary breakpoints characteristic for the cervid lineage. Using BAC-FISH mapping, we verified the breakpoints positions in a total of eight karyotypically different cervid species from subfamilies Cervinae and Capreolinae and specified the centromere-telomere orientation of their evolutionarily rearranged chromosomes.2. Materials and Methods2.1. Samples and Karyotype AnalysisSamples of whole peripheral blood of cattle (Bos taurus) and eight deer species including the red deer (C. elaphus) were obtained from captive born animals held in the Prague zoological garden and/or in deer enclosures in Bila Lhota and Frycovice (Czech Republic). The analysed species are listed in Table 1. Taxonomic nomenclature published by Groves and Grubb (2011) was used in this study [22]. Peripheral blood lymphocytes were cultured, harvested and fixed according to the previously described protocols [23]. Metaphase chromosome spreads for the karyotype and FISH analysis were prepared according to the procedures described previously [24]. GTG-banded karyotypes of B. taurus and C. elaphus were prepared using the standard trypsin/Giemsa method [25]. The karyotype of C. elaphus was arranged in accordance with the previously published deer karyotypes [13,26].2.2. Chromosome Orthology and Breakpoint Site PredictionOrthology between the red deer and bovine chromosomes was assessed by a comparison of B. taurus ARS-UCD1.2 and C. elaphus hippelaphus CerEla1.0 annotated genome assemblies available in the NCBI database (Accessed on 15 May 2020). Predicted locations of protein coding genes in the CerEla1.0 genome assembly (https://www.ncbi.nlm.nih.gov/genome/browse/#!/proteins/10790/321837%7CCervus%20elaphus%20hippelaphus/ accessed on 15 May 2020) were compared with positions of the corresponding genes in the bovine genome (https://www.ncbi.nlm.nih.gov/gene/advanced accessed on 15 May 2020). Briefly, we selected predicted protein coding genes separated by a distance of approximately 5 Mb along the length of the C. elaphus CerEla1.0 C-scaffolds and searched for their positions in the bovine ARS-UCD1.2 genome assembly. The 5 Mb distance was chosen to enable a reliable distinguishing of the mutual positions of BAC probes mapping to these regions when any incongruences would need to be solved by a dual colour BAC-FISH. To specify the breakpoints of the evolutionary chromosome fissions of ancestral chromosomes corresponding to bovine BTA1, 2, 5, 6, 8 and 9 [13,14], we predicted the putative ancestral breakpoint sites on the basis of the flanking gene positions in the deer and bovine genome assembly. The real positions of the evolutionary breakpoints were narrowed using a set of BAC probes and verified in all deer species available for this study.2.3. FISH ProbesBAC clones specific to proximal and distal chromosome regions, to regions flanking the predicted evolutionary breakpoint sites in Cervidae, and to regions showing incongruences between the deer and bovine genome assembly were selected from the CHORI-240 bovine BAC library (BACPAC Genomics, Emeryville, CA, USA) on the basis of their location along the bovine chromosomes in the ARS-UCD1.2 genome assembly. The chromosome positions of cervid evolutionary breakpoints were further narrowed using additional BAC clones located in neighbouring positions. The BAC clones used in this study are listed in Supplementary Tables S1–S3. The BAC DNA was isolated using Wizard Plus SV Minipreps DNA Purification System (Promega, Madison, WI, USA), labelled with Green-dUTP (Abbott, Abbott Park, IL, USA), biotin-16-dUTP (Roche, Mannheim, Germany), or digoxigenin-11-dUTP (Roche) using BioPrime Array CGH Genomic Labeling Module (Invitrogen, Carlsbad, CA, USA) and used for FISH.2.4. FISHA hybridization mixture containing 50% formamide, 2 × SSC, 10% dextran sulfate, 0.7 μg salmon sperm, 1.3 μg Bovine Hybloc DNA (Applied Genetics Laboratories, Melbourne, FL, USA) and 200 ng of the labeled DNA probe was prepared. Ten μL of the mixture were denatured at 75 °C for 10 min, preannealed at 37 °C for at least 30 min, and applied on slides with metaphase chromosomes denatured by 0.07 M NaOH as previously described [27]. After hybridization in a humid chamber at 37 °C overnight, the slides were washed in 0.7 × SSC at 72 °C for 2 min. The BAC probes labeled with biotin-16-dUTP or digoxigenin-11-dUTP were detected with Avidin-Cy3 (Amersham Pharmacia Biotech, Piscataway, NJ, USA), Streptavidine-Cy5 (Invitrogen/Molecular Probes, Camarillo, CA, USA) and antidigoxigenin-rhodamine (Roche) according to manufacturers’ instructions. If we used a combination of two probes labelled/detected by the same fluorochrome for the same chromosome, we performed two rounds of FISH, so that the position of each probe could be reliably determined. The slides were mounted in Vectashield mounting medium containing 1.5 mg DAPI (Vector Laboratories) and analysed using Zeiss Axio Imager.Z2 fluorescence microscope (Carl Zeiss Microimaging GmbH, Jena Germany) equipped with appropriate fluorescent filters and the Metafer Slide Scanning System (MetaSystems, Altlussheim, Germany). Images of well-spread metaphase cells were captured by CoolCube CCD camera (MetaSystems) and analysed using Isis3 software (MetaSystems). The reliability of the BAC probes was confirmed by their hybridization on bovine chromosomes prior to FISH in deer.3. ResultsComparing chromosomal positions of the predicted genes annotated to the CerEla1.0 C-scaffolds with their locations in the bovine ARS-UCD1.2 genome assembly, we assigned all red deer C-scaffolds to their bovine orthologues (Supplementary Table S4). Then we verified the deer-bovine chromosome orthology by BAC-FISH, which also enabled reliable physical anchoring of CerEla1.0 C-scaffolds to C. elaphus karyotype. Using BAC probes, we observed identical physical centromere-telomere orientation of orthologous red deer and bovine chromosomes. However, the orientation of CerEla1.0 C-scaffolds 2, 6, 8, 11, 12, 16, and 22 in the NCBI database was found to be reversed, and the deer chromosome CEL4 was found rearranged, when compared with the corresponding CerEla1.0 C-scaffold 19. The orthology between the G-banded red deer and cattle karyotypes are displayed in Figure 1 and Supplementary Figure S1. The relationships among the CerEla1.0 C-scaffolds and the red deer and cattle chromosomes are summarized in Table 2. The comparative FISH results in cattle and the red deer are documented in Figure 2. Karyotypes of the additional studied cervid species with indicated homologies with B. taurus are displayed in Supplementary Figure S2.Minor differences in the gene order revealed between CerEla1.0 C-scaffolds 4, 5, 6, 11, 12, 18, 19, 23, 27, 33 and X and their bovine orthologues in the ARS-UCD1.2 genome assembly are highlighted in Supplementary Table S4. We clarified the incongruences in eight of these regions on six red deer chromosomes, i.e., where the order of the BAC probes mapping to the incongruent regions was clearly visible. We observed an identical order of the BAC-FISH signals on cattle and red deer in all of the regions (Figure 3). Interesting results were obtained using the BAC probe CH240-134N9 targeted to the incongruent proximal region of the C-scaffold 11 corresponding to the distal part (82.9 Mb) of BTA11 according to CerEla1.0 and ARS-UCD1.2 comparisons, and to orthologous C. elaphus chromosome CEL9. Instead of BTA11 and CEL9, this probe hybridized to a distal part of other chromosome in both cattle and red deer. This chromosome was subsequently identified as BTA29, and CEL31, respectively, by FISH with the BAC probe CH240-384F12 specific to the proximal part (5.8 Mb) of BTA29 orthologous to CEL31 (Figure 3B).Regarding the X chromosome, we found that the sequences spanning 1–86 Mb of the CerEla1.0 X chromosome C-scaffold copy the gene order of the bovine X chromosome. However, a different order of the evolutionary conserved X chromosome segments was previously reported in studies using BAC-FISH in Cervidae [13,28]. Positions of the evolutionary chromosome breakpoints in chromosomes orthologous to BTA1, 2, 5, 6, 8 and 9 in the cervid ancestor were predicted on the basis of the genes located in the most proximal and distal positions of the corresponding CerEla1.0 C-scaffolds and, thus, flanking the assumed breakpoints. However, we revealed that the real breakpoints were located in a slightly different positions by a physical FISH-mapping with a series of BAC probes distributed along the chromosomes in the proximity of the predicted breakpoints (Figure 4 and Figure 5). The subsequent analysis of the breakpoint positions in additional cervid species showed similar results in all deer species analysed in this study (Supplementary Figure S3).The evolution of the BTA1 orthologue in the cervid lineage involved an initial fission followed by intrachromosomal rearrangements of one of the newly formed chromosomes. Two differentially rearranged types of the chromosome orthologous to the distal part of the BTA1 were observed in this study: An acrocentric chromosome common to Cervinae and C. capreolus and a submetacentric chromosome observed in the remaining Capreolinae in this study (R. tarandus, A. alces and O. virginianus) that was most probably derived from the previous by a pericentric inversion (Figure 5). Using CerEla1.0 and ARS-UCD1.2 genome assembly comparisons, the fusion site of the ancestral chromosomes corresponding to BTA17 and BTA19, which roughly represents the position of centromere, was found at 95 Mb of the deer C-scaffold 5 (CEL1) length. The evolutionary fission, giving rise to bovine separated BTA28 and BTA26, was located to 60 Mb of the CerEla1.0 C-scaffold 15 (CEL8).Finally, we used the bovine BAC probes to determine the centromere-telomere orientation of their evolutionarily fused chromosomes in seven additional deer species with rearranged karyotypes (Table 1). Except for the tandem fusion of BTA28;26 common to all Cervidae, the rearranged chromosomes were formed by evolutionary centric fusions in all studied species (Figure 6). 4. DiscussionThe recent publishing of the C. elaphus whole genome assembly (CerEla1.0) [16] brought a great resource for a research in the field of deer evolution, conservation and population genetics. However, the high automation in the genomic assembly construction may lead to errors. A verification and further improvements provided by molecular genetic and cytogenetic approaches are recommended for all newly established genome assemblies [18,19,20,21]. Inter- and intraspecies assembly comparisons supported by FISH enabled the detection and correction of misassembled sequences in genome assemblies of economically important bovid species (cattle, Bos taurus, sheep, Ovis aries and goat, Capra hircus) [29,30]. The combination of bioinformatic comparisons and BAC-FISH allowed identification of cryptic divergences between cattle and goat [15]. Using universal BAC sets, multiple scaffolds can be anchored to chromosomes of various species, as it was shown in birds [20].In this study, we focused on the verification of chromosome relationships among C. elaphus CerEla1.0 and B. taurus ARS-UCD1.2 genome assemblies and karyotypes of both species. Using bovine BAC probes, we physically anchored the CerEla1.0 C-scaffolds to C. elaphus and B. taurus karyotype (Figure 1). Similar approach exploiting BAC-FISH mapping technique was previously successfully used for an integration of cytogenetic landmarks or upgrading draft genome sequences to chromosomal level in other species [20,31]. The C-scaffolds of the CerEla1.0 genome assembly had been constructed according to the reference deer linkage map [17] and the well-established bovine (B. taurus) Btau_5.0.1 genome assembly [16]. The order, orientation and schematic length of the C-scaffolds in the NCBI database comply with the deer genetic linkage map [17] but do not correspond with their sequence length in Mb, nor the position of the chromosomes in the red deer karyotype [5,26,32].To document the results of this study, we arranged the G-banded red deer karyotype with regard to the chromosome morphology, physical lengths and G-banding patterns. Our G-banding and BAC-FISH showed concordant centromere-telomere orientation of the orthologous chromosomes in C. elaphus and B. taurus karyotypes. In compliance with the published paper on the CerEla1.0 assembly [16], we observed that the CerEla1.0 C-scaffolds 2, 6, 8, 11, 12, 16, 19 and 22 are presented in reversed centromere-telomere orientation in the NCBI database compared with the physical orientation of the red deer and bovine chromosomes.Comparing the gene order in the CerEla1.0 and ARS-UCD1.2 genome assembly, we observed differences in several CerEla1.0 C-scaffolds. Bana et al. [16] suggested that these red deer genomic regions represent inverted segments. We analysed eight of these regions by BAC-FISH and observed identical BAC probes order in the orthologous bovine and deer chromosomes in all studied regions (Figure 3). Nevertheless, we revealed that the BAC probe CH240-134N9, selected from the position 82.9 Mb of the BTA11 in ARS-UCD1.2 genome assembly, hybridised to a distal part of BTA29 and, correspondingly, to the BTA29 orthologue in the red deer (CEL31). Either the chromosome position of this BAC in the NCBI database is incorrect, or the region covered by this BAC probe in the bovine ARS-UCD1.2 genome assembly and probably the wider region at the start of the CerEla1.0 C-scaffold 11 showing several incongruences with ARS-UCD1.2 (Supplementary Table S4), actually represent sequences of the chromosome BTA29 and CEL31, respectively. The above-mentioned regions of the CerEla1.0 assembly need further thorough revision.In the published paper on the CerEla1.0 de novo genome assembly, the C-scaffold 33 was supposed to comprise sequences orthologous to parts of chromosomes BTA2 and BTA22 [16]. However, the bovine counterparts of all genes predicted to the CerEla1.0 C-scaffold 33 and selected for the CerEla1.0 - ARS-UCD1.2 comparisons in this study were found on BTA2. Our comparisons of the CerEla1.0 C-scaffold X with the bovine chromosome X in ARS-UCD1.2 showed that, despite several smaller discrepancies, the gene order on the CerEla1.0 C-scaffold X corresponds to that on the bovine X chromosome. However, it was previously published that cervid X chromosomes were shaped by complex evolutionary rearrangements, including neocetromere formation, that differentiated them to two distinct types characteristic for Cervinae and Capreolinae [13,28]. With regard to the previously published findings on the X chromosome structure in Cervidae [13,28], the first 86 Mb of the CerEla1.0 X chromosome C-scaffold need to be revised accordingly.Regarding another evolutionary chromosome changes, it is known that karyotypes of the current deer species derived from the pecoran ancestral karyotype (2n = 58) by fissions of six ancestral chromosomes orthologous to BTA1, 2, 5, 6, 8, 9 [12,13,14,17]. We used BAC probes selected on the basis of CerEla1.0 and ARS-UCD1.2 comparisons to hybridise to positions flanking the predicted evolutionary breakpoints, with the aim to physically verify the breakpoint sites. We revealed that the factual breakpoints differed from those predicted on the basis of CerEla1.0 C-scaffolds by up to 10 Mb, showing that the sequence span of the CerEla1.0 C-scaffolds needs to be properly adjusted. The newly assessed breakpoint locations were proved in all analysed species (four Cervinae and four Capreolinae) in this study.We also showed that the evolutionary history of the BTA1 orthologue in Cervidae was more complicated than a simple fission and involved also intrachromosomal rearrangements, as was previously suggested [13,16]. The actual evolutionary breakpoint sites on the ancestral BTA1 orthologue, approximated by the set of BAC probes used in this study, diverged from those predicted on the basis of the CerEla1.0 and ARS-UCD1.2 genome assembly comparisons, neither they corresponded to the schematic presentation of the B. taurus and C. elaphus chromosome differences shown in Bana et al. (2018). Using BAC probes at positions flanking the evolutionary breakpoints, we showed that the primary evolutionary fission of the ancestral chromosome orthologous to BTA1 occurred between 52 and 57 Mb of the BTA1 length. This led to the formation of two neochromosomes with different lengths. The smaller neochromosome orthologous to the proximal part of BTA1 corresponds to CEL16 and CerEla1.0 C-scaffold 31 is present in both Cervinae and Capreolinae. This indicates that this fission of the ancestral BTA1 orthologue together with fissions of BTA2, 5, 6, 8 and 9 orthologues probably represent a defining event of the karyotype evolution of Cervidae. The larger neochromosome orthologous to the distal part of BTA1 then underwent an intrachromosomal rearrangement with a breakpoint between 119 and 125 Mb of the BTA1 length in the common ancestor of C. capreolus and the current Cervinae. This rearrangement was followed by a pericentric inversion of the proximal part of the rearranged chromosome during a separate evolution of the lineage leading to R. tarandus, A. alces and O. virginianus (Figure 5). Because the BAC-FISH was proved to be an advantageous and sensitive tool for karyotype evolution studies [7,11,13,14,15,33,34,35,36], we used this method for verification of the evolutionary chromosomal rearrangements in Cervidae. The four species of Cervini analysed in this study share the fusion of BTA17;19 previously described on the basis of banding patterns and chromosome painting [7,13,14,37]. Using BAC probes, we proved that the ancestral chromosomes fused by their centromeres. Apart from the BTA17;19, five other centric fusions were proved in R. eldii, four in R. timorensis and one in C. albirostris by BAC-FISH in this study (Figure 6). As for Capreolini, the centric fusion BTA29;17 was confirmed in A. alces in this study. The chromosomes involved in the above mentioned fusions were previously identified by FISH with painting probes but their orientation in fused chromosomes could not be further specified by whole chromosome probes [7,13,14].In general, our analysis of chromosome evolution in the studied cervid species showed that centric fusions probably represented the main evolutionary mechanism shaping their karyotypes. In species analysed in this study, only the chromosome comprising BTA28;26 orthologues (CEL8) was shown to be formed by a tandem (centromere to telomere) fusion. The fact that the BTA28;26 fusion is common to all Cervidae and characteristic for all pecoran species excluding Bovidae [12,13,14,17] suggests that this chromosome probably represents an ancestral chromosome which underwent a fission at the origin of the Bovidae lineage [38]. Centric fusions are generally characteristic for the karyotype evolution in the family Bovidae [38,39]. However, in Cervidae, centric and tandem fusions dominate differentially in individual clades. In the subfamily Cervinae, centric fusions are relatively common in the tribe Cervini but the karyotypes of Muntjacini were diversified by extensive tandem fusions [4,5,11,33,38]. Among Capreolinae, presumed centric fusions occurred in the karyotype evolution of Ozotoceros bezoarticus, Blastocerus dichotomus and A. alces [3,4] (the latter one was proved in this study). On the other hand, both centric and tandem fusions were involved in the karyotype diversification of South-American Capreolinae species of the genus Mazama [40,41,42]. This suggests that the karyotype evolution has been driven by different mechanisms in the individual cervid lineages and cytogenetic studies employing BAC-FISH for the detailed differentiation of the evolutionary rearrangements can help in future studies focused on the reconstruction of the cervid phylogeny. 5. ConclusionsIn this study, we verified the red deer-cattle chromosome relationships, anchored the CerEla1.0 C-scaffolds to the red deer and cattle karyotype and proved the centromere-telomere orientation of the CerEla1.0 C-scaffolds. We indicated necessary adjustments to the CerEla1.0 genome assembly, including better specification of the sequence span of the chromosomes that underwent evolutionary chromosome fissions. Finally, we proved the location of the cervid evolutionary fissions and orientation of the fused chromosomes in a total of eight cervid species. Our results can serve as a basis for the CerEla1.0 genome assembly improvement, supporting, thus, future research in Cervidae.

animals : an open access journal from mdpi

10.3390/ani13061115

PMC10044701

Recently, increasing the efficiency of porcine embryo cultures by promoting oocyte maturation in vitro has attracted much attention. Brain-derived neurotrophic factor (BDNF) was beneficial to oocyte maturation and increased the developmental potential of porcine embryos. Although the effects of BDNF on porcine follicular development and the maturation of oocyte have been previously demonstrated, no literature was available, at the time of this work, relating to miRNA-regulated gene expression and signal pathways in mechanisms of BDNF, promoting porcine GCs proliferation. Therefore, this study explored the miRNAs involved in BDNF-induced proliferation of porcine GCs, as well as the involvement of the MAPK-ERK signaling pathway.

As a member of the neurotrophic family, brain-derived neurotrophic factor (BDNF) provides a key link in the physiological process of mammalian ovarian follicle development, in addition to its functions in the nervous system. The emphasis of this study lay in the impact of BDNF on the proliferation of porcine follicular granulosa cells (GCs) in vitro. BDNF and tyrosine kinase B (TrkB, receptor of BDNF) were detected in porcine follicular GCs. Additionally, cell viability significantly increased during the culture of porcine GCs with BDNF (100 ng/mL) in vitro. However, BDNF knockdown in GCs decreased cell viability and S-phase cells proportion—and BDNF simultaneously regulated the expression of genes linked with cell proliferation (CCND1, p21 and Bcl2) and apoptosis (Bax). Then, the results of the receptor blocking experiment showed that BDNF promoted GC proliferation via TrkB. The high-throughput sequencing showed that BDNF also regulated the expression profiles of miRNAs in GCs. The differential expression profiles were obtained by miRNA sequencing after BDNF (100 ng/mL) treatment with GCs. The sequencing results showed that, after BDNF treatment, 72 significant differentially-expressed miRNAs were detected—5 of which were related to cell process and proliferation signaling pathways confirmed by RT-PCR. Furthermore, studies showed that BDNF promoted GCs’ proliferation by increasing the expression of CCND1, downregulating miR-127 and activating the ERK1/2 signal pathway. Moreover, BDNF indirectly activated the ERK1/2 signal pathway by downregulating miR-127. In conclusion, BDNF promoted porcine GC proliferation by increasing CCND1 expression, downregulating miR-127 and stimulating the MAPK-ERK1/2 signaling cascade.

1. IntroductionOvarian follicle development is a fundamental process of reproductive physiology in female mammals. Germ stem cells differentiate into oogonia in the genital ridge and the oogonia divide and become primary oocytes via mitosis. The primordial follicles are formed during the process of differentiation. Thus, primordial follicle assembly is crucial for the acquisition of fertility during female mammalian reproduction [1]. Primordial follicles are composed of primary cells and surrounding GCs and basement membrane [2]. The proliferation of GCs can effectively promote follicle development and the formation of the antral cavity. As such, improving the effectiveness of in vitro porcine embryo production by promoting the maturation of oocytes has become a research hotspot.Neurotrophins (NTs) are soluble polypeptides known for their effects in regulation of synaptic growth, revival, differentiation and functionalization in the nervous system [3,4,5]. The NTs not only affect nervous system development, but also participate in ovarian development [6,7]. During embryogenesis in pigs, various NTs play important roles in folliculogenesis [8,9], oocyte maturation [10,11], steroid synthesis [9] and embryonic development [11]. Brain-derived neurotrophic factor (BDNF) is a well-known member of the NT group and may also increase oocyte maturation and porcine embryonic growing potential in vitro [10,11]. In adult mammalian ovaries, BDNF expression, and its high-affinity TrkB receptor, have both been detected [12]. Moreover, the expression of BDNF differed in porcine Germinalvesic oocytes and in vitro mature MII oocytes, implying that BDNF might be involved in porcine oocyte development and maturation [13]. Previous data from our laboratory showed that BDNF promoted GC growth in cocultured bovine early embryos [14] and regulated GC proliferation and progesterone synthesis via the TrkB-AKT signaling pathway [15]. Additionally, BDNF-TrkB signaling was involved in oogenesis, follicle recruitment, germ cell survival and oocyte nuclear and cytoplasmic maturation [16,17,18,19]. Thus, BDNF was shown to regulate the reproductive function of mammals along with TrkB.As an important epigenetic regulatory mechanism, microRNAs (miRNAs) inhibit target genes’ expressions by binding to their 3′UTR region and are related to the key regulation of many biological processes. Indeed, miRNAs play critical roles in spermatogenesis [20,21], follicle development [22,23,24], maturation of oocytes [25,26] and primary embryonic growth [26,27] during mammalian reproduction. Regulation of miRNAs with BDNF involvement has been demonstrated in the nervous system [28,29,30]. In the ovary, BDNF affects oocyte–cumulus cell interaction and granulosa cell function by regulating miRNAs. For example, miR-NA-10b inhibits goat GC proliferation by targeting BDNF [31]. Moreover, BDNF promotes the expansion of porcine cumulus–oocyte complex by downregulating miR-NA-205 [11]. miR-127, as a tumor suppressor, plays a role in inhibiting tumor cell growth and migration in various types of cancer [32,33,34,35]. In mammalian reproductive systems, miR-127 inhibits the progression of ovarian cancer by regulating MAPK4 expression [36]. In addition, miR-127 affects placental development by targeting Rtl1 [37,38,39]. Notably, miR-127 was differentially-expressed in follicular fluid of polycystic ovary syndrome (PCOS) patients, compared with normal controls [40]. Therefore, the influence and role of miR-127 in the ovarian follicle is worthy of attention.Although the impacts of BDNF on follicular development and maturation of oocytes in pigs have been demonstrated, there has been, to the best of our knowledge, no report on miRNA-regulated gene expression and signaling pathways related to BDNF mechanisms promoting porcine GC proliferation. Therefore, this study explored the mechanishms by which miRNAs were regulated by BDNF and MAPK-ERK1/2 signaling pathways in porcine GCs proliferation.2. Materials and Methods2.1. Cell Isolation and CulturePorcine ovaries, from prepubertal gilts, were gathered at a local slaughter abattoir and kept at 37 °C in sterile saline, with penicillin and streptomycin at concentrations of 100 IU/mL. The ovaries were transported within 2 h to the laboratory. GCs were isolated and collected from the follicular fluid, as previously stated [15]. Concisely, the collected ovaries were rinsed three times with water to eliminate blood stains, then placed in sterile normal saline after being washed three times with 75% ethanol for disinfection. Follicular fluid was aspirated between 3 mm and 5 mm follicles using a 10 mL sterile syringe, and about 8–10 mL was acquired from twenty porcine ovaries. The collected follicular fluid was transported into 20 mL PBS and centrifuged at 300× g for 1 min to eliminate the oocytes. The isolated supernatant was transferred into a 50 mL sterile centrifuge tube, followed by centrifugation at 500× g for 6 min to separate GCs. The trypan blue exclusion test was performed to examine the viability of GCS. Approximately 1 × 106 GCs per well were seeded in 6-well plates within 2 mL culture medium (Dulbecco’s Modified Eagle Media/Nutrient Mixture F12 (DMEM/F12)), comprising of 10% fetal bovine serum (FBS) for 24 h at 37 °C with 5% CO2.2.2. Immunofluorescence StainingFresh porcine follicle granulosa cells were cultured in 24-well (2 × 105 cells/well) optical bottom plates for 24 hand then triple-washed with PBS (5 min each). Cells were preserved for 20 min in precooled absolute methanol and then permeabilized with 0.1% Triton X-100 for 30 min. Cells were blocked with 1% goat serum at ambient temperatures for 30 min, followed by incubation at 4 °C overnight in the following primary antibodies: anti-FSHR (BS2618, Bioworld, Nanjing, China, 1:200 for rabbit anti-mouse), anti-BDNF (BS6533, Bioworld, 1:200 for rabbit anti-mouse), or anti-TrkB (BS94070, Bioworld, 1:100 for rabbit anti-mouse). They were then rinsed three times with PBS. After that, the cells were incubated in goat anti-rabbit immunoglobulin G (IgG) combined with fluorescein (FITC) antibody (BS10950, Bioworld, 1:1000) at ambient temperatures for 1 h and triple-washed with PBS. After that, cells were stained for 10 min using PI solution (1:1000; Beyotime, Shanghai, China) and observed with a fluorescence microscope (IX71; Olympus, Tokyo, Japan). The porcine GCs were presented in two colors: green for FSHR and red for nucleus.2.3. Cell TreatmentGCs were cultured in 6-well plates, with 106 cells per well, until cell density reached 70% and then washed twice with PBS to eliminate the suspended cells. The culture media were replaced with serum-free DMEM/F12 for 4 h, and then cells were treated with BDNF, K252α (TrkB-specific inhibitor; 100 ng/mL; Invitrogen, Carlsbad, CA, USA) or PD98059 (MAPK suppressor; 15 μM; Beyotime). BDNF (0, 10, 20, 50, 100 and 200 ng/mL; PeproTech, London, UK)-treated GCs were used, for different times (12, 24, or 48 h). K252α (100 ng/mL) were used for 24 h, PD98059 (15 μM), for 30 min.2.4. Cell TransfectionThe siRNA for BDNF (si-BDNF sequence: 5′-GCGGTTCATAAGGATAGAC-3′) was synthesized from RiboBio (Guangzhou, China). Transfection of porcine GCs with 50 nM siRNA was accomplished using FuGENE® HD (Roche, Basel, Switzerland) reagent for transfection and Opti-MEM medium (Gibco, Paisley, Scotland, UK), as directed by the manufacturing company. GCs were transfected with siRNA for 24 h to extract RNA and protein.The mimics and inhibitors of miRNAs (miR-185, miR-1273, miR-7047, miR-127 and miR-532) were acquired from Genepharma (Shanghai, China). Based on the manufacturer’s guidelines, mimics of the five miRNAs or miR-127 inhibitors were transfected into the GCs using FuGENE® HD transfection reagent for 24 h. The transfection concentration was 50 nM.2.5. Cell Viability AssayThe capacity of GCs’ viability was evaluated by CCK-8 kit (Dojindo, Shanghai, China). Briefly, 100 μL of medium, containing 1 × 104 GCs, were grown in 96-well plates. After treatment (BDNF, K252α or PD98059) or transfection (miRNAs or si-BDNF), CCK8 reagent was added into medium for 10 μL of each well. After 2 h of incubation, the absorbance was detected at 450 nm with a microplate reader (BioTek, Winooski, VT, USA).2.6. RNA Extraction and Reverse Transcripton-Quantitative Polymerase Chain Reaction (RT-qPCR)Total RNA was extracted from GCs using Trizol reagent (Invitrogen, Carlsbad, CA, USA) and reverse transcription was done using a PrimeScript RT Reagent Kit (Takara, Tokyo, Japan) based on manufacturer instructions. qRT-PCR was performed with a SYBR Green Master Mix (Takara, Tokyo, Japan) and real-time quantitative fluorescence PCR instrument (Mx3005P; Agilent, Santa Clara, CA, USA). The reaction criteria were as follows for temperature profile: denaturation at 95 °C for 15 min, then amplification for 40 cycles of 35 s at 95 °C and 40 s at 60 °C. Using GAPDH or U6 as reference genes, the relative mRNA expressions of indicated genes were calculated using 2−ΔΔCT methods. All sequences of primer synthesis were completed by Comate Bioscience Co., Ltd. (Changchun, China) as shown in Supplementary Table S1.2.7. Western BlottingCell lysis and protein collection were done using a RIPA buffer (Beyotime) containing phosphatase inhibitor cocktail A (Beyotime). The concentrations of proteins were determined using a BCA Kit (Beyotime). Then, 12% SDS-PAGE gel electrophoresis was used to separate a total of 20 μg protein in each sample and the separated proteins were transferred to polyvinylidene difluoride (PVDF) membranes (Millipore Co., Ltd., Norwood, OH, USA). The unspecific bands of membranes were blocked with 5% skim milk at ambient temperature for one hour. Subsequently, the membranes were stored at 4 °C overnight with the primary antibodies listed below: anti-BDNF (BS6533, Bioworld, 1:1000, Rabbit anti-mouse), anti-CCND1 (ab16663, 1:1000), anti-P21 (ab109520, 1:500, Rabbit anti-human), anti-BCL2 (ab182858, 1:50), anti-Bax (ab32503, 1:50), anti-ERK1/2 (ab184699, 1:100), anti-p-ERK1/2 (ab278538, 1:50) or anti-beta-actin (ab179467, 1:80). After rinsing with TBST (4 × 8 min), peroxidase (HRP), coupled with secondary antibody (1:800), was used to incubate the membranes for 1 h. The bands were visualized using a BeyoECL Plus kit (Beyotime) and tested using a chemiluminescent detector (Tanon, Shanghai, China), based on the manufacturer’s instructions. Unless otherwise noted, all antibodies were rabbit anti-mouse and purchased from Abcam (Cambridge, MA, USA). Intensity of bands were analyzed via grey scanning using the Tanon Gel Imaging System (Tanon).2.8. Flow Cytometric Analysis (FACS)FACS was used to measure the cell cycles of the porcine GCs. GCs were digested by trypsin and the collected cells were twice-rinsed with precooled PBS, then fixed and refrigerated in 70% ethanol for more than 24 h. The fixed cells were resuspended in 500 μL staining buffers containing PI (50 μg/mL) and RNaseA (1 mg/mL) after rinsing with PBS. Subsequently, after 30 min incubation in the dark, BD-LSR flow cytometry (BD Biosciences, Franklin Lakes, NJ, USA) was performed to determine cell cycle kinetics.2.9. Sequencing of miRNAsGCs were collected after treatment with BDNF (100 ng/mL) for 24 h. The library operation and sequencing experiments were performed following standard Illumina procedures. The Small RNAs Sample Pre Kit (Illumina, San Diego, CA, USA) was used to structure libraries of small RNA sequencing and Illumina HiseQ2000/2500 was performed for sequencing. A sequence length of 1 × 50 bp was collected for bioinformatics analysis.2.10. Bioinformatics AnalysisThe Illumina HiSeq 4000 sequencing platform (Illumina, Inc., San Diego, CA, USA) was used to sequence RNA from pig GC specimens and construct cDNA libraries. Denatured libraries were converted into single-stranded DNA molecules and sequenced over 51 cycles on Illumina HiSeq, based on the manufacturer’s directions. The analyses were performed with the help of Beijing Yuanyi Gene Technology Co., Ltd. (Beijing, China). After sequencing, Solexa CHASTITY [41] quality filtered reads were harvested as clean reads. Then, clean reads were screened for siRNA through BLAST and Rfam databases and classified siRNA were annotated or predicted as mature miRNA. The miRBase database was used for sequence alignment of mature miRNA. We calculated miRNA expression using the most readily available isoforms; miRBase was used to identify the mature miRNAs and all miRNA isoforms (5p or 3p). When miRNA profiles were differentially-expressed, comparisons between the two groups, fold changes, p-values and FDRs were calculated and used to identify significantly differentially-expressed miRNAs. log2 (fold change) > 1 or log2 (fold change) < −1 were selected for differentially-expressed miRNAs. The differences were statistically significant (p < 0.05) after R package comparison.2.11. Data AnalysisStatistical analyses were performed with SPSS (Windows version 19). Data analysis between the two groups was done using the unpaired t-test. Three or more groups of data were evaluated for significant differences using ANOVA. Statistical significance was shown at * p < 0.05 and ** p < 0.01. Data were shown as mean ± standard deviation (SD).3. Results3.1. Effects of BDNF on the Proliferation of Porcine GCsThe effects of BDNF on GC proliferation were determined. Figure 1A revealed that BDNF increased porcine GCs’ viability in a dose-dependent manner (p < 0.05) and a concentration of 100 ng/mL was the most effective after 24 h. Unless otherwise noted, all BDNF described in our results were treated at 100 ng/mL for 24 h.In addition, BDNF knockdown (Figure 1B) in GCs significantly decreased GC viability (Figure 1C). The proportion of S-phase cells in the si-BDNF group (7.55 ± 0.1%) was considerably reduced, compared with the si-NC group (9.47 ± 0.42%) (Figure 1D). The levels of proliferation-related genes’ (including CCND1, p21 and Bcl2) expression and apoptosis-related genes (Bax) in porcine GCs cells were evaluated. As shown in Figure 2E, after BDNF knockdown, the mRNA and protein expression levels of CCND1, p21 and Bcl2 were significantly reduced. On the contrary, the expression levels of Bax were considerably elevated (Figure 1E).3.2. The BDNF/TrkB Pathway Affects the Proliferation of Porcine GCsThe effects of BDNF on GC proliferation after K252α treatment were assessed. The results showed that the viability of porcine GCs (Figure 2A) and the distribution of cells in S-phase (Figure 2B) were significantly decreased with K252α treatment alone. Moreover, K252α reduced BDNF-induced porcine GC viability, and the proportion of cells in S-phase, when K252a and BDNF treatment were undertaken together in GCs (Figure 2A,B).The expression levels of genes connected with cell proliferation and apoptosis were measured. BDNF significantly increased mRNA (Figure 2C-a) and protein (Figure 2C-b,C-c) expression levels for CCND1, p21 and Bcl2. Additionally, the Bax mRNA and protein expression levels were reduced dramatically. Figure 2C shows that BDNF, in combination with K252a, weakened the effects of BDNF on the expression of the above genes in GCs.3.3. BDNF Promotes GCs Proliferation through Increase of CCND1 by Downregulating miR-127miRNA sequencing was performed. According to the sequencing results, 72 differentially-expressed miRNAs were detected, including 34 upregulated and 38 downregulated, as shown in Figure 3A-a,A-b (fold change > 2, p < 0.05). Additionally, 5 miRNAs (miR-185, miR-1273, miR-7047, miR-127 and miR-532) associated with cellular processes (Figure 3A-c) and cell proliferation signaling pathways (Figure 3A-d) were significantly downregulated in GCs after BDNF treatment (Figure 3B).Subsequently, to ascertain the consequences of these miRNAs on GC proliferation, GCs were transfected with the above 5 miRNA mimics for 48 h, respectively. Overexpression of miR-185 and miR-532 inhibited GC viability, while overexpression of miR-127 significantly reduced the viability of GCs and the proportion of cells in S-phase simultaneously (Figure 3C). Therefore, we selected miR-127 as the key miRNA in BDNF-induced GC proliferation to continue our research.BDNF significantly decreased miR-127 expression (Figure 3B) and increased CCND1 expression (Figure 2C). To confirm the regulatory relationship between miR-127 and CCND1, the mRNA expression relationship was determined. As shown in Figure 3C, miR-127 inhibited the expression of CCND1. The protein expression levels of CCND1 were compatible with RT- qPCR after miR-127 was overexpressed or inhibited in porcine GCs (Figure 3D,E).3.4. BDNF Promotes GC Proliferation via the ERK1/2 Signaling Pathway Mediated by miR-127To determine whether the MAPK-ERK1/2 pathway was entangled in proliferation by porcine GCs after BDNF treatment, the phosphorylation and total protein levels of ERK1/2 in porcine GCs were assessed. The results indicated the phosphorylation levels of ERK1/2 increased after BDNF supplementation for 15 (p < 0.05), 30 (p < 0.01) and 60 min (p < 0.05; Figure 4A). Moreover, GCs were treated with BDNF in combination with PD98059; the effects of BDNF on the viability of GCs (Figure 4B-a) and the distribution of cells in S-phase (Figure 4B-b) were diminished by a blockade of the MAPK-ERK1/2 pathway. Secondly, the expression levels of CCND1 were assessed. Inhibition of ERK1/2 decreased the expression levels of CCND1 mRNAs and proteins (Figure 4B-c–B-e). Thirdly, MAPK-ERK1/2 suppression had no impact on the expression levels of miR-127 (Figure 4B-f), but the phosphorylation of ERK1/2 was stimulated after transfection of miR-127 inhibitor with GCs. Then, after treatment with BDNF, the expression level of p-ERK1/2 was further increased (Figure 4C).4. DiscussionIn ovaries, BDNF is mainly presented in follicular granulosa cells and oocytes [16], while BDNF and its receptor TrkB are mainly expressed in granulosa cells and membrane cells of porcine follicles [42]. In this study, the expression levels of BDNF and trkB in GCs were verified by immunofluorescence (Supplementary Figure S1). BDNF regulates early follicle development in various mammals and directly affects ovulation, as shown in previous studies [10,17]. BDNF has also been associated with different causes of infertility during in vitro fertilization [12,43,44,45]. Additionally, BDNF regulates the maturation of cytoplasmic and nuclear in porcine oocytes by paracrine and/or autocrine signaling systems and promotes potential embryo growth following in vitro fertilization and somatic cell nuclear transfer [10]. Thus, elucidating the transcriptional mechanisms regulated by BDNF in GCs will be beneficial for improving the IVM of porcine oocytes.Transfection of specific siRNA interfering with BDNF expression in cells will be beneficial to studies of the role of BDNF in the proliferation of porcine GCs. With the extension of transfection time, the interference efficiency of siRNA was enhanced, but further inhibition of GCs proliferation had no significant effect. These findings suggested that the continuous decreases in the secretion of BDNF in porcine GCs did not inhibit cell proliferation at all times.miRNAs are small noncoding RNAs. They play a vital role in hormone-induced ovarian development [46,47]. In previous studies, the primary role of miR-127 was described as a tumor inhibitor, involved in a series of cellular processes—for instance proliferation, senescence, migration and invasion [48,49,50]. Moreover, miR-127 mediated the differentiation of mouse embryonic endoderms and promoted placental development [38,51]. However, the potential functions of miR-127 in porcine reproduction remain ambiguous. To our knowledge, this study revealed the expression of miR-127 in porcine GCs for the first time, and demonstrated that miR-127 was involved in porcine granulosa cell proliferation as a negative regulator.Cell proliferation and differentiation are regulated by the progression of the cell cycle. This progression is controlled by cyclins and Cdks complexes [52]. Cyclin D1 (CCND1), a member of the cyclin family, promotes the transition of cells from G1 phase to S phase by interacting with CDK4 and CDK6 [53,54,55,56,57]. CCND1 is also a key regulator of cell proliferation [58]. Due to the inhibition of miR-127 on the proportion of cells in S-phase, the regulatory relationship between miR-127 and CCND1 was examined. The overexpression and knockdown of miR-127 showed that miR-127 significantly negatively regulated CCND1. It also revealed the regulatory pathway, in which BDNF downregulated miR-127 to promote CCND1 expression during porcine GC cell proliferation.MAPK-mediated signal transduction is a key factor affecting cell fate processes [59]. As the core module of the MAPK signal cascade, the MAPK-ERK1/2 signaling pathway is highly conservative [60]. The extracellular signals are transmitted into the nucleus by ERK1/2 and trigger modifications in the expression of certain proteins in cells [61]. MAPK-ERK12 is also connected with several cellular activities, comprising differentiation, proliferation, apoptosis, transcription and adhesion [62,63,64,65,66,67]. Hence, it was necessary to examine this intracellular signaling pathway to further substantiate the molecular mechanisms underlying BDNF-induced cell proliferation and increased CCND1 expression in GCs. According to our data, phosphorylation of ERK1/2 was increased by BDNF, and porcine GC proliferation was inhibited by blockades of TrkB or ERK1/2. Thus, we concluded that BDNF-induced cell proliferation also depended on MAPK-ERK1/2 signaling activation and that BDNF promoted GC proliferation by regulating CCND1 through MAPK-ERK1/2 signaling cascade. A study on the promotion of bovine GC proliferation by the BDNF-MAPK-ERK1/2 signaling pathway was previously reported [15]. Notably, however, miRNAs (especially miR-127) were found to be involved in the transduction of the BDNF-MAPK-ERK1/2 signaling pathway in this study. Downregulating miR-127 facilitated the stimulation of the MAPK-ERK1/2 signaling pathway. This result suggested that CCND1 upregulation by BDNF indirectly actuated the MAPK-ERK1/2 pathway by downregulating miR-127. These findings provided novel insights for future studies on the function of BDNF and revealed the molecular mechanisms underlying mammalian follicle development. In addition, these findings could provide a new target for the treatment of follicular dysplasia, such as PCOS.5. ConclusionsIn conclusion, we determined that BDNF stimulated GC proliferation by increasing CCND1 expression through miR-127 downregulation and MAPK-ERK1/2 pathway activation (Figure 5). Notably, BDNF stimulated the MAPK-ERK1/2 pathway directly and indirectly by downregulating miR-127. Our current findings provided insights that could aid in efforts to effectively promote porcine oocyte maturation and improve embryo production efficiency in vitro.

animals : an open access journal from mdpi

10.3390/ani12010052

PMC8749932

Transport stress (TS) can impact the physiology and psychology of broilers, and this can be an important factor affecting liver iron metabolism in broilers. By establishing a transport model group, broilers (n = 144) reared under the same conditions were allocated into six groups and transported duration for 0, 0.5, 1, 2, 4, and 6 h. The results showed that the enrichment of iron content in the liver was the highest at a transport duration of 4 h, so the effect of transport duration of 4 h on iron metabolism was further investigated using TMT quantitative proteomic analysis. It was found that TS caused the enrichment of iron ions in the liver, TMT identified FTH1, IREB2, and HEPH as key proteins affecting iron metabolism, and key genes regulating iron homeostasis were validated using RT-PCR.

Abnormal iron metabolism can cause oxidative stress in broilers, and transport stress (TS) may potentially influence iron metabolism. However, the mechanisms by which TS affects iron metabolism are unclear. This study used quantitative proteome analysis based on tandem mass tag (TMT) to investigate the effects of TS on liver iron metabolism in broilers. Broilers (n = 24) reared under the same conditions were selected randomly into the transported group for 4 h (T2) and non-transported group (T1). Results showed that the serum iron level and total iron-binding capacity of broilers in the T2 were significantly higher than those in the T1 (p < 0.05). The liver iron content of broilers in the T2 (0.498 ± 0.058 mg·gprot−1) was significantly higher than that in the T1 (0.357 ± 0.035 mg·gprot−1), and the iron-stained sections showed that TS caused the enrichment of iron in the liver. We identified 1139 differentially expressed proteins (DEPs). Twelve DEPs associated with iron metabolism were identified, of which eight were up-regulated, and four were down-regulated in T2 compared with T1. Prediction of the protein interaction network for DEPs showed that FTH1, IREB2, and HEPH play vital roles in this network. The results provide new insights into the effects of TS on broilers’ liver iron metabolism.

1. IntroductionThe pre-slaughter transport process is an important integral part of poultry management [1]. Pre-slaughtering transport of market-age broilers from their geographically dispersed farms is an unavoidable common practice [2]. During transit, chickens are exposed to numerous potential stressors, including handling, feed withdrawal, noise, vibration, thermal extremes, social disruption, crowding, and restriction of movement [3], which may lead to undesirable changes in animal welfare and immunity. The transport stress (TS) responses comprise mainly autonomic responses via activation of the autonomic nervous system (ANS) mediated by adrenaline and noradrenaline including increased respiration and heart rate, elevated body temperature, and promotion of energy utilisation from body reserves [4], accelerating glycogenolysis and suppressing energy storage [5]. In addition, birds are more sensitive to temperature than other monogastric animals due to feather coverage and the absence of sweat glands [6]. TS also increases the concentration of circulating corticosterone hormone, via activation of the hypothalamic–pituitary–adrenal (HPA) axis, which has a significant impact on the hepatic glycogen, protein and lipid metabolism, and meat quality [7].The metabolic consequences of TS are mainly reflected in metabolic acidosis and oxidative stress, both of which lead to cytotoxicity, a free radical-mediated chain reaction [8]. The main source of ROS in tissues is the leakage of electrons in the mitochondrial respiratory chain during the conversion of molecular oxygen to water [9]. In addition, iron ions can promote the production of hydroxyl radicals. Evidence suggests that ROS were closely associated with iron because it could diffuse freely over cellular membranes and then interact with ferrous iron (known as Fenton reaction), which catalyzes a Fenton-type reaction to produce more reactive oxygen radicals [10]. Therefore, the condition of iron during pre-slaughter transport might be an important factor in the health of broilers and the quality of chicken meat. Iron, one of the essential trace elements for the growth and development of poultry, is utilized by most living cells and organisms for essential biochemical functions, such as oxygen transfer, DNA, RNA and protein synthesis, electron transfer, cellular respiration, cell proliferation and differentiation [11,12]. Iron is also involved in the function of catalases and peroxidases that protect the cells against the formation of free radicals, but iron concentration must be finely regulated because any excess of free iron is rapidly toxic. After all, iron is a transition metal with divalent and trivalent oxidation numbers and exerts its toxicity by catalyzing ROS generation. The most toxic hydroxyl radical (•OH) is produced in large quantities in the Fenton reaction. •OH may cause cell injury by inducing damage to the lysosomal, cytoplasmic, nuclear, and mitochondrial membranes, apoptosis through activation of the caspase cascade, and hyperoxidation of fatty acid chains [13]. Modern medical research has found that an imbalance in iron homeostasis can induce metabolic disorders and several diseases in the body, but this process is not limited to physiological processes [14]. Psychological stress can even activate the thalamic–pituitary–adrenal (HPA) axis system, causing an increase in serum adrenocorticotropic and adrenocorticotropic hormone levels and a decrease of serum iron levels, hepatic iron enrichment, and the development of iron overload [15,16].At the individual level, the liver maintains iron homeostasis by balancing iron supply with iron utilization and losses. The liver orchestrates systemic iron balance by producing and secreting hepcidin, inducing degradation of the iron exporter ferroportin, which regulates the import of iron from the bloodstream to the liver [17]. The liver has been shown to be a hub for iron regulation, but the effect TS on hepatic iron metabolism is unknown. So, does TS affect changes in chicken liver iron content? If so, how is this iron being transported? What proteins are involved after the iron homeostasis is disrupted? The process involves a range of proteins, biological processes, and pathways. Due to the complexity of iron metabolism, information on proteins is needed to fully understand the biological processes and pathways. The quantitative proteomics approaches based on full tandem mass tag (TMT) have been used to study animal metabolic processes [18,19]. The analysis of this resource has enabled us to examine the mechanisms of protein diversification, thereby expanding knowledge of the complexity of TS-induced iron metabolism.2. Materials and Methods2.1. Experimental Design, Chickens, Management, and Transport ConditionsThis study was approved by the Animal Care and Use Committee, Ningxia University. One-day-old male Chinese white-feathered broilers were obtained from a commercial hatchery and were reared until 72 d of age under normal conditions with free access to water and feed in the poultry farm of the Ningxia Haoshuichuan Breeding Co., Ltd. (Guyuan, China). Transport model group allocation: Seventy-two day-old Chinese, white-feathered broilers with consistent body weights (2.50 ± 0.10 kg) were randomly allocated to six treatment groups (n = 24 per group) with transport duration set at 0, 0.5, 1, 2, 4, and 6 h, respectively. Every 4 broilers were transported in a single crate with the dimensions of 52 × 43 × 26 cm (length × width × height). Fasting for 10 h, with access to water prior to transport, no drinking water was provided during transport. The truck was an opened truck designed for shipment of broilers, and the 36 crates were put randomly on the truck (4.8 m × 1.46 m × 0.95 m), with 3 rows × 3 columns × 4 layers and no apparent gaps among rows and layers. The pre-slaughter handling included catching the chickens, weighing, loading into containers, transport, unloading, weighing after the transport and slaughter. The 0 h broilers were slaughtered directly, and the rest were slaughtered after completing the transport time separately. Transport was undertaken on a ring road (approximately 78.81 km·lap−1) near the city of Yinchuan, and the average speed was 60 km·h−1. The truck’s average relative humidity and temperature were 34 to 45% and 30.1 to 35.6 °C, respectively. Distribution of broilers in T1 and T2 groups: Seventy-two-day-old Chinese, white-feathered broilers with consistent weights (2.50 ± 0.10 kg) were randomly allocated to two treatment groups (T1 and T2, n = 24 per group). T1 and T2 groups were set up slightly differently from the transport model group. The non-transport group (T1) were captured, loaded, and left in the transport vehicle for 4 h, and broilers in the transport group (T2) were captured, loaded, and transported for 4 h. At the end of transport, groups T1 and T2 broilers were slaughtered together. The pre-slaughter handling of the broilers, the transport vehicle, the speed of the vehicle, and the environment outside the vehicle were kept consistent for T1 and T2 with the transport model group.2.2. Slaughter and Sample CollectionAfter transportation, blood samples were immediately collected from the jugular vein and put into 5 mL EDTA-evacuated tubes, which were then centrifuged at 2500× g for 15 min at 4 °C, frozen in liquid nitrogen, and plasma samples were stored at −80 °C until analysis. Broilers were slaughtered using electrical stunning and exsanguination from the jugular vein and de-feathered. After dissection, the livers were rapidly cleaned with PBS, and fresh samples were placed in 4% paraformaldehyde fixative for haematoxylin-eosin (HE) and Prussian blue (PB) analysis, placed in liquid nitrogen, and stored at −80 °C for proteomic and biochemical analysis [20].2.3. Biochemical Indices of Plasma and LiverSerum concentrations of lactate dehydrogenase, glucose, serum iron, and liver iron were measured using an Chemray 800 automated biochemistry analyzer (Rayto Life and Analytical Sciences Co., Ltd., Shenzhen, China) using colorimetric methods and following the instructions of the manufacturer of the corresponding reagent kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). The concentrations of cortisol, corticosterone, glutathione peroxidase, and adrenocorticotropic hormone were determined with a commercial ELISA kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturers’ instructions.2.4. Determination of Iron Metabolism-Related Indicators in BloodA fully automated biochemical analyser (Myriad Biomedical Electronics Co., Shenzhen, China) was used to determine the serum’s transferrin-bound iron and unsaturated iron binding capacity. The total iron-binding capacity and transferrin saturation were then calculated, total iron binding capacity was expressed as the sum of serum iron content and unsaturated iron binding capacity, and transferrin saturation was expressed as serum iron content divided by total iron binding capacity, multiplied by 10,000.2.5. Haematoxylin-Eosin (HE) and Prussian Blue (PB) StainingThe liver samples were fixed with 4% paraformaldehyde at room temperature. Fixed tissues were dehydrated in 30% sucrose (v/v), paraffin-embedded, and sectioned (10 μm) with a sliding microtome (Leica Mikrosysteme Vertrieb GmbH, Solms, Germany). Sections were stained with hematoxylin or Prussian blue solution, dehydrated with ethanol and xylene, sealed with neutral gum, and then examined under a light microscope (Olympus Corporation, New York, NY, USA) [21,22].2.6. ROS Fluorescent StainingChicken livers were frozen, sectioned, and dried at room temperature (20 ± 1.5 °C). The sections were washed three times with PBS buffer (pH 7.4, 0.01 M) for 5 min each, followed by DAPI staining for 10 min in the dark to stain the nuclei, and the sections were then sealed with an anti-fluorescence quencher. The slices were observed under a fluorescence microscope, and images were collected (DAPI UV excitation at 330–380 nm, emission at 420 nm, blue light; FITC excitation at 465–495 nm, emission at 515–555 nm, green light; CY3 excitation at 510–560 nm, emission at 590 nm, red light) (Front-mounted Fluorescence Microscope Eclipse C1, Nikon Co., Tokyo, Japan).2.7. TMT Proteomics Analysis2.7.1. Sample PreparationThe chicken liver tissues were ground in liquid nitrogen. Hence, 400 µL of SDT lysate buffer (4% SDS, 100 mM DTT, 150 mM Tris-HCl pH 8.0) was added to each sample, ultrasonicated for 2 min in an ice bath, and centrifuged at 16,000× g for 20 min at 4 °C. The supernatant was collected and quantified with a BCA Protein Assay Kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).2.7.2. Protein DigestionDigestion of protein (300 μg for each sample) was performed according to the FASP procedure [23,24]. Briefly, 200 µL of UA buffer (8 M Urea, 150 mM Tris-HCl, pH 8.0) were added to the protein sample, mixed well, followed by repeat ultrafiltration (Microcon units 30 kD, 12,000× g for 15 min) facilitated by centrifugation. Then, 100 μL 0.05 M iodoacetamide in UA buffer was added to block reduced cysteine residues, and the samples were incubated for 20 min in the dark. The filter was washed with 100 μL UA buffer three times and then 100 μL 25 mM NH4HCO3 twice. Finally, the protein suspension was digested with 4 μg trypsin (Promega, Madison, WI, USA) in 40 μL 25 mM NH4HCO3 overnight at 37 °C, and the resulting peptides were collected as a filtrate. The peptides were desalted using a desalting spin column (Thermo Fisher Scientific, Waltham, MA, USA) for quantification.2.7.3. TMT Labeling of PeptidesBriefly, 100 μg of the peptide was taken from each sample and labeled according to TMT labeling kit instructions (Thermo Fisher Scientific, Waltham, MA, USA). Each aliquot (100 μg of peptide equivalent) was reacted with one tube of TMT reagent. After the sample was dissolved in 100 μL of 0.05 M TEAB solution, pH 8.5, the TMT reagent was dissolved in 41 μL of anhydrous acetonitrile. The mixture was incubated at room temperature for 1 h. Then, 8 μL of 5% hydroxylamine were added to the sample and incubated for 15 min to quench the reaction. The multiplex labeled samples were pooled together and lyophilized. The peptides of each fraction were dried and solubilized with 0.1% fomic acid for LC-MS analysis.2.7.4. LC-MS AnalysisLC- MS analysis was performed on a Q Exactive mass spectrometer coupled to Easy nLC 1200 (Thermo Fisher Scientific, Waltham, MA, USA). Peptide from each fraction was loaded onto the C18-reversed-phase column (12 cm long, 75 μm ID, 3 μm) in buffer A (0.1% formic acid) and separated with a linear gradient of buffer B (95% acetonitrile) at a flow rate of 300 nL/min over 90 min. The linear gradient was set as follows: 0–2 min, linear gradient from 2% to 8% buffer B; 2–42 min, linear gradient from 8% to 28% buffer B; 42–47 min, linear gradient from 28% to 40% buffer B; 47–52 min, linear gradient from 40% to 100% buffer B; 52–60 min, buffer B maintained at 100%. The peptides were separated and analyzed by DDA (Data Dependent Acquisition) mass spectrometry using a Q-Exactive HF-X mass spectrometer (Thermo Fisher Scientific, Waltham, MA, USA). The analysis time was 60 min, detection mode: positive ion, parent ion scan range: 350–1800 m/z, primary mass resolution: 60,000 @m/z 200, AGC target: 3e6, primary Maximum IT: 50 ms. The peptide secondary mass spectra were acquired according to the following method: each full scan triggered the acquisition of secondary mass spectra (MS2 scan) of 20 highest intensity parent ions, Secondary mass resolution: 15,000 @m/z 200, AGC target: 1e5, Secondary Maximum IT: 50 ms, MS2 Activation Type: HCD (Higher energy collisioninduced dissociation), Isolation window: 1.2m/z, and Normalized collision energy: 1.2 m/z. 2.7.5. Database Searching and AnalysisThe resulting LC-MS/MS raw data were imported into Maxquant (version 1.6.0.16, Thermo Fisher Scientific, Waltham, MA, USA) for data interpretation and protein search against the database Uni-prot_Hordeum-vulgare_201747–20180125 (downloaded on 25 January 2018, including 201,747 protein sequences), which was sourced from the protein database at https://www.uniprot.org/uniprot/?query=Hordeum-vulgare&sort=score (accessed on 22 December 2021). An initial search was set at a precursor mass window of 10 ppm. The search followed an enzymatic cleavage rule of Trypsin/P and allowed maximal two missed cleavage sites and a mass tolerance of 20 ppm for fragment ions. The modification set was as follows: fixed modification: Carbamidomethyl (C), TMT10plex(K), TMT10plex(N-term); variable modification: Oxidation(M) and Acetyl (Protein N-term). The minimum 6 amino acids for peptide, ≥1 unique peptides were required per protein. For peptide and protein identification, false discovery rate (FDR) was set to 1%. TMT reporter ion intensity were used for quantification. The database used for the search was Uni-prot_Gallus_gallus_34995_202108.fasta from the URL https://www.uniprot.org/taxonomy/9031/ (accessed on 22 December 2021) Protein Data Bank.2.7.6. Bioinformatics AnalysisNCBI BLAST + client software (https://www.ncbi.nlm.nih.gov/, accessed on 22 December 2021) and UniProtKB/Swiss-Prot (https://www.expasy.org/resources/uniprotkb-swiss-prot, accessed on 22 December 2021) were used to search the sequences of proteins. Differentially expressed proteins were screened with the cutoff of a ratio fold-change of >1.20 or <0.83 and p-values < 0.05 was carried out with Perseus software (Max-Planck-Institute of Biochemistry—Computational, Systems, Biochemistry) and Excel (Microsoft office 2019, Redmond, DC, USA) statistical computing software. Gene ontology (GO) terms were annotated using the software Blast2GO (BioBam Bioinformatics S.L., Valencia, Spain). After annotations, the proteins were blasted against the online Kyoto Encyclopedia of Genes and Genomes (KEGG) database (http://geneontology.org/, accessed on 22 December 2021). Hierarchical clustering analysis was conducted via R 4.1.2 (The University of Auckland, Auckland, New Zealand). GO and KEGG enrichment analyses were carried out with the Fisher’s exact test, and FDR correction for multiple testing was also performed. Enriched GO and Kegg pathways were nominally statistically significant at the p < 0.05 level. Protein-protein interactions were analyzed by string (http://string-db.org/, accessed on 22 December 2021) against the Sus scrofa database and considering a medium confidence score of 0.7 for interactions. 2.8. Analysis of Gene Expression by Real-Time Quantitative PCR (RT-PCR)Real-time PCR amplification was performed using an CFX fluorescent quantitative PCR instrument (Bio-Rad Laboratories, Hercules, CA, USA) in a total volume of 20 μL. The primer sequences are shown in Table S1. Actin was used as a reference gene. The PCR reaction mixture contained 4 μL 5× reaction buffer, 0.5 μL oligo (dT)18 primer (100 μM), 0.5 μL random hexamer primer (100 μM), 1 μL servicebio RT enzyme mix, and 110 μL total RNA. PCR amplification was performed using the initial heating for denaturation at 95 °C for 10 min, followed by 40 cycles at 95 °C for 15 s and 60 °C for 30 s. Then, the melting curve was analyzed at the end of every program [25]. Each PCR reaction was performed in triplicate. The 2−ΔΔCT method was used to calculate each gene’s relative expression level [26].2.9. Data Processing and Statistical AnalysisResults are presented as the means ± standard deviations (SD). All graphs were completed using Origin 2021 (Origin Lab Corporation, Northampton, MA, USA). Data were analyzed with one-way ANOVA using SPSS 25 (SPSS Inc., Chicago, IL, USA). A p < 0.05 was considered to be a statistically significant difference between the means.3. Results3.1. Changes in Blood Biochemistry and Liver Iron Content of Broilers in the Transport Model GroupAs shown in Table 1, as the duration of transport increased in broilers of the transport model group, the concentrations of cortisol, corticosterone, adrenocorticotropic hormone, lactate dehydrogenase, glutathione peroxidase, glucose, and serum iron were increased (p < 0.05), indicating that the duration of transport affected the degree of stress, energy metabolism, and oxidative status in the broilers. It is interesting to note that the transport duration of 4 h resulted in the highest enrichment of iron in the liver. Therefore, further analysis of the metabolic status of iron ions in broilers at 4 h of transport was conducted. 3.2. Changes in Iron Metabolism-Related Indicators in Blood and Tissue IronAs shown in Table 2, serum iron content and total iron-binding capacity in T2 were significantly higher than T1, unsaturated iron binding capacity in T2 was significantly lower than T1, and transferrin saturation in T2 tended to be higher than T1 (p < 0.05). The liver iron content in T2 was significantly higher than T1 (p < 0.05). 3.3. HE and PB StainingAs shown in Figure 1, the liver cells in both T1 and T2 were clearly visible, well-arranged, and dense, with no pathological features, such as cell rupture and histolysis. Prussian blue staining caused the cells and tissues to appear pink and iron to appear blue, and it was noteworthy that the number and size of blue spots were higher in T2 than in T1 (Figure 1). This is consistent with the results shown in Table 2.3.4. ROS Fluorescent StainingWe assessed the effect of TS on ROS production in the liver (Figure 2). In contrast to the fluorescence imaging findings, where the ROS always accompanied the cell’s nucleus, the liver ROS fluorescence intensity in the T2 was significantly higher than in the T1.3.5. Overview of the Chicken Liver Proteomic AnalysisChicken liver samples were used for protein extraction, followed by SDS-PAGE gel electrophoresis (1D gel). On the 1D pattern, the protein bands in different groups were similar (Figure 3a). Intra-group reproducibility was analyzed according to Pearson correlation, with R-values all greater than 0.8 (Figure 3b), indicating the reproducibility of the experiment. For comparison between T1 and T2 chicken liver samples, a protein exhibiting a fold change of >1.2 or <0.83 and a p-value of <0.05 was regarded as a differentially expressed protein (DEPs). Based on the two criteria, 1139 DEPs were identified. Moreover, the DEPs of each group were analyzed and displayed in the form of a hierarchical clustering heat map (Figure 3c, Table S2). 3.6. Annotation and Functional Enrichment of DEPsTo gain insights into the biological functions of DEPs, we performed GO functional enrichment analysis. The biological processes, cellular components, and molecular functions GO terms associated with DEPs are shown in Figure 4a, and the detailed data are shown in Table S3. These terms were associated with cell adhesion, biological adhesion, protein folding, supramolecular fiber organization, extracellular matrix, external encapsulating structure, protein-containing complex binding, actin filament binding, structural molecule activity, and calcium ion binding. KEGG enrichment analysis showed that DEPs were significantly enriched in the following functional categories: Phagosome, Focal adhesion, Ecm–receptor interaction, Cell adhesion molecules (CAMS), Ribosome, Proteasome, Glvcolvsis/Gluconeogenesis, and Pentose phosphate pathway (Figure 4b, Table S3).3.7. Liver Iron Metabolism-Related Proteins and Protein-Protein Interactions (PPI) NetworkTo further analyze the effect of TS on hepatic iron metabolism, 12 DEPs associated with iron metabolism were identified, eight of which were up-regulated, and four down-regulated in T2 compared with T1 (Figure 5a, Table S4). Physical and functional protein-protein interactions (PPI) networks were constructed for the screened DEPs. Prediction of the protein interaction network for DEPs showed that FTH1, IREB2, and HEPH play vital roles in this network. Iron-responsive element-binding protein 2 (IREB2) had the highest number of interactions (Figure 5b).3.8. Expression of Iron Homeostasis-Related Genes in the Liver of BroilersWe further investigated the expression of the genes associated with the regulation of iron homeostasis (Figure 6). Eight crucial genes related to iron homeostasis were analyzed. The mRNA expression levels of SLC40A1, TFRC, FECH, FTL, ACO1, IREB2, and HEPH in the T2 were significantly higher than those in the T1 (p < 0.05). The expression levels of FTL, TFRC, IREB2, and HEPH between T2 and T1 were up-regulated more than three-fold.4. DiscussionBroilers can provide abundant, cheap, and nutritious animal proteins for human consumption [27]. The transport process of broilers before slaughter activates the autonomic nervous system (ANS), and the hypothalamic–pituitary–adrenal (HPA) responses of the body to promote corticosterone hormone secretion and accelerate anaerobic glycolysis. In this study, the levels of cortisol, corticosterone, and adrenocorticotropic hormone in the T2 were significantly higher than the T1 (Table 1). This stressful process could lead to an imbalance in iron homeostasis, increased ROS, and tumor necrosis factor (TNF) levels, and induce the apoptotic process [28,29]. There was evidence that these can interact with each other. For example, TNF increased labile iron level and subsequently promoted the production of mitochondrial ROS [8]. Serum iron levels were measured in T1 and T2 to determine if disorders in iron metabolism were present. Tissue iron content reflected the degree of iron enrichment in cells or tissues. Serum total iron-binding capacity and unsaturated iron binding capacity were indirect measures of transferrin concentration. The serum iron metabolism index showed a significant increase in iron content in the blood of broilers in T2 (p < 0.05) (Table 2), and the binding of transferrin was enhanced. The liver iron content in T2 was significantly higher than T1 (p < 0.05) (Table 2), which may be because hepatocytes are the primary iron storage cells in the body, and the liver is an essential organ in regulating iron homeostasis. Several studies have found that the transport process caused the release of iron ions, leading to an increase in serum iron level [30], which was consistent with the findings of this study.Iron homeostasis in the cells was regulated by balancing iron uptake with intracellular storage and utilization, and iron metabolism was visualized by mapping the pathways of iron ion transport (Figure 7). The process of iron uptake by the cell mainly involves the binding of transferrin receptors (TFRC) on the cell membrane to ferroportin (FPN), followed by endocytosis to transport iron into the cell or through divalent metal transporter 1 (DMT1) on the cell membrane to transport iron across the membrane into the cell [12,14]. In addition, the solute carrier family 40 protein (SLC40A1) regulates the ferrous iron transmembrane transporter activity and iron ion transmembrane transporter activity in the body. In the study, the expression of SLC40A1 was significantly higher in the T2 than T1 (Table S4, Figure 6) (p < 0.05). Hephaestin (HEPH), as a ferroxidase for Fe2+ to Fe3+ conversion, may be implicated in iron homeostasis and may mediate iron efflux associated with FPN. At the cellular level, FPN is regulated by the iron regulatory proteins (IRPs), iron-responsive element (IREB2), and cytoplasmic aconitate hydratase (ACO1). Once the iron entered the cell, it entered a hypothetical low-molecular-weight pool, otherwise known as the chelatable iron pool. At the same time, a large amount of ferrochelatase (FECH) was needed to chelate iron and keep the chelatable iron pool in a stable state [8,31]. In the study, the protein and mRNA expression levels of IREB2, TFR, and FECH in the T2 were significantly higher than those in T1 (Table S4, Figure 6), indicating that the ferric ion transport activity of FPN may be activated. The liver might regulate iron uptake or release, and there might be an imbalance in iron homeostasis. The elevated serum iron level led to the binding of ferroportin to FPN, which caused FPN to enter the cell, degrade it, and finally decrease the iron excretion. This is achieved predominantly at the level of protein synthesis (translation of mRNA into protein) rather than at the transcription level (mRNA synthesis). In the study, serum iron level in T2 was significantly higher than T1 (Table 2). The liver needs to absorb and transport iron ions to ensure stable serum iron levels. Iron transport requires the involvement of TRF, FPN, ferritin (light polypeptide), and ferritin heavy chain 1. Therefore, the expression of these genes was enhanced, which is consistent with our TMT quantitative proteomics and RT-PCR results.Under normal physiological conditions, about one-third of transferrin is saturated with iron [31]. The buffering capacity of excess apo-transferrin ensures that each iron ion that enters the circulation remains shielded, and redox-active Fe2+ in the form of the labile iron pool (LIP) is maintained at low concentrations to sustain metabolic needs. In contrast, the excess is sequestered in proteins, including ferritin, to avert toxic repercussions [32]. However, when the organism is under physical or psychological stress, increased iron flux in the blood leads to gradual saturation of transferrin and accumulation of non-transferrin-bound iron. This is readily taken up by parenchymal tissue cells. The absorption process in the liver cells then causes an overload of iron in the liver. Under oxidative stress conditions, high levels of superoxide could induce Fe2+ release from iron compounds, including [4Fe-4S] cluster, heme, and ferritin, and cause iron-dependent accumulation of ROS. Interestingly, HEPH might function as a ferroxidase for Fe2+ to Fe3+ conversion [8,11]. The mRNA expression level of HEPH in the T2 was 3.61 times higher than in the T1, and the protein expression level was also up-regulated (Table S4, Figure 6). Iron could contribute to the ROS pool in the cell through the Fenton reaction in which Fe catalyzes the breakdown of H2O2 to yield hydroxyl radicals [19]. Fe2++H2O2→Fe3++OH−+OH·Imbalance in the ROS generation and clearance rate may lead to oxidative stress and the consequent production of free radicals. Hydroxyl radicals are the most reactive free radical species and they may react with a wide range of cellular constituents including amino acid residues and attack membrane lipids to initiate a free radical chain reaction known as lipid peroxidation, leading to histopathological damage to hepatocytes [31]. In the study, liver iron content in the T2 was significantly higher than in the T1 (p < 0.05) (Table 2, Figure 2), which was consistent with the result of ROS (Figure 2). Histomorphological observations of the liver did not reveal significant histopathological damage to hepatocytes (Figure 1), which might be because the degree of imbalance in iron homeostasis was at a primary stage and caused little cellular damage. Transport could cause the excessive production and accumulation of ROS, ultimately result in oxidative stress [33]. We have verified by TMT proteomics and RT-PCR that iron may be a potential maker of ROS. In the present study, oxidative damage to cells during pre-slaughter transport of broilers could not be self-repaired because these chickens would be slaughtered without sufficient time to repair the oxidatively damaged cells. It has been suggested that ROS may affect meat quality by interfering with collagen turnover, and/or may cause the deterioration of meat by lipid peroxidation and protein oxidation directly through blood transport [34] and may even result in PSE meat [35]. Therefore, these DEPs in the current study might help reveal the genetic mechanism of transport stress-mediated imbalance of iron homeostasis using TMT quantitative proteomics strategies. Proteins are the executors of physiological functions and participate in the specific pathway to complete their biological functions comprehensively rather than independently performing their functions [36]. Moreover, some strategies are offered. If the composition of the basal diet of broilers is changed a week or more before slaughter, some natural phenolic-rich plant ingredients could be added to improve iron-induced oxidative stress by using the chelating properties of polyphenols with iron. The metabolic processes of iron in poultry farming should be taken into account, e.g., the excess of iron in the composition of the basal diet and the recovery process of newly hatched chicks after transfer, because these chicks are less able to adapt and recover from their environment.5. ConclusionsIn summary, the findings of this study suggest that iron metabolism was disturbed during pre-slaughter transport in broilers, especially in the liver. The imbalance in iron homeostasis induced the production of cellular ROS. We identified FTH1, IREB2, and HEPH as key proteins that regulate iron metabolism by TMT quantitative proteomics. This study contributes to understanding the complex biological processes controlling the imbalance of iron homeostasis mediated by TS and provides new insights into improving the adverse effects of TS in broilers.

animals : an open access journal from mdpi

10.3390/ani12030387

PMC8833775

The European eel is a species with high commercial value for aquaculture, and it has suffered a drastic reduction of its natural stocks during the last decades; thus, breeding in captivity is nowadays considered essential to avoid the extinction of the species. In this sense, a new method to maturate the European eel (males and females) for eel aquaculture has been studied. For the first time in the European eel, a controlled release hormone system (osmotic pumps) has been tested, which allowed to induce the testis maturation and sperm production in part of the males as well as a very early maturation and ovulation in the females.

The European eel (Anguilla anguilla) is a commercially valued species for aquaculture. Over the past decades, it has experienced a drastic reduction in its natural stocks. Thus, breeding in captivity is considered essential, nowadays, to guarantee the eel aquaculture and to reduce pressure on natural populations. Traditionally, the European eel has been sexually matured by means of weekly hormonal injections, which cause stress to the fish. The purpose of this research study was to assess the use of osmotic pumps as a new method to induce sexual maturation in male and female European eels, without the weekly injection. The control groups were treated with weekly hormone injections (recombinant human chorionic gonadotropin for males and carp pituitary extract for females), and the implanted groups were treated with osmotic pumps (ALZET® osmotic pumps) loaded with the respective hormones. Regarding male European eels, this study shows that the use of controlled release systems was able to induce the maturation and spermiation, but without the necessary capacity to produce enough gametes with acceptable quality parameters that could meet the needs of a commercial eel hatchery. Concerning female European eels, the study demonstrates that the use of osmotic pumps loaded with CPE became an effective method, generating early maturations (4 to 10 weeks) in 50% of the females, so this method could become a viable alternative for eel hatchery procedures.

1. IntroductionThe European eel (Anguilla anguilla) is a commercially valued species, especially for the Japanese and European markets. Nevertheless, it is not yet possible to breed European eels in captivity, and current aquaculture production consists of fattening wild-caught glass eels. Over the past 50 years, populations of the European eel have been declining, and the recruitment of glass eels has collapsed since the early 1960s [1]. Overfishing, habitat reduction, pollution, and the swimbladder parasite (Anguillicola crassus) are some of the causes that have led to the decline of the European eel [2]. As a result, it was included in the Red List of the International Union for Conservation of Nature (IUCN) as a “Critically Endangered” species [3], and measures for the recovery of the stock have been established by the European Union (Regulation 1100/2007, 18 September 2007) [4].Nowadays, controlling the reproduction and ability to produce glass eels seems to be the only sustainable solution to reduce the pressure on natural populations. However, the complexity of its reproductive physiology makes it difficult to increase the supply of glass eels for aquaculture. The European eel is a catadromous fish with a very complex life cycle, which includes two transoceanic migrations. In order to overcome the lack of natural spawning stimuli in captivity, the sexual maturation of both males and females must be induced with long-term hormonal treatments [4,5,6].Regarding male European eels, human chorionic gonadotropin (hCG) has been the most widely used hormone for reaching spermiation, but it has been administered to the animals in different formats [7,8]. Currently, the use of weekly injections of recombinant human chorionic gonadotropin (hCGrec) is considered an effective alternative and showed better results than hCG in terms of sperm quantity and quality [9,10].With regard to female European eels, the traditional method consists of weekly injections with carp (CPE) or salmon (SPE) pituitary extracts at around 15 to 20 weeks [5,7]. When the oocytes reach the nuclear migration stage, a final treatment is administered using a dose of the maturation-inducing steroid (DHP) [11,12]. There are several studies on fertilization and hatching trials in the European eel, but the fact is that the individual response of females to treatments is highly variable and spawning percentage, egg quality, or larval hatching is affected [13,14].Despite the positive results obtained with the traditional hormonal treatments to induce the sexual maturation of the eels in both sexes, these are time-consuming processes, and the weekly injections of hormones require repetitive handling of the broodstock, causing stress, and susceptibility to the diseases [15,16]. In this sense, the study of hormonal pretreatments [17,18], new hormonal treatments, and hormone delivery systems that reduce the level of stress in fish during the induction of the sexual maturation, have been carried out to optimise the protocols for this species. Therefore, over the past years, several gonadotropin-releasing hormone delivery systems have been assessed, including osmotic pumps [19,20]. The mechanism of operation of the osmotic pumps is based on the osmotic difference between the inside of the pump (osmotic layer) and its environment (animal tissue). These delivery devices cause the sexual maturation of the fish by steadily releasing hormones, reducing stress in the fish by avoiding the weekly injections. In this way, the osmotic pumps have been used in several species of teleosts, but the effectiveness was quite variable (Chanos chanos: [21]; Lates calcarifer: [22]; Clarias batrachus and C. gariepinus: [23]). The use of osmotic pumps to induce sexual maturation in eels was studied for the first time by Kagawa et al. [24]. It was shown that the implantation of an osmotic pump loaded with hCG stimulates spermatogenesis and spermiation in male Japanese eels (Anguilla japonica). Subsequently, the osmotic pumps were used to induce the maturation of captive female Japanese eels. Pumps were loaded with SPE, hCG, and GnRHa, and SPE had better results compared with the other hormones [25].Therefore, the aim of the present study was to evaluate these hormone delivery systems as an alternative method to induce the sexual maturation of male and female European eels. The production and quality of gametes from both males and females of the different maturation methods were recorded.2. Materials and Methods2.1. Fish HandlingThirty male eels (mean body weight = 126.7 ± 17.9 g) from the fish farm Valenciana de Acuicultura, S.A. (Puzol, València; Spain) and twenty-one female eels (mean body weight = 771 ± 123.8 g) caught by local fishermen in the Albufera Lagoon (Valencia) during their migration to sea, were moved to our facilities in the Aquaculture Laboratory of the Universitat Politècnica de València, Spain. Each individual eel was tagged with passive integrated transponders (PIT tags) in order to identify the eels. Male eels were distributed into three 150-L aquaria equipped with separate recirculation systems and thermostats/coolers. Female eels were distributed into four 500-L aquaria equipped with IRTAmar® technology [26]. The fish were gradually acclimatised for 1 week from freshwater to sea water (salinity 37 ± 3 g/L). The fish were fasted during the experiment and the aquaria were covered to maintain constant shade (photoperiod: 0L:24D).This study was carried out in strict accordance with the recommendations given in the Guide for the Care and Use of Laboratory Animals of the Spanish Royal Decree 53/2013 regarding the protection of animals used for scientific purposes (BOE 2013). The protocol was approved by the Experimental Animal Ethics Committee from the Universitat Politècnica de València (UPV) and final permission (2019/VSC/PEA/0034) was given by the local government for managing endangered fish species (Generalitat Valenciana).2.2. Males2.2.1. Induction of Sexual Maturation: Experimental DesignThe male eels were divided into three experimental groups: control, OP-100, and OP-200 (n = 10 per group). Over 15 weeks, the control group was weekly anesthetised with benzocaine (60 ppm) and received an intraperitoneal injection of hCGrec (Ovitrelle®, 1.5 IU/g fish). In parallel, the eels from OP-100 and OP-200 groups were anesthetised with benzocaine (60 ppm) and an osmotic pump (ALZET® OsmoticPumps [27]; Figure 1A) was implanted into the peritoneal cavity of each eel after cutting the abdomen wall with a scalpel approximately 0.5 and 2 cm, respectively (Figure 1B). Finally, the wound was sutured (Figure 1C). The control group fish underwent the same surgery and a sterile object of similar size to the osmotic pump was introduced into their intraperitoneal cavities.The OP-100 group of males received the osmotic pump ALZET®-1004 (diameter = 6 mm, length = 15 mm, reservoir volume = 100 μL), while the OP-200 group of fish received the osmotic pump ALZET®-2006 (diameter = 7 mm, length = 30 mm, reservoir volume = 200 μL). To load the osmotic pumps, hCGrec was diluted in a saline solution (NaCl 0.9%) at a concentration of 13 IU/μL. In accordance with the instructions of the manufacturer, the ALZET®-1004 and ALZET®-2006 osmotic pumps can, respectively, release 0.05 and 0.07 μL of the solution per hour for approximately 10 weeks when the fish are maintained at a water temperature of 21 °C. These data were considered to calculate the volume of diluted hormone contained in each type of pump to guarantee that the three groups received a similar amount. After 10 weeks, the osmotic pumps stopped releasing hormone and the eels from the OP-100 and OP-200 groups received weekly intraperitoneal injections of hCGrec (1.5 IU/g fish) for 5 weeks to complete the treatment.2.2.2. Sperm Collection and SamplingBefore sperm collection, the eels were anesthetised with benzocaine (60 ppm). Sperm samples were collected weekly through the application of abdominal pressure 24 h after the administration of the hormone in the control group (following the protocol described by Pérez et al. [28]), taking special care to avoid contamination with faeces, urine, and sea water. Sperm volume was measured using graduated tubes and sperm density was determined by computer-assisted sperm analysis (CASA) systems (see the next section). After that, samples were diluted 1:24 (sperm:extender) in P1-medium [29] and kept in plastic tubes at 4 °C until the sperm kinetic analysis, which was carried out in the 2 h following sperm collection.2.2.3. Sperm Motility AssessmentSamples were activated by mixing 0.5 μL of P1-diluted sperm with 4.5 μL of artificial sea water (Aqua Medic, Meersalz, 37 g/L, with 2% BSA (w/v), pH adjusted to 8.2). All the motility analyses were performed in triplicate using the module of ISAS ® v1 (Proiser R+D, S.L.; Paterna, València, Spain), which uses CASA technology to evaluate the sperm. Video sequences of 0.5 s were recorded (at 60 fps) using a video camera (Nikon Digital Sight DS-5M) mounted on a phase-contrast microscope (Nikon Eclipse 80i) with a 10× objective lens. The chamber used throughout all the analysis was a SpermTrack-10® (Proiser, Paterna, València, Spain) with a 10× negative contrast phase lens on a Nikon Eclipse (E-400) microscope.The parameters considered in this study were density, defined as the number of spermatozoa/mL; total motility (MOT, %), defined as the percentage of motile cells after sea water activation; and progressive motility (pMOT, %), defined as the percentage of spermatozoa which swim forwards in 80% of a straight line. Spermatozoa were considered motile if their straight-line velocity (VSL) was >10 μm/s.2.2.4. Biometric ParametersAt the end of the 15-week experimental period, all males were euthanised and weighed. In order to evaluate the progression of maturation, some biometric parameters such as gonadosomatic index [GSI = (gonad weight/total body weight) ∗ 100] and pectoral fin colour (Black, Dark grey, Light grey, Transparent; [30]) were evaluated.2.3. Females2.3.1. Induction of Sexual Maturation: Experimental DesignThe female eels were divided into two experimental groups: the control group (n = 11) and the OP-2ML4 group (n = 10). The control group was treated with weekly intraperitoneal injections with carp pituitary extract (20 mg/kg; CPE, Catvis, Ltd., The Netherlands) for 15 to 25 weeks (or until ovulation happened). The CPE was diluted in a saline solution (NaCl 0.9%) to obtain a concentration of 0.1 g CPE/mL. The mix was centrifuged (3000 rpm, 10 min.) and the supernatant was stored at −20 °C until use. Once a week, the eels were anesthetised with benzocaine (60 ppm) and weighed before the injections to calculate the individual CPE dose [13]. In the OP-2ML4 group, osmotic pumps Alzet-OP-2ML4 (diameter = 14 mm, length = 51 mm, reservoir volume = 2000 μL) were used as the hormone delivery system. The osmotic pumps were implanted in a similar way to that previously described for the males (Section 2.2.1.), by cutting the abdomen wall with a scalpel approximately 0.5 cm. The control group fish underwent the same surgery and a sterile object of similar size to the osmotic pump was introduced into their intraperitoneal cavities.The osmotic pumps were loaded with the supernatant of CPE. In accordance with the instructions of the manufacturer, the used osmotic pumps can release 1.05 μL of the solution per hour for approximately 10 weeks when the fish are maintained at a water temperature of 20 °C. The volume of diluted CPE contained by the pumps was calculated to contain a similar amount of hormone than the received by the control group. After 10 weeks, the osmotic pumps stopped releasing hormone and the eels from the OP-2ML4 group received weekly intraperitoneal injections of CPE (20 mg/kg fish) for 15 weeks to complete the treatment.2.3.2. Sexual Maturation Assessment and Spawning InductionThe females were monitored on a weekly basis and the eels that seemed to be sexually mature (based on body weight increase and enlarged abdomen [11]) were moved to a separated tank to follow the maturation evolution over the following days. If the abdomen was quite large, a sample of oocytes was collected by intraovarian cannulation (cannula diameter = 1.6 mm). The oocytes were diluted in saline solution (NaCl 0.9%) and observed using the camera (Nikon Digital Sight DS-5M) mounted on the binocular loupe (Leica MZ16F). If the oocytes were in the 3rd or 4th phase of nucleus migration [11], the female was led to the final maturation stage, following the protocol described by Butts et al. [12]. Consequently, females received a final CPE injection (prime dose 20 mg/kg fish) and 24 h later, if the oocytes were in the 5th phase of development, the females were treated with 17α,20β-dihydroxy-4-pregnen-3-one (DHP) at 2 mg/kg fish. Over the following 12 h, the females were monitored until the spawning occurred. The ripe females were isolated in a separated tank and anaesthetised with benzocaine (60 ppm) and the genital area was cleaned with freshwater before abdominal stripping. The eggs were collected into laboratory trays and weighed to calculate the amount of sperm to be added for in vitro fertilization, using a sperm to egg ratio of a minimum of 1:25,000, according to Butts et al. [12].2.3.3. Biometric Parameters: Gonadosomatic Index (GSI)At the end of the 25-week experimental period, the females that did not ovulate during the experiment were euthanised and weighed. To evaluate the progression of maturation, the gonadosomatic index was calculated for females that did not ovulate.2.3.4. Fertilization, Hatching Rate, and Embryonic DevelopmentTwo hours before spawning, sperm from three or four males was collected (total motility >70%) and its quality was determined using the protocol and the CASA system previously mentioned. Sperm (diluted 1:99 in P1-medium) was first mixed with the eggs and then immediately activated with artificial seawater (37 g/L). After 5 min of gamete contact time, eggs were moved for incubation with sterile sea water (19 to 21 °C). During the egg incubation, the dead eggs were removed, and sea water renewals were carried out. Fertilization rate was determined between 4 to 6 h post-fertilization (hpf), by examining a sample of 80 to 100 eggs of each spawn with a FullHD camera (Moticam 1080) mounted on a binocular loupe (Leica MZ16F). Embryonic development was evaluated over the following 48 h.2.4. Statistical AnalysisThe mean ± standard error was calculated for all the parameters. Shapiro–Wilk and Levene tests were used to check the normality of data distribution and variance homogeneity, respectively. One-way ANOVA and Kruskal–Wallis tests were used to analyze the sperm quality parameters between groups at the same week. Significant differences between treatments were detected using the Student–Newman–Keuls (SNK) test. Differences between the gonadosomatic indexes in control group and OP groups for both males and females were analyzed using Student’s t-tests (for parametric data) and Mann–Whitney U tests (for non-parametric data). Significant differences were detected when p-value < 0.05.All statistical analyses were performed using the statistical package SPSS version 24.0 for Windows software (SPSS Inc., Chicago, IL, USA).3. Results3.1. Males3.1.1. Sperm Production: Spermiating Males, Sperm Volume, and Sperm DensityRegarding the sperm production, the control group showed the highest percentage of spermiating males (with motile cells) (Figure 2A), achieving values >80% from the 7th week of treatment. The group OP-100 showed more spermiating males than the group OP-200, reaching a value of 50% between the 11th and 12th weeks. In addition, males from the OP-200 group did not produce sperm until the 7th week of the treatment, but males from the control and OP-100 groups produced sperm from the 5th week. Regarding sperm volume, the control group showed significantly higher values during the whole spermiation period (Figure 2B), reaching a maximum value of 4 mL/100 g fish at the 13th week of treatment. The OP-100 and OP-200 fish showed similar results with each other during the full period, and significant differences between both groups were found only at the 11th week because the OP-200 group achieved a maximum value of 1 mL/100 g fish.Concerning the sperm density (Figure 2C), the control group showed higher values than OP groups. The OP-100 and OP-200 groups showed an increase that started earlier (9th week) and reached higher values in the OP-100 group, reaching the maximum value in the 12th week, while the control group did it in the 15th week (>15 × 109 spermatozoa/mL).3.1.2. Sperm Quality Assessment: Total Motility and Progressive MotilityRegarding sperm motility patterns (Figure 3A,B), the control group showed higher values than experimental groups, reaching maximum MOT values of 61.7% and maximum pMOT values of 40.9% in the 12th week. The OP-100 and OP-200 groups maintained lower MOT and pMOT values than 15 and 5%, respectively, throughout the treatment.3.1.3. Biometric ParametersRegarding the mean GSI values (Figure 4A), a significant difference was found between groups, where control (5.92 ± 1.02) showed higher values than experimental groups OP-100 (2.23 ± 0.81) and OP-200 (3.06 ± 1.43). Concerning the pectoral fin colour (Figure 4B), it was found to be different in shading, which is considered a maturation biomarker [31]. In sexually immature eels, the pectoral fin is transparent, and it becomes darker during the sexual maturation. Prior to the experiment, the eel fins showed a transparent colour. At the end of the experiment, 87% of the control group males showed a black fin that was present in only 50% and 33% of the males in the OP-100 and OP-200 groups, respectively. A dark grey fin was found only in 13% of the control group, but in 40% and 33% of OP-100 and OP-200 groups, respectively.In the osmotic pump groups, there were animals that did not show any signs of maturation. In this sense, some of these males had the osmotic pump strongly encapsulated inside the conjunctive tissue (Figure 5A). However, males that showed gonadal maturation in the osmotic pump groups presented the osmotic pump free along the peritoneal cavity (Figure 5B).3.2. Females3.2.1. Sexual Maturation EvaluationRegarding the percentage of females that responded to the hormonal treatments, 10 out of 11 females from the control group and 8 out of 10 females from the OP-2ML4 reached the ovulation stage. Regarding the spawning induction, six females were induced to a spawning event and subject to abdominal stripping.3.2.2. Weight Evolution during Sexual MaturationAltogether, all the females from the control and the OP-2ML4 groups that responded to the hormonal treatment showed a weight increment in the weeks prior to sexual maturation (reaching means of 20 ± 14 and 16 ± 6%, respectively; see Appendix A, Figure A1; females that did not respond to the treatment exhibited a weight decrease). Regarding the time of maturation, the females from control group matured after 13 to 17 weeks of treatment, while females in the OP-2ML4 group showed two maturation phases: an early phase (females matured after the 4th to 10th week) and a late phase (20th to 23rd week) (Figure 6). A female in the OP-2ML4 group was able to show two ovulations: the first occurred in the 4th week and the second in the 20th week.There were differences in the weight increase in both groups, with respect to their initial weights. The control group showed a mean increase of 28 ± 20%, with a range between 8% and 68% (Figure 7A). The OP-2LM4 group showed a mean increase of 16 ± 14%, reaching a maximum value of 36% and a minimum value of 3% (Figure 7B).3.2.3. Biometric Parameter: Gonadosomatic Index (GSI)Both control and OP-2ML4 groups showed similar results, showing ovulation and gonad development (Figure 8). The control group reached a GSI of 51.2 ± 9.8%, while the OP-2ML4 group had a mean value of 49.5 ± 2.8%.3.2.4. Fertilization, Hatching Rate, and Embryonic DevelopmentAt the end of the experiment, the eggs collected from the six spawning females (five from control group and one from the OP-2ML4 group) were incubated to evaluate their embryonic development (see Appendix A, Figure A2). Embryos survived up to 33 h post-fertilization, and no hatching was observed.4. DiscussionThe traditional hormonal treatments to induce sexual maturation in fish entails a continuous handling of the animals, which has a negative impact on the maturation process due to the stress suffered by the fish. As a result, in over the past decades sustained-release delivery systems have been used, principally to control oocyte maturation in females, but also to improve the spermiation in several species. There are different sustained-release delivery systems available, such as cholesterol pellets, ethylene-vinyl acetate implants, or biodegradable microspheres, which have been widely used to control the reproduction in fishes [19]. The osmotic pumps have shown positive results in several fish species (Gadus morhua, [32]; Clarias gariepinus [23,33]), including the Japanese eel [24,25]. To date, the efficacy of osmotic pumps to induce the sexual maturation has not been proven in the European eel.4.1. MalesThis study shows that the osmotic pumps (ALZET; OP-100 and OP-200) can induce the testis maturation and sperm the production in part of the males.However, the sperm volumes produced by these males were significantly lower compared to the control group. The results of control group were similar to those obtained by Herranz-Jusdado et al. [10], since, in both studies the males started to produce sperm in the 6th week of the treatment and with similar volumes and densities throughout the weeks of experiment. Additionally, the low-density values of the OP-100 and OP-200 groups were not enough to make up for the low volume values. In addition to density, sperm quality is crucial to fertilization trials. In several fish species, the fertilization and hatching rates are used to evaluate the sperm quality, but due to the reproduction limitations of the European eel, quality is assessed by kinetic parameters (motility and velocity) of the spermatozoa [34]. In the present study, the spermatozoa from the OP-200 samples never showed motility during the experiment. However, the samples from the OP-100 group showed motility from the 10th week and reached a maximum (around 20% of motile cells) in the 15th week. In previous studies, Kagawa et al. [24] tried several doses of hormone (hCG) and showed that the animals that received doses from 5 to 50 IU/day presented the highest motility values (25–35%). However, in the present study the European eels received a dose from 15 to 20 IU/day (OP-100 and OP-200 groups, respectively), and the motilities were lower than the obtained in Japanese eels. Furthermore, we assessed the sexual maturation at the end of the experiment using biometric parameters such as GSI and fin colour. The mean GSI value for males from OP-100 and OP-200 groups were significantly lower compared to the value reached by the control group males. However, both experimental groups showed matured and spermiating males, which means that the osmotic pump loaded with hCGrec had the capacity of inducing the male’s sexual maturation. It was observed that in males that did not mature, the osmotic pump was encapsulated inside the conjunctive tissue, thus, the hormone could not be released properly. The encapsulation of the osmotic pumps inside the eels could be due to a fibrotic response by the animal’s immune system, as a measure to isolate the body from a foreign material [35]. We think that this phenomenon occurred in the present study because the male eels were very small (mean body weight of OP groups = 116.7 ± 11.1 g), compared to those in the study by Kagawa et al. [24], where they were larger (mean body weight = 336.7 ± 10.3 g) and did not have the problem of encapsulation.Regarding the fin colour, it became darker throughout the course of maturation coinciding with previous results [30]. The results showed that the highest percentage of fish with a dark fin belonged to control group (88%), followed by OP-100 (50%) and OP-200 (33%), which seems to be related to the percentage of spermiating males in each group (control, >80%; OP-100, 50%; OP-200, 40%). Differences in the final maturation stage reached by the OP-100 and OP-200 groups may have been due to the larger size of the pump in the OP-200 group (ALZET® OP-2006), which caused greater encapsulation in the males of this group.4.2. FemalesThis is the first study obtaining sexually matured European eel females using osmotic pumps (ALZET; OP-2ML4), generating developed ovaries and even spawning events. Fifty per cent of our implanted females reached the final maturation, which is similar to the results obtained by Kagawa et al. [25] (63.4% of females). The GSI values from OP-2ML4 group (49.5 ± 2.8%) showed promising results and demonstrate that sexual maturation occurred. In Kagawa et al. [25], GSI values (27.5 ± 4.1%) were lower than in the present study. This difference may be because CPE contains higher levels of LH and FSH than SPE [36]. In addition, the amount of CPE/day released by the pumps was higher than in Kagawa et al. [25]. Therefore, the present study has proved that the osmotic pumps are an alternative method to induce the sexual maturation in European eel females, but the high variability between females poses a problem for the standardization of the method.The female European eels need several weeks of standard hormonal treatment to reach the sexual maturation, between 16 to 25 weeks using SPE as hormonal treatment [37,38] and 12 to 25 weeks using CPE [11,39]. In the present study, sexual maturation was achieved between 13 to 17 weeks using CPE injections (the control group), which coincides with the aforementioned studies. In our study, two periods of sexual maturation were observed. A faster sexual ovulation was observed in some females in the OP-2ML4 group (50%) between 4th to 10th week, which represents the earliest sexual maturation obtained for European eel females, previously achieved by Pedersen [36], who reported ovulations between 7.5 to 11.5 weeks using a treatment of two doses of SPE per week. In addition, the weekly double dose of SPE [36] worked better than standard doses (one per week) because there was a shorter lapse of time between the injections and the levels of hormone maintained high throughout this time. In this sense, the osmotic pumps release the hormone continuously. Thus, the good results obtained in the present study coincide with the previous idea. To summarise, an in-depth study of the osmotic pumps as an alternative to hormonal injections should be necessary since the continuous hormone-releasing seems to work better than the weekly injections. Secondly, some of the implanted females reached a late sexual maturation between the 20th to 23rd week. That indicates that these females did not respond to the osmotic pumps, and they really responded to the weekly CPE injections after the 10th week. In this group it is important to note that one of the females (OP-1) showed two maturation peaks. The first one occurred in the 4th week due to the osmotic pump, while the second one occurred in the 19th week, apparently due to the CPE injections. This result shows that European eels could display a group-synchronous maturation of oocytes. Eels have been traditionally considered to have synchronous ovaries [38] because they spawn once in their life and then die. However, previous studies have demonstrated that ovaries could display a group-synchronous development when the female’s maturation is artificially induced [11,37,39,40], which agrees with the results of the present study. With respect to the weight evolution of the females, several studies have reported that a weight increase of 10% according to the initial weight, is a signal to induce the ovulation in female Japanese eels [41,42]. However, studies in the European eel showed that the individuals presented a higher variation, and the use of weight increment to induce the ovulation is not the more suitable method [11,36]. In the present study, the results showed a mean weight increase of 28 ± 20% (control group) and 16 ± 14% (OP-2ML4) from the beginning of the experiment. Nevertheless, this increment was lower than 10% in some of the females, and one of them even showed weight reduction. Females were fasted during the hormonal treatments, and consequently some of the animals suffered a weight reduction with respect to their initial weight. Thus, we consider that the weight gain to be examined is the one from the week previous to maturation, and not since the experiment began. In this sense, all the females showed a weight increment in comparison with the week prior to maturation. Furthermore, it would be recommended to conduct regular cannulations to assess the oocytes stage of development as a method to predict the optimal moment to induce the ovulation, which would be more accurate than the weight changes.Concerning embryonic development, in recent years the technology and incubation methods to produce European eel larvae have advanced considerably, allowing the production of large batches of viable eggs and larvae that reach the first feeding stage [43,44,45] despite the early development stages being sensitive to biophysical parameters [46,47]. Although it was not one of the objectives of our project, some of the females that achieved maturation and developed the oocytes in the required stage were induced to final ovulation. A total of six clutches were obtained, whose embryos were incubated reaching the stage of embryo formed with head, eyes, and somites [48] as maximum. In this study, no hatching was obtained, probably due to the lack of proper incubators to maintain the culture conditions in a precise way for a correct embryonic development. 5. ConclusionsWith regard to European eel males, this study showed that the controlled release systems (ALZET® osmotic pumps) were able to induce the maturation and spermiation, but not a sufficient amount of sperm (volume, density) with acceptable kinetic features (motility) for achieving a reasonable amount of high-quality sperm for aquaculture and research purposes.Concerning European eel females, our study demonstrated that the use of osmotic pumps loaded with CPE was an effective method to induce sexual maturation, generating extremely early maturations (from the 4th to 10th weeks) in 50% of the females. However, the high variability between females poses a problem for using this technique for aquaculture goals, so further studies with controlled release systems must be an ongoing task for that the method standardization.

animals : an open access journal from mdpi

10.3390/ani11082164

PMC8388502

Sea cage farms dominate European aquaculture production of seabass (Dicentrarchus labrax) and gilthead seabream (Sparus aurata). It means that to complete the commercialization process, fish must be crowded in a net, lifted from the rearing cage, and placed in a stunning/slaughtering tank during the extraction procedure. Brailing and pumping are the two techniques used. The brailing involves the use of a large net that is hoisted by a crane, and the fish and water are released from the brail by opening the closed end of the net with a release. The fish enter water through a pipe and pass through a grid that removes the water before being placed in the stunning/slaughtering tank. This paper examines the scientific progress made in these areas over the last two decades in relation to farmed seabass and seabream describing the consequences of different methodologies on the time fish takes to reach the unconscious stage, the different concentrations of stress indicators in plasma, and the evolution of flesh quality related to spoilage during fish shelf-life.

The behavioural responses of fish to a stressful situation must be considered an adverse reaction caused by the perception of pain. Consequently, the handling prior to stunning and the immediacy of loss consciousness following stunning are the aspects to take into account during the slaughtering process. The most common commercial stunning method in seabream and seabass is based on hypothermia, but other methods such as electrical stunning, carbon dioxide narcosis or anaesthetic with clove oil, are discussed in relation to the time to reach the unconsciousness stage and some welfare indicators. Although seawater plus ice slurry is currently accepted in some guidelines of fish welfare well practices at slaughter, it cannot be considered completely adequate due to the deferred speed at which cause loss of consciousness. New methods of incorporating some kind of anaesthetic in the stunning tank could be a solution to minimize the impact on the welfare of seabass and seabream at slaughtering.

1. IntroductionAnimal welfare evaluation should be promoted so that decisions are made based on scientific evidence rather than emotion, with the understanding that the concept of welfare is a characteristic of an animal, not something given to it, and can be precisely measured [1]. In the case of fish production, the objective measurement of animal welfare is an issue that must be addressed in order to promote adequate guidelines for the levels of acceptability management [2], particularly during the stunning and slaughter processes. The most contentious research area in non-mammalian welfare is the debate over whether fish feel pain [3], but given that fish exhibit behavioural and physiological responses similar to those found in mammals, there is a large scientific consensus that there is no adequate basis for denying them conscious pain experiences [4].If pain is defined as an unpleasant sensory and emotional experience associated with actual or potential harm, an animal must have sentience in order to experience pain. According to Chandroo et al. [5], fish suffer in ways similar to tetrapods because anatomical, pharmacological, and behavioural data indicate that affective states such as pain, fear, and stress are likely to be experienced. Nevertheless, arguments against the fact that fish feel pain repeatedly appear over the capacity for non-mammalian species to experience the discomfort or suffering rather than a nociceptive reflex. Pain is caused by neural processing in the brain that necessitates structural connectivity and the presence of a cortex, which fish lack [6]. Fish may not have the complex brains of the higher mammals, but they do have a nervous system that can detect noxious stimulation. Such an experience does not need a cortex because the experience is raw, tied directly to the immediate damage, and is an objective extension of that damage, which drives the aversive behavioural responses [7,8]. In fact, fish have nociceptors that detect noxious stimuli and brain pathways that process nociception signals in the same way that vertebrates do [9], so their behaviour responses after noxious stimuli administration are not simply reflexes but rather indicators of pain perception [10].The possibility that fish are sentient and, as a result, experience pain and suffering has become a major topic in aquaculture in order to provide appropriate conditions during slaughter, i.e., to be unconscious and insensible when slaughtered [11]. Simple risk analysis on a simple neural system shows that the probability that fish can feel pain is not negligible [12], and a precautionary principle for welfare consideration should still advance animal welfare protection [13]. As a result, the best practice would be to provide fish with the same level of protection that any other vertebrate receives [14]. Despite the fact that fishes are very different from us and are unlikely to have a capacity for awareness of pain or emotional feelings that meaningfully resemble our own [15], a strong alternative view is that complex animals with sophisticated behaviour probably have the capacity for suffering, though it may differ in degree and kind from the human experience [16]. In any case, sentient animals in our care must be kept in comfortable conditions that maximize their health and welfare, and they must be slaughtered as quickly and painlessly as possible [11]. The aquaculture industry should be governed by ethical principles ensuring the health and welfare of fish, including humane slaughter [12], which should incorporate them into a holistic assessment for fish management, not as a purely scientific analysis, assess available alternatives and take into account new knowledge to recalculate ethical stress in the new perspectives [17].Taking into consideration that the total aquaculture production of seabass and seabream increased from just under 8 thousand tons in 1990 to 522 thousand tons in 2019 [18], it seems pertinent to discuss the scientific evidence on slaughter methodologies to try to guarantee the best welfare conditions for both species.2. Impact of Stunning on Farmed Fish WelfareThe goal of optimal fish slaughter is to eliminate needless stress and agony during the procedure [19]. Humane slaughter methods are designed to bring about the rapid loss of consciousness and, ultimately, a complete loss of brain function in animals destined for use as food. This means minimizing or eliminating anxiety, pain, and distress associated with terminating the lives of the fish [20]. In addition, the approach utilized to kill the fish should do so fast after successful stunning to avoid regaining consciousness [2].Clearly, the development of appropriate commercial technologies to promote a humane slaughter for farmed fish is an active area of research, linked to a growing awareness in the aquaculture industry about the importance of ensuring that stunning takes place under the best conditions to cause loss of consciousness until the fish dies. Thus, it is necessary to identify stressful situations early enough so that an intervention can take place before detrimental effects occur [21] and determine how quickly the fish is rendered insensible, which is difficult in practice [22]. Finally, the induction of unconsciousness should not cause suffering even if the methodology used does not result in an immediate loss of consciousness [23].It is understood that consumer acceptance of aquaculture products will be increasingly influenced by the extent to which the industry is perceived to be dealing with fish welfare, obviously including the time of slaughter [24], with a growing insistence among consumers that the animals they eat were well treated [11]. The challenge for the fish farming industry in this context is to demonstrate that this activity is conducted in an ethical and humane manner [25].2.1. Pre-Slaughter HandlingThe application of slaughter technology varies by species, but it is well-established in the majority of segments of the food fish industry to achieve product quality control, efficiency, and processor safety [26]. The slaughtering process in a fish farm, in this case rearing the fish in sea cages, consists of a starvation period to empty the gut, crowding and collecting or pumping to remove the fish from the water, stunning, and killing. Despite the fact that most research focuses on the stress experienced during the slaughtering process, the negative prior handling is frequently overlooked [17]. As a result, rough handling during crowding and repeated catching cause additional stress, resulting in increased cortisol and haematocrit levels [27]. The impact of entire processes on welfare varies significantly depending on the species [28], and it also affects fish quality because pre-slaughter handling and slaughter methods start an irreversible process of flesh degradation [29]. Indeed, a multidisciplinary approach that considers animal behaviour as well as the various biochemical and physiological ante mortem and post-mortem processes could be the best strategy for determining fish welfare during stunning/slaughtering procedures and their impact on product quality [24]. Hence, techniques for pre-slaughter and slaughter should be used to reduce the level of evoked stress response and physical activity [30].2.2. Stunning MethodsThe most important aspects of the slaughtering process are the handling prior to stunning and the immediacy of loss consciousness following stunning. Thus, stunning methods that cause immediate loss of consciousness and reduce exposure to aversive situations are considered fast methods, while the methods that do not cause immediate loss of consciousness are considered slow methods [31]. Therefore, due to the relationship between an animal’s welfare and subsequent meat quality, methods that cause a slow loss of sensibility have a negative impact on the carcass’ overall quality, whereas methods that cause a rapid loss of sensibility have a positive impact on the carcass’ overall quality [31].Any killing protocol must be monitored to ensure its effectiveness, ensuring the least amount of fish suffering and allowing for the improvement of current methods, including the adoption of new techniques and improvements that take into account the quality of the flesh or operational costs [28]. In this sense, fish responses to stimuli and reflexes appear to be capable of distinguishing with reasonable certainty the state of awareness of a variety of species as they are killed [22,26]. Van de Vis et al. [32], on the other hand, believe that using behavioural indicators alone, even if they have been shown to correlate with brain activity, may not be sufficient, especially in ice slurry due to the muscle paralysis caused by cooling [31].Thus, the goals of farmed fish slaughtering can be summarized as follows: reduce fear and pain, ensure product quality, and enable process efficiency [33]. Additionally, management during crowding prior to harvest should be conducted carefully, fish should be quickly unconscious, a large number of animals can be slaughtered in a short period of time without compromising welfare, and finally, consider the final destination of fish.3. Criteria to Evaluate the Welfare Impact of Stunning Methods in Farmed Seabass and Seabream3.1. Delay to Reach the Unconsciousness StageSlaughter methods, which do not cause immediate loss of consciousness, are primarily used in farmed seabass and seabream stunning/slaughtering. This is because they are simple and inexpensive to apply [31] and more appropriate in a wide range of fish. The most common process to render the fish less active and easier to manage during slaughter is the cooling of animals. Because the hypothermia procedure for stunning appears to be temperature-dependent [34], it may be more effective in the case of warm species such as seabass and seabream. Adding sedative agents, such as clove oil, to pre-slaughter cooling would go a step further in extending the effectiveness of fish stunning and reducing pain, though it is unclear whether this method would be acceptable to consumers, both in terms of sensory appeal and food safety. Some slaughter guidelines [35] accept the use of water plus ice slurry to stun seabass and seabream. The fish move around for a short time after being placed in the stunning tank before slowing down and becoming paralyzed as their muscles cool [31]. Despite the fact that the time required for sure stunning is not short, it does not appear to be particularly stressful in terms of causing pre-slaughtering cooling, reduced movements at death and gave good responses related to hematic and muscular stress indicators [36] with decreasing breathing amplitude and movement loss on the stunning tank’s bottom after 3 min [34]. Furthermore, fish are not asphyxiated in the ice slurry because they can breathe [31], and the paralysis caused by rapid cooling is reversible, with fish returning to rearing conditions quickly regaining muscular movement. In this sense, the stunning/slaughtering tank’s uniform water temperature of around 0 °C ensures that the fish do not die of asphyxia but rather of thermal shock [37].The time required to detect the state of unconsciousness by using ice slurry to stun seabass varies between authors, ranging from 10 min [38] to 20 min [36,39]. The results for seabream are very similar, with time allocations ranging from 15 to 20 min [40]. Liquid ice or binary ice has been used to stun seabream [41,42,43] and seabass [44]. Fish are cooled faster than those stunned with ice slurry due to the physical properties of liquid ice (microscopic size of the ice crystals) and the low temperature of the stunning conditions. However, if the fish are kept in the ice slurry tank after stunning, the time to death would be around 25 [45] or 34 min [46], possibly reaching 40 min if the fish were crowded during pre-slaughter handling [44]. When compared to seabass, the time to death in seabream may be delayed by more than five minutes [44].Another method used for seabass stun is the diffusion of gases into the stunning tank, faster to reach unconsciousness than ice slurry. In the case of carbon dioxide, the time it takes to reach unconsciousness will be 7 min, possibly as little as 4 min if the gas is insufflating into the ice slurry [36]. Aside from carbon dioxide, a mixture of nitrogen in various proportions has also been used but with no discernible differences [34]. Thus, the seabass dies after 16 [46] or 20 min [47] in CO2-supersaturated seawater, with the gas mixture reducing this time to 10 min [34].The addition of anaesthetic has also been tested for stunning. The clove oil anaesthesia will render the fish unconscious before transferring them to an ice slurry for slaughter [38]. Recently, an innovative new option for reducing gilthead seabream stress at slaughter has been proposed: the inclusion of nanoencapsulated clove oil in the ice used to stunning in order to improve their water solubility [42].Some electric stunning trials have also yielded promising results in seabream and sea-bass [31,34,36,48,49,50]. Following the application of an electric current, the fish are transferred to a tank filled with ice slurry for slaughter. In all cases, the fish are unconscious before being transferred to the slaughtering tank, though the recovery time varies depending on the methodology used, ranging from less than one minute [49] to more than twenty minutes [34]. This is especially important because a humane slaughter is achieved if the fish becomes insensible very quickly after being exposed to an electric field and remains insensible until death occurs [51].The last three procedures for stunning seabream and seabass are asphyxia in air, spiking, and percussive stunning. Asphyxia should not be considered. It would not meet current animal welfare requirements for aquaculture [46] due to the lengthy time required for stunning (loss of movements only after more than an hour) and the violent reactions in the first minutes [36]. Spiking and percussive stunning, on the other hand, are not practical for batches of small-sized species, despite being the fastest and least stressful of all previously cited methodologies [36,52].3.2. Metabolic Indicators of StressThe stunning/slaughtering methods cause significant changes in plasma stress indicators, such as osmolality, glucose, lactate, and cortisol, reaching higher levels than the undisturbed fish in all cases [46]. However, depending on the method of stunning/slaughtering and the species, the magnitude of these changes varies. Even the stocking density and time spent in confinement prior to harvesting are factors to consider [53]. Thus, there were no differences in mean plasma cortisol, glucose, or lactate concentrations in seabream after electrical stunning followed by immersion in ice slurry or ice slurry alone [50]. Cortisol and lactate levels were higher in seabass stunned by electricity than in those immersed in ice slurry [54].When it comes to using carbon dioxide to stun seabass, the results vary between authors, with some reporting higher levels of cortisol and lactate [36] and others reporting lower levels of lactate [46] when using the gas instead of ice slurry. This last finding is supported by a longer time to death, which could lead to increased metabolic activity in muscles and, as a result, the accumulation of metabolites such as plasma lactate. When the gas, nitrogen, or a mixture of nitrogen and carbon dioxide, was dissolved in the ice slurry, glucose and lactate did not differ significantly with or without gas, but cortisol was higher in the ice slurry alone, most likely due to the longest time for the fish to become stunning [39].The use of anaesthetics in ice slurry produced different results in seabass and seabream. While plasma lactate levels in seabass stunned with clove oil were higher than in ice slurry [55,56], they were lower in seabream (Lopez-Cánovas et al., 2019). Furthermore, both plasma glucose and cortisol levels were lower after adding clove oil to seabream [42] or seabass [56], but only under experimental conditions, with no significant variations in serum cortisol levels detected under industrial farm conditions, most likely due to greater individual variability.3.3. Flesh QualityA way to evaluate the fish welfare during the slaughter is to analyse the impact on related aspects such as flesh quality parameters. The slaughter methods have a significant impact on the quality of the flesh, changing its physical properties, spoilage processes during ice storage, and sensory attributes [57]. Taking into account the impact of pre-slaughter stress on commercially harvested fish; however, the expected variations in quality parameters will be difficult to record and will most likely go unnoticed alongside the changes caused by slaughter [58]. As a result, various authors have investigated the effects of the stunning/slaughter method on the onset and resolution of rigor mortis, the evolution of post-mortem pH, freshness indicators such as the K-value (based on ATP breakdown and subsequent by-products) or the QIM (Quality Index Method) and, in some cases, other physical parameters, such as texture and colour.In terms of rigor mortis, Knowles et al. [48] concluded that electrical stunning accelerated the pattern of onset and resolution of rigor mortis in seabass when compared to immersion in ice slurry, despite the fact that the time to stun is shorter [34]. The process was even faster when carbon dioxide was used [36]. This gradual onset of rigor mortis can provide information on the fish’s stress status prior to death while also preserving cellular energetic reserves [36]. Given the lower muscle activity and energy reserves consumption, the ice slurry would be appropriate for seabass [34]. In this regard, combining clove oil anaesthesia with immersion in ice slurry reduced anaerobic glycolytic activity and delayed the onset of rigor mortis [38], with a significant effect on ultimate muscle pH [58,59].The evolution of pH is linked to rigor and energy consumption and is influenced by factors such as overcrowding and oxygen availability prior to slaughter [60]. When compared to ice slurry, seabass stunned by carbon dioxide showed a sharper decrease in muscle pH during the first hours [46]. Giuffrida et al. [40], on the other hand, found no significant differences in muscle pH values among seabream stunned/slaughtered with carbon dioxide or ice slurry. Electrical stunning seabass yielded contradictory results, with an initially lower pH compared to ice slurry or no differences between the two methods [48]. Percussive stunning [34] and asphyxia [39,61] were used to reach the extremes of highest and lowest pH values, respectively.The K-value, which measures the concentration of muscle ATP and its degradation metabolites, can be used to detect slaughtering distress [24]. During storage of seabream using ice slurry, higher values of the ratio ATP/IMP are observed than when using carbon dioxide [40]. Sea bass stunned/slaughtered with ice slurry had higher IMP concentrations and lower levels of inosine and hypoxanthine than fish stunned/slaughtered with electricity, seawater, or flake ice saturated with a mixture of nitrogen and carbon dioxide, indicating a better freshness condition in ice slurry fish [34]. When compared to ice slurry, liquid ice maintains higher K-value values in seabream after slaughter [43]. Furthermore, the K-value of seabream killed by immersion in ice slurry or a blow to the head after anaesthesia with clove oil did not differ during chilling storage [62].Depending on the authors, the results of the sensory evaluation of freshness, whether using the QIM methodology or similar protocols, were slightly different. While no differences were found between fish stunned/slaughtered on an ice slurry and fish stunned/slaughtered on carbon dioxide [46] or electricity [48], in some experiments with seabass, the fish stunned/slaughtered on an ice slurry maintained higher freshness scores during the shelf-life in others [34,36]. Seabass slaughtered in liquid ice showed more distinct differences, with significantly lower spoilage rates than fish slaughtered in ice slurry [44]. After percussive stunning, the QIM scores in seabream were similar to those obtained after immersion in an ice slurry (van de Vis et al., 2003).Electrical stunning has been found to have a lower hardness than ice slurry in studies of seabass flesh texture variations [50]. The addition of clove oil to the rearing tank reduced pre-slaughter harvesting stress in seabream, but the hardness of the fillet, both raw and cooked, did not differ significantly from the stress condition caused by harvest net crowding [59]. Despite the fact that intense exercise prior to slaughter alters post-mortem muscle degradation processes via changes in myofibrillar proteins [63,64], the stunning/slaughtering method had no effect on skin lightness and colour in seabass [48,49] or seabream [43].4. ConclusionsConsumers are becoming increasingly concerned about the treatment of the animals that provide us with food. Aquaculture’s higher efficiency in obtaining protein when compared to other farmed terrestrial animals has made it not only an important source of protein for human consumption but also a rapidly growing activity around the world. In this context, self-imposed good practice guidelines in aquaculture farms may be considered consistent with fish welfare outcomes, ensuring reproductive success, good growth performance, and product quality from a market standpoint. However, the most important aspect of animal welfare, the stunning/slaughter procedure, which is well-known and studied in other farmed terrestrial animals, is not well-studied in fish, especially seabass and seabream. In these species, some peculiarities arise in some stages of the process since, during the crowd and harvest procedures in the marine cages, the fish are transferred to the stunning tanks exhibiting a frenzied escape behaviour. Only the reduction of the time available in both processes helps to reduce the suffering of the fish.The scientific evidence from the last two decades in these species supports stunning with ice slurry to promote hypothermia while paying attention to the flesh quality parameters as indicators of metabolic stress, which offer similar results to those associated with electrical stunning. Other methods that could be used on a commercial scale, such as asphyxia or carbon dioxide (alone or in combination with nitrogen), had more negative consequences. However, the most important question to consider is whether a slow slaughter method can be accepted due to the time it takes for the animal to become unconscious. The use of ice slurry in conjunction with a nanoencapsulated anaesthetic can help reduce fish stress during stunning and achieve an appropriate state of unconsciousness in seabass and seabream more quickly, all while remaining compliant with animal welfare standards.

animals : an open access journal from mdpi

10.3390/ani12030309

PMC8833531

In Aotearoa–New Zealand, the helicopter application of the toxin sodium fluoroacetate (1080) is a common method for controlling invasive mammals. However, the application of 1080 using current methods leaves some surviving mammals, meaning eradication cannot be achieved. A new application method, called 1080-to-zero, aims to eradicate target mammals or reduce them to near-zero levels. This study monitored the response of invasive black rats (Rattus rattus) to a 1080-to-zero application and a standard 1080 application. In this case it found that the 1080-to-zero method did not improve rat removal compared to the standard application, and did not reduce rats to near-zero levels. However, these results differ from a 1080-to-zero application in another part of the country, which did achieve near-zero abundance for rats. Questions remain about how local factors affect this tool, and how it can be further improved.

Aotearoa–New Zealand has embarked on an ambitious goal: to completely eradicate key invasive mammals by 2050. This will require novel tools capable of eliminating pests on a large scale. In New Zealand, large-scale pest suppression is typically carried out using aerial application of the toxin sodium fluoroacetate (1080). However, as currently applied, this tool does not remove all individuals. A novel application method, dubbed ‘1080-to-zero’, aims to change this and reduce the abundances of target pests to zero or near-zero. One such target is black rats (Rattus rattus), an invasive species challenging to control using ground-based methods. This study monitored and compared the response of black rats to a 1080-to-zero operation and a standard suppression 1080 operation. No difference in the efficacy of rat removal was found between the two treatments. The 1080-to-zero operation did not achieve its goal of rat elimination or reduction to near-zero levels, with an estimated 1540 rats surviving across the 2200 ha treatment area. However, 1080 operations can produce variable responses, and the results observed here differ from the only other reported 1080-to-zero operation. We encourage further research into this tool, including how factors such as ecosystem type, mast fruiting and operational timing influence success.

1. IntroductionInvasive mammals present a global threat to biodiversity, especially in island ecosystems such as those of Aotearoa–New Zealand [1]. Native New Zealand species are particularly vulnerable, as they evolved in the absence of terrestrial non-volant mammals [2]. Predation and herbivory from introduced mammals has been a leading cause of extinctions and habitat degradation [3,4]. In response, New Zealand has made significant developments in invasive mammal control, pioneering techniques such as eradication from islands [5,6] and fenced mammal-free sanctuaries [7]. Recently, New Zealand has begun pursuing a new era of management: the eradication of invasive mammals from large, contiguous landscapes.In 2016, the government announced Predator Free 2050 (PF2050): an initiative to eradicate rats, possums and mustelids from all of New Zealand by 2050. This encompasses seven species: black rat (Rattus rattus), brown rat (R. norvegicus), Polynesian rat (R. exulans), brushtail possum (Trichosurus vulpecula), ferret (Mustela furo), the least weasel (M. nivalis) and stoat (M. erminea). This has received strong support but represents a sharp departure from historic management practices [8]. Typically, eradication has been employed in small, isolated areas where the chances of success are relatively high and reinvasion relatively low. Larger landscapes have received suppression management: the control of pests to low but non-zero abundances [9,10]. At 11.5 and 15 million hectares each, New Zealand’s two ‘mainland’ islands are far from small, and the management areas within them are not isolated. The complex, contiguous landscape makes it challenging to meet two of the key requirements for eradication: the total removal of individuals within the treatment area and the preclusion of reinvasion [11,12].In response to these challenges, many of the emerging mainland eradication projects have adopted what is termed a ‘remove and protect’ management strategy [13]. This strategy aims to remove resident pests and protect against reinvasion by detecting and removing invaders as they arrive. Using this strategy, pest elimination zones can be established and expanded, ultimately enabling mainland eradication [13]. The remove and protect strategy is best served by treating large areas, ideally in the tens of thousands of hectares. This reduces reinvasion pressure by shrinking the boundary-to-treatment-area ratio and improving the flexibility to use natural barriers to movement [14,15].As treatment area increases, financial cost becomes a limiting factor. Although ground-based traps and bait stations can be effective at eradicating mammals [6], they are labour-intensive, especially when targeting pests with small home ranges [13]. Rats have the smallest home ranges of the PF2050 target species (typically < 1 ha), and their elimination can require device densities of at least one device every 50 m [6,16]. In areas without access to ample volunteers, these labour costs are typically prohibitive over a large scale, even when considering the multi-trap devices currently on the market [17]. Additionally, parts of the New Zealand landscape (e.g., cliffs) make ground-based access extremely challenging.In contrast, the aerial application of toxins is comparatively cost-effective [13,18,19]. However, there are critical gaps in what this tool can achieve. In New Zealand, the only approved toxins for aerial application are brodifacoum and sodium fluoroacetate (also known as 1080). Brodifacoum can be reliably used to eradicate mammals from large islands (e.g., 11,300 ha Campbell Island [20]), but its use is restricted on the mainland due to its long environmental persistence [9]. In comparison, 1080 has a low environmental persistence and is regularly used for the mainland suppression of possums, rats and stoats [10]. However, it tends to leave small surviving populations, making eradication unachievable [9].Motivated by PF2050, attention has turned to developing 1080 application methods capable of achieving results more typical of brodifacoum [9,21]. The success of aerial toxin applications is known to be highly dependent on the precise operational methods followed [22], and the past few decades have seen significant improvements in the safety, reliability and efficacy of 1080 applications [18]. Currently, owing to their different toxicology and purposes, brodifacoum and 1080 application methods differ, with the former often including features such as two toxic applications and higher bait sow rates [23]. A novel eradication-focused 1080 application may be able to provide a cost-effective, large-scale elimination tool for mainland eradications.Zero Invasive Predators, a research NGO, has developed a new 1080 application method which borrows heavily from brodifacoum applications. Dubbed ‘1080-to-zero’, this method features several improvements: pre-feeding with non-toxic bait twice instead of once; overlapping helicopter flight paths to reduce gaps in bait distribution; and applying more bait per hectare (in contrast with Dilks [24]). If post-treatment monitoring indicates the presence of survivors, the pre-feed and toxic bait are re-applied several weeks later, producing a ‘double 1080′ operation [21].In 2019, two large-scale 1080-to-zero operations were carried out: one in the Perth River Valley, on the West Coast of the South Island, and one in the Kaitake Ranges of Taranaki, in the North Island. To critically evaluate the success of a 1080-to-zero operation, and indeed any 1080 operation, it is necessary to know the density of pests prior to the 1080 application, the extent and duration of the reduction to zero density, and the rate at which the population recovers without ongoing management. The 2019 South Island short-term results for possums, black rats and stoats have been published elsewhere [21]. Here, we evaluate the short- and long-term effectiveness of the 2019 North Island 1080-to-zero application on black rats, one of the PF2050 target species most difficult to eradicate using ground-based tools, and thus most in need of aerial alternatives [25].2. Materials and Methods2.1. Study SitesThe study took place in Te Papakura o Taranaki, a national park on the west coast of the North Island of New Zealand (Figure 1). The 34,000 ha park is centred on Taranaki Mounga, a large dormant volcano hereafter referred to as the Mounga. To the north are the 2200 ha Kaitake Ranges, included in the national park. One pest monitoring site was selected in the Kaitake, and one at the base of the Mounga. These were matched by approximate elevation (330 m at Kaitake vs. 615 m at Mounga), distance from bush edge (837 m vs. 1260 m respectively) and habitat type. Both sites were in mixed podocarp hardwood forests, but the fine-scale vegetation composition varied [26]. The Kaitake site was situated in a tawa (Beilschmiedia tawa)-dominated lowland forest, with rewarewa (Knightia excelsa), pukatea (Laurelia novae-zelandiae) and hinau (Elaeocarpus dentatus) being common [26]. The Mounga site was located in a lower montane forest, dominated by kahikatea (Dacrycarpus dacrydioides), rimu (Dacrydium cupressinum) and kamahi (Pterophylla racemose) [26]. In March and April of 2019, unusually heavy fruit production was observed at both sites. This is consistent with Department of Conservation seed sampling, which indicated 2019 was a heavy mast year [27]. Mass synchronised seed and fruit production was observed nationwide, including in beach and rimu forests and tussock grasslands [27].2.2. 1080 Aerial Application TreatmentsIn 2019, as part of a PF2050 possum elimination project called Restore Kaitake, the Kaitake Ranges were treated with a 1080-to-zero aerial toxin application. Although primarily targeting possums, this application used the same methods as 1080-to-zero operations targeting possums, rats and stoats [21]. Non-toxic pre-feed was applied on 4 and 17 April 2019, followed by toxic bait on 1 May 2019. Subsequent possum monitoring by the Restore Kaitake project identified numerous survivors: 21 days post-toxin application monitoring detected possums at 27 of 60 trail cameras, with one camera per 42 ha throughout the Kaitake Ranges (personal communication [28]). This level of survivorship triggered a second application, with pre-feed applied on 28 June and 18 September, and toxic on 31 October. Bait sow rates were reduced between the first and second operations to reflect the reduction in target pest numbers, dropping from 2 to 1.5 kg/ha for pre-feed, and from 4 to 2 kg/ha for toxic. As part of the 1080-to-zero protocol, the bait lure was also changed to reduce bait-shyness. The first operation used cinnamon-flavoured 6 g RS5 cereal pellets, while the second used orange lured 6 g Whanganui #7 cereal pellets. Helicopters used a 50% flight path overlap for both pre-feed and toxic applications to provide an even distribution of bait and reduce gaps (Figure 2).In the same year, as part of a 3-yearly pest control plan, the rest of the national park received a standard 1080 application [24] following the Department of Conservation’s protocols [29]. Non-toxic pre-feed was applied once on 18–19 July and toxic bait on 26–27 June. Both the pre-feed and toxin applications used cinnamon-flavoured RS5 6 g cereal pellets with a sow rate of 2 kg/ha. The helicopter paths had no overlap.For both the 1080-to-zero and standard applications, the bait had a lure concentration of 0.3%, and toxic pellets had a 1080 concentration of 0.15%. Bait was supplied by Orillion from Whanganui, New Zealand, and manufactured between 2.3 and 5.6 months before application. Bait was only distributed in fine weather, and all operations had dry weather in the week following each application.2.3. Pest-Monitoring TimelineMonitoring of black rats (Rattus rattus) was repeated 10 times between February 2019 and February 2021: twice prior to the 1080 applications, 3 weeks after each toxic application and every 3 months thereafter until February 2021. This provided pre-1080, immediately post-1080 and approximately 15 months of recovering pest densities. Although the rate of toxin breakdown is variable depending on environmental conditions [30], we selected a 3-week monitoring stand-down period after each toxic application to reduce the potential for ongoing toxin exposure and to minimise the potential for device avoidance [31,32] due to acute stress from the operation. Monitoring after the standard operation was slightly delayed due to logistical complications, taking place 4.8 weeks after the toxic application compared to 3.3 and 2.7 weeks for the two 1080-to-zero operations. For each of the 10 trips, monitoring was carried out at both study sites, except for May 2019, when flooding prevented access to the standard treatment site. For logistical reasons, both sites could not be observed at once, and measurements for each site were on average 13 and no more than 20 days apart.2.4. Rat Mark-RecaptureRat density was measured using mark-recapture over 5 consecutive nights, except for one instance where trapping was cut short at 4 nights, and another instance where trapping was suspended for 2 nights, both at the standard treatment site due to heavy rain. Live capture Tomahawk model 201 traps were spaced 25 m apart in a 7 × 7 grid (n = 49 traps), baited with a 50:50 Pic’sTM peanut butter and oat mixture. For the first 1080-to-zero monitoring session, the traps were covered with vegetation for rain shelter, and hay bedding was provided for warmth. However, the rats frequently pulled the vegetation inside the trap, and the hay became cold when wet. Despite trapping in late summer, this was not sufficient to prevent cold exposure. For all trapping thereafter, a white corflute rain cover was fitted over the back of the traps, and the hay bedding was replaced with water-resistant DacronTM stuffing. With these measures, evidence of cold exposure was rare even in winter. Traps were checked every morning as per the conditions of our animal ethics permissions (University of Auckland Animal Ethics Committee Approval No. 8442).Each captured rat was identified according to its species and transferred to a plastic bag where it was weighed using a spring scale and then anaesthetised to effect using hand-pumped isoflurane gas [33]. While unconscious, its sex was identified, and females were checked for fur wear around the nipples—an indicator of suckling pups and thus breeding. While still unconscious, the animal was then fitted with a uniquely numbered metal ear tag for individual identification and returned to the cage to recover. The anaesthesia provided pain relief during ear tagging and had minor memory-loss effects, which improve the recapture rate [34]. Upon full recovery from the anaesthesia, the animal was released at the site of capture. For each recapture, the trap number and ear tag ID were recorded.2.5. Relative Abundance IndicesRelative abundance indices (i.e., percent of devices detecting rats) were obtained using two different devices: wax tags and tracking tunnels. Wax tags were used to supplement the rat mark-recapture monitoring [35]. On the first night of each trapping session, unscented wax tags were attached to a tree or tree fern, 30 cm off the ground, within 1 m of each rat mark-recapture trap (n = 49 tags). Wax tags were recovered after 7 nights, except for two instances at the standard treatment site where flooding disrupted site access, in which case they were recovered after 5 nights. Wax tags were assessed for rat chew. Rat chew was differentiated from mouse chew by having an incisor width of 0.6 mm or more, determined using a 0.6 mm wire for a reference in the field [36].From November 2019 onward, tracking tunnels were added to the study to improve the detection of rats after the 1080 treatment [37]. Tracking tunnels were thus present for six of the ten monitoring events at each site. At each site, we used the closest existing Department of Conservation tracking tunnel line (one line per site, 297 metres from the standard treatment site and 235 metres from the 1080-to-zero site). Each line was approximately 1 km in from the forest edge and had 10 tunnels spaced 50 m apart. Cards were deployed for one dry night during mark-recapture trapping. Pre-inked Black TrakkaTM cards were installed baited with peanut butter in the centre of the card and retrieved the next day.2.6. Mark-Recapture Statistical AnalysisMark-recapture data were analysed using the package ‘secr’ in R software to model spatially explicit capture-recapture with half-normal detection curves and full-likelihood [38]. Captures of non-target species were excluded from analysis. Covariates across mark-recapture sessions were applied to density, g0 (capture probability) and σ (scale of movement) to estimate pre-1080 densities, the 1080 impact and the recovery thereafter at each site.The covariates of interest were: 1080 effect (0 for sessions prior to the first 1080 application and 1 thereafter), linear recovery (days since the most recent 1080 application), accelerating recovery (the square of days since 1080), diminishing recovery (the square root of days since 1080), season, mark-recapture trip (numbered 1 to 10) and site (1080-to-zero vs. standard). Accelerating and diminishing recovery were always applied in additive combination with linear recovery. The effect of learnt behaviour ‘b’ was also investigated for g0.Due to the large number of covariates, it was not possible to model every covariate combination. Instead, six sequential model sets were used to target different aspects of the model. The difference in the Akaike Information Criterion adjusted for small samples (dAICc) was used to compare models. At each stage, models with a dAICc > 10 were removed from further analysis to avoid excessive model proliferation.Density was investigated first. Holding g0 and σ constant, all additive covariate combinations for density were modelled (model set 1a, 51 models). A subset of these models, only considering temporal covariates, was also assembled (model set 1b, 27 models). This was followed by a preliminary investigation of g0 and σ. Density was modelled using the best-ranked covariate combination from model set 1a, and each covariate was added to g0 (model set 2a, 8 models) or σ (model set 2b, 7 models) while holding the other constant. At this stage, site was identified as a highly ranked covariate, with the potential to affect all of density, g0 and σ (see results). For each of the most supported models (dAICc < 10) from sets 1a and 1b, every combination of site across density, g0 and σ was generated (model set 3, 40 models).The most supported models from model set 3 became the ‘base models’ for further refinement of g0 and σ (model set 4, 104 models). For each base model, each covariate was added independently to g0 or σ. Any addition which was supported (AICc less than or within 2 of the base model) was noted, and all additive combinations thereof were generated. All models from this set were then compared to identify the most supported covariate combinations overall.From there, interactions with site were explored using the most supported models from model set 4 (model set 5, 54 models). Where site and another covariate influenced the same variable, the interaction between them was modelled. Where there were multiple possible interactions with site, each was modelled pair-wise independently. Any interactions which improved the model (AICc within 2 of its exclusion) were then applied together as multiple pair-wise interactions with site. All models from this set were then compared.Finally, a new covariate was generated: ‘method’, with a value of 1 for the first 1080-to-zero session and 0 for all others. This reflects the slightly different trap set-up used in that monitoring session. The ‘method’ covariate was added to all of the most supported models from set 5 (model set 6, 28 models). The final model with the lowest AICc score was selected from this group.2.7. Relative Abundance Indices Statistical AnalysisThe relative abundance index (i.e., percent of devices detecting rats) was calculated separately for wax tags and tracking tunnels, and for each monitoring session at each site. The results for each site were graphed for visual comparison with density estimates. The results from the two sites were then pooled for correlation testing. The Pearson correlation test in R software was used to assess the linear correlation between wax tag index and density estimates, tracking tunnel index and density estimates, and wax tag and tracking tunnel indices [39]. The relationship between the density estimates and tracking tunnel and wax tag indices was then visualised using a fitted least-squares linear regression line drawn with the package ggplot2 in R software [40].3. Results3.1. Mark-Recapture TrappingAll rats captured were identified as black rats (Rattus rattus). There was no apparent sex bias. Excluding sessions with fewer than 10 individuals, females composed one to two thirds of captures per session, and overall half of all captures in the study. Based on the presence of nipple fur wear for females and descended tests for males, 70 g represented the lowest recorded weight for sexual maturity, but reproduction was not common below 120 g. Breeding occurred primarily in summer and autumn, with fur wear around females’ nipples twice as common during these seasons than in winter and spring. Smaller juveniles (30 to 70 g) representing recent breeding composed 15% of captures in summer and 16% in autumn, but only 3% in winter and 6% in spring. Larger juveniles (70 to 120 g) increased from 14% to 26% of captures from summer to autumn, then decreased to 20% in winter and 0% in spring. These demographic trends did not differ notably between years or sites.The 1080-to-zero and standard treatment sites, respectively, recorded 544 and 216 rat captures of 203 and 126 unique individuals. Monitoring immediately after 1080 returned one, two and zero rats for the two 1080-to-zero applications and one standard application, respectively. No individuals captured before the standard application or first 1080-to-zero application were recaptured afterwards. However, one individual captured between the two 1080-to-zero applications was recaptured after the second application. In total, there were three sessions with zero rats captured, all at the standard treatment site after 1080.3.2. Relative Abundance IndicesThe Pearson linear correlation tests returned high correlation coefficients (r) between density estimates and relative abundance indices (Figure 3). Rat density and tracking tunnel index were more strongly correlated (r = 0.92, p < 0.001) than rat density and wax tag index (r = 0.64, p < 0.01), but both were significant. Tracking tunnels and wax tags also showed a significant correlation (r = 0.82, p < 0.01). There was evidence of differential performance of relative abundance indices at very low or very high densities. At low density, on three occasions, wax tags returned no rat sign when rats had been identified via trapping or tunnels. On one occasion, tunnels returned rat sign when rats had not been identified via trapping. At high density, tracking tunnels were saturated, returning 90% or 100% rat detection. Wax tags did not saturate but may lose linearity at very high densities, although this is based on only a small number of such observations (n = 2).3.3. Density EstimatesIn total, 242 models were run across 6 model sets (the exhaustive list is in the Supplementary Materials). During early analysis, the accelerating recovery covariate returned exceptionally poor results (dAICc > 650) when applied to density, g0 or σ alone or with another temporal covariate. This covariate was thus removed from the analysis.The investigation of density while holding g0 and σ constant returned two competitive models when including site (model set 1a) and three when excluding site (1b) (Table 1). These tended to vary density by 1080, diminishing recovery and site. Where site was excluded (set 1b), it was replaced by trapping or season. One model excluded the 1080 covariate and one supported a linear recovery.The exploration of g0 and σ using the most supported model from set 1a revealed a strong effect of site on g0 and σ (model sets 2a and 2b, Table 2). When compared, the effect of site on g0 was stronger than σ (dAICc 32.58). Further exploration of site across density, g0 and σ used all four density configurations identified in set 1a and 1b (model set 3, Table 2). Six competitive models were identified, all of which included site on g0, and some of which also included site on density, σ or both. These included five density configurations: all shared the diminishing recovery scenario, most included the 1080 suppression effect, and some included an effect of site, trip or both.Model set 4 added further covariates to the competitive models from set 3 (Table 3). The resulting models ranked quite closely, with 17 modes with dAICc less than 10 and 8 with less than 5. The model configuration varied: the only universal features were an effect of site on g0 and diminishing recovery on density. The addition of trip, linear recovery or b often ranked well for g0, and sometimes trip or 1080 for σ.When investigating interactions with the site covariate, most competitive models returned an interaction with 1080 and/or linear recovery for density (model set 5, Table 4). There was no difference between including a site interaction for 1080 and linear recovery (dAICc 0), or just for 1080 (dAICc 0.07). This continued into the final model set, where three models returned a dAICc of less than two, and the most supported of these failed to achieve a model weight (AICwt) of more than 0.5. Of these three models, the two most supported included an interaction between site and 1080, and the third also included an interaction with linear recovery. All three models supported the addition of method to g0. The highest ranked model was selected for graphing and analysis.Throughout the model building process, a diminishing recovery rate of density was consistently supported. A site effect on g0 was also supported, with the standard treatment site having a lower capture probability than the 1080-to-zero site. The movement parameter σ tended not to be influenced by covariates, and by the final model set all of the most supported models (dAICc < 10) held σ constant. Based on the most competitive (dAICc < 2) models from the final model set, there was strong support for an interaction between site and 1080 on density. There was weak support for an interaction between site and linear recovery, present in one of the three models. There was moderate support for a linear and learnt behaviour effect on g0, or a trip effect. However, the final g0 covariate configuration had little effect on g0 values.In the final most supported model, σ was constant, at a value of 22.24 (20.63–23.98, 95% lcl to ucl). The standard treatment site had a much lower g0 than the 1080-to-zero site, with a capture probability ranging between 0.03 and 0.05 versus 0.09 and 0.16 across all ten sessions. Capture probability increased over time at both sites, regardless of 1080 (Figure 4). The model found an effect of method on g0: the different capture method used in the first monitoring session at the 1080-to-zero site returned a much higher capture probability than the second (0.16 vs. 0.09), but with g0 increasing over time the final capture probability was similar to the first (0.15).Density was higher at the standard site than the 1080-to-zero site before 1080 (25.84 vs. 14.15 rats/ha) but similar 3 weeks after treatment (0.51 vs. 0.72 and 0.77 rats/ha for the standard and two 1080-to-zero applications) (Figure 5). Using the density estimates and effective sampling area (4.35 ha), an estimated 62 and 112 individuals were present at the 1080-to-zero and standard treatment site before 1080 application. Based on these estimates, 65% and 38% of the resident 1080-to-zero and standard site populations were captured in the final mark-recapture session prior to treatment. The 1080 treatment resulted in an estimated 95% and 98% reduction in population at the 1080-to-zero (first application) and standard site. At both sites, population growth occurred rapidly between January and August 2020 and then plateaued, 9 months after the second 1080-to-zero application, and 14 months after the standard application. At the 1080-to-zero site, a slight density increase was seen in the last capture session, but it is unknown if this indicates further population growth after the monitoring ended. The standard treatment site may have been slightly slower to recover, but large error bars preclude a fine-scale comparison, and the final model did not support an interaction between site and recovery rate. This indicates that any site difference in recovery was minimal. The highest densities recorded following 1080 were slightly lower at the standard site (11.78 vs. 13.04 rats/ha, standard vs. 1080-to-zero), although the variation was high and the error bars overlap. These represent a return to 92% and 43% of pre-1080 densities at the 1080-to-zero and standard sites. By the end of the study, although final densities at each site were similar, only the 1080-to-zero site had returned to densities similar to those recorded before 1080, while the standard treatment site had not.4. DiscussionThe assessment of invasive species management tools ideally requires a clear measurement of pest densities immediately before and after treatment (‘short term’), and the rate of recovery thereafter (‘long term’). Without this information, it is not possible to determine the magnitude or duration of effect the tool had. We used extended rat monitoring to compare the short- and long-term efficacy of a novel 1080-to-zero aerial toxin application to a standard 1080 application, in the context of black rat (Rattus rattus) management.Rat density can be influenced by a number of factors, including habitat [41], predation [42] and food abundance [10]. Densities also fluctuate seasonally and yearly [42]. The maximum densities observed both before and after 1080 were generally consistent with the upper estimates previously recorded for North Island’s mixed and podocarp forests (0.29 to 13.61 rats/ha [42]; 3 to 12 rats/ha [43]). However, the pre-1080 density at the standard treatment site was amongst the highest ever recorded on the New Zealand mainland (17 rats/ha, in preparation [44]; 22 rats/ha [45]). The reason for this is unclear, but may reflect the 2019 mast year [27]. At both sites, fruit was observed in greater abundance in 2019 than 2020 or 2021, with the standard treatment site displaying a greater diversity of fruit than the 1080-to-zero treatment site. Masting has been observed to elevate rat abundance in beech and southern rimu forests [10,44,46], although the effect in North Island mixed podocarp forests is less clear [42,47].Both the 1080-to-zero and standard 1080 treatments achieved a >95% reduction in rat densities and a reduction of wax tag detections to zero, on par with common benchmarks for success (>80% [48] or >90% reduction in abundance [29]; <5% post-treatment relative abundance indices [10]). Rat density then recovered rapidly before plateauing, suggesting a stabilisation of the populations and a final recovery time of 9 and 14 months since the 1080-to-zero and standard applications, respectively. This recovery time is within the mid-range observed for this habitat type, with previous studies reporting rat populations recovering within 4 months [49] to 2.5 years [47]. Interestingly, according to the final density model, rat movement patterns did not change with the application of 1080 or across the wide range of densities observed. In other mammal species, home range is known to change with density [50]. Black rats are known to display varying home ranges between sites, although it is unclear to what extent this is an influence of density or food abundance [51,52]. However, in this study there may have been too few captures at low densities to reliably detect a change in movement activities.A combination of live traps, tracking tunnels and wax tags was used to improve the detection of survivors. This redundancy proved useful, with tracking tunnels detecting individuals when traps and wax tags did not. The apparent sensitivity of tracking tunnels contrasted with concerns from others, who observed that some rats will avoid entering them [31]. Despite the high sensitivity observed here, it is likely there will always be some individuals that avoid devices and will only be detected by devices that do not require interaction (e.g., cameras) [32,33]. Interestingly, the modifications to traps between the first and second sessions at the 1080-to-zero site showed strong effects on capture probability. Covering traps with brush may reduce neophobic avoidance and be a preferred trapping method when monitoring low-density or trap-avoidant rat populations [53] (but see [54,55]). Overall, the relative abundance indices obtained in this study appeared highly correlated to density. However, there was evidence that this relationship may wane at very high densities through the saturation of devices. The wax tags also tended to return false absences. This is consistent with other studies that show strong correlation of relative abundance indices with density and trapping rates at moderate densities [35,37], but poor correlation at very high or low densities [56,57].Overall, monitoring indicated that the 1080-to-zero applications did not achieve rat elimination, with at least one rat captured 3 weeks after each 1080-to-zero application. However, by itself, this does not necessarily indicate a failure to meet the objectives. A secondary goal of the 1080-to-zero application, where it fails to eliminate pests, is to reduce them to near-zero levels that can potentially be eradicated, afterwards, with more targeted methods, such as spot toxin treatment [21,58]. However, the results obtained here cannot be construed as near-zero, with post-treatment densities of approximately 0.7 rats/ha translating to approximately 1540 rats across the 2200 ha Kaitake Ranges. The second 1080-to-zero application did not improve circumstances, with at least one confirmed survivor and similar post-treatment densities to the first application. Ultimately, the 1080-to-zero treatment did not appear to improve rat suppression relative to the standard method. Post-1080 rat densities were similar at the two sites and the final model did not support a difference in recovery rate. One out of the three most supported models suggested a slower recovery at the standard treatment site, but it seems unlikely that this difference in recovery is meaningful. Even if it were, it would indicate an improved outcome at the standard treatment site.1080 operations can be influenced by local factors [10], and the possibility that site differences may have masked an improved impact from the 1080-to-zero operation must be considered. The only differences highlighted by the final density model were a higher pre-treatment density and lower capture probability at the standard site than the 1080-to-zero site. High pest densities are sometimes associated with worse outcomes for 1080 operations [10,48], and a low capture probability could indicate neophobic behaviour that translates to reduced bait consumption [55,59]. However, if these factors were influencing the 1080 outcome, this should have only worsened the outcome at the standard site, and so they are unlikely to have influenced our conclusions. Thus, although clear differences among the sites existed, for the parameters we measured they should not have predisposed the 1080-to-zero site to a higher failure rate. However, that is not to say that some other site-specific factors may not have contributed to the unsuccessful outcome of the 1080-to-zero application.A key question in pest elimination monitoring is whether the individuals observed soon after treatment are survivors or invaders. With no barrier to movement, the primary protection against reinvasion was the distance from the study site to the treatment boundary (837 m for the 1080-to-zero site). The invasion behaviour of black rats is still not well understood. While some studies indicate that movement over this scale could take weeks to months [9,24,47], others suggest it could take days [60]. Reinvasion could not be ruled out in this study, as no rats captured prior to the first 1080-to-zero application were recaptured later. However, only 65% of the population was captured before treatment. Additionally, at least one individual survived the second 1080-to-zero application, having been previously captured after the first application. Furthermore, rats are known to display varying responses to management devices, and rats which do not enter traps may also be pre-disposed to avoid toxic bait and so survive 1080 applications [61]. Future studies may benefit from supplementing mark-recapture monitoring with non-invasive detection devices (e.g., trail cameras) in the first few weeks immediately after treatment [54,62].The results seen here differ from the only other large-scale 1080-to-zero operation completed so far: a South Island operation which did not achieve full removal of rats, but did reduce them to near-zero levels [21]. That operation also saw improved rat removal from a second toxin application [21], something echoed in prior studies which explored double-1080 applications using standard application methods [9]. It is not unheard of for similar 1080 operations to produce variable outcomes from site to site for no apparent reason [10]. The results observed here highlight the complexity of aerial toxin applications, and the need for repeated testing to assess the variability in performance across a range of environments.Without replication, our study cannot generalise which factors may influence elimination success for 1080-to-zero operations, but can point to key features that should be considered in future research. The study sites showed features associated with difficult-to-manage locations. This includes high food abundance [63], high pest density and rugged terrain [10]. The influence of these factors on 1080 operations is poorly understood, although they are associated with a higher risk of failure [10,63]. Additionally, the timing of the operations was not optimal. The first 1080-to-zero application was performed in autumn following a large mast fruiting event, with fruit visible on the ground. The high food abundance [63] and the application of bait in summer or autumn instead of winter or spring can reduce bait uptake [64]. The second 1080-to-zero application also had long delays between the two pre-feeds (82 days) and pre-feed and toxin (43 days). These were longer than recommended (5 to 14 days in mast years [29]) or used in other operations (e.g., 5 to 10 days [22]; 26 days [24]; 15 to 28 days [21]). Pre-feeding is critical in a double toxin application as a mechanism for reversing bait shyness in survivors of previous treatments [9]. Whether such a delay would have impacted the efficacy of the pre-feed is unclear. Future studies should continue to investigate the influence of factors such as habitat, pest density, food abundance and operational timing on 1080-to-zero and other eradication operations [65].Finally, further modification of the 1080-to-zero method may improve results. In particular, while bait switching between the first and second application is a standard method for reducing bait shyness [21], Nugent et al. recently found that this is most effective if individuals are exposed to both baits before being exposed to toxin [66]. Operations may thus benefit from pre-feeding with both bait types before any toxin is applied. Additionally, there is evidence that targeted bait deployment can improve bait distribution and pest reduction [48]. Although targeted bait deployment was initially explored as a method for reducing bait sow rate while maintaining adequate rat suppression, improved results may be seen from combining it with the increased sow rate of the 1080-to-zero method. This may be especially relevant for rugged terrain, such as that of the Kaitake 1080-to-zero treatment area, where the uneven topography may inhibit even bait distribution [10].5. ConclusionsThis study assessed the short- and long-term outcomes on rat density of a novel eradication 1080-to-zero application technique, in comparison to a standard 1080 application. In these Taranaki study sites, the novel method did not appear to achieve elimination of rats nor improve results relative to the standard method. However, with only one 1080-to-zero treatment site, this study cannot make generalised statements regarding the efficacy of this tool. Operations using 1080 can produce variable outcomes, often for unknown reasons, and the results here differ from those in a similar South Island operation. Future studies should investigate how habitat and operational design influence the efficacy of this and other eradication tools.

animals : an open access journal from mdpi

10.3390/ani13111858

PMC10252004

Despite studies conducted with other seaweed species on in vitro gas production, until now, no studies have been reported on the use of pelagic Sargassum as potential ruminant feed. Thus, the objective of the present study was to determine the effect of Sargassum inclusion, using tropical grass as substrate, on in vitro gas production kinetics. Additionally, heavy metals and macro- and microminerals were determined in Sargassum. For that, in vitro incubations were performed with different levels of Sargassum inclusion on a basal substrate (Stargrass hay). In vitro results showed that up to 30% pelagic Sargassum could be included in hay-based substrates from tropical grasses.

This study determined the effect of pelagic Sargassum on in vitro dry matter and organic matter degradation, total gas production (TGP), and protozoa population. The treatments were different levels of Sargassum inclusion on a basal substrate (Stargrass hay; Cynodon nlemfuensis) as follows: T0 (control treatment based on Stargrass hay), T10 (90% Stargrass hay + 10% Sargassum), T20 (80% Stargrass hay + 20% Sargassum), and T30 (70% Stargrass hay + 30% Sargassum). Ruminal fermentation kinetics and protozoa population were determined during 72 h of in vitro incubations. Compared to control, dry matter degradability at 48 and 72 h and organic matter degradability at 24 and 48 h were higher in Sargassum treatments. TGP was lower with T20 at 48 h. The total population of protozoa and the concentration of Entodinium spp. were lower at T20 at 48 h and T30 at 72 h. Cl, S, Ca, K, and Zn (103, 5.97, 88.73, 285.70 g/kg, and 15,900 mg/kg) were high in Sargassum, reaching twice or even nine times higher than the contents in Stargrass (11.37, 1.60, 43.53, 87.73 g/kg, and 866.67 mg/kg). Overall, up to 30% pelagic Sargassum could be included in hay-based substrates from tropical grasses without negative effects on in vitro dry matter and organic matter degradability.

1. IntroductionThe growing animal protein demand driven by the increase in the world population, and the adverse effects of change climate, put pressure on the agricultural production systems to increase its productive efficiency [1]. In relation to the negative impacts on the environment, greenhouse gas emissions (GHG) are the most related to ruminant production systems [2]. Within the GHG generated from livestock, the most important are methane (CH4), carbon dioxide (CO2), ammonium (NH4+), and nitrous oxide (N2O) [3,4,5]. GHG produced in the rumen can be mitigated through feeding strategies which, on the one hand, improve the digestibility of organic matter (OM); and on the other, modify the rumen microbiome [2,6,7] using secondary metabolites, by-products, and essentials oils.In the Mexican Caribbean, one possible alternative is the use of seaweed as a source of bioactive substances that could help to reduce CH4 production through modulation of the ruminal microbiome [1,8,9]. Sargassum is found in relatively large amounts in the Mexican coasts of the Yucatan Peninsula, which represent a serious problem for the tourism sector, and this has led to the development of projects aiming to make use of this pollutant [10,11,12]. In this regard, pelagic Sargassum that reaches the coasts of the Mexican Caribbean is mainly composed by two species of algae brown: Sargassum natans and Sargassum fluitans [13]. These marine species could be alternative feeds that could improve OM fermentation in ruminants, as shown in in vitro studies with other species of marine algae [14,15,16]. The nutritional value of seaweed is related to its content of minerals, polysaccharides, and phenolics compounds such as phlorotannin that can modify ruminal microbiome toward reducing CH4 production [17,18,19,20,21]. Some in vitro studies conducted with brown and red seaweed have shown improvements on fermentation characteristics and reductions in CH4 [20,22,23,24]. However, until now, studies have not been conducted on the use of pelagic Sargassum as potential ruminant feed. Based on the above, the objective of the present study was to determine the effect of Sargassum inclusion, using a tropical grass as substrate, on in vitro gas production kinetics. Additionally, heavy metals and macro- and microminerals were determined in Sargassum.2. Materials and Methods2.1. LocationThe study was carried out in the digestive physiology laboratory of the Technological Institute of Conkal, Yucatan, Mexico. Located at coordinates 21°04′45.9″ N 89°29′57.7″ W, at 7 m above sea level, with an Aw0 climate according to the Köppen climate classification, as modified by García (1988). The highest proportion of rainfall occurs during the months of June to October, with 900 mm of precipitation, an average annual temperature of 29 °C, and annual average relative humidity between 66 and 89%.2.2. Management of Donor AnimalsThe animals were handled according to the animal handling and welfare standards of the Technological Institute of Conkal (project 15135). Five male lambs of the Pelibuey breed (four years of age) with a body weight of 40 ± 3 kg were used as donors of rumen contents. Lambs were fed solely on Stargrass (Cynodon nlemfuensis) from grazing paddocks. For 15 consecutive days, grazing had a daily duration of 8 h (from 8:00 am to 4:00 pm). After grazing, animals were housed in individual roofed pens (3 × 3 m) with free access to water. At the beginning of the adaptation period, the sheep were dewormed with Closantel 5%® (Wyeth LLC, Madison, NJ, USA) at a dose of 10 mg/kg of body weight.2.3. Experimental Design and TreatmentsA completely randomized design with four treatments and four replications was used. In vitro gas production (IVGP) was recorded at 0, 3, 6, 9, 12, 24, 36, 48, and 72 h of incubation; while pH and ruminal protozoa population, dry matter degradation, and organic matter degradation, were analyzed at 24, 48, and 72 h of incubation. The treatments consisted of different levels of Sargassum inclusion on a basal substrate (Stargrass hay; Cynodon nlemfuensis) as follows: T0 (control treatment based on Stargrass hay); T10 (90% Stargrass hay + 10% Sargassum); T20 (80% Stargrass hay + 20% Sargassum); and T30 (70% Stargrass hay + 30% Sargassum).2.4. Sample PreparationThe Sargassum was collected in September 2021 on the beach of San Miguelito, Municipality of Cancun, State of Quintana Roo, Mexico. Subsequently, the Sargassum was cut into small pieces and dried naturally in the shade and then ground in a Wiley mill (Thomas Wiley Laboratory Mill, Swedesboro, NJ, USA) with a sieve size of 2 mm to convert it into flour. The Stargrass was harvested on day 28 of growth and immediately dried in a forced air oven at 60 °C for 24 h and ground in a Wiley mill (Thomas Wiley Laboratory Mill, Swedesboro, NJ, USA) with a sieve size of 2 mm.2.5. In Vitro TrialRumen contents (solid to semisolid phases) were obtained through an oesophageal probe as described by Ramos-Morales et al. [25] at 08:30 h before grazing. This was performed to reduce the variation of inoculum composition and activity and to minimize the influence of the diet fed to the donor animals [26]. Contents were kept in thermos at a constant temperature of 39 °C to be transferred to the digestive physiology laboratory where they were filtered through four layers of gauze to obtain only the liquid fraction, which was saturated with CO2 when mixed with reduced mineral solutions according to Menke and Steingass [26].The tropical grass was used as the basal substrate in the four treatments with different percentages of inclusion of Sargassum that resulted from the mixtures of Stargrass hay and Sargassum. One gram of the mixture resulting from each treatment was placed in 48 amber glass bottles (n = 4 times), with a capacity of 100 mL [26]. In addition, 24 flasks were used as blanks (only with ruminal fluid) to correct for gas production. Once the vials with the samples were prepared, they were filled with the inoculum and sealed with their respective aluminum rings and rubber stoppers to be incubated in a water bath at 39 °C, and the pressure and gas volume readings were recorded at 3, 6, 9, 12, 24, 36, 48, and 72 h, in addition to the variables described below.2.6. In Vitro Gas ProductionThe volume of gas generated (mL/g of incubated DM and OM) was measured according to the procedure proposed by Theodorou et al. [27], namely:V = (P − 21.016)/16.132, where: V = volume of gas (mL);P = measured pressure (psi).To measure pressure changes, a pressure transducer was used. The kinetics of gas production was evaluated using the Gompertz model [28]:Y = A exp {− exp [1 + be/A (LAG − t)]}, where:Y = Cumulative total gas production (mL);A = Theoretical maximum gas production (mL);b = Maximum gas production rate (mL/h), which occurs at the inflection point of the curve;LAG = Lag time (h), defined as the time axis intercept of the tangent line at the inflection point;t = time.Parameters a, b, and LAG were estimated by means of a non-linear regression analysis, for which the Origin 8 program was used. These parameters were used to evaluate the kinetics of gas production in vitro according to the methodology described by Machado et al. [26].2.7. Fermentation ParametersThe inoculum was sampled to analyze pH and protozoa count at 24, 48, and 72 h of incubation. The pH was measured with a pH meter (ECOTESTR, Thermo Scientific Eutech Instruments, Mundelein, IL, USA). For protozoa count, 2 mL of the inoculum was taken after 24, 48, and 72 h of incubation and mixed with 2 mL of methyl green formalin saline solution, composed by 100 mL of 35% formaldehyde solution, 0.6 g of methyl green, 8.0 g of NaCl, and 900 mL of distilled water [29]. The protozoa were counted under the microscope using a counting chamber (Neubauer, Fuchs-Rosenthal). Each sample was counted six times, and when the mean of the repetitions differed by more than 10%, the counts were repeated.2.8. DegradabilityIn vitro degradability of DM and OM was estimated using the digestion nylon bags method (Dacron® fabric). After incubation, the nylon bag with the substrate was washed three times and then dried in an oven at 60 °C for 48 h. Subsequently, the bags were weighed to obtain the dry weight of the remaining sample and to obtain the DM degradability. The following formula was applied according to Choi et al. [20]:Dry matter degradability g kg−1DM=IDM weight−FDM weightIDM weight×1000, where:IDM = initial dry matter;FDM = final dry matter.To determine OM degradability, the remaining samples of the nylon bags were incinerated in a muffle at 560 °C for 8 h to obtain the ash content and determine the OM of each incubated sample. Subsequently, the same formula was applied to calculate the OM degradability developed by Choi et al. [20]. Organic matter degratability g kg−1DM=IOM weight−FOM weightIOM weight×1000, where:IOM = initial organic matter;FOM = final organic matter after incubation.2.9. Chemical AnalysisContents of dry matter (DM; method 934.01), crude protein (CP; method 954.01), crude fiber (CF; method 962.09), ethereal extract (EE; method 920.39), ash (AC; method 942.05) of Sargassum, and Stargrass hay were determined as described by the AOAC [30]. Neutral detergent fiber (NDF) and acid detergent fiber (ADF) were determined according to the procedures of Van Soest et al. [31]. Non-fibrous carbohydrates (NFC) were calculated as 100 − (CP + NDF + EE + Ash).The determination and quantification of macronutrients (Cl, P, K, S, Ca, Mg, and Na), micronutrients (Mn, Fe, Zn, Cu, Al, and Si), and heavy metals (As) from Sargassum and Stargrass was performed by μ-X-Ray Fluorescence (μ-XRF) analysis, with the methodologies described by Morales-Morales et al. [32], using an M4Tornado 100 equipment (Bruker, Germany). The chemical and mineral composition of Sargassum and Stargrass are described in Table 1, Table 2 and Table 3.2.10. Statistical AnalysisData were analyzed by one-way analysis of variance (ANOVA), with the PROC GLM procedure for a completely randomized design in the SAS statistical software (SAS, 1999). In vitro gas production kinetics were analyzed using the Gompertz model [28]. Tukey’s test was used for comparisons of means between treatments. The results were considered statistically significant at a value of p < 0.05.3. Results3.1. Chemical Composition of Sargassum and StargrassChemical composition revealed differences between the Stargrass hay and Sargassum mostly in CP contents (10.48 vs. 6.73%) and fiber fractions such as NDF (76.70 vs. 23.12%), and ADF (41.61 vs. 17.18%) (Table 1).When mixing Stargrass hay and Sargassum (Table 2), it was observed that the concentration of ash increased as the level of Sargassum increased from 7.46 to 17.42%, and vice versa in the concentration of CP, OM, EE, NDF, and ADF, which were reduced as the inclusion of Sargassum increased from 10.48 to 9.36%; from 92.54 to 82.58%; from 1.34 to 1.11%; from 76.70 to 60.63%; and from 41.61 to 34.28%, respectively (Table 1 and Table 2).animals-13-01858-t001_Table 1Table 1Chemical composition of Sargassum and Stargrass.IngredientsDM (%)Ash (%)OM (%)CP (%)NDF (%)ADF (%)EE (%)Lignin (%)NFC (%)TFCTTStargrass93.467.4792.5310.4876.7041.611.34ND4.01NDNDSargassum87.7640.0759.936.7323.1217.180.5720.3529.510.070.04NFC: non-fibrous carbohydrates = 100 − (CP + NDF + EE + Ash); TFC: Total phenolic compounds; TT: tannins totals; ND: not determined.animals-13-01858-t002_Table 2Table 2Ingredients and chemical composition of dietary treatments.ComponentsTreatmentsTCT10T20T30Inclusion (%) Stargrass100908070 Sargassum 0102030Chemical composition (%) Dry matter93.4692.8792.5691.67Organic matter92.5488.7187.4482.58Ash7.4611.2912.5617.42Cru de protein10.4810.109.739.36Ether extract1.341.261.191.11Neutral detergent fiber76.7071.3465.9860.63Acid detergent fiber41.6139.1736.7234.28Differences were observed in the concentrations of macro and micro minerals. Cl, S, Ca, K, and Zn (103, 5.97, 88.73, and 285.70 g/kg and 15,900 mg/kg) were high in Sargassum, reaching twice or even nine times higher than the contents in Stargrass (11.37, 1.60, 43.53, and 87.73 g/kg and 866.67 mg/kg). The elements Al and As (500 and 530 mg/kg) were only found in Sargassum (Table 3).animals-13-01858-t003_Table 3Table 3Macro and micro mineral contents of Stargrass and pelagic Sargassum. IngredientStargrass Sargassum Macrominerals (g/kg)Cl11.37103.20Na6.901.80Mg0.900.0S1.605.97Ca43.5388.73P0.600.0K87.73285.70Microminerals (mg/kg)Fe10,033.331600Mn233.33100.00Zn866.6715,900Cu66.670.0Al0.0500Si3333.331266.67Heavy metals (mg/kg)As-530FNDNDCdNDNDCrNDNDPbNDNDHgNDNDAll values represent the mean of triplicates. NA: not determined.3.2. Total Gas Production and Characteristics of In Vitro FermentationThe pH was similar between treatments at 24 h of incubation (p = 0.0875). At 48 h of incubation, there was a dose-dependent effect, where the lowest pH value (p < 0.0001) was recorded at T30; while at 72 h, the lowest value (p = 0.0507) was observed in T20 compared to the other levels containing Sargassum, but they were similar to the control values (TC). Despite the differences in the pH values reported in the present study, they are still within the optimal range (5.66–7.47) (Table 4).The total gas production (TGP) had differences at 48 h (p = 0.0137), and the lowest production was found at T20 with respect to TC (127.56 vs. 107.77 mL/g DM) (Table 5 and Figure 1). The DM degradability (DMD) was affected by the addition of Sargassum at 24, 48, and 72 h of incubation (p = 0.0105; p = 0.0056; p = 0.0055). At 24, the treatments T10 and T30% registered the highest DMD (p = 0.0105) with respect to the control (32.22 and 33.90 vs. 24.67%). At 48 h of incubation, T30 was the one that showed the highest (p = 0.0056) degradability (37.91 vs. 29.36%), and the other treatments were similar to each other and to the control. At 72 h of incubation, T20 and T30 had a higher DMD (p = 0.0055), while T10 was similar to the control. Regarding the OM degradability (OMD), differences were observed at 24 and 48 h of incubation (p = 0.0487, p = 0.0141); the highest values were recorded at T20. At 72 h; and no differences were observed (p = 0.2077) in OMD.3.3. Protozoa PopulationTotal protozoa population was similar between treatments at 24 h of incubation (p = 0.1215), while at 48 and 72 h, there was an observed effect of the treatments (p = 0.0291; p = 0.0130), namely, T20 at 48 h and T30 at 72 h, i.e., those that showed the lowest concentrations of protozoa (Table 4).4. Discussion4.1. Total Gas Production and In Vitro Fermentation KineticsThe inclusion of algae in ruminant diets modifies digestion, fermentation kinetics, proteolysis, and nitrogen metabolism, which causes changes in rumen microbial communities [8,16,22,33,34,35]. In this regard, ruminal pH is considered the main factor that influences microbiome and the degradability of DM and OM, the concentration of NH3, and molar proportions of volatile fatty acids [9,20]. In this study, although the pH was affected by Sargassum inclusion (p < 0.0001; p = 0.0507), the values obtained were in the optimal range for adequate microbial growth (5.5–7.5) (Table 5). Therefore, the inclusion of Sargassum in the basal substrate could provide a stable and adequate environment for rumen microorganisms’ growth [23,35,36]. Compared with control, at 48 h of incubation, the highest pH (p < 0.0001) was obtained with T10; however, all other values were above a pH of 6, results that do not coincide with those reported in other studies evaluating Sargassum fusiform and Sargassum fulvellum at 10% inclusion to a substrate based on Timothy grass (Phleum pretense), where no differences in pH were observed [20,35]. On the other hand, Choi et al. [23] evaluated extracts of five species of algae added at a level of 5% and found that the highest pH at 72 h was obtained with Sargassum fusiform, and pH values recorded at 24, 48, and 72 h were above 6, which coincides with those obtained in the present study.Total gas production is related to substrate degradability, VFA production, and microbial growth in the rumen [37,38,39]. Therefore, the addition of algae in the substrate is directly related to an increase in the populations of bacterial species such as Fibrobacter succinogenes, and Ruminococcus flavefaciens, which are responsible for degrading dietary fiber [23,24]. In this study, a reduction in TGP was observed with T20 at 48 h incubation (p = 0.0137). However, it did not influence DMD, and this was affected by the level of Sargassum in the basal substrate, with T30 being the one that showed the highest DMD at all incubation times (p = 0.0105; p = 0.0056; p = 0.0055), showing that there was a linear trend for DMD with the increase in the inclusion of Sargassum. These results coincide with those of Widiawati and Hikmawan [40], who observed a linear increase in DMD with increasing inclusion at 48 h of incubation of Eucheuma cottonii.The results of the present study also agree with those obtained by Choi et al. [15], who observed a dose-dependent effect on DMD with the inclusion of increasing levels of U. pinnatifida to a Timothy grass-based substrate. For their part, Choi et al. [24] argued that the greater degradability of the DM obtained in treatments with brown algae was due to the increase in the abundance of fibrolytic bacterial populations. In this study, T10 at 48 and 72 h of incubation did not show differences compared with the control, and these results are similar to those reported by Choi et al. [20] and Choi et al. [35] with Sargassum fusiform and Sargassum fulvellum at 10% inclusion, where they found no differences in DMD compared to the control. Maia et al. [16] evaluated three species of algae (Ulva rigida, Gracilaria vermiculophylla, and Saccharina latissima) as supplements added at 25% to a basal substrate and reported increases in DMD with all algae compared to the control. However, the findings of Rjiba-Ktita et al. [41], with increasing inclusion levels of green algae up to 40% to concentrated feed as substrate, yielded a linear decrease in DM degradability with increasing inclusion levels of Ulva lactuca and Chaetomorpha linum.The OMD was influenced by the addition of Sargassum to the basal substrate; however, unlike the DMD, the highest value was observed at T20 (p = 0.0141). Some studies evaluating the composition and chemical characterization of pelagic Sargassum reported ash and OM contents of 46.94 and 30.61%, respectively [24,39]. In this study, the ash concentration and consequently low OM level of Sargassum was expected to negatively affect the OMD, which contains 40.07 and 59.93% ash and OM, respectively (Table 1). However, this situation did not occur; therefore, it is possible that high levels (20–30% DM) of inclusion of this unconventional ingredient can be used under in vitro conditions. The results observed in this study agree with those reported by Maia et al. [16], who included 25% Saccharina latissima to a mixed total ration where OMD was increased.For their part, Widiawati and Hikmawan [40] observed a linear increase in OMD with increasing inclusion of Eucheuma cottonii at 48 h of incubation added to a substrate based on elephant grass (Pennisetum purpureum). The improvement in the digestibility of nutrients is due to the bioactive compounds (i.e., polysaccharides such as fucoidan, alginate, laminarin and mannitol) from brown algae, which can favor changes in metabolic pathways with an increase in fibrolytic activity [20,23]. These polysaccharides cause changes in ruminal microbiome because they are highly available for microbial growth and favor the production of acetate and butyrate, which are directly related to fiber degradation in the rumen [15,35]. In this sense, the degradability of nutrients is influenced by algal species [37]. This species effect is directly related to the chemical composition and secondary metabolite content of the algae, which includes polysaccharides, polyphenolic compounds, halogenated compounds, minerals, and fatty acids, which confer various biological properties [14,16,33,41,42,43,44,45].The composition of the cell wall influences in vitro degradability, apparent digestibility, and availability of nutrients from algae as feed for ruminants [46]; however, there is also an interaction effect between algae species and harvesting season on in vitro nutrient degradability [38]. Regarding pelagic Sargassum, according to the study by Saldarriaga-Hernandez et al. [47], the harvest season influences the composition of carbohydrates, proteins, and total phenolic compounds due to several factors, such as light intensity and solar radiation, that affect the growth of these algae.This study was carried out with the purpose of using non-conventional additives or ingredients in ruminant feed and verifying their effects on rumen fermentation kinetics. The results of this research contribute valuable information to reduce dependence on grains for formulating diets, which would reduce production costs [1,5,8,48,49,50,51,52]. However, it is convenient to carefully analyze the level of inclusion of these ingredients according to the species of seaweed, since the purpose is not to negatively affect fermentation parameters that could affect productivity and animal performance. This is related to changes in metabolic hydrogen (H) fluxes in ruminal fermentation and in the post-absorption metabolism of the animal, caused by CH4 inhibitors [53,54,55]. In ruminants, reductions in the intake and digestibility of diets with seaweed have been reported; this is due to the increase in the mineral contents of the diet and the increase in H levels in the rumen due to the inhibition of methanogenesis [56,57,58].4.2. Population of ProtozoaMarine macroalgae contain a wide variety of bioactive compounds depending on the species; among these, we can highlight bromoform and phlorotannin that have antimicrobial properties [1,21,59,60]. According to Choi et al. [23], phlorotannin modify the abundances of cellulolytic bacteria, methanogenic archaea, and methanogens associated with ciliate protozoa. The antibacterial mechanism of action of these phenolic compounds is mediated by their ability to affect cell wall permeability. Phlorotannin change the shape of the bacterial cell membrane, leading to cell lysis; nevertheless; they also suppress bacterial reproduction through their union with bacterial proteins, RNA and DNA, inhibiting cell replication [61]. In this study, effects were observed at 48 and 72 h of incubation on the population of protozoa, which decreased in treatments T20 and T30 (Table 5). Although the reduction of protozoa was not linear, the inhibitory effect of brown algae on the protozoa population of ruminal fluid was demonstrated. These results agree with those obtained by Choi et al. [20], who evaluated Sargassum fusiform at levels of 1 to 10%, and with those reported by Prayitno and Hidayat [62], who evaluated Sargassum sp. at levels of 1 to 5%, which were accompanied by a CH4 reduction of more than 40 and 80%, respectively. They also agree with the study by Belanche et al. [14] when evaluating Ascophyllum nodosum and Laminaria digitata added to 5% DM, which showed reductions in the concentration of methanogens and protozoa without affecting the bacterial population and anaerobic fungi.Other studies with red algae have reported similar results, such as the study by Widiawati and Hikmawan [40], which evaluated Eucheuma cottonii with inclusion levels of 4, 8, and 12%. They observed a linear decrease in the population of protozoa and in the concentration of CH4 in the ruminal liquid with the increase in the dose. Roque et al. [8] reported reductions in the abundance of methanogens in ruminal fluid and, consequently, a 95% reduction in CH4 production with the addition of 5% Asparagopsis taxiformis to a good quality substrate. Contrary to these studies, Molina-Alcaide et al. [37], when evaluating various species of red algae (Mastocarpus stellatus, Palmaria palmata, and Porphyra sp.), reported no effects on the microbial population.Some studies with brown seaweed extracts have reported differential effects, for example, Choi et al. [24] reported an increase in the population of ciliate protozoa with Sargassum fusiform, while with Undaria pinnatifida and Sargassum fulvellum, there was a decrease. This effect was also reported by Belanche et al. [18], who demonstrated antiprotozoal activity with Ascophyllum nodosum compared to Laminaria digitata which did not show antiprotozoal effect. These differential effects are probably due to differences in the number of polyphenolic compounds, especially in the concentration of phlorotannin present in each species [14]. Based on the above, and on the results obtained in this study, it is important to pay special attention to the concentration of polyphenols, the chemical structure, and the molecular weight of phlorotannin in the different species of brown algae to know their mode of action [61]. In addition to the above, the characterization of the sulfated polysaccharides from pelagic Sargassum, based on their bioactive properties in ruminants, is another point of interest since there are no studies in this regard so far.For the reasons stated, it is necessary to carry out more studies to determine the composition of fatty acids, minerals, potentially toxic compounds, and secondary metabolites in pelagic Sargassum, and their effects on the microbial communities of the rumen. Furthermore, further in vitro work with different base substrates and inclusion levels is of vital importance [63,64]. Moreover, long-term in vivo studies are required with different species of ruminants, and with different feeding regimens to verify and rule out harmful effects of Sargassum on the health of animals and on the quality of meat and milk [49,56,57,58,65,66,67,68,69]. The findings and their implications should be discussed in the broadest possible context. Future research directions may also be highlighted.5. ConclusionsPelagic Sargassum has great potential as ruminant feed. The results of this in vitro study showed that the inclusion of up to 30% pelagic Sargassum in hay-based substrates from tropical grasses does not have negative effects on rumen fermentation kinetics, nor on the degradability of dry matter and organic matter. The use of this unconventional natural resource in ruminant production systems would have important economic benefits since it provides a route for the management of marine algae residues in the Mexican Caribbean. This would reduce the negative impact of pelagic Sargassum on the tourism sector, which is the primary source of income for families living in is region.

animals : an open access journal from mdpi

10.3390/ani12030298

PMC8833732

The domestic cat is the only species within the felis genus to have transitioned from a wild, solitary species to one of the most popular human-companion animals globally. In stark contrast to their closest wild ancestors, the domestic cat displays an impressive capacity to cohabit successfully with both humans and other cats. However, at an individual level, domestic cats demonstrate substantial variability in their sociability towards both species. Such variability may be influenced by a range of factors including their early life experiences, genetic selection, and individual cat and human characteristics, in addition to various factors associated with their social and physical environment. The impact of these factors may have important implications regarding a cat’s social relationships, their adaptability to various social contexts, and, ultimately, their wellbeing. In line with modern pet-keeping practices, domestic cats may often be exposed to lifestyles which present a range of complex social and environmental challenges, although it is unclear how much cats have been selected by humans for traits that support adaptability to such lifestyles. This review aims to summarise what is currently known about the various factors that may influence domestic cats’ sociality and sociability towards both humans and cats, with a predominant focus on populations managed by humans in confined environments. Current limitations, knowledge gaps, and implications for cat wellbeing are also discussed.

Sociality can be broadly defined as the ability and tendency of individuals to reside in social groups with either conspecifics and/or other species. More specifically, sociability relates to the ability and tendency of individuals to display affiliative behaviours in such contexts. The domestic cat is one of the most globally popular companion animals and occupies a diverse range of lifestyles. Despite an arguably short period of domestication from an asocial progenitor, the domestic cat demonstrates an impressive capacity for both intra- and interspecific sociality and sociability. At the same time, however, large populations of domestic cats maintain various degrees of behavioural and reproductive autonomy and are capable of occupying solitary lifestyles away from humans and/or conspecifics. Within social groups, individuals can also vary in their tendency to engage in both affiliative and agonistic interactions, and this interindividual variation is present within free-living populations as well as those managed in confined environments by humans. Considerable scientific enquiry has focused on cats’ social behaviour towards humans (and conspecifics to a much lesser extent) in this latter context. Ontogeny and human selection, in addition to a range of proximate factors including social and environmental parameters and individual cat and human characteristics, have been highlighted as important moderators of cats’ sociability. Such factors may have important consequences regarding individuals’ adaptability to the diverse range of lifestyles that they may occupy. Where limitations to individuals’ social capacities do not enable sufficient adaption, compromises to their wellbeing may occur. This is most pertinent for cats managed by humans, given that the physical and social parameters of the cats’ environment are primarily dictated by people, but that positive human-selection for traits that enhance cats’ adaptability to such lifestyles appears to be limited. However, limitations in the availability and quality of evidence and equivocal findings may impede the current understanding of the role of certain factors in relation to cat sociability and associations with cat wellbeing, although such literature gaps also present important opportunities for further study. This review aims to summarise what is currently known about the various factors that may influence domestic cats’ sociality and sociability towards both humans and conspecifics, with a predominant focus on cats managed by humans in confined environments. Current limitations, knowledge gaps, and implications for cat wellbeing are also discussed.

1. IntroductionIn a relatively brief period of evolutionary time, the domestic cat has transitioned from a wild solitary species to one of the most popular companion animals globally. During their initial domestication (from wild populations of F. silvestris lybica [1]), natural selection pressures are likely to have favoured bolder individuals, as well those with a greater tolerance to human and conspecific proximity [2]. Subsequently, as the value of cats as a source of human companionship increased, a degree of active selection by humans for cat tractability likely followed [3]. However, even within modern-day domestic cat populations, it is unclear of the strength to which (both human and conspecific) sociality has been selected (either naturally or artificially), given that (i) domestic cats may still be motivated to seek out a primarily solitary existence, and can survive under such conditions [4,5,6,7], that (ii) socialised cats from companion populations may live and/or readily interbreed with unsocialised cats from free-living populations, and additionally that (iii) the most intensive period of humans’ cat selection has occurred within the last century, with aesthetic features largely prioritised over traits that might enhance sociability towards conspecifics and/or humans [1,8,9].Despite these possible constrains to cats’ sociality, the domestic cat is still capable of residing within social groups and may actively choose to associate with conspecifics and/or humans, in each case potentially developing positive social relationships. In both free-living populations, as well as those managed by humans in confined environments, cats may display a range of affiliative behaviours. Affiliative behaviours directed towards conspecifics include vertical tail raising on approach (e.g., the ‘tail up’ signal), as well as initiating various forms of physical contact including nose touching, play, allo-grooming, allo-rubbing, tail wrapping, and sleeping and resting together or in close proximity [10,11,12]. Vocalisations such as purring and meowing between mother and offspring dyads occur frequently and are thought to serve important communicative and care solicitation functions [13,14]. Similar social behaviours are often directed towards humans during cat–human interactions [15,16,17,18,19] and cats are well documented as having the capacity to develop affiliate social relationships with people [20,21,22,23].At a species level, the domestic cat occupies a diverse range of lifestyles with varying degrees of association with conspecifics and humans [24]. Broad lifestyle categories include cats that can be described as free-living (i.e., feral, street, or stray cats) and those that are living in confined environments managed by humans (e.g., the domestic home, shelter or rehoming centre, and research facility). In many cases, close associations with conspecifics and/or humans may promote health advantages for cats. For individuals under some form of human management, these typically include the provision of primary or supplemental feeding, veterinary care, and access to warm, sheltered, and safe environments. For cats outside of human management, the benefits of living with conspecifics may include communal raising of offspring and shared access to clumped resources of value such as food and shelter. At the same time, cats are considered to be ‘socially flexible’ rather than ‘socially obligate’, meaning they possess the potential to adapt to different forms of social living, but that group living (with conspecifics and/or humans) is not necessary for their survival. Degrees of sociability may be highly variable at the individual level, even amongst those occupying similar lifestyles [22,25]. Within lifestyle categories, the availability and quality (i.e., distribution and abundance) of the cats’ physical resources can vary greatly [26,27,28], as can the characteristics (e.g., age and sex, personality, previous experiences) of the cats and/or humans with which individuals cohabit [29,30]. Variability in cats’ social behaviour towards both humans and conspecifics, as well as the impacts of cohabitation on their wellbeing, may be influenced by a range of ontogenetic [31] and genetic factors [32] and their potential interaction [33] in addition to the various social and environmental parameters associated with their specific lifestyle [12,34,35,36,37,38]. Thus, in certain instances, close associations with humans and/or conspecifics may be detrimental rather than beneficial to the cats’ wellbeing, especially where individual social flexibility or adaptability is limited [39,40]. These discussion points form the basis of this review, with the intention that such knowledge can be particularly useful in supporting optimal wellbeing outcomes for domestic cats across the various contexts where they are managed by humans. Inherent difficulties associated with the reliable, practical, and valid measurement of ‘wellbeing’ and variation in how it is defined, operationalized, and its measures subsequently interpreted, are well established [41,42,43]. As such, where evidence is considered relevant to cat wellbeing within this review, a broad definition of this term is applied to include any measures or outcomes that might potentially provide useful information about the mental and/or physical health state of a cat.2. Proximate Factors and Their Links to Conspecific Social Behaviour in Free-Living PopulationsIn free-living populations, cat density and active group living is primarily determined by resource abundance and distribution, with cat densities increasing, and groups typically forming, where resources are plentiful and localised (e.g., refuse areas, farm buildings, or supplemental feeding stations) [26,29,44]. The nature of group-living and conspecific social interactions may be influenced by a range of factors such as relatedness and familiarity, age, sex and sex ratio, and individual personality, in addition to human intervention (e.g., resource provisions and neutering). Where group living occurs, these may often be matrilineal in structure, being comprised of related females, their offspring, and immigrant adult males [29,44,45]. Agonistic interactions amongst free-living individuals are described as rare, but occurring more frequently towards unfamiliar/unrelated cats (i.e., non-group members), particularly those that are female [44,46]. Amongst group-living adults, individuals are also described as having ‘preferred associates’ with which they spend proportionality more time in proximity to and perform affiliative behaviours towards [11,29,44,47]. Amongst offspring, observations of related juveniles and kittens suggest individuals prefer to associate with conspecifics of a similar age category to themselves and littermates over non-littermates [29]. Rates of agonistic behaviour between group members have been explained as a function of variation in age and sex ratios in males, and of neuter status in both males and females. For example, in a (presumably intact) colony with larger male to female and adult-male to juvenile-male ratios, adult males were reported to initiate aggressive behaviour towards other adult males more often than other sex-age classes [29]. In contrast, in a colony where adult male to female and adult-male to juvenile-male classes were smaller, adult males were observed to initiate aggressive behaviour most frequently towards juvenile males [29]. In a study of group-living populations of female cats located around supplemental feeding stations in Israel [48], intact females were found to engage in significantly larger proportions of agonistic interactions with conspecifics (e.g., chasing, hiss, yowl and growl, threat approach, stare, and physical attacks) than neutered females. Similar associations with neuter status and social interactions were reported in populations of group-living cats in Italy [49], where the neutering of both males and females was associated with a significant decrease in rates of aggression (e.g., striking with a paw, biting, assuming threatening postures, chasing, ritualized vocal duels, and physical fighting), but also a reduction in conspecific proximity. Although differences in pre- and post-neutering frequencies of affiliative behaviours were not significantly different overall, rates of nose sniffing and rubbing were noted to decrease in frequency for most dyads, with the exception of two specific male-male dyads, where rates increased. Collectively, these results would suggest that neutering may generally decrease the tendencies for individuals to engage in agonistic interactions (and social interactions in general), although for some specific dyads, affiliative interactions may increase. Amongst free-living cats, personality differences may also influence the general behavioural styles of individuals towards conspecifics, although there is limited research within this area. In a study of unneutered male cats across several colonies [50], cats were described as having either predominantly ‘proactive’ or ‘reactive’ personalities. ‘Proactive’ males were defined as those engaging in more agonistic (e.g., threating, fighting, and chasing), but also affiliative (e.g., ‘nose sniffing, passive contact, and rubbing’), interactions with other group members. In contrast, ‘reactive’ individuals were mostly characterised by displaying avoidant (e.g., ‘avoiding, crouching, flying, hissing’) and less agonistic behaviour. 3. Proximate Factors and Their Links to Conspecific Social Behaviour in Confined PopulationsFor cat populations managed in confined environments, group-living and group composition are primarily dictated by humans, and, thus, cats have little choice over this aspect of their sociality. Despite this, cats may still display individual variation in the conspecifics they choose to associate with or avoid [51,52]. Variations in the nature of conspecific social interactions have been linked to similar factors to those highlighted in free-living populations, although these relationships and their direction of effects are not consistent across studies. For example, in a USA-based survey of cat owners that had recently introduced a cat into their household [53], the provision of outdoor access (but not cat age, sex, or cat group size) was associated with increased rates of fighting amongst cats. In contrast, a more recent (USA) study [12] based on a substantially larger population of cats and owners reported that age (i.e., younger cats), sex (i.e., female), multi-cat group size (i.e., larger), as well as the recent addition of an unfamiliar cat to the home, were all associated with increased rates of cat conflict, whilst the provision of outdoor access was not. Similar trends were reported in a UK survey [54] where again, provision of outdoor access was not predictive of increased aggression, but sex and neuter status (i.e., neutered females) were. However, it is important to note here (and elsewhere in the review where results of owner-completed surveys are mentioned), that owner reports of cat’s behaviour obtained via this method are subjective and potentially contain various sources of bias. Their results should thus be interpreted with this in mind, especially where survey data are compared between studies. Little is currently known about the impact of variations in environmental provisions on the social dynamics of multi-cat groups residing in the various different confined environments that are managed by humans. Nonetheless, the relative quantity and distribution of resources are considered central components in the management and prevention of conspecific conflict [55,56,57]. Their importance is certainly plausible, given the impact of variations in resource abundance and distribution on cat sociality in free-living populations [44]. Despite this, in previous research [12], no significant relationships were identified between the frequency of (group level) conspecific conflict and the quantity of various resources provided in the home (e.g., scratching posts, food stations and litterboxes, outdoor access, available indoor space). However, greater quantities of litter boxes and food stations were associated with increased frequencies of affiliative behaviours. While this study provides some useful ‘top level’ insights into the relationships between conspecific social behaviour and resource provision, further detailed studies on this topic are needed, and across different confined environments. These should include the consideration of pertinent variables such as resource quality (i.e., size and suitability) and relative distributions, in addition to individual cat characteristics such as age, sex, neuter status, relatedness, and personality. Investigations should also account for intra-group variation in dyadic social relationships and consider their interactions with these other factors, rather than quantifying social behaviour at the group level [12].Initial behavioural responses during introductions are considered important determinants of the future relationships between cohabiting cats. For example, the same study [12] reported greater rates of conflict and lower rates of affiliative behaviours between cats in households where owners indicated initial introductions ‘did not go well’, with similar results reported in [53]. In general, gradual methods of cat introductions are considered to help promote more amicable conspecific relationships. Such methods incorporate scent swapping between cats before progression to (slowly increased) periods of visual, and then supervised, physical access. However, while these methods are widely recommended [56,58,59], their benefits are currently based on anecdotal observations rather than empirical research. In the one study where current rates of fighting were assessed relative to methods of cats’ initial introductions (i.e., either ‘gradual’ or ‘immediate’) [53], no significant differences were reported between owners using either method. However, the author highlighted issues associated with the way that ‘gradual’ and ‘immediate’ methods were categorised, potentially resulting in less than meaningful statistical comparisons. For example, ‘gradual’ introductions encompassed a broad range of time periods, from hours to weeks, and did not include any details of the specific methods used (e.g., scent swapping or visual access initially). Further study into the potential benefits of gradual approaches during cat introductions and their impact on long-term relationships are therefore warranted. Familiarity, relatedness, and social exposure during the sensitive period (i.e., approximately between 2–7 weeks of age) have also been linked to differences in conspecific social behaviour. In observations of mostly unrelated cat dyads within the domestic home, longer lengths of cohabitation were found to negatively correlate with rates of conspecific aggression [60]. In observations of cat dyads housed together in a cattery [10], litter mates (i.e., related and together since birth) spent more time in physical contact with each other and were more likely to feed together and allo-groom, compared to unrelated pairs (that had previously lived together in a home for at least a year). Rather than simply their genetic relatedness, the authors attributed these greater rates of affiliative behaviours to littermates having been socialised to each other during their sensitive periods and then also experiencing an extended duration of cohabitation due to their remaining together into adulthood. In a large group of confined cats residing in a private property, duration of cohabitation was similarly positively associated with proximity and allo-grooming between conspecifics [61]. In this latter study, however, where duration of cohabitation was controlled, related individuals (but not necessarily littermates) performed affiliative behaviours more frequently than unrelated individuals, suggesting that both duration of cohabitation and general relatedness may be important modifiers of conspecific social behaviour. To date, it is unclear which of these factors (i.e., genetic relatedness, cohabitation during the sensitive period, total duration of cohabitation), and their interactions, may have most impact on the social behaviour of cats towards conspecifics as adults. However, it is anticipated that genetically related cats that reside together during their sensitive period and have subsequently lived together for longer periods (within a stable environment) might be more likely to have long-term affiliative social relationships with each other compared to other classes.Within the domestic home, rates of agonistic behaviour have been linked with aspects of cat’s personality. In [54] a (weak) positive relationship between owner ratings of their cat’s level of fearfulness and conspecific aggression (e.g., growls, hisses, bites, scratches) was reported. In [12] a (weak) positive relationship between cats described as ‘sedentary and shy’ and rates of conflict behaviours (e.g., ‘flee’, ‘hiss’ and ‘twitch tail’) was also identified.4. The Impact of Cohabiting with Conspecifics in Confined Environments on Cat WellbeingGiven the potential links between increased group size and conspecific conflict [12], it is logical to assume that living in larger multi-cat groups may be associated with more negative wellbeing outcomes for cats. However, systematic reviews on this topic highlight a lack of cross-study consensus within both the domestic home [38] and shelter environments [36]. In the home, for example, greater numbers of cohabiting cats have been significantly linked to greater rates of owner-reported ‘behaviour problems’ and anxiety [62] and increased house soiling [28,63,64,65], but also fewer ‘behaviour problems’ [66], lower Cat Stress Scores or CSS (a posture and behaviour based scoring system) [67,68], and less negative interactions with humans [69,70,71]. Additionally, several studies have reported no significant links between cat group size and ‘behaviour problems’ [72], house soiling [69,71], obesity [71], or physiological stress [30,73].These equivocal findings potentially arise due to several pertinent factors. Issues associated with the variability of study methodologies and limitations in their designs and analytical approaches mean that the relative effects of cat group size may not be clear and/or comparable between studies, making it difficult to reach a consensus [38]. Furthermore, the varied, multi-faceted nature of housing and husbandry conditions to which domestic cats are exposed, and subsequently studied under, in addition to the potential diversity of the individual characteristics of the cats within each multi-cat group, means that cat populations may be very demographically diverse. This may occur both across separate study populations, but also within a single comparison group, within a single study [38]. This diversity makes it difficult to isolate the impact of an individual variable, such as group size, on cat wellbeing, without considering and (where suitable) controlling for the influence of important cat- and environment-based covariates. For example, a study focused on cats in the domestic home [74] highlights the importance of individual cat characteristics and their potential moderating role in the relationship between group size and increased adrenocortical activity (thus physiological stress), measured via faecal glucocorticoid metabolites (GCM). While in this study, no main effects of cat group size on variation in GCM were detected, several age-related, within-group effects were identified. For cats housed in groups of 3–4, those aged 2 years or older had significantly higher GCM values than cats under 2 years of age. Younger cats from single cat households also had higher GCM concentrations than those housed in groups of 3–4. These findings suggest that the relative impact of cat group size on physiological stress may be age dependent, with younger cats having lower, and older cats higher, adrenocortical activity when housed in groups. However, relationships between GCM and stressors can vary depending on the nature of the stressor, [75,76], sampling method, and time period [77], making it difficult to interpret GCM values in relation animal wellbeing, particularly when these are considered in isolation to other relevant physiological and behavioral measures [77,78]. In the cattery environment, similar limitations regarding evidence quality and a lack of cross-study consensus were evident [36]. Findings from this review also highlighted the importance of both cat and environment features when investigating links between group size and cat wellbeing. For example, in one study, no significant differences were identified where Cat Stress Scores (CSS) were compared between cats housed singly or with one or two familiar conspecifics, although the CSS of these cats were significantly lower compared to cats housed in large, unfamiliar groups [79]. In another study, singly housed cats in barren conditions were found to have significantly higher CSS compared to cats housed in more enriched environments, either alone or in a large group of unfamiliar conspecifics [80]. In a third study [39] comparing CSS between cats housed singly and those housed in a group with unfamiliar cats, no differences were found for cats considered socialised towards conspecifics, although CSS were higher for cats considered unsocialised. Additionally, when ‘unsocialised’ cats were added to group housing, their presence caused the CSS of other group members to increase. Relative environmental provisions, the socialisation status of the individual, and that of cohabiting conspecifics, may therefore determine whether residing with conspecifics is more, less, or similarly stressful for the individual than being housed alone within a cattery context.5. Proximate Factors and Their Links to Cat–Human Social Interactions in Confined EnvironmentsSeveral studies have highlighted various human characteristics as important determinants of cats’ social behaviour during human-cat interactions (HCI). Observational studies taking place in the domestic home [81,82] suggested cats demonstrate preferences for social interactions with adults (particularly females) over children. These differences in cats’ responses may be explained by variations in humans’ interaction styles, given that children (in particular males) may be more likely to approach resting cats, pick them up, and follow retreating cats than adults, behaviours which are likely to be perceived as threatening by the cat, or to at least induce a degree of discomfort [81,82]. In contrast, adults (in particular women) may be more likely to vocalise to cats and crouch down to their level, postures and behaviors which may be perceived as a less threatening and more encouraging of the cat to engage in social interactions [82]. In [34] observations of HCI with adult owners in the home suggested that cats interacted with owners for longer durations when the interactions were initiated by the cat as opposed to the owner. Additionally, cats were found to be more likely to engage in interactions when initiated by the owner, if owners were generally responsive to the cats’ requests for interactions [34]. Regions of the cats’ body that are touched by humans may also impact the nature of their behavioural responses. Human stimulation to the cats’ caudal region may produce greater rates of human-directed aggression (e.g., hissing, biting, smacking, or scratching) as well as behaviours indicative of discomfort (e.g., flattening the ears, flicking and/or swishing the tail) [83,84]. In contrast, stimulation to cat’s temporal regions may lead to greater rates of affiliative behaviours (e.g., closing or half-closing the eyes, “kneading” with paws, purring, rubbing against the human, and dribbling), with stimulation to the perioral regions, flank, stomach, and back producing much greater between-individual variation in responses [83]. Understanding of the impact of human behaviour styles on cat comfort and social behaviour during HCI (incorporating findings from the aforementioned studies) was recently formalised into a set of ‘best practice’ guidelines for humans [85]. Compared to control HCI, cats were found to exhibit significantly more affiliative and positively valanced behaviour (e.g., tail waving, kneading, sniffing and rubbing, ears forwards), as well as less agonistic behaviours (e.g., hiss/growl, cuff/swipe, bite) and fewer signs of conflict (e.g., tail swishing, ears rotated/flattened, paw lift, rapid groom, head/body shake, freeze/crouch, avoid/move/turn away), when humans followed the guidelines [85]. In addition to general styles of interaction, certain facial and postural cues displayed by humans may promote more positive social responses from cats. For example, companion cats were found to approach unfamiliar humans significantly more often when they performed a ‘slow-blink sequence’ towards the cat, rather than when they adopted a neutral expression [86]. Cats were also found to spend a longer time in contact with their owners when they displayed a ‘happy’ rather than an ‘angry’ posture and facial expression [87].In a study examining the underlying structure of human-cat interactions taking place in owners’ homes [17], links with owner personality were identified. HCI were reported to be less patterned and structured where owners scored higher for the personality trait Neuroticism. In contrast, owners scoring higher in Conscientiousness had more complex styles of interactions with their cats, meaning the structure of their interactions involved a more diverse range of cat and owner behavioural elements. Owner mood has also been linked to differences in intent to interact with the cat [88,89]. For example, owners rating themselves as more anxious and touchy were found to display greater intents to initiate interaction, whilst those that were more extroverted or depressed showed less intents [89]. In the same study [89], the more extroverted and agitated the person was, the more the cat was found to approach the person during an ongoing HCI, whereas the more numb the owner reported they were, the less the cat approached. In another study by the same authors [88], cats were found to head and flak rub more during interactions with owners that reported a more depressed mood. The personality of owners has also been linked to more general, longer-term aspects of cats’ human-directed social behaviour [37], with the direction of results sharing parallels with those identified within the parent–child [90,91] and owner-dog dynamics [92,93]. In a large UK-based survey of owners and their cats [37], higher owner Neuroticism was associated with more aggressive and fearful cat behavioural styles as well as greater reported ‘problem behaviours’. In contrast, higher owner Conscientiousness, Agreeableness, and Openness were associated with less aggressive and aloof cat behavioural styles, and higher Conscientiousness with less fearful, but more gregarious, styles. However, across these various studies [17,37,88,89], specific details of the handling styles exhibited by owners during HCI and/or their associations with the cats’ behavioural responses were either not quantified or reported, making the results hard to interpret in relation to the cats’ experience and comfort during HCI. Further studies investigating the relationship between human personality/mood, their HCI styles and cats’ subsequent reactions are therefore warranted.In addition to owner personality, various characteristics of the cat and their environment may impact on the nature of their social interactions with humans. For example, in a survey of Brazilian cat owners [35], several risk factors were associated with increased rates of human-directed aggression across various contexts (i.e., when petted, when startled, during play, and when around unfamiliar people and animals). These factors included the cat living in a household described by their owners as ‘frenetic’ (i.e., busy and unpredictable), the cat being of mixed breed rather than a pedigree, being described as “disliking” rather “liking” being petted, and having “poor” relationships with other animals in the household. Cats’ individual characteristics (i.e., see further sections) may also affect their perceived value of human-social interaction relative to other non-social stimuli categories such as food and toys. For example, preference tests performed on cats within a shelter center context indicated substantial variation in their preferences, and while overall, more cats preferred interacting with humans, a substantial sub-population were found to prefer the non-social stimuli [94].6. The Impact of Cohabiting with Humans in Confined Environments on Cat WellbeingDuring typical social interactions with humans, a cats’ exposure to a single HCI that it finds aversive may induce an acute (and thus potentially short lived) negative experience (see previous section). In contrast, the repeated, frequent exposure to aversive HCIs is much more likely to induce chronic (and thus longer term) negative states within individuals, potentially leading to their compromised wellbeing. While this line of scientific enquiry remains largely uninvestigated, a preliminary study [74] in the domestic home reported that cats that were described by their owners as generally “tolerating” being stroked had higher faecal glucocorticoid metabolite levels, compared to cats that were described as actively “liking” or “disliking” being petted. Such findings highlight the potential impact of a cats’ experiences during petting on their physiological stress response, with cats “tolerating” petting potentially at greater risk of increased stress. However, further research, which considers a range of additional (suitable) behavioural and physiological measures, is warranted in order to understand the implications of aversive HCI to cats’ wellbeing more broadly. In the previous section, busier, less predictable households [35] and owners scoring higher in Neuroticism [37] were associated with greater owner reports of cat human-directed aggression [35,37] and ‘problem behaviors’ and anxiety [37]. These human-based factors have also been associated with several other cat health- and welfare-linked outcomes, supporting their potential validity as risk factors for cat wellbeing. For example, in [73], cat urinary cortisol concentrations were positively correlated with the total number of humans living in a home, as well as human density. In another study [30], cats from single households were found to have higher faecal glucocorticoid metabolites where their owners reported being more socially active with other humans. Such findings might suggest that greater human presence and levels of human-social activity within the home create more stressful environments for cats, leading to their compromised wellbeing. However, again, more detailed investigations are required to explicitly test this hypothesis.In [37], cats of more neurotic owners were reported to display greater stress-linked ‘sickness behaviours’ (a composite score representing poorer coat condition, greater frequencies of cystitis, vomiting, diarrhoea, and constipation). These cats were also more likely to have an existing medical condition, to be of an unhealthy weight, and were more likely to have either restricted or no access to the outdoors. In humans, Neuroticism has been associated with decreased empathy, more authoritarian and over-protective parenting, as well as the provision of harshly controlled, but poorly structured, environments [95,96,97]. It is therefore important to determine if similar dynamics might be present in human-cat relationships, given that harsher, more forceful, and less predictable formal handling and husbandry styles have been associated with more negative wellbeing outcomes for cats [98,99]. While comparable studies in the domestic dog [100,101] reported similar owner-Neuroticism–pet-wellbeing relationships to those in [37], the current evidence base remains primarily correlational rather than causal (and based on subjective owner reports). Therefore, further investigations examining the causal mechanisms underpinning the relationships between owner personality, human-cat interactions and cat wellbeing are needed.7. The Ontogeny of Human and Conspecific Sociability in Cats Managed by Humans in Confined EnvironmentsPositive human-social experiences during early developmental periods appear important predictors of friendliness (and its generalisation) towards humans later in life. Two studies [102,103] have investigated the effect of kittens being handled by humans during different periods of their development (e.g., from 1–5 weeks of age, or from 2–6 weeks, 3–7 weeks, or 4–8 weeks), and for different amounts of time each day (e.g., 15 min versus 40). The authors reported that kittens handled within the 2–7 week time period were generally more amenable to being handled and were quicker to approach people, as were those that were handled for longer periods each day. This 2–7 week ‘sensitive period’ was identified as the stage where kittens were deemed most receptive to social learning regarding humans, with handling commencing towards the later stages of this period reported to produce less effective results. Increased handling following a specific ‘socialisation and habituation’ programme (as opposed to more basic, limited handling) within a shelter context was also found to produce cats that were reported as displaying less fear-based behaviours towards their adoptive owners, as well as providing them with more ‘social support’ when assessed at a year of age [31]. In addition to the timing, quality, and quantity of handling, the number of different handlers may also impact on kittens’ subsequent behaviour towards unfamiliar humans. In one study, kittens regularly handled by five different people (as opposed to a single person or not handled at all) were observed to make fewer attempts to retreat from a stranger [104]. Potentially heritable traits may also interact with kittens’ early social experiences to impact the nature of their social behaviour towards humans. In the following studies [102,103], kittens considered to be more ‘timid’ were anecdotally noted as being initially less amenable to handling, although this did improve if they were then adequately socialised to humans during 2–7 weeks of age. A study of cats residing in a research facility [33] suggested the genetic influence of paternal ‘boldness’ might enhance the degree to which (human socialised) offspring demonstrate sociability, via a greater receptibility to human-socialisation. In [33] while kittens that had been regularly handled between 2 and 12 weeks of age and sired by unfriendly, less ‘bold’ fathers were friendlier than unsocialised kittens from both friendly and unfriendly fathers, those that were socialised and sired by friendly males were the most friendly and confident of all groups. When tested at a year of age, these latter cats were most likely to approach and explore novel objects, approach and interact sociably with a person, and were also least likely to behave aggressively when approached and handled. Such results would suggest these genetic/ontogenetic effects also demonstrate temporal stability. Interactions between heritable maternal traits, early social exposure to humans, and subsequent human-sociability have not been investigated, although it is likely that similar relationships are present. However, genetic (inherited) maternal influences are more difficult to study practically, given the challenges of controlling for maternally induced epigenetic effects that may impact kittens during their pre- and post-natal periods, prior to weaning. For example, modifications to kittens’ gene expressions which affect their behavioural phenotypes may occur as a result of maternal stressors, maternal deprivation, and the dynamics of the mother-offspring relationship [105].Cats’ social experiences with conspecifics during their sensitive period appear similarly important in influencing subsequent social behaviour towards other cats. In a study of kittens within a research facility [106], it was noted that kittens’ behavioural responses to an unfamiliar kitten varied depending on whether they had been reared with/without their mother and with/without their siblings until six weeks of age. Kittens were regularly tested from 2 to 20 weeks of age and kittens that were reared in isolation from their littermates (either with or without their mother) were found to display more frequent biting, wrestling, approaching, lowering their ears, as well as sideways stances with piloerect fur, when in an unfamiliar kitten’s presence. These littermate-deprived kittens were also described by the authors as less able to engage in appropriate social play with a conspecific, and that the nature of their social interactions was more agonistic than play-like. These kittens were also reported to have their claws extended more often and were less able to display bite inhibition. For example, such kittens were described as intensifying their biting and attacking when the test kitten performed a distress cry or tried to escape, compared to the kittens reared with their littermates, who would usually retreat at this point. The kittens reared without littermates were also described as generally being more ‘hyperactive’ when exposed to both objects and the test kittens. Other studies suggest that a cats’ early familial environment may impact their future behaviour towards conspecifics as well as humans. In another lab-based study [107], groups of kittens were either separated from their mother and litter mates at 2, 6, or 12 weeks of age. Based on experimenter observations of the cats over a 9-month period, the kittens separated from their mothers at 12 weeks of age were anecdotally described as being more ‘docile and friendly’ towards conspecifics as well as humans, while the kittens isolated at 2 weeks of age were described as generally more anxious and likely to behave aggressively when frightened. Anecdotal observations from other studies suggest that the presence of both the mother [108] and littermates [109] during socialisation towards humans may reduce anxiety and increase the kittens’ receptivity to humans’ attempts to socialise them. Additionally, in a large survey of pet cats and their owners [110], cats that were (retrospectively) reported as being removed from their mothers earlier (i.e., before 8 weeks of age) were more likely (as adults) to behave aggressively towards unfamiliar humans than those weaned later (i.e., between 12–15 weeks of age). The authors also reported that cats that either had not been separated from their mothers at all, or had not been separated until reaching adulthood (i.e., at least one year or older), were less likely to behave aggressively towards both conspecifics as well as familiar and unfamiliar people. Under certain conditions, such as limited amounts of human-socialisation and barren environmental conditions, impacts of a cats’ early social environment on their subsequent behaviour towards humans may be subsumed by paternal genetic effects. In a study of cats within a research facility [111], differences in kittens’ latency to approach a person, durations they held their tail above the horizontal, and level of tolerance to human restraint during blood sampling were assessed. When tested at 20 weeks of age, no significant differences were reported in kittens’ behaviour relative to whether they had experienced early separation from their mother (i.e., 5 weeks) and subsequent individual caging with either basic (i.e., 15 mins three times a week for 3 weeks) or no human handling, or whether they had experienced later separation from their mother (i.e., at 6–7 weeks) and been subsequently housed with conspecifics. However, variation in kitten’s behaviour during tests was associated with the identity of their father. Kittens sired by two particular fathers were reported as behaving more sociably towards humans and being more tolerant to restraint, whilst kittens sired by other fathers were reported to be less sociable and more aggressive. While it is certainly plausible that these behavioural differences amongst kittens could be due to differences in the boldness/shyness amongst their fathers [33,112], no data on paternal personality were reported in the study. In general, adequate social exposure towards humans during the cats’ sensitive period (i.e., 2–7 weeks of age) and beyond, appear central to the development and stability of human-sociability, although inherited traits (e.g., boldness/shyness) may also play important mediating roles. Later separation (i.e., beyond the cats ‘sensitive period’ for human-socialisation) from mother and siblings may also enhance both conspecific and human-cat social relationships later in life. Related conspecifics that remain together from birth may be more likely to engage in affiliative interactions than non-related cohabiting conspecifics. However, whether cats have a similar ‘sensitive period’ for the development of sociability to conspecifics, if this generalises to unfamiliar, unrelated cats in the same way that it might in humans, and the role of heritable personality traits as moderating factors remain unclear. 8. Lifestyle Variation, Social Flexibility, and Welfare ConsiderationsWithin a cats’ lifetime, some individuals may transition from an independent free-living lifestyle to cohabiting with humans (and also conspecifics) within the confines of a domestic home. In some instances, the cat might actively facilitate this transition, for example they may start spending time near to human dwellings and eventually decide to ‘move in’. In most cases, however, this process occurs due to human intervention. Typically, a cat deemed to be ‘stray’ or ‘unowned’ is physically removed from its original location, temporarily housed in a shelter environment, before then being placed into a domestic home of humans’ choosing. Indeed, such cats can represent a large proportion of the total shelter population, with a UK study suggesting up to as many as 42% within a given year [113]. It is worth noting that prior to a period of free living, some of these stray or unowned cats may have originated from domestic homes, and thus been appropriately socialised (to humans and potentially conspecifics) during their sensitive period. However, many of these free-living cats may have had no, or very limited, social experiences with either species, and thus potentially a much more limited capacity to adapt well to future environments that require close cohabitation with them. Within the shelter environment, a cats’ previous social experiences (and inferred degree of socialisation) appear to be important determinants of their wellbeing. In [39] cats that were deemed to be ‘unsocialised’ towards humans had higher Cat Stress Scores (CSS) than those considered to be ‘socialised’, and those considered ‘unsocialised’ to conspecifics had higher CSS when housed in groups. In another study [40], cats that had previously resided in single, rather than multi-cat, homes were found to have significantly higher CSS during the first few days of their arrival at a shelter, even though they were housed individually. A lack of previous socialisation to humans and conspecifics is therefore likely to compromise the wellbeing of cats when housed within a shelter environment. While the impact of conspecifics on cats unsocialised to cats can potentially be mitigated by proving them with single housing, an equivalent is obviously more difficult to provide for cats unsocialised to humans within such human-managed confined environments. Comparable data regarding the wellbeing of cats within the domestic home, based on their lack of previous socialisation to humans and/or cats, is missing. However, it is anticipated that similar relationships are likely present, with such previously unsocialised cats struggling more (or generally failing) to adapt.Attempts to human-socialise unfriendly and fearful cats that are already outside of their ‘sensitive period’ (i.e., 2–7 weeks of age) can be common practice within shelters [114,115]. While the aim of this process may be to try to enhance a cats’ capacity to adapt well to future cohabitation with humans, such practices may pose important ethical, as well as practical, considerations. The cats’ early developmental or ‘sensitive period’ is characterised by enhanced neurobiological plasticity, outside of which neuronal circuits are typically less susceptible to modification via experience [116]. Additionally, innate responses to human proximity in non-socialised cats are primarily characterised by fear, attempted avoidance, and defensive behaviour [117]. Therefore, both the facilitating of non-fear inducing and positive social experiences with humans, and their ability to be rapidly generalised to other humans and contexts, is likely to be more difficult to achieve, beyond this initial period of plasticity. Instead, it is likely that acceptance of humans in previously non-socialised, fearful cats occurs through the process of stimulus-flooding. This process involves the exposure of an individual to negative, acutely stressful stimuli from which it cannot avoid, until cessation of the initial behavioural response (i.e., fearful or aggressive reactions) is achieved [118]. Rather than producing a positive end result, this process can potentially induce the experience of ‘learned helplessness’ within individuals, due to their lack of perceived ability to control their exposure to the fear-inducing stimulus [118,119,120]. It is uncertain if cats that have undergone such stimulus-flooding processes are then able to develop generalised positive associations with humans in the same way as cats socialised towards humans during their ‘sensitive’ period are. However, anecdotal reports suggest these cats tend to struggle to cope with domestic living and also typically fail to meet owner expectations for companionship (Oral communication, Battersea Dogs and Cats Home). This may subsequently lead to reduced owner satisfaction [121,122] and potentially cat relinquishment [123,124]. However, given the potential moderating role of individual personality (i.e., inherited traits) during cats’ initial socialisation to humans [33,103,111], it is possible that individuals of a certain temperament may be more receptive to developing positive relationships with people, even when these commence outside of their ‘sensitive period’. For example, very bold, free-living, food-motivated cats that are not innately fearful of humans (sensu [3], see next section) may have a greater capacity for positive human-social learning. However, their subsequent adaptability to domestic living may still be limited, considering their lack of previous experience with this lifestyle and its associated challenges [28,125].9. Domestication, Selection, and Implications for Sociability in Modern Day Domestic CatsIn a comparative review of sociality across Felidae, [126] the author notes that while increases in sociality within a species are usually associated with increases in brain to body mass ratios (i.e., encephalisation), this relationship has not been evident during the domestication of the cat. At a genomic level, however, notable differences between domestic cats and their wild progenitors are evident. In a comparative analysis of the domestic cat, European (F. silvestris silvestris) and Eastern (F. silvestris lybica) wildcat genes [3], positive selection for enriched neural-crest-related genes within domestic cat samples was identified. In mice, these genes are linked with a reduction in fear responses, enhanced memory, and the ability to learn based on positive rewards such as food [3,127,128,129]. These genetic signatures potentially explain the cognitive mechanisms through which initial domestication and increased sociality were able to occur, with less fearful individuals more readily able to tolerate human proximity and create positive associations with humans. Compared with their early domesticates, however, modern-day companion cats are likely to experience more socially complex and potentially challenging cat-caregiver dynamics [23,37,66,130]. Despite this, it is unclear whether much active selection for traits that enhance successful adaption to modern human-social relationships and domestic living have been undertaken [2]. For example, in domestic dogs, genetic signatures suggestive of intense selection for prosocial traits such as those associated with enhanced responsiveness towards humans, attention-seeking, and initiation of prolonged social contact are evident [131,132,133]. While in domestic cats, this area of research has received less attention, similar evidence of genotypic-phenotypic relationships for prosocial behavior and their positive human-selection are currently lacking. For example, in [32], genetic surveys on microsatellite polymorphisms in a population of domestic cats were conducted, focusing on those linked to receptors (associated with oxytocin release) and previously linked with sociability in other species [131,134]. While in [32], positive associations between microsatellite length and the caretaker-rated friendliness of cats were identified, pedigree cats were found to have shorter alleles compared to non-pedigree cats, implying a less enhanced genetic predisposition for human-sociability in pedigree populations. Additionally, while levels of oxytocin and cortisol have previously been found to increase in domestic dogs following social interactions with their owners [135], a recent preliminary study in domestic cats reported that both oxytocin and cortisol levels were actually decreased following human-social interactions [136]. In a recent survey of cat owners [25] investigating the heritability of cat behaviour traits, considerable diversity in the social tendencies of cats across different breeds was reported, with no obvious trends towards increased human or conspecific sociability, compared to non-pedigrees. Certain breeds also scored particularly highly for unsociable behaviours, including limited contact with people (e.g., British short hair) and aggression towards both conspecifics and humans (e.g., Turkish van). Therefore, rather than enhancing human and/or conspecific sociability, in certain cases, selective breeding practices may have produced the opposite effect. It is, however, important to highlight that the data used to assess cats’ behaviour in these studies were based on the subjective reports from owners and caretakers and from two distinct populations of cats and owners (i.e., Finnish [25] and Japanese [32]). Breed types were also confirmed by owners [25] or vets [32] rather than via objective genetic or morphometric methods. Given that humans’ perceptions of different cat breeds and their behaviour seem to vary between studies and demographic populations [137,138,139], and that only a limited number of pedigree cats (n = 40) and breed types (n = 10) were included in the genetic analysis conducted in [32], the generalisability of these findings may be limited. Further studies that incorporate standardised behavioural observations, in addition to more comprehensive genetic analyses, are likely to facilitate a more comprehensive understanding of the potential differences in sociability between pedigree and non-pedigree cats at both the behavioural and genetic level.Little is known about whether sociability towards humans and conspecifics might share similar trait pathways within domestic cats. However, latent behavioural variables derived from various surveys of owners/caretakers reporting on cats suggest that human-directed and cat-directed social behaviours form their own separate structures, and, as such, could be considered individual traits [140,141,142]. Additionally, while in [25], some breeds demonstrated similar rates of expression of particular social behaviours towards both species, this was not consistent across all breeds. For example, the Turkish Van had a comparatively high probability of aggression towards both cats and humans, whilst the Devon Rex had a comparatively high probability of aggression towards humans, but comparatively low probability towards other cats. The apparent emphasis on the physical characteristics of pedigree breeds (over traits that may convey social advantages) might have other limiting consequences regarding their social capabilities. For example, it is possible that breed-based characteristics that cause discomfort or limit the physiological or communicative functionality of an individual may also negatively impact their social relationships with both humans and/or conspecifics. For example, highly brachycephalic breeds such as the modern Persian and Exotic short hair can experience a range of health conditions associated with their eyes, skin, and respiration [143,144]. In the Scottish Fold, their characteristic ‘folded’ ears are the result of selection for heritable gene mutations, which also cause abnormal bone and cartilage development, leading to chronic pain and mobility issues [145]. The presence of such chronic conditions has the potential to induce poor mood, increased irritability, and human- and conspecific-directed aggression, in addition to a range of other behaviours which owners tend to find problematic [146,147,148]. Such individuals may also be more likely to occupy sedentary lifestyles [144] and generally have a reduced desire or tolerance for positive social interactions with humans and/or conspecifics. Breeds with highly exaggerated morphology may also struggle to effectively communicate during social interactions. A recent study [149] has suggested that breeds with more extreme facial features, such as those that are very brachycephalic (e.g., Exotic short hairs, modern Persians) or dolichocephalic (e.g., Oriental short-hair and Sphynx), may have more limited abilities to produce clearly identifiable and differentiable facial expressions. Given the importance of facial expressions during the expression of emotions and intentions [150,151], and thus the maintenance of social relationships, such limitations might have important consequences regarding the nature of cats’ social interactions. Links between cats’ altered morphology and their communicative abilities may also extend beyond the face. For example, the shortened limbs and proportionally elongated spine of the Munchkin may limit their ability to effectively alter their posture for communicative purposes, while the largely absent tail of the Manx cat prohibits their ability to perform important affiliative behaviours such as conspecific tail wrapping and the vertically raised ‘tail up’ signal [11,48,152]. Further studies are required to fully understand how breed-linked morphology might impact on the behavioural repertoires utilised by individuals during social interactions with both humans and conspecifics, and the subsequent effect this might have on the nature of their social relationships. Likewise, research is needed to assess the potential negative impact of breed-linked health conditions on cats’ tendencies to engage in different styles of social interaction (e.g., affiliative, agonistic, tolerant, or avoidant) with both species. 10. ConclusionsAt a species level, the domestic cat displays an impressive capacity to cohabit successfully with both humans and conspecifics, despite their recent asocial ancestry and apparent limited selection for traits that might enhance their social capabilities. At the individual level, however, cats may demonstrate substantial diversity in their human and conspecific sociability. This may have an impact on their ability to cope with the various social challenges to which they are exposed when under human management. This diversity may largely be explained by interactions between early social experiences and inherited behaviour traits, although further studies which more broadly and explicitly test these hypotheses are needed. Additionally, various proximate factors, including individual cat and human characteristics, appear important mediators of cats’ social interactions with both species. However, the availability and quality of evidence varies, and in some instances equivocal findings limit current understanding.As urbanisation continues and companion animals are increasingly relied upon for social and emotional support [23,153], domestic cats are likely to be exposed to increasing environmental and social challenges whilst cohabiting with humans and conspecifics in confined environments [125]. Thus, active selection for traits that enable cats to adapt to these lifestyles are likely to be beneficial for their wellbeing, as well as their future relationships with both species. Additionally, adequate exposure of cats to positive social experiences during their early development, in combination with suitable housing and handling practices, represent important aspects of their husbandry that may support similar outcomes.

animals : an open access journal from mdpi

10.3390/ani13061065

PMC10044549

The keeping and use of horses have become of increased interest to the public due to welfare concerns. It is therefore vitally important to better understand the impacts on the horse’s emotional state and how to measure and use observed behaviours to determine the effects of common husbandry practices on the horse. This will enable steps to be taken to improve equine quality of life and ensure the social licence to operate within the horse industry. In order to achieve this, reliable animal-based behavioural indicators of welfare that can be used in industry are needed. The behaviour of horses kept on day-stabling routines was compared to that of horses kept on night-stabling routines. Eight behaviours including ear movement and locomotory leg movements as well as yawning, recumbency, and non-nutritive chewing occurred significantly more often in horses on a night-stabling routine. These behaviours have been identified as potential indicators of affective state (the animal’s underlying emotional state) and equine welfare that can be used in dynamic quality of life assessments.

Increasing interest in equine welfare has emphasised the need for objective and reliable behavioural indicators of horses’ affective state. However, research has yielded mixed results regarding behaviours suited for industry use largely because they are subject to anthropomorphic interpretation. Stabling is commonly used to manage domesticated horses despite research indicating that it can negatively impact horse welfare, but its effect on their affective state is yet to be quantified. Ten adult horses (11.8 ± 4.4 years) were observed either on a day- (DS) or night-stabling (NS) schedule over two consecutive 24 h periods. NS horses were kept confined for significantly longer (13.60 ± 0.04 h) than DS horses (7.73 ± 0.07; t7 = 5.70; p = 0.0004). Eight behaviours occurred significantly more often during NS than DS: forward ears (t7 = 3.32; p = 0.001), neutral ears (t7 = 3.47; p = 0.001), stepping forward (t7 = 2.62; p = 0.001), stepping laterally (t7 = 2.39; p = 0.001), sternal recumbency (t7 = 2.64; p = 0.001), yawning (t7 = 2.69; p = 0.001), non-nutritive chewing (t7 = 2.49; p = 0.001), and closing eyes (t7 = 2.71; p = 0.001). These behaviours may be candidates for indicators that can be used to determine the affective state in horses and subsequently be used to assess equine quality of life and to optimise individual horse welfare.

1. IntroductionThe welfare of horses has been subject to increasing scrutiny by the general public due to high-profile sporting events where horse management and use have at times appeared to be substantially less than optimal. Failure to appropriately manage horse welfare leaves the industry with a tenuous social licence to operate (SLO). It is therefore increasingly important to review contemporary and emerging management practices and their alignment with the promotion and safeguarding of good welfare [1]. Whilst much of the equine industry and the scientific community recognise the need for a reliable and robust method of welfare assessment [2,3,4], the complexity of this task is evident from the still-evolving nature of applicable welfare assessment tools that will enable reliable assessment of the horse’s physiological and psychological needs [5,6]. Despite a growing appreciation of animal-based indicators of welfare, difficulties remain with accurately determining objective measures of an animal’s emotional state in general, and of positive affective state in particular [1]. In the absence of validated measures for accurately identifying equine affective state, the most common method of evaluating equine welfare is behavioural observation, therefore making the identification and validation of behaviours reliably linked to mental state critically important [7].Although keeping horses in the paddock full-time is often considered preferable [8], this may not be possible due to restrictions on resources such as space and feed, resulting in many horses being stabled for at least part of the day. Some horse keepers consider stabling a suitable alternative, or in some cases, the ideal method of housing for horses [9]. Common management considerations of owners include available land, space, cost, access to riding facilities, and proximity to events or veterinary clinics. This may be compounded by a lack of knowledge regarding practices that optimise horse welfare, leading to impacts on equine quality of life. Meanwhile, constraints may also be placed upon horses to manage their dietary or physical requirements, e.g., yarding (small, confined outdoor pen) [10,11].Stabling has been linked to significant impacts on the horse’s physiological health, as evidenced by the increase in clinical conditions such as equine asthma [12,13,14] and equine gastric ulceration syndrome (EGUS) [15]. However, individuals and businesses offering housing for horses may not find it easy to provide more appropriate housing or understand the need to do so [9]. Many facilities subdivide the land available into small areas to increase the number of horses that can be provided for [16], and as Yarnell et al. [17] observed, this can lead to a reduction in group/herd size and may even favour keeping horses singly, which has also been reported to be contrary to good welfare. Therefore, it is important that modern horse management practices are routinely scrutinised and investigated for opportunities to optimise the horse’s home environment. Whilst there has been research conducted regarding the behaviour of stabled horses [18,19,20,21], particular emphasis has been put on stereotypies [22,23,24] or specific behaviours [25], with little regarding the effect of different stabling routines on equine behaviour and welfare. In addition to this, the effects of light and circadian rhythm on horse behaviour lack sufficient research to be quantifiable.When using evidence-based research outcomes, the optimal form of horse housing is commonly regarded as large paddocks that enable locomotory behaviour and provide access to conspecifics, with adequate shelter and sufficient nutritional content in the grass to minimise or negate the requirement for supplementary feeding [9,19]. When comparing the time budgets of horses kept in modern housing situations, a dramatic contrast to their feral counterparts is shown, with significantly reduced locomotion, grazing behaviours, and conspecific interactions [18,26]. The Five Domains was proposed in 1994 as a comprehensive method of welfare assessment due to its ability to provide a systematic and structured approach to animal welfare assessment and management with an increased focus on the impact of animal affective state [6]. The adoption of this welfare framework has allowed researchers to further consider an individual’s mental state [5], and current evidence exists that stabling can have a detrimental effect on the mental welfare of the horse [21]. These combined effects on the physiological and psychological wellbeing of the horse can significantly impact their overall welfare and quality of life [17] and highlight the need to investigate the effects of stabling on equine welfare and, consequently, quality of life.There is currently no standardised equine quality of life (EQoL) assessment framework in place, and importance has been put upon the development of a system that uses evidence-based and observable behaviours that can reliably reflect the individuals’ affective state [27]. This will allow for the preferences of the individual to be better recognised and catered for, meaning that common husbandry procedures, such as stabling, can be optimised to better suit the individual, focusing on a ‘best fit’ model rather than a rigid focus on a ‘best practice’ approach [28].The aim of this study was to determine the effect of stabling routines on equine behaviour in order to identify potential indicators of their affective state that may be suitable measures of equine quality of life with further focused research and validation. Two common stabling routines where turnout is limited were compared: horses stabled during the day and turned out overnight and horses turned out during the day and stabled overnight. The hypothesis that stabling routines would not have a significant effect on horse behaviour was investigated.Behaviours were observed via the use of continuously recording video cameras placed discretely in the stable and classified according to an ethogram developed during an earlier pilot study of five horses. The significant behaviours identified by this study were then investigated in the literature for potential links to affective states in horses as well as other species.2. Materials and Methods2.1. SubjectsThe horses used in this study were recruited from the existing pool of student horses boarded at the Charles Sturt University Equine Centre. Participation in the study was on a voluntary basis from the owners, for which written consent was obtained. Ten geldings with a mean (standard deviation) age of 11.8 (4.4) and various breeds were used for this study (Table 1). All horses were considered to be pleasure and/or low-level performance horses in light to heavy dressage and jumping work as described by the National Research Council (NRC) 2007 guidelines [29]. The mean (SD) body condition score (BCS) was 3.35 (0.45) at the commencement of the study. The horses included in the study were not suffering from any illness, disease, injury, or lameness as defined by the owners.2.2. Horse Housing and ManagementTwo stabling routines were investigated in this study. Five horses were stabled overnight from approximately 17:00 to 09:00 and turned out into a two-acre paddock shared with 3–5 conspecifics for the remainder of the day. The remaining five horses were stabled during daylight hours from approximately 09:00 to 17:00 and turned out into the same two-acre shared paddocks with three to four conspecifics overnight. The stabling facility used was located at the Charles Sturt University Equine Boarding Centre. Stables were 4 m × 4 m and bedded approximately 25 cm deep with sawdust. Horses were able to make olfactory, visual, and restricted tactile contact with neighbouring horses on two to three sides of the stable. All horse management decisions including stabling routine and timings, nutrition, and workload were determined by the owner, and no changes were made to the normal routine as a result of participation in this study. All horses were stabled at the CSU Equine Centre prior to their involvement in the study and were observed in their familiar home environment/stable. Stabling allocations, paddock allocations, and routines were predetermined by the CSU Equine Centre staff, and no changes were made to this when participating in the study. Throughout this study, all horses had at least one horse in a neighbouring stable whilst being observed.When stabled, forage was provided in the form of hay via hay nets or on the floor of the stable according to owner preference, and fresh water was provided in one or more buckets. Boarding facility regulations necessitated that horses be turned out for half the day, and when turned out in the paddock, horses had access to three or four conspecifics and ad libitum access to water but limited pasture. The available pasture was known to be insufficient in nutritional content, quantity, and quality to adequately meet the forage requirements of an individual horse, and owners were not permitted by the facility to provide supplementary feeding during turnout to avoid incidental feeding of other horses and the potential risk of increased injury.Due to the temperate climate of the area, the stabling facility had an open and back-to-back layout, meaning that stables were not enclosed to the outer environment and had an open view of the surrounding facility. Therefore, during the day, natural lighting was the main source of light. At night-time, there were lights available to handlers for use when at the facility, which were extinguished from 10:00 p.m. Sunrise was at approximately 7:00 a.m. and sunset at approximately 5:00 p.m. during the time of this study.2.3. MaterialsA 4-channel Techview 720p AHD DVR recorder with four indoor/outdoor 720p cameras equipped with infrared LEDs to allow for night-time recording was used. Data from the Techview DVR were transferred onto a 2 TB Seagate external hard drive to allow for the viewing of footage on a Dell Latitude 5400 computer via the VLC Media Player application.2.4. ProcedureAs this study was a purely observational study, there was no intervention by researchers, and horses were observed in their familiar stabled environments. All horses were filmed over a 48 h period, allowing data capture of two successive occurrences of stabling per horse.A preliminary investigation was conducted on five horses to determine the number of cameras needed per animal in order to effectively capture behavioural data and to aid in the development of the ethogram for this study. The data generated in the preliminary investigation was not included in the final study due to changes made to the experimental procedure and design.Two cameras were used per stable, and cameras were set up at the highest point of the stable walls in diametrically opposing corners to minimise interference with or by the horse as well as allow for the maximisation of space captured. This allowed for two horses to be recorded per 48 h period. Researchers contacted the owners prior to the setup and installation of the cameras to coordinate a time to do so during which the horse was not being stabled as well as to notify them when the recording would begin. Recording began from the moment camera installation was completed and was kept on for the 48 h period in which the horses were being observed.Once camera installation had been completed, horse owners were notified that the recording period had begun and advised to continue with their normal stabling and husbandry routines. Recording ceased after 48 h at the conclusion of the observation period, and horses were removed from the stable. Recordings were dated and timestamped by the Techview DVR recorder to allow for identification when removed and backed up onto an external hard drive.2.5. Data Extraction and CollationRecordings were partitioned by the Techiew DVR system (Jaycar Electronics, Sydney, Australia) into 45 min blocks of footage. These video data files were downloaded from the DVR system onto the external hard drive where the data files were then organised by horse, date, and time once uploaded to the computer.A continuous time sampling method was used to assess the behaviour exhibited by all horses during the stabling period and viewed using VLC Media Player. The samples were analysed in ten-minute blocks, commencing with the first ten minutes of the horse in the box and post the owner departing the stable upon completion of normal husbandry practices and ending with the last ten minutes before either the horse was removed from the box or the stable was entered by the handler. For each hour between entry and departure from the stables, an interim block was chosen for scrutiny using the Google online random number generator. The total number of blocks analysed per horse was dependent on the length of time spent in the stable, as dictated by owner management preferences.Each sample video was observed continuously, and the subjects’ behaviours were recorded instantaneously and classified according to the ethogram comprising 52 possible behaviours developed during the pilot study of 5 horses, with behaviours categorised as per Kiley-Worthington (1997) [18] (Table 2). Individuals were also observed for their response to humans entering their stable for voluntary approach. Frequencies of behavioural occurrences were derived for each sample video and recorded into an MS EXCEL v16.0 (Microsoft, Redmond, WA, USA) spreadsheet for collation.2.6. Data AnalysisBehaviours were categorised for time budget analysis (Table 2) for comparison with existing published equine time budgets [30]. This was performed with the aim of identifying potential differences between the two stabling routines as well as potential differences between the horses used for this research and full-time stabled horses observed by Kiley-Worthington in 1989 [30]. The 52 observed behaviours were categorised into one of seven possible groups (eat, stand, lie, locomotory, non-locomotory, elimination, or other); however, this analysis ultimately excluded eight of the behaviours that were categorised as “Unspecified”. Seven of the behaviours were excluded as they were able to be performed within any behavioural classification, for example, ears forward whilst head up, whilst the eighth represented no visual contact with the horse and was therefore unable to be considered a performed behaviour.2.7. Statistical AnalysisThe data analysis involved tabulating the results in MS Excel v16.0, followed by statistical analysis in R version 3.6.2 [31]. The normality of the data was verified through a Shapiro–Wilk normality test. To compare the behaviours exhibited by day- and night-stabled horses, a series of unpaired t-tests were used. The equality of variance was assessed using an F-test, and a Holm–Bonferroni correction was implemented to minimise the risk of Type I errors. A significance level of p < 0.05 was consistently applied. Moreover, a hierarchical clustering analysis method was employed using the hclust function of R to identify potential clusters of behaviours observed in the two test conditions, i.e., stabled at night and stabled during the day.2.8. Ethical ApprovalThe research undertaken was approved by the Charles Sturt University (CSU) Animal Care and Ethics Committee, authority no. A18043.3. ResultsObservations of 52 different behaviours across nine horses over 12 days and 576 h of footage resulted in 13,778 behavioural occurrences recorded. The mean (SD) number of behaviours performed was 14.8 (1.1) per time sample blocks analysed per horse in the night-stabled routine and 8.5 (1.7) per time sample blocks analysed for horses in the day-stabled routine. This yielded a mean (SD) of 1531 (688.8) observations per horse. Horse 10 was excluded from the study due to owner mismanagement that resulted in stable confinement for the entire 48 h period, a decision that was made independently of the study and meant that observations would have not been consistent with other horses receiving turnout.3.1. Time BudgetsThe 52 observed behaviours were organised into seven categories (Table 2). Behaviours listed as belonging to the unspecified category were excluded from the time budget analysis as these could be performed in conjunction with other behaviours. A bar chart was constructed to illustrate the time budgets of stabling routines and the total population over their respective 48 h sampling period (Figure 1).3.2. Significant BehavioursHorses kept in the stables overnight were housed for a significantly longer period (13.60 ± 0.04 h) than horses kept in stables during the day (7.73 ± 0.07 h; t7 = 5.70; p = 0.0004) and displayed significantly more total behaviours on average per hour of stabling (t7 = 2.70; p = 0.03).A series of unpaired t-tests were also used to compare the frequencies of behaviours occurring with stabling routines, and 11 behaviours were found to have a significant difference. This was reduced to eight after the Holm–Bonferroni correction was applied. No horses demonstrated stereotypical behaviours such as windsucking, weaving, or box walking, whilst all horses demonstrated a positive approach when humans entered the stable. Significant increases in the frequency of ear movements (forward ears, t7 = 3.32, p = 0.001; neutral ears, t7 = 3.47, p = 0.001; locomotory behaviours (stepping forward, t7 = 2.62, p = 0.001; stepping laterally, t7 = 2.40, p = 0.001); and behaviours classified as ‘Other’ (yawning, t7 = 2.70, p = 0.001; non-nutritive chewing, t7 = 2.49, p = 0.001) were seen in night-stabled horses. Other behaviours that showed significant increases for night-stabled horses were closing of the eyes (t7 = 2.69; p = 0.001) and sternal recumbency (facing sideways, t7 = 2.640; p = 0.001). Significant behaviours and the number of occurrences were compared using a bar chart (Figure 2).Hierarchical cluster analysis was conducted to identify potential behavioural clusters that occur under the different stabling conditions and may help indicate relative affective states of behaviours of significance with better researched and understood behaviours. Figure 3 and Figure 4 visually illustrate clusters of behaviours for the night-stabling and day-stabling routines, respectively, generated by hierarchal clustering analysis.4. DiscussionFor a behavioural observation to be useful in welfare assessments, it needs to be easily observed and validated [1], which has meant that high-intensity behaviours such as overt aggression or stereotypies such as box-walking and weaving are easily classified and associated with compromised welfare, suggesting negative affective states [22]. However, given that welfare is experienced differently by individuals, the absence of obvious negative affective states does not indicate that the individual is experiencing ‘good’ welfare, nor does the absence of obvious positive affective states suggest poor welfare [32]. Unsurprisingly, much difficulty has been encountered in determining the possible affective state associated with more subtle equine behaviours [33]. Therefore, behavioural observations were used in acknowledgment that other physiological measures such as heart rate (HR), heart rate variability (HRV), and salivary cortisol are not accessible to the average horse owner and still require sufficient validation when used to measure behaviour, making it important to see if observations and subsequent welfare assessments could be used independently.Whilst the video capacity did not permit an accurate assessment of eye blink rate in this study, instances of eyes closed were able to be captured and analysed. If the spontaneous blink rate (SBR) is required in future studies, the study design would require adjustment to accommodate the limitations experienced in this instance. However, horses stabled at night displayed significantly more instances of eyes closed than those stabled during the day. This appears to coincide with the significantly greater number of night-stabled horses that would sleep in a sternal recumbent position. When examining the dendrogram for behavioural clustering of blinking behaviours, day-stabled horses did show some correlation between eyes closed and stretching behaviours. However, when examining the dendrogram of night-stabled horses, there was a stronger correlation between the eyes closed behaviour and the behaviour of facing away from the stable door with their heads lowered (negative internal). A likely explanation for this is that the horse is better able to achieve a restful state when external stimulation is at a minimum, allowing the horse to feel comfortable enough to face into the box.Unlike humans, horses have adapted to minimal sleep patterns and only sleep for approximately four hours a day. Whilst horses have evolved to sleep standing up via the use of the stay apparatus, effective rapid eye movement (REM) sleep only occurs during sternal or lateral recumbency and is an important requirement for overall horse health, both physiologically and psychologically [34]. Multiple factors affect recumbency in horses including bedding substrate [34,35] and stable size [36] and may also be influenced by equine stress levels [34]. Seven of the nine horses were observed in sternal or lateral recumbency during the time periods analysed. Horses stabled overnight were significantly more likely to display sternal recumbency, which may indicate a positive affective state given the option to perform this behaviour whilst in shared turnout with conspecifics, but were relaxed enough with their surrounding environment to pursue a full rest state when stabled. This suggestion is supported by the findings of Mazzola et al. in 2021 [37], which found horses that were brought in overnight for stabling had better sleep quality and reduced cortisol levels compared to horses kept naturally. When considering sternal or lateral recumbency as a potential cluster or paired behaviour, the highest level of correlation in the dendrogram for night-stabled horses is seen between standing negative facing a companion and sternal recumbency facing a companion. It is plausible that this is a vital component of the individual ethological need for companionship in order to achieve an effective resting state. When viewing the dendrogram for day-stabled horses, the strongest correlation was between a form of sternal recumbency (performed facing out of the stable) and head nodding; however, it is not yet clear how or why these behaviours may be related.The behavioural analyses revealed that the stabling regime impacts horse activity. Horses stabled at night spent significantly longer in the stable than when turned out in the paddock with herd mates than horses stabled during the day. It could be suggested, therefore, that if owners are looking for a way to maximise a horse’s access to turnout, it may be beneficial to stable horses during the day and turn them out. This may be of further benefit to horses housed in hot or humid conditions as it allows the horse to be turned out during the coolest part of the day. This may also support the use of other beneficial amenities such as misting equipment and fans in warmer climates, pending availability and stable design, without completely restricting the horse from turnout.Horses kept in the stable overnight were stabled for an average of 13.5 h, whilst horses kept stabled during the day were only stabled for an average of 7.75 h. Whilst time budgets of night- versus day-stabled horses did not show substantial differences in the types of behaviours or stereotypical behaviour, horses stabled overnight displayed more leg movement, such as stepping forward and laterally, which may indicate a milder version of barrier frustration behaviours such as box walking and weaving [38] and therefore indicate links to a negative affective state. When considering the dendrogram for night-stabled horses shown in Figure 4, the closest relationships were between stepping forward and increased tail swishing and between lateral stepping and increased lateral head movements. This may suggest increased movement and activity from the horse overall, which is perhaps indicative of increasing restlessness or frustration due to confinement. In contrast to this, Figure 3 for day-stabled horses showed that forward and lateral steps had closer relationships to rested hind legs and foraging, respectively, behaviours more commonly associated with rest and relaxation [39]. However, when comparing the time budget of the sample horses in this study to traditional time budgets of full-time stabled horses published by Kiley-Worthington in 1989 [30], it is interesting to note that the proportion of time spent eating is far greater for the sample horses. This may be due to changes in management practices since 1989 that have allowed horse owners to be more aware of the need for adequate access to roughage sources during confinement. Eight of the nine sample horses were provided with roughage during stabling, and half of those horses received their roughage in a slow-feeding hay net, which increased the average time spent eating.Ear positioning is often discussed when observing horse behaviour as a relative indicator of mood or affective state [40,41]; however, the frequency of ear movement displayed by horses in their resting environment has received limited investigation. When investigating the literature on other species, ear position and frequency of movement have been highlighted as being indicative of a negative affective state in previous studies with sheep [42]. Given the proposed relationship between ear position and assumed emotional states such as backward pinned ears being associated with discomfort, pain, or stress in horses [40] and sheep [42], rapid shifting between positions may provide further insight into the horse’s affective state. Increased ear activity was noted in horses stabled overnight, with ears moving between the forward and neutral positions with significantly increased frequency compared to those stabled during the day. Previous research into the positioning of the ears focused primarily on placement when performing behaviours under saddle [41,43,44], experiencing fear or pain [3,41], or as an indicator of the horse’s attention [45]. However, studies have linked the decreased frequency of ear movement to positive attitudes towards humans [45], and therefore, it should not be discounted in the identification of positive affective states. It is important to consider if the increased frequency of ear movement for night-stabled horses may be potentially explained by the decreased visibility conditions increasing the horse’s sensitivity to auditory stimuli. This proposal aligns with Hartman and Greenings’ demonstration in 2019 [25] of the positive influence of auditory stimulation on behaviours such as recumbency in stabled horses.Non-nutritive chewing behaviour has been previously highlighted in horses as being a response to stress [46], contravening industry associations that this behaviour is indicative of relaxation, learning, cognitive processing, or submission in the horse. However, research conducted by Lie and Newberry (2018) [46] showed that non-nutritive chewing behaviour was seen following aggressive conspecific interactions by both parties, with particularly higher prevalence when observing the aggressor, refuting the suggestion that the behaviour is based on submission. Instead, it is believed that non-nutritive chewing occurs when parasympathetic activity increases following a stressful event. Therefore, the increased prevalence of non-nutritive chewing behaviours observed in night-stabled horses may be indicative of increased stress levels and, subsequently, a negative affective state. When considering the dendrogram for day-stabled horses in Figure 3, non-nutritive chewing was most closely related to drinking, lending further support to the idea that non-nutritive chewing occurs with reactivation of the parasympathetic nervous system, which prioritises digestive processes over the fight/flight response [47]. Similarly, Figure 4 for night-stabled horses shows non-nutritive chewing as being most closely related to head-lowering behaviour, further linking it to an activity performed after experiencing a stressful state. Day-stabled horses also demonstrated a moderate link to yawning, which also has links to negative affective states.Incidences of yawning were similarly seen to increase in frequency for night-stabled horses; however, its role in the identification of affective state remains complex due to the determinants of yawning behaviour remaining unknown [48]. Whilst yawning behaviours have been shown to have a relationship to sleep in multiple species [49,50,51], similar research exists to show an increase in the prevalence of yawning behaviours following stressful and emotional interactions as well [51,52,53]. Fureix et al. (2011) [48] investigated the relationship between yawning behaviour and stereotypic behaviours in an effort to further quantify the significance of yawning in horses. The results indicated there was a positive correlation between yawning frequency and the performance of stereotypical behaviours, suggesting potential links to poor welfare or a negative affective state. This is further supported by Gorecka-Bruzda et al.’s (2016) [53] findings, in which yawning behaviours were investigated in both domesticated horses and Przewalski horses for potential causation, with findings demonstrating a positive relationship between the frequency of yawning behaviour and stressful social situations.Whilst stabling has been linked to decreased welfare states in horses due to the inability to engage in natural grazing behaviour and interactions with conspecifics, the horses used in this study did not display any stereotypies commonly associated with stabling, such as weaving, box walking, or crib biting. This may suggest that the provision of even limited turnout access could improve the negative underlying factors that are believed to initiate stereotypical behaviour in some horses, therefore preventing them from originating and becoming consolidated reactions to suboptimal conditions [54,55]. Further to this, all horses used in this study displayed voluntary approach behaviour to the handlers retrieving them from the stable. The voluntary approach response in horses has previously been used as an indicator of a positive affective state in horses and is indicative of their previous human–horse interactions [56]. Whilst a voluntary approach to humans can be considered ideal and a potential indicator of the horse’s affective state, more research needs to be carried out to verify whether this response is due to a positive affective state related to their housing and management, due to their positive association of the presence of a human with their removal from the stable, or simply a conditioned response to human approach regardless of affective state or environmental conditions and management.The behaviours identified in this study may assist with the identification of affective states in horses with further research and validation to ensure behavioural indicators are robust and reliable across contexts. With increased knowledge of behavioural indicators of affective state, horse owners and managers may be able to better identify individual perceptions of common husbandry practices and make adjustments to better suit the needs of the individual, therefore improving the day-to-day quality of life of the animal.Some limitations to this study were experienced due to the variability in owner management practices, which may have had an overall effect on behaviour. One horse was removed from the study due to the owner changing their management routine to that of a full-time stabled horse partway through the study, thereby necessitating data generated by that horse for exclusion. The variability between horses in the type of feed/forage provided, how the feed was delivered, exercise routines, and total time spent stabled could be controlled in future studies, resulting in a more homogenous sample population. The small sample size of the population was also considered to be a limitation of the study design.5. ConclusionsGiven the fact that stabling is a commonly practiced, accepted, and even promoted or required component of modern horse management, efforts should be made to optimise approaches to housing horses in order to alleviate the stress experienced by the horse rather than focusing on the inherently negative aspects associated with stabling. Whilst it is not yet possible to reliably categorise all behaviours as indicators of positive or negative affective states, specific behaviours such as non-locomotory leg movement as well as ear position and frequency of movement are already being identified for future use in dynamic contemporary welfare assessments that are applicable to all horses and across multiple contexts in which they are displayed.

animals : an open access journal from mdpi

10.3390/ani11061598

PMC8227379

This study aims to investigate how epidermal growth factor (EGF) attenuates the effect of lipopolysaccharide (LPS) on the growth performance, nutrient digestibility, microelement absorption of early-weaned pigs. A total of 48 early weaned piglets were randomly distributed to four groups consisting of a 2 × 2 factorial design. The main factors were the level of LPS (HLPS = high LPS: 100 μg/kg body weight; ZLPS = low LPS: 0 μg/kg body weight) and EGF (HEGF = high EGF: 2 mg/kg diet; ZEGF = low EGF: 0 mg/kg diet). Each group had four replicates and each replicate consisted of three piglets. The results showed that HLPS level decreased the growth performance and the apparent digestibility of crude fat, while HEGF level increased the average daily feed intake. The concentration of most microelements in the gastrointestinal tract chyme and feces were increased by HLPS level and decreased by HEGF level. The expression levels of most microelement transport-relative genes in the mucosa of gastrointestinal tissues were decreased by HLPS level and increased by HEGF level. In conclusion, dietary EGF could attenuate the negative effect of LPS exposure on the apparent digestibility of crude fat and microelement absorption through changing the expression levels of microelement transport-relative genes. EGF can be used as an additive to increase the essential trace elements absorption in the early weaning piglets.

Epidermal growth factor (EGF) plays an important role in nutrients absorption. However, whether it can be an effective additive to improve the growth performance and nutrients absorption in lipopolysaccharide (LPS) challenged early weaning piglets is still unknown. A 14-days trial was conducted to investigate how EGF attenuates the effect of LPS on the growth performance, nutrient digestibility, microelement absorption of early-weaned pigs, and study the underlying mechanism. A total of 48 early weaned piglets, aged 25 days, were randomly distributed to four groups (control, EGF, LPS and EGF + LPS groups) consisting of a 2 × 2 factorial design. The main factors were the level of LPS (HLPS = high LPS: 100 μg/kg body weight; ZLPS = low LPS: 0 μg/kg body weight) and EGF (HEGF = high EGF: 2 mg/kg diet; ZEGF = low EGF: 0 mg/kg diet). Each group had four replicates and each replicate consisted of three piglets. The results showed that piglets injected with HLPS level significantly decreased the average daily gain (ADG), and significantly increased the feed conversion ratio (FCR) compared with the piglets injected with ZLPS level, while piglets fed HEGF level significantly increased the average daily feed intake (ADFI) compared with the piglets fed ZEGF level (p < 0.05). Piglets injected with HLPS level significantly decreased the apparent digestibility of crude fat compared with the piglets injected with ZLPS level (p < 0.05). Piglets injected with HLPS level significantly increased the concentration of most microelements in the gastrointestinal tract chyme and feces, and significantly decreased the expression levels of most microelement transport-relative genes in the mucosa of gastrointestinal tissues compared with the piglets injected with ZLPS level (p < 0.05). Piglets fed HEGF level significantly decreased the concentration of microelement in the gastrointestinal tract chyme and feces, and significantly increased the expression levels of the microelement transport-relative genes in the mucosa of gastrointestinal tissues compared with the piglets fed ZEGF level (p < 0.05). In conclusion, dietary EGF could attenuate the negative effect of LPS exposure on the apparent digestibility of crude fat and microelement absorption of early-weaning piglets. EGF and LPS influenced the absorption of essential trace element through changing the expression levels of microelement transport-relative genes in the mucosa of gastrointestinal tissues. In the early weaning piglets, EGF can be used as an additive to increase the essential trace elements absorption.

1. IntroductionEssential trace elements are the indispensable nutrients for animals, and especially Cu, Fe, Zn, and Mn are required for the normal growth, development, and many physiological functions in animals [1,2,3,4]. Cu is a part of Cu-transporting P-type ATPase and Cu/Zn superoxide dismutase [5]. Fe as the part of hemoglobin and myoglobin plays an important role in delivering the oxygen, and it also plays a vital role in the host immunity [6]. Zn takes part in the growth, oxidation resistance and immunity [7]. Mn as the part of phosphoenolpyruvate carboxykinase takes part in the gluconeogenesis, and it is related to the neuronal health [8].Pig (Sus scrofa) is one of the most raised animals in the world. Piglets are weaned early to increase the reproductive performance of the sow and to reduce pathogen transmission [9]. However, as the digestive system of piglets is immature, early weaning will lead to maldigestion [10]. Meanwhile, because of rapid and dramatic change of the living environment and exposure to the bacteria [11], early weaning piglets easily suffer from stress, which reduces the growth performance and feed intake [12] and decreases the nutrient digestibility through digestive disorders [13]. It leads to the resources waste and environment pollution and limits the sustainable development of animal husbandry. Thus, it is urgent to look for an effective additive to relieve early weaning stress and improve the absorption of nutrition. The absorption of nutrition is closely related to the intestinal health, however, early weaning stress increases the intestinal permeability of piglets which has a negative effect on the absorption of nutrients [14]. Lipopolysaccharide (LPS) is the primary component of Gram-negative bacteria outer cell walls [15] and it can induce severe bacterial diarrhea, apoptosis [16], inflammatory responses [17], intestinal barrier damage [18], and then inhibits the growth performance and decreases the nutrients absorption of the animal [19]. Due to its good repeatability, the LPS stress mode is widely used in research.Many of growth factors exist in milk, such as insulin, nerve growth factor (NGF), and epidermal growth factor (EGF) which can improve the intestinal development of piglets and thus improve their growth performance [20]. Early weaning prevents the supply of those growth factors from milk to piglets. Interestingly, EGF is one of the most abundant growth factors in milk [21,22], which indicates its important role for young mammals. EGF was first isolated by Dr. Cohen from the mouse (Mus musculus) submaxillary gland in 1962 [23]. It is a polypeptide comprising 53 amino acids [24]. It is found in many body fluids such as the milk, blood, saliva, and intestinal fluid [25], and it plays important roles in the regulation of cell growth, proliferation, apoptosis and tumorigenesis [26,27,28]. Previous studies showed that EGF could improve the growth performance of broiler chicks (Gallus gallus) [27] and rats (Rattus norvegicus) [29]. Dietary EGF can augment the intestinal length and villus height by activating the phosphatidylinositol-3-kinases/protein-serine-threonine kinase (PI3K/AKT) and RAS/mitogen-activated protein kinase (RAS/MAPK) signaling pathways [30,31]. Meanwhile, EGF can also promote the proliferation of goblet cells [10] and increase the activity of digestive enzymes in the intestine [32]. However, the effect of EGF on growth performance and nutrients absorption in LPS challenged early-weaning pigs is unclear. Whether it can be added as an effective additive in the feed of early weaning piglets is still unknown. In this experiment, a model of LPS stress was established to examine how EGF attenuates the effect of LPS on the growth performance, nutrient digestibility, microelement absorption of early-weaned pigs, and study the underlying mechanism.2. Materials and Methods2.1. Experimental DesignA total of 48 Duroc × Landrace × Large White early weaned piglets (castrated male pigs, average initial weight was 7.84 ± 0.30 kg), aged 25 days, were randomly distributed among four groups (control, EGF, LPS and EGF + LPS groups) which consisted of a 2 × 2 factorial design. Each group had four replicates and each replicate consisted of three piglets. The main factors were the level of LPS (HLPS = high LPS: 100 μg/kg body weight; ZLPS = low LPS: 0 μg/kg body weight) and EGF (HEGF = high EGF: 2 mg/kg diet; ZEGF = low EGF: 0 mg/kg diet). Piglets in the LPS and EGF + LPS groups were intraperitoneally injected with the 100 μg/kg body weight LPS (Sigma-Aldrich, Saint Louis, MO, USA) at 7 and 15 days during the experiment [33]. Meanwhile, the control and EGF groups were injected with the corresponding volume physiological saline (Nanjing Jiancheng Biotechnical Institute, Nanjing, China). The control and LPS groups were fed the basal diet (diet 1) which met the nutrient requirements of pigs according to NRC 2012 (Table 1). The piglets in the EGF and EGF + LPS groups were fed the basal diet supplemented with 2 mg/kg EGF (diet 2, Peprotech, Rocky Hill, CT, USA). The experiment lasted for 14 days and the pigs had ad libitum access to feed and water during this period. The humidity ranged from 50% to 70%, and the temperature ranged from 18 to 22 °C. The pigs were fasted for 24 h and were weighed in the morning at 1 day and 15 days during the experiment, and feed intake was recorded every day. At the end of the trial, initial body weight (IBW), final body weight (FBW), average daily feed intake (ADFI), average daily gain (ADG), and feed conversion ratio (FCR, feed/gain) were calculated.2.2. Sample CollectionFeces were collected from days 11 to 14 during the trial and were stored at −20 °C. At the end of the experiment, all pigs were slaughtered 4 h after the final injection of LPS. Before slaughter, all piglets were euthanized with Zoletil (active compound: tiletamine and zolazepam, Virbac, Beijing, China) at 15 mg/kg body weight. The chyme samples from the stomach, jejunum and ileum were collected and immediately frozen at −20 °C. The stomach, duodenum, jejunum and ileum samples were washed with saline solution, and then the mucosa of these samples was collected by glass slide and immediately frozen at −80 °C for Q-RT-PCR analysis.2.3. Nutrient Digestibility and Essential Microelements ConcentrationThe diet, feces and chyme samples were dried at 105 °C. Then, they were ground into a fine powder and passed through a 40 μm mesh. Gross energy, crude protein, crude fat, crude fiber, and P were tested according to the methods of the Association of Official Analytical Chemists International, 2007. The digestibility of nutrients was calculated as follows:Digestibility (%) = 100 − (Id ÷ Is) × (Ns ÷ Nd) × 100%,(1) where the Id and Is are the concentration of the acid-insoluble ash in the diet and the feces, respectively, and Ns and Nd are the concentration of the nutrient in the feces and the diet, respectively.Samples of diet, feces and chyme were digested in the concentrated nitric acid and perchloric acid mixture solution (the addition ratio of concentrated nitric acid and perchloric acid was 4:1) to dissolve the Cu, Fe, Zn and Mn (GB/T 23942-2009), and the concentration was analyzed by electron coupled plasma atomic emission spectrum (Ke Jie Instrument Limited Company, Nanjing, China).2.4. Quantitative Real-Time PCR (Q-RT-PCR) AnalysisThe relative expression levels of zrt-irt-like protein 4 (Zip4), zrt-irt-like protein 7 (Zip7), zinc transporter 1 (ZnT1), zinc transporter 4 (ZnT4), copper transport protein 1 (Ctr1), cytochrome c oxidase copper chaperone 17 (Cox17), antioxidant 1 (Atox1), copper-transporting P-type 7A (ATP7A), copper-transporting P-type 7B (ATP7B), copper chaperone for superoxide dismutase (CCS), divalent metal transporter 1 (DMT1), cytochrome b (CYTB), hephaestin (Hp), and transferrin (Tf) in the mucosa of the stomach, duodenum, jejunum and ileum were detected by Q-RT-PCR. The primers (Sangon Biotech, Shanghai, China) used are listed in Supplementary Table S1. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene was chosen as the reference gene for sample normalization. Total RNA from the intestinal tissue was extracted using the TRIzol reagent (Invitrogen, Carlsbad, CA, USA). The integrity of each RNA sample was estimated by 1% agarose gel electrophoresis (Sangon Biotech, Shanghai, China). The cDNA was synthesized using a SMART cDNA Synthesis Kit (Clontech Laboratories, Palo Alto, CA, USA) by following the manufacturer’s protocol. Q-RT-PCR reactions were carried out in a BIO-RAD CFX96 touch Q-PCR system (Applied Biosystems, Foster City, CA, USA) in 20 μL volumes that contained the following components: 10 μL of SYBR Green Mix (Takara, Changsha, China), 2 μL cDNA (1000 ng·μL−1), 0.4 μL of each primer (10 mM) and 7.2 μL dH2O, followed by 40 cycles of 95 °C for 30 s, 55 °C or 58 °C for 30 s, and 72 °C for 30 s. Finally, a melt curve analysis was used to detect the single product (temperature from 65 to 95 °C). All samples were tested in triplicate. The 2−ΔΔCT method was used to analyze the relative expression level. The standard curve was obtained by using 5-fold serial dilutions of cDNA (in triplicate), and the amplification efficiencies of all primes ranged from 0.90 to 1.00.2.5. Statistical AnalysisThe experimental design was a 2 × 2 factorial design while the main factors were the level of LPS and EGF. Data were analyzed by 2-way ANOVA using SPSS 23.0 (SPSS. Inc., Chicago, IL, USA), which included the main effects of LPS level, EGF level and their interaction (LPS level × EGF level). Tukey’s multiple range test was used to analyze the differences. All data were further subjected to one-way ANOVA. When overall differences were significant, the differences were tested by Duncan’s multiple-range test (SPSS 22.0). The data about the concentration of essential microelements in the diets were subjected to independent-samples T test (SPSS 22.0). The level of significance was set at p < 0.05. The results are presented as the mean values and standard error of mean (SEM).3. Results3.1. Cu, Fe, Zn, and Mn Concentration in DietsThe concentration of essential microelements in diets 1 and 2 are shown in Table 2. There were no significant differences of Cu, Fe, Zn, and Mn concentration between diet 1 and diet 2 (p > 0.05).3.2. Growth PerformanceThe effect of LPS and EGF levels on growth performance are shown in Table 3. The LPS level affected ADG and FCR of piglets, and EGF level affected ADFI (p < 0.05). The LPS and EGF levels displayed a significant interaction effect on ADFI (p < 0.05). Piglets injected with HLPS level significantly decreased the ADG and significantly increased the FCR compared with the piglets injected with ZLPS level (p < 0.05). The ADFI of piglets fed HEGF level was 12% higher than the piglets fed ZEGF level (p < 0.05). The lowest ADFI was observed in the control group, which significantly differed from the other three groups (p < 0.05)3.3. Nutrient Apparent DigestibilityExcept for the apparent digestibility of crude fat, no differences were observed in the nutrient apparent digestibility, and there were no interactions between LPS and EGF levels regarding the apparent digestibility (p > 0.05, Table 4). Compared with ZLPS level, HLPS level significantly decreased the apparent digestibility of crude fat (p < 0.05). The lowest apparent digestibility of crude fat was observed in the LPS group, which was significantly differed from the other groups (p < 0.05), and there was no significant difference between the EGF + LPS and control groups (p > 0.05).3.4. Concentration of Cu, Fe, Zn, Mn in the Gastrointestinal Chyme and FecesThe concentration of Cu, Fe, Zn, Mn in the gastrointestinal chyme and feces are shown in Table 5. The present study revealed significantly interactions between the EGF and LPS levels regarding the Cu concentration in the stomach, jejunum and ileum chyme (p < 0.05), and there was no interaction in the feces (p > 0.05). Piglets injected with HLPS level decreased the Cu concentration in the ileum chyme and increased the Cu concentration in the jejunum chyme compared the piglets injected with ZLPS level (p < 0.05). The Cu concentration in the jejunum, ileum chyme and feces of piglets fed HEGF level were 18%, 55%, and 28% lower than those of piglets fed ZEGF level, respectively (p < 0.05). The LPS group had a significantly greater Cu concentration in the jejunum chyme compared with the other groups (p < 0.05), while there was no significant difference between the EGF + LPS and control groups (p > 0.05). The control and LPS groups had significantly greater Cu concentration in the ileum chyme compared with the EGF and EGF + LPS groups (p < 0.05).The LPS and EGF levels displayed a significant interaction effect on the Fe concentration in the feces (p < 0.05). Piglets injected with HLPS level decreased the Fe concentration in the ileum chyme and increased the Fe concentration in the jejunum chyme and feces compared with the piglets injected with ZLPS level (p < 0.05). The Fe concentration in the ileum chyme and feces of piglets fed HEGF level were 35%, and 31% lower than those of piglets fed ZEGF level, respectively (p < 0.05). In the feces, the LPS group had a significantly greater Fe concentration compared with the other groups (p < 0.05), and the EGF + LPS group had a significantly lower Fe concentration compared with the control group (p < 0.05).The LPS and EGF levels displayed significant interactions on the Zn concentration in the stomach and ileum chyme, and feces (p < 0.05). The Zn concentration in the jejunum, ileum chyme and feces of piglets injected with HLPS level were 12%, 53%, and 19% higher than those of piglets fed ZLPS level, respectively (p < 0.05). The Zn concentration in the jejunum, ileum chyme and feces of piglets fed HEGF level were 12%, 68%, and 28% lower than those of piglets fed ZEGF level, respectively (p < 0.05). In the ileum chyme, a significantly greater Zn concentration was observed in the LPS group, which significantly differed from the other groups (p < 0.05), and the EGF + LPS group had a significantly lower Zn concentration compared with the control group (p < 0.05). In the feces, the LPS group had a significantly greater Zn concentration compared with the other groups (p < 0.05), while there was no significant difference between the EGF + LPS and control groups (p > 0.05).Piglets injected with HLPS level significantly increased the Mn concentration in the stomach chyme and feces compared with the piglets injected with ZLPS level (p < 0.05). The Mn concentration in the jejunum, ileum chyme and feces of piglets fed HEGF level were 29%, 15%, and 27% lower than those of piglets fed ZEGF level, respectively (p < 0.05). In the stomach chyme, the LPS and EGF + LPS groups had significantly greater Mn concentration compared with the control and EGF groups (p < 0.05). In the feces, the LPS group had a significantly greater Mn concentration compared with the other groups (p < 0.05), and the EGF + LPS group had a significantly lower Mn concentration compared with the control group (p < 0.05).3.5. Expression of Cu Transport-Relative Genes in the Mucosa of the Gastrointestinal TissuesAs shown in Table 6, the expression levels of the Cu transport-related genes in the mucosa from the gastrointestinal tissues were affected by LPS and EGF levels (p < 0.05). In the stomach, the LPS and EGF levels displayed significant interaction effects on the expression levels of Cox17, Atox1, ATP7A, and ATP7B (p < 0.05). Piglets injected with HLPS level significantly decreased the expression levels of Atox1 and ATP7B compared with the piglets injected with ZLPS level (p < 0.05), and piglets supplied with HEGF level significantly increased the expression levels of Ctr1, Cox17, Atox1, ATP7A, and ATP7B compared with the piglets supplied with ZEGF level (p < 0.05). The LPS and EGF + LPS groups had a significantly lower expression level of Atox1 compared with the control and EGF groups (p < 0.05), whereas there was no significant difference between the EGF + LPS and LPS groups (p > 0.05). The LPS group had a significantly lowest expression level of ATP7A compared with the other groups (p < 0.05), and EGF + LPS group had a significantly higher expression level compared with the LPS and control groups (p < 0.05).In the duodenum, there were significant interactions between LPS and EGF levels in the expression levels of Cox17, Atox1, and ATP7B (p < 0.05). Compared with the piglets injected with ZLPS level, piglets injected with HLPS level significantly decreased the expression levels of Cox17, ATP7A, ATP7B, and CCS (p < 0.05). Piglets supplied with HEGF level significantly increased the expression levels of Cox17, ATP7A, and ATP7B compared with the piglets supplied with ZEGF level (p < 0.05). The LPS group had the significantly lowest expression level of Cox17 compared with the other groups (p < 0.05), whereas there was no significant difference between the EGF + LPS and control groups (p > 0.05). The LPS group had a significantly lower expression level of ATP7A compared with the control and EGF groups (p < 0.05), while there was no significant difference between the EGF + LPS and control groups, or between the EGF + LPS and LPS groups (p > 0.05).In the jejunum, the LPS and EGF levels displayed a significant interaction effect on the expression level of CCS (p < 0.05). The HLPS level significantly decreased the expression levels of Ctr1, Atox1, and CCS compared with the ZLPS level, and the HEGF level significantly increased the expression levels of Ctr1, Atox1, ATP7A, ATP7B, and CCS compared with the ZEGF level (p < 0.05). The LPS group had the significantly lowest expression levels of Ctrl and Atox1 compared with the other groups (p < 0.05), while there was no significant difference between the EGF + LPS and control groups (p > 0.05).In the ileum, there was a significant interaction between LPS and EGF levels in the expression level of Atox1 (p < 0.05). Piglets injected with HLPS level significantly decreased the expression levels of Cox17, ATP7B, and CCS compared with the piglets injected with ZLPS level (p < 0.05), and piglets supplied with HEGF level significantly increased the expression level of Atox1 compared with the piglets supplied with ZEGF level (p < 0.05). The LPS and EGF + LPS groups had significantly lower expression levels of Cox17, ATP7B, and CCS compared with the control and EGF groups (p < 0.05), whereas there was no significant difference between the EGF + LPS and LPS groups (p > 0.05). The LPS group had the significantly lowest expression level of Atox1 compared with the other groups (p < 0.05), and the EGF + LPS group had a significantly greater expression level compared with the control group (p < 0.05).3.6. Expression of Fe Transport-Relative Genes and DMT1 Gene in the Mucosa of the Gastrointestinal TissuesThe expression levels of Fe transport-related genes and DMT1 gene in the mucosa of the gastrointestinal tissues are shown in Table 7. In the stomach, the LPS and EGF levels displayed a significant interaction effect on the expression level of CYTB (p < 0.05). Piglets injected with HLPS level significantly decreased the expression level of Tf compared with the piglets injected with ZLPS level (p < 0.05), and piglets supplied with HEGF level significantly increased the expression levels of CYTB, Hp, Tf, and DMT1 compared with the piglets supplied with ZEGF level (p < 0.05).In the duodenum, there were significant interactions between LPS and EGF levels in the expression levels of CYTB and DMT1 (p < 0.05). Compared with the piglets injected with ZLPS level, piglets injected with HLPS level significantly decreased the expression levels of Tf and DMT1 (p < 0.05). Supply with HEGF level significantly increased the expression levels of CYTB and DMT1 compared with the piglets supplied with ZEGF level (p < 0.05). The LPS and EGF + LPS groups had significantly lower expression levels of Tf and DMT1 compared with the control and EGF groups (p < 0.05), whereas there was no significant difference between the EGF + LPS and LPS groups (p > 0.05).In the jejunum, the HLPS level significantly decreased the expression levels of CYTB, Tf, and DMT1 compared with the ZLPS level, and the HEGF level significantly increased the expression levels of Hp and DMT1 compared with the ZEGF level (p < 0.05). The LPS and EGF + LPS groups had significantly lower expression level of CYTB compared with the control and EGF groups (p < 0.05), whereas there was no significant difference between the EGF + LPS and LPS groups (p > 0.05). The LPS group had a significantly lower expression level of DMT1 compared with the control and EGF groups (p < 0.05), while there was no significant difference between the EGF + LPS and control groups, or between the EGF + LPS and LPS groups (p > 0.05).In the ileum, there were significant interactions between LPS and EGF levels in the expression levels of Hp and DMT1 (p < 0.05). Piglets injected with HLPS level significantly decreased the expression level of DMT1 compared with the piglets injected with ZLPS level (p < 0.05), and piglets supplied with HEGF level significantly increased the expression levels of CYTB and Hp compared with the piglets supplied with ZEGF level (p < 0.05). The LPS group had the significantly lowest expression level of Hp compared with the other groups (p < 0.05), and the EGF + LPS group had a significantly greater expression level compared with the control group (p < 0.05). The LPS group had a significantly lower expression level of DMT1 compared with the other groups (p < 0.05), but there was no significant difference between the EGF + LPS and control groups (p > 0.05). 3.7. Expression of Zn Transport-Relative Genes in the Mucosa of the Gastrointestinal TissuesThe expression levels of Zn transport-related genes in the mucosa from the gastrointestinal tissues are shown in Table 8. The present study revealed significant interactions between EGF and LPS levels regarding the expression level of Zip4 in the stomach, and the expression levels of Zip7 in the stomach and ileum (p < 0.05). Injected HLPS level significantly decreased the expression levels of Zip4 and Zip7 in the stomach and ileum compared with the ZLPS level (p < 0.05) and supplied HEGF level significantly increased the expression level of Zip4 in the stomach, jejunum and ileum compared with the ZEGF level. The LPS group had a significantly lower expression level of Zip4 in the ileum compared with the other groups (p < 0.05), whereas there was no significant difference between the EGF + LPS and control groups (p > 0.05). The LPS and EGF + LPS groups had significantly lower expression levels of Zip7 in the stomach and ileum compared with the control group (p < 0.05), while there was no significant difference between the LPS and EGF + LPS groups (p > 0.05). For ZnT1, injected HLPS level significantly increased the expression level of it in the stomach compared with the ZLPS level (p < 0.05), and supplied HEGF level significantly increased the expression level of it in the stomach compared with ZEGF level (p < 0.05).4. DiscussionRecently, the application of EGF has received increasing amounts of attention due to its positive impacts on animals [26,27,28]. Previous studies had shown that LPS significantly decreased the ADG and ADFI of weaned piglets [34], and EGF could increase body weight gain of broiler chickens and early-weaned mice [27,35], and increase gain/feed of early-weaned pigs [36]. Our results indicated that injected HLPS level significantly decreased the ADG and significantly increased the FCR, and dietary HEGF level significantly increased the ADFI of early-weaned piglets, which were in agreement with the previous studies. Our results also indicated that injected HLPS level significantly decreased the apparent digestibility of crude fat. The changes of growth performance induced by LPS was related to the changes of nutrients absorption. LPS leads to partial loss and sloughing of ileal villi and decreases the intestinal barrier function in mice [37]. LPS also increases the intestinal epithelial cell permeability [38]. LPS reduced the apparent digestibility maybe through reducing the intestinal health. Previous study had showed that dietary EGF had no significant influence on the apparent digestibility of crude protein, gross energy and P [39]. Our results also indicated that dietary HEGF level had no significant influence on the apparent digestibility, which was in agreement with the previous studies.Indispensable microelements take part in the regulation of the body physiological functions, such as participating in the redox active [1,2,3,4], oxygen transport, DNA biosynthesis [40], cellular signal recognition [41,42], and nutrients metabolism [43,44]. A higher concentration of microelements in the gastrointestinal chyme and feces means a lower absorption level. Cu as a cofactor plays an essential role in redox-active, pigmentation, oxidative phosphorylation and neuropeptide biogenesis [45,46]. Our results showed that injected HLPS level increased the concentration of Cu in the jejunum chyme, and dietary HEGF level decreased the concentration of Cu in the jejunum, ileum chyme, and feces. ATP7A, ATP7B, Cox17, Ctr1, Atox1, and CCS genes are Cu transport-related genes. Ctr1 is a major Cu extracellular uptake protein and involved Cu transport across membranes [47]. Atox1, CCS, and Cox17 are metallochaperones: Cox17 transport Cu to the mitochondria, and CCS transport Cu to combine the SOD in the cytoplasm and mitochondria, and Cox17 transport Cu to combine the ATP7A and ATP7B [48]. ATP7A and ATP7B are P-type Cu-ATPases which transport Cu to the ceruloplasmin and lysyl oxidase and take part in the Cu exportation from the cell [49]. Our results showed that in the mucosa of gastrointestinal tissues, injected HLPS level decreased the expression levels of the ATP7A, ATP7B, Cox17, Ctr1, Atox1, and CCS genes, while dietary HEGF level increased the expression levels of these genes. These results explained how LPS and EGF regulated the absorption of Cu in the gastrointestinal tissues. However, injected HLPS level decreased the concentration of Cu in the ileum, and the underlying reason needs further analysis.Fe is the transporter of oxygen, and it takes part in the redox reaction, electron transport, cell growth, and energy production [50,51]. Our results showed that injected HLPS level increased the concentration of Fe in the jejunum chyme and feces, while dietary HEGF level decreased the concentration of Fe in the ileum chyme and feces. CYTB, Tf and HP are Fe transport-related genes. CYTB and HP are oxidoreductases: CYTB as ferric reductase changes ferric iron to ferrous iron [52], while Hp is expressed in the enterocyte and oxidizes ferrous iron [50,53]. CYTB co-operate with DMT1 to transfer Fe from the duodenal lumen to the enterocyte, while Hp co-operate with ferroportin to transfer Fe from the basolateral membrane to the systemic circulation [50]. Tf is synthesized almost in the liver [54], and it can regulate iron homeostasis and erythropoiesis [55]. As an important iron carrier in the blood [53], Tf delivers iron to the tissues [54]. Our results showed that in the mucosa of gastrointestinal tissues injected HLPS level decreased the expression levels of the CYTB and Tf genes, and dietary HEGF level increased the expression levels of the CYTB, Tf and HP genes. Consequently, LPS and EGF affected the absorption of Fe in the gastrointestinal tissues by regulating the expression levels of Fe transport-related genes.Mn as enzymatic cofactors or structural centers takes part in a plethora of biological processes, such as glycosylation, signal transduction, phosphorylation, and hydrolysis [56,57]. It also takes part in the host immune system [58]. Our results showed that injected HLPS level increased the concentration of Mn in the stomach chyme and feces, while dietary HEGF level decreased the concentration of Mn in the jejunum and ileum chyme and feces. DMT1 is a multiple divalent metals transport gene, and it can transfer the Cu, Mn, Zn [59]. Our results showed that in the mucosa of gastrointestinal tissues injected with HLPS level downregulated the expression level of the DMT1 gene, while dietary HEGF level upregulated the expression level of the DMT1 gene. That was the one reason why the LPS and EGF changed the absorption of Cu, Mn, and Zn in the piglets.Zn is the cofactor for many enzymes and takes part in many biological processes [60], and it plays the important roles in protein synthesis, growth, and immunity [61]. Our results showed that in the jejunum and ileum chyme, and feces, injected HLPS level increased the concentration of Zn, while dietary HEGF level decreased the concentration of Zn. In the body of vertebrates two kinds of zinc transporter family proteins exist, ZIP and ZnT family [62]. Zip4, Zip7 and ZnT1 are the Zn transport-related genes. Zip4, Zip7 belong to the ZIP family and take part in the import of Zn to the cytoplasm. Zip4 is a primary importer for the absorption of Zn in the enterocyte, and it can transfer Zn from intestine lumen to the epithelial cells [63]. Zip7 exists in the membrane of endoplasmic reticulum and Golgi apparatus, and transfers Zn to the cytosol [64]. ZnT1 belongs to the ZnT family, which predominantly localizes in the basolateral membrane [65]. In the intestinal epithelial cells, it takes part in the export of Zn from the cytoplasm to the portal vein [65]. Our results showed that in the mucosa of gastrointestinal tissues, injected HLPS level decreased the expression levels of the Zip4, Zip7 and ZnT1 genes, while dietary HEGF level increased the expression levels of the Zip4 and ZnT1 genes. It is implied that the bioavailability of Zn was affected by LPS and EGF through regulating the expression of Zn transport-related genes.5. ConclusionsIn conclusion, the present findings suggested that intraperitoneal injection with HLPS level increased the FCR, and decreased the ADG, apparent digestibility of crude fat, and absorption of Cu Fe, Zn, Mn in the early-weaned pigs. Dietary EGF could reduce the adverse effect of LPS exposure on apparent digestibility of crude fat and microelement absorption of early-weaning piglets. EGF and LPS influenced the absorption of essential trace element through changing the expression levels of Zip4, Zip7, ZnT1, Ctr1, Atox1, CCS, Cox17, ATP7A, ATP7B, DMT1, CYTB, Hp and Tf genes in the mucosa of gastrointestinal tissues. Hence, EGF can be used as an additive to increase the essential trace elements absorption in the early weaning piglets.

animals : an open access journal from mdpi

10.3390/ani13101644

PMC10215717

In commercial pork production, piglets are weaned at a fairly young age. Piglets are removed from the sow and are subject to sudden changes in their diet, environmental conditions and social grouping. For the piglet, this causes major upheaval and disruption to eating and behavioral patterns, leading to distress, gastrointestinal tract dysfunction and behavioral disorders after weaning. This contrasts with natural conditions where nest-leaving of the sow and her offspring and joining the herd with increasing interactions with non-littermate piglets occurs between approximately 6.5 and 15 days after birth, with weaning typically completed between 14 and 18 weeks of age. Strategies that could be adapted to a commercial setting and that allow socialization of non-littermate piglets before weaning and gradual separation of the litter from the sow in the period leading up to weaning may reduce the stress occurring after weaning and improve performance, welfare and gastrointestinal function. This review summarizes current knowledge concerning such strategies.

The weaning of pigs in most commercial pork production systems is an abrupt event performed at a fairly young age, i.e., mostly between 2.5 and 5 weeks of age. This practice induces a stress response, and its impact on behavior, performance and the gastrointestinal tract has been well described. Historically, there has been a focus on pre- and post-weaning nutritional strategies and post-weaning housing conditions and medication to improve production and reduce mortality after weaning. However, alternative pre-weaning housing and management systems that promote the development of natural social behaviors of piglets before weaning have recently received more attention. Co-mingling of non-littermates before weaning is a strategy that aims to initiate social interactions prior to weaning. The separation of the litter from the sow in the period leading up to weaning, termed intermittent suckling, aims to enhance the gradual separation from the sow. In addition, these practices encourage the young pig to learn explorative nutrient sourcing. Altogether, they may reduce weaning-associated stress. In this review, these strategies are defined, and their effects on behavior, performance, mortality, gastrointestinal function and immunocompetence are described. Though these strategies may be adapted to a commercial setting, it also becomes clear that many factors can contribute to the success of these strategies.

1. IntroductionThe weaning of piglets in most commercial pork production systems worldwide is an abrupt event and is performed at a fairly young age. Piglets are simultaneously removed from maternal care, mixed and moved into a new environment, and offered solid feed devoid of the antibodies and other gut protective factors and stimulants found in sow’s milk. This series of events causes an upregulation of the hypothalamic-pituitary-adrenal (HPA) axis and, accordingly, an acute increase in cortisol levels, irrespective of weaning age, indicative of physiological and psychological stress [1]. The weaning-associated stress response and its impact on performance and the gastrointestinal tract (GIT) have been well described [2,3,4,5,6]. Historically, there has been a focus largely on pre- and post-weaning nutritional strategies and post-weaning housing conditions and medication to improve production and reduce morbidity and mortality after weaning [7,8,9]. However, alternative pre-weaning housing and management systems that promote the development of natural social behaviors in piglets before weaning have recently received more attention for their potential to ease the weaning transition and improve performance and health.With this in mind, the general aim of this review was to summarize the literature related to two techniques, namely co-mingling of non-littermates before weaning and intermittent suckling (IS), on the effects of weaning-associated stress and post-weaning performance. Co-mingling of non-littermates before weaning is a strategy that aims to initiate social interactions prior to weaning. The separation of the litter from the sow in the period leading up to weaning, termed IS, aims to enhance the gradual separation from the sow. In addition, these practices encourage the young pig to learn explorative nutrient sourcing, which may better prepare it for weaning. Together, they may reduce weaning-associated stress. The factors that contribute to the success of these strategies will be discussed in detail. In order to fully understand how these two pre-weaning techniques may affect the pig and its response upon weaning, an in-depth understanding of the development of social and eating behaviors of pigs in (semi-)natural conditions and of the effects of commercial weaning practices on piglets’ behavior and stress response is warranted. Therefore, first, the related literature is summarized in the two following sections.2. Development of Social and Eating Behaviors of Pigs in (Semi-)Natural ConditionsIn the late 20th century, an elegant series of studies was conducted to provide insights into the social and eating behaviors of domestic pigs when kept in (semi-)natural conditions (Table 1). The weaning-associated behaviors described in these studies are believed to be very similar to those of wild pigs. The following section describes these studies’ major findings and possible implications for piglets raised commercially.2.1. Nest Occupation and Maternal CareSows isolate themselves from the herd to build a nest, to farrow and to care for their young [10]. After birth, the first interaction between the sow and her offspring, i.e., nose-to-nose contact and sniffing, occurs before piglets suck at the teat for the first time and serves to quickly establish the maternal bond. This first contact takes place earlier when it is initiated by the sow than when initiated by the piglet (20.2 vs. 35.7 min post-partum) [11]. This selective bonding allows the mother and her young to quickly identify each other and is crucial to ensuring maternal care and survival [12]. Newborn piglets have functional sensorial systems (e.g., auditory and visual) for early learning, and, through neurobiological mechanisms and neurotransmitters, behavioral preference for filial individuals is established. Although newborn piglets have a high degree of locomotory development, they will remain in the nest for the days to come [13]. During the first day after farrowing, more than 85% of all sucklings are initiated [14,15] and less than 5% of all sucklings are terminated [15] by the sow. The latter occurs to prevent fights at the udder due to competition for teats between littermates [16]. The synchronization of suckling bouts is already developed during colostrum provision, and complete suckling synchronization is achieved by day 3 [17]. The sow initiates suckling by grunting, approaching and nosing her piglets, after which the piglets awaken, approach the udder and start sucking [16]. Within 3 days of birth, piglets will develop a teat preference [11]. Piglets sniff the udder and sample different teats, and fighting between littermates is often involved [16]. Piglets stay in the nest during the first three days after birth, only leaving the nest to urinate and defecate [18]. From day 4 post-partum, piglets follow the dam outside the nest, but they still remain close to her. As piglets age, their activity increases and they follow their mother over larger distances [18,19]. Hence, the first encounters with other members of the herd take place. These approaches are mainly nose-to-nose contact and occur peacefully [20].2.2. Social Integration within the HerdBetween approximately 6.5 and 15 days after birth, the sow leaves the nest together with her offspring [10] and occupies a new sleeping ground every night in proximity to the herd [21], finally joining the herd at about 9 days post-partum [15,18]. Approximately 10 days after nest leaving (up to week 8), piglets spend less time at the udder (on average 17% of the observation time) and undertake other activities more, such as walking and standing (on average 21% of the observation time) [22]. When pigs become older, they need more food than the sow can provide or is willing to supply. Therefore, piglets try to initiate suckling by approaching the udder, whining, sucking and massaging the teats. However, the mother still determines the precise timing of suckling [16], and sucklings are still ended by the sow. Jensen (1988) [23] concluded that the percentage of sucklings terminated by the sow significantly increased between 1 and 4 weeks post-partum. For example, Jensen et al. (1991) [15] showed that, on day 1 post-partum, less than 5% of the sucklings were terminated by the sow, whereas this was about 60% on day 10 post-partum. The ratio of time spent at the udder versus other activities starts to change. Piglets are more confident in exploring greater distances from the sow, which presents more opportunities for social contact with non-littermates. This is especially evident during sow feeding events. Between days 7 and 18 after farrowing, the piglets attend, with their mother, for the first time, the feeding sites [10,20] and then continue to do this every morning. When the sow is eating with other group members, piglets are resting at a distance of 10–20 m from the feeding site, where the occasion for social interaction with other group members again increases [20]. Since this feeding takes place in a small area, the likelihood of social interaction with non-littermates increases accordingly. Hence, between weeks 2 and 7 post-partum, social interactions between piglets and other members of the group are numerous and frequent. It seems that, after week 7, the process of social integration, which can be defined as the process whereby pigs become accepted and form stable social bonds within the herd, is completed. According to Petersen et al. (1989) [20], social interactions between weeks 2 and 5 mainly occur amongst littermates, whereas between 8 and 12 weeks, interactions with non-littermates and other members of the group become as important. In addition, this study demonstrated that 60.9% of all observed social interactions amongst piglets and other animals between weeks 2 and 17 post-partum consisted of nose-to-nose contact. Only 17.3% of all interactions were considered aggressive, i.e., head knocking, shoulder-to-shoulder movements and neck biting. As piglets mature, their activity increases, but the frequency of social interactions decreases. Petersen et al. (1989) [20] showed that, in week 2 after farrowing, there were 25.7 interactions per hour, while in weeks 7, 8–12 and 15–17, it declined to 5.3, 5.6 and 3.9 interactions per hour, respectively. During this time, piglets also start to explore foods other than milk. For example, between days 24 and 36 after farrowing, piglets start grazing and eating at the feeding place [20].2.3. Slow and Gradual WeaningPiglets are weaned naturally between 14 and 18 weeks of age, according to observations by Jensen (1986) [10], Jensen and Recén (1989) [24], Jensen and Stangel (1992) [25], Newberry and Wood-Gush (1985) [16] and Petersen et al. (1989) [20]. The weaning process of piglets in (semi-)natural conditions is slow and gradual. Suckling-related activities slowly decline and consequently disappear when the weaning process is completed [23,24], in part caused by changes in dam behavior. Jensen (1988) [23] observed that the proportion of sucklings initiated by the sow decreased significantly between weeks 4 and 10 post-partum. From week 10 post-partum, the decrease in the number of sucklings is even more pronounced, further stimulating the piglets to feed themselves with sources other than milk. Jensen (1986) [10] and Jensen and Recén (1989) [24] indicated that weaning involves a rapid decline or atrophy of the udder of the sow. When there was no suckling for two consecutive days, they observed a clear degeneration of the udder 2–3 days afterwards [24]. The weakening of the family bond between the sow and her litter is also an indication of the weaning process [25]. To our knowledge, no study has looked at the stress response under (semi-)natural conditions of weaning, but it is probable that pigs may not experience a growth check during this period [26], as observed following abrupt weaning as practiced commercially.animals-13-01644-t001_Table 1Table 1Literature overview of characteristics of pigs in (semi-)natural conditions.ReferenceBreedGroup Size and CompositionEnvironmentDistance from Farrowing Nest to the Group (m)Litter Size (Mean and [Range])Weaning Age (Mean and [Range]) (Weeks Post-Partum)Castrén et al., 1989 [14]Swedish Landrace-Group 1:4 gilts -Groups 2 and 3:1 old boar, 4 adult sows and offspring, 4 giltsTovetorp Research Station, Stockholm, Sweden3 adjacent enclosuresEnclosure 1: group 1, primary forest, 1 haEnclosures 2 and 3: groups 2 and 3, several biotopes: open field, swamps, mossy areas, primary forest, large fir plantations, 7 and 13 ha [5–13] Jensen, 1986 [10]Swedish Landrace-Group 1: 1 boar, 2 gilts -Group 2:1 boar, 1 sow and offspring, 2 giltsTovetorp Research Station, Stockholm, SwedenHilly area with few steep slopes2 adjacent enclosures:several different biotopes, comprising open fields, swamps, mossy areas, primary forest, large fir plantations, 7 and 13 ha225–550 [14–17]Jensen, 1988 [23]Swedish Landrace2 groups each consisting of 1 older and 1 younger boar, 3–4 adult sows and offspring, 4 giltsThe same as in the study of Jensen (1986) [10] 7.5 Jensen et al., 1987 [27]Swedish Landrace2 groups each consisted of 1 older and 1 younger boar, 3–4 adult sows and offspring, 4 giltsThe same as in the study of Jensen (1986) [10] Jensen et al., 1991[15]Swedish Landrace-Group 1: 1 boar, 4 sows and offspring, 4 gilts -Group 2:1 older and 1 younger boar, 4 adult sows and offspring, 4 giltsThe same as in the study of Jensen (1986) [10] Jensen and Recén, 1989 [24]Swedish Landrace2 groups each consisting of 5–10 sows and offspring, 1–2 adult boarsThe same as in the study of Jensen (1986) [10] [7–10]17.2 [15.6–19.5]Jensen and Redbo, 1987 [21]Swedish Landrace2 groups each consisting of 1 old and 1 young boar, 2–3 adult sows and offspring, 4 giltsThe same as in the study of Jensen (1986) [10]150–250 Jensen and Stangel, 1992 [25]Swedish Landrace1 boar, 5–8 sows and offspring, 0–4 giltsHilly areaEnclosure contains several biotopes: pastures, swamps, forests, large fir plantations, small streams and other sources of natural waterDifferent shelters 18.9 [15–22]Newberry and Wood-Gush, 1985 [16]Large White1 adult boar, 5 adult sows and offspring, 10 juveniles from previous littersPig Park, Midlothian, ScotlandHilly areaEnclosure contains running streams, open grassy areas, gorse-covered areas and pine trees [8.6–14.3]Newberry and Wood-Gush, 1986 [19]Large White1 boar, 5 sows and offspringThe same as in the study of Newberry and Wood-Gush (1985) [16] Petersen, 1994[22]Swedish LandraceYear 1-Group 1: 1 boar, 5 sows, 1 gilt -Group 2:2 boars, 4 adult sows, 4 giltsYear 22 groups each consisting of 1 boar, 4 sows, 4 giltsThe same as in the study of Jensen (1986) [10] Petersen et al., 1989 [20]Swedish Landrace-Group 1 2 sows, 1 older and 1 younger boar, 4 gilts -Group 23 sows and offspring, 1 older and 1 younger boar, 4 giltsThe same as in the study of Jensen (1986) [10] [15–17]Petersen et al., 1990 [11]Swedish Landrace-Group 1: 1 boar, 5 sows, 1 gilt -Group 2:2 boars, 4 sows, 4 giltsThe same as in the study of Jensen (1986) [10] 10.2 [6–17] Stangel and Jensen, 1991 [18]Swedish Landrace-Group 1 1 boar, 4 sows and offspring, 4 gilts -Group 21 older and 1 younger boar, 4 sows, 4 giltsThe same as in the study of Jensen (1986) [10]50–100 3. Effects of Commercial Weaning Practices on Piglet Behavior and Stress Responses3.1. Factors Leading to Weaning Distress and Post-Weaning Piglet BehaviorIn contrast to the slow and gradual nature of the weaning process in (semi-)natural environments, under most commercial conditions worldwide, piglets are typically weaned abruptly between 2.5 and 5 weeks of age. This weaning practice is, therefore, undoubtedly a critical phase in the pigs’ lives that causes distress and behavioral disorders [28]. It is well accepted that the sudden change in nutrition and the pig’s social and physical environment is associated with physiological and psychological stress. It is clear that different stressors culminate upon weaning (reviewed by [29]), but to what extent single factors and their interactions contribute to the upregulation of the stress response and subsequent productivity is far less clear. Funderburke and Seerley (1990) [30] reported that plasma cortisol levels were higher in pigs that faced nutritional changes (from milk to dry feed) as compared to psychological (removal of sow) or cold stress (from 13 °C and below), which may suggest that the dietary switch had the greatest influence. In this study, though, piglets remained in the farrowing pen in all treatments, with only the sow being moved. When piglets are mixed and moved to the nursery, as occurs mostly commercially, the relative contributions of nutritional, psychological and environmental stressors to the post-weaning malaise are more difficult to establish. As has been described many times, a major consequence of weaning is a reduction in voluntary feed intake [5,31]. Common strategies to ease the dietary switch, therefore, are to provide solid creep feed or milk liquid feed during the suckling period and to offer the same feed(s) in the days following weaning. Creep feeding can be an effective strategy to familiarize piglets with solid food [7,32]. However, the intake of solid feed during lactation is generally small and highly variable within and between litters and may only be significant if weaned later than 3 weeks of age [33,34,35].More relevant to this review is the effect of maternal separation. Its contribution to post-weaning stress in piglets cannot be ignored. Maternal separation is a well-characterized model of early-life stress in rodents. For example, Gareau et al. (2006) [36] demonstrated immediate transient effects on colonic mucosal barrier function and long-lasting effects on bacterial–host interactions in separated neonatal rat pups. Repeated 2 h daily isolation from days 3 to 11 of age in piglets caused changes in behavioral, neuroendocrine and immune regulation and produced long-term effects, not only on the activity of the HPA axis but also on the immune–brain circuitry [37]. Early separation from the mother is often manifested by prolonged vocalizations. Colson et al. (2006) [38] showed an increase in vocalizations and lying in litter cohesion, which was later followed by increased aggressive and (belly) nosing behavior in piglets weaned either at 21 or 28 days of age, and yet more intense in the first group. These behaviors could also be related to anorexia following weaning. However, low grunts and calls during several days after the weaning event could be attributed to pigs searching for their mother.At weaning, piglets from different litters are usually brought together in larger groups, after which they are redistributed across different pens, often based on body weight and sex. Stress caused by grouping piglets has short-term and long-term consequences [39]. In the short term, the mixing of non-littermates induces an increase in body temperature and fighting behavior [39]. Piglets fight in order to impose dominance [40]. Ekkel et al. (1995) [41] and Pluske and Williams (1996) [42] demonstrated that piglets that are mixed at weaning show significantly more agnostic behaviors during the first hour after mixing than those that are not mixed at weaning. Moreover, it has been demonstrated that the mixing of litters at weaning not only intensifies interactions between unfamiliar piglets but also the recurrence of agnostic interactions between littermates [43]. Therefore, the sudden mixing of unfamiliar piglets results in high numbers of skin lesions and scratches. Bohnenkamp et al. (2013) [44] found that grouping and redistribution of non-littermates at weaning based on weaning weight may influence the agonistic behavior of piglets after weaning. They indicated that piglets with a weaning weight of 8 kg showed more agonistic behavior in the 40 h after weaning than piglets with a weaning weight of 6 or 10 kg. Mei et al. (2016) [45] found that, in a group where three pigs from one litter were mixed with three pigs from another litter, more fights occurred during the 5 h after mixing compared with the group where four pigs from one litter were mixed with two pigs from another litter, or where no piglets were mixed (respectively, 25.3 ± 4.8; 15.3 ± 2.2; 2.0 ± 0.5 fights). In addition, the time spent fighting was higher in the first group. Similarly, the number of aggressive interactions was higher in groups with four piglets from three different litters compared to groups with six piglets from two litters [46]. Blackshaw et al. (1987) [47] reported that mixing piglets from multiple litters caused severe agonistic behaviors during the 1.5 h after mixing. Moreover, they observed that mixing piglets from three litters at weaning caused less agonistic behavior during the 1.5 h after mixing, which is an important period for the establishment of a hierarchy, than mixing piglets from two or four litters. Interestingly, a recent study suggests that the frequency of agnostic interactions showed high fluctuation across groups of mixed weaners, yet for most pens, stable social networks were established by the third day after weaning [48]. Aggressive behavior caused by grouping unacquainted litters is associated with activation of the HPA axis and an increase in plasma cortisol concentrations [49,50,51]. The studies by Colson et al. (2012) [49] and Merlot et al. (2004) [51] demonstrated that social stress at weaning also causes a higher occurrence of lying or resting behaviors. Colson et al. (2012) [49] found that combining social and housing changes at weaning resulted in an increased behavior of the pigs lying inactive in an awake state during the 7 h after weaning. Moreover, the study by Merlot et al. (2004) [51] reported that the effect on cortisol and behavior is influenced by the social rank of the pig within the group. In the study, a dominance index was assigned to each piglet in the pen by observing the agonistic behavior of the piglets after grouping. The behavioral responses of dominant piglets differed from the behavioral responses of subordinate piglets. The researchers suggested that mixing piglets causes stress but that piglets can adjust to stressful situations by developing behavioral strategies. In the long term, the social stress caused by mixing unfamiliar piglets at weaning causes immunological and physiological disorders [52] and reduced performance of the animal [41]. For example, in the study by Ekkel et al. (1995) [41], the average daily gain of piglets that were mixed at weaning was lower compared to piglets that were not mixed or were moved to another pen at weaning (479 ± 12 vs. 528 ± 14 g/d). The effect of weaning age on post-weaning behavior has been well acknowledged [53]. The detrimental effects are more substantial when weaned earlier [53,54]. For example, a higher level of aberrant behaviors, such as belly nosing, vocalizations and escape attempts, was noticed in 7-day-old weaned pigs as compared to weaning at 14 or 28 days of age [53].3.2. Weaning-Induced Stress ResponsesSeveral neuroendocrine responses occur during this weaning-induced stress response via the sympathetic nervous system activation of the adrenal medullae and the HPA axis. The locus coeruleus/norepinephrine system in the brain releases the neurotransmitter norepinephrine, which affects many organs and causes the release of catecholamine hormones (epinephrine/norepinephrine and dopamine) from the adrenal medullae. Collectively, they initiate several physiological events to create the ‘fight-or-flight response’. Yet, the effects of weaning on catecholamines are equivocal, and their role in the regulation of the metabolism of the weaned pig is still unclear [55]. The HPA axis releases the corticotrophin-releasing hormone (CRH) during stressful events, which stimulates the release of the adrenocorticotropic hormone (ACTH) and glucocorticoids such as cortisol. Weaning results in a transient increase in CRH and cortisol levels, irrespective of weaning age, but they tend to be higher when weaned at a younger age [1,56,57]. Though activation of the HPA axis is a critical survival mechanism to respond to a stressor and return to homeostasis, e.g., by stimulation of gluconeogenesis and lipid metabolism, it is also to some extent responsible for compromised GIT structure and function after weaning.At the intestinal level, the expression of CRH receptors is upregulated by early weaning [57] and chronic mixing and crowding [58]. The secretion of CRH from the hypothalamus and activation of intestinal CRH receptors cause hyperplasia and degranulation of mast cells in the intestinal mucosa [57,59,60]. Mediators of these mast cells, such as proteases, biogenic amines and cytokines, amongst others, are released, which negatively affect epithelial barrier function by influencing paracellular permeability [61,62]. For example, tryptase, the most abundant secretory granule-derived serine protease contained in mast cells, cleaves and activates a protease-activated receptor (PAR2) expressed on the apical and basolateral membranes of epithelial cells. This leads to Ca2+ mobilization, β-arrestin and ERK1/2 MAPK activity, perijunctional F-actin redistribution, zonula-1 delocalization and protein junctional adhesion molecule-A downregulation. Weaned piglets, therefore, may show a transient increased intestinal permeability to macromolecular markers lasting only a few days after weaning, which gradually declines over the subsequent 2 weeks [57,63]. Nevertheless, elevated CRH may also cause delayed gastric emptying, slow intestinal transit, increased colonic transit, and increased baseline secretory state and responsiveness to secretagogues of the intestine [62]. Regarding the latter, Moeser et al. (2007) [57] demonstrated that the neural inhibitor toxin, tetrodoxin, prevented weaning-induced elevations in pigs’ transepithelial short circuit current, a marker of intestinal ion secretion. Further, stress neuroendocrine mediators such as catecholamines and ACTH have also been shown to affect the adherence of Escherichia coli O157:H7, a serotype of Shiga toxigenic E. coli, to porcine cecal and colonic mucosa [64,65]. McCracken et al. (1999) [66] hypothesized that the increase in plasma cortisol may also inhibit the activity of AP-1 transcription factors, responsible for the induction of genes of the major histocompatibility complex class I (MHC-I) [67]. A compromised immune response, because of the reduction in MHC-I expression, may render pigs more susceptible to enteric infections, particularly viral infections, often resulting in diarrhea and reduced performance [66]. Moreover, in piglets aged 2–4 weeks, CD8+ T cells are largely absent in the lamina propria of the gut tissue [68]. Barrier dysfunction, secretory hypersensitivity and suppressed immunocompetence as a result of weaning stress may aggravate the post-weaning growth check and elevate susceptibility to enteric infections.4. Socializing Piglets during Lactation in Commercial SettingsWhile pork industries typically cannot delay weaning for an excessively long period, techniques that can be adapted to a commercial setting to promote a more gradual weaning process have been explored and will be discussed below. Ultimately, pre-weaning socialization may reduce the risk of mixing stress and promote social eating behaviors (Section 4.1), while intermittent suckling (IS), a form of gradual weaning, can increase exploratory behaviors and allow the piglets to become accustomed to maternal separation (Section 4.2). Finally, a period of separation from the sow would also present an ideal window for piglets from different litters to co-mingle, allowing producers to combine the two techniques (Section 4.3).4.1. Co-Mingling of Piglets before Weaning4.1.1. Co-Mingling as Management StrategyThe co-mingling (mixing) of non-littermate piglets before weaning is the subject of many studies and is of commercial interest due to its propensity to lower post-weaning stress responses. Kanaan et al. (2012) [69] described the co-mingling of non-littermates before weaning as a management strategy or process in which two or more unfamiliar litters can interact during the lactation period, allowing pre-weaning socialization and enabling piglets to cope better with future social and non-social challenges. Since it is common to socialize piglets that are similar in age, co-mingling is most applicable in batch-farrowing systems.4.1.2. Different Settings for Co-Mingling of Piglets before WeaningMany variations of co-mingling exist, with most systems differing by individual versus group housing of the sow, the timing of when the piglets are co-mingled during lactation, the size of the co-mingled group, and the weaning age of the co-mingled piglets. Table 2 summarizes the different studies that exist according to these criteria and were all conducted in the last three decades.Concerning the housing of the sow, the first approach of co-mingling non-littermates before weaning is applied in a conventional farrowing system (Table 2). Shutters in solid partitions between farrowing pens can be opened, solid barriers between farrowing pens can be removed, or solid partitions can be replaced by partitions with an opening a few days after farrowing, all allowing piglets from different litters to interact with each other. The sow remains confined in her farrowing crate. Either way, this intervention is relatively easy to implement in traditional farrowing facilities. Interestingly, Weary et al. (1999) [70] created a central communal area for the piglets by leaving two of the five farrowing pens vacant, whereas Turpin et al. (2017) [71,72] combined IS with co-mingling non-littermates by housing two different litters in an empty farrowing pen in an adjacent room. Another variant was described by Parratt et al. (2006) [73], where piglets from multiple litters had access to a common corridor after the removal of the separations along the corridor. Secondly, co-mingling of non-littermate piglets can also be implemented where sows are loose housed. Due to the absence of a farrowing crate or the removal of the farrowing crate shortly after partus, the sow can freely move in the farrowing pen. In a similar way as in conventional farrowing pens, the piglets can be co-mingled, yet the sow remains within her own pen. In the study of Kutzer et al. (2009) [74], for example, a loose-housing system with four farrowing pens was used. Thirdly, a multi-litter system, also referred to as a multi-suckling, group-housing or group lactation system, is a farrowing system in which a group of lactating sows is housed in one large pen together with their respective litters. More space, free movement for dams and progeny and social interactions between animals are the key elements of the system. Usually, a multi-litter system has a communal area, either divided or not into feeding, dunging and resting areas, and a number of farrowing pens. By removing partitions behind farrowing pens, removing the barriers and piglet hatches from the farrowing pens, removing the entire farrowing pen, or alternatively, rearranging the multi-litter area after farrowing, pigs obtain access to the entire area. A separate straw-bedded area for the piglets without access for the sows can be included [75,76]. In two studies, piglets were restricted from entering the communal area of the sows, but the piglets could move to other farrowing pens via a separate area that the sows could not enter [77,78]. Finally, a few days or weeks post-partum, sows and their litters could also be transferred from a conventional farrowing or loose-housing system to the multi-litter system.Next to housing, the time of starting co-mingling can vary (Table 2). This moment is often expressed as a function of the age of the piglets at the start of co-mingling, even though the farrowing dates of the sows within the group may range over a few days. This makes the interpretation of the exact age of piglets at co-mingling less easy to define. Alternatively, reference is taken to the average farrowing date of the sows in the group, the average expected farrowing date of the sows, the farrowing date of the last sow in the experiment, or the day of weaning. In some cases, piglets are in the open-access area readily from birth [79]. Grouping of pigs from multiple litters can be performed early, i.e., 3 [80] or 5 days after farrowing [81,82]. In several studies, the co-mingling of non-littermates takes place around 10 days after partus, whereas fewer studies applied a co-mingling system where non-littermates are mixed from around day 14 after partus or beyond. Surprisingly, in the literature, minimal arguments are provided for the choice of the co-mingling date. Interestingly, in the study of Schrey et al. (2019) [83], non-littermates were allowed to freely choose the moment of socializing, which occurred on average at 10.6 days post-partum. Nonetheless, in most studies, piglets are grouped somewhere between 7 and 14 days after partus, which corresponds to the period of nest-leaving of the sow and her offspring and joining the herd, commensurate with increasing interactions with non-littermates in (semi-)natural pigs. Nevertheless, Parratt et al. (2006) [73] grouped unfamiliar piglets only 5 days before weaning, or approximately 16 days after birth. They hypothesized that the implementation of co-mingling of suckling piglets closer to weaning would have more benefits on pre-weaning performance and behavior since fewer suckling disruptions would occur. Unfortunately, few studies were designed to study the effect of time of co-mingling on aggression, stress, the occurrence of estrus during lactation, weaning-to-estrous intervals, sow performance, nursing–suckling interaction or piglet performance [84,85,86,87,88,89,90,91,92]. Yet, in practice, other considerations, such as vaccination schemes, disease prevalence, data collection from breeding sows and the provision of creep feed, may encourage later starting dates.Another point of difference between studies is the size of the group, i.e., the number of litters (Table 2). In conventional farrowing systems or loose-housing systems, the number of grouped litters is usually small, i.e., non-littermates from two to four litters are grouped, whereas the multi-litter system mainly consists of four or more litters. The largest groups (8–22 sows and litters) were described by Hultén et al. [93,94,95,96].animals-13-01644-t002_Table 2Table 2Overview of different variants of pre-weaning co-mingling of non-littermates 1.ReferenceHousing of Sowin the Farrowing UnitTime StartCo-Mingling ofSuckling PigletsGroup Size and CompositionWeaning Age(Days)Individual/GroupConventional Crates/Loose HousedSalazar et al., 2018 [84]IndividualConventional crates7 days post-partum2 litters25D’Eath, 2005 [97]IndividualConventional crates10 days post-partum2 litters30 ± 3.0Kanaan et al., 2012 [69]IndividualConventional crates10 days post-partum2 litters18Hong et al., 2017 [98]IndividualConventional crates10 days post-partum3 litters/Pluske and Williams, 1996 [42]IndividualConventional crates10 days post-partum3 litters28.5Kutzer et al., 2009 [74]IndividualConventional crates10 days post-partum4 litters28Morgan et al., 2014 [99]IndividualConventional crates10 days post-partum4 litters21–24Klein et al., 2016 [100]IndividualConventional cratesYoungest litter 10 days old4 litters28Wattanakul et al., 1997 [101]IndividualConventional crates11 days post-partum3 litters28Weary et al., 1999 [70]IndividualConventional crates11 days post-partum3 litters28Hessel et al., 2006 [102]IndividualConventional crates12 days post-partum3 litters28Kanaan et al., 2008 [103]IndividualConventional crates13 days post-partum2 litters/Camerlink et al., 2018 [104]IndividualConventional crates14 days post-partum2 litters26Camerlink et al., 2019 [105]IndividualConventional crates14 days post-partum2 litters26Ko et al., 2020 [106]IndividualConventional crates14 days post-partum2 litters25Salazar et al., 2018 [84]IndividualConventional crates14 days post-partum2 litters25Weller et al., 2019 [107]IndividualConventional crates14 days post-partum2 litters28Parratt et al., 2006 [73]IndividualConventional crates16 days post-partum(5 days before weaning)2–3 litters21Van Kerschaver et al., 2021 [89]IndividualConventional crates6, 11 and 16 days before weaning3 litters21Turpin et al., 2017 [72]IndividualConventional crates7 days before weaning2 litters22 ± 1.7Turpin et al., 2017 [71]IndividualConventional crates7 days before weaning2 litters25.3 ± 0.7Illmann et al., 2007 [108]IndividualLoose housed10 days post-partum2 litters/Kutzer et al., 2009 [74]IndividualLoose housed10 days post-partum4 litters28Arey and Sancha, 1996 [79]GroupLoose housedAt birth4 sows and litters/Naya et al., 2019 [80]GroupLoose housed3 days post-partum6 sows and litters35Bohnenkamp et al., 2013 [81]GroupLoose housed5 days post-partum6 sows and litters26Bohnenkamp et al., 2013 [109]GroupLoose housed5 days post-partum6 sows and litters26Bohnenkamp et al., 2013 [44]GroupLoose housed5 days post-partum6 sows and litters26Grimberg-Henrici et al., 2016 [110]GroupLoose housed5 days post-partum6 sows and litters27.8 ± 0.15Lühken et al., 2019 [111]GroupLoose housedYoungest litter 5 days old6 sows and litters26Nicolaisen et al., 2019 [112]GroupLoose housedYoungest litter 5 days old6 sows and litters28Nicolaisen et al., 2019 [113]GroupLoose housedYoungest litter 5 days old6 sows and litters28Schnier et al., 2019 [82]GroupLoose housedYoungest litter 5 days old6 sows and litters26Gentz et al., 2019 [114]GroupLoose housed5 days post-partum10 sows and litters27Gentz et al., 2020 [115]GroupLoose housed5 days post-partum10 sows and litters27Lange et al., 2020 [116]GroupLoose housed5 days post-partum10 sows and litters26.45 ± 0.97van Nieuwamerongen et al., 2015 [117]GroupLoose housedYoungest litter 6 days old5 sows and litters27.1 ± 0.3Grimberg-Henrici et al., 2018 [118]GroupLoose housed6 days post-partum10 sows and litters26 ± 1Grimberg-Henrici et al., 2019 [119]GroupLoose housed6 days post-partum10 sows and litters26 ± 1Verdon et al., 2019 [90]GroupLoose housed6.9 ± 1.2 days post-partum5–7 sows and litters26.7Arellano et al., 1992 [120]GroupLoose housed7 days post-partum5–10 sows and litters28Pedersen et al., 1998 [77]GroupLoose housedYoungest litter 7 days old4 litters28Goetz and Troxler, 1995 [121]GroupLoose housedYoungest litter 7 days old4 sows and litters28Rantzer et al., 1997 [122]GroupLoose housed7.2 (2–15) days post-partum4 sows and litters35Verdon et al., 2020 [92]GroupLoose housed7.3 ± 1.2 days post-partum5–7 sows and litters25.5 ± 2.1 Verdon et al., 2019 [91]GroupLoose housed7.4 ± 1.1 days post-partum5–7 sows and litters25.5van Nieuwamerongen et al., 2017 [123]GroupLoose housed7.9 ± 0.3 days post-partum5 sows and litters27.1 ± 0.4van Nieuwamerongen et al., 2018 [124]GroupLoose housed7.9 ± 0.3 days post-partum5 sows and litters27.1 ± 0.4van Nieuwamerongen et al., 2017 [125]GroupLoose housed8.1 ± 0.3 days post-partum5 sows and litters28 (A4 treatment) and 63 (IS9 treatment)Maletínská and Špinka, 2001 [126]GroupLoose housed10 days post-partum3–4 sows and litters/Kutzer et al., 2009 [74]GroupLoose housed10 days post-partum4 sows and litters28Hillmann et al., 2003 [75]GroupLoose housed10 days post-partum5 sows and litters28Marchant et al., 2000 [127]GroupLoose housed10 days post-partum5 sows and litters25.2Schrey et al., 2019 [83]GroupLoose housed10.6 days post-partum5 sows and litters35Verdon et al., 2020 [92]GroupLoose housed10.1 ± 1.2 days post-partum5–7 sows and litters25.5 ± 2.1 Li et al., 2010 [128]GroupLoose housed10 days post-partum8 sows and litters28–35Li and Wang, 2011 [129]GroupLoose housed10 days post-partum8 sows and litters35Hultén et al., 1995 [94]GroupLoose housedYoungest litter 10 days old8–22 sows and litters38.8Hultén et al., 1998 [96]GroupLoose housedYoungest litter 10 days old8–22 sows and litters36.8 ± 6.9Li and Johnston, 2009 [130]GroupLoose housed10 ± 3 days post-partum8 sows and litters28 ± 3Dybjær et al., 2001 [131]GroupLoose housed11 days post-partum6 sows and litters32Olsen et al., 1998 [132]GroupLoose housed11 days post-partum6 sows and litters32Li et al., 2012 [133]GroupLoose housed12 ± 1.3 days post-partum8 sows and litters33 ± 1.3Verdon et al., 2020 [92]GroupLoose housed13.5 ± 1.4–13.9 ± 0.9 days post-partum5–7 sows and litters25.5 ± 2.1 Verdon et al., 2019 [91]GroupLoose housed13.6 ± 1.3 days post-partum5–7 sows and litters25.5Verdon et al., 2019 [90]GroupLoose housed13.9 ± 1.1 days post-partum5–7 sows and litters26.7Weary et al., 2002 [78]GroupLoose housed14 days post-partum3 litters28Wattanakul et al., 1998 [134]GroupLoose housed14 days post-partum4 sows and litters28Verdon et al., 2016 [135]GroupLoose housed14 days post-partum6 sows and litters27.3Tang et al., 2023 [85]GroupLoose housed8 and 13 days post-partum5 sows and litters49Šilerová et al., 2006 [136]GroupLoose housed11–20 days post-partum(in one case, 7–10 days post-partum)6–11 sows and litters33–42Šilerová et al., 2010 [137]GroupLoose housed11–20 days post-partum6–11 sows and litters33–42Greenwood et al., 2019 [138]GroupLoose housed21 days post-partum6 sows and litters28Thomsson et al., 2016 [87]GroupLoose housed1–3 weeks post-partum2–6 sows and litters44 ± 1.6Thomsson et al., 2015 [86]GroupLoose housed1–3 weeks post-partum5–8 sows and litters42Thomsson et al., 2018 [88]GroupLoose housed1–3 weeks post-partum5–8 sows and litters43.6 ± 0.6–44.2 ± 0.6Hultén et al., 1995 [93]GroupLoose housedYoungest litter 2 weeks old11–22 sows and litters37.7Wattanakul et al., 1997 [76]GroupLoose housed2 weeks post-partum5–6 sows and litters31 (trial 1) and 29 (trial 2)Hultén et al., 1997 [95]GroupLoose housed2–3 weeks post-partum12–22 sows and litters39.21 References are ordered according to housing of the sows in the farrowing unit, time of start of co-mingling of suckling piglets, group size and composition.4.1.3. Occurrence of Cross-Suckling or Allo-Suckling in Piglets by Co-Mingling before WeaningCo-mingling of non-littermates before weaning gives piglets the opportunity to suckle sows other than their mother [126]. Lower milk production is the main reason that piglets become ‘cross-sucklers’ [132], and cross-sucklers are able to search for sows with higher milk production (as assessed by piglet weight gain) and to have a preference for more productive teats [132]. This may be beneficial for their growth [70], but it has also been reported to have no effect on performance when compared to non-cross-sucklers [108,132]. More specifically, piglets that cross-suckled in more than 50% of their nursings (i.e., permanent allo-sucklers) were found to have an average weight gain of approximately 2300 g, whereas their biological littermates that were not classified as cross-sucklers had an average weight gain of approximately 2900 g between day 10 post-partum (the day of co-mingling) and day 24 post-partum. Dybjaer et al. (2001) [131] observed that the daily growth of piglets was negatively correlated with the percentage of cross-sucklers. Indeed, overt cross-suckling may cause lost nursings [77] and more disturbances of the orderly suckling and continuation of teat order [70]. In some studies, the incidence of cross-suckling appeared to be very low [81,101], while in the experiment of Maletínská and Špinka (2001) [126], more than one-third of the piglets would cross-suckle at least once. In some studies, the incidence of cross-suckling was even higher, e.g., more than 50% [76,134].This high variability in cross-sucking behavior is surprising. The occurrence of cross-suckling is affected by several factors [126,131] and may depend on the manner of implementation of the system of co-mingling non-littermates [78,113,134]. For example, the incidence of cross-suckling may decrease when sows and their respective litters have the opportunity to establish clear dam–progeny bonds prior to implementing co-mingling [78,113,134]. Due to intense sow–piglet interactions immediately after initiating the co-mingling of non-littermates, sows seem to be highly discriminant against other, alien piglets (potential cross-sucklers), which leads to a reduced number of successful sucklings. However, 3 h after grouping, cross-sucklers seem to be accepted by the sow [101], which is reasonable since it is hypothesized that piglets pick up the same odor after grouping. Horrell and Hodgson (1992) [139] observed that sows cannot distinguish their own piglets from other piglets until the age of 2 weeks, when the odors are masked. In contrast, piglets seem to be less able to distinguish their littermates from non-littermates [140], suggesting they may unintentionally join non-littermates at the udder of their sow during milk letdown.4.1.4. Effects of Co-Mingling before Weaning on Social Behavior before and after WeaningThe behavioral effects of pre-weaning socialization have been the focus of many recent studies, with observations of the behaviors of socialized piglets both before and after weaning. When piglets are first socialized with non-littermates during lactation, aggressive behaviors, reflected in the occurrence of skin lesions or damage, are often observed immediately after litters are mixed [89]. This type of fighting usually occurs between unfamiliar pigs and is most likely related to the establishment of a social hierarchy. This type of competitive behavior between non-littermates is also evident at the udder when cross-suckling occurs. Interestingly, Salazar et al. (2018) [84] observed an increase in aggression 1 day after mixing in piglets co-mingled at 7 days of age, which was not the case when piglets were co-mingled at 14 days of age. In contrast, Verdon et al. (2020) [92] found no effect of age on co-mingling and suggested that this discrepancy could be related to differences in the studied systems. In a study by Kanaan et al. (2008) [103], co-mingled piglets had more ear injuries 2 days after socialization compared with conventionally housed piglets, but the differences disappeared by day 5 after socialization. Kutzer et al. (2009) [74] found no effect of the housing system on skin lesion scores 4 days after allowing unfamiliar litters to co-mingle during lactation. Only on the last day in the farrowing pens did the grouped piglets have more skin lesions. Furthermore, in some studies, no differences in lesion scores between socialized and non-socialized piglets were observed (anymore) close to weaning [92,106,135] or at any timepoint [69]. The size and layout of the pens might explain why the co-mingled piglets had more aggression than others across the different studies [74]. In this respect, the size of the co-mingled group also seems to be a factor for aggression. Grimberg-Henrici et al. (2018) [118] reported more skin lesions on the carpus and body of co-mingled piglets of group-housed sows at the end of the lactation period compared with conventionally reared piglets. They indicated that an increase in social interactions due to a large (n = 10 litters) group size of co-mingled litters might be the reason since, in the study by van Nieuwamerongen et al. (2015) [117], where only five litters were co-mingled before weaning, no differences in skin lesions on the body were found at the end of the lactation period. However, in contrast to the study of Grimberg-Henrici et al. (2018) [118], the multi-suckling piglets in the study of van Nieuwamerongen et al. (2015) [117] had more skin lesions at the snout compared to the control piglets at that time, which, according to the authors, was due to the increased competitive behaviors of the piglets at the udder and, in consequence, the incidence of cross-suckling.It is after weaning that the behavioral benefits of co-mingling are observed. Hillmann et al. (2003) [75] demonstrated that piglets that socialized with non-littermates before weaning adapted better to social and non-social challenges during the weaning event. In addition, numerous studies have demonstrated a reduced level of aggression among socialized piglets immediately after weaning, independent of the housing system, group size and age at the start of co-mingling. For example, Van Kerschaver et al. (2021) [89] showed that co-mingled piglets had fewer skin lesions at the shoulders and the flanks compared to controls in the days following weaning. Socialized piglets are already familiar with one another as, prior to weaning, the social skills of the piglets were stimulated and the piglets had already established a new social hierarchy. Some studies demonstrate that co-mingling prior to weaning reduces pig aggression at weaning even when familiarity at weaning is not a factor, i.e., when piglets that are mixed at weaning originate from different co-mingled groups. In addition, pre-weaning familiarity may also have benefits in the longer term for social development, as evidenced by, for instance, less fighting [129], a shorter duration of fights [97,129], fewer skin lesions [104,105] and fewer shorter tails caused by tail-biting in the fattening period [100], even when socialized piglets are regrouped with unfamiliar piglets at a later stage during rearing. However, other studies report that aggression after weaning did not differ between socialized and control pigs [99,137]. Additionally, Naya et al. (2019) [80] and Gentz et al. (2020) [115] found no clear effect on tail biting in the early social environment occurring before weaning. Differences in the group size, a strong batch effect, inter-farm differences, weaning age (and moment of observations), method of weaning (without mixing litters), stress and space allowance and the method of assessing social behaviors may explain different outcomes across studies. Indeed, it is important to understand that, in the literature, various methods and techniques, such as (skin) lesion scoring, social (challenge) tests such as resident–intruder tests, video recordings and behavioral observations, are used to assess the social skills of co-mingled piglets before and after weaning.4.1.5. Effects of Co-Mingling before Weaning on Performance before and after WeaningThe main objective of the co-mingling of non-littermates prior to weaning is to ameliorate the weaning transition in order to reduce GIT dysbiosis and maintain optimal production. The co-mingling of non-littermates does not commonly affect the performance of piglets before weaning [97,104,138], although some studies have reported lower weaning weights and pre-weaning weight gains in socialized piglets compared with non-socialized piglets [44,76,92,127]. According to the authors, this could be due to a myriad of factors, including: the timing of co-mingling; higher space allowances and, therefore, more exercise; lower creep feed intake; disruption of suckling; cross-suckling; enhanced active behaviors; greater occurrence of fighting behaviors (skin lesions); relocation of the sows during co-mingling within the co-mingling group and, consequently, a high incidence of cross-suckling; and/or stress on the piglets due to the introduction of the piglets into new accommodation and environments after moving the sows and their litters from conventional housing to group lactation. In contrast, Arey and Sancha (1996) [79] observed a higher piglet daily weight gain in week 2 post-partum in piglets born in a multi-suckling system and suggested that this was due to the more successful nursing behavior in the multi-suckling system.Several studies have demonstrated a positive influence on the post-weaning performance of piglets, particularly in the immediate post-weaning period, suggesting that co-mingling might reduce weaning stress and contribute to a better adaptation of the piglets to weaning. For example, in the study of Hessel et al. (2006) [102], piglets co-mingled in a conventional housing system tended to gain 290 g more after weaning than conventionally reared piglets during the first week after weaning, and gained, in total, 1.09 kg more over the entire study (9 weeks). Feed efficiency in the first week after weaning tended to be improved in piglets that were co-mingled 6 days before weaning [89]. In research by van Nieuwamerongen et al. (2015) [117], piglets reared in a multi-litter system gained 69% more weight between day 1 before weaning and day 2 after weaning compared with control piglets, and the weight gain of these piglets was 24% higher in the total nursery period, i.e., between days 1 and 35 after weaning. The piglets from the multi-litter system also consumed 81% more feed between days 1 and 2 after weaning compared with the conventionally reared piglets. In the study of Weary et al. (2002) [78], co-mingled piglets consumed more creep feed before weaning and one day after weaning, resulting in higher weight gain in the immediate post-weaning period. Co-mingling non-littermates before weaning may thus also positively affect feed intake in pigs. However, these substantial benefits have been contradicted by others. Other studies showed that co-mingling did not affect post-weaning performance [42,109]. Although the body weight was higher at weaning in the co-mingled piglets, the average daily gain during the 14 days after weaning did not differ between the piglets in the co-mingled group and in the conventionally reared group in the study by Pluske and Williams (1996) [42] (367 vs. 378 g/day). Bohnenkamp et al. (2013) [109] observed no differences in body weight at weaning and after rearing (7 weeks after weaning) between co-mingled piglets and control piglets and suggested, by comparing other studies, that the timing of grouping non-littermates during lactation might have had an impact. Rantzer et al. (1997) [122] observed an even lower growth rate during the 4 weeks after weaning of piglets from a multi-suckling system compared with conventionally reared piglets, which the authors ascribed to the larger weaner pen and the lower feed intake following weaning of the piglets from the multi-suckling system.4.1.6. Effects of Co-Mingling before Weaning on Mortality before and after WeaningMortality and, more specifically, pre-weaning mortality of piglets may depend on the housing system used to achieve co-mingling. Pre-weaning piglet mortality is typically higher in multi-litter systems (group lactation) and is often caused by a greater incidence of crushing. For example, in the study of Marchant et al. (2000) [127], 14–17% of the piglets born alive died in the multi-suckling systems as a result of crushing, compared to 8% in the conventional farrowing system. In addition, van Nieuwamerongen et al. (2015) [117] found that more piglets died before weaning in the multi-litter system compared with the conventional farrowing system (respectively, 3.22 ± 0.42 vs. 1.52 ± 0.25 piglets per litter), particularly because of crushing (respectively, 2.34 ± 0.44 vs. 0.20 ± 0.09), which is likely associated with the loose farrowing pens in the multi-litter system since mortality mostly occurred before co-mingling. Interestingly, in some studies, no differences in piglet losses during lactation between a multi-suckling system and a single housing system were observed, which was likely due to housing sows and their litters individually for a short period following birth before transferring to a multi-suckling system [76,87,135], the presence of farrowing crates in the multi-suckling system [44] or the enhanced maternal behavior of the sows in an enriched environment [79]. In the study of Grimberg-Henrici et al. (2016) [110], piglet losses before weaning were even lower in the group-housing system compared with conventional farrowing crates. The researchers suggested that the experience of the stockpersons and the free movement of the sows before farrowing explained this. The mortality rate of piglets in conventional housing systems where co-mingling of non-littermates is applied seems to be lower and does not differ from conventional housing [97,102,104]. After weaning, no significant differences in piglet mortality between co-mingling and conventional housing systems have been found.4.1.7. Effects of Co-Mingling before Weaning on GIT Structure and Function and Immunocompetence of Weaned PigletsCo-mingling of non-littermates before weaning may also positively influence the GIT structure and function and immunocompetence of weaned piglets. In a study by van Nieuwamerongen et al. (2015) [117], the average fecal consistency score was significantly lower (i.e., more solid feces and less diarrhea) during the 2 weeks after weaning for the piglets reared in the multi-litter system compared with the conventionally reared piglets. In addition, the study showed that titers of IgM-binding keyhole limpet hemocyanin, as a measure for natural antibodies of the adaptive immune system, were higher for group-housed suckling piglets compared with conventionally housed piglets on the day of weaning. They speculated that this was due to an accelerated development of the pigs’ immune system, triggered by early contact with materials such as feed and/or enrichment in the pre-weaning environment. However, titers of IgG-binding keyhole limpet hemocyanin and haptoglobin and leukocyte concentrations were not affected by the housing system. In the experiment by Rantzer et al. (1997) [122], in which a multi-suckling system was compared with a conventional housing system, the average diarrhea score did not differ after weaning, despite the fact that the peak excretion of hemolytic E. coli was delayed in piglets from the multi-suckling system. The fecal consistency scores after weaning also did not differ between pigs reared in a multi-litter system and a conventional system in the study of van Nieuwamerongen et al. (2018) [124]. However, in the study, lower plasma sugar concentrations such as mannitol and galactose were observed on day 5 after weaning in pigs raised in the multi-litter system. The authors speculated that the lower concentrations of mannitol would probably indicate a lower intestinal permeability and, therefore, a less compromised intestinal barrier function as a result of weaning. Further, it was speculated by the authors that the lower concentrations of galactose were due to a better adaptation to weaning and the pre-weaning experience with solid feed of the pigs in the multi-litter system since no differences in post-weaning feed intake were observed.4.2. Gradual Separation from the Sow4.2.1. Sow-Controlled Housing to Promote Gradual WeaningTo overcome the negative consequences of abrupt weaning, housing systems that allow the sow to leave her piglets voluntarily have been investigated. In sow-controlled housing, sows can express their natural tendency to spend time apart from their offspring as lactation advances by leaving the conventional farrowing pens to mingle with other sows and/or eat in a common area, whereas the piglets cannot leave the farrowing pen. In these types of systems, sows will generally mimic the pattern of spending less time with their litters during the latter part of lactation [26,141,142]. However, there is still a considerable amount of variation in the response patterns, with some sows either spending all their time away from the litter or those that rarely leave the nest [143]. Further to this, no consistency from one lactation to another has been seen [143]. Compared with animals in conventional systems, piglets from sow-controlled housing generally consume more solid food during lactation [141,144], which can translate into an increase in solid feed intake and growth rate after weaning [144,145]. However, there is a risk that the large variation in the use of piglet-free areas may lead to inadequate maternal care, and it has therefore been suggested that housing systems that allow for more consistent use of piglet-free areas may be more beneficial [144].4.2.2. Intermittent Suckling to Promote Gradual WeaningSimilar in concept to sow-controlled housing, IS is a more structured, gradual weaning process that involves the daily separation of sows and piglets for a specified period during the last part of lactation. There are two main reasons why IS has been examined over the years. The first reason is to induce estrus in lactation, allowing for mating in lactation, which presents an opportunity to uncouple estrus cycling from weaning, therefore extending the weaning age (e.g., to 5–8 weeks) without decreasing the number of litters produced per sow per year ([146,147]; for reviews). Second, IS has also been studied as a strategy to promote supplementary (e.g., creep) feed intake in piglets during lactation. In this scenario, conventional weaning ages (i.e., 3 to 4 weeks) have been used in combination with intermittent separation of the litter from the sow to promote exploratory behaviors and encourage piglets to explore nutritional options other than milk. Table 3 describes the different variations in IS regimens that have been examined since 1961.4.2.3. Intermittent Suckling and Extended Lactation: Effects on Litter PerformanceWeaning under current commercial conditions can have a negative impact on production, health and welfare in the immediate post-weaning period. An older weaning age (i.e., greater than 5 weeks) has been shown to improve post-weaning feed intake along with a reduction in stress [148] and a reduction in post-weaning diarrhea [149]. However, given that sows are naturally in anestrus during lactation, producers rely on shortening the length of lactation to maximize the number of litters born per sow per year, thereby creating a conflict between sow and piglet welfare and the profitability of sow performance [150]. Previous studies have reported that 10–16% of commercial sows housed individually will ovulate spontaneously during lactation [151]. This can be increased if interventions, such as group lactation housing, boar exposure and/or exogenous gonadotropins, are introduced [151,152]. Reducing the suckling stimulus in a controlled manner (i.e., not relying on sows to regulate nursing frequency) by subjecting litters to interrupted suckling techniques such as IS has been used on its own or in combination with the techniques mentioned above to increase gonadotrophin release and can result in a 100% occurrence of estrus in lactation. However, the results are still highly variable, most likely due to factors such as sow parity, metabolism, litter suckling intensity, seasonal effects, and litter creep feed intake [147,150].While the effects of IS on sow reproductive performance are well documented, less is known about the impacts that IS in combination with extended lactation can have on piglet performance. A reduction in weight at the start of the intervention followed by compensatory growth a week later, most likely due to an increase in creep feed intake, is consistent across most of the early studies [153,154,155]. The effects of IS in combination with extended lactation were compared with control groups weaned at a conventional weaning age in a series of studies. In summary, combining IS with an extended lactation period caused a more gradual adaptation to weaning with respect to piglet growth and feed intake, which significantly ameliorated the post-weaning growth check [156,157,158]. However, a mild growth check was evident at the start of IS when IS started at 19 days of age or less [156,157]. This growth check was not evident when IS started at 26 days of age or older [157,158]. Collectively, these data suggest that, when IS starts at an older age, creep feed intake is likely to be sufficient to compensate for the reduction in milk consumption during the time of separation from the sow.Studies examining the effect of extended lactation on GIT morphology before weaning are limited, with only Turpin et al. (2016) [158] reporting that IS, with or without an extended lactation, did not affect the absorption of mannitol, a measure of absorptive surface area, compared with conventionally weaned controls 3 days before weaning (i.e., 4 days after the start of IS). After weaning, IS in combination with an extended lactation (33 days) prevented villus atrophy compared with conventionally weaned piglets [157]. Results from Nabuurs et al. (1996) [159] supported this finding and suggest that creep feed intake before weaning is an important factor in this outcome. Four days after weaning, Turpin et al. (2016) [158] reported an improvement in galactose absorption but no improvement in mannitol absorption, despite a marked increase in solid feed intake before and after weaning in litters subjected to IS with an extended lactation compared with conventionally weaned pigs. The potential limitation of mannitol as a measure of the intestinal surface area needs to be considered when interpreting these results [158], but the reported increase in galactose absorption highlights the importance of solid feed intake in GIT function, given that a change from a milk-based diet to solid feed increases the Na+-K+ ATP pump activity needed for galactose absorption via a sodium–glucose-linked transporter [160].animals-13-01644-t003_Table 3Table 3Overview of different variants of intermittent suckling (IS).ReferenceStart Time of Gradual Weaning 1Separation Length 2DurationWeaning AgeCombined with Another TechniqueSmith, 1961 Experiment 1 [153]Days 31–3512 h per day25–21 days56 daysExtended lactation + mating in lactationSmith, 1961 Experiment 2 [153]Day 2112 h per day35 days56 daysMating in lactation + extended lactationKirkwood and Smith, 1983 [161]Day 1412 h per day14 days28 daysMating in lactationHenderson and Hughes, 1984 [154]Day 1012 h per day25 days35 daysMating in lactation + extended lactationGrinwich and McKay, 1985 [155]Day 213 or 22 h per day14 days35 daysMating in lactation + extended lactationChapple et al., 1989 [162]Day 144 × 5 h intervals per day with 1 h back with the sow in between14 days28 days+/− hydrocortisone injectionsCosta and Varley, 1995 [163]Day 123 h per day9 days21 daysMating in lactationNabuurs et al., 1993 [164] Experiment 3 onlyDays 16–188 h per day14 days30–32 daysControl piglets not offered creep feedNabuurs et al., 1996 [159]Day 188 h per day14 days32 daysWith and without creep feed (experiment 2)Control piglets did not receive creep feedKuller et al., 2004 [165]Day 1612 h per day11 days27 days Kuller et al., 2007 [166]Day 1412 h per day11 days25 days Berkeveld et al., 2007 [156]Day 1412 h per day and 2 × 6 h intervals per day with 6 h back with the sow in between27 to 31 days41 to 45 days(controls were 21 days)Extended lactationMillet et al., 2008 [167]Day 147 h per day14 days28 daysFlavor recognitionBerkeveld et al., 2009 [157]Day 19 or 2610 h per day7 or 14 days26 or 33 daysExtended lactation for two of the four treatment groupsKluivers-Poodt et al., 2010 [168]Day 1412 h per day and 2 × 6 h intervals per day with 6 h back with the sow in betweenNot specified23 days after ovulationMeasurements in sows onlyDowning et al., 2011 [169]Day 14, 16 or 1816 h per day3 days26 daysMating in lactationMcDonald et al., 2013 [170]Day 2116 or 8 h per day3 days28 daysMating in lactationFrobose et al., 2015 [171]Day 1812 h per day7 days25 daysCo-mingling and alternating between two sows every 12 h, split weaning, 24 h separationTurpin et al., 2016 [158]Day 21 or Day 288 h per day7 days28 days or 35 daysExtended lactation for one of the three treatment groupsTurpin et al., 2016 [172]Day 228 h per day7 days29 days Turpin et al., 2017 [71]Day 188 h per day7 days25 daysOne of the three treatment groups exposed to co-minglingTurpin et al., 2017 [72]Day 158 h7 days22 days2 × 2 factorial with IS and co-mingling as main effectsTurpin, 2017 [173]Day 20 or 238 or 16 h6 days or 3 nights26 days van Nieuwamerongen et al., 2017 [125]Day 3510 h7 days63 daysMulti-suckling, sow-controlled housing, mating in lactation and extended lactation1 Expressed as day of lactation. 2 Pattern continued from start time to weaning unless otherwise specified.4.2.4. Intermittent Suckling and Conventional Lactation: Effects on Litter PerformanceMany studies have combined conventional weaning ages (i.e., 3 to 4 weeks) with an IS regimen to improve creep feed intake by promoting exploratory behavior when the piglets are separated from the sow. However, the length of time that the piglets are separated from the sow, the style of separation housing and the age at which IS is initiated all seem to play a role in the outcomes. Litters subjected to separation times of 12 h for 7 days consistently had an increase in creep feed intake compared with control litters that remained with the sow [166,166,171], and this translated into improvements in solid feed intake after weaning [165,166]. Non-significant increases in creep feed intake have also been reported with 12 h separation times [154]. There are a limited number of studies examining separation times greater than 12 h. Chapple et al. (1989) [162] used a total separation time of 20 h per day for 14 days and did not find a difference in creep feed intake between limited nursing piglets and controls, but the piglets did return to the sow every 5 h for 1 h. While the authors postulated that this could have been due to the acceptability of the diet or litter-to-litter variability of feed intake (discussed below), this result supports findings from Berkeveld et al. (2007) [156], where segmenting the separation time over multiple periods in a day resulted in lower feed intake levels during lactation, probably due to a greater dependency on milk. Turpin (2017) [173] examined separation times of 16 h over 3 days and found an increase in creep feed intake, although these results need to be interpreted with some caution because creep feed was already stimulated in the treatment group before the start of the IS intervention, making it difficult to conclude that IS was the only contributing factor for these piglets.When separation times were less than 12 h, the ability of IS to stimulate creep feed intake was not guaranteed, with IS litters either consuming more [71,72,158] or the same amount of creep feed as the control litters that remained with the sow [71,157,158,172,174]. Only one study reported a reduction in feed intake in IS litters compared with continuously suckled litters [167]. Collectively, these outcomes suggest that IS litters are more likely to remain dependent on milk for their growth when separation times are less than 12 h. One of the challenges to consider when assessing strategies to improve creep feed intake is that individual creep feed consumption is highly variable [33,34,35]. Studies that have looked at the individual eating patterns of piglets exposed to IS have reported that IS increases the creep intake of piglets that were already eating before the period of separation but does not increase the percentage of piglets eating creep feed within a litter [166,173]. However, when litters were categorized as either high or low creep feed intake litters, the piglets in the low creep feed intake group that were subjected to IS visited the feeder more frequently than their control counterparts, which translated to more time spent at the feeder on the second day after weaning [175].In conventional production systems, increases in creep feed consumption can be achieved as lactation progresses, with studies reporting 60–80% of total creep feed intake occurring in the 6 days before weaning [7,176] and higher feed intake rates achieved (in a non-linear pattern) as the piglets’ age increases [177]. This phenomenon also occurs with gradual weaning strategies, where postponing the onset of IS for 1 week (14 days vs. 21 days) caused a greater increase in creep feed intake [178]. In addition to the length of the daily separation period and the age at which separation from the sow starts, the style of separation housing has not received much attention in studies to date, but housing IS litters in the same room as their control counterparts will likely have a negative impact on creep feed intake since nursings are usually synchronized within a room and are accompanied by noise, which could distract the IS litters [165].Most IS studies using conventional weaning ages with IS starting in the week before weaning have observed a reduction in piglet growth rate at the start of IS when the separation period is greater than 7 h per day [72,157,158,162,165,166,170,173], with the exception of only a couple of studies where IS had no impact on growth rates before weaning compared with continuously suckled litters [71,172]. In most cases, piglets compensated for the reduction in growth at the start of IS during the latter part of the first week after weaning [72,156,163,166,172], but in contrast to studies where IS is combined with extended lactation, it seems that the weaning-associated growth check in the first few days after weaning is not necessarily prevented or reduced when IS for one week or less is implemented with a conventional weaning age [71,72,157]. Furthermore, there has been no evidence to suggest that gradual weaning through IS improves longer-term growth or market weights [158,166,171].4.2.5. Intermittent Suckling and Conventional Lactation: Impacts on the GIT, Behavior and WelfareDespite varying results regarding pre-weaning creep feed intake and growth, no studies have reported gradually weaned pigs performing worse in the immediate post-weaning period compared with pigs subjected to abrupt weaning. This suggests that the beneficial effects of gradual weaning might also be mediated by factors other than the potential for improvements in feed intake. This is supported by findings from Berkeveld et al. (2009) [157], where one week of IS (10 h of separation from the sow per day) before weaning at day 26 prevented post-weaning villus atrophy, even though feed intake before and within the first 2 days after weaning was not improved compared with conventionally weaned pigs. To date, only a few studies have examined the behavior of gradually weaned pigs after weaning. More focus has been placed on piglet behavior before weaning and during IS because of the perceived welfare implications of intermittent sow and piglet separation and udder damage from excessive piglet attention after re-joining with the sow. Results from continuous and instantaneous scan sampling studies have reported that IS pigs showed the lowest levels of manipulative behavior [71,174] and the highest levels of relaxed behavior [71] on the day of or the day after weaning compared with abruptly weaned pigs. These differences in behavior were only supported by reductions in the stress marker cortisol when the blood samples were taken on the day of or 24 h after weaning [71,72,174].The act of repeated maternal separation associated with gradual weaning during lactation on the piglets’ stress response is a potential welfare concern [179]. Berkeveld et al. (2007) [156] reported a peak in total piglet activity and vocalization on the first day of IS during the 12 h separation, but 2 days later, the total piglet activity had decreased and then stabilized for the remainder of lactation. This result was supported by the study of Turpin et al. (2016) [172], in which there was an elevation in plasma cortisol after the first 4 to 6 h of separation for the IS piglets compared with the continuously suckled piglets, but this effect disappeared by the next measurement point 6 days later. A similar pattern was observed in sow cortisol levels during the first separation period from the piglets using 12 h separation times [168]. Piglets subjected to long periods of separation (16 h overnight for 5 nights at 2 to 3 weeks of age) had an increased probability of passive lying behavior during separation times compared with continuously suckled controls, which can be interpreted as a stress behavior [38]. This suggests that extended separation times could have a detrimental effect on piglet welfare [180]. In addition to cortisol, other markers of piglet well-being, including acute phase proteins and cytokines, have been measured before and after weaning in piglets subject to IS across a number of studies using mostly an 8 h separation for 6 to 7 days, but with no obvious patterns in response [71,72,157,172,173].4.3. Gradual Weaning in Combination with Co-MinglingGroup lactation systems that promote piglets spending gradually less time with their sows in combination with pre-weaning socialization generally result in improvements in post-weaning behavior and performance compared to conventionally weaned pigs. Similar results have also been achieved when gradual weaning and piglet socialization are applied in a more structured way. Turpin et al. (2017) [71] investigated the behavior and performance of litters that were separated during IS treatments (8 h per day for 7 days) and then socialized with another litter during the separation period. The addition of socialization with another litter before weaning improved pre- and post-weaning feed intakes and post-weaning growth rates in an additive manner when the socialized groups could remain together after weaning. This result highlights the advantage of familiarity within a group rather than the development of social skills since similar findings were not reported in other studies where co-mingling was combined with IS using a conventional weaning age (i.e., 3 to 4 weeks) [71,171]. Applying co-mingling intermittently may have had an impact on these outcomes, and the requirement for extra labor and potentially specialized housing to achieve IS in combination with co-mingling may limit the commercial application of these techniques.Nevertheless, there have been promising performance, health and behavioral results from a study that has combined all the techniques described in this review: multi-suckling, IS, sow-controlled housing, and extended lactation [125]. In this study, all piglets were subjected to a multi-suckling system with five sows in each group. One treatment group was abruptly weaned at 4 weeks of age, and the other was subject to IS, involving 10 h of separation per day during the fifth week of lactation. After this, the sows were then able to separate themselves from the piglets voluntarily until weaning at 9 weeks of age. The authors chose to compare results by age, but it must be recognized that both the duration of the suckling period (4 weeks vs. 9 weeks) and the physiological age will have had an influence on the measured parameters. Nonetheless, the combination of multi-suckling with IS and extended lactation prevented the growth check at the start of IS and at weaning [125]. The multi-suckled IS pigs also showed a lower level of belly-nosing and a more solid fecal consistency, indicating a lower occurrence of diarrhea, compared with the piglets that were only exposed to multi-suckling and weaned at 4 weeks of age. Regarding reproductive performance, 83% of the treatment sows showed an estrus in lactation, and of those sows, 95% were pregnant following insemination during lactation [125]. These results highlight that some adjustments to current multi-suckling housing designs can facilitate the implementation of IS and/or sow-controlled housing; therefore, removing the need to wean early without sacrificing sow performance.5. ConclusionsIt was hypothesized in this review that pre-weaning socialization presents an opportunity to reduce the risk of mixing stress and promote social eating behavior, while intermittent suckling (IS), a form of gradual weaning, can increase exploratory behavior and allow the piglets to become accustomed to maternal separation. Numerous studies demonstrated a reduced level of aggression among socialized piglets immediately after weaning, independent of the housing system, group size and age at the start of co-mingling prior to weaning. Studies have generally shown a positive influence on post-weaning performance, particularly in the immediate post-weaning period. Yet, attention is needed to prevent reductions in pre-weaning weight gains due to the timing of co-mingling, a higher space allowance and, therefore, more exercise, a lower creep feed intake, the disruption of suckling, cross-suckling, enhanced active behaviors and others. In the majority of publications, piglets from different litters were co-mingled between 7 and 14 days of age, which seems to be consistent with their (semi-)natural behavior. The size of the group seemed to depend on the housing system; in multi-litter systems, groups consisted mainly of four or more litters, whereas, in other systems, it was usually restricted to four litters or less. It is not clear whether this is due to practical limitations or scientific grounds, yet larger groups may cause more aggression between piglets and competitive behavior at the udder. Overall, few studies have consistently studied main variables, such as housing system, group size and age at the start of co-mingling prior to weaning, which prevents us from drawing firm conclusions on the effect of these factors.From the piglet perspective, although there are well-known impacts of IS on sows’ reproduction during lactation, IS has also been studied as a strategy to promote creep feed intake during lactation. Many studies have combined conventional weaning ages (i.e., 3 to 4 weeks) with an IS regimen to improve creep feed intake by promoting exploratory behavior when the piglets are separated from the sow. However, the length of time and frequency with which the piglets are separated from the sow, the style of separation housing and the age at which IS is initiated all seem to play a role in the outcomes. Intermittently suckled litters are more likely to remain dependent on milk for their growth when separation times are less than 12 h per day. Nonetheless, no studies have reported gradually weaned pigs performing worse in the immediate post-weaning period compared with pigs subjected to abrupt weaning. This suggests that the beneficial effects of IS might also be mediated by factors other than the potential for improvements in feed intake. If IS is applied with an extended lactation period, creep feed intake is likely to be sufficient to compensate for the reduction in milk consumption during the time of separation from the sow.Pre-weaning socialization, or co-mingling, and IS are techniques that are potentially adaptable to a commercial setting to promote a more gradual weaning process and reduce post-weaning stress for piglets with better welfare outcomes and more optimal performance and GIT structure and functioning. The economic benefits of either or both practices, if implemented commercially, will be determined by the level of performance and/or welfare improvements and impacts on other variables, for example, a reduction in medication use after weaning.

animals : an open access journal from mdpi

10.3390/ani13050951

PMC10000214

Equine leaky gut syndrome is characterized by gastrointestinal hyperpermeability and may be associated with adverse health effects in horses. The purpose was to evaluate the effects of a prebiotic Aspergillus oryzae product (SUPP) on the stress-induced leakiness of the gut. For 28 days, 8 horses received a diet containing the prebiotic or an unsupplemented diet (CO). On Days 0 and 28, horses were dosed with a compound (iohexol) that should only leak out of the gastrointestinal tract if the gut walls become leaky. Immediately following iohexol administration, four horses from each feeding group underwent 60 min of transport immediately followed by a moderate-intensity exercise bout of 30 min (EX), and the remaining horses were maintained as sedentary controls (SED). Blood was sampled before iohexol, immediately after trailering, and at 0, 1, 2, 4, and 8 h post-exercise. Blood was analyzed for iohexol, as well as lipopolysaccharide (a compound found in the gastrointestinal tract that can leak out) and serum amyloid A (a marker of inflammatory response). EX resulted in a significant increase in plasma iohexol in both CO and SUPP groups on Day 0; this increase was not seen in SED horses. On Day 28, EX increased plasma iohexol only in the CO feeding group; this increase was completely prevented by the provision of SUPP. It is concluded that combined transport and exercise induce leaky gut. Dietary SUPP prevents this and therefore may be a useful prophylactic for pathologies associated with gastrointestinal hyperpermeability in horses.

Equine leaky gut syndrome is characterized by gastrointestinal hyperpermeability and may be associated with adverse health effects in horses. The purpose was to evaluate the effects of a prebiotic Aspergillus oryzae product (SUPP) on stress-induced gastrointestinal hyperpermeability. Eight horses received a diet containing SUPP (0.02 g/kg BW) or an unsupplemented diet (CO) (n = 4 per group) for 28 days. On Days 0 and 28, horses were intubated with an indigestible marker of gastrointestinal permeability (iohexol). Half the horses from each feeding group underwent 60 min of transport by trailer immediately followed by a moderate-intensity exercise bout of 30 min (EX), and the remaining horses stayed in stalls as controls (SED). Blood was sampled before iohexol, immediately after trailering, and at 0, 1, 2, 4, and 8 h post-exercise. At the end of the feeding period, horses were washed out for 28 days before being assigned to the opposite feeding group, and the study was replicated. Blood was analyzed for iohexol (HPLC), lipopolysaccharide (ELISA), and serum amyloid A (latex agglutination assay). Data were analyzed using three-way and two-way ANOVA. On Day 0, the combined challenge of trailer transport and exercise significantly increased plasma iohexol in both feeding groups; this increase was not seen in SED horses. On Day 28, EX increased plasma iohexol only in the CO feeding group; this increase was completely prevented by the provision of SUPP. It is concluded that combined transport and exercise induce gastrointestinal hyperpermeability. Dietary SUPP prevents this and therefore may be a useful prophylactic for pathologies associated with gastrointestinal hyperpermeability in horses.

1. IntroductionLeaky gut syndrome (LGS) is characterized by gastrointestinal hyperpermeability and increased accessibility of the systemic environment to compounds that are normally sequestered within the gastrointestinal lumen [1]. The contribution of LGS to equine disease is poorly understood, and its mitigation by dietary interventions has not been described in the literature. An MSc thesis from Michigan State University [2] describes a study in which oral phenylbutazone contributed to the development of gastrointestinal hyperpermeability in 18 Arabian horses, suggesting that gastric ulceration, phenylbutazone administration, or both, contribute to the development of LGS in horses. Evidence also implicates diets high in starch as complicit in gastrointestinal hyperpermeability [3]. Exercise is another likely candidate as an LGS risk factor but has not been clearly described in horses. Research in humans, however, provides evidence for a positive correlation between exercise intensity/duration and hyperpermeability of the gastrointestinal tract [4,5,6]. A recent study in eight horses reports that the combination of exercise and trailer transport induces an increase in gastrointestinal permeability, as well as increased serum amyloid A and lipopolysaccharide [7]. Whilst the pathophysiological consequences of LGS are as vaguely characterized as its triggers, there is evidence that, depending on the degree of inflammatory response to luminal toxins, LGS may impair skeletal muscle metabolism [8], and contribute to metabolic dysfunction [9,10], allergies [11,12], and inflammatory diseases such as arthritis [13]. Dietary interventions with evidence for an ability to protect against the development or clinical consequences of LGS will make an important contribution to preserving robust equine health.Perhaps due (at least in part) to the incomplete picture defining the cause-and-effect of LGS, interventions tend to rely heavily on the management of downstream clinical consequences. To the authors’ knowledge, there are currently no feed supplements or pharmaceutical drugs that have been evaluated against the gastrointestinal hyperpermeability that is the cornerstone of LGS. A commonly reported feature of LGS in non-equine species is gastrointestinal dysbiosis, and there is evidence that this dysbiosis contributes to the development of hyperpermeability [14,15,16,17]. Dysbiosis is likely in horses receiving a high-starch diet [3,16], and in horses experiencing physiological stress [16]. Thus, interventions with potential to stabilize gastrointestinal microbiota may protect against the development of hyperpermeability under conditions of stress. Aspergillus oryzae is a filamentous fungus, which has demonstrated the ability to amplify the abundance of probiotic microbes (particularly Bifidobacterium pseudolongum) whilst protecting DSS-challenged mice against colitis [18]. The fermentation product of A. oryzae also promotes fiber-degrading bacteria in the rumen and hindgut when fed to lactating dairy cows [19]. In addition to evidence for a prebiotic-like effect, A. oryzae also exerts a marked anti-inflammatory effect in LPS-stimulated polymorphonuclear cells and improves the structure of gastrointestinal lumen (i.e., villus height–crypt ratio) in broiler chickens [20]. Furthermore, the administration of a postbiotic from A. oryzae to calves prevented the increase in intestinal permeability associated with exposure to high ambient temperature [21]. These data support the hypothesis that A. oryzae protects against stress-induced hyperpermeability by amplifying the abundance of a healthy gastrointestinal microbiome. Accordingly, the purpose of the current study was to evaluate the effects of a fungal prebiotic produced through a proprietary fermentation process with A. oryzae (SUPP; BioZyme Inc.; St. Joseph, MO, USA) on equine gastrointestinal hyperpermeability induced by a combination of trailer transport and moderate-intensity exercise horses. The objectives were to characterize the effect of a dietary A. oryzae prebiotic on the appearance and disappearance of an oral permeability marker (iohexol) in the blood of horses challenged with combined transport and exercise stress, and to correlate observed effects with those on downstream evidence of inflammation (serum amyloid A (SAA)) and translocation of enteric endotoxin (lipopolysaccharide (LPS)).2. Materials and MethodsCare and use of animals was reviewed and approved by the University of Guelph Animal Care Committee in compliance with the guidelines published by the Canadian Council on Animal Care (Approval Number 3800).2.1. HorsesEight (8) healthy mares (Age: 14.2 ± 3.7 years; body weight: 570 ± 47.4 kg) from the Arkell Equine Research Station, University of Guelph, were included in the randomized, partial cross-over trial. The horses were group-housed in an open turnout area, with unrestricted access to a large covered shelter bedded with straw, 1st cut Timothy hay, water, and trace mineral salt. Two hundred and fifty (250) g of a 12% maintenance pellet rationa was provided once per day (morning) (Table 1). Horses were all accustomed to a lifestyle that did not include forced exercise. At the beginning of the study, all 8 horses were randomized into one of two feeding groups (n = 4 per group): Group A: unsupplemented control diet (CO); Group B: diet containing A. oryzae prebioticb (SUPP; 0.02 g/kg BW). SUPP was a textured, unpelleted product and was top-dressed onto the horse’s individual pelleted feed once per day. Horses consumed their pelleted ration with or without SUPP once per day in individual stalls. Once their feed was completely consumed, they were returned to the outdoor turnout area. Within each feeding group, horses were further divided into stress-challenged (EX—see below for details) or non-challenged sedentary controls (SED) (n = 2 per group per replicate). Horses received their assigned diet for 28 days. On Days 0 and 28, one SED and one EX horse were evaluated in the morning, and a second SED and second EX horse were evaluated in the afternoon. At the end of the 28-day feeding period, horses were washed out for 28 days, and then assigned to the opposite feeding group for an additional 28 days. The trial was then repeated, for a final ‘n’ of 8 per feeding group (i.e., 4 × EX and 4 × SED per feeding group). Horses were tested at the same time of day (morning or afternoon) in both study periods.On study days, horses remained in their turnout area with unrestricted water access, but from which all feed had been removed. Following 12 h of fasting, horses were stalled and administered via nasogastric tube an indigestible marker of gastrointestinal permeability (iohexolc; 5.6% solution, 1.0 mL/per kg BW; 56 mg/kg BW) by a licensed veterinary professional [7]. The procedure was conducted in the absence of any sedation, so as not to interfere with normal gastrointestinal motility [22].2.2. Stress ChallengeHorses were challenged with combined trailer transport and exercise, which we have previously demonstrated to produce a measurable and significant increase in gastrointestinal hyperpermeability [7]. Briefly, following the administration of iohexol, one EX horse was walked onto a 2-horse trailer for a 60 min drive to the Equine Sports Medicine and Reproduction Centre, University of Guelph. Once at the facility, a heart rate (HR) monitord was attached to the horse using a flexible belly-band, and the horse was free-lunged around an indoor arena (5 min’ walk, 10 min trot (left), 10 min trot (right), and 5 min’ walk) on a sand footing for 30 min. Horses were encouraged to achieve an exercise intensity that resulted in a HR of approximately 150 bpm during the trot, in order to encourage the horse to work at or beyond the anaerobic threshold [23]. At the cessation of exercise, EX horses returned to the group housing yard directly and were turned out with unrestricted access to hay and water. This challenge has previously been demonstrated to produce gastrointestinal hyperpermeability in horses [6].Following the application of topical lidocaine at the jugular groove, blood was sampled from the jugular vein immediately before iohexol administration (P1), immediately after trailering (P2), immediately after exercise (P3), and then 1 (P4), 2 (P5), 4 (P6), and 8 h (P7) post-exercise. Blood samples were cooled on ice, centrifuged within 2 h of collection, and the recovered plasma was frozen (−20 °C) until analysis. Manure samples were collected within 2 min of voiding before the horse walked into the trailer, at the end of 60 min of transport, and the first manure after exercise. 2.3. Non-Challenged ControlsSED horses received iohexol at the same time as the EX horses, and blood was sampled at the same time as the EX horses. After receiving iohexol they were returned to the group housing area with free access to water. Hay was provided upon return of the EX horse from transport and exercise. 2.4. Sample AnalysisAll chemicals and reagents were purchased from Sigma Aldrichf, unless otherwise stated. Plasma samples were analyzed for systemic inflammation (serum amyloid A and lipopolysaccharide (LPS)) biomarkers, and an exogenous marker of gastrointestinal permeability (iohexol). Plasma iohexol was determined via HPLC (Agilent 1200 series HPLC gradient system), which was used to quantify plasma iohexol (μ g/mL) with UV detection at 254 nm, as previously described [7] (intra- and inter-assay CV: 3.106 and 4.217%, respectively). SAA was determined by Eiken Serum Amyloid A latex agglutination assay at a commercial laboratory (Animal Health Laboratory, University of Guelph). Plasma samples, acclimated at room temperature, were analyzed in duplicate for LPS (pg/mL) using an equine-specific quantitative sandwich ELISA kit according to manufacturerh instructions (inter- and intra-assay coefficient of variability: 1.5 and 1.6%, respectively). A standard curve was used to generate a linear regression equation, which was used to calculate LPS concentrations in each sample.2.5. Data AnalysisData analysis was conducted using SigmaPloti (Version 14.2). Data are presented as mean ± SD unless otherwise indicated. Normality of data was determined using the Shapiro–Wilk test. Three-way ANOVA was used to detect interactions between feeding groups, stress challenge, and time after iohexol administration. Two-way ANOVA was used to identify significant differences between feeding groups in SED and EX horses on Day 0 and Day 28 with respect to stress challenge and time after iohexol administration. The Holm–Sidak post-hoc test was used to identify significantly different means when a significant F-ratio was calculated. Significance was accepted at p < 0.05. 3. Results3.1. Gastrointestinal Barrier Function3.1.1. Control Diet (Figure 1)Day 0: In SED horses receiving the CO diet, there was no significant change in plasma iohexol at any time between P1 (0.56 ± 0.02 ug/mL) and P7 (0.69 ± 0.04 ug/mL) (p = 0.26). EX horses demonstrated a significant increase in plasma iohexol between P1 (0.52 ± 0.03 ug/mL) and P3 (1.14 ± 0.08 ug/mL) (p = 0.02). Plasma iohexol was significantly higher in EX horses than in SED horses at P2 (SED: 0.71 ± 0.06 ug/mL; EX: 1.02 ± 0.18 ug/mL) (p = 0.04) and P3 (SED: 0.75 ± 0.09 ug/mL; EX: 1.14 ± 0.08 ug/mL) (p = 0.01) (Figure 1). Day 28: In SED horses receiving the CO diet, there was no significant change in plasma iohexol at any time between P1 (0.48 ± 0.04 ug/mL) and P7 (0.60 ± 0.06 ug/mL) (p = 0.44). EX horses demonstrated a significant increase in plasma iohexol between P1 (0.58 ± 0.09 ug/mL) and P3 (1.07 ± 0.06 ug/mL) (p = 0.006). Plasma iohexol was significantly higher in EX horses than in SED horses at P2 (SED: 0.54 ± 0.06 ug/mL; EX: 1.01 ± 0.12 ug/mL) (p < 0.001), P3 (SED: 0.56 ± 0.07 ug/mL; EX: 1.07 ± 0.12 ug/mL) (p < 0.001) and P4 (SED: 0.59 ± 0.04 ug/mL; EX: 1.00 ± 0.10 ug/mL) (p < 0.001) (Figure 1). Day 0 vs. Day 28: In SED horses, plasma iohexol was significantly higher on Day 0 than on Day 28 at P3 and P5 (p = 0.04 and 0.05, respectively). There were no significant differences between Day 0 and Day 28 in EX horses (p = 0.23) (Figure 1).3.1.2. Supplemented Diet (Figure 2)Day 0: In SED horses receiving the SUPP diet, there was a significant increase in plasma iohexol between P1 (0.51 ± 0.03 ug/mL) and P2 (0.87 ± 0.04 ug/mL) (p = 0.005), P3 (0.82 ± 0.06 ug/mL) (p = 0.02) and P4 (0.97 ± 0.09 ug/mL) (p < 0.001). EX horses demonstrated a significant increase in plasma iohexol between P1 (0.70 ± 0.15 ug/mL) and P3 (1.75 ± 0.19 ug/mL) (p = 0.01). Plasma iohexol was significantly higher in EX horses than in SED horses at P3 (SED: 0.82 ± 0.06 ug/mL; EX: 1.75 ± 0.19 ug/mL) (p < 0.001) (Figure 2). Day 28: In SED horses receiving the SUPP diet, there was no significant change in plasma iohexol at any time between P1 (0.49 ± 0.05 ug/mL) and P7 (0.70 ± 0.05 ug/mL) (p = 0.43). There was also no significant increase in plasma iohexol in EX horses at any time between P1 (0.87 ± 0.23 ug/mL) and P7 (0.56 ± 0.12 ug/mL) (p = 0.36)(Figure 2). 3.1.3. Day 0 and Day 28 in Supplemented and Control DietsOn Day 0, iohexol tended to be higher in SUPP than CO horses (p = 0.053). Overall iohexol was significantly elevated in EX horses at P2, P3, (p < 0.001) and P4 (p = 0.02) compared with P1, but there were no differences between treatment groups (Figure 2) On Day 28, iohexol was significantly higher overall in CO horses compared with SUPP horses (p = 0.008). Overall, iohexol was significantly higher at P3 than P1, but there were no significant differences between treatment groups (Figure 2). 3.2. Systemic Inflammation3.2.1. Serum Amyloid A (SAA; Table 2)Control DietDay 0: In SED horses receiving the CO diet, there was no significant change in SAA at any time between P1 (0.10 ± 0.1 μg/mL) and P7 (0.10 ± 0.1 μg/mL) (p = 0.78). There was also no significant change in EX horses in SAA between P1 (0.22 ± 0.16 μg/mL) and P7 (0.86 ± 0.56 μg/mL) (p = 0.70). Overall, SAA was significantly higher in EX than in SED horses (p = 0.01), but there were no significant differences between groups at any specific time point (Table 2). Day 28: In SED horses receiving the CO diet, there was no significant change in SAA at any time between P1 (0.0 ± 0.0 μg/mL) and P7 (0.10 ± 0.10 μg/mL) (p = 0.92). There was also no significant change in EX horses in SAA between P1 (0.15 ± 0.15 ug/mL) and P7 (0.20 ± 0.20 μg/mL) (p = 0.96). In horses receiving the CO diet, SED horses had significantly lower SAA than EX horses overall (p = 0.04), but there were no significant differences at individual time points (Table 2). Supplemented DietDay 0: In SED horses receiving the SUPP diet, there was no significant change in SAA at any time between P1 (0.33 ± 0.33 μg/mL) and P7 (0.15 ± 0.15 μg/mL) (p = 0.71). There was also no significant change in EX horses SAA between P1 (0.08 ± 0.08 μg/mL) and P7 (0.30 ± 0.30 μg/mL) (p = 0.70). There were no significant differences between SED and EX at any specific time point on Day 0 (Table 2). Day 28: In SED horses receiving the SUPP diet, there was no significant change in SAA at any time between P1 (0.17 ± 0.17 μg/mL) and P7 (0.35 ± 0.15 μg/mL) (p = 0.59). There was also no significant change in EX horses in SAA between P1 (0.35 ± 0.25 μg/mL) and P7 (1.00 ± 0.53 μg/mL) (p = 0.96). Overall, SAA was significantly higher in EX than in SED horses (p = 0.02), but there were no significant differences between groups at specific time points (Table 2).Day 0 and Day 28 in Supplemented and Control DietsOn Day 0, there were no differences in SAA between SUPP and CO horses (p = 0.257). Overall, SAA was significantly higher in EX than SED horses (p = 0.015), primarily owing to significantly higher SAA in EX than SED horses in CO horses (p = 0.002) that was not observed in SUPP horses (p = 0.826) (Table 2). On Day 28, SAA was significantly higher overall in SUPP horses compared with CO horses (p = 0.01). There was no significant difference in SAA between EX and SED horses overall, but SAA was significantly higher in SED horses than EX horses in horses receiving the supplemented diet (p = 0.05) (Table 2).3.2.2. Lipopolysaccharide (LPS; Table 2)Control Diet Day 0: In SED horses receiving the CO diet, there was no significant change in LPS at any time between P1 (2.10 ± 0.09 pg/mL) and P7 (2.13 ± 0.12 pg/mL) (p = 0.71). There was also no significant change in EX horses in LPS between P1 (2.18 ± 0.06 pg/mL) and P7 (2.21 ± 0.10 pg/mL) (p = 0.99). Overall, LPS was significantly higher in EX than in SED horses (p = 0.02), but there were no significant differences between SED and EX at any specific time point (Table 2). Day 28: In SED horses receiving the CO diet, there was no significant change in LPS at any time between P1 (2.1 ± 0.09 pg/mL) and P7 (2.1 ± 0.05 pg/mL) (p = 0.94). There was also no significant change in EX horses in LPS between P1 (2.14 ± 0.03 pg/mL) and P7 (2.10 ± 0.08 pg/mL) (p = 0.94). Overall, LPS was significantly higher in EX than in SED horses (p = 0.004), but there were no significant differences between groups at specific time points (Table 2). Supplemented DietDay 0: In SED horses receiving the SUPP diet, there was no significant change in LPS at any time between P1 (2.15 ± 0.04 pg/mL) and P7 (2.17 ± 0.04 pg/mL) (p = 0.91). There was also no significant change in EX horses LPS between P1 (2.06 ± 0.04 pg/mL) and P7 (2.13 ± 0.01 pg/mL) (p = 0.98). LPS was significantly higher in SED than EX horses (p = 0.03), but there were no significant differences between groups at specific time points (Table 2).Day 28: In SED horses receiving the SUPP diet, there was no significant change in LPS at any time between P1 (2.20 ± 0.08 pg/mL) and P7 (2.18 ± 0.07 pg/mL) (p = 0.90). There was also no significant change in EX horses in LPS between P1 (2.06 ± 0.04 pg/mL) and P7 (2.06 ± 0.05 pg/mL) (p = 0.97). LPS was significantly higher in SED than EX horses overall (p < 0.001), as well as at P5 (p = 0.01) and P6 (p = 0.05) (Table 2). Day 0 and Day 28 in Supplemented and Control DietsOn Day 0, there were no differences in LPS between SUPP and CO horses (p = 0.346). There was also no significant difference between EX and SED horses overall (p = 0.268). LPS was significantly higher in EX than SED horses in the CO group (p = 0.003), but there were no significant differences in LPS between EX and SED horses in the SUPP group (p = 0.068) (Table 2).On Day 28, there were no differences in LPS between SUPP and CO horses (p = 0.674). There was also no significant difference between EX and SED horses overall (p = 0.392). LPS was significantly higher in EX than SED horses in the CO group (p = 0.004) and significantly lower in EX than SED in the SUPP group (p < 0.001) (Table 2).4. DiscussionThe purpose of the current study was to quantify the effect of a dietary A. oryzae prebiotic on gastrointestinal permeability in horses challenged with combined transport and exercise stress. The main finding was that 28 days of supplementation with the A. oryzae prebiotic completely eradicated stress-induced gastrointestinal permeability in this group of horses.We have previously demonstrated that the combination of transport and exercise stress model utilized in the current study produces gastrointestinal hyperpermeability and an increase in blood biomarkers that evidence transient, low-grade systemic inflammation [7]. Like our previous study, we report herein that 60 min of trailer transport immediately preceding half an hour of moderate-intensity exercise is a clear, reproducible model of gastrointestinal hyperpermeability. On Day 0 for both feeding groups, the stress model resulted in a significant uptick in the systemic appearance of orally administered iohexol that was not seen in unstressed controls. That this spike in the systemic appearance of iohexol was absent in stressed horses in the SUPP feeding group on Day 28 provides strong evidence for the role of A. oryzae prebiotic in protecting gastrointestinal barrier function in horses during stress. The mechanism for this blockade is not known but may be associated with an effect of A. oryzae prebiotic on the enteric microbiome. A. oryzae strongly increases the relative abundance of anti-inflammatory bacterial strains such as Bifidobacterium [18,24] and important fiber-degrading bacteria such as Ruminococcaceae [19]. Dietary provision of Bifidobacterium-based probiotics to obese humans results in a marked decrease in gastrointestinal hyperpermeability [25], which provides support for the hypothesis that A. oryzae prebiotic protects the enteric barrier from stress-induced hyperpermeability via its modulation of the gastrointestinal microbiome. This hypothesis should be tested in future studies.When dietary groups were combined, there was an overall increase in SAA in response to our stress challenge, consistent with our previous study [7], but this effect was not observed when analyzing dietary groups individually. SAA is the major acute phase protein in the horse. While it is a highly sensitive indicator of an inflammatory event, it is not specific, and its production can be markedly increased in the presence of almost any inflammatory challenge [26]. The vast majority of SAA is produced by hepatocytes, but small amounts may also be produced by enterocytes [27]. Our small sample size, together with SAA fluctuations in both EX and SED groups that were unrelated to our stress challenge, likely contributed to the lack of statistical increase in SAA within groups. Consequently, the effect of A. oryzae prebiotic on this biomarker remains unknown. Owing to the highly plastic nature of SAA in vivo, future studies to evaluate the effects of the A. oryzae prebiotic on this outcome measure may benefit from controlled in vitro assessment of enterocyte-specific production of SAA [27].The marked gastrointestinal hyperpermeability that was observed in the current study in EX horses in the control feeding group on Days 0 and 28 was not associated with a significant time-dependent increase in circulating LPS, and like SAA, this may have been due, at least in part, to our small sample size. But the overall serum LPS concentration of EX horses was significantly higher than SED horses. Surprisingly, however, serum LPS was significantly lower in EX than in SED horses for the A. oryzae feeding group. This result is probably not associated with the supplement because it was observed both on Day 0 (prior to beginning supplementation) and on Day 28, so instead is more likely an artifact of randomizing a small number of animals to the feeding groups. Furthermore, our maximum LPS concentration of 2.24 pg/mL in either feeding group is well within the reference interval for the normal flux of systemic LPS in healthy horses [26]. Future studies designed to detect the effect of the dietary A. oryzae prebiotic on the translocation of enteric LPS at levels expected to be associated with disease will require a stronger stress challenge such as non-steroidal anti-inflammatory drugs [2,27].The current study had fewer animals in each treatment group than our previous study, which may have resulted in the current study being underpowered to detect the effects of stress and/or diet on SAA and LPS.5. ConclusionsIn conclusion, the data presented herein provide compelling evidence for a protective effect of A. oryzae prebiotic on stress-induced gastrointestinal hyperpermeability. This supplement may be a useful dietary ingredient for horses undergoing combined transport and exercise stress as a prevention for gastrointestinal hyperpermeability. Future studies should explore the effects of A. oryzae prebiotic on the equine gastrointestinal microbiome as a potential mode of action.

animals : an open access journal from mdpi

10.3390/ani11041137

PMC8071557

Prion diseases are neurodegenerative disorders affecting humans and animals. The development of in vitro cellular models from naturally susceptible species like humans or ruminants can potentially make a great contribution to the study of many aspects of these diseases, including the ability of prions to infect and replicate in cells and therapeutics. Our study shows for the first time how ovine mesenchymal stem cells derived from bone marrow and their neural-like progeny are able to react to scrapie prion infection in vitro and assesses the effects of this infection on cell viability and proliferation. Finally, we observe that the differentiation of ovine mesenchymal stem cells into neuron-like cells makes them more permissive to prion infection.

Scrapie is a prion disease affecting sheep and goats and it is considered a prototype of transmissible spongiform encephalopathies (TSEs). Mesenchymal stem cells (MSCs) have been proposed as candidates for developing in vitro models of prion diseases. Murine MSCs are able to propagate prions after previous mouse-adaptation of prion strains and, although ovine MSCs express the cellular prion protein (PrPC), their susceptibility to prion infection has never been investigated. Here, we analyze the potential of ovine bone marrow-derived MSCs (oBM-MSCs), in growth and neurogenic conditions, to be infected by natural scrapie and propagate prion particles (PrPSc) in vitro, as well as the effect of this infection on cell viability and proliferation. Cultures were kept for 48–72 h in contact with homogenates of central nervous system (CNS) samples from scrapie or control sheep. In growth conditions, oBM-MSCs initially maintained detectable levels of PrPSc post-inoculation, as determined by Western blotting and ELISA. However, the PrPSc signal weakened and was lost over time. oBM-MSCs infected with scrapie displayed lower cell doubling and higher doubling times than those infected with control inocula. On the other hand, in neurogenic conditions, oBM-MSCs not only maintained detectable levels of PrPSc post-inoculation, as determined by ELISA, but this PrPSc signal also increased progressively over time. Finally, inoculation with CNS extracts seems to induce the proliferation of oBM-MSCs in both growth and neurogenic conditions. Our results suggest that oBM-MSCs respond to prion infection by decreasing their proliferation capacity and thus might not be permissive to prion replication, whereas ovine MSC-derived neuron-like cells seem to maintain and replicate PrPSc.

1. IntroductionTransmissible spongiform encephalopathies (TSEs) or prion diseases are fatal neurodegenerative disorders that affect humans and animals [1]. These diseases are caused by the conformational conversion of the cellular prion protein (PrPC) to an infectious isoform that is partially resistant to proteases and prone to forming aggregates called PrPSc [2]. The accumulation of this isoform in the central nervous system (CNS) causes spongiform neuronal degeneration, activation of glial cells and neuronal loss [3]. Scrapie, which affects sheep and goats, was the first reported TSE [4] and it is considered the prototype of these diseases [5].Cell culture systems are useful tools to study prion protein propagation in TSEs and to identify new prion therapeutics [6]. However, only a few cell lines can be infected and display PrPSc accumulation and/or infectious capacity [7]. In most cases, murine cell lines are used, requiring a previous mouse-adaptation of the prion strain to eliminate the problem of the species barrier [8].Mesenchymal stem cells (MSCs) are fibroblast-like cells characterized by their capacity for both self-renewal and differentiation in mesodermal tissues (osteoblasts, adipocytes, chondrocytes and myocytes) [9]. These cells can also transdifferentiate in vitro into neuron-like cells [10,11] and undifferentiated cells expressing PrPC [12], which seems to play a key role in the neuronal differentiation process of MSCs [13,14,15].Murine compact bone-derived MSCs (CB-MSCs) are able to migrate to brain extracts from prion-infected mice in vitro and significantly prolong the survival of mice infected with the Chandler prion strain when injected in vivo [16]. Furthermore, murine bone marrow-derived mesenchymal stem cells (BM-MSCs) can be infected with a Gerstmann–Sträussler–Scheinker strain adapted in mice ex vivo [17] and maintain the infectivity along passages. The susceptibility of these cells to prion infection makes them good candidates for use in developing in vitro models for prion research [18]. Therefore, the development of in vitro models from naturally prion-susceptible species like humans or ruminants, which would avoid the adaptation process, would be very useful for cutting-edge prion research. Although in recent studies, human cerebral organoids [19] and astrocytes [20], both derived from human induced pluripotent stem cells (iPSCs), have been described to maintain and propagate prion infectivity in vitro, in domestic species like sheep, the reprogramming of somatic cells to iPSCs might require adjustments of standard protocols.We have previously described the isolation of ovine MSCs from peripheral blood (oPB-MSCs), which express PrPC at the transcript level [21]. Our group also reported the presence of PrPC in ovine bone marrow-derived MSCs (oBM-MSCs) at both transcript and protein levels [18]. However, in contrast to BM-MSCs obtained from individuals with sporadic Creutzfeldt-Jakob disease (CJD), who are positive to PrPSc [12], the pathogenic prion protein was not detected in oBM-MSCs isolated from scrapie sheep [18]. In addition to the lack of PrPSc, these cells displayed diminished proliferation potential compared to oBM-MSCs derived from healthy sheep. To the best of our knowledge, the susceptibility of oBM-MSCs to scrapie infection in vitro and their potential to replicate prions have never been investigated. The aim of the present study was to assess the susceptibility of oBM-MSCs and their derivative neuron-like cells to scrapie prion infection in vitro, their potential to replicate prions and the effects of this infection on cell viability and proliferation.2. Materials and Methods2.1. Animals and Sample CollectionBone marrow samples were obtained from 11 adult female (n = 7) and male (n = 4) sheep, aged from 1 to 7 years and carrying different genotypes for the PRNP gene (Table 1). After animal sedation (Xylazine) and local anesthesia (Lidocaine), bone marrow aspirates were harvested from the humeral head using a 13 G Jamshidi needle and 10-mL syringes previously loaded with 0.5 mL of sodium heparin. All procedures were carried out under Project Licence PI06/12, approved by the Ethical Committee for Animal Experiments from the University of Zaragoza. The care and use of animals were performed in accordance with the Spanish Policy for Animal Protection, RD53/2013, which meets European Union Directive 2010/63 on the protection of animals used for experimental and other scientific purposes.The animals used in this study were maintained in an experimental flock in which the prevalence of scrapie was high. Although none of the animals displayed clinical signs compatible with scrapie, an in vivo test for PrPSc determination using third-eyelid biopsies was performed as previously described [22,23] to identify any scrapie-infected preclinical sheep. Two males were positive to scrapie but their cultures were maintained in the study to evaluate if these cultures could react differently to those obtained from negative sheep, although in previous studies infectivity was not detected in oBM-MSCs derived from scrapie sheep [18]. Negative animals were those that did not show scrapie compatible symptoms and were negative to PrPSc based on a third-eyelid biopsy.2.2. Ovine Mesenchymal Stem Cell Isolation and CultureMSC isolation from bone marrow aspirates (3–5 mL) was performed following the previously described protocol [18,21,24]. This protocol is based on the separation of the mononuclear fraction after density gradient centrifugation in Lymphoprep (Atom) and further isolation thanks to the ability of MSCs to adhere to plastic. After isolation, cells were expanded up to passage 3 in basal medium, consisting of low glucose Dulbecco’s modified Eagle’s medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% fetal bovine serum (FBS), 1% L-glutamine (Sigma-Aldrich) and 1% streptomycin/penicillin (Sigma-Aldrich).In addition to plastic-adherence in standard culture conditions, the minimal criteria to define MSCs are the expression of certain cell surface markers and the ability to differentiate into adipocytes, osteoblasts and chondroblasts in vitro [25]. The ability to differentiate to mesodermal lineages and the expression of mesenchymal and hematopoietic markers of these cultures have been evaluated previously [18]. After characterization, the expression of PrPC in oBM-MSCs was confirmed by RT-qPCR and dot blotting [18].2.3. Neurogenic DifferentiationTo study whether neural differentiation increased the susceptibility to prion infection, oBM-MSC cultures were seeded at 1500 cells/cm2 and differentiated into neuron-like cells using HyClone neurogenic medium (Thermo Scientific, Waltham, MA, USA) according to the manufacturer’s instructions. The differentiation process lasted three days. Neural differentiation was monitored and confirmed by observing the cultures through an inverted optical microscope. The formation of neuron-like cells was seen within 24 h, peaking at 72 h (Figure 1a,b), as previously described [18].2.4. Scrapie Inocula and Infection of oBM-MSC and Neuron-Like CulturesInocula were prepared using CNS samples from one healthy (negative controls) and one classical scrapie-infected sheep carrying the ARQ/ARQ genotype and preserved at the tissue bank of the Center of Encephalopathies and Emerging Transmissible Diseases (CEETE; University of Zaragoza). The presence/absence of PrPSc in the tissues was confirmed following protocols reported in other works [26], using two rapid diagnostic tests (Prionics-Check Western blotting; ThermoFisher Scientific and Idexx HerdChek; IDEXX, Westbrook, ME, USA) and confirmation by immunohistochemical examination of CNS tissue. CNS samples were homogenized and diluted 1:10 (g/mL) in physiological saline solution (Braun). Afterwards, samples were treated at 70 °C for 10 min before adding streptomycin sulphate (100 µg/mL) and benzylpenicillin (100 µg/mL). In order to check the safety of the inocula once generated, samples were incubated in blood agar plates, and the absence of any bacterial growth was confirmed.To determine the effect of prion infection on the proliferation potential and the ability of prion replication, oBM-MSCs cultures were seeded at 5000 cells/cm2 for proliferation conditions and at 1500 cells/cm2 for neurogenic conditions. In both cases, three groups were established: positive, negative and control cultures. Positive cultures were infected with inocula from a scrapie-infected sheep, negative cultures with inocula from a healthy sheep and control cultures were kept in standard conditions. After adhesion for 24 h, basal media was substituted by inocula diluted 1:10 in DMEM media (10% FBS, 1% L-glutamine and 1% streptomycin/penicillin) for the oBM-MSC cultures and in HyClone media for the oBM-MSC cultures in neurogenic differentiation. Cells were maintained in this medium for 48 h to analyze the proliferation potential and cell viability and for 72 h for the MTT/ELISA assays. Afterwards, the medium was changed twice a week.2.5. Proliferation Potential and Cell ViabilityTo determine the effect of prion infection on oBM-MSC proliferation potential, cultures from three different donors (biological replicates) were seeded in 6-well plates at 5000 cells/cm2, inoculated with scrapie and control inocula and maintained until passage 3 post-infection; every passage was performed at around 80% confluence. Adherent cells were counted every passage and the cell doubling number (CD) and cell doubling time (DT), used to determine the time it takes for a population of cells to double in size, were calculated as previously described [21,24]. The results were evaluated using the paired Student’s t-test.To assess early prion toxicity, cell viability was also evaluated using MTT in oBM-MSC from 8 donors at 3, 7 and 10 days post-inoculation (dpi), seeding 4 technical replicates for each culture. oBM-MSC cultures were seeded in 96-well plates at 5000 cells/cm2 in growth conditions and at 1500 cells/cm2 in neurogenic differentiation conditions. Briefly, the MTT assay was performed by adding 25 µL of MTT solution (2 mg/mL) per well. Then, the plates were incubated at 37 °C for 4 h. Afterwards, the content of each well was removed and was substituted with 150 µL of HCl solution (HCl 40 mM in isopropanol) per well. Plates were then incubated for 1 h at room temperature protected from light. The absorbance was measured at 570 nm in an Infinite F200 microplate reader (Tecan Ibérica Instrumentación, Barcelona, Spain). A calibration curve was prepared with different amounts of cells. Since oBM-MSCs in growth conditions are seeded in a higher density than the ones in neurogenic differentiation conditions, two calibration curves were prepared: a more concentrated one to compare oBM-MSCs in growth conditions (Figure S1a) and a more diluted one to compare cultures in neurogenic conditions (Figure S1b). In both cases, the calibration curve enabled us to establish the relationship between absorbance and the amount of cultured cells. The toxicity of the prion was studied in three conditions (inoculated with scrapie-positive brain homogenates, negative brain inoculum and non-inoculated controls) and at three different stages (3, 7 and 10 dpi). As cells were kept in contact with the inoculum for 72 h, the stage 3 dpi corresponds to the moment just after inoculum removal. The moment of the infection with the inocula was considered as day 0. The normality of the results was evaluated with Shapiro–Wilk and D’Agostino–Pearson tests. Differences in cell viability and proliferation were evaluated with Student’s t-test. Statistical significance was defined as p < 0.05.2.6. PrPSc DetectionCells from the three biological replicates analyzed in the proliferation assay were used to evaluate if PrPSc was increased or maintained along the passages in MSC cultures infected with scrapie and maintained under grown conditions. Approximately 106 cells of passages 1, 2 and 3 post-infection were frozen at −80 °C for further PrPSc determination by Western blotting. Pellets of frozen cells were homogenized in 100 µL of PBS. Afterwards, samples were analyzed using the BSE Scrapie Discriminatory Kit (Bio-Rad Laboratories, Hercules, CA, USA) and treated following the manufacturer’s recommendations. Electrophoresis was developed in 12% SDS-PAGE gels. Protein was then transferred to a 0.20-µm nitrocellulose membrane (Bio-Rad). CDP-Star substrate (ThermoFisher Scientific, Westbrook, ME, USA) was used to determine chemiluminescence in a Versa-Doc Imaging System (Bio-Rad Laboratories). Chemiluminescence signals were evaluated using ImageJ 1.4.3.67 (Psion Image), as described previously [27].Neurogenic differentiation of MSCs requires seeding cells at low density and differentiated cells cannot be maintained along passages. To test the ability of these cells to replicate PrPSc, we quantified the amount of the pathogenic protein soon after prion infection at three different stages (3 dpi, which corresponds to inoculum removal, 7 and 10 dpi) in oBM-MSCs in growth and neurogenic differentiation conditions. We used a more sensitive test, the ELISA kit EEB-Scrapie HerdCheck kit (IDEXX), following the manufacturer’s recommendations. oBM-MSCs cultures from 5 donors were seeded in 6-well plates and the retrieval of the cells was performed by means of trypsinization and subsequent centrifugation. To quantify the PrPSc detection range of the ELISA kit, a calibration curve was performed using different concentrations of scrapie inocula (Figure S2). PrPSc was detected in all inoculum concentrations, showing that this kit is suitable to detect PrPSc in oBM-MSC cultures, as the amount of inocula used in oBM-MSC infection was higher than the most diluted concentration of the calibration curve. The inoculum used in the calibration curve was the same used to infect oBM-MSCs in growth and neurogenic conditions. For infection, a volume of 100 µL of scrapie inoculum per well was employed, which would correspond to >3 units of absorbance.3. Results3.1. Proliferation Potential of Infected oBM-MSCThe effect of scrapie infection on the proliferation capacity was analyzed in oBM-MSCs, calculating the CD and DT. Significant differences between cultures infected with scrapie and control inocula were found for both CD and DT at the first passage post-infection. CD was higher and DT was lower in the cultures treated with control inocula compared to those inoculated with scrapie brain cells (Table 2).3.2. Cell Viability of Infected CulturesThe effect of prion infection on cell viability was studied in three conditions (scrapie positive inoculum, healthy/negative inoculum and control without inoculum) and at three different stages (3, 7 and 10 dpi) in oBM-MSCs in growth conditions and in neurogenic differentiation.Proliferation was evidenced in oBM-MSC cultures maintained in growth medium under the three conditions. Inoculated cultures displayed higher number of cells than controls at the three stages (3, 7 and 10 dpi). Proliferation was significantly lower in scrapie infected cells than in cultures treated with negative inocula at 3 dpi but, in subsequent stages (7 and 10 dpi), the positive cultures displayed significantly more cells than the negative cultures (Figure 2a).In neurogenic differentiation conditions, the number of cells also increased over time in the three conditions, whereas inoculated cultures showed a higher growth than controls. Comparing between the inoculated cultures, the number of cells was significantly higher in cultures that were in contact with negative inoculum than the ones infected with scrapie, and this difference was statistically significant at 10 dpi (Figure 2b).3.3. PrPSc Detection in Infected oBM-MSCs, Analyzed by Western BlottingAfter inoculation of oBM-MSCs in growth conditions, surviving cells retained their ability to proliferate and were expanded until passage 3 post-infection. Western blotting analysis revealed the presence of PrPSc in the cultures during these three passages although the intensity of bands decreased with the number of passages (Figure 3, Figure S3) and was lost in further subcultures.3.4. Prion Detection by ELISA in oBM-MSC and Neuron-Like Cultures Infected with ScrapieTo test the ability of MSC-derived neuron-like cells to replicate prions, the presence of PrPSc was studied by means of ELISA immediately after infection at three different stages (3, 7 and 10 dpi) in oBM-MSCs infected with positive inocula in growth and neurogenic differentiation conditions.In oBM-MSCs maintained under growth conditions, a decrease in ELISA absorbance was observed, which could be associated with a loss of the PrPSc signal. Significant differences were found between 3 dpi and 7 dpi (p < 0.05) and 3 dpi and 10 dpi (p < 0.01) (Figure 4a).In contrast, the majority (four out of five) of neuron-like differentiated cultures showed an increase in absorbance over time, which could be associated with a progressive increase in the PrPSc signal. Even though most of the cultures had an increasing signal pattern, as one of them (BMO4) displayed decreased absorbance over time, no significant differences were found between any of the three stages (Figure 4b).Although we used the same inoculum in all cultures, the initial amount of PrPSc was different in each culture at 3 dpi, suggesting a heterogeneity in the ability to retain prions. Higher cellular density under growth conditions would explain the higher absorbance observed in this condition, compared to neurogenic conditions.4. DiscussionPrion diseases are fatal neurodegenerative disorders affecting humans and animals. Over the years, a substantial effort has been made to develop in vitro models for the study of these pathologies. Most of the cellular models are based on the culturing of murine cell lines [8] and require a previous adaptation of the strain to mice, due to the well-known phenomenon of the species barrier. Therefore, in vitro models with a natural host background would be very useful tools for research into many prion topics, e.g., prion replication, toxicity, genetic susceptibility, differences in strain susceptibility, early mechanisms of infection and new treatment testing.MSCs can be easily collected from several accessible adult tissues like bone marrow or peripheral blood [8,28] and they show the ability to transdifferentiate into neuronal elements in vitro [10,29]. Several works have described the ability of murine stromal cells to propagate prion infectivity [12,17,30,31] and expanded MSCs obtained from sporadic Creutzfeldt-Jakob disease CJD patients have been shown to be positive for PrPSc [12]. The infectivity of MSCs obtained from sick individuals does not seem to be a pan-species characteristic, as oBM-MSCs from scrapie infected sheep did not show PrPSc [18]. Although MSCs derived from human, cattle and sheep express PrPC [12,18] to the best of our knowledge, the potential of MSCs derived from these naturally susceptible species to propagate prion infection in vitro has never been investigated. In the present work, we infected ovine bone marrow-derived MSCs and their neuron-like derivatives with scrapie-infected sheep isolates to study the response of these cells to prion infection during a certain period of time.MSCs can migrate to prion-affected neurological tissues as a response to secreting trophic factors that activate endogenous restorative reactions in the injured brain [32,33,34]. In our study, cell viability was higher in both oBM-MSCs and neural-differentiated cultures after inoculation compared with non-inoculated control cultures in the three monitored stages (3, 7 and 10 dpi), which suggests that brain inocula, independently of their origin, may contain factors that stimulate oBM-MSC proliferation.Murine stromal cells are able to propagate prions for many passages [17,30]. On the contrary, oBM-MSCs do not seem to be permissive to PrPSc infection. The loss of PrPSc signal over time detected by ELISA in oBM-MSC cultures soon after infection with positive inocula suggests that these cells, if infected, are unable to replicate the prions, unlike what happens in mice. Similarly, Western blotting revealed the presence of PrPSc in scrapie-infected oBM-MSC cultures three passages after inoculation, but the presence of the pathologic protein seemed also to be weakened between passages 2 and 3, suggesting that PrPSc may be taken up by oBM-MSCs without leading to a successful prion infection. In some works, murine BM-MSCs infected with prions in vitro showed few or no PrPSc production during the first 10 or even 50 passages [12,31] and stable and detectable (by Western blotting) production of PrPSc afterwards. We could not explore this possibility because, contrary to murine cells, MSCs obtained from humans or unconventional model organisms including sheep are able to be maintained in culture for far fewer passages [35,36].Moreover, after inoculation with scrapie, oBM-MSC cultures displayed a high proliferation rate, with an average doubling time during the three passages that was lower than the DT described previously for BM-MSC cultures derived from scrapie and healthy sheep [18]. Cell division modulates prion accumulation in cultured cells [37] and direct proximity between donor and recipient cells increases the infection in other cell culture models [38]. The high proliferation rate observed in oBM-MSCs could help to avoid the transmission of PrPSc from infected cells to non-infected ones because the cells are not in contact for a long enough time. Therefore, only the cells infected during the inoculation process and their daughters would show infection and this would be diluted in successive passages. Changes in culture conditions focused on slowing down the proliferation rate or increasing their contact in spheroid cultures could facilitate the propagation.On the other hand, we cannot discard the possibility that infection with scrapie-infected brain cells could be toxic for the oBM-MSC cultures. Even though no toxicity was observed in murine MSCs infected with the CJD agent [12], we have to take into account that our cells come from a naturally susceptible host. In our study, CD was significantly higher in cells infected with healthy brain extracts and, accordingly, DT was higher in cells infected with scrapie inocula. Therefore, those cells exposed to scrapie prions showed lower proliferation potential, similar to the findings observed in MSCs obtained from scrapie sheep [18]. This could be a consequence of the loss of infected cells due to prion toxicity. In the MTT assay, the effect of prion infection on cell viability was evaluated during the first passage after prion infection. Toxicity seems to be an early effect of prion infection, as 3 days after inoculation, viability was lower in scrapie cultures than in healthy infected cultures. In contrast with the first assay, at the end of this passage (10 dpi), the number of cells in scrapie-infected cultures was higher than that in the cultures inoculated with control brain cells. This could be a consequence of differences in the inocula, as brain tissues used in the different experiments were different and could have contained different amounts of PrPSc and therefore exhibited different degrees of toxicity. Nevertheless, throughout all passages, CD and DT differences were lower and, similarly, the number of cells in scrapie-infected cultures increased after early toxicity. In both cases, this increase in proliferation and viability was accompanied by the loss of PrPSc detection, which might indicate a recovery of the cell culture conditions after the elimination of PrPSc infected cells, increasing the proportion of non-infected cells, which display higher proliferation potential.Regarding the differentiation of oBM-MSC cultures into neuron-like cells, although certain toxicity was also observed 3 days after prion infection, four out of the five cultures analyzed seemed to be infected and possibly displayed the ability to replicate the pathological prion protein. Although we did not obtain statistical support for this observation, the ELISA assay showed that these cells maintained the PrPSc signal and this signal increased progressively over time. Similarly, astrocytes derived from human induced pluripotent stem cells are capable of replicating prions from brain samples of CJD patients, generating prion infectivity in vitro [20]. Taking this into account and knowing that cells from the central nervous system are the target of the pathological prion protein, oBM-MSC-derived neuron-like cells may have a greater ability to capture and replicate PrPSc than oBM-MSCs in growth conditions. The lack of statistical significance in our results was due to the existence of variability in prion replication, as one culture (BMO4) failed to replicate prions. This culture displayed a PRNP genotype identical to other three cultures (ARQ/ARQ), suggesting that, in addition to the PRNP genotype, other factors influence prion replication. Despite all the cultures being infected with the same amount of inoculum, BMO4 was the one that showed the highest absorbance for PrPSc under both growth and differentiation conditions. Differences in the ability to access prions could explain differences in toxicity and prion replication.The observed differences in undifferentiated and differentiated oBM-MSCs suggest that the latter possess a competence for infection that it is not present at the MSC stage, even though they share a genetic background for each given animal. This system, using oBM-MSC-neuron-like derivates, could serve for the investigation (in an isogenic context) of the molecular trigger that sustains scrapie infection in vitro, specifically in the neural lineage. In addition, differences between cultures harboring the same PRNP genotype could help in the identification of other factors related to prion susceptibility.5. ConclusionsThis work describes for the first time the infection with scrapie agents of bone marrow-derived MSCs obtained from sheep, which is a natural host of prion diseases. Culturing ovine MSCs with CNS extracts in growth and neurogenic conditions induced cell proliferation, although some toxicity was observed in scrapie-infected cultures. Inoculated oBM-MSCs in growth conditions were not permissive to prion infection, whereas most cultures under neurogenic differentiation conditions seemed to retain and replicate the pathological prion protein. oBM-MSC-derived neuron-like cells could be a good candidate for developing in vitro studies in species for which iPSC reprogramming is not standardized, like sheep. Further studies focusing on elucidating the molecular mechanisms implicated in retaining prion infectivity and inducing prion toxicity in mesenchymal stem cells and MSC-derived neuron-like cells are warranted.

animals : an open access journal from mdpi

10.3390/ani11082303

PMC8388372

Transportation is a stressful event for all animal species, but some species may be subjected to worse welfare consequences than others due to their ethological characteristics and specific coping strategies. Among equines, horses with a low level of tameness are at higher risk for transport-related disease and injury. For this reason, in Europe, regulations for the protection of animals during transport (Regulation EC 1/2005) are stricter for unbroken (untamed) vs. broken (tamed) horses. However, in practice, official veterinarians cannot verify regulatory compliance as there is no valid tool for the classification of horses as broken or unbroken. This study proposes the Broken/Unbroken Test (BUT) for assessment and scoring of horse behaviour during approach, haltering, and handling. After a validation process, our study has shown that the BUT is a reliable, valid, and feasible tool for determining whether a horse is broken or unbroken. The use of this tool would allow simple verification of compliance with Regulation 1/2005, and would help to ensure that transport procedures for unbroken horses are more respectful of their ethological and physiological characteristics. This may reduce the incidence of adverse welfare consequences for horses during transportation.

Regulation EC 1/2005 has stricter rules for transportation of unbroken (untamed) vs. broken (tamed) horses, but does not provide adequate tools for their identification. This study aimed to develop and validate such a tool. A behavioural test (Broken/Unbroken Test (BUT)) based on approaching, haltering, and leading was applied to 100 horses. Physiological and additional behavioural data were also collected, and the horses’ status (broken/unbroken) was assessed by the expert who administered the BUT. Each horse’s behaviour during the BUT was scored by four trained observers blinded to the horse’s history. The BUT score showed excellent inter-observer, intra-observer, and test–retest reliability (all intraclass correlation coefficients (ICCs) > 0.75). It was also negatively associated with respiratory rate, avoidance distance, and time needed to approach, halter, and lead the horse (p < 0.05 for all). The optimal BUT score cut-off for discrimination between broken and unbroken horses (gold standard: expert judgment) showed 97.8% sensitivity and 97.3% specificity. There was almost perfect agreement between BUT-based and expert classification of horses (ICC = 0.940). These findings confirm the BUT’s construct and criterion validity. The BUT could provide officials with a feasible, reliable, and valid tool to identify a horse’s broken/unbroken status and, consequently, direct stakeholders towards correct transport procedures.

1. IntroductionEvery year, millions of horses are transported over long distances by road, sea, and air [1,2]. Horses may be transported for various purposes and, unlike other farmed species, many times in their lives [3], and travel conditions and related welfare consequences differ depending on the situation. Although transport is a potential source of physical and psychological stress for all horses, the risks of mortality, disease, and injury are generally higher for low-value animals, such as meat horses, which are often transported in unsatisfactory conditions [4,5]. Moreover, several countries have no plants that slaughter horses for human consumption. As a result, each year, hundreds of thousands of horses are subjected to a gruelling, cross-border journey (>8 h of travel) that ends in slaughter [2,3,6,7,8]. Long journeys increase the risk of welfare issues and often lead to blurring of information related to transport conditions [8].The effects of transportation on the welfare of horses include anxiety-related behaviours, aggression, exhaustion, injury, respiratory and gastrointestinal disease, dehydration, pyrexia, and immunosuppression [3,9,10]. About 1% of horses die en route [4] but a greater percentage of animals are euthanised later due to severe injuries sustained during the journey or have non-visible injuries such as bruising, which is only recognisable post-mortem [11]. Behavioural and physiological responses, as well as injury rates, are affected by management factors such as vehicle specification, journey duration, and driver experience [12,13,14], and also by the physical fitness of the horse, its temperament, and its coping strategies [3,5,10,15,16]. For example, Grandin et al. recommended that aggressive horses should be segregated during transport because fighting was documented as a major cause of injury [17]. Fazio et al. evaluated the stress responses of stallions with different temperaments and found that nervous stallions had poor capacity to adapt to transport, probably as a result of adrenocortical depletion [18]. Conversely, changes in physiological parameters and problem behaviours decrease with repeated transport, suggesting that transport-induced stress responses are reduced in horses habituated to the situation [19,20].The influence of prior handling on the degree of transport stress experienced by horses has not been extensively studied. It is known, however, that abilities and coping strategies differ for broken (tamed) and unbroken (untamed) horses [21,22,23,24]. The response of unbroken horses to challenging situations is characterised by high arousal, fear, excitability, and a negative emotional state, increasing their risk of distress and travel-related pathologies. Knowles et al. [5] confirmed that transport management of unbroken ponies should take into account their ethology and physiological response to stressors. For example, unlike handled horses, it is preferable for untamed ponies to travel as a group [5]. It is likely that their strong herding instinct underlies this difference [5]. The same authors realised the importance of having a tool that would predict individual horse’s responses to transport stress before embarking on the journey. These authors failed to find a strong relationship between pre-transport behaviours and aggressive behaviour during transport. However, rating the reactions of untrained horses during a novel object or handling test may be a better predictor of their temperament, emotional responses, and coping strategies [23].The European Union recognises that broken and unbroken horses have different needs and adaptability skills. For this reason, Regulation EC 1/2005 on the protection of animals during transport includes stricter rules for the transportation of unbroken animals: they must not be transported on journeys over eight hours, tied during transport, or transported in individual bays, but must instead travel in groups of ≤4. An unbroken horse is defined in the Regulation as a horse that “cannot be tied or led by a halter without causing avoidable excitement, pain or suffering”. However, this definition is not accompanied by verification procedures; therefore, in practice, there is still no test to identify whether a horse is broken or unbroken. Thus, even when violations are identified during on-road inspections, nobody can be fined because official veterinarians do not have a test to identify whether a horse is broken or not. The European Parliament has expressed serious concerns about horse welfare during transportation and admits that there is still a high level of regulatory noncompliance, mainly related to unbroken horses [2,4]. It is therefore essential to provide official veterinarians with a tool that allows them to categorise horses and, consequently, direct transporters towards the correct transport procedures. Provision of a reliable test and the resultant reduction in the number of horses that travel under inappropriate conditions would avoid many injuries and substantial suffering.The objective of this study was to develop and validate a behavioural test to identify whether a horse is broken or unbroken. The study was based on the hypothesis that horses show different behavioural and physiological responses to being approached, haltered, and led, depending on prior level of tameness.2. Materials and Methods2.1. Sample Size CalculationThe sample size calculation was based on intraclass correlation coefficient (ICC) procedures, as previously reported [25]. The sample size was calculated using the confidence interval (CI) with the desired precision approach proposed by Bonnet [26] with a 95% CI of width Wk = 0.15, assuming a planning value of ICC (ICCplan) = 0.7 and k = 4 raters and an expected drop-out rate of 3–5%. The minimum sample size required was 100 horses.2.2. Farms and AnimalsExperiments were carried out during March and April 2021 on nine farms in Italy, with an average temperature and humidity of 18 °C (range: 9–26 °C) and 55% (range: 30–68%), respectively. The altitude of the farms ranged from 4 masl to 522 masl. The total number of horses reared on each farm ranged from 3–25 (mean, 11) and the paddocks where the horses were kept ranged from 15–10,000 m2. Only 11 horses were kept in individual housing (usually stallions) while the majority (78%) were kept in groups of ≥3 (maximum group size, 21). More than 50% of the horses had >133 m2 of space allowance (Table S1). In two paddocks, a stallion was kept with mares. Ten horses were confined in indoor stalls (stallions or mares before or after foaling) while the others were housed in outdoor paddocks with shelter. No horse was tied. Water and hay were provided ad libitum on all farms.Italian Heavy Draft horses (n = 103) were selected for this study. Demographic data (i.e., age and sex) of the horses were provided by the owner (Table S2). Horses were born on the farm where they were assessed or had been kept there for at least three months prior to assessment. They were reared for various purposes (breeding, meat production, family enjoyment, and showing). Three horses were excluded from the study before the start of the test due to lameness. Thus, the final group consisted of 100 healthy horses (16 males and 84 females, including 14 broodmares with foals), without behavioural problems (i.e., aggressivity and stereotypy), with a mean age of 7 years (range 1–25 years). A subset of these horses (n = 65; 6 males, 59 females; age (mean/range), 7 years/1–25 years) was evaluated twice, with a 3-week interval between evaluations. Before the start of the testing procedures, one author (LM) asked each owner whether their horses were broken or unbroken, based on how much handling and training they had undergone. According to the owners’ opinions, 55 of the 100 horses (55%) were broken (6 males, 49 females; age (mean/range), 8 years/1–25 years) and 45 (45%) were unbroken (10 males, 35 females; age (mean/range), 5 years/1–18 years).2.3. Testing ProcedureEach horse underwent a Broken/Unbroken Test (BUT), which consisted of two phases (Approaching and Haltering Test (AHT); Handling Test (HT)) to evaluate their behaviour during approaching, haltering, and leading. The BUT was always performed by the same tester (BP), a veterinarian with more than 20 years of experience in horse behaviour, handling, and learning theory, and who was trained for the procedure. The tester was unfamiliar to the horses and blinded to their owner-assessed level of handling. Each horse was tested in the paddock/pen where it was usually kept (test area). Horses were usually kept without a halter on. For biosecurity reasons, halters were provided by each farmer. Thus, two different types of halter were used; all were adjustable at the headpiece, but 17 had a clip at the throatlash and 83 did not.The BUT was a modified version of the tolerance test proposed by Wulf et al. [27]. Briefly, the tester entered the test area and walked towards the horse slowly with the halter in her hand, approaching and then trying to halter the horse (AHT; Figure 1a–d). The maximum time allowed for approaching and haltering the horse was 5 min. If it was not possible to approach and halter the horse within this time, the test ended. If it was possible to halter the horse within the maximum time, the tester started the second phase of the test (HT). During HT, the tester tried to lead the horse three steps forward (about 2 m distance) and three steps backwards (Figure 1e,f). When leading, the handler kept the horse relatively close to her right-hand side. The tester used a negative reinforcement procedure, applying light pressure on the lead rope and releasing the pressure as soon as the horse showed the desired behaviour (e.g., a step forward). The maximum time allowed for HT was 5 min. Thus, the maximum total time for BUT was 10 min. However, the procedure was stopped at the first sign of pain or if the horse showed a high level of distress (shaking movements of the head and tail, rearing, or moving away vigorously) or signs of aggression (ears laid backwards, attempting to bite or kick). A collaborator with experience in horse behaviour and handling used a stopwatch (Onstart 310, Kalenji, Decathlon, Villeneuve d’Ascq cedex, France) to record the time required for completion of each phase of the test and alerted the tester if the maximum time had elapsed or if the horse was displaying the aforementioned behaviour consistent with pain, distress, or aggression. All tests were filmed using a digital video camera recorder (HDR-CX115E, Sony, China) for later analysis. To evaluate test–retest reliability, the BUT was repeated 3 weeks later, by the same tester and under the same experimental conditions, on a subset of the original group of animals.At the end of the BUT, the tester judged whether the horse was broken or unbroken based on her own experience and knowledge. This ‘expert judgment’ was based on the horse’s ability to respond to pressure, allowing the handler (tester/expert) to control the horse’s head and front and back legs in accordance with learning theory [28]. The expert’s judgment as to whether the horse was broken or unbroken was treated as the ‘gold standard’ in the analyses.2.4. Observers and TrainingAll BUT videos were renamed and coded using casual numbers by an author (LM) not involved in scoring, and then analysed by four condition-blind (e.g., broken or unbroken, ID of horses and farmer) observers. One observer (main observer, EDC) was a senior veterinarian with long experience in evaluating horse welfare and behaviour, while the last three were veterinary or animal science students (non-expert observers). The latter were all familiar with horses but had no experience in formal behavioural observations. The non-expert observers were trained by the main observer on the behavioural scores to apply (Table 1) using 30 videos. The observers were trained on the fact that if a phase of the test (AHT or HT) was not completed within the maximum time or was stopped due to pain and/or distress, a score of 0 (worst situation) must be attributed to that phase. Moreover, when AHT received 0, HT must receive 0 as well. The training ended after approximately 15 days when an excellent level of inter-observer agreement for each score (kappa > 0.75%; [29]) was obtained.2.5. Test Scoring and Classification of Horses According to Regulation EC 1/2005Each of the four observers assessed each of the AHT and HT videos recorded during the data collection phase of this study, using the criteria in Table 1. These assessments were made independently. The behavioural score for each phase of the BUT was based on a three-point scale (0–2). The AHT and HT scores were then summed to obtain an overall BUT score (0–4).After watching the videos, the observers also classified the horses as broken or unbroken according to the definition provided by Regulation EC 1/2005: “An unbroken horse is a horse that cannot be tied or led by a halter without causing avoidable excitement, pain or suffering” [30].The observers scored 100 videos showing the first BUT and 65 videos showing the repeated BUTs on a subset of the same horses, to evaluate test–retest reliability. Of these 165 videos, 20 were randomly selected, renamed, and analysed for a second time to estimate intra-observer agreement. Therefore, overall, the observers scored a total of 185 videos.2.6. Physiological and Behavioural Parameters for ValidationData relating to several physiological and behavioural parameters were collected during the BUT procedures for validation of the behavioural scores. During the approach phase of the AHT, the distance between tester and horse, at the moment when the horse showed any avoidance behaviour (e.g., moving away, turning the head away), was measured, as reported by Dalla Costa [31], using a laser distance meter (LV5800-50M, LOMVUM, China). This was defined as the ‘avoidance distance’ (Figure 1a,b). If the tester touched the horse, the avoidance distance was considered as 0 m. When possible, the tester also evaluated the horse’s heart rate (HR) (facial artery) and respiratory rate (RR) (flank movement) [32] at the end of the BUT. These parameters required touching the animal and could not be evaluated in all horses. In particular, HR data were missing for most of the horses, which were judged by the expert as unbroken (41/49 and 25/26 of horses during the first and second tests, respectively).When possible, three infrared thermography images of the horse’s head were taken at the end of the BUT, by the same person (LM), using a portable camera (FLIR E76 24°; FLIR Systems AB, Danderyd, Sweden), from the right side of the horse (Figure 2). The camera was calibrated using the environmental temperature and relative humidity recorded on that day by a weather station (Kestrel 4000, USA). Illuminance was recorded using a lux meter (BTMETER 881E, Aquarium Photography, China). The resolution of the camera was 320 × 240 pixels, and the accuracy was ±2°C or ±2% at environmental temperatures ranging from 15 to 35 °C. The camera was positioned at 90° to the sagittal plane and, in order not to scare the less well-handled horses, at a distance of approximately 3 m from the horse’s side. The exact distance from the horse was recorded using the camera’s distance laser meter. The images were analysed using the FLIR Tools® software (FLIR Systems, Inc., Täby, Sweden). In particular, an oval area was traced around the eye, including the eyeball and ∼1 cm around the outside of the eyelid, and the maximum temperature (°C) in this region was recorded, as previously reported [33,34,35]. Eye Temperature (ET) was defined as the mean of the three maximum temperatures (one per image). Similar to HR values, ET could not be calculated for all horses because many of them drifted away during the test and could not be re-approached within 3 m. This parameter was calculated for 91/100 and 60/65 horses for the first and second tests, respectively.In addition to being scored by each of the four observers, the BUT videos were assessed by one author (LM) for the following timings using a stopwatch (Onstart 310, Kalenji, Decathlon, Villeneuve d’Ascq cedex, France):Approach time: This was defined as the time taken for the tester to move from a distance of about 3 m from the horse to the point that she was able to gently touch the horse’s shoulder. Usually, the tester gently and slowly raised her hand to signal the start of the procedure (and therefore the start of the approach time).Haltering time: This was defined as the time that elapsed between the end of the approach time (i.e., the tester touching the horse’s shoulder) and the halter being correctly positioned. This latter moment was generally apparent in the recording as the sound generated by clipping the halter, or by the tester saying ‘OK’ to signal the end of the haltering procedure.Handling time: This was defined as the time required for the tester to lead the horse forwards and backwards. The time was measured from the end of haltering (i.e., correct positioning of the halter) to the end of the handling test (i.e., the horse having completed three steps forwards and three steps backwards). Usually, the tester saying ‘OK’ to signal the end of the test.Total time: This was calculated by summing the approach, haltering, and handling times.If, during the AHT or HT, the horse showed pain and/or a high level of distress (e.g., moving away vigorously, intending to kick), the test was terminated, and the maximum time (5 min) was automatically assigned to the test.2.7. Analytic ApproachStatistical analyses were performed with SPSS Statistics version 25 (IBM, SPSS Inc., Chicago, IL, USA). The level for statistical significance was set at p < 0.05. Statistical methods for validation of the BUT scale are summarised in Table 2, in agreement with the literature [36,37,38].Initially, descriptive statistics were performed, and values were expressed as numbers and percentages and as mean ± standard error.Validation of the BUT score started with the assessment of inter-observer, intra-observer, and test–retest reliability, as well as internal consistency [36]. Inter-observer reliability for the AHT and HT scores was estimated using the Fleiss’ kappa (Fk) [40], as this test allows evaluation of the agreement between >2 observers and provides an estimate of the agreement for each score (0, 1, and 2) of the scale. The 95% CI for Fk was also calculated. Fk, was also used to evaluate inter-observer agreement for horse classification (broken/unbroken) based on the definition written in the EU legislation. The BUT score was treated as a continuous variable and analysed using the consistency type ICC using a two-way mixed model approach [38]. Values < 0.4 indicate poor reliability, those between 0.40 and 0.75 represent fair-to-good reliability, and values > 0.75 indicate excellent reliability [41].Intra-observer and test–retest reliability for the AHT and HT scores and the classification based on the Regulation’s definition were evaluated using concordance rate and Kendall tau-b correlation coefficient (τ), a measure of association based on the number of concordances and discordances in paired observations. The τ could range from 0 (no concordance) to 1 (perfect concordance). Associations were considered weak if τ < 0.30, moderate if 0.30 ≤ τ ≤ 0.50, and strong if τ > 0.50 [42]. The ICC using the two-way mixed model approach was used to evaluate intra-observer and test–retest reliability for the BUT score [43].Internal consistency was estimated by the inter-test correlation (i.e., inter-item correlation) evaluated using Spearman’s rank-order coefficient (ρ) as Cronbach’s coefficient alpha is inappropriate and meaningless for two-item scales [39]. Correlations of each item and BUT score were also calculated (i.e., item-BUT score correlation). On a reliable scale, each item should have a ρ > 0.30 [44].The second phase of validation focused on validity. Validity was evaluated by using the BUT score of the main observer while the expert’s judgment (broken/unbroken) was treated as a criterion measure.Construct validity was determined using the hypothesis that there would be a negative correlation between BUT score and ET, HR, RR, avoidance distance, as well as approach, haltering, and handling times. In other words, we expected that these parameters would decrease as the level of previous tameness increased, and, consequently, as BUT score increased. These associations were evaluated using Spearman’s coefficient (ρ). This was only performed using data from the first test session, to avoid possible bias due to repeated measures. The correlation was considered poor if ρ < |0.3|, medium if |0.3| ≤ ρ < |0.5|, and large if ρ ≥ |0.5| [44]. The association between BUT score and physiological and behavioural parameters was also analysed using ordinal logistic regression. Generalised Linear Models (GLMs) procedures with a multinomial probability distribution and cumulative logit link function were used, with BUT score as the dependent variable and physiological and behavioural parameters as predictors. Moreover, since ET could be affected by many intrinsic and exogenous factors, especially when measured under field conditions [34], this was further investigated using GLMs with a normal probability distribution and identity as link function. ET was included as a dependent variable while the following multiple fixed effects were evaluated: BUT score, age, sex, exact distance between camera operator and horse, environmental temperature, and lux. In these GLMs, the horse was included as the subject and the session (first or second) as the within-subject effect. Results were expressed as odds ratio (OR), 95% CI, and the P-value of the Wald statistic.Criterion validity indicates the accuracy in predicting scores on a ‘gold standard’ criterion measure (expert’s judgment) [36]. The sensitivity of the BUT score as a predictor of horses’ tameness level was evaluated by binary logistic regression, using GLM procedures, and expressed as OR, 95% CI, and P-value. The hypothesis was that the BUT score would increase in line with the tameness level of the horse (i.e., the odds that the horse was broken would increase as the BUT score increased). The expert’s judgment (broken/unbroken) was included as the dependent variable and the main observer’s BUT score as the predictor. The horse was also included in the model as the subject and the session (first or second) as the within-subject effect.Receiver operating characteristic (ROC) analysis was used to estimate the ability of the BUT score to discriminate between broken and unbroken horses and to determine the optimal threshold value (cut-off) for discriminating between the two groups. The condition of ‘broken’, as defined by the expert, was set as the positive actual state and larger values of the BUT score indicated stronger evidence for a positive actual state. Based on the area under the curve (AUC), the BUT score may be considered to be uninformative (AUC = 0.50), poorly accurate (0.50 ≤ AUC ≤ 0.70), moderately accurate (0.70 ≤ AUC ≤ 0.90), very accurate (0.90 ≤ AUC < 1) or perfect (AUC = 1). The optimal cut-off was determined using Youden’s index [45,46].Finally, we evaluated agreement in horse classification (broken/unbroken) between the expert’s judgment, the owner’s opinion, the classification according to the Regulation as assessed by the observers after viewing the videos, and the BUT score according to its optimal cut-off. In order to define the agreement between each of these classifications, a correlation matrix was built reporting the Cohen’s kappa (Ck) values for each pair of comparisons [47]. Only the first session was included in this analysis. As for the reliability analyses, values below 0.4 were considered as poor agreement, values above 0.75 as excellent agreement, and values between 0.4 and 0.75 as fair-to-good agreement [41].3. Results3.1. Descriptive StatisticsDescriptive statistics of AHT and HT scores, BUT scores, and classification of horses according to the Regulation EC 1/2005 definition by the four blinded observers are shown in Table S3. Figure 3 shows the relative distribution of the main observer’s BUT score for horses judged by the expert as being broken vs. unbroken. None of the horses classified as broken had a BUT score of 0, while > 40% had a BUT score of 3. Among the horses classified as unbroken, none had a BUT score ≥ 3, while 85.3% had a BUT score of 0.3.2. Reliability of Test Scores, BUT Score, and Classification of Horses According to Regulation EC 1/2005Table 3 shows the results of the inter-observer agreement analyses for the AHT and HT scores. Inter-observer reliability of test scores ranged from 0.652 (score 1 of AHT) to 0.961 (score 0 of HT). The overall inter-observer reliability was, however, excellent for both tests (Fk > 0.750, p < 0.001).Table 4 shows indices for intra-observer and test–retest reliability for the AHT and HT. Concordance rates for intra-observer reliability exceeded 90%, and correlations were very high (τ > 0.8) for both tests. Concordance between the first and second sessions was approximately 75%, but all concordances were ‘strong’ according to the τ values (τ > 0.5).Table 5 presents the reliability indices for the BUT score. ICC ranged from 0.792 for test–retest agreement to 0.916 for inter-observer agreement, with all indices indicating excellent reliability (ICC > 0.750).A strong correlation was found between the two AHT and HT scores (ρ = 0.825; p < 0.01) as well as between each of these test scores and the BUT score (Approaching and Haltering-BUT: ρ = 0.948; Handling-BUT: ρ = 0.951; p < 0.01). These findings support the internal consistency and the homogeneity of the BUT score.Finally, the reliability indices for classification of broken/unbroken according to the Regulation showed poor inter-observer agreement (Fk = 0.536, 95%CI = 0.477–0.596; p < 0.001), excellent intra-observer agreement (concordance rate = 93.8%; τ = 0.864; p < 0.001), and good test–retest reliability (concordance rate = 79.6%; τ = 0.592; p < 0.001).3.3. Validity of Total ScoreDescriptive statistics of physiological and behavioural parameters are shown in Table 6, while Table 7 shows their correlations with each other and with BUT score. BUT score was negatively correlated with most of the physiological and behavioural parameters investigated. In particular, BUT score increased as RR, avoidance distance, and all recorded times reduced (for all; p < 0.01). In contrast, no correlation between BUT score and HR or ET was found. However, GLM revealed that ET was influenced by exogenous factors: it increased as environmental temperature increased (OR = 1.096, 95%CI = 1.025–1.099; p = 0.001) and as distance from the horse decreased (OR = 0.632, 95%CI = 0.521–0.766; p < 0.001; Table S4).Ordinal logistic regression also showed that reductions in RR and all the recorded times were associated with an increase in the odds of higher BUT scores (OR < 1.0; p < 0.05; Figure 4 and Table S5). The narrow confidence intervals of these ORs indicate a high level of precision of the estimated effect.Criterion validity was evaluated using binary logistic regression to determine the nature of the relationship between BUT score and the ‘gold standard’ criterion. This analysis showed that, for every one-point increase in BUT score, the likelihood that the horse had been classified by the expert as broken increased more than 90-fold (OR = 94.719, 95% CI = 19.748–454.308; p < 0.001).Moreover, ROC analysis showed that the BUT score was a very accurate method to discriminate between broken and unbroken horses (AUC = 0.993, 95%CI = 0.984–1.000; p < 0.001; Figure S1). The optimal BUT score cut-off value to discriminate between the two groups was 2 (i.e., ≥2 indicates ‘unbroken‘, <2 indicates ‘broken’). This cut-off has a sensitivity of 97.8% and a specificity of 97.3%. Figure 5 summarises the procedure for assessing whether a horse could be identified as broken or unbroken according to the optimal cut-off.The agreement between horse classification based on the optimal BUT score cut-off (broken if BUT score ≥ 2) and expert judgment is shown in Table 8. Based on this cut-off, 50/100 horses were classified as unbroken, and the classification of only 3 horses was not in agreement with the expert’s judgement. The BUT score showed, thus, the highest degree of accuracy. Contrariwise, the lowest agreements were found for the Regulation definition.4. DiscussionThis study describes the development and validation of a test (BUT) that—for the first time—allows the classification of horses as broken or unbroken. Our results confirmed our hypothesis that horses with different levels of prior handling would react differently to being approached, haltered, and handled. The BUT is based on scoring the horse’s behaviour when it is approached, haltered, and handled in a standardised way. Each horse receives a score that ranges from 0 to 4, where 0 indicates the worst situation while 4 indicates the best situation. Thus, low BUT scores were assigned to horses that showed nervousness and avoidance behaviours, as well as those that could not be approached, haltered, and led, whereas high BUT scores were assigned to horses that exhibited fewer avoidance behaviours and could be approached, haltered, and led easily. We established a threshold that allows classification of a horse as broken (BUT score ≥ 2) or unbroken (BUT score < 2). This simple test could fill a legislative gap as, although Regulation EC 1/2005 includes different rules for transport of broken and unbroken horses, no tool for the classification of horses has previously been available. Due to their greater reactivity, unbroken animals are at increased risk of injury and disease. However, if the BUT was included in future transportation regulations worldwide, it would ensure that the correct transport procedures were followed for these animals and would help officials to verify regulatory compliance. Regular application of the BUT before a journey to a sub-group (randomly selected) of the horses in departure could therefore safeguard the horses’ welfare.The current study followed the rigorous validation process that is required to confirm the reliability and validity of behavioural rating scales [36,37,38]. Agreement analyses showed that the AHT, the HT, and the BUT all have excellent inter-observer and intra-observer reliability. The highest agreement indices were obtained for score 0, while the lowest were obtained for score 1. These findings are not surprising because a score of 0 indicates not only high arousal levels (e.g., aggressive responses, fear, excitation) but also test failure (animal not haltered or led), a situation that is well-defined and unambiguous. Conversely, a score of 1 indicates an intermediate arousal level (moderate reluctance and one or more avoidance behaviours) where subjective judgment could have a greater influence. This is the first test developed to assess the horses’ level of prior tameness, and there is no literature with which to compare the results directly. However, similar values of inter-observer agreement have been reported for tests evaluating the human-animal relationship [31] and for some pain scales [25,48]. Czycholl et al. [49] have recently evaluated inter-observer reliability of the indicators proposed by the AWIN protocol for horses, including behavioural tests scored on a 3- or 2-point scale. The authors reported acceptable-to-good agreement for all the indicators but highlighted that behavioural responses such as fear and avoidance, as well as approach tests, may show low reliability in horses because—similar to other species [50,51]—they may show incongruent behaviours (e.g., the simultaneous existence of curiosity and fear could lead them to approach, then run away, and then return). However, although defining an exact level of arousal may be problematic, our experience shows that rating horses’ behavioural responses with the BUT is a reliable and easy way to judge the level of prior handling and tameness of unfamiliar horses.Test–retest reliability data were obtained by repeating the BUT on a subset of horses three weeks after the first session. Although test–retest reliability was lower than inter- and intra-observer reliability, it was good for the AHT and the HT, and excellent for the BUT. This result was expected as it is commonly accepted that test–retest could be influenced by many factors including changes in test conditions, physical or mental state of the animal, and its learning experience [36,52,53]. In the present study, the main sources of variability affecting test–retest agreement were likely to be intra-observer variation and changes in the horses’ behavioural responses. Even in a well-defined test situation, test–retest reliability is very sensitive to the animal’s affective state and mood on the day. For example, a positive emotional state could imply a lower latency time in approaching humans and less aggressive and fearful behaviours [52,54], leading to a higher BUT score. Conversely, a negative emotional state leads to fearful and cautious reactions [52], which could result in a lower BUT score. In the repeated BUTs, although many days had passed between the two tests, a positive or negative emotional response could also be linked to the horse’s memory of its previous BUT experience [50,53]. Pain is another factor that could influence the horse’s response to human approach [31,55]. To limit this bias, only horses that appeared to be healthy on visual clinical evaluation were used. Moreover, although tester and test area were the same in the two repetitions and no major changes had occurred within the farms, identical test conditions could not be guaranteed. For example, changes in social dynamics of the herd or the farmer’s handling between the two sessions could have affected the horses’ behaviour and, therefore, the test scores. Most horses were tested in the presence of other animals, and this could also confound the results. Unfortunately, none of these factors could be controlled in the present study, and this may explain the relatively low test–retest reliability. In spite of these caveats, the agreement indices indicate that the horses’ responses to the BUT were consistent over time. We suggest, however, that assessors in the field should take into account the environmental and psychological context in which the test is conducted during their scoring, as the stability of the BUT across different situations has not been confirmed. We also suggest that the tester wear protective equipment and always stop the test at the first signs of distress or aggressive behaviour.Our BUT demonstrated both construct and criterion validity. The correlation between BUT and the recorded physiological and behavioural measures, which are known to be related to the level of taming [21,22,23], confirmed its construct validity [36,37]. Some authors have indeed claimed that the different reactions to stimuli of “naïve” horses compared to trained ones are related to activation of different areas of the brain [22]. In particular, broken horses do not usually show negative reactions when exposed to humans or novel environments, whereas unbroken horses typically show a high level of emotion and display different behavioural (aggression, fear, vocalisation, defaecation, and so on) and physiological (changes in heart rate, blood pressure, hormones, respiration rate, and so on) stress related responses [21,24,56,57]. Several authors [23,58,59] have shown that, compared with unbroken horses, broken horses have a lower increase in heart rate, approach the tester sooner, and are caught more quickly than unbroken horses during novel object and handling tests. It has also been shown that a reduction in emotional reactivity and improvement in the human–horse relationship continue as the number of training and handling sessions increases [60,61]. Our correlation analyses have confirmed that a high BUT score indicates a broken horse, as this was associated with lower respiratory rates, avoidance distances, as well as the time taken for approach, haltering and handling. Conversely, horses achieving a low BUT score could be defined as unbroken because they showed high RR, avoidance distance, and test times. However, there was no correlation between BUT score and HR or eye temperature. HR has been used previously to assess the personality and reactivity of horses [23,60], including those that are unbroken [58,59]. The inconsistency between our results and those of previous studies is likely to be due to the amount of missing HR data in our study. Its measurement required close contact with the animal and therefore could not be collected for many horses, particularly those that were unbroken, as the test was stopped if the horse showed signs of distress (e.g., flight response). Moreover, HR in horses increases during physical exercise, and, after the BUT, it could have been difficult to distinguish between emotional and physical reasons for an increase in HR. Eye temperature, on the other hand, has been used as an indicator of arousal for horses [33,35,60], but this can be influenced by many factors, especially when measured in the field [34]. We tried to standardise some of these factors; however, this was not always possible under field conditions. Indeed, ET was still strongly influenced by environmental temperature and measurement distance, and these could represent confounding factors that mask its association with the animal’s level of reactivity. In the context of the present study, parameters that do not require close contact with the animal seem not only feasible but also fit for purpose. Higher respiratory rates, the presence of avoidance behaviour, and longer approach times for unbroken horses indicate activation of the sympathetic nervous system and a “fight-or-flight” response. They also suggest that these horses perceive approach by a human as a dangerous situation that triggers a negative emotional state and a high degree of arousal [24]. Fearful and stressed horses are more likely to develop transport related respiratory disease after 8 h journey [62], so unbroken horses may be at higher risk when travelling over this lenght. The criterion validity was investigated by choosing the expert’s judgment as the ‘gold standard’ criterion measure against which to compare the BUT score. The expert evaluated the horse in the test area and defined it as broken or unbroken based on its ability to respond to pressure, used as negative reinforcement [28]. Our findings confirm the responsiveness and predictive value of the BUT on the basis that the likelihood of the horse being broken increased substantially with each additional BUT score point. The responsiveness of the BUT score also suggests that it could be used to indicate different levels of prior taming. For example, a BUT score of 0 could be defined as “no taming,” while scores of 2 and 4 could be defined as “moderate level” and “good level” of taming, respectively.The utility of the BUT as a tool for widespread use is that it can discriminate, with high sensitivity and specificity, between a broken horse and an unbroken one. Specifically, a horse would be defined as broken if its BUT score was ≥2 and as unbroken if it was <2. Our statistical approach thus confirmed the criterion validity of the BUT and suggested a rigorous procedure for applying it. After adequate training, official veterinarians would be able to score a horse’s behaviour objectively using the BUT, decide whether the horse is broken or unbroken, and advise on the transport procedures that should be put in place. In addition to BUT’s binary classification (broken vs. unbroken), which should direct personnel towards using specific transport procedures, the BUT score may indicate the horse’s level of taming. This information could accompany the animal throughout its transport and could also be relevant to human safety, because the fearful and aggressive reactions that characterise a low level of taming have been identified as the major cause of horse-related accidents [21].The results achieved so far proved that the BUT could be a reliable and valid tool. In contrast, when the observers were asked to classify horses using the definition proposed in Regulation EC 1/2005, all the agreement analyses showed that this classification system had poor reliability. It follows from this that the definition of unbroken horses, as written in the current legislation, is unclear. This could have led to confusion and consequently to the transport of unbroken horses over long distances and in inappropriate transport conditions [2,3]. This highlights the need to include, within the ongoing revision of the current legislation, a better definition of unbroken horses. However, we would also like to question the terminology that is currently used to define horses’ prior level of handling and training (i.e., taming). The term ‘unbroken’ was used in this study because it is the term used in Regulation EC 1/2005 to describe untamed horses. However, the converse state (‘broken’) implies that the animal has been ‘defeated’, ‘beaten’, ‘overpowered’, or ‘vanquished’, a terminology that is outdated at a time when humane animal handling and training procedures are prevailing worldwide. The term ‘broken’ is also used to indicate that the horse has been trained for riding or driving, something that is irrelevant in the context of this legislation. We, therefore, suggest that the term ‘unbroken’ should be replaced with ‘unhandled or untamed’ in the updated version of Regulation EC 1/2005.Our findings need to be interpreted with caution because this study has several limitations. The BUT was applied to a draught horse breed, and all horses were tested in their paddock. Consequently, our findings need to be confirmed by applying the BUT on a larger population of horses housed in both familiar and unfamiliar environments. Untamed meat horses are often conducted in unfamiliar pens and kept there with a low space allowance before loading, so the BUT should be also re-conducted in a real setting; during the application of the BUT in a real-world, many other problems could happen, which may require a refinement of the described procedures. However, even if our results are preliminary, they confirmed that tamed and untamed horses have a different reaction when approached, haltered, and led. These differences in reactivity and their relationship with humans suggest that—as is currently the case under Regulation EC 1/2005—different transport procedures must be followed in these two groups of horses. This would help to reduce the distress that can be associated with the transport of horses with different levels of taming prior to transport. Since there is a need for a robust procedure that allows identification of these animals, based on our findings, it may be suggested to include the BUT in the legislation on the protection of welfare during live animal transport. This will allow personnel to define, prior to shipping, whether a horse is broken/tamed or unbroken/untamed, thus paralleling the current requirement for pre-transport assessment of fitness for travel. BUT test would take a bit of time during the preparation phase of transport; however, this little time investment may be crucial to safeguard the welfare of the travelling horses as well as the horse handlers, who often get injured during loading and unloading procedures. Horses and human health and welfare are indeed interconnected, and the application of BUT may therefore enhance both.5. ConclusionsThis study described the development of a reliable and valid behavioural test (BUT) that seems to be able to estimate a horse’s prior level of taming and classify it as broken or unbroken. The BUT was associated with excellent inter- and intra-observer agreement and test–retest reliability, and its construct and criterion validity are supported by its association with physiological and behavioural parameters and expert judgment. By contrast, our study suggests that the current definition of unbroken, as written in Regulation EC 1/2005, is not fit for purpose. Inclusion of the BUT in future codes for animal transportation and regular application of the BUT before shipping could help transporters to direct horses towards the correct transport procedures, help competent authorities to verify compliance with the Regulation, and—most importantly—allow horses to travel under conditions that are appropriate for their prior experience, thereby avoiding many injuries and much suffering. Ultimately, widespread adoption of the BUT could safeguard the welfare of millions of horses during transport and also minimise horse-related injuries to humans. However, further studies are needed to confirm our findings and to optimise transport strategies for unbroken horses and evaluate the effects of handling, loading, and travelling training on the horses’ emotional state and the incidence of transport-related problems.

animals : an open access journal from mdpi

10.3390/ani11041061

PMC8069115

This review summarizes the clinical background, possibilities, and limitations of twin pregnancy diagnosis in cattle, with a special emphasis on pregnancy loss aspects. Due to the fact that reproductive performance is strongly affected by twin pregnancy, clinical veterinarians should become familiar with the correct diagnosis of this phenomenon. Thus, each herd must plan a herd-specific management practice to detect twin pregnancies. A highly accurate diagnostic tool is required to detect such pregnancies, already during the first month of gestation. This commentary review focuses on the diagnostic possibilities and limitations on the field.

Twin pregnancies are an economically unwanted phenomenon in dairy cattle, not only because they increase pregnancy losses, but also because antibiotics usage and culling rate of the dam are also dramatically increased due to them, furthermore animal welfare issues are also affected through them. In cattle, under field conditions using an early pregnancy determination tool, the first accurate diagnosis from the pregnancy status is available from around day 28, although further confirmations of pregnancy are required. Twin pregnancy diagnosis is available either by rectal palpation or ultrasonography. The measurement of pregnancy specific proteins are also available to determine gestation, but there is still a long way to go to properly identify twin pregnancies. In this commentary, we compared our own results with the literature data in this field with a special emphasis on the clinical practices.

1. IntroductionThe diagnosis of twin pregnancy is one of the key factors for reproductive performance in well managed dairy herds. Among the undesirable consequences are higher percentage of pregnancy losses [1,2,3,4], shorter duration of gestation, increased stillbirth and dystocia rates [5,6,7], and increased frequency of postpartum complications (retained placenta, metritis), and consequently the increase of the use of antibiotics [6,8] should be mentioned. Twin pregnancy also has negative consequences on the newborn calves (decreased birth weight, increased mortality) [8]. Another disadvantage is freemartinism [9], due to sex chromosome chimerism resulting in grossly abnormal internal genitalia of freemartin heifers [8,10]. Twin pregnancy also shortens the length of gestation [8,11].Earlier observational data in Hungary including more than 13,000 calvings in five years reported a 3.4% incidence rate of twin pregnancies [12], with a maximum of eight to nine percent in some years [7,12] at calving time. A much higher twin pregnancy rate can be found at the time of pregnancy diagnosis. A study from Spain reported 20% twin pregnancy rate at the time of ultrasound pregnancy diagnosis, at the end of the first month of gestation [13]. The same study reported an increased frequency of unilateral twins. However, our own data [14] showed an almost equal distribution of unilateral and bilateral twins.A recent Hungarian study [15] highlighted that about 70% of the dairy farms introduced some technological tool to detect early pregnancy. Another study from the United Kingdom reported more common application [16] and highlighted the importance of evaluation the pregnancy diagnosis. However, in Hungary about one third of the farms still use rectal palpation to detect pregnancy diagnosis. For that reason, nowadays there are no data about regular screening for twin pregnancy.2. Clinical Diagnosis of Twin PregnancyDuring the last four decades, the rate of twin calving has increased [1,17] due to the increase of multiple ovulations associated with high milk yield [13,18] and with the use of different synchronization protocols [2]. Double ovulation occurs more frequently in multiparous cows with high milk production than in primiparous cows [13,18,19].A recent study advised to set up ‘pivotal periods’, by taking into consideration the practical point of view of the regular reproductive management possibilities. Under field conditions, the application of an early pregnancy determination tool as the first accurate diagnosis is available from around day 28, and further confirmations of pregnancy are required [20]. In our own previous study, pregnancy diagnosis was carried out between days 29 and 42, with an average of day 33.6 [21].It seems that fulfilment of the metabolic requirements of high-yielding cows stands behind double ovulations (during the first 8 weeks of lactation when milk production peaks). In a study [13], cows above a milk production of 50 kg/day showed a rate of multiple ovulations higher than 50 percent.2.1. Rectal PalpationPregnancy diagnosis can be performed in several ways, such as palpation of the amniotic vesicle [22]. The vesicle(s) of about 1–2 cm in diameter can be palpated on the same side where the ovaries carry the corpora lutea. Experienced practitioners can palpate the amniotic vesicle as early as on day 30 of pregnancy. The other possibility for palpating pregnancy is the technique of fetal membrane slipping. It is not commonly used in Europe, while it is more widely applied in the United States [23].Most of these techniques carry the risk of damaging the amniotic vesicle [24,25,26,27,28]. According to Day et al. [29] and Karlsen et al. [30], rectal manipulation carries the highest risk of damaging the embryo. The general pregnancy loss in the above-mentioned reports varied between 3% and 10%. Romano [23] examined cows in early gestation by ultrasonography, and then also used the fetal membrane slipping technique between days 34 and 41 of pregnancy. It was found that repeated ultrasound examinations—when carrying out the confirmational examination with ultrasound—did not increase the pregnancy loss. When analyzing the data of all cows giving data to the study, the total pregnancy loss ratio was 14%. It was concluded that the increased number of rectal palpations and/or the possible endogenous release of prostaglandins might have contributed to this.In Hungary, palpation of fluctuation in the pregnant uterine horn is widely used at days 42–48 of pregnancy [15]. Positive pregnancy diagnosis due to the accumulation of allantoic fluid means an enlarged/asymmetric pregnant uterine horn, while experienced practitioners can successfully perform this examination already between days 30 and 40 of gestation [23,31]; however, pathologic conditions such as pyometra and metritis will present the same signs; therefore, this method should never be used as a stand-alone pregnancy diagnosis method [28].By rectal palpation, bilateral twins can be detected by enlargement of the uterine horns. On the contrary, unilateral twin pregnancies remains unnoticed. For these reasons, rectal palpation cannot be used as a routine screening technique of twin pregnancies.Concerning twins, a result of the ovulation of co-dominant follicles [30] two corpora lutea (CL) are formed, and two embryos will start to develop in the uterine horn(s). The occurrence of monozygotic twins, originating from a single ovulation and spontaneous division of the embryo, is rare, around 5% among all twins [32]. Triplets or higher-order twins are extremely rare in cattle. Although two CL can be present on the same ovary (meaning usually two embryos in the same uterine horn), bilateral twins are more common than unilateral ones, therefore the presence of two CL not necessarily means twins, while double ovulations are much more common than twin pregnancy [13].2.2. Transrectal Ultrasonography (TRUS)Ultrasonography was first introduced into bovine reproduction management in the 1980s. Curran et al. [33] determined the times when different fetal organs of the embryo could be detected. In their study, embryonic heartbeat was detectable as early as on days 22–24 post AI. Nowadays, this is the diagnostic criterion for a positive pregnancy diagnosis. It is difficult to evaluate fetal heartbeat at that early stage in practice, therefore the most implemented period to diagnose pregnancy with transrectal ultrasonography is between 28–35 days of gestation. [34,35,36]. The allantoic fluid is first visible from days 25–26 in multiparous cows and from days 23–24 in heifers [23]. Furthermore, the presence of the embryo and the active heartbeat are the main criteria for a positive early pregnancy diagnosis by means of TRUS under practical conditions. As regards the location of the embryos, they can be located either unilaterally or bilaterally, unilateral twins used to have an echogenic line connecting the two embryos, acting as a diagnostic tool [10].In the case of twin pregnancy, both embryos heartbeats, and the amniotic vesicles should be detected in TRUS examination [37]. In case of doubts in twinning or fetus viability, the evaluation of ovarian structures could be assessed: (1) Number, usually two (30) due to the fact that mostly of twin pregnancies come from the ovulation of co-dominant follicles not from monozygotic twins [38]; (2) Quality of CL, meaning that the echogenicity and the size of the CL, which should be at least 17 mm in the largest diameter [39]. Our data confirmed this [14]; furthermore, we found a low number of twin pregnancies with three corpora lutea. The proportion of cows carrying singleton pregnancies with two corpora lutea was around 10%. With the diagnostic possibilities, it is hard to differentiate between singleton pregnancies and a co-dominant pregnancy or a possible partial embryonic mortality [40].2.3. Examination of Pregnancy ProteinsFrom day 22 of pregnancy, mononuclear cells originating from the trophoblast migrate into the endometrium. Cell migration originating from the trophoblast can be observed throughout the gestation. During this process, they turn into binucleate and, in some cases, trinucleate cells. Because of this migration, the bovine placenta is called a synepitheliochorial placenta [41,42].Pregnancy-associated glycoproteins can be differentiated into two subgroups: the PAG-2 subgroup [43] was known to mainly be localized at the fetal-maternal borderline, although some reports recently showed them in the maternal plasma [44,45]. The PAG-1 subgroup is expressed mostly on the bi- and trinucleate cells of the trophoblast, although mononucleate cells also secrete it [46,47]. The measurement of them was used to discriminate singleton and twin pregnancies [48]. Molecular cloning studies showed that the amino acid sequence of PSP-B is homologous to that of PAG1, and they are inactive members of the aspartic acid proteinase family [49]. The isolated preparations differed in carbohydrate and sialic acid content. These characteristics may explain the minor differences in their profile and in their disappearance from the maternal circulation after calving or embryonic mortality [50].Despite 30 years of intensive research, the clear function of the production of these proteins has remained unknown. The welfare and viability of the fetus, as well as pregnancy loss during gestation, can be monitored by their use [45,48,50]For more than 20 years, different methods were used for measuring pregnancy proteins: blood sera [51] and milk [52,53]. Serum is the method of election because these proteins have lower concentrations in milk [20]. Besides the biological fluid, numerous factors may affect the concentrations of pregnancy proteins in serum. Yániz et al. [54] showed significant differences in PAG-1 levels between individuals depending on the sire, the number of embryos, and the season.Several studies showed that in dams carrying two embryos, higher concentrations of PAG-1 could be measured, and with the progress of gestation, the elevation of PAG-1 concentration was also higher than in cows carrying singletons [55,56]. Despite this, PAG-1 cannot identify partial pregnancy loss or embryo reduction [56]. Moreover, the shorter half-life of PAG compared to PSP-B is also remarkable [57]. One study determined that TRUS is needed for twin pregnancy diagnosis [58] because clinically applicable cut-off value for the diagnosis of twin pregnancy could only be reached from day 85 of pregnancy.3. The Effect of Twin Pregnancy on Pregnancy Losses under Practical ConditionsThere are several factors influencing the pregnancy loss associated to twin pregnancy. One of the most important factors is the embryo’s location in the uterus. Bilateral twin pregnancies resulted in a lower number of pregnancy losses than unilateral twin pregnancies [56], indicating that possibly the physical extensions of two embryos can induce pregnancy loss. Heat stress effect was studied to influences pregnancy losses of dams pregnant with twins [32], and in another study, heat stress also effected pregnancy loss with the same odds as carrying a twin pregnancy [59]. The method of pregnancy diagnosis can also be a source of pregnancy loss; however a recent study, when veterinary students under training were performing rectal palpation [60], could not confirm this. In case of increased losses after early pregnancy diagnosis, to date there is no data and most of the recent scientific results supported the ineffectiveness of the careful examination.The primary pregnancy diagnoses must be confirmed later [20,56,61] to achieve reliability and to manage herd health issues. This confirmation is usually performed around day 60 of pregnancy when placentation is completed [24]. Most of the losses in singleton and twin pregnancies usually occur between the time of early pregnancy diagnosis and the confirmation of gestation [20,62]. Twin pregnancy loss also showed a peak in late gestation [14].4. Management of Twin PregnanciesThe individual animal performing of induced reduction in case of twin pregnancy is a described method [63,64]. The manual embryo reduction is considered to be the most economic option to mitigate the negative impacts of twinning [65]; however, practical application is not widespread at the moment. The maintaining strategy is based on medical treatments [66] highlighting that the economic circumstances must be taken into consideration as well. The total herd level termination of twins by using prostaglandin F2alpha is a possible strategy [67], but farmers’ resistance should be noted when performing it under practical circumstances. In dairy herds with low incidence rate of twinning, termination of pregnancy can be an option.In case of a dairy herd, the early diagnosis of twin pregnancies is the primary goal. Previously, it was shown in two independent studies under practical circumstances [14,68], that the application of different synchronization protocols prior AI decreased the percentage of the occurrence of twinning. The possible explanation was that the stronger regulation of the ovarian cycle was decreasing the possibility of the ovulation of coordinated follicles [67]. An alternative solution at herd level can be the screening for twins at the early diagnosis with a selected diagnostic method, after that the identified twins undergo confirmational diagnosis every second week until day 60 of pregnancy in order to rule out partial or total losses. At the time of the initial twin diagnosis, medical support should be given. It was also demonstrated in clinical practice, that calving time stillbirth events dramatically increase due to twinning [14], therefore careful management in the calving barn is also part of the defense on one hand, with the training of the assistance to handle twin calvings and on the other hand to support the newborn animals.5. ConclusionsBased on the data presented above, nowadays all the pregnancy examination methods seem to be feasible in detecting twin pregnancies in dairy cows. The rectal palpation is able to distinguish between unilateral and bilateral twin pregnancies. Transrectal ultrasonography is the most widespread method under practical circumstances, because both unilateral and bilateral twins can be detected. Moreover, heartbeat (and through this fetal viability) and corpora lutea can be monitored. The measurement of pregnancy-specific proteins from biological fluids has been a promising tool for 20 years, but at the moment, the detection of twins is limited. Further studies should be performed in order to increase the accuracy of this technique.Pregnancy loss also affects reproductive performance with special regard to twin gestations in cattle. While more authors demonstrated an increased occurrence of losing twin pregnancies, especially in the case of unilateral twins, the distribution of laterality shows high variability, resulting under field circumstances various results.

animals : an open access journal from mdpi

10.3390/ani11030741

PMC8001781

Oocyte-secreted factors play an essential role in oogenesis and fertility through bidirectional crosstalk between oocytes and somatic cells. Interleukin-7, known as an oocyte-secreted factor, has recently been shown to improve oocyte developmental competence through interaction with cumulus cells around the oocytes. This study aimed to investigate the effects of interleukin-7 on porcine cumulus-oocyte complexes during in vitro maturation. Our results showed that supplementation with interleukin-7 during in vitro maturation exerted beneficial effects on porcine oocyte meiotic maturation by upregulating antioxidant-related genes and enhanced the subsequent developmental potential of porcine embryos after parthenogenetic activation.

Interleukin-7 (IL-7) is a cytokine essential for cell development, proliferation and survival. However, its role in oocyte maturation is largely unknown. To investigate the effects of IL-7 on the in vitro maturation (IVM) of porcine oocytes, we analyzed nuclear maturation, intracellular glutathione (GSH) and reactive oxygen species (ROS) levels, and subsequent embryonic developmental competence after parthenogenetic activation (PA) under several concentrations of IL-7. After IVM, IL-7 treated groups showed significantly higher nuclear maturation and significantly decreased intracellular ROS levels compared with the control group. All IL-7 treatment groups exhibited significantly increased intracellular GSH levels compared with the control group. All oocytes matured with IL-7 treatment during IVM exhibited significantly higher cleavage and blastocyst formation rates after PA than the non-treatment group. Furthermore, significantly higher mRNA expression levels of developmental-related genes (PCNA, Filia, and NPM2) and antioxidant-related genes (GSR and PRDX1) were observed in the IL-7-supplemented oocytes than in the control group. IL-7-supplemented cumulus cells showed significantly higher mRNA expression of the anti-apoptotic gene BCL2L1 and mitochondria-related genes (TFAM and NOX4), and lower transcript levels of the apoptosis related-gene, Caspase3, than the control group. Collectively, the present study suggests that IL-7 supplementation during porcine IVM improves oocyte maturation and the developmental potential of porcine embryos after PA.

1. IntroductionPorcine embryos derived from in vitro maturation (IVM) are useful for research regarding reproduction [1,2]. In particular, as pigs have an organ structure that is physiologically similar to that of humans, many studies use porcine embryos in the field of biomedical research [3,4,5]. To obtain consistent results in these fields, such as the production of transgenic disease models, the ability of the oocyte must be satisfied at an early stage of oogenesis [6,7]. The improvement in oocyte capacity is closely related to IVM. To improve the efficiency of porcine IVM, addition of various growth factors and other molecules to the IVM medium has been suggested [8,9,10]. Although these efforts have led to significant advances in the porcine IVM system, the in vitro matured oocytes exhibit lower developmental competence compared with matured oocytes in vivo, as the in vitro environment is not as optimum as in folliculogenesis [11].Environmental factors around the oocyte play an important role in the developmental competence of the oocyte. For example, during folliculogenesis, bidirectional somatic cell–oocyte signaling is essential to simultaneously alter follicle development with oocyte maturation [12]. In particular, oocyte-secreted factors (OSFs), which are soluble growth factors secreted from the oocyte, play a critical role in oogenesis and fertility by regulating the functions of granulosa cells and cumulus cells (CCs) [13]. Among the OSFs, growth differentiation factor 9 (GDF9) and bone morphogenetic protein 15 (BMP15) are known to be vital for the initiation of primordial follicles and follicular development in many species [14]. In mice, deletion of GDF9 and BMP15 is detrimental to ovarian follicular development and fertility [15,16]. In addition, GDF9 and BMP15 co-treatment during IVM increases the embryonic development of porcine oocytes by enhancing the expansion of CCs [17].IL-7, a cytokine primarily secreted by stromal cells in bone marrow and thymus, plays a pivotal role in cellular differentiation of lymphocytes [18]. IL-7 binds to the IL-7 receptor complex that consists of the IL-7Rα chain (IL-7R) and the common cytokine γ chain (γc) [19]. The IL-7R binds explicitly to IL-7 and leads to Janus kinase 1 (JAK1) activity, a receptor related to tyrosine Janus kinases, whereas the γc activates Janus kinase 3 (JAK3) [20]. Several signaling pathways that are involved in proliferation, survival, cell cycling, and metabolism, are affected when IL-7 binds to these receptors [21]. In naïve and memory T cells, IL-7 acts as an essential factor for survival [22] and homeostatic proliferation by downregulating apoptotic activity and upregulating growth activity [23]. In the reproductive system, oocyte-secreted IL-7 has been shown to act on granulosa cells as a survival factor by inhibiting apoptosis and enhanced oocyte maturation in rats [24]. In addition, IL-7 can potentially enhance the developmental competence of oocytes by stimulating proliferation of mouse CCs [25]. Recent research reported that low IL-7 concentrations increased favorable intracellular reactive oxygen species (ROS) levels and significantly improved oocyte maturation rate by downregulating the apoptotic process in sheep [26].Although these studies have reported a role for IL-7 in other species, the function of IL-7 in the porcine reproductive system, including its effect on oocytes and CCs, has not been investigated. Therefore, we hypothesized that supplementation with IL-7 during IVM may enhance the oocyte quality by replenishing insufficient intracellular signaling in the in vitro environment. In this study, we investigated whether IL-7 is effective in both porcine IVM and the subsequent pre-implantation embryonic development following parthenogenetic activation (PA).2. Materials and Methods2.1. Chemicals and ReagentsRecombinant human IL-7 was purchased from PeproTech (London, UK). Unless otherwise stated, all chemicals and reagents used in the present study were purchased from Sigma-Aldrich (St. Louis, MO, USA).2.2. Measurement of IL-7 in Porcine Follicular Fluid (FF)Follicular fluid was collected from different ovary pairs of prepubertal three-way cross pigs (mixed Yorkshire, Landrace, and Duroc breeds) obtained in a local abattoir. Porcine follicles were aspirated from three groups according to their diameters (small (1–2 mm), medium (3–7 mm), or large (≥8 mm)) [27]. All collected porcine FF were centrifuged at 1000 rpm at 4 °C for 20 min to eliminate debris and blood. Supernatants were separately filtered through 1.2 μm syringe filters (Sartorius Stedim Biotech, Aubagne, France) and then frozen at −80 °C until analysis. Concentrations of IL-7 in porcine FF were measured using an ELISA kit (Fine Test, Wuhan, China), according to the manufacturer’s instructions. All ELISA experiments were performed in quadruplicate, and a mean value was used to determine the IL-7 concentration in porcine FF.2.3. Oocyte Collection and IVMPorcine ovaries were obtained from a local abattoir and transported to the laboratory within 1 h in 0.9% NaCl at 37 °C. Thereafter, cumulus-oocyte-complexes (COCs) were aspirated into 15 mL conical tubes from antral follicles of 3–7 mm in size using an 18-gauge needle and a 10 mL disposable syringe. The debris was allowed to settle at the bottom of the tube at 37 °C for 5 min. Subsequently, the porcine FFs were removed and the sediments were resuspended twice in HEPES-buffered Tyrode’s medium containing 0.05% (wt/vol) polyvinyl alcohol (TLH-PVA). The COCs surrounded with intact compact CCs layers and with evenly granulated cytoplasm were chosen using a stereomicroscope for IVM and then washed twice in TLH-PVA. After washing once in IVM medium, approximately 50–60 arbitrarily selected COCs were cultured in IVM medium in a four-well Nunc dish (Nunc, Roskilde, Denmark). IVM medium (TCM-199; Invitrogen Corporation, Carlsbad, CA, USA) supplemented with 0.6 mM cysteine, 0.91 mM sodium pyruvate, 10 ng/mL epidermal growth factor, 75 μg/mL kanamycin, 1 μg/mL insulin, and 0.1% (wt/vol) PVA was added to each well. The IVM process occurred over 42 h. Collected COCs were incubated with 10 IU/mL equine chronic gonadotropin (eCG) and 10 IU/mL human chorionic gonadotropin (hCG) (Intervet, Boxmeer, Netherlands) at 39 °C in a humidified 5% CO2 atmosphere for the first 22 h. Thereafter, COCs were cultured in a hormone-free maturation medium for the remaining 20 h. During the entire IVM period, IL-7 was added to the media at concentrations of 0 (control), 0.1, 1, and 10 ng/mL for each group. The concentration of IL-7 was decided based on a previous IL-7 study in rat granulosa cells [24].2.4. Assessment of Nuclear MaturationOocytes from COCs were obtained by mechanically denuding surrounding CCs using 0.1% hyaluronidase after IL-7 supplementation post-IVM. The denuded oocytes were washed twice in TLH-PVA medium and immediately transferred into 30 μL TLH-PVA containing 5 μg/mL Hoechst-33342, and then stained in the dark for 10 min to assess the nuclear maturation rates. The stained oocytes were observed under a fluorescence microscope (Nikon Corp., Tokyo, Japan) with a UV filter (370 nm). Oocytes from each group were classified into germinal vesicle (GV), metaphase I (MI), anaphase-telophase I (AT), and metaphase II (MII) stages according to the classification by Naito and Toyoda [28].2.5. Measurement of Intracellular GSH and ROS LevelsDenuded oocytes of MII-stage were collected using 0.1% hyaluronidase after 42 h IVM to evaluate the intracellular GSH and ROS levels. The GSH and ROS levels were determined using methods according to You et al. [29]. Briefly, CellTracker Blue 4-chloromethyl-6,8-difluoro-7-hydroxycoumarin (CMF2HC; indicated by blue fluorescence; Invitrogen) and 2′,7′-dichlorodihydrofluorescein diacetate (H2DCFDA; indicated by green fluorescence; Invitrogen) were used to measure GSH and ROS in the oocyte cytoplasm, respectively. Twenty oocytes were placed in 30 μL TLH-PVA supplemented with 10 μM CMF2HC or 10 μM H2DCFDA and incubated in the dark for 30 min. Thereafter, the stained oocytes were washed thrice with TLH-PVA and placed into a 8 μL drop of TLH-PVA, which was subjected to fluorescence microscopy using an epifluorescence microscope (TE300; Nikon) with UV filter (370 nm for GSH and 460 nm for ROS). Adobe Photoshop CS6 was used to examine the fluorescence intensity of oocytes and normalized to control oocytes. The independent experiment was repeated three times (GSH samples, N = 60; ROS samples, N = 60).2.6. Gene Expression Analysis Using Real-Time qPCRMatured oocytes and CCs were sampled from 50–60 COCs from each group by gently pipetting with 0.1% hyaluronidase after the 42 h IVM. Denuded oocytes of MII-stage and their respective CCs were sampled into 1.5 mL microfuge tubes and frozen at −80 °C until assayed. The gene expression analysis was performed using real-time qPCR for 13 specific genes associated with different functions, such as apoptosis: BCL2 associated X (BAX), BCL2 like 1 (BCL2L1), and caspase-3 (CASP3); mitochondrial-related genes: transcription factor A (TFAM) and NADPH oxidase 4 (NOX4); antioxidant-related genes: glutathione-disulfide reductase (GSR) and peroxiredoxin 1 (PRDX1); IL-7 associated genes: phosphoinositide-3-kinase regulatory subunit 1 (PIK3R1), AKT serine/threonine kinase 1 (AKT1), and solute carrier family 2 member 1 (SLC2A1; GLUT1); developmental competence-related genes: proliferating cell nuclear antigen (PCNA), KH domain containing 3 like (KHDC3L; Filia), nucleophosmin/nucleoplasmin 2 (NPM2), hyaluronan synthase 2 (Has2), and TNF alpha induced protein 6 (TNFAIP6). The mRNA expression levels of BAX, BCL2L1, CASP3, TFAM, NOX4, GSR, PRDX1, PIK3R1, AKT1, GLUT1, PCNA, Has2, and TNFAIP6 were measured in CCs, and those of BAX, BCL2L1, CASP3, NOX4, GSR, PRDX1, PIK3R1, AKT1, GLUT1, PCNA, Filia, and NPM2 were measured in oocytes. All primer sequences are provided in Table S1.Total RNA was isolated from stored CCs and oocytes using TRIzol reagent (TaKaRa Bio, Inc., Otsu, Shiga, Japan) according to the manufacturer’s protocol. Complementary DNA (cDNA) was prepared from 1 μg of total RNA using Reverse Transcription Master Premix (Elpis Bio, Inc., Daejeon, Korea). In total, 1 μg of the synthesized cDNA was amplified using 2× SYBR Premix Ex Taq (Takara Bio, Inc.) and specific primers (5 pmol) by real-time qPCR (CFX96 real-time qPCR cycler (Bio-Rad, Hercules, CA, USA). The cycling parameters were as follows: 95 °C for 5 min, followed by 40 cycles of 15 s at 95 °C, 15 s at 56 °C and 30 s at 72 °C. The expression of each target gene in CCs and oocytes was quantified relative to a housekeeping gene, glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and 18S ribosomal RNA (RN18S). The relative quantification was determined by comparing the threshold cycle (Ct) at constant fluorescence intensity. The relative mRNA expression (R) was calculated using the equation, R = 2−[ΔCt sample − ΔCt control]. The R values were normalized using GAPDH for cumulus cells, and RN18S for oocytes. The mean of three replicates was analyzed for statistical analysis.2.7. Parthenogenetic Activation and In Vitro Culture of Porcine EmbryosAfter IVM, CCs were removed from oocytes as described above. The MII-stage oocytes from each group were selected to conduct PA as described by Jeon et al. [30]. The oocytes were washed twice in activation solution containing 280 mM mannitol, 0.01 mM CaCl2 and 0.05 mM MgCl2. The activation chamber connecting electrodes was filled with 2 mL of activation solution. The oocytes were placed in the chamber and then activated with two direct-current pulses of 120 V/mm for 60 μs. After PA, oocytes were transferred to IVC medium [31] supplemented with 5 μg/mL of cytochalasin B for 4 h under a humidified atmosphere of 5% CO2, 5% O2 and 90% N2. After incubation, the activated embryos were washed thrice in the IVC medium, placed for seven days in 25 μL droplets of IVC medium (10 gametes per drop) covered with mineral oil, and then cultured under the same atmospheric conditions. The culture media were refreshed at 48 h (Day 2) and 96 h (Day 4, with 10% FBS) after PA.2.8. Evaluation of Developmental Competence and Total Cell CountDay 0 was regarded as the day on which PA was initiated. Day 2 after PA, the cleavage formation was analyzed and embryos were categorized into three groups (2–3 cells, 4–6 cells and 7–8 cells). Blastocyst formation was assessed at seven days post PA and the blastocysts were categorized into three groups according to their morphology (early, expanded, and hatched), as reported in a previous study [8]. To calculate the total cell number of blastocysts at Day 7, the blastocysts were washed in TLH-PVA and fixed in 4% paraformaldehyde in PBS-PVA and stained for 5 min using 5 μg/mL Hoechst-33342. Next, the blastocysts from each group were transferred to a drop of 100% glycerol on glass slides and gently covered with a coverslip. The stained blastocysts were observed using a fluorescence microscope (Nikon, Tokyo, Japan) at 400× magnification. The experiment was repeated three times.2.9. Statistical AnalysisEach experiment was repeated at least three times. The COCs used in each experiment were collected in the same abattoir on the same day and then were randomly used in each group. The rates of cleavage and blastocyst formation from the activated embryos were analyzed on the same day. Statistical analysis was conducted using SPSS 12.0 (SPSS, Inc., Chicago, IL, USA). Percentage data (rates of nuclear maturation and embryonic development) and average data (ELISA, intracellular GSH and ROS levels in oocytes and total cell number in blastocyst) were analyzed using Duncan’s multiple range test after one-way ANOVA. The data are presented as the mean or the mean ± standard error of the mean (SEM). The values from the ELISA experiments are presented as the mean ± standard deviation (SD). Values of p < 0.05 were considered statistically significant.3. Results3.1. Detection of IL-7 in Porcine FFELISA was performed to determine the concentration of IL-7 in porcine FFs at each ovarian follicle size (small, medium, and large). The maturation medium, M199, was used as a negative control. IL-7 was detected in all porcine FFs obtained from different follicle sizes. Interestingly, the concentration of IL-7 in the medium FF group (64.2 ± 39.2 pg/mL) was significantly higher than that of the small FF group (6.8 ± 5.0 pg/mL) (p < 0.05). There was no significant difference in the large FF group (44.0 ± 22.8 pg/mL) compared to other groups (Table 1).3.2. Effect of IL-7 Supplementation during IVM on Oocyte Nuclear MaturationTo evaluate the effect of IL-7 treatment on the nuclear maturation of porcine oocytes during IVM, matured oocytes were evaluated at stages GV, MI, AT, or MII. After 42 h of IVM, the 1 ng/mL IL-7-supplemented group (97.3%) showed a significantly higher (p < 0.05) MII rate compared with the control (91.6%). However, no significant differences were observed in other IL-7-treated groups (0.1 and 10 ng/mL group: 92.2 and 92.1%) compared with the control (Table 2).3.3. Effect of IL-7 Supplementation during IVM on Cytoplasmic MaturationTo assess the effects of IL-7 supplementation on cytoplasmic maturation during IVM, MII stage oocytes from each group were selected and stained as described above (Figure 1A). The intracellular GSH levels were significantly increased (p < 0.05) in oocytes from all IL-7-treated groups compared to the control group (Figure 1B). The 1 and 10 ng/mL IL-7-treated groups showed significantly (p < 0.05) lower intracellular ROS levels than the control group (Figure 1B). In this study, the IL-7 supplemented during IVM significantly affected intracellular GSH and ROS levels in matured oocytes.3.4. Effect of IL-7 Supplementation on Gene Expression Levels in CCs and Oocytes during IVMTo examine the effect of IL-7 on the expression of apoptosis-, mitochondrial-, and antioxidant-related genes, the mRNA expression levels of Bax, BCL2L1, CASP3, TFAM, NOX4, GSR, and PRDX1 were assessed in CCs and oocytes. As shown in Figure 2A, the CCs supplemented with 1 ng/mL IL-7 showed significantly higher anti-apoptotic gene BCL2L1 levels and lower pro-apoptotic gene CASP3 levels than the control group (p < 0.05). In the mitochondrial-related genes, the transcription levels of TFAM were significantly increased in the CCs treated with 10 ng/mL IL-7 compared with control (p < 0.05). However, the TFAM transcript levels in oocytes were too low to be accurately quantified. In addition, NOX4 transcript levels were significantly higher in all IL-7-treated CCs than in the control group (p < 0.05). Levels of the antioxidant-related genes GSR and PRDX1 were significantly increased in the CCs treated with 10 ng/mL IL-7 compared with controls (p < 0.05). Further, oocytes supplemented with 0.1 ng/mL IL-7 displayed significantly higher (p < 0.05) transcript levels of GSR than the control group (Figure 2B). PRDX1 mRNA levels appeared significantly increased in oocytes treated with 10 ng/mL IL-7 (p < 0.05).To investigate the role of IL-7 on the expression of IL-7-associated and developmental competence-related genes, mRNA expression levels of PIK3R1, AKT1, GLUT1, PCNA, Has2, TNFAIP6, Filia, and NPM2 were determined in CCs and oocytes from each group. As shown in Figure 3A, transcription levels of PIK3R1 significantly decreased in CCs supplemented with 10 ng/mL IL-7 compared to the control (p < 0.05). However, no significant difference was observed in PIK3R1, AKT1, and GLUT1 transcript levels in IL-7-supplemented oocytes compared with the control group. In developmental competence-related genes, PCNA transcript levels were significantly higher in 0.1 ng/mL IL-7-treated oocytes than in the control (p < 0.05). Transcript levels of Filia and NPM2 were significantly increased (p < 0.05) in oocytes treated with 0.1 or 10 ng/mL IL-7 compared with the control group (Figure 3B).3.5. Effect of IL-7 Supplementation in IVM Media on Developmental Potential after PAAfter PA, cleavage and blastocyst rates were significantly higher (p < 0.05) in all IL-7-treated groups than in the control group (Table 3, Figure S1). In addition, the total cell number in the blastocyst was significantly (p < 0.05) increased in the 1 ng/mL IL-7-treated group compared to the control (Table 3). The cleavage patterns differed in IL-7-supplemented groups. As shown in Figure 4A, 4–6 cell cleavage rates of the PA embryos were significantly increased in all IL-7-supplemented groups (p < 0.05). The 1-cell and fragmentation group and 7–8 cell cleavage rates were significantly lower in all IL-7-treated groups than in the control group (p < 0.05).4. DiscussionBidirectional communication between the oocyte and its surrounding somatic cells is essential for follicle development, oocyte maturation, and acquisition of developmental competence for oocytes in mammals [13,32]. OSFs, such as GDF9 and BMP15, are involved in this interaction [33]. IL-7 is an OSF detected in human FF [34,35]. However, the physiological function of IL-7 during oogenesis and folliculogenesis is still unknown. This is the first study to demonstrate the presence of IL-7 in porcine FF. Notably, supplementation with IL-7 during IVM exerted beneficial effects on nuclear maturation and improved cytoplasmic maturation of porcine oocytes by upregulating antioxidant-related genes. IL-7 also enhanced the subsequent developmental potential of PA porcine embryos. Further, expression levels of various genes were significantly changed in CCs and matured oocytes post-IL-7 treatment.Cytokines are essential for successful progression during folliculogenesis, such as primary-to-antral follicle transition [36]. The intraovarian/perifollicular cytokine environment is closely related to oocyte quality and viability [37]. This study showed that the cytokine IL-7 is present in porcine ovarian FF. Furthermore, the concentration of IL-7 in medium follicles (3–7 mm) is significantly higher than that observed in small follicles (1–2 mm). In patients with abnormal ovarian follicles, a lower IL-7 level in FF is correlated with lower embryo quality and lower in vitro fertilization outcome [34]. As previously reported, oocytes collected from small follicles (≤ 3 mm) reveal inadequate cytoplasmic maturation due to lower levels of stored mRNA and proteins related to developmental competence [38,39]. Hence, higher IL-7 levels found in medium follicles provide evidence of its role in porcine follicle development.Oocyte meiotic maturation is an essential process during in vitro production (IVP) of embryos and determines whether an oocyte is competent to undergo fertilization and embryogenesis. This process contains developmental programs such as nuclear and cytoplasmic maturation. For the successful IVM of oocytes, the oocytes must undergo synchronization between nuclear and cytoplasmic maturation [40,41]. Previous studies reported that IL-7 elevates nuclear maturation of pre-ovulatory oocytes in rats and sheep [24,26]. Consistently, the 1 ng/mL IL-7 treatment significantly increased the number of porcine oocytes at the MII stage. This study suggests that IL-7 supplementation during IVM can positively affect the nuclear maturation of porcine oocytes.After IVM, GSH levels in oocytes are regarded as an indicator of oocyte cytoplasmic maturation [42]. Further, GSH plays a vital role in protecting the cell against ROS toxic activity [43]. Results indicate that IL-7 supports the production of intracellular GSH and suppresses ROS levels in matured oocytes. ROS often induce cellular damage, leading to apoptosis, and cause DNA breakdown due to the formation of free radicals [44,45]. ROS play a concentration-dependent role in meiotic resumption of the oocyte during mammalian reproduction—lower ROS concentration assists meiotic resumption, whereas a higher ROS level causes meiotic cell cycle arrest and apoptosis. Therefore, an adequate balance is vital for oocyte quality [45]. In human T-cell acute lymphoblastic leukemia cells, IL-7-induced generation of ROS is reportedly involved in PI3K-AKT signaling, which aids survival [46]. In ovine studies, the appropriate treatment of IL-7 increases favorable intracellular ROS levels in matured oocytes [26]. In contrast, the results of this study indicate a decrease in ROS levels in mature oocytes, which is thought to be due to differences between species and cell source. Consistent with GSH and ROS results, the 10 ng/mL IL-7-supplemented oocytes and CCs displayed a significant up-regulation of GSR and PRDX1 mRNA. The H2O2 produced by ROS can be converted to H2O by PRDX1 and GSH [47]. GSH homeostasis is regulated by GSR, which catalyzes the reduction of glutathione disulfide (GSSG) to the sulfhydryl form GSH [48]. Therefore, the GSH increase is likely to be exerted by increased expression of GSR and PRDX1 transcript levels by IL-7, along with its direct antioxidant effect to improve the cytoplasmic maturation during porcine IVM.Mammalian studies reported that an increase in GSH and decrease in ROS levels in oocytes during IVM are related to developmental competence and affect female reproductive outcome [49,50,51]. Consistent with the effects on GSH and ROS, the IL-7-treated groups exhibited significantly increased mRNA expression levels of the developmental-related genes PCNA, Filia, and NPM2 in matured oocytes. It was reported that PCNA determines the fate of the oocyte by increasing its expression in oocytes during the initiation of primordial follicle development [52]. Furthermore, accumulation of Filia and NPM2 mRNA, as maternal effect genes, in oocytes activates the embryonic genome and are involved in oocyte maturation, fertilization, and early embryonic development [53,54]. However, Has2 and TNFAIP6 transcript levels, as cumulus expansion-related genes [55], did not affect the CC expansion upon IL-7 treatment. Consistent with a previous report, IL-7 does not appear to be involved in CC expansion [25]. In the PA embryonic development analysis, the groups treated with IL-7 displayed significantly increased cleavage and blastocyst formation rates. Therefore, IL-7 supplementation during IVM improved embryonic development after PA by upregulating developmental competence genes in porcine oocytes.Several previous studies have mentioned the importance of communication between oocytes and CCs to acquire oocyte developmental competence [13,56,57]. In particular, increased apoptosis in CCs causes detrimental nuclear maturation in oocytes and reduces the fertilization rate [58]. IL-7 reportedly suppresses precursor B cell apoptosis by providing a survival factor via the BCL-2 family proteins [59]. Moreover, IL-7 decreases apoptosis of granulosa cells via suppression of caspase 3/7 [24]. As previously reported, CASP3 and caspase-7 transcript levels were increased in the CCs of patients with polycystic ovary syndrome, which can lead to higher apoptosis in these cells and negatively affect embryo quality [60]. This study observed that transcript levels of BCL2L1 increased, while that of CASP3 decreased in CCs treated with 1 ng/mL IL-7, compared with controls. The caspase-3 cascade, a mitochondrial apoptotic signaling pathway, is regulated by cytochrome c released from mitochondria. Mitochondrial cytochrome c is repressed by the anti-apoptotic BCL-2 protein [61]. Therefore, this result indicated that IL-7 may inhibit mitochondrial-associated apoptosis in porcine CCs. However, it is unclear how IL-7 increased the transcript levels of BCL2L1 in porcine CCs, and further study is necessary. Furthermore, this study found that CCs treated with IL-7 significantly increased transcript levels of TFAM and NOX4. The production of TFAM mRNA is related to mitochondrial DNA transcription and supports mitochondrial respiratory function [62,63]. In addition, NOX4 is located in the mitochondria and produces ROS [64]. It was previously reported that ROS induced by mitochondrial respiration and NADPH oxidase contributes to the survival of T-cell acute lymphoblastic leukemia cells by being associated with IL-7-mediated signaling [46]. These results indicated that IL-7 may potentially enhance oocyte developmental competence by interacting with mitochondrial-related apoptosis and survival signals in porcine CCs. Therefore, further study is necessary to investigate the potential for mitochondrial function following IL-7 supplementation during porcine IVM.5. ConclusionsIn conclusion, for the first time, the present study demonstrated that IL-7 is present in porcine FF, using ELISA. The present study showed that IL-7 supplementation enhances porcine oocyte meiotic maturation by simultaneously improving nuclear maturation and cytoplasmic maturation due to the antioxidant effect. In particular, the enhanced oocyte quality improved the subsequent developmental potential of porcine embryos after PA. Furthermore, IL-7 treatment during IVM may improve the developmental competence of oocytes by inhibiting mitochondrial-related apoptosis and improving survival in porcine CCs. These findings may provide further insight into the effects of IL-7 on follicular development or oocyte meiotic maturation and might aid in enhancement of porcine IVM system and related techniques.

animals : an open access journal from mdpi

10.3390/ani13081406

PMC10135110

We observed a forage allowance by forage type effect on improvements in the daily milk yield of dairy cows grazing swards sown with perennial ryegrass, white clover and plantain relative to those sown to perennial ryegrass only. Improvements in milk yield were evident at lower forage allowances of 14 to 20 kg of dry matter/cow per day (inclusive), diminishing at the highest allowance of 25 kg of dry matter/cow per day. At the lower forage allowances, energy intake would have been a limiting factor for milk production, potentially highlighting the nutritive advantages of plantain and white clover. Increasing species diversity is a possible strategy for overcoming seasonal nutritive challenges present in perennial ryegrass monocultures.

We tested for a forage allowance effect on the milk yield of early lactation dairy cow herds grazing swards sown with perennial ryegrass (Lolium perenne L.), white clover (Trifolium repens L.) and plantain (Plantago lanceolata L.) relative to perennial ryegrass alone. The examined allowances consisted of offering 12, 14, 16, 18, 20 or 25 kg of dry matter (DM)/cow per day of grazeable herbage, with diverse swards sown as mixtures and spatially adjacent monocultures. After adapting cows to their assigned forage type for 8 days, treatment effects on milk yield and composition, blood metabolites (beta-hydroxybutyrate, non-esterified fatty acids and urea concentrations), body weight change, forage intake and selection differentials for forage species and certain nutrients were monitored over 7 days. We confirmed a forage allowance effect on milk yield improvements in dairy cows grazing diverse swards relative to perennial ryegrass monocultures. Improvements in milk yield were evident at forage allowances of 14 to 20 kg of DM/cow per day, diminishing at the highest allowance of 25 kg of DM/cow per day. Improvements in milk yield for the mixture and spatially adjacent monocultures peaked at forage allowances of 18 and 16 kg of DM/cow per day, equalling increases of 1.3 and 1.2 kg of milk/cow per day, respectively.

1. IntroductionSouth-eastern Australian and New Zealand dairy systems are predominantly grazing-based [1], with perennial ryegrass (Lolium perenne L.) being the primary sown forage species [2]. A disadvantage of the existing perennial-ryegrass-based feedbase is its strong seasonality of growth and nutritive value [3,4]. Up to 60% of annual forage growth occurs in spring [5], which is associated with the transition from high-nutritive-value vegetative growth to low-nutritive-value reproductive development [6,7]. Spring forage exceeding herd requirements is often conserved as silage or hay, while paddocks retained in the grazing rotation often require pre- or post-grazing mechanical defoliation to achieve correct stubble heights for optimum forage quality and growth [8]. The disadvantages of these strategies include forage conservation costs [9,10] and dry matter (DM) losses [11,12]. An infrequently used strategy to reduce seasonality in nutritive value and to shift growth outside of spring is incorporating alternative perennial forage species into the perennial-ryegrass-based feedbase [13]. The success of this strategy depends on the selection of suitable species for the specific site and situation, rather than simply increasing sward diversity per se [14,15]. Both forb and legume species may have a role in achieving these goals. Milk yield responses to the inclusion of forbs and legumes into perennial-ryegrass-based swards have ranged from being negligible [16,17,18] to increases of 1.3 [19], 1.4 [20], 1.7 [21] and 2.0 kg/cow per day [22]. Differences in forage allowance may explain some of this variability [14], with milk yield improvements restricted to experiments allocating < 25 kg of DM/cow per day of grazeable forage (i.e., forage above a target post-grazing forage biomass residual). Understanding this potential interaction is required to identify the forage allowances at which the inclusion of forbs and legumes into perennial ryegrass-based swards benefit milk production.A forb with significant potential for many temperate dairying regions is forage-type plantain (Plantago lanceolata L.). Not only can plantain achieve equal or greater levels of annual forage production than perennial ryegrass [23,24], but owing to its superior tolerance of hot and dry conditions [25,26], plantain can be especially beneficial in increasing summer–autumn growth [24,27,28]. Relative to perennial ryegrass, plantain herbage typically has a lower structural fibre content [29], which can permit individual cow DM and resultant metabolisable energy (ME) intake to be elevated sufficiently to increase milk production [30]. Plantain herbage is also known to be mineral-rich [31,32,33] and have a negative dietary cation–anion difference [34], which can reduce the incidence of the economically important metabolic disorder hypocalcaemia (milk fever) [35]. Aside from these obvious production advantages, incorporating plantain into swards can have the added benefit of reducing environmental nitrogen (N) losses [22,36].Improvements in forage nutritive value (i.e., lower fibre and higher digestible protein content) and DM intake can also be achieved by incorporating white clover (Trifolium repens L.) into perennial-ryegrass-based swards [37,38]. The realisation of these advantages into increased individual cow milk production necessitate that white clover constitutes more than the typical 10–20% of grazed forage on a DM basis [38,39,40]. Achieving higher white clover levels in swards has the added environmental advantage of increasing biological N fixation [41,42], thus reducing the need for fossil-fuel-based synthetic N fertilisers [43]. A disadvantage of elevating the white clover content in swards is the concomitant increase in bloat risk [44], which must be considered when analysing the cost–benefit of white clover.It is not yet clearly understood how best to incorporate forbs and legumes into the perennial-ryegrass-based feedbase. Incorporating forbs and legumes in a mixture/polyculture with perennial ryegrass provides each species with the chance to exploit niches within the sward [45,46], maximising their complementary aspects. Alternatively, species may be sown in spatially adjacent monocultures within the same field, which has the potential advantages of minimising interspecies competition and allowing for species-specific fertiliser and herbicide management [20]. A disadvantage of spatially adjacent monocultures is the potentially greater ability of grazing ruminants to exhibit a predilection for specific species, which in the case of legumes could exacerbate the risk of bloat.Our experiment tested the hypothesis of a forage allowance by forage type interaction effect on improvements in the milk yield of early lactation dairy cows grazing functionally diverse swards sown to perennial ryegrass, white clover and plantain relative to perennial ryegrass only. Both mixtures and spatially adjacent monocultures of these species were evaluated to test the implications of these methods in increasing sward diversity on milk production.2. Materials and Methods2.1. Ethical StatementAll procedures involving cattle were approved by the University of Tasmania Animal Ethics Committee (A0012629).2.2. Experimental Site DescriptionThis experiment was conducted over 15 d during mid spring (September/October 2014) using the More Milk from Forage (MMFF) experimental site [20] at the Tasmanian Institute of Agriculture Dairy Research Facility (41°08′ S, 145°77′ E; 155 m above mean sea level), Elliott, northwest Tasmania, Australia. The location is characterised by a cool–temperate climate and winter-dominant rainfall pattern (mean annual rainfall, 1200 mm). During the 15 d experimental period, mean maximum and minimum daily ambient temperatures were 1.7 °C and 2.0 °C warmer than the long-term 40 year average for this period, while mean daily wind speed was 0.33 m/s lower (Table 1). Mean daily relative humidity and total rainfall deviated minimally from the long-term average. Soil at the experimental site is classified as a clay loam red ferrosol (Humic Etrodox) soil [47,48].The MMFF field site consisted of ten discrete experimental areas (paddocks), ranging in size between 1.20 and 2.16 ha. Each paddock was divided into three plots of equal area, with each plot occupied by a well-established sward of one of three forage types and all forage types present in each paddock. Forage types had been randomly assigned to plots at sowing (April 2012) and consisted of a perennial ryegrass monoculture (PRG); perennial ryegrass, white clover and plantain mixture (RCPM); and spatially adjacent monocultures (SAM) of perennial ryegrass, white clover and plantain (Supplementary Figure S1). Each SAM plot contained a single strip of each forage species, with white clover always occupying the centre strip. All strips in a SAM plot were of equal width (mean ± SD; 14.3 ± 4.6 m) and ran parallel to each other along the full length of the plot. Cultivars used to construct forage types included perennial ryegrass cv. Base® (AR37, non-toxic endophyte), ladino-type white clover cv. Grassland Kopu II® and forage-type plantain cv. Ceres Tonic®. Further details regarding the establishment and management of forage types is provided in Pembleton et al. [20].2.3. Experimental Design and ProceduresThe experiment was conducted over 15 d and consisted of 72 early lactation dairy cows allocated equally between 18 treatments, with treatments including each combination of the three forage types (described above) by six forage allowance levels. Each treatment was allocated four early lactation dairy cows for the experiment duration, with previous research showing this lactation stage to be when the effects of forage type on milk production are greatest [20]. A treatment herd size of four was selected to give a high probability of detecting treatment effects (power = 0.8). Power analysis was conducted as outlined by Cohen [49] assuming a large treatment effect via the pwr.f2.test function in R (df1 = 17, df2 = 54, Cohen’s f2 = 0.35, p = 0.05) [50].Forage allowance treatments were imposed for the final 7 d of the experiment (response period), with cows offered 12, 14, 16, 18, 20 or 25 kg of DM/cow per day of grazeable herbage from their respective forage type. An 8 d pre-response period preceded the response period, when all cows were allocated 14 kg of DM/cow per day of their respective forage type to facilitate adaptation of the rumen. During the pre-response period, all cows assigned to a forage type were managed as one herd (i.e., three herds of 24 cows), whereas during the response period (when forage allowance treatments were imposed), cows were managed in their treatment herds (i.e., 18 herds of 4 cows) (Figure 1). After each daily milking event (0700 and 1500 h), cows received half of their daily forage allowance. During each of these grazing periods, all cows grazed within the same paddock, with portable electric fencing (Kiwitech Ltd., Bulls, New Zealand) used during the response period to divide forage type plots crosswise into forage allowance subplots. This resulted in the treatment herds grazing SAM subplots being offered equal areas of each species component. The ordering of forage allowance subplots in each plot was randomly determined, with forage allowances achieved by manipulating the size of the subplots provided to treatment herds between 72.5 and 374.4 m2.During the week preceding the experiment, each plot (n = 30) was yield-mapped at 2 m intervals using a C-Dax pasture meter integrated with a GPS console (C-Dax Ltd., Palmerston North, New Zealand). Manifold GIS software was then used to construct a yield map [51]. A day before cows grazed a new paddock, eight randomly placed 0.5 by 0.5 m quadrats were cut to ground level in the PRG plot, RCPM plot and each species component of the SAM plot. Cut material was removed and immediately weighed, with DM content (%) determined (for experiment management purposes only) by repeatedly drying a subsample of known weight in a microwave oven until constant weight was achieved. Pre-grazing forage biomass (kg of DM/ha) of each quadrat was then calculated and integrated into the yield map to determine the required subplot area for each treatment cow herd. Forage was allocated above a post-grazing residual of 1250 kg of DM/ha for the PRG, RCPM and perennial ryegrass monoculture component of the SAM; 500 kg of DM/ha for the white clover monoculture component of the SAM; and 800 kg of DM/ha for the plantain monoculture component of the SAM.2.4. AnimalsEach treatment herd consisted of four early lactation multiparous crossbred dairy cows (Bos taurus). Cows in each treatment herd were balanced for age (mean ± SD; 4.7 ± 0.5 years old), milk production (mean ± SD; 32.3 ± 3.9 kg/cow per day), days in milk (mean ± SD; 38.6 ± 3.7 d), body weight (mean ± SD; 497 ± 55 kg) and breed. An in-bail feeding system supplied each cow with 4 kg of DM/d of concentrate (wheat-based pellet with additional mineral and vitamin supplements). The daily concentrate allowance was equally split between morning and afternoon milking events, and consisted of the following: crude protein (CP), 11.2% of DM; neutral detergent fibre (NDF), 12.8% of DM; acid detergent fibre (ADF), 6.0% of DM; lignin, 1.8% of DM; water-soluble carbohydrates (WSC), 1.8% of DM; crude fat, 2.1% of DM; and estimate metabolisable energy (ME), 12.5 MJ/kg of DM. During grazing, each cow had ad libitum access to bloat blocks that contained 10% alcohol ethoxylate teric 12A 23 (Bloat-Liq, Olsson’s Industries, Melbourne, VIC, Australia). Furthermore, white clover monoculture components of the SAM treatment were sprayed immediately before grazing with 15 L/ha of an emulsifiable anti-bloat oil (BP Pasture spray, BP Australia Pty Ltd., Melbourne, Australia). All cows had ad libitum access to fresh water while grazing via Kiwitech portable 100 L water troughs (WT R100).2.5. MeasurementsAt each milking event, milk yield (kg) of each cow was recorded by a DeLaval Alpro milk metering system (DeLaval International AB, Tumba, Sweden). On the day preceding (covariate) and the final three days of the response period, milk samples were collected at both daily milking events. Milk samples were analysed by TasHerd Pty Ltd. (Haspen, Tasmania, Australia) using a Bentley B2000 Infrared Milk Analyzer (Bentley Instruments Inc., Chaska, MN, USA) for fat, true protein and lactose concentration. At each milking event (i.e., twice daily), automatic walk-over scales (DeLaval AWS100 automatic weighing system) recorded body weight as cows exited the milking parlour. Twice-daily weighing of cows permitted for the effect of changes in rumen fill on daily body weight measurements to be averaged. Blood samples were collected from each cow via coccygeal venepuncture using 10 mL Vacutainers® (Becton, Dickinson and Company, Plymouth, UK) containing sodium heparin. Samples were collected after morning milking events on the day preceding (covariate) and for each of the final four response period days. Blood samples were centrifuged (1125× g) for 10 min at 4 °C, with plasma then collected and frozen at −20 °C until laboratory analysis. Plasma was analysed by the Western Australian Department of Agriculture and Food Animal Health Laboratory (South Perth, Western Australia, Australia) using an Olympus AU400 Chemistry Analyser (Olympus, Tokyo, Japan) for blood urea nitrogen (BUN), beta-hydroxybutyrate (BHBA) and non-esterified fatty acid (NEFA) concentrations.Immediately pre- and post-grazing, forage biomass in each subplot was estimated from compressed forage height (mm), measured with an electronic rising plate meter (Farmworks Systems, Fielding, New Zealand). Compressed forage height values were converted to forage biomass on a hectare basis (kg of DM/ha) using separate linear regression equations for PRG and RCPM treatments, and for each species component of the SAM treatment [52]. Each calibration equation was developed in the week preceding the experiment by measuring the compressed height of forage contained in 80 randomly selected square quadrant (0.25 m2) samples. Forage in each quadrant was harvested to ground level and immediately weighed, with a subsample of known weight dried to constant weight at 60 °C in a fan-forced drying oven (Unitherm drying oven; S & T Engineering Company, Birmingham, UK). Dried subsamples were reweighed to determine DM content and calculate forage biomass (kg of DM/ha). Regression equations were then developed, with forage biomass as the dependent variable and compressed forage height as the independent variable (Supplementary Table S1).Forage intake from each subplot was estimated from the pre- and post-grazing estimates of forage biomass using the following equation:Estimated forage intake (kg of DM/cow per day) = (Mpre − Mpost)/4, where Mpre = pre-grazing forage biomass (kg of DM) and Mpost = post-grazing forage biomass (kg of DM). Daily estimated forage intake was the sum of estimated forage intake from morning and evening grazing periods (i.e., the two subplots allocated daily to the treatment herd). Predicted forage intakes were calculated empirically using forage nutritive value data and known energy requirements for maintenance, body weight change, pregnancy and production [53].Prior to grazing a paddock, hand-cut forage samples were collected from each forage type plot to determine the botanical composition and nutritive value of offered forage. In both PRG and RCPM plots, all forage above ground level was collected from twenty-five points (40 by 40 mm) along a defined transect. This sampling process was repeated for each species component of SAM plots. On the day preceding (covariate) and the final four days of the response period, forage samples were similarly collected from each subplot immediately pre- and post-grazing. Samples were used in conjunction with forage biomass data to calculate selection differentials exhibited by treatment herds for the sown forage species and certain nutritive value parameters, including CP, NDF and ADF. Botanical composition was determined by hand-separating forage samples into their individual botanical components (perennial ryegrass, white clover, plantain and other volunteer species (weeds)), which were dried at 60 °C for 72 h. Each dried botanical component was weighed to calculate botanical composition on a DM basis. Forage for nutritive value analysis was similarly dried and then milled through a 1 mm screen before being assayed. All nutritive value samples were analysed via wet chemistry procedures by the Dairy One Forage Laboratory (Ithaca, New York, NY, USA). In vitro DM digestibility (IVDMD) was determined using an ANKOM DaisyII Incubator (ANKOM Technology Corporation, Fairport, NY, USA), with dry-milled forage samples anaerobically incubated at 39 °C for 48 h in a medium containing Van Soest buffer solution and rumen fluid [54,55]. Rumen fluid was collected from fistulated lactating dairy cows, which were total mixed ration (TMR)-fed. Further details regarding these procedures and an estimation of ME are provided in Pembleton et al. [20]. Selection differentials for selected forage species and nutritive parameters were calculated as the ratio of the species or nutrient concentration in herbage consumed relative to the herbage on offer using the following equation, as described by Jacobs et al. [56]:Selection differential = Nsel/Npre where Nsel = [(Mpre × Npre) − (Mpost × Npost)]/(Mpre − Mpost), where Mpre = pre-grazing forage biomass (kg of DM), Mpost = post-grazing forage biomass (kg of DM), Npre = pre-grazing species component or nutrient concentration (g/kg DM) and Npost = post-grazing species component or nutrient concentration (g/kg DM).2.6. Statistical AnalysisEffects of forage allowance and forage type on milk yield and composition, body weight change, blood metabolite levels and estimated forage intake were examined using multiple regressions analysis. Separate regression models were developed for each measured parameter, with the choice of regression manually pre-set based on graphical observations of the data. Forage allowance was included in all models as a continuous predictor (explanatory) variable, with individual cows being the unit of analysis. Forage type was only included in models as a categorical predictor variable if it statistically significantly improved the variance accounted for by the model. This was determined by an ANOVA comparison of models developed for each measured parameter with and without the inclusion of forage type. Inclusion of forage type in the model resulted in the generation of a separate regression equation for each forage type. Despite using historical milk production records to balance treatment herds, significant treatment herd differences in milk yield and composition were detected in data collected immediately before commencement of the experiment. Consequently, these values were used as covariates prior to fitting multiple regressions to these measurements.Pre- and post-grazing forage biomass and selection differentials were analysed via split-plot ANOVA, with forage type the main plot, forage allowance the subplot and the monitored response days treated as blocks. The split-plot ANOVA statistical model was as follows:Yijk=μ+αi+γk+ηik+βj+(αβ)ij+εijk in which Yijk is the measured parameter (forage biomass or selection differential), αi is the fixed effect of forage type, γk is the fixed effect of response day (block), ηik is the whole-plot error, βj is the fixed effect of forage allowance, (αβ)ij is the interaction between forage type and forage allowance and εijk is the split-plot error.All statistical analyses were undertaken using the R statistical package (R Core Team 2014). Unless otherwise stated, differences discussed are significant at p < 0.05.3. Results3.1. Botanical Composition and Nutritive ValueForage offered to cows grazing PRG and the perennial ryegrass component of the SAM treatment was greater than 97% perennial ryegrass, with the remainder being weeds (Table 2). Forage offered in the RCPM treatment was approximately two-thirds perennial ryegrass, one-third plantain, with white clover and weeds being only minor components (1.6 and 4.1%, respectively). While the forage offered in the plantain component of the SAM treatment was primarily plantain (77.8%), there was a sizable weed component (17.4%). Weeds contributed 52.3% of the offered forage in the white clover component of the SAM treatment.Analysis of the nutritive content of offered forage (Table 2) revealed that herbage obtained from the white clover component of the SAM treatment contained the lowest fibre (NDF and ADF) concentrations and the highest available protein concentration, IVDMD and estimated ME content. The forage offered in PRG and the perennial ryegrass component of the SAM treatment had the highest NDF concentration. Compared to these perennial ryegrass monocultures, the forage offered in the RCPM treatment had lower NDF and higher lignin and WSC contents. This reflected the inclusion of plantain in the RCPM treatment, as evidenced by the nutritive value of the plantain component of the SAM treatment. High WSC levels resulted in the forage offered in the RCPM and plantain component of the SAM treatment having the highest WSC:CP ratios (≥0.61). White clover contributed minimally to the nutritive value of the RCPM treatment, as it only represented 1.6% of the offered forage. Of all forages, the plantain component of the SAM treatment had the lowest estimated ME content, reflecting its higher ash and lignin content and resultantly lower IVDMD than all other forages. Supplementary Table S2 contains additional details on the chemical composition of each forage type.3.2. Milk Yield and CompositionAcross all treatments, the milk harvested during the experiment consistently contained 4.8% fat, 3.0% protein and 5.1% lactose. Milk yield increased as forage allowance increased (Figure 2). At p = 0.08 significance level, we observed a forage type effect on this relationship. Fitted regressions predicted that the milk yield of cows grazing RCPM and SAM treatments would peak at forage allowances of 24.5 and 25 kg of DM/cow per day, respectively. At the highest tested forage allowance of 25 kg of DM/cow per day, the milk yield of cows grazing the PRG treatment was still increasing. At forage allowances between 14 and 20 kg of DM/cow per day (inclusive), cows produced up to 1.3 and 1.2 kg/cow per day more milk when grazing RCPM or SAM treatments relative to the PRG treatment, respectively. The greatest difference in milk yield between RCPM and SAM vs. PRG treatments occurred at forage allowances of 18 and 16 kg of DM/cow per day, respectively.3.3. Blood Metabolite LevelsBlood NEFA concentration decreased as forage allowance increased (Figure 3A), with this relationship being independent of forage type. In contrast, blood BHB concentrations were independent of both forage type and allowance (Figure 3B), averaging 0.57 mmol/L. The concentration of BUN increased as forage allowance increased (Figure 3C), with the rate of increase consistent across all forage types. At each forage allowance, BUN concentrations for cows grazing the RCPM and SAM treatments were 0.591 and 1.036 mmol/L higher than their contemporaries grazing the PRG treatment, respectively. 3.4. Body Weight ChangeThe body weights of cows grazing RCPM and SAM treatments did not decline (Table 3). At forage allowances ≤ 18 kg of DM/cow per day, the body weight of cows grazing the PRG treatment declined by between 0.4 and 0.7 kg/cow per day.3.5. Forage Intake and Selection DifferentialsWithin each forage type (including each species component of the SAM treatment), pre-grazing forage biomass was similar across tested forage allowances (Table 4). As forage allowance increased, cows left greater levels of post-grazing forage biomass. As would be expected from the target residuals, post-grazing forage biomass was lower in the white clover relative to the plantain component of the SAM treatment, with both components having a lower post-grazing forage biomass than the PRG, RCPM or perennial ryegrass components of the SAM treatment.The estimated forage intake was independent of forage type and increased at greater forage allowances (Figure 4A), although the rate of increase declined as forage allowance increased. The fitted regression predicted that the maximum estimated forage intake (14.4 kg of DM/cow per day) occurred at a forage allowance of 24.5 kg of DM/cow per day. Estimated forage intake averaged 4.5 kg of DM/cow per day less than the predicted forage intake (Figure 4A,B).Across each forage type, cows receiving forage allowances of 18, 20 or 25 kg of DM/cow per day exhibited greater selectivity for perennial ryegrass than cows allocated 16 kg of DM/cow per day (Table 5). In contrast, cows grazing RCPM and SAM forage types exhibited greater selectivity for plantain at lower forage allowances. Within RCPM and SAM treatments, cows consistently showed a selective preference for white clover (mean selection differential, 1.57).Calculated selection differentials for CP were independent of forage type or allowance (Table 6). Cows showed minimal selective preference for CP, as indicated by the mean selection differential across treatments being close to 1 (1.06). Cows grazing the SAM, relative to PRG treatment, exhibited a significantly greater selection against NDF. At forage allowances of 20 and 25 kg of DM/cow per day, cows grazing the RCPM and SAM treatments showed a greater selection against ADF than cows grazing the PRG treatment. At a forage allowance of 12 kg of DM/cow per day, cows again showed a greater selection against ADF when grazing the SAM relative to the PRG treatment.4. DiscussionThis experiment confirmed our hypothesis of a forage allowance by forage type interaction effect on improvements in the daily milk yield of dairy cows grazing functionally diverse swards (RCPM and SAM) relative to perennial ryegrass monocultures (PRG). Such effects were independent of milk composition, which across the evaluated forage allowance by forage type treatments consistently contained 4.8% fat, 3.0% protein and 5.1% lactose. The experiment was conducted in spring, which is known from previous experiments to be the seasonal period when feeding evaluated diverse swards to spring-calving dairy cows has the greatest chance of increasing their daily milk production above that possible with perennial ryegrass only [20]. In the current experiment, such milk yield advantages of the diverse forage types were only evident at forage allowances resulting in below maximum estimated forage intake (i.e., forage allowances < 24.5 kg of DM/cow per day). Maximum daily milk yield improvements occurred at forage allowances of 18 kg of DM/cow per day for the RCPM and 16 kg of DM/cow per day for the SAM, equalling daily milk yield increases of 1.3 and 1.2 kg/cow per day, respectively. The advantages of the diverse forage types for body weight change followed a similar pattern to daily milk yield, with forage allowances ≤ 18 kg of DM/cow per day resulting in the cows grazing the PRG treatment losing weight, whereas cows grazing the RCPM and SAM treatments gained weight.The energy intake of cows with an estimated forage intake below maximum would have been a limiting factor for milk production, potentially highlighting the nutritive value advantages of the diverse forage types. Assuming that cows receiving the highest forage allowance of 25 kg of DM/cow per day ate to requirements, cows allocated between 12 and 20 kg of DM/cow per day had an average forage intake deficit of 2.4 kg of DM/cow per day. The results of this calculation are supported by blood NEFA concentrations, which were higher for cows allocated less forage, indicating a more significant lipolysis of adipose tissue to supply energy [57].Our study was not a highly controlled confinement-based feeding study (e.g., individual metabolic stalls), but instead compared the milk production potential of evaluated forage types when grazed to provided commercially relevant results for grazing-based dairy systems. A consequence of grazing systems is that cows may not always eat their full allocation and may exhibit a predilection for certain forage species present in the sward. At the lower forage allowances associated with the diverse forage types having a milk production benefit, cows grazing the RCPM and SAM treatments consumed relatively more plantain (selection differential data). While this could have simply resulted from cows grazing further into swards, where relatively more plantain may have been present, grazing at lower forage allowances would have resulted in plantain having a greater potential to influence cow production.A benefit of including plantain in the RCPM treatment was an increase in the WSC:CP ratio of forage above PRG levels (0.61 vs. 0.50), as plantain contributed 26.5% of offered forage DM in this treatment. Evidence is provided by the higher WSC:CP ratio of forage obtained from the plantain component of the SAM relative to PRG treatment (0.69 vs. 0.50). This may have improved the efficiency of N conversion into microbial protein, due to the improved synchrony of energy and protein within the rumen [58,59]. Indeed, the WSC:CP ratio for the plantain component of the SAM treatment was only slightly below the benchmark (>0.70) known to significantly improve ruminant N utilization efficiency [59,60]. Advantages of white clover were not evident in the RCPM treatment because of the minimal contribution of white clover on a DM basis (1.6%). Lower temperatures early in the growing season, as is typical for the experimental site, would explain the low prevalence of clover during our experiment [61]. This would also help to explain the high weed burden and lower-than-desired prevalence of clover in the white clover component of the SAM treatment (47.7% on a DM basis). The nutritive advantages of the white clover component of the SAM relative to PRG treatments thus cannot be attributed solely to white clover but included a higher estimated ME (11.5 vs. 11.3 MJ/kg of DM) and available protein content (23.1 vs. 15.8%).Across forage types, selection differential values for CP were similar and close to one, indicating that cows did not exhibit any selection for or against CP. An explanation is provided by BUN concentrations remaining within the expected range of 2–4 mmol/L for morning-sampled dairy cows in grazing systems with adequate CP intake [62]. Recorded BUN levels were lower than levels often prescribed in the literature for cows receiving an adequate CP supply because cows used in these studies were TMR-fed [63,64]. Unlike TMR-fed cows with regular access to feed, cows in grazing systems experience a period of fasting overnight due to forage depletion, which is well known to depress BUN concentrations [65,66].The inclusion of plantain and/or white clover in the diverse forage types resulted in offered forage having a lower fibre (NDF and ADF) content than the PRG treatment and permitted cows to select a diet lower in fibre. Reduced fibre intake is typically associated with an increased digestibility of consumed forage, decreased rumen retention time and a concomitant increase in voluntary forage intake [67,68]. Forage obtained from the white clover component of the SAM had a greater IVDMD than the PRG treatment, but the opposite was true for the plantain component. Forage in the RCPM relative to PRG treatment subsequently had a lower IVDMD due to containing a significant plantain component. These observations can be partly explained by the higher lignin and ash contents of plantain relative to perennial ryegrass forage [69]. However, higher lignin content in plantain forage may have been advantageous in maintaining the reservoir of buffering exchangeable cations in the rumen and the provision of coarse fibre necessary for rumen function [70]. The lower IVDMD of plantain forage may also stem from the rumen fluid used for this assessment being obtained from TMR-fed donor cows. Previous research showed that rumen fluid obtained from TMR vs. pasture-fed donor cows resulted in 19.4% lower IVDMD values for pasture samples [71]. The use of rumen fluid from unacclimated cows may have been particularly problematic for assessing the IVDMD of plantain herbage, as plantain forage contains antimicrobial compounds known to interfere with ruminal fermentation [26].Despite the white clover component of the SAM treatment having the highest IVDMD of all forages, the estimated forage intake remained independent of forage type. This observation contrasts previous research showing that legumes are often associated with higher forage intake than grasses [72]. This discrepancy may result from the considerable intratreatment variation for estimated forage intake (as indicated by the error bars on Figure 4A). Differences in predicted forage intake amongst the forage types and allowances (Figure 4B) ranged between 3.6 and 0.1 kg of DM/cow per day. Low differences would be challenging to detect using pre- and post-grazing biomass measures with a rising plate metre. Unfortunately, animal-based methods (e.g., N-alkanes or other tracers), while potentially more accurate, are generally unsuitable for use in diverse swards [73].Across forage types there was a diminishing rate of increase in estimated forage intake with increases in forage allowance. Similar responses were reported by Moate et al. [74] and Peyraud et al. [75], with our observations in line with the theoretical response of forage intake to forage allowance proposed by Minson [76]. Our observation of maximum estimated forage intake occurring at a forage allowance of 24.5 kg of DM/cow per day is on the lower end of the range reported in previous studies [74,75,76]. Many factors other than forage allowance influence forage intake, including supplementary feeding level, cow size and prevailing environmental conditions [77], preventing direct comparison between experiments. Sward structure and species composition are also known to affect forage intake by influencing both bite size and rate [77].5. ConclusionsThis experiment has confirmed our hypothesis of a forage allowance effect on improvements in the daily milk yield of cows grazing functionally diverse swards relative to perennial ryegrass monocultures. Advantages were evident at lower forage allowances, when energy intake was limiting for milk production, potentially highlighting the nutritive value advantages of these diverse forage types. The most significant milk production benefits occurred at forage allowances between 16 and 18 kg of DM/cow per day. These forage allowance levels align with current grazing practices and consequently the successful integration of either RCPM or SAM treatments into grazing-based dairy systems will not require changes in daily forage allowance.

animals : an open access journal from mdpi

10.3390/ani12040447

PMC8868392

Paratuberculosis, caused by Mycobacterium avium ssp. paratuberculosis (MAP), is widely spread among ruminants worldwide. After a long-lasting incubation period, infected animals suffer from chronic granulomatous enteritis. Economic losses are caused by premature culling, reduced milk yield and slaughter value in the dairy and beef industry, triggering attempts to control the disease. Paratuberculosis is a listed disease (category E), according to the European Animal Health Law, and intended to be monitored within the European Union. Evaluation of several herd-level monitoring approaches including the testing of environmental fecal samples to detect the infectious agent have been evaluated, proving environmental sampling to be a useful monitoring tool on herd level. This study comprises the application of environmental sampling within a herd prevalence estimation study in German dairy herds. Based on regional differences in the structure of livestock farming, Germany was divided into three regions where a representative number of farms were visited for sample collection. The results clearly indicate a different regional MAP herd level prevalence. The highest percentage of affected herds is found in the eastern part with large dairy herds, and the lowest in the south with the smallest average herd size. We conclude that the regional differences in MAP prevalence imply different approaches to control the disease.

On-farm environmental sampling is an effective method for herd-level diagnosis of Mycobacterium avium ssp. paratuberculosis (MAP) infection and between-herd prevalence estimation. So far, no prevalence study enrolling important livestock-farming regions has been conducted. As the structure of dairy farming differs between main livestock-farming regions in Germany, our objective was to assess the between-herd prevalence of paratuberculosis for these regions in a standardized approach. Methods: In total, 457 randomly selected dairy farms from three regions of Germany (North: 183, East: 170, South: 104) were sampled between 2017 and 2019. Environmental samples (boot-swabs, aggregate feces and/or liquid manure samples) were cultured and analyzed using an IS900-qPCR for MAP determination. Of the 457 selected farms, 94 had at least one MAP-positive environmental sample with significant differences between regions regarding the apparent (North: 12.0%, East: 40.6%, South: 2.9%) or corrected true (North: 14.8%, East: 50.1%, South: 3.6%) between-herd prevalence. In conclusion, regional differences of between-herd prevalence of paratuberculosis are substantial in Germany, indicating the need for control approaches with different aims. Taking into account regional MAP prevalence, MAP-control programs should focus on on-farm prevalence reduction or on mitigating the risk of between-herd transmission, depending on region.

1. IntroductionParatuberculosis, or Johne’s disease, is a bacterial infection of the small intestine caused by Mycobacterium avium ssp. paratuberculosis (MAP). MAP is an acid-fast rod-shaped organism, which can survive in the environment for several months [1]. The infection is mainly present in farmed cattle, sheep and goat around the world but is also found in wild living species such as deer. Fecal shedding of MAP occurs in infectious animals, whereas oral uptake of MAP from the environment by susceptible animals leads to transmission of the disease [2]. Due to the long incubation time and the slow disease progression, early identification of infected animals is difficult. Infected animals start shedding MAP intermittently before clinical symptoms are present [2].On an animal level, paratuberculosis can be diagnosed using ELISA for antibody detection in milk and blood samples or by using PCR and culture methods to detect MAP DNA in fecal samples [2]. In addition, for MAP herd status determination, environmental sampling has been described as an easy, cheap and effective sampling strategy [3].Paratuberculosis has been described as early as 1895 by Johne and Frothingham [4]. It has spread around the world with reported herd-level prevalence between 7% and 83%, depending on the production type of cattle herds studied and the diagnostic methods used [5,6]. In general, herd-level prevalence tends to be lower in beef herds compared with dairy herds and for dairy herds it has been found to be positively related to herd size [7,8]. The comparison of MAP-prevalence between studies should be interpreted with caution when determined by different diagnostic methods. So far, data to compare MAP-prevalence between German regions using the same test approach are currently missing. These knowledge gaps hamper prevention and control of paratuberculosis [9]. Nevertheless, based on either laboratory or non-laboratory data herd-level prevalence was estimated to be above 40% in dairy cattle in most developed countries. In addition, prevalence is believed to be underestimated in most countries mainly because of the type of tests used in combination with a lack of monitoring [8]. Paratuberculosis is a listed disease according to the European Animal Health Law (category E), and intended to be monitored within the European Union. Diagnostic methods for monitoring are not defined yet (Regulation (EU) No. 2018/1882). In Germany, a passive monitoring system is in place where MAP-positive results have to be reported to the authorities under certain conditions. However, so far, there is no active, monitoring program to facilitate regular animal testing, and a prevalence study using a standardized approach is missing. Therefore, available prevalence estimates for federal German states can only be compared with caution. Estimates for apparent herd-level prevalence data based on different study designs and test methods exist for a few federal states, varying from 8.9% in Hessia up to 30.2% in Thuringia and Saxony [7,10,11,12]. The available data suggest regional differences in MAP prevalence but due to different study designs and especially because of different test approaches, numbers should be interpreted cautiously. Nonetheless, marked regional differences of livestock farming throughout Germany [13] indicate that different regional between-herd MAP prevalences can be expected. The objective of this study was to determine between-herd MAP prevalence in the three main livestock-farming regions of Germany. Participating farms were identified using a stratified sampling procedure. 2. Materials and Methods2.1. Herd SelectionWithin the framework of a joint project (PraeRi study) of the University of Veterinary Medicine Hannover Foundation, the Faculty of Veterinary Medicine of Freie Universität Berlin and the Faculty of Veterinary Medicine of Ludwig-Maximilian-Universität Munich dairy farms were visited, and animal health, biosecurity and animal housing were scored [14]. Participating farms were located either in Schleswig-Holstein and Lower Saxony (region north [RN]), in Brandenburg, Mecklenburg-Western Pomerania, Thuringia, Saxony-Anhalt (region east [RE]) or in Bavaria (region south [RS]) (Figure 1). Enrolled farms were randomly selected for the PraeRi study from the German database on animal identification and registration (HI-Tier) [14,15] and stratified according to herd size. Participating farmers of the PraeRi project were invited to participate in the MAP prevalence study (PraeMAP) by written consent. Participation in both studies was voluntary. Three research teams based in either Hanover, Berlin or Munich performed farm visits. In RN and RE environmental samples for this study were collected during the initial farm visit. In RS sampling on participating farms was performed by veterinarians of the Bavarian Animal Health Service during a separate farm visit. Between July 2017 and July 2019, the research teams contacted farmers and conducted the on-farm visitation and sampling. Due to an a priori training in environmental sampling, further permanent training and the usage of standard operation procedures (SOP), observer and sampler effects were considered negligible during the study.2.2. Sampling MethodsEnvironmental sampling was performed using boot-swabs, slurry and environmental fecal sampling [11,16,17], as laid down in a written SOP. In short, boot-swabs were collected by walking between 100 and 200 steps in high cow traffic areas, e.g., waiting pen and main cow alleyway as described before [16]. After walking, the boot-swabs were placed in an aseptic plastic bag and labelled. Slurry was mixed before a sampling container was submerged in the liquid storage area to collect a sample of 150 mL slurry. Environmental fecal samples were collected by sampling 5 to 10 locations in high cow traffic areas. Boot-swab and slurry samples were the preferred way of sampling. Whenever a slurry sample was not available, an environmental fecal sample was collected. After collection, samples were stored at −20 °C with a maximum of three months at the respective university until transportation to and analysis at the laboratory of the Thuringian Animal Health Service in Jena (RN and RE) or at the Bavarian Animal Health Service in Poing (RS).2.3. Laboratory AnalysisSamples were submitted to fecal culture for detection of viable MAP as well as to IS900 real-time PCR or detection of direct MAP DNA. Before analysis, boot-swabs were washed in NaCl to extract fecal material as described before [17]. Supernatants of the boot-swabs, slurry and environmental samples were analyzed by fecal culture according to the official guidelines of diagnostic procedures published by the Friedrich-Loeffler-Institut (FLI), the German Federal Research Institute for Animal Health [18,19]. Characteristic colonies were heated and sonicated according to a standardized method recommended by the FLI for DNA isolation [18] and confirmed by a conventional IS900 PCR [20]. Only samples with typical grey-white colonies with mycobactin-dependent growth, and a positive IS900 PCR were characterized as MAP-culture-positive [18]. In addition, a commercial IS900 real-time PCR was applied on each fecal sample for direct detection of MAP without a preceding cultivation step. For direct DNA extraction (QIAmp DNA Mini kit, Qiagen, Hilden, Germany) and sample pre-concentration (Adifil 100, Adiagene, Bio-X Diagnostics S.A., Rochefort, Belgium) for MAP DNA detection commercial kits were used according to the manual provided by the manufacturer. For MAP DNA detection a commercial IS900-based real-time PCR method was used (Adiavet Paratb, Adiagene, Belgium) according to the manufacturer’s instructions. Farms were categorized as MAP-positive when at least one sample had either a MAP-positive culture or direct PCR outcome or both.2.4. Data AnalysisData analysis was performed using the statistical software package SAS [21] and EPITOOLS [22]. Farms were categorized as MAP-positive if at least one of the collected samples tested MAP-positive. Apparent between-herd prevalence (AP) for each region was calculated. Herd-level sensitivity of the MAP-detection method was determined by applying fecal culture and/or the Adiavet Paratb qPCR test on a set of a boot swab and a slurry sample derived from a data set of a previous study [23] as follows: fecal culture—0.72%, real-time—PCR 0.78%, both tests combined—0.81%, and herd-level specificity of fecal culture was 1.00. The specificity of the Adiavet Paratb PCR test was determined using the validation data given in the manual of the manufacturer (1.00). True prevalence estimates (TP) [24] and 95% confidence intervals for AP and TP [25] were calculated for each region. In order to account for sporadic false positive results that may be due to pre- or postanalytical sample handling in practical application, we performed a sensitivity analysis for the calculation of the TP using the specificity estimates of 0.99 and 0.98.3. Results3.1. Study FarmsA total of 457 (RN = 183; RE = 170; RS = 104) farmers agreed to participate in the study. Farm characteristics differed per region (Table 1). Median herd size between regions differed substantially with the largest farms in RE and the smallest farms in RS. Holstein Friesians represented the typical breed in farms located in RN and RE, whereas in RS, Simmenthal was the main breed. The percentage of cattle housed in free stalls was highest in RN compared with RE and lowest in RS. Results regarding animal health in general, biosecurity and animal housing will be published elsewhere.3.2. Boot-Swabs and Environmental Samples In total, results of 367 environmental samples were available of RN, 341 of RE and 205 of RS. An overview regarding the detection of MAP DNA and viable MAP per kind of environmental sample is given in Table 2. As expected, for most samples, both detection methods showed positive results (column ‘both’) and only for a minority of samples’ PCR and culture results were not congruent (column ‘PCR pos’ and ‘culture pos’).3.3. Between-Herd Prevalence Apparent MAP between-herd prevalence differed markedly between regions (Table 3). In the eastern part of Germany 40.59% [Confidence Interval (CI) 33.21; 47.97] of herds were detected MAP-positive whereas in the southern region only 2.88% [CI 0; 6.09] were MAP-positive. The true apparent prevalence was only slightly higher compared with the AP (Table 3). However, the assumption that test specificity in field samples is lower (0.98) than 1 led again to minor changes in TP. The corrected between-herd TP of RE decreased to 48.9 [39.5; 58.2] and in RS to 1.1 [0; 5.2].4. DiscussionThis study aimed at estimating the prevalence of MAP-positive dairy herds in three main livestock-farming German regions using an environmental sampling protocol as an easy-to-use and cost-effective sampling approach [26]. Both fecal culture and a MAP IS900 real-time PCR were used to detect the presence of viable MAP and/or MAP DNA, as previously reported [19,23]. The need of this study resulted from missing a prevalence study regarding paratuberculosis at national level where farms were selected using a stratified selection process and the same protocol for sampling and analysis. The three chosen regions relate to the three main livestock farming regions (North, East, South) as determined by Merle et al. [13], which represent the structural diversity of the German dairy industry. Sharing the sampling frame of the joint project PraeRi enabled us to perform a study with uniform stratified random farm selection ensuring representativeness for each region. As expected, a clear north–south divide of MAP between-herd prevalence was detected, with more affected farms in the northern part of Germany than in the southern part, and an excessively high percentage of MAP-positive herds was identified in the eastern region. These regional differences of MAP prevalence were expected because of the regional diversity of farming [13]. In addition, existing regional studies with convenient herd selection and sampling as well as different diagnostic methods had already pointed in that direction. However, comparison of those data had to be performed cautiously due to the different approaches. For example, by testing pooled milk for MAP-specific antibodies within the mandatory MAP-control program of Lower Saxony an apparent prevalence of 20% was calculated [12], which is slightly higher than the point estimate of TP determined here, but lies within the 95% CI. A Bavarian study reported a higher herd-level seroprevalence, namely 12% (95% CI: 5–15%), using an ELISA test with a specificity of 99% compared with 3.56% between-herd prevalence established here, presumably due to sampling bias of the convenience sampling in the former study [27]. Based on our results, we assume that the differences between our results and results reported by the former study may be caused by the limited specificity of the used tests and the missing calculation of TP. Another reason for overestimating prevalence might have been convenience sampling with an overrepresentation of affected herds, because those herds may be more willing to test their herd and therefore introducing bias. This underlines the value of the uniform stratified random farm selection performed in our study. Two studies based on serological testing were published for regions in the eastern part of Germany reporting an apparent between herd-level prevalence of 85% or 90% [28,29]. These findings were also biased by low specificity of the applied ELISA test with 77% or 95%, respectively. Furthermore, these studies used data from voluntary control programs in the respective regions. As mentioned above, the affected herds are more likely to be enrolled in voluntary control programs. Both effects may have led to an overestimation of herd-level prevalence in these studies. A study [7] based on environmental sampling (boot swab) and applying a commercial real-time PCR reported 30% MAP-positive herds resulting in herd-level TP estimates of 56% for the federal state Thuringia and 34% for Saxony. Nonetheless, the results presented here are in line with those of previous studies in Germany and confirm the suitability of environmental sampling to estimate the herd-level prevalence of paratuberculosis. Using the same diagnostic approach, prevalence studies at national level were performed in the United States of America (USA) and Canada as well, substantiating a true between-herd prevalence of 91% for USA and 24–66% for dairy herds in different Canadian regions [30,31]. Herd size has been identified to be positively related to the risk of a MAP-positive herd-status [7,10,31,32]. This is assumed to be a relevant factor for the difference between the regions in our study, where the southern region with the lowest mean herd size had the lowest herd-level prevalence, and MAP was most present in herds of the eastern region with a median herd size of 199 cattle. This effect may be explained by the higher propensity of large herds to buy cattle, or a denser occupation of the calving pen may play a role, which is also linked to within-herd transmission of MAP [33,34]. Furthermore, keeping dairy cows in free stalls is positively associated with a positive MAP status when compared with tie stalls [10,31], which again is more common in small herds in southern Germany compared with large herds in the eastern part. Based on the fact that we used the same sampling scheme and MAP detection method in all three regions, we assume a good internal validity of our study. Between-sampler variability was minimized using written SOP protocols for collecting slurry and environmental samples and for the use of boot swabs. Evaluating inter-observer reliability was not considered because the additional herd-level data used in this study to describe the different structure of the dairy industry in the respective regions (herd size, breed, housing) was easy to record. Furthermore, categorizing these traits in dichotomous variables (German Holstein breed or not, free stalls or not) reduced the risk of misclassification. Due to sample management samples collected in this study were stored at −20 °C up to three months before analysis. It has been discussed that MAP viability might be reduced by longtime freezer storage and repeated freezing and thawing cycles. Since we limited storage time to a maximum of three months and samples were only thawed once before analysis the influence on MAP survival can be neglected [35]. With respect to the laboratory methods, a high level of quality assurance is established in both laboratories that performed the fecal culture and the real-time PCR test. Both laboratories are accredited veterinary test laboratories under license of the German Accreditation Body according to the quality standards of the German and European Standard DIN EN ISO 17025. The standard includes ring trails and comparative test between laboratories. The same commercial real-time PCR test system was applied in both laboratories, and fecal culture was performed according to the official method as laid down in the manual of official test methods published by the Friedrich-Loeffler-Institut, the German federal research institute of Animal Health [18]. The question whether real-time PCR positive samples contained viable or non-viable MAP could not be answered in this setting. However, since the real-time PCR results as well as the culture results were used to determine the MAP status of the particular herd, and results were not used to determine the role of an infected environment on MAP transmission, the applied analysis is valid. Since test results were interpreted in parallel, test sensitivity was increased; therefore, we assume a high validity of the sensitivity for the whole sampling and testing approach used for the calculation of true prevalence. It was calculated from the data of a previous study using the same sampling approach and the same culture and PCR method [23]. The reference method in that study was a whole-herd testing of all cows by individual fecal culture. Analytical specificity as given by the manufacturer of the commercial PCR test was 1.0, which is not necessarily the diagnostic specificity. To account for a reduced diagnostic specificity due to pre- or postanalytical sample handling or laboratory contamination, we performed a sensitivity analysis using the values of 0.99 and 0.98, which is a common routine when using diagnostic tests, resulting in TP estimates, which were in the same range as calculated with a perfect specificity. Therefore, we conclude a high validity of our TP estimation. The nearly 80% difference between AP and TP reflects the 81% herd-sensitivity of the test approach when perfect herd-specificity is presumed.The stratified random farm selection method used in this study ensured an acceptable level of external validity, i.e., representativeness for each region involved. The prevalence estimates presented here are valid for the respective target population, which are the dairy herds in the respective region. Although some of the respective regions with the relevant dairy industry were missing (e.g., the federal states North Rhine-Westphalia in the northern region, Baden-Württemberg in the southern region and Saxony in the eastern region), the provided prevalence estimations for the respective region are valid. The agricultural structure is comparable in the missing federal states of each region; therefore, results can be generalized for each particular region. Due to the limited number of herds, we have to accept the remarkable confidence intervals of the AP and TP estimators. Nonetheless, there is no overlap of the confidence intervals; therefore, they allow a clear distinction of different levels of between-herd prevalence. This supports our main conclusion that regions differ in their between-herd prevalence regarding MAP. Furthermore, this finding discourages the calculation of an average ‘national’ between-herd prevalence across the regions. This would even out MAP prevalence and thus not reflect the real situation regarding paratuberculosis infection in German dairy herds. A minor limitation of our study is the absence of two other relevant agricultural regions in Germany according to Merle et al. [13]. Since both regions comprise only small livestock populations with little relevance for the dairy sector in Germany these regions were not included. For the central region, a study was performed in Hesse in 2017 [10]. Boot swabs and PCR testing were used, but farms were not selected by stratified random sampling. Results showed an AP of 8.9% [10], which is between the determined MAP prevalence of RS and RN. Another limitation of our study arises from the disproportionality of sample size and number of herds in the study population. Although the southern region is characterized by a large number of small-sized herds, the eastern region includes a small number of large herds, with the northern region in between. Applying approximately the same sample size to these regions might result in an unequal variance for quantitative estimates, which could affect statistical comparison. However, the goal estimating the between-herd prevalence of different German regions was unaffected and could be reached by using the same methodology. Although the sample size was rather small, data are representative for the situation in each region since they were selected according to a random stratified sampling protocol per region. These estimates provide a robust basis for further epidemiological assessment regarding paratuberculosis. At the same time, they represent a starting point for future research or even national active monitoring on MAP prevalence in dairy cattle in Germany. 5. ConclusionsFrom our results, we conclude that different approaches are needed to deal with the disease in regions with different between-herd prevalence ranging from on-farm prevalence reduction to mitigating the risk of spreading between herds. Measures for prevalence reduction and repeated herd status monitoring for success control are important for regions of high prevalence (east and north). Regions of low prevalence (south) should use those data to monitor MAP prevalence. Strategies reducing between-farm transmission of MAPs are important for all regions. Furthermore, a nationwide active monitoring would help to identify suitable control measures for all dairy production regions in Germany.

animals : an open access journal from mdpi

10.3390/ani13091470

PMC10177615

The yak is a unique livestock species living in the Qinghai–Tibet Plateau. Investigating the morphological differences among different breeds of yaks is of paramount importance. However, due to the lack of effective communication of yak genetic information under natural and artificial selection, the genetic diversity of regional yaks is not effectively utilized for new breedings, and it is difficult for the existing analysis models to analyze such complex multi-species populations. Therefore, we extended the application scope of the current statistical model to perform whole-genome association analysis on multiple yak breeds and identified four genes significantly associated with body height. The findings of this study are of great significance for the development and improvement of yak morphological traits, as well as the expansion of statistical models.

Yaks have evolved several breeds or genetic resources owing to their geographical and ecological environment, and investigating the genetic construction of body size among breeds is key for breeding. Here, a genome-wide association study (GWAS) was performed for five body size traits in 31 yak breeds and genetic resources. The information from clustering individuals according to their habitats was used for kinship grouping in the compressed mixed linear model (CMLM). We named this approach the pCMLM method. A total of 3,584,464 high-quality single nucleotide polymorphisms (SNPs) were obtained, and six markers were found to be significantly associated with height by pCMLM. Four candidate genes, including FXYD6, SOHLH2, ADGRB2, and OSBPL6, were identified. Our results show that when CMLM cannot identify optimal clustering groups, pCMLM can provide sufficient associated results based on population information. Moreover, this study provides basic information on the gene localization of quantitative traits of body size among yak breeds.

1. IntroductionYak (Bos grunniens), a unique large livestock species of the Qinghai–Tibet Plateau and surrounding Hengduan Mountains, provide a basic resource for the livelihood of plateau farmers and herders [1]. Due to different geographical and climatic environments, ecological conditions, grassland types, feeding levels, breeding levels, and social and economic structures in the main producing areas, China has formed 12 yak breeds: Qinghai Plateau yaks, Gannan yaks, Tianzhu White yaks, Bazhou yaks, Zhongdian yaks, Jiulong yaks, Muli yaks, Maiwa yaks, Niangya yaks, Xizang Alpine yaks, Pali yaks, and Sibu yaks [2]. Tibet is one of the main yak-producing areas in China accounting for 30% of the total number. Yaks are distributed in 71 counties in Tibet, forming many local yak breeds and populations, including the Niangya yak, Pari yak, Sibu yak, Sangsang yak, Sangri yak, Baqing yak, Dingqing yak, Kangbu yak, Jiangda yak, Leiwuqi yak, and Gongbujiangda yak [3]. This rich genetic diversity reflects adaptation to the external environment and is a vital genetic resource for breeding new breeds or strains. However, due to the lack of active communication of yak genetic information during natural and artificial selection, obtaining differential genetic information of different yak breeds and genetic resources has become very important. The yak body size trait is the most critical genetic index. It is an essential reference index to determine meat production performance and one of the most direct breeding selection parameters [4]. Some yaks are more aggressive due to mixing with wild blood, and extracting relevant biological traits is more complicated. However, technology for extracting animal body size traits by image recognition has become more advanced [5,6]. At present, digital images of beef cattle acquired by a Microsoft Kinect device can be used to establish model equations for predicting body weight, carcass weight, and body fat content; this has facilitated rapid and easy body size trait determination [7].With the development of high-throughput genotyping technology, opportunities have been provided for identifying new genetic variants associated with economic traits in cattle, where single nucleotide polymorphisms (SNPs) distributed throughout the genome have become the genetic markers of choice. Genome-wide association study (GWAS) is a common experimental approach to study SNP markers associated with various economic traits in animal production by linking phenotypic and genotypic data and using statistical models to investigate genetic variant loci causally associated with the target trait [8]. The GWAS approach has successfully revealed genetic determinants associated with disease susceptibility and resistance in humans [9], animals [10], and plants [11]. At the same time, the total genome data of yaks is 2.7 G, indicating genetic diversity among populations and the use of random mating within populations. This results in a large number of effective SNPs within the yak population. This will increase the computational burden. The compressed Mixed Linear Model (CMLM) was reported to improve the statistical power and computational speed of GWAS by clustering individuals into groups based on kinship among individuals [12]. When the likelihood values of the testing model in the CMLM are not able to identify the optimum clustering group, the individuals are grouped in a group with only one individual. This result has been proved by simulation results. However, previous population clustering [13] is widespread in animals, especially in yaks, and application of previous population structure is key for animal GWAS.In this study, we performed GWAS using yak datasets from different geographical areas in order to (1) detect SNP markers associated with body size among yak breeds, and (2) develop animal GWAS methods using the information provided on the populations.2. Materials and Methods2.1. Individual Samples and SequencingA total of 94 yaks were collected from different regions of the Qinghai–Tibet Plateau in China, including 17 Tibetan regions, 4 Qinghai regions, 4 Sichuan regions, 2 Gansu regions, 2 Xinjiang regions, 1 Yunnan region, and 4 wild yaks; in total, there were 31 yak breeds and genetic resources (Table 1). Detailed information on these yak genetic resources and their distribution areas are shown in Table S1. Population samples were obtained from the Key Laboratory of Qinghai–Tibetan Plateau Animal Genetic Resource Reservation and Utilization, Sichuan Province, and Ministry of Education, Southwest Minzu University. The sequence files of 94 yaks [13] were all obtained from the sequencing results of DNA extracted from blood samples using the Illumina Hiseq 2000 sequencer (Illumina, San Diego, CA, USA). Individuals of each yak (more than 2.5 years old, male) breed or genetic resource were measured for body height (BH, cm), body length (BL, cm), body weight (BW, kg), chest circumference (CC, cm), and circumference of the cannon bone (CCB, cm). The phenotypic values used for GWAS analysis in this study were replaced by the overall mean values of the measured phenotypes for each breed or genetic resource. This allowed for a better representation of breed characteristics [14].2.2. Genotyping Quality Control and FilteringData filtering was performed using the FASTP [15] software (version 0.20.1). Double-end sequencing reads were aligned using the Burrows–Wheeler Alignment Tool (BWA) [16] software (version 0.7.15), and high-quality reads were compared with the yak reference genome (BosGru3.0); the resulting BAM files were sorted using the sort command of the SAMtools [17] software (version 1.11) and de-duplicated using the rmdup command (all duplicate reads were removed directly). Local recombination of reads and comparison of near enhanced indel polymorphisms were performed using the Genome Analysis Toolkit (GATK) [18] software (version 4.0.1). Then the HaplotypeCaller command in GATK was used for SNP calling, CombineGVCFs command for VCF file merging, GenotypeGVCFs command for variant detection, and VariantFiltration command for initial filtering. The variant filtering conditions were set as follows: QD (QualByDepth, variant loci confidence divided by the number of unfiltered non-reference reads) < 2.0; FS (FisherStrand, Fisher exact test to assess the probability that the current variant is a strand bias, this value is between 0 and 60) > 60.0; MQ (RMSMappingQuality, square root of the matching quality in all samples) < 40.0; MQRankSum (MappingQualityRankSumTest, assesses the confidence based on the matching quality of the read of REF and ALT) < 12.5; ReadPosRankSum (ReadPosRankSumTest, evaluate the variation confidence by the position of the variation in the read, usually the error rate is higher at both ends of the read) < 8.0, and SOR (StrandOddsRati, comprehensive assessment of the likelihood of strand bias) > 3.0. Maker filtering using the PLINK [19] software (version 1.90) with the variant filtering conditions were set as follows: maf (minor allele frequency) 0.05, max-missing (maximum deletion rate of genotype) 0.05, and hwe (deviations from Hardy–Weinberg equilibrium) 1 × 10−6. The Genome Associated Prediction Integrated Tool (GAPIT) [20] was used for heterozygosity analysis of all markers.2.3. Population StructureThe neighbor-joining (NJ) tree was constructed using the P distance matrix calculated by the VCF2Dis [21] software (version 1.46), tree beautification was performed on the online site iTol (https://itol.embl.de/, accessed on 3 October 2022), and principal component analysis (PCA) was performed and plotted using the GAPIT package in R. Population clustering analysis was performed using the Admixture [22] software (version 1.30). Kinship and differentiation between samples from different regions were viewed by joint analysis, and linkage disequilibrium (LD) decay analysis was performed using PopLDdecay (version 3.41) [23].2.4. Association StudyGenome-wide association analysis was performed using a compressed mixed linear model (CMLM) [12] in the GAPIT (version 3.0) software, where PCA and the kinship matrix were added as covariates, p-values were corrected with Bonferroni, and the cutoff was set to 0.05/number of all markers. The general expressions of CMLM are consistent:(1)Y=Wv+SNPi+Zu+e where, Y is the phenotypic vector (n × 1); W is the covariate design matrix of vector v, and v is the corresponding coefficient vector, which is the non-marker effect among the unknown fixed effects (we used the effects of the first 3 PCs as fixed effects in this study); SNPi is the testing marker genotype; Z is the random design matrix (n × n) of u, where we re-defined the Z matrix as n × n′, n′ being the number of groups compressed, and u being the random effect vector of individuals, which obeys u~N(0,KVg), of which K is the n × n kinship genetic matrix (in this study, we used the group kinship matrix n × n′ to replace K); Vg is the additive genetic variance, and e is the random residual and obeys e~N(0,IVe), in which I is the n×n design matrix and Ve is the residual component.2.5. Population Index BuildingKinship in the model was calculated using all the markers. Combining test markers with kinship in MLM may lead to confusion between test markers and the genetic effects of individuals defined by kinship. In CMLM, Zhang et al. [12] introduced a variable now called the compress group, which clusters individuals with closer kinship into groups and uses the kinship between groups instead of the kinship between individuals for the operation. In this study, 94 individual yaks were artificially clustered into 7 populations based on their geographical relationships: Sichuan, Qinghai, Tibet, Gansu, Yunnan, Xinjiang, and wild yaks. We added a “compress_z” variable to the GAPIT built-in function GAPIT.Compress.R to store the real geographical groupings; this provides the groupings in CMLM in advance instead of the compressed groupings being estimated. We named this approach the provided compressed mixed linear model (pCMLM). The specific parameter compress_z under this method was marked with the number 1 to be identified as the same group of objects, and the output parameter group.membership was marked with the same number to be the same group. Currently, the parameter file for pCMLM has been uploaded to GitHub (https://github.com/liu-xinrui, accessed on 8 February 2023), which can provide a real or artificially defined grouping file for association study, as well as add grouping information to CMLM.2.6. Identification of Candidate GenesBased on the physical location of the target trait association loci on the yak reference genome and combined with the LD decay distance of the yak genome (~20 kb), associated genes were screened on both sides of the SNP loci; we adjusted the LD decay distance to 100 kb when no linked genes were detected. If the gene on the reference genome had only an Ensembl ID, the sequence of the gene was extracted for Basic Local Alignment Search Tool (BLAST) comparison using the wild yak reference genome (BosGru_v2.0) for functional analysis. For multiple significant SNP loci, haploid block mapping was performed using LDBlockShow [24] (version 1.40) for all SNPs within 100 kb upstream and downstream of the lead SNP, and a box plot of independent significant SNPs and phenotypic values of each individual within a block was created.3. Results3.1. Phenotypic DistributionWe analyzed five body size traits of 94 adult yaks and summarized the descriptive statistics (mean, variance, maximum, minimum, and coefficient of variation) for different body size traits (Table 2). Most records of the body heights ranged from 100 to 205 cm. The mean value of body height was 118.2 cm, the mean value of body length was 138.6 cm, the mean value of body weight was 284.9 kg, the mean value of the chest circumference was 168.8 cm, and the mean value of circumference of the cannon bone was 17.24 cm. All body size data of wild yaks were higher than those of domestic yaks in all regions. The t-test results showed that there was a significant difference (p-value < 0.05) between the body sizes of the domestic yaks and wild yaks. Among them, the body size trait of yaks in the Tibetan region was in the upper level among domestic yaks.3.2. SNP Calling and Population StructureA total of 47.15 million markers, including SNPs, indels, and other variants, were detected using the BWA-SAMtools-GATK pipeline [25] program with default parameters. Overall, 3,584,464 SNPs remained after filtering by the GATK and PLINK software; on average, they were distributed over 29 autosomes and 1 X chromosome, and the SNP density of most windows was >1 kb/Mb (Figure 1A). In addition, the heterozygosity of most individuals’ and SNP markers’ was low (Figure 1B,C).In this study, yak populations were derived from 31 yak breeds and genetic resources from different regions of the Qinghai–Tibetan Plateau, and consisted of several yak populations from several provincial areas of China (Figure 2A); moreover, this sample had complex population structures. To analyze the population structures of the 94 yaks, we performed PCA, population stratification, and NJ-tree analysis on seven yak populations (including six regions and wild yak populations) using 3.58 million high-quality SNP data points obtained through filtering. The two-dimensional scatter plot of PCA clustering showed that the population structure of the yak populations was relatively weak, and it was difficult to distinguish between population structures. After excluding outliers, the six populations were in a mixed state excepted for Tibetan individuals that could be roughly clustered into a population (Figure 2B). The genetic variance contribution explained by the first two principal components was 2.85% and 1.82% (Figure S1). The NJ-tree clustering results showed that most of the breeds or genetic resources shared a recent common ancestor with both the Tibetan and Gansu populations, and all individuals were not independent; these NJ-tree clustering results presented approximately the same population structure as PCA did (Figure 2C). The optimal CV value of 1 in admixture population structure analysis cannot accurately reflect the actual grouping of the population. To analyze yak breeds and genetic resources more accurately, we forced grouping of Admixture, and classified yak breeds and resources based on geographical population (seven populations in total). However, the results only showed that most Tibetan individuals could be clustered into one population, while the remaining six populations were in a mixed state where specific clustering situations could not distinguish (Figure 2E). The LD analysis showed that the breed-based yak populations had more rapid decay of LD and lower LD levels. The most immediate decay was in the Tibetan breeds or genetic resources, followed by the Sichuan yaks, Qinghai yaks, Gansu yaks, Xinjiang yaks, Wild yaks, and Yunnan yaks.3.3. GWAS and Candidate GenesThe CMLM model was used to associate loci with body size traits, and two statistical strategies were constructed. The first strategy was a tight grouping constructed by splitting all individual kinship matrices into several inter-group kinship matrices using CMLM, named the CMLM group. The second was CMLM using a prior population clustering parameter, named the pCMLM group. Comparing the negative twice likelihood (-2LL) of the two strategies for detecting BH traits, the results showed that the difference between pCMLM and CMLM was less than 10% of the mean value. Genotypic and phenotypic data of 94 individual yaks were analyzed using the GAPIT software with the first three principal components as fixed effects. Manhattan and quantile–quantile (QQ) plots are shown in Figure 3. After Bonferroni correction, six SNPs were found to pass the 5% threshold line (p-value < 1.39 × 10−8) and these were associated with BH only in pCMLM (Figure 3A), whereas no significant loci associated with body sizes were detected in any of the CMLM (Figures S2–S6). The QQ plot (Figure 3C) of pCMLM shows that all points corresponding to observed and predicted values were in the middle or above the diagonal line. Although there are some deviations, the model still has a meaningful corrective effect on population stratification relative to traditional CMLM. The likelihood value used to determine the best compression ratio in traditional CMLM was not significant. CMLM does not catch the best clustered groups and cannot reflect the true population structure; this could lead to a large number of false negative results (Figure 3B). The QQ plot of traditional CMLM with BH show most points below the diagonal (Figure 3C), which indicates that the observed p-value for most loci was less than the expected value and that the model overcorrected for this group. The significant SNPs detected in pCMLM were rs769892 (located at 4,883,046 bp) on chromosome 4, rs2659279 (located at 59,118,279 bp) and rs2659285 (located at 59,119,427 bp) on chromosome 13, rs310769 (located at 16,165,590 bp) and rs477265 (located at 57,505,237 bp) on chromosome 2, and rs2910497 (located at 129,838,648 bp) on chromosome 15. The significant SNPs with their associated candidate genes did not have annotation information in the 100 kb linkage disequilibrium interval information either upstream or downstream. Therefore, we extracted the FASTA sequences of these candidate genes that were not annotated with gene names on the reference genome, for sequence alignment with the reference genome of wild yaks using BLAST. Except for the rs769892 locus, all the candidate loci obtained gene names with high similarity on the wild yak reference genome (Table 3).3.4. Genotype Correlation in the LD BlockIn order to further determine the haplotype effect of association markers in BH, we used phenotype values distribution of significant alleles across all individuals to show the influence between BH and alleles. The results of haplotype analysis of 100 kb before and after the independently significant SNP showed that there was linkage disequilibrium in chr15-rs2910497-57505237 (chr: chromosome of the SNP; rs: SNP numbering on the genome; the third digit indicates the relative physical position of the chromosome on which it is located). The nearby LD block region was small, and the SNP was located on the outer side of the transcript. The analysis of allele haplotype based on this SNP showed that only GA and GG genotypes existed in 94 individuals, and no homozygous AA genotype was observed. In the results of the BH traits, yak individuals show higher body height (Figure 4A) when they have allele A at this SNP position. The same situation was also observed for four SNPs: rs2659279 (Figure 4B), rs477625 (Figure 4C), rs310769 (Figure 4D), and rs769892 (Figure 4E); the four haplotypic alleles with strong additive effects on BH were G, C, T, and C, respectively. There was a strong LD between chr13-rs2659279-59118279 and chr4-rs769892-4883046, two SNPs in a LD block, but the block segment was small. Among them, rs2659279 and rs2659285 on chromosome 13 had strong linkage disequilibrium, and the two SNPs were physically close to each other. rs2659279 was used as a benchmark when screening candidate genes. Except for rs769892, the other four SNPs with statistical significance were located near the intron region where at least one unknown transcript existed, and the sequences of the physically closest transcripts were compared by blast. The results showed that the transcript near rs2659279 was matched to FXYD gene, rs310769 to ADGRB2, rs2910497 to SOHLH2, and rs477265 to OSBPL6 (Table 3).4. DiscussionMost GWAS models are effective in detecting populations, which is usually conducted by using genome-wide SNPs in a large number of individuals in the same population [26]. They have good statistical power in a wide range of applications for locating human disease loci [27] and developing molecular markers [28], and for breeding selection of major economic plants and animals [29]. On the Qinghai–Tibet Plateau, it is difficult to obtain a large number of individual yaks of the same breeds or genetic resources. In this study, we performed GWAS with a large number of non-homologous resource individuals. From the results, even if there is significant population stratification, it is difficult to reduce the stratification effect by marker estimation through traditional GWAS models. However, population structure is often a fundamental factor affecting the accuracy of association results, which inevitably leads to false positive and false negative association results [30]. In addition, all of these yak populations were sampled from each kernel population. There is wide inbreeding in each kernel population to keep pure lines between families. All samples were clustered into several groups. Hence, heterozygosity is not abundant.Variation in body size traits is mainly based on distant cross-breeding among yak populations or genetic resources. In particular, to rejuvenate current domestic yak production performance, the wild yaks with superior traits were used for cross-breeding by local herders. These wild-blooded yaks (cross-breeding offspring of wild yaks and domestic yaks) tend to have a stronger body. Therefore, yak populations are more complex throughout the whole Qinghai–Tibet Plateau region. None of the 31 yak breeds and genetic resources in this study showed significant population clustering in multiple cluster analyses, with only the Tibetan breeds and genetics resources having a more similar genetic structure owing to their domesticated origins in Tibet [13]. Such populations with complex population structures are not conducive to statistical analysis by traditional CMLM. The grouping process of CMLM, which assigns individuals with similar characteristics to the same group, uses the elements in the kinship matrix as a similarity measure [31], replacing the kinship between individuals [12]. The genetic principle utilizes intra-group balance, which reduces the variance of the model’s residual part and improves the statistical power of GWAS. The difference between pCMLM and CMLM is that pCMLM provides clustering relationships of breeds or populations from previous and known study, and the kinship is compressed directly. In contrast, CMLM needs to filter the best compression levels by optimizing the likelihood values. However, sometimes the likelihood values between different compression levels are insignificant. The above situation would cause CMLM to use individuals to represent groups and kinships without any compression, taking CMLM back to MLM. Population structure or population origin in yak populations usually implies similarity in nutritional level and growth environment [32], and so using these factors to force CMLM to be compressed is beneficial for detecting candidate genetic markers. However, the complex structure makes the statistical power of CMLM closer to that of standard MLM. Therefore, when traditional CMLM cannot identify optimal clustered groups, our proposed pCMLM can reduce the variance of the residuals in the model by providing the real kinship and group structure. In summary, pCMLM can provide us with more adequate association results. The results show a small difference in -2LL between the two strategies. The -2LL is used to determine how well the model fits the variables, and when there is a large difference in the -2LL values, the model’s fit can be used to explain the training optimality of the model itself. However, this training optimality is not absolute because the training optimality of the model itself does not fully represent the detection ability and prediction ability. In 2010 and 2014, Zhiwu Zhang and Meng Li, respectively, confirmed a negative correlation between -2LL values and GWAS detection efficiency using a large amount of simulated data, but this relationship was derived from a large number of statistics and no specific conclusion was given in a single experiment [12].GWAS results based on pCMLM identified six SNP loci that were statistically significant in association with BH. Based on the annotation information provided by the 100 kb upstream and downstream of the yak reference genome, we obtained relevant annotation information on only four SNPs, but these contained only Ensembl ID. Probably because the yak reference genome is not yet complete, the sequences of these genes were compared to the wild yak reference genome by the BLAST tool. These candidate genes show association of rs2659279 (chr13-59118279) with FXYD6 (domain containing ion transport regulator 6), rs2910497 (chr15-57505237) with SOHLH2 (spermatogenesis- and oogenesis-specific basic helix-loop-helix 2), rs310769 (chr2-16165590) with ADGRB2 (adhesion G protein-coupled receptor B2), and rs477265 (chr2-129838648) with OSBPL6 (adhesion G protein-coupled receptor B2). Among these genes, the FXYD6 gene is a member of the FXYD family encoding a transmembrane protein, a specific protein encoding the hippocampus phosphate that has been shown in humans to be involved in mediating the Na/K ion pump [33]. The FXYD6 gene significantly accelerates Na+ deactivation and Na+ pump conversion rates [34] and alters the selectivity of the intracellular ion pump [35]. The SOHLH2 gene belongs to the b-HLH (basic helix-loop-helix transcription factor) family, which encodes a testis-specific transcription factor essential for spermatogenesis, oogenesis, and folliculogenesis [36]. The b-HLH family is involved in numerous biological processes in the organism, including cell differentiation, cell cycle arrest, and apoptosis [37]. The SOHLH2 gene has also been shown to play an important regulatory role in the reproductive gonadal axis, pituitary, hypothalamus, ovary, and testis of buffalo [38], pigs [39], and mice [36,40,41]. The ADGRB2 gene acts as a transcriptional repressor through GA-binding protein, regulates vascular endothelial growth factor, and is significantly associated with its growth traits in grouper [42]. The OSBPL6 gene is a member of the family encoding hydroxysteroid-binding protein (OSBP), an intracellular lipid receptor [43]. The OSBPL6 gene contributes to the maintenance of cholesterol homeostasis in vivo by regulating cholesterol transport in humans through miR-33 and miR-27b [44]. In studies of OSBPL6 in juvenile DePaul dwarf horses, it was shown that OSBPL6 is an essential factor affecting body height in DePaul dwarf horses [45] by showing variable splicing of ES type in the pituitary gland. Moreover, ES variable splicing causes GH1 third exon jumping resulting in a 17.5 kD GH isoform, which is an essential factor contributing to height defects in patients with autosomal dominant growth hormone deficiency (type II) [46], and the gene may be associated with multiple epiphyseal dysplasias [47]. Therefore, it is hypothesized that OSBPL6 and ADGRB2 genes are the most likely candidates to affect body height traits in yaks. However, whether and how the studied localized genes affect body height and size traits in yaks needs to be further explored, providing new directions and ideas for later validation studies.5. ConclusionsIn conclusion, when traditional CMLM cannot provide effective compressed grouping by obtaining the best likelihood value, pCMLM can obtain better results by using the previous population clustering information provided. From the GWAS results of pCMLM and LD analysis, four candidate genes (FXYD6, SOHLH2, ADGRB2, and OSBPL6) were provided in association with yak’s body height. This study will help us to develop better biostatistical model optimization ideas and a deeper understanding of the relationship between genes and body height. These results may provide basic information for quantitative trait gene localization or candidate gene cloning in the yak body height formation mechanism.

animals : an open access journal from mdpi

10.3390/ani11113054

PMC8614377

Honeybees have attracted considerable scientific and public interest in recent years. Besides pesticides and pathogens, failure or loss of the queen have been considered the most important factors leading to colony losses worldwide. The poor quality of the queen is a factor that ranks among the top reasons for bee colony failure. There are traits that can indicate the quality of a queen bee. This study aims to investigate the diversity in reproductive and morphological traits that can be useful in selective breeding programs for improving colony performance and survivability. The studied animals came from a population bred by a professional queen breeder in Northern Italy. Heritability and genetic correlations were estimated. According to our results, some of the traits showed good variability and they could be included as breeding goals in selection programs. Improving the quality of queens could directly impact honeybee colonies’ performance and survivability. Ultimately, it represents an added value to a queen bee-breeder company.

The quality of the honeybee queen has an important effect on a colony’s development, productivity, and survival. Queen failure or loss is considered a leading cause for colonies’ mortality worldwide. The queen’s quality, resulting from her genetic background, developmental conditions, mating success, and environment, can be assessed by some morphological measures. The study aims to investigate variability for traits that could assess the quality of the queen. Related animals were enrolled in this study. Variance components were estimated fitting a mixed animal model to collected data. Heritabilities of body and tagmata weights ranged from 0.46 to 0.54, whereas lower estimates were found for the tagmata width and wing length. Heritabilities estimated for the spermatheca diameter and volume, number of ovarioles, and number of sperms were 0.17, 0.88, 0.70, and 0.57, respectively. Many phenotypic correlations related to size were high and positive, while weak correlations were found between morphology and reproductive traits. Introducing a queen’s traits in a selection program could improve colonies’ survivability. Further research should focus on better defining the correlations between the individual qualities of a queen and her colony’s performance.

1. IntroductionHoney bees (Apis mellifera) are social insects who live in colonies characterized by a cooperative system of brood care, overlapping generations, and reproductive division of tasks [1]. In such an organized bio-social structure, the queen is the fertile female whose main duty is to lay eggs [2]. Moreover, the queen maintains the colony cohesion through a continuous production of a bouquet of pheromones which are actively spread within the nest. They prevent the workers from substituting the queen and developing their ovaries [3,4,5,6]. The colony development, productivity, and survival depend substantially on the health and fitness of its queen and the drones which she mated with [7,8,9,10]. A bee colony is adversely affected if the queen shows any defects or becomes ill and ceases to lay eggs [11,12,13,14]. Besides agrochemicals, parasites, and pathogens, failure or loss of the queen have been considered the most important factors leading to colony losses worldwide, especially when it occurs outside the natural queen rearing season [15,16,17,18]. A poor quality queen is a factor that consistently ranks among the top reasons for bee colony failure [18,19].There are many measures that can be correlated to queen “quality”, which results from her genetic background, her developmental conditions, mating success, and adult environment including the beekeeper’s management [20,21,22]. The most intuitive are physical measures of the queen, such as the body weight, which was found to be significantly correlated with her fitness and colony productivity [23,24,25,26,27,28,29,30]. Weight was also found to be positively associated with higher acceptance of queens in new colonies [30,31,32,33]. Bodyweight was also positively correlated with reproductive organs of the queen such as ovaries and number of ovarioles, the diameter of the spermatheca, and the number of stored spermatozoa [23,24,27,28,34]. Amiri et al. [35] concluded in their review that the body weight of a queen could represent an integrative measure of the size and physiological condition. Therefore, it could be considered one of the most informative indicators of the queen’s quality. Researchers investigated any possible association between weight and reproductive organs [24,26,34,36]. In Delaney et al. [24], thorax width was found positively correlated with the number of stored sperm and mating frequency. Meanwhile, other studies reported no correlation between thorax width and ovarioles number, ovary weight, or mating number [26,34,36].In a mated and egg-laying queen, the ovaries are the organs involved in the production of eggs. They are present in couples and occupy most of the abdominal cavity [2]. They consist of a bundle of ovarioles (ca. 150 each), which are long tubules containing egg cells, nurse cells, and follicle cells [37]. Ovary development takes place soon after the mating flights and it is associated with distinct gene-expression patterns in the brain and ovaries, and physiological and behavioral changes in the queen [38,39,40]. The weight of the ovaries in a mature egg-laying queen not only depends on the number of ovarioles but also on the number and developmental stage of eggs they contain [35]. The ovaries’ weight was reported as one of the internal physical criteria to assess the reproductive potential of honey bee queens [24,41,42]. Ovary dimensions and fertility are reported to be positively correlated [43]. The number of ovarioles can be evaluated at any time during the life of a queen [37]. Queen size, ovary size, and symmetry are affected by larval nutrition [27,44]. If the queen is artificially reared, the age of the grafted larvae is critical and also influences natural queen supersedure [27,43,44,45].Besides the ovaries, the queen’s reproductive system includes one spermatheca. The spermatheca is a small spherical shaped organ that preserves living sperms after mating for a lifelong period of time [2]. The spermatheca’s size is another measure of internal physical queen quality, under the assumption that a larger spermatheca could hold a larger volume of semen [35]. Its size can be measured with or without the tracheal nets, and its diameter should be larger than 1.2 mm for high-quality queens [25,37]. This measurement was used as a direct estimation of the volume and as an indirect estimation of the theoretical maximum number of spermatozoa stored in spermatheca [26,28,34,37]. The size of the spermatheca is influenced by rearing conditions and genetics, and it is inversely proportional to the larval age at which the queen was reared from [26,43]. Queens raised from newly hatched larvae showed larger spermatheca [42,43,45]. However, the spermatheca is rarely filled completely, as the semen’s occupied volume in experimental queens was reported to be on average 47% [26,34].From a hypothetical perspective, a “high quality” queen should therefore be morphologically defects-free, and it should have a large body, spermatheca, and ovaries in order to store a high number of spermatozoa and lay a copious number of eggs, preferably over 2000 eggs per day [14,46].In this study, both external and internal physical queen traits were investigated. Such traits include: body weight, weight and width of the tagmata (head, thorax, and abdomen), length of the right forewing, diameter and volume of the spermatheca, number of sperms in the spermatheca, and number of ovarioles. Ovarioles were counted instead of weighed since the ovary’s weight could be influenced by the developmental stages of the eggs they contain, as pointed out by Amiri et al. [35].The aim of this research was to investigate phenotypic and genetic variability of the above-mentioned traits for queen quality in a small population bred by a professional queen breeder in Northern Italy.2. Materials and MethodsThe queens were provided by an Italian queen-breeding and beekeeping company that produces and sells about 400–600 queens per week, from spring to late summer. The rearing of the queens was characterized by a standardized production system and by traceability of both maternal and paternal lines of each queen (the pedigree). The standardized rearing system consists of using only queen-less finisher colonies. These colonies are fed and treated in a standardized way to provide uniform quality. Specifically, these colonies receive new brood from one single apiary on a regular basis. The “brood-donor” apiary is composed of genetically uniform colonies. The grafts are inserted in the finisher colonies for a week. Afterwards, the royal cells are collected and incubated at 34.5 °C for 11 days. After incubation, the cells are brought to the mating station where they are inserted in the mating nuclei. This process was carried out by the same operators following a strict timetable along the season, which minimizes any potential error variance due to management practices.This study was conducted on 147 queens bred during spring/summer seasons of 2017 (n = 70) and 2018 (n = 77). The analyzed queens were bred at different times of the production season reported as the ordinal number of the week of the year in which the mated queen was harvested from the mating nucleus. All queens naturally mated in the same area within the year. The queens were bred in groups of sisters from 10 maternal lines in 2017, and 7 maternal lines in 2018. Furthermore, the maternal lines shared common ancestors. The maternal lines queens mated at an isolated mating station where each year a group of 12–15 drone-producing colonies provided drones. In this way, a pedigree could also be traced from the paternal side.2.1. Animal Sampling and TransportThe newly mated queens were harvested by the operators after assessing that each queen was successfully mated at the mating station. This was accomplished by checking the presence of viable female brood in the mating nucleus. The queens were shipped to the laboratory in suitable queen cages with feed and a sufficient number of attending worker bees (Figure 1, left). During transport, they were kept in a cardboard box with spacer support for the cages and holes for aeration (Figure 1, right). This transport system is one of the most common ways to transport or ship living bees in Italy.2.2. Freeze ImmobilizationFor easier handling during the first inspection, the queens were stored in the freezer at −20 °C for 15–17 min. The cold anesthetized the animals and induced them into an immobile state.2.3. Morphological MeasuresAfter cold immobilization, the first analysis was to evaluate the exterior state of the insect for the detection of any macroscopic defects e.g., missing legs, wings, antennae, or any visible injury of the body. These assessments were carried out under a stereomicroscope. Afterwards, queens were euthanized by decapitation and processed. The way morphological traits were measured differed between 2017 and 2018.In 2017, after the sacrifice, the insect was pinned onto a dissection dish previously filled with paraffin in order to form a basal layer to permit the fixing of the animal on the paraffin surface with entomological needles. The right forewing was detached from the body and laid beside the insect. The dish with the insect was numbered and photographed for post hoc morphological measurements with ImageJ software (U.S. National Institutes of Health, Bethesda, MD, USA) [47]. Every image contained a reference scale (graph paper) to determine the pixel/mm ratio. After the acquisition of the image, every queen was weighed on an analytical scale recording the weight of the entire insect body (bw), and separately also the weight of the head (hw), abdomen (aw), and thorax (tw). Other metrics were recorded with image analysis as follows: the head width (hwi) was assessed measuring the distance between the two compound eyes in their widest frontal part; the thorax width (twi) was recorded measuring the distance between the two tegulae, which are the scales on the mesothorax that overlaps the root of the forewing; the abdomen width (awi) was assessed measuring the width of the first apparent abdominal tergite in its wider part. Finally, the length of the right forewing (wl) was measured from the humeral plate to the apex. In 2018, the above-described measurements were taken manually using a digital caliper right after the sacrifice of the insect. Figure 2 shows how measurements were taken both with image analysis and manually.2.4. Abdomen DissectionIn order to assess the reproductive characteristics of the queen, the abdomen was dissected following the methods described in Porporato et al. [46]. The abdomen was kept in a ventral position by two proximal needles and one distal needle. Each abdomen was dissected cutting the junction between dorsal and ventral tergites with a scalpel and pulling away one by one the dorsal tergites with tweezers and fixing them beside it with needles (Figure 3). As soon as the abdominal cavity was opened, the abdomen was submerged in physiological solution (NaCl 0.9%) to prevent the drying out of the internal organs.2.5. Spermatheca Extraction and AnalysisAs soon as the abdomen was opened, the spermatheca was extracted and put on a glass slide. In 2017, the spermatheca was photographed on a slide under the microscope, and its diameter (ds) was measured with ImageJ software. In 2018, the diameters were measured by hand with a digital caliper. Since the spermatheca of a queen does not have a perfect spherical shape, the diameter was calculated as the average of three measurements. The volume of the spermatheca (vs) was calculated as the volume of a sphere.To assess the concentration of semen, the spermatheca was popped in 1 mL of physiological solution (NaCl 0.9%). The concentration of sperm (sp) was assessed using a standard haemocytometer chamber (Burker Camera) under a light microscope [48].2.6. Ovarioles CountOnly the ovarioles of the right ovary were counted [37]. The number of ovarioles (o) was estimated with a method derived from Porporato et al. [46]. The right ovary was removed from the abdominal cavity and kept in staining solution for ca. 10 min (methylene Blue Trihydrate, 0.5%, Sigma-Aldrich, St. Luis, MO, USA). After dying, the ovary was washed with physiological solution, placed on a slide, and analyzed under a stereomicroscope. The ovarioles were separated with the help of a dissecting needle and counted one by one, without cutting the ovary (Figure 4).2.7. Statistical AnalysisFor the analysis, estimates of the genetic relationships among the studied queens are required. The methods of Brascamp and Bijma [49] were used to estimate relationships assuming the queens mated with 12 drones. The pedigree file was built following the procedure described in Brascamp et al. [50]. To estimate heritability and genetic correlations, the statistical package ASReml and the pin function of the nadiv package were used in the computing environment R [51,52,53].First, a univariate mixed animal model [54] for each trait was fitted, using the following model:(1)yijk=μ+wyi+αj+εijk, where μ is the overall mean of the trait, wyi represents the fixed effect of the ith combination of the week of the year in which the queen was harvested (i = 1, 15, specifically 6 weeks in 2017, 9 weeks in 2018), αj represents the random genetic effect of the jth queen (j = 1147), εi,j,k represents the random error term of the kth observation. This model allowed us to estimate the heritability of each measured phenotype.Secondly, a bivariate approach was used for each combination of traits fitting the same model above described. The bivariate analysis allowed us to estimate both phenotypic and genetic correlations between the measured traits.3. Results and Discussion3.1. Defects and Descriptive StatisticsEvery analyzed queen was free from macroscopic external defects. However, some internal defects were observed. The most frequent defect observed in the 16% of the queens was an abnormal intestinal tract, which appeared swollen and brownish/yellowish in color; in the 9% melanosis of the ovary was observed; in the 8% of queens enteroliths were observed, which are small stones in the intestinal tract also described in Porporato et al. [46]; in 2.7% atrophy of at least one ovary was observed. Finally, less than 1% of cases showed empty or dark-colored spermatheca. Potential causes for the aforementioned defects were not investigated further. It can be assumed they arose from either abiotic factors (transport conditions, handling at harvest) or from biotic factors (pathogens). Mean, standard deviation (SD), and coefficient of variation (CV) of recorded traits are presented in Table 1.The number of sperm and volume of spermatheca were very variable, as their CV resulted 82.8% and 41.7%, respectively. Other measures such as tagmata weights, diameter of spermatheca, and number of ovarioles were less variable and their CV ranged from 10.1% to 17.9%. Concerning body weight, our result was in good agreement with values reported by the literature [24,25,26,46,55]. Records from Italian studies were of 186 ± 24 mg average body weight reported by Porporato et al. [46] and 221 ± 3.09 mg reported by Hatjina et al. [25]. Head and thorax widths were in agreement with Tarpy et al. [26], Delaney et al. [24], and Hatch et al. [45]. While no references are currently available on abdomen dimensions, the average width of the first apparent tergite was 4.8 ± 0.21 mm. Furthermore, queen tagmata weights were not found in the literature. Our data suggest that the abdomen’s weight was the most variable among the three tagmata. It could be argued that the eggs’ developmental stages may differ along each ovariole and the nutritional state of the queen at the moment of the analysis (filling of the intestinal tract) might be the source of such variation. The length of the right forewing was in agreement with previous reports [24,55]. Concerning the reproductive organs of the queen, our result on the number of ovarioles was in range considering the 74 ± 14 reported by Porporato et al. [46] and the 174 reported by Hatijna et al. [25]. Results of spermatheca size and number of sperm in the spermatheca were in agreement with the majority of previously reported studies [24,25,26,46].3.2. Heritabilities, Genetic and Phenotypic CorrelationsGenetic parameters are reported in Table 2; heritabilities were estimated with univariate analyses of the traits, while genetic and phenotypic correlation were estimated with a bivariate approach.Heritabilities ranged from 0.17 to 0.88 with rather high standard errors which might be explained by the sample size. In particular, for body weight and for individual tagmata weight, heritabilities resulted to be in a narrower range between 0.46 and 0.54. On the other hand, for tagmata width, the higher estimate was found for thorax width (0.42 ± 0.32), followed by the head width (0.26 ± 0.27) and the lowest was the estimate for the abdomen width (0.13 ± 0.26). The estimate for the length of the right forewing was 0.30 ± 0.29. Estimates for diameter and volume of the spermatheca resulted in 0.17 ± 0.34 and 0.88 ± 0.39 respectively. The increase of sv heritability compared to sd heritability could be explained by the way sv was calculated from ds. Indeed, sv was estimated from sd applying the formula of the volume of the sphere. This transformation increased the total phenotypic variance of trait sv by a coefficient (4/3 × π), without affecting the sv error variance. Therefore, the higher the genetic variance, the higher the heritability of sv trait. This result should be verified as the volume of spermatheca was approximated to that of a sphere, although this organ may show different shapes [46]. The estimates for the number of ovarioles and for the number of sperm in the spermatheca were 0.70 ± 0.35 and 0.57 ± 0.35, respectively. There are no results in the literature so far on the heritabilities for the traits measured in this study. Overall, the results show considerable genetic variability in the studied population.Concerning morphological traits (bw, hw, tw, aw, hwi, twi, awi, wl), phenotypic correlations ranged from 0.16 to 0.88. Highly correlated traits were body weight with abdomen weight (0.88 ± 0.03) and thorax weight (0.67 ± 0.06); moreover, a correlation of 0.80 ± 0.04 resulted between abdomen and thorax widths. Lower correlations were observed between morphological traits and the length of the right forewing. We observed very low or close to zero phenotypic correlations among reproductive traits (ds, vs., o, sp) except for the correlation between sd and sv (0.97 ± 0.01). The latter is likely explained by the fact that sv is derived by sd using the formula to estimate the volume of a sphere. Remarkably, according to our results, reproductive traits do not seem to be associated with morphological measures. Our results are in agreement with Corbella and Gonçalves [56], Hatch et al. [45], and Jackson et al. [36] who also reported the lack of phenotypic correlation between the body weight of a queen and the number of ovarioles. In addition, no phenotypic correlation between the number of sperms and spermatheca diameter was found, as previously reported by Jackson et al. [36].As expected, our genetic correlations estimates were affected by large standard errors due to the limited number of individuals in our dataset. Therefore, estimates should be considered with caution. Among morphological traits, tagmata weights showed high genetic correlations which ranged from 0.61 to 0.99. High and positive genetic correlations were found between the weight and width of the head (0.97 ± 0.46) and thorax (0.83 ± 0.31), respectively. There was also a high genetic correlation between thorax width and abdomen width (0.74 ± 0.34). In addition, high genetic correlations were observed between wl and hw (0.92 ± 0.71), tw (0.74 ± 0.38), hwi (0.96 ± 0.48), and twi (0.98 ± 0.31). The high relationship between lw and thorax dimensions may be explained by the fact that wings grow on the thorax, the locomotive tagma which groups the legs and the wings. It could be argued that during the insect’s development they are tightly and jointly regulated. The dimension of the spermatheca positively correlated with tw (0.78 ± 0.39) and hwi (0.89 ± 0.48). As expected, the diameter and volume of the spermatheca correlation were near to one (0.99 ± 0.02). Positive genetic correlations were found between wing length and number of ovarioles (0.79 ± 0.59) and number of sperms (0.57 ± 0.52). The association between wing length and these reproductive traits may be explained by the paramount function of the wings during the mating flight, which depends on the wings’ movements. The very high negative genetic correlation between the dimension of the spermatheca and the number of sperms it contains was surprising (−0.96 ± 0.72 with ds, −0.70 ± 0.56 with vs), previous reports showed that the spermatheca is often not totally filled after mating [26,34].4. ConclusionsBeekeepers have long selected queens, choosing the “best” on the basis of phenotypic desired features, mainly body size and color. The purpose set at the beginning of this study was to investigate a series of traits that could be useful for the evaluation and the selection of any qualities of a queen. Statistical analysis confirmed the existence of certain rather intuitive correlations between some morphological measures. Other results shed light on aspects that were counterintuitive in principle, i.e., the absence of phenotypic correlation between morphological measures and reproductive traits. Heritability is a parameter that generally describes how easily parents transmit to their offspring a certain phenotype and which has, on the other hand, a practical operational utility for breeders and selection programs. The heritability of body weight, spermatheca volume, number of ovarioles, and number of stored sperm look like promising breeding goals, given the observed genetic variability. It would be interesting to reproduce and extend this study to other traits, and to a larger number of individuals (also considering different breeds), to confirm or update the conclusions from this study. Moreover, it would be pivotal to extend such studies to colonies’ performances. In fact, phenotypic or genetic correlations between colony performance and queen quality, estimated with the animal model, are completely lacking in the literature.In conclusion, monitoring and introducing queen’s traits as breeding goals in a genetic improvement program represent an appealing plan of approach to the overall decline of queens reported in recent years from both the beekeepers and the scientific community. Improving the overall quality and reproductive traits of queens could directly impact colonies’ performance and survivability. Ultimately, it represents an added value to a queen bee-breeder company.

animals : an open access journal from mdpi

10.3390/ani11113308

PMC8614475

Cats have been closely linked to humans for thousands of years. Nowadays, stray cats are frequently hosted in colonies, protected, and enrolled in programs of trap–neuter-–return to control population increase. Italian public veterinary services work in collaboration with voluntary colony caretakers and are responsible for neutering and monitoring the health of colony cats. This retrospective study, conducted by the Anatomical Pathology Unit of the Teaching Veterinary Hospital of Milan in collaboration with the public veterinary services, was undertaken because of the limited information available regarding causes of death of colony cats. The study reports on and statistically analyzes the causes of death of colony cats in the city of Milan as assessed by necropsy. Inflammatory processes including those consistent with the most relevant feline infectious diseases were most common in kittens and young cats. Trauma was more frequent in adult cats, while organ failure was the most common cause of death in aged cats. Considering the possible animal welfare issues deriving from colony cats, awareness of the most common causes of death and collaboration between university veterinary pathologists and public veterinary services represent an essential contribution to health monitoring of colony cats and can assist in the rapid detection of possible emerging animal welfare concerns.

The presence of cats in urban environments has a long history. In Italy, stray cats are protected by national and regional laws, and programs of neutering and reintroduction to colonies are ongoing. Colony cats have been widely studied from a behavioral perspective, while surveys regarding their causes of death are limited, although they may provide relevant information related to public health and cat welfare. This retrospective study provides pathological descriptions and statistical analyses of the causes of death of 186 cats from 100 colonies in the city of Milan. Inflammatory processes represent the primary cause of death (37.7%) and include common feline infectious diseases such as feline panleukopenia (67.5%), particularly in kittens, and feline infectious peritonitis (32.5%), most common in adult cats. Trauma was found to be a common cause of death of young/adult cats (14%) with a generally good body condition, while severe parasitosis was less represented (2.6%). The death of old cats was statistically associated with organ failure (24.7%), particularly renal failure, and tumors (11.8%). Knowledge of the most common causes of death of colony cats could make an important contribution to the health monitoring of these cats and sanitary control of their habitats and provide information on possible related emerging animal welfare concerns.

1. IntroductionPopulations of unowned cats exist throughout the world [1], and cats have been closely linked to human society for thousands of years [2]. The most ancient evidence of cat–human co-existence, dating to 7000 years BC, was discovered in Cyprus, followed by Egyptian evidence dating to 2000 years BC [3,4]. In Italy, the presence of domestic cats in urban environments has a long history, with ancient Roman documents reporting stray cats living around monuments and in public gardens [5]. Many public issues related to stray cats include health concerns with regard to zoonotic diseases, the spread of diseases to pet cats, public nuisance, and the welfare of the cats themselves [2,6,7]. Stray cats are considered a threat to native wildlife, but when they are grouped in managed colonies, studies have revealed minimal bird predation by colony cats [8]. A cat colony is nowadays mainly defined as a group of three or more sexually mature cats living and feeding close to one another [9]. Cat colonies are not mere aggregations of individuals around sources of food, but represent truly structured and functional social groups [10,11], even if a consistent source of adequate food provided by caretakers is essential for a colony to remain in one location [2].Italy signed and ratified the European Convention for the Protection of Pet Animals and delegates regions to organize birth control [12], protecting stray cats in feline colonies that are regularly counted and registered. Recent data refer to around 70,000 colonies of cats, 11,000 of which are in Lombardy, the region where Milan is the main town. In Milan, approximately 1000 cat colonies have been registered by the veterinary public service (VPS). Cat colonies are provided with food and shelter by caretakers, who also participate in the no-killing, trap–neuter–return (TNR) program. Since 1991, TNR is the only legal policy in Italy for limiting the cat population [13]. With TNR, veterinarians identify cats with microchips, and their neutering status is indicated by a small cut on the ear. Caretakers play a fundamental role in the TNR process for the welfare and health management of cats by alerting municipal veterinarians in case of cat illness or death. The Veterinary Teaching Hospital (VTH) of the University of Milan is involved in TNR and the health management of colony cats through an agreement signed with the VPS (ATS Città metropolitana di Milano). The VTH provides hospitalization for severely ill or traumatized cats, and allows submission of dead cats from registered colonies for necropsy. Regular post-mortem examinations of colony cats, performed at the pathology unit of the VTH, makes a significant contribution to the health monitoring of the cat population in this urban territory.Several studies on unowned cats have been published in the last three decades, however, most of these have mainly focused on either the TNR program [1,7,14], the complete eradication of stray cat populations in specific areas [15,16], or an analysis of the prevalence of single infectious/parasitic diseases in certain regions or groups of colonies [17,18,19]. On the other hand, causes of death of colony cats have only been occasionally mentioned [6,20,21]. In a recent paper focusing on the possible negative effect of stray cats on bird populations, the stomach contents of cats were analyzed, but the causes of death of the examined cats were not reported [22].The aim of the present study is to report on and critically and statistically analyze the causes of death of colony cats in registered colonies in the municipality of Milan.2. Materials and MethodsNo animals were killed for this retrospective study. Necropsies of the colony cats were performed as part of the health monitoring programs according to the agreement between the Veterinary Public Service of the city of Milan and the VTH/Pathology Unit.2.1. AnimalsFor the present retrospective study, the archives of the VTH Pathology Unit were searched for colony cats submitted for necropsy by the municipal VPS between 2014 and 2018.For each cat, together with gross and histological findings, the following data were mostly reported: breed, age, sex, colony of origin, body condition, and cause of death. The age was estimated at necropsy and in the database was reported as kitten, young, adult, or old. Body condition evaluated at necropsy was recorded as emaciated, very thin, underweight, ideal, overweight, or obese. The town colony location was defined as central, middle, or peripheral, according to Abakumova (2013) [23].Causes of death were grouped as follows: (1) trauma; (2) inflammatory process (infectious and non-infectious diseases); (3) organ failure (kidney, liver, heart); (4) tumors, (5) severe parasitic disease; and (6) pulmonary edema (with no other apparent cause). For causes of death of single cats, a seventh group was created, named “miscellaneous causes”, which included all causes that were observed either occasionally or in single individuals. The eighth group included cases for which the cause of death remained undetermined.2.2. Statistical AnalysisDescriptive statistics, chi-square, and Fisher’s exact test were used to describe the causes of death and the associations with the different recorded categorical factors: age, sex, colony of origin, and body condition. Statistical analysis was performed using JMP® 13.0.0 (SAS, 2016, Cary, NC, USA).In addition, logistic regression analysis was performed on a subsample consisting of the 89/186 cats for which information related to all categorical factors considered in the study were available in the database. In the logistic regression model, cause of death as the dependent variable using age, sex, neutered status, and body condition, and colony location as the independent variable were analyzed.3. Results3.1. CatsNecropsy records from 186 cats were retrieved from the electronic archives (Table S1, Supplementary Material).3.2. ColoniesThe location of the colony was reported for 98 out of 186 cats (52.69%). For 10 of these colonies, multiple cats (two to five) were submitted for necropsy. The town distribution of colonies was central for 7/98 (7.1%), middle for 10/98 (10.2%), and peripheral for 81/98 (82.7%).3.3. Breeds and SexAll cats were domestic shorthair breed, and the sex was reported in 181 cases: 92 were male and 89 female. Eighty cats (80/181) were neutered, 43 (53.75%) male, and 37 (46.25%) female.3.4. AgeThe age was indicated for 170 cases: 32 (18.8%) were kittens, 24 (14.2%) were young, 44 (25.88) were adult, and 70 (42.1%) were old cats. The approximate age was not reported for 16 cats.3.5. Body ConditionBody condition was reported for 180 out of 186 cats: 47 of the 180 were in ideal body condition, 25 were underweight, 11 were overweight, 22 were emaciated, and 75 were very thin.3.6. Causes of DeathTrauma. Traumatic lesions were recorded for 27 out of 186 cats (14%); 12 were male and 15 female. Considering the age distribution, traumatic injury was the cause of death for three kittens, seven young, 11 adult, and two old cats. Age was not reported for eight cats. Body condition was reported for 26/27 traumatized cats. No cat was emaciated, seven were very thin, two were underweight, 13 were in ideal body condition, and four were overweight. Lesions observed included large hemorrhagic suffusions/hematomas, alone or associated with severe muscle ruptures and/or multiple bone fractures, frequently exposed or comminuted. The most commonly injured areas were the dorsal and hip regions and the hind legs. In one case, a diaphragmatic hernia was diagnosed.Inflammatory processes. Severe inflammatory disease was the cause of death for 70/186 cats (37.7%): 30 male, 39 female, and in one for which sex was not reported. Age was reported for 64/70 cats: 26 were kittens, 10 were young, 12 were adult, and 16 were old. In 40/70 cats, gross lesions observed at necropsy were consistent with either feline panleukopenia (feline infectious enteritis) (27/40) or feline infectious peritonitis (13/40). These diagnoses were histopathologically confirmed.Feline panleukopenia (FP) was diagnosed in 27/40 cats (67.5%): 19 female and eight male; 19 were kittens, five were young, and for three cats, the age was not reported. Body condition was reported in 25/27 of FP cases: 15 cats were very thin, five were underweight, and five were in ideal body condition. Grossly, severe hyperemia, and hemorrhagic enteritis of small intestine, jejunum, and ileum associated with edema of the mesenteric lymph nodes was observed in all cases (Figure 1). Histologically, crypt epithelial attenuation/necrosis with crypt ectasia or loss, villous atrophy, and collapse were observed in the small intestine. Peyer’s patches were severely depleted or regenerative depending on the stage of the disease. Edema and follicular depletion were detected in mesenteric lymph nodes. Lymphocytic depletion was also observed in the spleen and thymus.Feline infectious peritonitis (FIP) was diagnosed in 13/40 cases (32.5%): five cats were female, seven were male, and the sex was not recorded for one cat. Age was reported in 12 cases: one kitten, one young, five adult, and five old cats; three cats were emaciated, one was underweight and two overweight, four were very thin, and three were in ideal body condition. Five cats were affected by non-effusive (dry form) FIP. In these cases, histopathology identified pyogranulomatous lesions in various organs, with typical vasculitis and perivasculitis, especially of small to medium-sized venules of the renal cortex, lungs, and liver. Macrophages predominated and were associated with varying numbers of neutrophils, lymphocytes, and plasma cells. In 3/5 cats, only the abdominal organs were affected, while in two catsm both abdominal and thoracic organs were affected. Eight cats had effusive FIP (wet form). In seven cats, peritoneal effusion was observed, and in one cat, pleural effusion was also present. In all cases (8/8), an abundant, viscous, clear, and pale to deep yellow exudate, occasionally containing strands of fibrin, was documented (Figure 2). Strands of fibrin adhering to the omentum and serosal surfaces were also present. Pyogranulomatous lesions were occasionally seen in the parietal peritoneum, in the omentum, and on the surface of the main abdominal organs. In one case, pyogranulomas affected the visceral pleura and mediastinal lymph node.Pneumonia was reported in 19 cats, 10 male and nine female. Body condition was not reported for one cat, four cats were emaciated, three were underweight, eight were very thin, and three were in ideal body condition. Age was not reported in one case, five cats were old, six were adult, three were young, and four were kittens. Lungs were overall increased in volume and weight, with altered consistency. Suppurative pneumonia was most commonly diagnosed and was grossly characterized by hyperemic and irregular pulmonary surface with fibrous thickening of the interlobular septa. Areas of consolidated parenchyma were visible, together with gray nodules scattered throughout, revealing purulent exudate on cut surfaces. Histologically, alveolar and bronchiolar lumina were filled with edematous fluid and numerous degenerate and non-degenerate neutrophils, a few mature lymphocytes and plasma cells, scattered foamy reactive macrophages, and scant necrotic debris; similar inflammatory infiltration was frequently observed within bronchial/bronchiolar walls and expanding alveolar septa.Inflammatory processes other than pneumonia were diagnosed in 11 cats, six female and five male. In 10 cases, the age was reported: one adult, six old, two kittens, and one young cat. Regarding their weight, six of these cats were very thin, three were in ideal body condition, and two were underweight. Four cats were affected by enteritis: two of these four were old, one was young, and one was adult; three of the four were in poor body condition and one was underweight. Multifocal to coalescing facial subcutaneous abscesses were reported in two old cats.Organ failure. Forty-six cats died because of severe organ failure involving one or multiple organs.Renal failure due to chronic kidney disease was reported in 32 cases (32/46, 70%): 19 were male and 11 female, and sex was not recorded for two cats. Among these cats, 24 were old, seven were adults, and age was not reported for one; eight cats were emaciated, 16 were very thin, three were underweight, and five had ideal weight. Chronic renal failure was grossly associated with grayish, small, firm kidneys (end-stage kidneys) with multifocal to coalescing irregular depression of the surface. On cut surfaces, cortex was thinned, firm, and sclerotic. Histologically, interstitial fibrosis, glomerulosclerosis, atrophic nephrons, and ectatic tubules and dilated Bowman’s spaces were present. Multifocal, interstitial, severe lymphoplasmacytic inflammation was observed in most cases, while tubular mineralization was rarely observed.Hepatic failure was reported in 6/46 cases (13%), four males and two females, two adults and four old cats. Cats were very thin in two cases. In one case, the body condition was not reported, one was emaciated, one was underweight, and one was in ideal body condition. Cats were frequently icteric. Grossly, the liver was enlarged, greasy, and fragile. The liver was yellow in four cases histologically characterized by severe diffuse macrovescicular intrahepatocytic lipidosis (fatty degeneration). In another case, the liver was pale brownish and histologically characterized by interstitial deposition of eosinophilic amorphous homogenous material that stained positive with Congo Red (amyloidosis). In the sixth case, the liver was reduced and firm, with an irregular surface. Microscopically, lobules of degenerated hepatocytes were separated by septa of fibrous connective tissue containing multiple proliferating small bile ducts (cirrhosis).Pancreatic failure was recorded in one very thin, old male cat. In this case, the pancreas was reduced in size, firm, and pale. In addition, small multifocal dry yellowish plaques were detected on the surrounding mesentery (fat necrosis). Histologically, severe diffuse fibrosis of the interstitial pancreatic septa was detected.Heart failure was reported in five cats, four male and one female generally in good body condition. In these cases, severe concentric cardiac hypertrophy with reduced left ventricular lumen was observed. Ventricular myocardial interstitial fibrosis was histologically detectable in two cases.Multiorgan failure was observed in two old female cats. One was emaciated and showed end-stage kidney disease, severe diffuse hepatic degeneration, and left myocardial hypertrophy. The other cat was very thin, and pancreatic and severe hepatic fibrosis were detected.Tumors. Neoplastic lesions were observed in 22 cats, 12 female and eight male (in two cases, sex was not recorded). The age of two cats was unknown, 15 were old, three were adult, and two were young. Four cats were emaciated, five underweight, four were in ideal body condition, and six were very thin. The most common tumors were squamous cell carcinoma (6/22 cases) and lymphoma (5/22 cases). Two cats were diagnosed with osteosarcoma, while in the other cats, the following tumors were diagnosed: feline injection site sarcoma, hepatic carcinoma, pancreatic carcinoma, pulmonary carcinoma, ceruminous gland adenocarcinoma, histiocytic sarcoma, oral fibrosarcoma, myeloid leukemia, and thyroid adenocarcinoma.Parasitosis. Parasites were reported as the cause of death in five cats, four male and one female; one was adult, one was young, and one was a kitten, and the age of the fifth was unknown. Three of these cats were very thin, one was underweight, and for one, the body condition was not reported. Three cats were affected by Aelurostrongylus abstrusus. In these cases, multiple consolidated pulmonary areas and small nodules protruding on the lung surface and scattered in the parenchyma were grossly observed. Histologically, eggs and larvae in various stages were observed in both the alveoli and bronchioles, surrounded by eosinophilic and granulomatous reactions. One cat showed the massive presence of nematodes (Toxocara cati) in the intestinal lumen. In the other case, hookworms, morphologically consistent with Ancylostomadidae, were detected in the small intestine.Pulmonary edema. This was the only lesion observed in 10 cats, six male and four female. The age was reported in nine cases: four were adult, two were old, two were young, and one was a kitten. Body condition was ideal in four cats, three cats were very thin, one cat was underweight, one emaciated, and for one cat, it was not recorded.Miscellaneous causes of death. Three male cats were classified in this category. In two cats, cardiovascular collapse was suspected, while one cat was affected by ischiatic thrombosis.Indeterminate causes of death. For one old male cat and a female kitten in advanced putrefaction, the cause of death remained indeterminate.3.7. Statistical AnalysisThe distribution and interaction of the variables for the cats (sex, age, and body condition) with causes of death were analyzed with the chi-square test and logistic regression, as described in the Materials and Methods.No statistical differences were observed between male and female cats by the chi-square test. There were statistically significantly (p < 0.05) fewer neutered cats (n = 80, 44%) than non-neutered ones (n = 100, 56%), however, the neutered rate increased with age, reaching 56 and 66% in adult and old cats, respectively. No statistically significant differences between the sexes with regard to neutering was observed in the sample analyzed. A significant difference (p < 0.05) was observed between sex and age, with a larger number of dead male adult cats (n = 30, 68%). Cause of death was not significantly related to sex (Figure 1), although the higher impact of organ failure as a cause of death in males was evident (Figure 3). A significant difference (p < 0.001) was observed between age and causes of death (Figure 4). Specifically, inflammatory processes were the main statistically significant cause of death for kittens (81%), and the most common cause of death for 59% of young and adult cats, even if not significant. The main cause of death in old cats was organ failure (55%), followed by tumors (25%). As expected, organ failure was negatively associated with young age; indeed, none of the young cats died from organ failure. Death of young and adult cats was significantly associated with trauma, which was involved in 21 and 25% of recorded deaths, respectively. Within organ failure, renal failure was the most prevalent cause of death at 67%, significantly represented by neutered old male cats. A statistically significant (p = 0.002) association between body condition and cause of death was also observed (Figure 5). Specifically, ideal and overweight cats were significantly more likely to die of traumatic events. Conversely, emaciated cats were frequently affected by organ failure, although the chi-square test was not significant.The logistic regression model (R2 = 0.397, n = 87), similar to the chi-square test, showed that only age (p = 0.00467) and body condition (p = 0.0209) were statistically significant. No statistical effect was observed for neutering with this model, even if a statistically significant difference was observed with the chi-square test considering only the interaction between cause of death and neutering. This result is mainly associated with a confounding effect due to the different ages of non-neutered and neutered cats. In fact, as reported in Figure 6, cats that died from inflammatory processes were mainly kittens (non-neutered), and cats affected by organ failure were frequently old and thus generally neutered.4. DiscussionThe presence of stray cats in the urban environment has long been known. In Italy, stray cats are protected by national and regional laws [24]. In particular, national law No. 281 (1991) has two key points: stray cats have the right to live undisturbed freely, gathered in colonies, and they must be neutered by the local veterinary public service (VPS) and then reintroduced to their colony. In addition, colony cat caretakers become institutionalized figures who gather in associations, and can obtain the official assignment of managing a cat colony. Feeding, trapping, neutering, and releasing stray cats and allowing them to live in colonies is an answer to the overpopulation problem [6], and TNR programs are considered the most practical, effective, and humane way to control stray cats [1,13]. The finding of a substantial representation of old cats in the current study vouches for the success of the monitoring provided by VPS and caretakers. In addition, collaboration between VPS and the VTH plays an important role in monitoring cats’ health [24,25].Because of the scarce literature, the present retrospective study was aimed at reporting and statistically analyzing the causes of death of colony cats submitted for necropsy in the urban environment of Milan.Although it must be considered that cats can move from colonies and die far from them with no possibility of being monitored, the number of cats retrieved from the database was representative of the colony cat population supervised by VPS.As expected, most cats (83%) in this study came from colonies located at the town periphery, where there are green spaces and old abandoned buildings and colonies are more tolerated. Unfortunately, we had data on the colony location for just half of the cats considered. We suggest that it is important to register colony locations in order to investigate how social differences might correlate not only with cause of death, but also with animal welfare.A previous study regarding cat colonies in Rome [10] reported no prevalence of one sex over the other, while an Australian study reported a prevalence of male cats [26]. In the present study, there were significantly fewer neutered cats than non-neutered ones. This was probably due to the large number of kittens and young cats, as statistics revealed that the neutered rate increased with age, reaching 56 and 66% in adult and old cats, respectively. These findings are in line with the VPS policy of neutering at a minimum age of four months. In addition, the constant availability of food and care attracts stray cats [6,13], therefore, the presence of a certain percentage of newly entered, non-registered, and non-neutered cats must also be taken into account. Moreover, it must be also noted that one of the most relevant reported problems concerning the reduced success of TNR seems to be that the continuing increase in the number of cats because of the presence of colonies, where cats are sustained and monitored, can “morally justify” the abandonment of cats [2,6,27,28].Regarding the body condition, 46% of cats were overweight, slightly underweight, or in ideal body condition, while 54% were emaciated or very thin. Of note, 75% of the colony cats included in this study were affected by inflammatory diseases, organ failure, neoplastic disease, and parasitic diseases, which have been reported as the main causes of weight loss.Concerning causes of death, inflammatory processes such as pneumonia and feline infectious viral diseases were the most frequent in the cat population examined (70/186). Severe pneumonia was detected in 19/70 cats and was not significantly associated with age, being diagnosed in kittens as well as young, adult, and old cats. Colony cats are exposed to climate factors and pathogens, which can be predisposing factors for pneumonia. Moreover, colony cats can contract respiratory diseases that in owned cats are prevented by routine vaccination. Pneumonia is one of the most common causes of death in cats. Viral pneumonia can be caused by felid herpesvirus-1 or feline calicivirus (FCV) infection [29] and can be complicated by bacterial superinfection [29,30]. In addition, in cats, primary bacterial suppurative pneumonia/bronchopneumonia is also frequently caused by Bordetella bronchiseptica [29] or other agents such as Pasteurella and Mycoplasma spp. [31,32]. Colony cats can also be affected by agents such as feline leukemia virus (FELV) and feline immunodeficiency virus (FIV), ref. [28] which compromise the immune system, predisposing the animals to infection.Other inflammatory processes occurred in 40/70 cats examined and were consistent with common feline infectious diseases such as FP and FIP.In the present study, FP was significantly more common in kittens and young cats, in agreement with other reports describing FP as a highly contagious and lethal disease that is less frequent in adult cats [33,34]. FP virus transmission occurs by orofecal or indirect contact, and due to its long persistence (one year or more), the contaminated environment could represent a source of infection for both colony and owned cats [22].Another common feline infectious disease observed in the present study was FIP, a progressive fatal disease caused by feline coronavirus (FCoV) infection [35], diagnosed in 13 cats. Similar to the current literature, cats affected by FIP in this study showed severe systemic inflammation of serosal membranes and widespread pyogranulomatous lesions [36]. FIP tends to occur most frequently in cats < 2 years of age and less commonly in geriatric cats [35]. In our study, FIP cases also occurred in adult or aged cats. This is not surprising for colony cats, because, as reported for catteries/shelters, the incidence of FIP seems to be directly correlated with the number and density of cats. Moreover, stressors such as pregnancy, parturition, and surgical procedures such as spaying in TNR can increase the risk of developing the disease [37].The second most common cause of death in the cats considered was organ failure, detected in 46/186 cats (25%). Organ failure was the main cause of death in old neutered male cats (67%), renal failure in particular (32/46). Although in other studies sex did not appear to be a factor related to the onset of chronic kidney disease, a strong prevalence of old cats, as in the present study, has been similarly described [38,39,40,41,42].Renal failure is often associated with chronic kidney disease (CKD), which has been identified as an irreversible and progressive loss of renal function that occurs most commonly in geriatric domestic cats (>12 years of age) [43,44]. Generally, in CKD pathogenesis, a primary renal disease initiates parenchymal damage with loss of nephrons, followed by other factors intrinsic to the affected animal that lead to self-perpetuating renal injury [45]. Several primary renal diseases have been identified in cats including congenital disorders that have a typical breed predisposition [42,45]. All cats included in this study were domestic shorthaired, a breed not strictly predisposed to renal congenital disorders. On the other hand, it is important to consider that colony cats can develop other acquired renal disorders, since they are exposed to several risk factors including viral or bacterial infections and exposure to toxicants that may be present in the environment [45,46,47].Other types of organ failure were less common in the present study (14/46). According to the literature [48], pancreatic failure is rare, and in this study was diagnosed in only one cat. Heart failure, generally related to hypertrophic cardiomyopathy (HCM), was reported in five cases. This low number was expected, since HCM is more common in specific breeds such as Maine Coon and Norwegian [49,50] than in shorthaired cats.The second most prevalent cause of death in old cats was tumors (25%), consistent with the current literature reporting tumors as one of the main causes of death in aged cats [51]. The most common tumors were squamous cell carcinoma (six cases) and lymphoma (five cases), referred to in previous studies as the most common neoplastic diseases in cats [52,53,54]. The cause of most cutaneous SCC is chronic exposure to ultraviolet (UV) light, and in fact, these tumors are seen almost exclusively on the head, with white cats and colored cats with white areas being at greatest risk [52]. Lymphoma in cats commonly affects the gastrointestinal tract and seems to occur more frequently in cats seropositive for FIV and FELV, viruses that are easily transmitted among colony cats [53,54]. In the present study, only five cases of lymphoma were diagnosed, and this result is supported by the reduced seropositivity described in a recent study on colony cats of Milan (6.6% of cats positive for FIV, 3.8% positive for FELV) [55].Regarding trauma as the cause of death, in the present study, young (21%) and adult male (25%) cats in ideal body condition were significantly more prone to dying from trauma than very thin and emaciated cats. Young and adult cats, fighting off various other infections/infestations, have more opportunity to die from trauma. Conversely, very thin and emaciated cats, most likely already affected by other diseases, tend to remain closer to the colony, coming into contact with fewer potential traumatic events.Other studies have mentioned trauma as a common cause of death in kittens, which are reported to die or disappear from colonies within six months of birth [2,21]. In the present study, only three kittens died from traumatic lesions; these were more likely to die from infectious disease. However, traumatic death cannot be excluded for kittens that eventually disappear.The least prevalent cause of death in the present study was parasitic disease. Pathological findings of verminous pneumonia due to Aelurostrongylus abstrusus were observed in three cats. A. abstrusus (Nematoda, Strongylida), a lungworm, is reported to be the most common lung parasite in cats worldwide [56,57]. Cats become infected by ingesting intermediate (snails) or paratenic (e.g., rodents, frogs, lizards, snakes, and birds) hosts [58,59]. Stray cats have greater exposure to intermediate and paratenic hosts, and actually, in the present study, all five cases of lung parasites were in colonies located at the extreme periphery of the town. In these areas, very close to the countryside, both intermediate and paratenic hosts are present, increasing the risk that colony cats will initiate the infection cycle. It must also be considered that A. abstrusus infection as well as other pulmonary diseases is cited as a risk factor for anesthetic-associated death at neutering [28,60].Considering other possible causes of death in colony cats, there were no suspected cases of poisoning in the current caseload. In fact, as poisoning is a relevant public health concern, since 2008 in Italy, the carcasses of suspected poisoned animals must be referred for necropsy to specific veterinary public health centers named Zooprophlylactic Institutes, which rely on the Ministry of Health. Regarding Milan, during the period 2014–2018 considered in the present study, 11 carcasses of colony cats suspected of being poisoned were referred by the VPS to Istituto Zooprofilattico Sperimentale della Lombardia e dell’Emilia Romagna (IZSLER) and, in three of the 11, toxicological exams were positive for anticoagulant rodenticides. The use of anticoagulant rodenticides is common worldwide including in Italy [61,62]. Exposure occurs through ingestion of the rodenticide from bait containers or from the environment in which the rodents have carried the bait [63]. Relay toxicosis, which is intoxication by the ingestion of a previously poisoned animal, is unlikely to occur with rodenticide agents, because the amount in the rodent is small; however, it is possible in colony cats, as in other animals, that eventually prey upon rodents in areas where such bait is commonly used [63].5. ConclusionsIn conclusion, the present study shows that inflammatory processes (including those typical of common feline infectious diseases), traumatic lesions, organ failure, and tumors are the main causes of death among colony cats. Statistical analysis revealed a significant association between cause of death and age, with the most common causes of death in kittens and healthy young/adult cats being inflammatory processes and trauma, respectively. For older cats, the statistically significant causes of death were organ failure and tumors.Since colony cats represent a possible animal welfare problem as well as a target, collaboration among caretakers, VPS, and veterinary faculties is important for monitoring them. In this scenario, knowledge of the most common causes of death could make an important contribution to the monitoring of cat populations and their environment and could alleviate emerging animal welfare concerns, allowing rapid intervention.

animals : an open access journal from mdpi

10.3390/ani11040945

PMC8066830

Cleanliness is important for the health and welfare of cattle, but also for farm profitability, as dirtiness increases the risk of sick animals and can compromise milk and meat production. Swedish legislation states that all animals must be ‘clean enough’, but dirty cattle are commonly recorded in official inspections in Sweden. This study investigated the reasons for dirtiness and how inspectors handle cases of dirty cattle. Of the 371 cattle farms inspected, 49% had dirty cattle. However, the inspectors did not categorize all farms with dirty cattle as non-compliant, mainly using the argument that only a few animals were dirty. Therefore, in addition to knowing what characterizes ‘clean enough’ cattle, both inspectors and farmers need better guidance on when a farm is compliant, or non-compliant, with animal welfare legislation. Dirtiness in cattle was found to depend mainly on management routines on the farm, which is a promising result since routines can be improved.

Dirty cattle have been commonly recorded in official animal welfare inspections in Sweden for years. The relevant authorities have initiated work to better understand the causes of dirty cattle, in order to improve compliance and standardize the grounds for categorizing a farm as non-compliant with welfare legislation when dirty animals are present. This study investigated the occurrence of dirty cattle in official animal welfare controls, on Swedish cattle farms, and examined farmers’ views on the reasons for non-compliance and on key factors in keeping animals clean. The data used were collected by animal welfare inspectors at the county level during the regular official inspections of 371 dairy and beef cattle farms over two weeks in winter 2020. In addition to completing the usual inspection protocol, the inspectors asked farmers a set of questions relating to why their animals were clean or dirty. Dirty cattle were found on 49% of the farms inspected, but only 33% of the farms were categorized as being non-compliant with Swedish welfare legislation. According to inspectors and farmers, dirtiness in cattle depends mainly on management routines, which is a promising result since routines can be improved. The results also revealed a need for better guidance for inspectors and farmers on when dirtiness should be categorized as non-compliance with animal welfare legislation.

1. IntroductionDirty cattle are commonly recorded by animal welfare inspectors in Sweden during official inspections [1]. Inspectors have found dirty animals during 20–25% of cattle inspections every year since at least 2013, without no decreasing trend [1]. When analyzing inspection reports, Keeling [2] found dirty animals to be one of the most common forms of non-compliance with animal welfare legislation in official inspections on Swedish farms, while Lundmark Hedman and co-workers [3] found dirty cattle or dirty lying areas to be the most common form of non-compliance in official inspections and in private inspections by the dairy company Arla. A high incidence of dirty cattle has also been reported in other countries, for example, Finland [4], Norway [5], Brazil [6], Spain [7], and India [8]. Within the European Union (EU), there is no species-specific animal welfare legislation for cattle, except for calves. However, some member states, such as Sweden, include all cattle in their national legislation. Both EU and national animal welfare legislation are typically mainly resource- and management-based, i.e., consisting of requirements related to risk factors in the animal husbandry system, interior design, and management procedures [9,10]. However, Sweden has introduced a few animal-based requirements, i.e., requirements focusing on the appearance of the animal [11]. For example, Swedish national regulations for cattle farming state that ‘animals shall be kept clean enough’ [12]. Cleanliness is important for the animal, but also for the farmer. Dirty cattle can be expensive for the farm, since dirtiness clearly increases the risk of developing health problems, and most likely impairs milk yield and meat production. Cow cleanliness is essential for ensuring hygienic milk production and the good welfare of dairy cows [5]. Cow cleanliness is also important for maintaining the udder health in the herd [13,14], since failures in farm hygiene resulting in dirty udders pose a risk of mastitis [13,14,15,16,17]. In addition, if the coat of the animal is dirty and wet it will lose some of its insulating capacity, leading to animals having difficulties with thermoregulation [18]. Hence, more energy is needed to maintain thermal comfort and the animals need more feed in order to maintain growth and production rates. In addition to poor welfare due to the risk of health issues or thermoregulation difficulties, dirtiness per se is a severe welfare problem, since urine and manure can cause painful skin burns and dermatitis [19]. Barrientos and co-workers [20] showed that cows in dirty stalls were more likely to suffer from hock injuries, which can result in lameness [21]. Therefore, it is important to not underestimate the impact of farm hygiene and animal cleanliness on animal welfare. Dirty animals at slaughter can also compromise hygiene in the abattoir [5]. According to Hauge and co-workers [22], it is generally accepted that it is more difficult to avoid carcass contamination and preserve good hygiene when dirty cattle enter the slaughter facility.The causes of dirtiness in cattle are probably multifactorial and several risk factors may be involved, e.g., type of flooring/lying substrate [23,24,25], husbandry system and interior design [3,26], temperature and humidity [5], stocking density [27], parity and outdoor access [28], and deficiencies in hygiene management [5]. These risk factors can be partly eliminated by management actions and good routines by the farmer; for instance, it has been shown that stockperson attitudes can affect animal handling in general, and very likely also the handling of cattle [29,30]. However, while dirtiness in cattle has been proven to have negative consequences for farmers and for cattle, few efforts have been made to rectify the situation. As already described, dirty cattle have been the most common form of non-compliance with Swedish animal welfare legislation for years [1], and are a common welfare issue in other countries, so it is surprising that little seem to be known about prevention. The central competent authority responsible for animal welfare in Sweden is the Swedish Board of Agriculture (SBA), while the 21 regional County Administration Boards (CABs) are responsible for carrying out official animal welfare control. The CABs have a joint reference group aiming at improving animal welfare control in Sweden, and this reference group recently initiated a project to deal with the high numbers of dirty cattle observed in official animal welfare inspections. This was done at the request of the Animal Welfare Council, which is made up of representatives from the SBA, the CABs, and the Swedish Food Agency, and which has the task of making animal welfare control more legally secure, equitable, and effective. The project investigated the causes of dirtiness in cattle, with the aim of formulating strategies for future prevention, i.e., improving the compliance rate among farmers. A second aim was to identify why inspectors categorized some farms as compliant even though individual animals were assessed as being ‘not clean enough’.AimThe objectives of this paper are to examine the occurrence of dirty cattle on Swedish cattle farms and compliance with Swedish animal welfare legislation recorded in official animal welfare inspections during the described CAB project, and to assess farmers’ opinions on the reasons for dirtiness and key factors in keeping their animals clean. 2. Material and MethodsInspectors from 20 of the 21 Swedish CABs participated in the project. A total of 371 farms with dairy and/or beef cattle were inspected within the regular inspection activities of these CABs during two weeks in January/February 2020. Hence, farms were not selected specially for investigating purposes, but would have been inspected during the CABs regular activities regardless of this project. During inspections, particular focus was placed on animal cleanliness. The inspector recorded the number of individual cattle considered to be dirty among the total number of animals on the farm, which is currently not included in regular inspection protocols. To support their on-farm assessments, the inspectors used an inspection protocol and control guidelines issued by the SBA. These documents are already used during animal welfare inspections in Sweden and were not developed specifically for the project. The inspection protocol consists of a checklist where every legal demand is listed in different control points. For each control point, the inspector ticks if the farm is compliant or non-compliant, or if this control point was not assessed or is not applicable on this farm. The control guidelines consist of a written document that describes how an inspector should reason around and assess different requirements and situations [31]. For example, different kinds of dirtiness have different levels of severity, with e.g., mud being categorized as a less severe form of dirtiness than manure and urine dried in the coat. The guidelines also suggest different scales for rating the extent to which the body is dirty, but do not specify how these scales should be interpreted in relation to the legislation. The definition of ‘clean enough’ according to the guidelines is that “animals that are managed in such a way that manure, urine, and sometimes mud does not stick and dry on their bodies”. In the CAB project, the SBA complemented the ordinary guidelines with an additional written guideline stating that “an animal shall be assessed as ‘not clean enough’ if that animal is exposed to, or risks being exposed to, suffering, sickness or abnormal behaviours” [32].In addition to completing the ordinary inspection protocol, the CAB inspectors were also asked to fill in a questionnaire with questions relating to why the animals were clean or dirty (see translated version in the Supplementary Material, Survey S1). These questions were answered by the inspectors in dialogue with the farmer. On farms where ‘not clean enough’ animals were found, the farmers were also asked about what could be done to get the animals clean and what they would need to do to prevent dirty cattle. On compliant farms with no dirty animals, the farmers were instead asked about factors they believed to be important for successfully keeping their animals clean. If inspectors categorized a farm as compliant even though dirty animals were observed during inspection, they were asked to state the reason for that categorization. For each question, a set of possible answers was provided and a box for free text was available when the options were not sufficient to describe the inspector’s or farmer’s opinion. The inspectors were not asked to record type of housing system, category of animal (except beef or milk), outdoor access, type of main forage, or type of bedding material. The data and information written in the questionnaire was then transferred from the paper sheets to Excel 2016. For descriptive statistical analyses Stata/IC 14.2 (StatCorp, College Station, TX, USA) was used. 3. Results3.1. Types of Farms and InspectionsIn total, 371 farms were inspected, out of 14,798 Swedish cattle farms in total in the 20 regions [33]. Of these, 240 were dairy farms and 117 had beef cattle (Table 1). Nine farms had both dairy and beef cattle, while for the remaining five farms the type of production was not stated. The median herd size was 72, ranging from 2 to 2800 heads. Twenty-five percent of the inspections were performed in herds with ≤30 animals (n = 93). The majority of the inspections (n = 265; 76%) were regular animal welfare inspections, while the next largest group was follow-ups on previous inspections (n = 67, 19%). Only 13 inspections (4%) were initiated due to notifications of suspected mistreatment of animals. 3.2. Number of Dirty Animals and Legal Compliance on Animal CleanlinessFor 51% of farms inspected (n = 188), all animals were assessed as clean enough at the time of inspection (Table 1). For the remaining farms (n = 183, 49%), from one to 280 animals were assessed as not clean enough by the inspectors. The majority of these farms had ≤10 dirty animals (n = 122). However, in the inspection protocols completed by the inspectors, the cleanliness of the animals was assessed as compliant with Swedish animal welfare legislation on 248 farms (67%) and as non-compliant on 120 farms (33%). Thus, on 59 occasions (16%), the animal welfare inspector assessed the farm as compliant, although ‘not clean enough’ animals were observed during inspection. In three cases, information on compliance with cleanliness was missing. The most common reasons stated by the inspectors for categorizing a farm as compliant, despite having ‘not clean enough’ animals present, were: ‘few animals not clean enough’ (n = 43, 75%), followed by ‘minor non-compliance’ (n = 18, 32%) and ‘temporary non-compliance’ (n = 13, 23%) (Figure 1). On 26 occasions (46%), two or more of these reasons for compliance were stated. Information about the reason for compliance was missing for three farms, leaving 56 answers in total. On farms where ‘few animals not clean enough’ was mentioned as a reason for compliance, up to 15 animals that were not clean enough were found in the inspection. It was more common for a farm to be assessed as compliant if three or fewer animals were affected, but there were also occasions when a single dirty animal caused the farm to be assessed as non-compliant with the Swedish legislation (Figure 2). 3.3. Reasons for Having Dirty Animals in the HerdAccording to the inspectors and farmers, ‘insufficient coat trimming routines’ (n = 69, 38%) and ‘absence of routines for grooming dirty animals’ (n = 66, 36%), followed by ‘lack of bedding material’ (n = 52, 31%) and ‘insufficient cleaning of lying areas or walking alleys’ (n = 55, 30%) were the most common reasons for ‘not clean enough animals’ being found during inspections (Figure 3). In most cases, several likely reasons for insufficient cleanliness were stated. The behavior of individual animals in choosing a suboptimal place for lying was often given as an explanation when ‘other’ was recorded in the questionnaire (n = 20). Bad weather conditions were also a common ‘other’ reason mentioned, especially when animals were housed mainly outdoors (n = 14). A third group of explanations identified had to do with feeding (n = 7), for instance a sudden change in feed causing diarrhea in some animals, making it more difficult to keep the animals clean. 3.4. Farmers’ Needs and Keys to Success in Keeping Cattle CleanThe responses of farmers with animals categorized as ‘not clean enough’ to the question “What would you need to do to keep your animals clean?” and the responses of farmers with only clean animals to the question “How do you succeed in keeping your animals clean?” are shown in Figure 4. Several answers could be given. On both types of farms, management routines was the factor most commonly mentioned. On farms with clean animals, the trimming and/or grooming of animals was commonly mentioned as an example of a good management routine in keeping animals clean (n = 23). The largest discrepancy was found in the answers related to access to bedding material. Only 14% of the farmers with dirty animals (n = 26) mentioned ‘better access to bedding material’ as a need, while 63% of the farmers with all clean animals (n = 119) mentioned ‘good access to bedding material’ as a key factor in maintaining a good animal hygiene level on their farm. Better weather conditions were most often mentioned (n = 10) when the option ‘other’ was ticked for farms with dirty animals in the herd. Keeping a low or not too high stocking density was the most commonly stated ‘other’ factor in success for farms with clean animals (n = 15). 4. Discussion 4.1. Dirty Cattle Is Common This study showed that dirtiness in cattle is still a frequent form of non-compliance with animal welfare on Swedish farms, with no major change compared with previous reports [1]. In this study, the number of farms with dirty animals was almost 50%. However, the data do not show whether there were differences between different categories of animals, e.g., milking cows, dry cows, bulls, heifers, young cattle, and calves, or between housing systems, e.g., tie-stall, loose-housed, or outdoor cattle. Previous studies have shown a higher level of non-compliance [3] or worse welfare [8,34] in tie-stall housing compared with free-stall housing of dairy cows. In Sweden, tie-stall housing is still permitted for some categories of cattle, mainly dairy cows, with the total proportion of cattle kept in this system reported to be lower than 30% [35]. As mentioned previously, dirtiness is a welfare issue not only in Sweden, but also in other countries. Most countries do not have specific animal welfare regulations for cattle and do not set specific requirements on the appearance of individual animals, i.e., animal-based measures. This may be one reason why this welfare issue receives so little attention, as no violations are committed if there are no such requirements in the regulations. However, there are countries having national legislation requiring animals to have access to dry and clean lying areas. A resource-based requirement that is closely connected to the cleanliness of animals. However, in the future, the increasing focus and demand for animal-based indicators in both legislation and private assurance schemes [36,37,38] could perhaps draw attention to how animal-based requirements can best be implemented, measured, and enforced. Dirty cattle still receive little attention in the industry, despite having an effect on farm profitability. From the perspective of the farmer, the level of knowledge of the individual farmer on the effects of dirty animals may be insufficient. From the perspective of the industry, the issue has not been regarded as important to date, although in 2020, the media in Sweden began drawing attention to dirty cattle on dairy farms, which has created some movement within the industry. It has also been shown that animal welfare issues raised in the media can have an effect on consumer demand [39]. From the perspective of animal welfare inspections and the SBA, the measures taken, and why they are not functioning, can be discussed. Since this form of non-compliance returns year after year, the enforcement of compliance with the legislation needs to be considered. Enforcement work should examine why it is so difficult for cattle farmers to have ‘clean enough’ animals and how better compliance can be achieved. The SBA very recently presented an animal welfare strategy for Sweden where better compliance is one of the main goals, together with an official control that is effective and uniform [40]. 4.2. Challenges in Interpretation and Assessment The animal welfare legislation clearly sets a minimum level of animal welfare. Thus, at the time of an animal welfare inspection, animals do not need to be perfectly clean, but ‘clean enough’. Even when cattle in this study were assessed as ‘not clean enough’ during on-farm inspection, the inspector did not always categorize the farm as non-compliant with animal welfare legislation. On the contrary, on multiple occasions the inspector made the assessment that, despite not having clean enough animals, the farm was in compliance with the legislation since the herd as a whole was assessed as ‘clean enough’. The most common comment made in relation to this assessment was that ‘only a few animals’ were too dirty. Assessing only a proportion of animals in a herd is becoming more common in animal welfare assessments, e.g., in the Welfare Quality® system [41]. This way of measuring animal welfare is useful for detecting systematic problems at farm level but is in contravention of animal welfare legislation in Sweden, which does not focus on measuring overall animal welfare level at a farm, but on protecting the individual animal [42]. This is also clearly apparent from the SBA control guidelines, which state that “Suitable measures must be taken if one animal is affected, since the legislation is written from the perspective of an individual animal” [31]. Hence, if only one animal is not clean enough, this is clearly a case of non-compliance and measures must be taken by both the inspector and the farmer. This study revealed that animal welfare assessments can be rather subjective. This indicates a need for better guidance from the SBA on animal welfare inspections, in order to achieve high inter-observer agreement. This is in line with findings in France [43] and Denmark [44]. Use of vague wordings, such as ‘clean enough’, in regulations is quite common [11,45], and increases the risk of subjective assessments [45,46,47]. Further, compared with resource- and management-based requirements, animal-based requirements are more time-consuming to measure [48] and more difficult to assess in a uniform way [49], and hence require much training to perform [50,51]. However, animal-based requirements and indicators are important in coming closer to measuring the actual welfare status of the individual animal, and not only considering risk factors for welfare through resource- and management-based requirements [51,52]. Use of animal-based indicators is one way to evaluate whether animal housing and management are functioning as the legislation intends. The European Commission is considering increased use of animal-based measures in EU legislation [36]. Animal-based requirements are valuable and important but need special attention when assessing compliance. The written SBA guidelines currently used in Sweden do not seem to provide adequate support for animal welfare inspectors or are perhaps not fully accepted by all inspectors. If inspectors question the methods they are required to use, some might be expected not to strictly follow the guidelines [43]. Therefore, new strategies and methods need to be developed and implemented in order to achieve more uniform assessment of compliance and non-compliance in relation to the cleanliness of cattle. Apart from official animal welfare inspectors, farmers also need to understand how to interpret ‘clean enough’ and when ‘not clean enough animals’ are considered to be in compliance or non-compliant with the legislation. In the free text space in the questionnaire, one of the inspectors commented ‘Farmer disagrees on animals being dirty’, showing that the interpretation of cleanliness is not always straightforward. Such statement also raises questions concerning why some farmers does not see dirtiness as a problem. Is it due to lack of education or a case of “home blindness”? In addition, professional farm advisors and veterinarians need to know the legally acceptable level of cleanliness in cattle. Staaf Larsson and co-workers [53] found that professionals with different backgrounds and experiences viewed and assessed dirtiness in cattle differently. Further, with the use of private standards and quality assurance schemes increasing in society, it is important to include these actors in the discussion, since private standards are commonly based at least partly on the requirements set in national legislation [11,54]. There will also be others making assessments on what is ‘clean enough’ and when a farm is compliant or not. Previous studies have shown that even when a requirement in a private standard is expressed in the same way as in national legislation, the way of measuring may be quite different [3,11]. According to Lundmark and co-workers [55], policymakers need to consider the whole animal welfare inspection arena, and not simply their own regulation or assurance scheme as an isolated entity.4.3. Actions and AttitudesThere is a risk that the uncertainty in how ‘clean enough’ is assessed will result in passive farmers. In this study, farmers with dirty cattle cited different management routines, such as insufficient coat trimming and grooming, insufficient cleaning of lying areas or walking alleys, and lack of bedding material as the most common reasons for dirty cattle. Farmers with ‘clean enough’ animals believed to a higher degree that management routines, sufficient staffing levels, bedding material, and good physical/mental health are key factors in keeping clean animals. Farmers with ‘not clean enough animals’ believed that shorter queues to slaughter would solve the problem of dirty animals to a higher degree. Dirtiness in cattle is a multifaceted problem according to the literature, and previous studies have identified risk factors in housing and interior design that contribute to hygiene problems [5,23,26]. Interestingly, and somewhat surprisingly, in this study housing conditions were not mentioned as often as various management routines when farmers were asked about the reasons for having clean or dirty animals. Previous studies have concluded that cattle need enough resting time for prevention of lameness and hock lesions [21,56,57], indicating that the issue of dirty cattle should be taken very seriously. Dirty cattle could be an indicator of animals not having a suitable dry bedding area, which can result in animal welfare-related skin lesion issues, and higher prevalence of lameness and hock lesions. Thus, dirtiness in cattle is a negative economic factor for farmers and can also be an indicator of poor overall herd health.In this study, several farmers blamed dirtiness on the behavior of individual animals in choosing a suboptimal place for lying. If cattle are free to choose, they will choose a lying area that is soft [58,59], dry and clean [60,61,62], and has a lot of bedding material [63]. Since lying behavior in cattle is strongly linked to comfort and welfare [64,65,66], it is unlikely that cattle will voluntarily choose to lie down somewhere wet and dirty. Therefore, the reasons behind animal behavior in this regard must be evaluated and suitable measures taken by the farmer.Finally, there is a possible interaction between the farmer’s attitude to animals and the handling regime. Several studies have shown that attitudes to animals influence the way in which animals are handled [30]. Hemsworth and colleagues [30] showed that attitudes, and handling, can be improved by providing species-specific education on animal care. Glanville and co-workers [67] suggested that in efforts to achieve a change in attitudes, one important component must be to inform farmers about the consequences of their handling and about the impacts of a change in handling regime on the animals, but also on the farmer’s situation. It is possible that the issue of dirtiness in cattle can be prevented through suitable education and training. 5. ConclusionsThis study showed that cattle in Swedish dairy and beef production are currently not clean enough, and dirty cattle was present at 49% of the farms Questionnaire responses clearly indicated that inspectors and farmers both believe that cleanliness in cattle is primarily connected to routines on the farm, e.g., trimming and grooming of animals. Both the farmers with clean cattle and farmers with cattle not clean enough thought management routines was the main key factor to keep animals clean. This is a promising result from an animal welfare point of view since management routines and strategies can be improved with the right measures. However, there was also some discrepancies between farmers’ views. Farmers with clean animals saw good access to bedding material as one of the keys to success for having clean animals, while few farmers with not clean enough animals did see better access to bedding material as a key factor to improve cleanliness.Despite dirty animals being present, animal welfare inspectors did not always categorize the farm as non-compliant. We conclude that a better official guidance for inspectors and farmers is needed, in order to support their assessments of when dirtiness in cattle represents non-compliance with animal welfare legislation.

animals : an open access journal from mdpi

10.3390/ani11113232

PMC8614345

Clinical cases associated with staphylococci infections are common among dogs and cats. There is evidence to suggest that staphylococci infections are increasingly becoming unresponsive to commonly used antimicrobials. This negatively impacts the ability of these infections to be treated successfully. Although resistance among these organisms has been linked to several factors, including sharing the same mechanism of action or belonging to the same group, there is evidence to suggest that cross resistance can occur between unrelated antimicrobials. The findings of this study not only confirm that antimicrobials that belong to the same group share the same mechanism of resistance and similar antimicrobial efficacy against staphylococcal infections, but also show that cross resistance occurs between unrelated antimicrobials. This should be taken into consideration when selecting antimicrobials for inclusion in the susceptibility testing panel as well as for the treatment of staphylococci infections.

Cross-resistance occurs between antimicrobials with either similar mechanisms of action and/or similar chemical structures, or even between unrelated antimicrobials. This study employed a multivariate approach to investigate the associations between the efficacy profile of antimicrobials and the clustering of eleven different antimicrobial agents based on their efficacy profile. Records of the susceptibility of 382 confirmed Staphylococcus species isolates against 15 antimicrobials based on the disc diffusion method were included in this study. Tetrachoric correlation coefficients were computed to assess the correlations of antimicrobial efficacy profiles against Staphylococcus aureus. Principal components analysis and factor analysis were used to assess the clustering of antimicrobial susceptibility profiles. Strong correlations were observed among aminoglycosides, penicillins, fluroquinolones, and lincosamides. Three main factors were extracted, with Factor 1 dominated by the susceptibility profile of enrofloxacin (factor loading (FL) = 0.859), gentamicin (FL = 0.898), tylosin (FL = 0.801), and ampicillin (FL = −0.813). Factor 2, on the other hand, was dominated by the susceptibility profile of clindamycin (FL = 0.927) and lincomycin-spectinomycin (FL = 0.848) and co-trimazole (FL = −0.693). Lastly, Factor 3 was dominated by the susceptibility profile of amoxicillin-clavulanic acid (FL = 0.848) and cephalothin (FL = 0.824). Antimicrobials belonging to the same category or class of antimicrobial, tended to exhibit similar efficacy profiles, therefore, laboratories must choose only one of the antimicrobials in each group to help reduce the cost of antimicrobial susceptibility tests.

1. IntroductionInfection with Staphylococcus species is common in domestic animals including dogs and cats [1,2]. These organisms cause various clinical conditions that include pyoderma, otitis, and wound infections [3,4,5,6]. However, there are increasing reports of antimicrobial resistance among Staphylococcus isolates in veterinary settings [1,7]. These are likely to complicate treatment outcomes due to treatment failures and the resultant poor prognosis, high morbidities, and mortalities [8].The increasing prevalence of resistance has been attributed to over prescription, improper prescription, and acquisition of resistant genes through a number of mechanisms including plasmids [9,10,11]. There is also evidence of association between the resistance profile of antimicrobials that belong to the same category and also with antimicrobials that belong to other categories [12,13,14]. For example, cross resistance has been reported between members of the β-lactams, fluroquinolones, and aminoglycosides [15]. Cross-resistance has also been reported against antimicrobial drugs to which bacteria have not previously been exposed [16]. This may develop without the target mutations or may develop mediated by mutation in the target resistance protein like in the case of fluoroquinolone. The latter has been associated with resistance among multiple non-quinolone. Research by Pal et al. [17] on cross-resistance among unrelated antimicrobials provides information on the long-term efficacy of novel antimicrobial compounds. In view of this, studies that investigate cross resistance among related and unrelated antimicrobial groups are needed.Although several studies have investigated associations between the efficacy profiles of antimicrobials in relation to Staphylococcus isolates in human medicine, similar studies are lacking in veterinary medicine in South Africa. In addition, there are limited studies that have investigated this phenomenon using statistically rigorous methods such as multivariate techniques. This study investigated the association between the efficacy profiles of antimicrobials evaluated against Staphylococcus clinical isolates. Study findings contribute to improved understanding of cross resistance among clinical isolates and can be used to determine antimicrobials for use in veterinary practice, especially in low resource settings.2. Materials and Methods2.1. Data SourceThis study used retrospective secondary data of the susceptibility profile of 382 confirmed Staphylococcus species isolates from canine clinical cases presented at a veterinary academic hospital located in Pretoria between January 2007 and December 2012. The culture and sensitivity analysis were conducted by the bacteriology laboratory of the veterinary academic hospital. The dataset was assessed for duplicate entries, missing data, and inconsistencies, such as improbable values.2.2. Antimicrobial Susceptibility TestingAll the isolates were subjected to antimicrobial susceptibility testing (AST) against a panel of 15 drugs using the disc diffusion method following the guidelines of the Clinical and Laboratory Standards Institute [18,19,20,21,22,23,24]. The panel included the following antimicrobials: 30 μg amikacin, 30 μg doxycycline, 5 μg enrofloxacin, 10 μg gentamicin, 10 μg ampicillin, 10 μg penicillin G, 25 μg trimethoprim-sulfamethoxazole (co-trimoxazole), 30 μg chloramphenicol, 30 μg cephalothin, 30 μg kanamycin, 2 μg clindamycin, 100 μg lincospectin (LS100), 5 μg orbifloxacin, 20/10 μg amoxicillin/clavulanic acid and 15 μg tylosin. The laboratory that supplied the data classified the susceptibility profile of the isolates into three categories (i.e., susceptible, intermediate, or resistant) in accordance with the Clinical and Laboratory Standards Institutes [18,19,20,21,22,23,24]. However, for the purposes of this study, intermediate and resistance isolates were recoded as nonsusceptible for all subsequent analyses.2.3. Data Analyses2.3.1. Correlation AnalysisTetrachoric correlation coefficients were computed to assess the relationship between the susceptibility profiles of different antimicrobials. The tetrachoric correlation coefficients were computed in this study because of their appropriateness for assessment of correlations between dichotomous variables [25] and are indicated in situations where consistency measures of reliability are preferred to agreement measures [26]. In this study, pairs of antimicrobials with correlation coefficients of ≥0.7 were classified as highly correlated. If this was observed between agents belonging to the same antimicrobial category, only one of the two was selected for inclusion in the subsequent principal component analysis (PCA) and factor analysis.2.3.2. Principal Components and Factor AnalysesPrincipal components analysis (PROC PRINCOM) and factor analysis (PROC FACTOR), implemented in SAS 9.4 (SAS Institute Inc., Cary, NC, USA), were used to assess the relationship between the efficacy profiles of antimicrobials against Staphylococcus species. Eigenvalues >1 were used to determine the number of factors to be retained in the factor analysis. In addition, the scree plot was used to visualize the factor numbers and associated eigenvalues. Orthogonal axis rotation (varimax) was applied to the factors to allow for easy interpretation of the interrelationships between the antimicrobial agents. The reliability of the items was assessed using the McDonald’s omega coefficient test implemented in JASP software version 0.14.1.0 (University of Amsterdam, Amsterdam, the Netherlands) [27]. Variables with low communality values were removed from the PCA.3. ResultsHigh numbers of nonsusceptible isolates were observed for ampicillin (58.9%), penicillin (55.5%), lincospectin (44.5%), and clindamycin (37.43%). However, low numbers of nonsusceptible isolates were observed for aminoglycoside (9.2%), tetracyclines (15.7%), fluoroquinolones (%), potentiated sulfonamides (17.02%), amphenicols (11.34), amoxicillin-clavulanic acid (12.57%) and macrolide (10.47%) (Table 1).Strong correlations were observed between efficacy profiles of the following antimicrobials: amikacin vs. gentamycin (r = 0.79), amikacin vs. kanamycin (r = 0.72), kanamycin vs. gentamycin (r = 0.93), ampicillin vs. penicillin (r = 0.96), enrofloxacin vs. orbifloxacin (r = 0.91), and lincospectin vs. clindamycin (r = 0.79) (Table 2).Three factors had eigenvalues >1 and were, therefore, extracted (Table 3). These factors together accounted for 85% of variation in antimicrobial nonsusceptibility (Figure 1, Table 3).Factor 1 was dominated by enrofloxacin (factor loading (FL) = 0.859), gentamicin (FL = 0.898), tylosin (FL = 0.801), and ampicillin (FL = −0.814). Factor 2 was dominated by clindamycin (FL = 0.927), lincomycin-spectinomycin (FL = 0.848) and co-trimazole (FL = −0.693). Lastly, Factor 3 was dominated by amoxicillin-clavulanic acid (FL = 0.848) and cephalothin (FL = 0.824) (Table 4). McDonald’s omega values indicated good internal reliability of the items (Table 5).4. DiscussionThis study investigated the interrelationships between the efficacy profiles of antimicrobial agents against Staphylococcus isolates. Strong correlations between the efficacy profiles were observed between antimicrobials that belong to the same category. For example, strong correlations were observed between efficacy profiles of members of each of the aminoglycoside, fluoroquinolone, and penicillin categories of antimicrobials. This was expected because antimicrobials that share a similar mechanism of action or chemical structure are known to exhibit cross resistance [11,13,28,29,30]. Therefore, if a member of one of the categories ceases to be efficacious against a pathogen, other members of that category are most unlikely to be efficacious against the same organisms. In view of this, during in vitro testing, antimicrobials belonging to the same group or class of antimicrobials should not be included in the testing panel. Only one antimicrobial from the group should be selected to represent other members of the group sharing a similar chemical structure and/or mechanism of action. This would be a cost-cutting measure with the potential to make antimicrobial sensitivity testing more accessible and affordable, especially in low resource settings. Furthermore, the results of the study suggest that clinicians should not consider antimicrobials for the treatment of Staphylococcus aureus if such antimicrobials belong to the same group of antimicrobials against which low efficacy against Staphylococcus isolates have been observed. This is likely to result in treatment failure.Results of factor analysis showed that enrofloxacin, gentamycin, tylosin and ampicillin clustered together, suggesting a similarity in the efficacy profiles of these groups of antimicrobials. However, ampicillin compared to the other antimicrobials tended to load negatively. This suggests that unlike the other groups of antimicrobials, ampicillin had an efficacy profile that was opposite in relation to Staphylococcus species. This could be explained by the high proportion of Staphylococcus isolates in this study that were resistant to ampicillin as compared to the other three antimicrobials. Furthermore, this disparity could be due to differences in the mechanisms of action, with ampicillin targeting the cell-wall while the others targets nucleic acid or protein synthesis [29].The clustering of antimicrobials from different classes and with different mechanisms of action observed in the preceding paragraph, suggests collateral sensitivity. Collateral sensitivity or cross resistance has previous been reported in other studies [12,13,14,15] and may develop without the target mutations or via mutation in the target resistance protein. Therefore, findings of this study support the evidence of cross resistance or collateral sensitivity among clinical isolates.Amoxicillin-clavulanic acid and cephalothin also clustered together. This was expected, given that both antimicrobials belong to the β-lactam group and both are highly efficacious towards β-lactamase producing Staphylococcus species [31,32]. Likewise, the clustering of clindamycin and lincospectin was anticipated, given that both antimicrobials belong to the same category of antimicrobials called lincosamide. These two antimicrobials are known to be highly efficacious against methicillin-resistant Staphylococcus aureus (MRSA) and multidrug resistant staphylococci [32,33].5. ConclusionsIn this study it was observed that antimicrobials in the same category or class, share similar efficacy profiles with respect to Staphylococcus species. Therefore, it is recommended that when performing susceptibility analysis, laboratories should only include one member of each category or class of antimicrobials to help reduce the cost of antimicrobial susceptibility tests, especially in low resource settings. Likewise, to minimize treatment failures, clinicians are advised not to prescribe antimicrobials belonging to the same group of antimicrobials if one member of that category exhibits reduced efficacy against Staphylococcus species.

animals : an open access journal from mdpi

10.3390/ani11061836

PMC8235278

Uncoupling protein-1 (UCP1) plays important roles in the energy balance and regulation of metabolism and in the body temperature regulation. In this survey the correlation among UCP1, body weight, rectal temperature and lipid profile was assessed in the adult horse. The findings gathered from the current survey showed that UCP1 values are not related with body weight and temperature in studied animals, but they seem to be linked to pathways involved in lipid and lipoprotein metabolism.

This study aimed to evaluate the possible relationship among UCP1, body weight, rectal temperature and lipid profile in the horse. Thirty clinically healthy Italian Saddle geldings (6–10 years old) were enrolled after the informed owners’ consent. All horses were blood sampled and their body weight and rectal temperatures were recorded. On the sera obtained after blood centrifugation the concentration of UCP1, total lipids, phospholipids, non-esterified fatty acids (NEFAs), triglycerides, total cholesterol, high density lipoproteins (HDLs), low density lipoproteins (LDLs) and very low density lipoprotein fraction (VLDLs) was evaluated. Pearson’s correlation analysis was applied to assess the possible relationship between serum UCP1 concentration and the values of body weight, rectal temperature and lipid parameters. Serum UCP1 concentration showed no correlation with body weight, rectal temperature, HDLs and LDLs values, whereas it correlated negatively with serum total lipids, phospholipids, NEFAs, total cholesterol, triglycerides and VLDLs values (p < 0.0001). The findings suggest that in the adult horse the role of UCP1 is linked to the lipid metabolism rather than to thermoregulation.

1. IntroductionThe thermogenin, also named uncoupling protein-1 (UCP1), is involved in energy balance, metabolism regulation and thermogenesis pathways [1,2]. This protein is mainly expressed in brown adipose tissue (BAT), the main site of adaptive thermogenesis known also as non-shivering heat production, in response to environmental temperature or diet [3,4].Unlike the white adipose tissue (WAT), containing large lipid droplets and few mitochondria [2,4], the BAT, heavily innervated by sympathetic nerves, consists of fat cells with numerous mitochondria and small lipid droplets [2]. Although it had been established that BAT declines after puberty become rare in adults, nowadays the hypothesis that BAT can be found in adults has been consolidated [5]. Moreover, recently it has been demonstrated that BAT is both functionally and metabolically highly active in adults, especially after chronic exposure to cold [5]. Particularly, following proper stimulation, white adipocytes can acquire typical features of brown fat cells, including UCP1 expression [6], which may be used by the organism to increase metabolic energy expenditure. This process is termed “browning” of WAT and the implicated white adipocytes are indicated as beige fat cells which gain UCP1 expression ability and thermogenic potential [7,8,9,10,11]. Though the contribution of beige fat cells to systemic adaptive thermogenesis is under debate, the specific uncoupling of adipocyte mitochondria remains an attractive target for the development of anti-obesity drugs in humans and animals, and also to improve and renew the knowledge on the metabolic strategies implemented by organisms under certain conditions [1,12].In the veterinary field, the investigation of UCP1 levels and their relationship with an animal’s metabolism has sparked a great interest, especially in livestock, under certain physiological conditions, such as growth or pregnancy, when metabolic pathways are constantly changing due to the metabolic adjustments needed for homeostasis maintenance. For instance, lower UCP1 levels are found in growing kids compared to adult goats [13], suggesting that this protein may serve to prevent the excessive energy loss in young animals. Also, as Arfuso et al., [14] described, UCP1 is involved in the lipid metabolism of periparturient mares, contributing to their metabolic adaptation.Although data from experiments with UCP1 in reconstituted systems tend to indicate that fatty acids are essential for UCP1 function and that this protein increases lipolysis and fatty acid oxidation [15,16,17], there is a paucity of information regarding the relationship between the serum UCP1 levels and lipid profile in adult mammals including the horse.To bring an additional contribution the aim of the current study was to assess whether there is a correlation between the UCP1 concentration and the values of body weight, rectal temperature and the various lipid components in the adult horse.2. Materials and Methods 2.1. Animals and Experimental Design. All the treatments, housing and animal care during the study were in accordance with the standards recommended by the European Directive 2010/63/EU for animal experiments. After the informed consent of the owners, 30 Italian Saddle geldings (6–10 years of age; mean body weight 449 ± 10 kg) were included in the study. All horses were managed similarly in the same place (a horse training center in Sicily, Italy, latitude 38°10′ 35′′ N; longitude 13°18′14′’ E), housed in individual boxes (3.5 × 3.5 m), subjected to the same environmental conditions (natural photoperiod, mean temperature 25 ± 4 °C, mean relative humidity 65 ± 5%).The animals were fed twice per day (08.00 a.m.; 06.00 p.m.) with a total amount of approximately 2.5% of their body weight in dry matter intake (forage:concentrate ratio 70:30) and water was available ad libitum.All the animals were clinically healthy (based on a thorough clinical examination) and free of internal and external parasites.2.2. Sampling Procedures and Laboratory Analysis The same operator measured the body weight of each horse by means of a weighting platform (PS3000HD Heavy Duty Floor Scale, Breckwell, UK), the rectal temperature (RT) of the horses using a digital thermometer (HI92704, Hanna Instruments Bedfordshire, UK) inserted 15 cm in the rectum and collected blood samples in two 10 mL tubes with clot activator (Terumo Corporation, Tokyo, Japan) by jugular venipuncture at the same time, before the morning feeding of the animals. The tubes were transported to the laboratory in refrigerated bags.The samples from the first tube (containing clot activator) were allowed to clot for 2 h at room temperature, then they were centrifuged at 1000× g for 20 min. The obtained sera were stored at −20 °C until analysis. These serum samples were analyzed to estimate the concentration of mitochondrial uncoupling protein-1 (UCP1) using an ELISA kit specific for equine species (Horse Uncoupling Protein-1, Mitochondrial (UCP1) ELISA Kit, Cat.No: MBS066225, MyBioSource, Inc. San Diego, California, USA) with a microwell plate reader (Sirio, SEAC, Florence, Italy). All calibrators and samples were run in duplicate and the samples exhibited parallel displacement to the standard curve for ELISA analysis. The sensitivity of this kit was 10 pg/mL and both the intra- and the inter-assay coefficients of variation for UCP1 were at <15%.The samples from the second tube (containing clot activator) were centrifuged at 1300× g for 10 min, within 30 min from the collection, and the obtained sera were stored at −20 °C until analysis. The obtained samples were analyzed to estimate the concentration of total lipids, phospholipids, triglycerides, total cholesterol, high density lipoproteins (HDLs) and low density and lipoproteins (LDLs) using commercially available kits with an automated analyzer UV Spectrophotometer (model Slim SEAC, Florence, Italy). Very low density lipoprotein fraction (VLDLs) was estimated as one-fifth of the concentration of triglycerides [18].2.3. Statistical analysis. All results were expressed as mean values ± standard deviation (±SD).The normal distribution was proven by the Kolmogorov-Smirnov test performed on all data (p > 0.05). Pearson’s correlation coefficients were computed to evaluate the relationship between serum UCP1 concentration and the values of body weight, rectal temperature, serum lipid and lipoprotein indices in enrolled horses. A linear regression model (y = a + bx) was applied to determine the degree of correlation between these parameters. p values less than 0.05 were considered statistically significant. The statistical analysis was performed using the statistical software Prism v. 4.00 (Graphpad Software Ltd., San Diego, CA, USA, 2003).3. ResultsThe serum UCP1 concentration did not correlate with body weight, rectal temperature, or HDLs and LDLs values, whereas it resulted negatively correlated with the serum values of total lipids, phospholipids, NEFAs, total cholesterol, triglycerides and VLDLs (Table 1). These findings were confirmed by the linear regression model results (Figure 1 and Figure 2).4. DiscussionAlthough the role of UCP1 in regulation of metabolic and energy balance and in body temperature control is widely recognized [19], conflicting results on the presence of this protein in adult animals are shown by the scientific community. However, the recent discovery of the existence of the beige adipocytes showing both an expression panel partially typical of white adipocytes and a role in thermogenesis via expression of UCP1 [20,21,22], opened new insights in this field. The function, origin and localization of these fat cells are still not fully described and the literature currently available is limited to humans and mice.According to the findings gathered in the current study, the UCP1 levels are not correlated with body weight values recorded in the investigated horses. In agreement with these results, a survey on genetically modified mice in which the UCP1 gene has been silenced (Ucp1−/− mice) suggested that the effect of UCP1 thermogenesis on body weight might be a consequence of the mechanisms carried out to body temperature maintenance [23].The rectal temperature values recorded in the investigated horses lacked of correlation with the values of UCP1. Contrary to the findings herein observed, a statistically significant negative correlation between the values of rectal temperature and UCP1 was reported in a previous study carried out on adult goats and growing kids [13], where it has been hypothesized that, due to the UCP1 enrolment in the thermogenic process, higher levels of this protein would indicate a higher thermogenesis level leading to a body temperature rise. The lack of correlation between the values of UCP1 and rectal temperature found in the current study could suggest that other thermogenesis tools are involved in the regulation of body temperature in horses. Moreover, the environmental temperature conditions throughout the study (approximately 25 °C) were within the thermo-neutral zone, and this could have limited the expression of UCP1 and, consequently, the results regarding the relationship between UCP1 and body temperature.Noteworthy, in our previous study carried out on horses, a negative relationship between UCP1 levels and the circadian clock gene Per2 involved in cold-induced adaptive thermogenesis has been found [24]. It has been supposed that this correlation is linked to the role of Per2 and UCP1 in the lipid metabolism, rather than to their role in thermoregulation. As a matter of fact, UCP1 is one of the targets involved in lipid metabolism within the peroxisome proliferator-activated receptor gamma (PPARγ). Per2 is the only clock gene protein able to interact with PPARγ, a master regulator of adipogenic differentiation and lipid metabolism [24].In the current study UCP1 values displayed a significant correlation with serum total lipids, phospholipids, NEFAs, total cholesterol, triglycerides and VLDLs values, suggesting that thermogenin could be an integral component of cellular energy control, and that mechanisms of coordinated regulation may exist for UCP1 and other enzymes of oxidative metabolism in adult horses [19]. The correlation results were confirmed by a linear regression model applied on the data that highlighted a negative correlation between UCP1 concentration and the values of lipid parameters. The negative correlations found between the values of UCP1 and lipids and lipoproteins herein investigated could be explained by the fact that the main physiological activators of UCP1 are the fatty acids resulting from hormone-stimulated lipolysis [19,25]. Particularly, the reaction cascade activated by sympathetic nervous system terminals ends with the activation of protein kinase A leading to triacylglycerol lipase phosphorylation and activation that leads to the conversion of triacylglycerols into free fatty acids, which, ultimately, activate UCP1 [23]. An in vitro research, focused in the study of direct metabolic effects of forced UCP1 expression in the white adipocytes [22], reported a metabolic phenotype consistent with the energy dissipative function of UCP1, as suggested by previous reports [26,27,28,29,30], and by the current study. Indeed, a reduced accumulation of triglycerides has been observed and it has been hypothesized that the reduction in intracellular lipid by UCP1 expression could reflect a down-regulation of fat synthesis [22]. Although studies carried out in humans suggested that UCP1 gene polymorphisms are associated with the obesity pathogenesis and with the disequilibrium of lipid metabolism [26,27,28,29,30], these insights are not confirmed in other studies carried out on humans [31,32,33], as well as by the findings obtained in the current study, which showed no correlation between the values of UCP1 and the levels of HDLs and LDLs in investigated horses.5. ConclusionsThe findings gathered in the current study showed that UCP1 values are not correlated with body weight and temperature in adult horses, but they seem to be associated with lipid and lipoprotein metabolism pathways in adult horses. It might be hypothesized that thermogenin is an integral component of cellular energy control, and that mechanisms of coordinated regulation may exist for UCP1 and other enzymes of oxidative metabolism. Further investigations on this special but still little-known protein are worthy of interest in order to better clarify its functions and its possible involvement in metabolic/energetic adjustments that occur during the different life-phases of the animals and/or under stress conditions.

animals : an open access journal from mdpi

10.3390/ani11072071

PMC8300343

Precociousness of heifers kept in mixed beef herds with young and adult bulls leads to unwanted pregnancies. Inbreeding, premature calving followed by dystocia and a high stillbirth rate are the consequences. As an alternative, such heifers are slaughtered during the resulting pregnancy. The slaughtering of pregnant animals is an ethical problem, which is strongly criticized by consumers and animal welfare organizations. Therefore, the aim of this study was to postpone puberty in female beef calves housed in Swiss mixed herds to avoid pregnancy until scheduled slaughter at 11 months of age. We used a vaccine (Improvac®) that induces antibody production against sexual hormones, thereby suppressing the reproductive cycle. Monthly progesterone analysis in blood was performed to decide whether the cycle had already started. The results proved, that vaccinated female beef were not coming in heat until 11 months of age compared to the unvaccinated control group, which came in heat earlier. In conclusion, vaccination with Improvac® is an animal-friendly, non-invasive and reliable method to avoid early pregnancy in heifers as well as the slaughter of pregnant cattle.

Precocious puberty in beef heifers can result in unwanted pregnancies due to accidental breeding by farm bulls. Inbreeding, premature calving followed by dystocia and a high stillbirth rate or slaughtering of pregnant heifers are the consequences of this behaviour. The aim of the study was to postpone puberty by using Improvac®, an anti-GnRH vaccine. Therefore, n = 25 calves were twice vaccinated, once at the age of 5 and then at 6.5 months. n = 24 calves served as unvaccinated case controls. The onset of puberty was assigned if progesterone analysis in the blood exceeded 1 ng/mL. Progesterone values were excluded if the corresponding serum cortisol levels were ≥60 nmol/L. Our target was met, as in the vaccinated group none of the calves exceeded a progesterone value >1 ng/mL until the scheduled age of slaughter at 11 months and only 12.5% of the animals exceeded a progesterone value of 1 ng/mL over the whole measuring period (>400 days) compared with 56.5% of the calves in the control group. In conclusion, the favourable results from our study using the vaccine Improvac® represent an animal-friendly, non-invasive and reliable way to avoid early pregnancy in heifers as well as the slaughter of pregnant cattle.

1. IntroductionThe herd size of beef cow-calf operations in Switzerland are small where male and female animals of all age groups are housed together. There is often no possibility to split up the herd (only one free stall housing available, not many different pastures per farm). This often leads to young heifers being inadvertently impregnated by farm bulls. Animal welfare and ethics are of concern when these young pregnant heifers are presented at slaughter [1]. A German cross-sectional study revealed that slaughtering pregnant cattle is a widespread practice with many fetuses in the second or third trimesters [1]. Due to increased consumer concern, a study in a Swiss abattoir was initiated that reported a pregnancy prevalence of 5.67% cattle pregnant > 5 months and 0.67% 7 to 9 months pregnant (BLV: Projekt Schlachtung von trächtigen Rindern-Prävalenz und Gründe der Schlachtung). Based on these results from that study, at present, farmers have to declare the pregnancy status in cows later than 5 months post-partum (p.p.) and heifers older than 15 months when slaughtered (Proviande: Fachempfehlung zur Vermeidung des Schlachtens von trächtigen Tieren der Rindviehgattung).Puberty in a calf is defined as a measure of a physical development leading to sexual maturity. In cattle, puberty mostly occurs between 8 and 12 months of age, but there is a large variation depending on the breed [2,3]. Therefore, declaring pregnancy status at an age > 15 months of age might be too late in early mature beef heifers. The onset of puberty in calves is defined as the age at first ovulation and an increase in plasma progesterone concentrations above 1 ng/mL [4]. As rectal palpation is not feasible in calves/young heifers, analysis of serum progesterone might be used to determine the beginning of puberty (serum levels above 1 ng/mL [4,5]). However, it was also demonstrated in ovariectomised cattle, that under a certain stress condition, animals might secrete additional progesterone (1.8 ± 0.7 ng/mL) and cortisol (11.5–44.3 ng/mL (≈31.7–122.2 nmol/L)) from the adrenal cortex [6].In dairy heifers, the early onset of puberty is one of the breeding goals to reduce costs for the rearing period [7]. Different studies focused on intensifying the feeding of calves in the rearing period in order to advance the age at puberty [3,8,9,10]. In beef heifers intended for slaughter, though, precocious puberty is not desired due to the risk of unwanted pregnancy. In addition, in replacement heifers, inadequate skeletal maturity can be a problem if the age at first calving is <24 months [11] resulting in increased rates of dystocia [3].If puberty in calves should irreversibly be avoided, surgical methods, such as ovariectomy might be used [12,13]. As an alternative, the suppression of the hypothalamo-gonadal axis using an anti-GnRH vaccine is an animal-friendly option that is even reversible [14,15,16]. Immunocastration suppresses sexual behaviour in male and female cattle, sheep, pigs and horses [17,18,19,20,21,22,23,24]. It also reduces aggressive or undesirable behavior of pigs [25] mares [26] and bulls [27].Our goal was to postpone puberty in beef calves by using an anti-GnRH vaccine at 5 and 6.5 months of age compared to untreated calves from the same herd. Consequently, the vaccinated group should not be pregnant when slaughtered at ±11 months.2. Materials and Methods2.1. Animals, Care and Housingn = 49 female calves from 14 farms were included in this study (n = 24 in the control group (C) and n = 25 in the vaccinated group (V)). The breeds were Limousin, Angus, Swiss Fleckvieh and mixed breeds. All herds consisted of female and male calves with breeding bulls and teaser bulls and were housed in pens with straw bedding or cubicle housing systems. The cattle had free access to water and to pasture, depending on weather and temperature. Diets consisted of grass, hay, corn and/or grass silage. Data were collected from May 2019 to August 2020.2.2. Treatment with Anti-GnRH VaccineImprovac® is an anti-GnRH vaccine (Zoetis Schweiz GmbH, 2800 Delémont, Switzerland) containing an analog of GnRH linked to a carrier protein combined with a synthetic aqueous adjuvant (200 μg of GnRH-protein-conjugate per mL). Heifer calves in group V, received 2 doses (initial and booster) of Improvac® 6 weeks apart (initial vaccination was at 5 months ± 14 days). The dosage used for both the initial and the booster vaccination was 400 μg of GnRH-protein-conjugate (2 mL of Improvac®). All injections were administered subcutaneously on the right side of the neck. In group C, the animals received 2 mL of 0.9% saline solution subcutaneously on the right side of the neck (twice, 6 weeks apart).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 council): BE 73/19.2.3. Including and Excluding CriteriaCalves were randomly assigned to either group (1–2 calves per group and farm). Before the injections, the calves were subjected to basic clinical examination including rectal body temperature, heart rate, respiratory rate, auscultation of lungs and gastrointestinal tract, and presence of umbilical remnants. If the rectal temperature was >39.5 °C and/or auscultation revealed pathological findings (bronchopneumonia, audible heart anomalies), the respective calf was excluded. After vaccination, the injection site was inspected for signs of tissue reaction by the farmers twice daily. If swelling was observed and if appetite decreased (food intake subjectively judged by farmers) to ≤75% of the normal amount, the veterinarian was notified.2.4. Blood Sampling, Progesterone and Cortisol ValuesIn all calves, blood was sampled during both vaccination visits and then every 4 weeks until slaughter or until the end of the study (4–9 samples/animal). The blood samples were collected by venipuncture from the jugular vein or the V. caudalis mediana into serum tubes (S-Monovetten 9 mL with Clot Activator, Sarstedt, Nümbrecht, Deutschland). After a clotting time of 1–3 h at room temperature, samples were centrifuged (4000× g, 10 min), and serum was then stored at −18 °C for later analysis. Progesterone concentration was measured by chemiluminescence assay on an Immulite 2000 XPi (Siemens Healthcare, Zürich, Switzerland). The analytical sensitivity was 0.1 ng/mL and the measurement range was 0.2–40 ng/mL. The assay was performed according to the manufacturer’s instructions. Cortisol concentration was measured by chemiluminescence assay on an Immulite 2000 XPi (Siemens Healthcare, Zürich, Switzerland). The analytical sensitivity was 5.5 nmol/L and the measurement range was 0.99–1380 nmol/L. The assay was performed according to the manufacturer’s instructions. If serum cortisol levels were >60 nmol/L [28] the corresponding serum progesterone value was excluded from further analysis.2.5. Statistical AnalysisThe primary endpoint was the number of calves not revealing a cycle at the time of slaughter (10–12 months of age). The study was planned as a case-control study with an assumption of 20% of the calves vaccinated confirmed in the cycle and 80% of the not vaccinated calves in the cycle (difference 60%) and a power to be at least 80%. The number of calves per group was calculated to be n = 25, each (resulting power = 88%). For categorical data, the frequency of categories was determined. For metric variables, mean, median, sd, 25% and 75% quantiles, minimum and maximum were calculated. Survival times were estimated by Kaplan–Meier estimation and compared by a frailty model because the data were clustered within a farm. The events of a progesterone value above 1 ng/mL during the whole measuring period were compared between the two groups with a generalised linear mixed model. The McNemar test was applied for the frequency table whether a vaccinated calf and a control calf within a farm had a progesterone value above 1 ng/mL during the whole measuring period.A p-value < 0.05 indicated a significant result. Data were analysed using the statistical software SAS® version 9.4 (SAS Institute Inc., Cary, NC, USA).3. Results3.1. Treatment with Anti-GnRH VaccineAge at first vaccination/placebo administration was 157 d in group V (median; 25%/75% quartiles: 151 d/164 d) compared to 159 d (148 d/169 d) in group C. Median heart rate, respiratory rate and rectal body temperature at first and second vaccination and duration of the period between vaccinations are listed on Table 1. Age at slaughter was comparable in both groups (median 326 days in group V compared to 320 days in group C).3.2. Side Effects, Including and Excluding CriteriaSlight swelling at the injection site as is described to occur frequently after vaccination with Improvac® [14] was not observed. Reduction of food intake ≤75% of the normal amount judged subjectively by farmers was observed in n = 1 calf of group V for 1 day (without elevated rectal body temperature). Retrospectively, a pair of calves (n = 1 from group V, n = 1 from group C, same farm) had to be excluded because all measured serum cortisol values exceeded 60 nmol/L [28]. In group V, a total of 27/169 progesterone samples in serum had to be excluded (corresponding cortisol sample > 60 nmol/L). In group C, a total of 15/164 progesterone samples in serum had to be excluded (corresponding cortisol sample > 60 nmol/L [28]). The farmers in this study subjectively judged herds to be calmer, even if only one of a pair of calves was vaccinated. They also mentioned that fewer accidents were occurring.3.3. Progesterone and Cortisol ValuesDifferences in progesterone concentrations during the study period are shown in Figure 1a,b and Table 2. The number of calves with progesterone concentration above 1 mg/mL varied between V and C groups during the study period (p = 0.0028). In group V only 3/24 animals (12.5%) exceeded a progesterone value of 1 ng/mL in all samples measured (age: 346 d, 362 d, 363 d), whereas there were 13/23 (56.5%) calves in group C with a progesterone value of >1 ng/mL, beginning at the age of 286 d. The Logrank test (p = 0.0038) showed a highly significant difference between calves within groups V and C, versus the duration of the period between second vaccination and progesterone value > 1 ng/mL (Figure 2).Median (25th/75th quartiles) cortisol values with included progesterone values in group V and C were 18.4; 8.4/35.3 nmol/L and 21.5; 8/36.9 nmol/L, respectively. Cortisol values with excluded progesterone values (>60 nmol/L) in group V and C were 97.7 (81/115.8) nmol/L and 86.8 (65.8/108.8) nmol/L, respectively.3.4. Age at SlaughterThe target age at the slaughter of the calves of this study was 11 months. In group V n = 16 calves were slaughtered at 326 days (median), but there were n = 7 animals slaughtered at a median age of 617 days (1 animal still alive). In group C, n = 17 calves were slaughtered at 320.5 days (median), but there were n = 4 animals slaughtered at a median age of 603 days (3 animals still alive). The reason for keeping these animals longer was due to better meat prices (special meat label).4. DiscussionOur objective was achieved since none of the calves in group V exceeded a progesterone value > 1 ng/mL until the age of 11 months (scheduled age of slaughter). Furthermore, only 12.5% of the animals in group V exceeded a progesterone value of 1 ng/mL during the entire period of measurement compared with 56.5% of the calves in group C. Different studies already proved cycle suppression in adult cows [14,15,29,30] and reduction of serum levels of testosterone in bulls resulting in decreased sexual and aggressive behaviour [20,31,32,33]. The most simple and reliable parameter for clinical veterinarians in the field to predict a return to estrus after vaccination was the discovery of class III follicles (>9 mm) [14,34]. However, in growing calves, rectal palpation is not feasible. Different authors relied on progesterone values to define the beginning of puberty [4,35]. Progesterone values, however, are not reliable if concomitant cortisol values in serum are high (>60 nmol/L [28]), as cortisol and progesterone then originate from the adrenal glands. This was demonstrated in dehorned steers [36], white-tailed deer [37] and also in a cold stress test in women [38]. We therefore included only progesterone values if cortisol values were ≤60 nmol/L [28]. Weekly or biweekly progesterone analyses would have been preferable. The decision for blood sampling every 4 weeks was a concession to logistics and the farmers’ time expenditure and the risk of handling the animals living in herds. It cannot be excluded that a progesterone value > 1 ng/mL was possibly missed.Vaccination with an anti-GnRH vaccine of very young animals (2, 6 and 13 weeks of age) could not postpone the age of puberty in cattle (52–54 weeks) compared to unvaccinated control animals [30]. Alternatively, some studies were performed in older heifers with a bodyweight of 230–480 kg [29,39,40], which was too late for the objective of our study. We decided to vaccinate calves at 5 and 6.5 months of age intending to postpone their puberty until the age of 11 months. If heifers are kept longer on farms, a third vaccination should be performed to extend the time until the beginning of puberty (as performed in cattle to extend the period of cycle suppression [16]), but this was not in the present study’s perspective. The influence of the presence of bulls on age at puberty was another discussion point, as there were adult bulls in all farms included. Neither a study in young heifers (140–402 d of age) [41] nor in heifers of 287 d of age [42] indicated that the presence of an adult bull altered the incidence of precocious puberty.The vaccine Improvac® was used, as in Europe the cattle-specific vaccine Bopriva® is not allowed due to the preservative Thiomersal. Therefore, the use of Improvac® (preservative Chlorocresol) was allowed for cattle by the Institute of Virology and Immunology (responsible for the approval and monitoring of animal vaccines and immune sera for use in veterinary medicine in Switzerland).In Switzerland, the number of beef cows is increasing compared to the decreasing number of dairy cows (Agristat–Statistik der Schweizer Landwirtschaft 2021). Therefore, an increasing number of pregnant young heifers at slaughter is expected. Additionally, interactions among animals in heat can lead to accidents and impair carcass quality [43]. It was described that cattle during estrus were mounted four times in stables and seven times in pasture [44]. Farmers stated that with an increasing number of heifers in heat, herds were more restless, suffered more injuries and were more difficult to handle. In conclusion, the favourable results from our study represent a step forward in solving this problem in an animal-friendly, non-invasive and reliable way to avoid early pregnancy in heifers as well as the slaughter of pregnant cattle.

animals : an open access journal from mdpi

10.3390/ani11082389

PMC8388699

The microbiological safety and quality of commercial animal feed for laboratory animals, produced in Costa Rica, was assessed. Analysis of the animal feed included general microbial markers (total coliforms and molds) and the behavior over time of two specific feed contaminants (Salmonella spp. and mycotoxins). Results from the study suggest that there is a low risk of contamination from viable microorganisms but the product contains important levels of mycotoxins. Current preventive measures (UV light disinfection) are not effective and additional handling protocols should be considered.

Safety and quality of compound feed for experimental animals in Costa Rica is unknown. Some contaminants, such as Salmonella spp. and mycotoxins, could elicit confounding effects in laboratory animals used for biomedical research. In this study, different batches of extruded animal feed, intended for laboratory rodents in Costa Rica, were analyzed to determine mycotoxin and microbiological contamination (i.e., Salmonella spp., Escherichia coli, total coliform bacteria, and total yeast and molds enumeration). Two methods for Salmonella decontamination (UV light and thermal treatment) were assessed. Only n = 2 of the samples were negative (representing 12.50%) for the 26 mycotoxins tested. Enniatins and fumonisins were among the most frequent toxins found (with n = 4+ hits), but the level of contamination and the type of mycotoxins depended on the supplier. None of the indicator microorganisms, nor Salmonella, were found in any of the tested batches, and no mold contamination, nor Salmonella growth, occurs during storage (i.e., 2–6 months under laboratory conditions). However, mycotoxins, such as enniatins and fumonisins tend to decrease after the fourth month of storage, and Salmonella exhibited a lifespan of 64 days at 17 °C even in the presence of UV light. The D-values for Salmonella were between 65.58 ± 2.95 (65 °C) and 6.21 ± 0.11 (80 °C) min, and the thermal destruction time (z-value) was calculated at 15.62 °C. Results from this study suggest that laboratory rodents may be at risk of contamination from animal feed that could significantly affect the outcomes of biomedical experiments. Thus, improved quality controls and handling protocols for the product are suggested.

1. IntroductionThe quality and safety control of the diets for laboratory animals is crucial for the welfare of the animals and to ensure that experimental outcomes are not biased by unintended nutritional or contamination factors [1,2]. Chemical and biological contaminants of feeds, such as bacteria, bacterial toxins, and mycotoxins could influence toxicological, physiological, immunological, reproductive, and other types of research experiments [2,3]. In fact, animal feed, or the raw materials used for its fabrication, are prone to contamination, which can occur at any stage of processing or during storage [4].Similar to pet food, rodent feed belongs to the group of foods called “foods with low moisture content,” whose main characteristic is the low aw (values below 0.85) [5]; it consists primarily of cereals, such as wheat, maize, barley, and other legumes such as soybean. Additionally, in Costa Rica, this type of feed usually contains animal (poultry, beef, and pork) by-product meals, which have shown a high prevalence of Salmonella contamination [6,7]. This is important, as foods with reduced aw levels are believed to be less permissive for microbial growth, but some pathogens can survive for extended periods under these conditions [8]. Bacterial infections of laboratory animals (both clinical and subclinical) can lead to abnormal responses to experimental treatments and interfere with research outcomes [9,10]. Salmonella spp. is a foodborne pathogen highly related to the contamination and survival in foods with low aw content [6,11]; it can survive under harsh conditions for extended periods, even when food products are subjected to high temperatures [12]. In fact, Salmonella recontamination of animal feed, after high-temperature treatment, has been reported [5,13] and, as a consequence, pet food has been documented as one of the leading causes of salmonellosis in humans and animals [14,15]. For example, in Canada, a total Salmonella prevalence of 12.5% was reported in pet food analyzed between 2002 and 2009 [14]. Based on this information, it can be hypothesized that Salmonella contamination could pose a problem in animal feed for laboratory purposes, as these bacteria can affect mice similarly to humans. Though scarce, salmonellosis reports in laboratory animals exist [9,16,17,18], and even zoonotic spread of the infectious disease has been reported [19].Animal feed is also vulnerable to mycotoxin contamination in the field (in the case of raw materials) or during storage and this depends on pre- and post-harvest edaphoclimatic and environmental conditions [20]. Some factors, such as improper processing, packaging, drying techniques, and transport activities may also influence fungal growth and increase the risk of mycotoxin production [21]. Mycotoxins have also demonstrated to be refractory toward processing operations [22], meaning that they could be present in the final product used to feed laboratory animals. Local feed surveillance has already shown a considerable prevalence of mycotoxins in raw ingredients and compound feed [23,24]. Specifically, for rodent feed, considerable concentrations of deoxynivalenol (DON), nivalenol (NIV), ochratoxin (OTA), and zearalenone (ZEA) have been found in commercial feed samples [25]. More recently, Escrivá and coworkers found multiple toxins, including enniatin B, B1, and ZEA [26]; all these compounds may exert additional stress on laboratory animals, and results of scientific studies could be biased, thus leading to wrong conclusions. For example, ZEA has a demonstrated immunomodulatory effect in murine models [27]. The importance of mycotoxin contamination of animal feed is increasing and recent improvements in analytical methodologies to determine and confirm several toxins simultaneously in animal feed is the proof of that [28]. Considering these data, it is clear that the quality control of food, for laboratory animals, is becoming more critical to ensure that reliable in vivo results could be obtained.Few institutions in Costa Rica house animals for laboratory purposes and scientific experiments. One example is the Laboratory of Biological Assays (LEBi) at the University of Costa Rica, whose main objective is to provide healthy animals for the purpose of scientific studies. To fulfill its objective, the animal feed used by LEBi must comply with the highest standards in terms of food quality and safety results, so experiments performed using these animals, are repeatable and scientifically sane [29]. However, given the low demand for experimental animals locally, the internal market of feed that complies with the requirements for scientific experiments is too small, and just two local producers and suppliers are in charge of providing the product. These suppliers, whose main activity is the processing of pet food, do not provide the microbiological data, if any, that they could generate internally in their facility, supporting the absence of pathogenic bacteria, quality, and indicator microorganisms. In addition, it has been noted that there could be inadequate management of the inventory of the product, as the batches are stored for long periods, which may increase the risk of contamination in the facility.Rodent feed manufacturers do not provide information on their product safety controls and official controls in this matrix are scarce, as surveillance programs are mainly focused on feed for productive animals. Hence, in general, the food safety of the product used to feed these animals in Costa Rica is ignored. For this reason, the vivarium takes extra measures to ensure the quality of the product is acceptable for feeding the rodents. Some of these measures include a basic microbiological analysis (coliforms and total aerobic counts) and the exposition of the animal feed to UV light during storage to eliminate or reduce potential microbiological contamination. However, validation studies, to know the efficacy of this treatment, have not been conducted, and no information is available regarding other contaminants such as pathogens, toxins, or chemicals. As the safety characterization of the animal feed used at vivariums may not be thorough due to the absence of regular surveillance programs that include mycotoxins and Salmonella analysis, the real risk for the experimental animals is unknown, and potential corrective actions may not be in place.The objectives of this study were (i) To generate information on the microbiological risk profile of the animal feed for laboratory animals used at LEBi through the analysis of relevant microorganisms (Salmonella, Escherichia coli, total coliform bacteria, and total yeast and molds) and mycotoxins. (ii) To raise awareness about this potential risk for biomedical research in developing countries with similar facilities. (iii) To evaluate the efficacy of current intervention methods for the product (UV light) and alternative procedures (thermal treatment) in case contamination occurs. The information was used to design recommendations for the management of microbiological risks on the feed used for experimental animals in settings with limited resources.2. Materials and Methods2.1. Sampling and Study ObjectsIn this study, different batches of extruded/pelletized animal feed intended for laboratory rodents fabricated in Costa Rica were sampled and analyzed. Feed samples are routinely stored in stacked 1–3 mm thickness polypropylene woven 30 kg sack bags and destined for animals prepared for biological testing by the Laboratory for Biological Essays (LEBi) at the University of Costa Rica (San José, Costa Rica).2.1.1. Sampling for Mycotoxin AnalysesFor mycotoxin analysis, n = 16 different rodents extruded feed batches were sampled. The samples were acquired from two different local suppliers (n = 9 and n = 7 different feed batches from suppliers 1 and 2, respectively).Mycotoxin sampling was performed according to the parameters underlined by the Feed Inspectors Manual [30]. To improve sampling representability and compensate for possible contamination hotspots, compound 5 kg samples from each batch of feed were obtained by collecting n = 50 increments of 100 g each. Additionally, the 5 kg sample was quartered, and a subsample milled using an mm sieve (Retsch ZM 200, Haan, Germany).2.1.2. Sampling for Microbiological AnalysisFor Salmonella analysis, two representative samples per batch from n = 12 different batches were taken for feed dispatched from supplier 1. To analyze Escherichia coli, total coliform bacteria, enumeration of yeast, and molds n = 16 different feed batches were sampled from supplier 1 and 2. The sampling method, described in the previous section, was adjusted to be performed aseptically, so the same samples could be used for microbiological analysis. Hence, corers and spatulas were previously autoclaved. Microbiological analysis subsamples were recovered before preparation for toxin analysis.2.2. Safety and Quality in Animal Feed2.2.1. Mycotoxin AnalysisThe assays were performed using a multitoxin targeted MS-based LC approach (n = 26 toxins). Briefly, 25 g of each sample was extracted using 100 mL of an ACN:H2O:CH3CO2H solution (74:25:1). The mixture was then dispersed for 2 min using an Ultra-turrax® homogenizer set at 18,000 rpm (T25, IKA, Werke GmbH & Co. KG, Staufen in Breisgau, Germany). The supernatant was recovered and then gravity-filtered through Whatman® 541 ashless filters (GE Health Life Sciences Little Chalfont, Buckinghamshire, UK). A 2 mL aliquot was pipetted to a 25 mL volumetric flask its volume made-up with phosphate buffer solution (PBS, containing NaCl, 150 mmol L−1; KCl, 2.50 mmol L−1; Na2HPO4, 4.50 mmol L−1; KH2PO4, 1.50 mmol L−1 and adjusted at a final pH of 7.4). During sample cleanup, solid phase extraction cartridges were activated and conditioned with 2 mL MeOH and equilibrated with 2 mL of a MeOH/H2O solution (5:95 volume ratio, Oasis® HLB columns, WAT094226, 3 cc, 60 mg, 30 μm particle size, Waters Corporation, Milford, MA, USA). Solvents and sample extracts were transferred through the columns at a maximum flow rate of 1 mL min−1 with the aid of a an SPE 12 port vacuum manifold (operating at 15 mmHg, 57.044, Visiprep™, Supelco Inc., Bellefonte, PA, USA). To recover the analytes, columns were washed with a 2 mL MeOH/H2O (5:95) solution, and 2 mL of MeOH was used to elute analytes. The resulting eluates were concentrated to dryness under vacuum at 60 °C (Centrivap, LABCONCO, Kansas City, MO, USA), reconstituted with 300 μL of MeOH, and transferred to HPLC polypropylene vial insert (300 μL, polymer feet, 5182–0549, Agilent Technologies, Santa Clara, CA, USA).The analytical determinations were accomplished using chromatographic system equipped with a 1260 infinity quaternary pump (61311C), a column compartment (kept at 40 °C during analysis, G1316A), and an automatic liquid sampler module (injection volume set at 10 μL, ALS, G7129A), and equipped with Zorbax Eclipse Plus chromatographic column (3.0 mm ID × 100 mm, 3.5 μm, P/N 959961–302, Agilent Technologies). The apparatus was coupled with a single quadrupole mass spectrometer with electrospray ionization ion source (6120B, Agilent Technologies). The drying gas, nebulizer pressure, drying gas temperature, and capillary voltage were set to 10.0 L min−1, 50 psi, 350 °C, and 4000 V, respectively, for positive ion mode electrospray ionization (ESI+).A gradient analysis based on A ACN and B H2O as solvents, both acidified with formic acid at 0.1 mL/100 mL, was employed to separate the mycotoxins quantitatively at a flow rate of 0.15 mL min−1. The established gradient was as follows: at 0 min 10% A, at 4 min 10% A, at 22 min 100% A, at 25 min 10% A, and finally at 35 min 10% A. Toxins were assessed using a Selected Ion Monitoring/SIM mode, peak width and cycle time were set to 0.1 min and 0.60 s cycle−1, respectively. The molar mass, target ions, retention times, cone voltage, and obtained limits of detection and quantification of each analyte were described previously [23]. Finally, a naturally contaminated cornmeal (TR-MT100, Multitoxin Reference Material MT-C-999-G, Trilogy®, Washington, MO, USA) was tested parallel for quality control purposes during each batch of analysis.2.2.2. Water Activity (aw)As previously described [31], the Aqualab® chilled mirror technique was used to determine water activity in the samples (Aqualab® 4TE, Metter Food, Pullman, WA, USA). A small portion of ca. 1 g of sieved material was placed in plastic cups and placed inside the chamber until equilibrium was reached (2.5 min on average). The aw values were registered with ±0.003 accuracies at 24.50 ± 0.24 °C. Water activity was measured for each feed immediately after sampling and when attained and was also monitored for six months during storage. AAFCO check sample 201921 (Equine Feed) and a verification solution (0.500 aw, 8.57 mol kg−1 LiCl) were used for quality control during measurements.2.2.3. Microbiological AnalysisSamples were homogenized using a laboratory blender (InterscienceTM BagMixer® 400, Saint-Nom-la-Bretèche, France). Subsequently, Salmonella, Escherichia coli, total coliform bacteria, and total yeast and molds enumeration were performed. The protocol to isolate Salmonella was performed according to the Bacteriological Analytical Manual (BAM) method: Salmonella, Chapter 5 (FDA). Suspicious colonies on solid media were confirmed with the VITEK 2 system (Biomérieux, Durham, NC, USA) following the manufacturer’s instructions.For the quantification of Salmonella, the samples were homogenized with 90 mL of 0.1 g/100 mL Sterile Peptone Water (PW; Oxoid Ltd., Basingstoke, UK) using a stomacher blender; additional decimal dilutions were applied to this original suspension using PW-containing tubes, and a volume of 100 µL of proper dilutions were used to inoculate Tryptic Soy Agar (TSA; Oxoid Ltd., Basingstoke, UK) and Xylose Lysine Deoxycholate Agar (XLD; Oxoid Ltd., Basingstoke, UK) plates. Agar plates were incubated at 35 °C for 24 h, and typical Salmonella colonies were quantified to obtain the population from each sample (log CFU g−1).Coliform bacteria were analyzed following APHA/CMMEF methods 9.91–9.94 based on an MPN technique.The enumeration of yeast and molds was analyzed following the respective method of the Bacteriological Analytical Manual [32].2.3. Contaminants Behavior in Animal Feed during Storage2.3.1. Mycotoxins and Molds Growth during StorageA randomly elected sample from each batch of extruded animal feed and supplier [batches n = 8 and n = 1 from suppliers 1 (S1) and 2 (S2), respectively] was conserved under laboratory conditions [i.e., from 21.5 to 28.7 (mean 24.5) °C and 29.7 to 56.1 (mean 37.8) %RH, respectively] and monitored for aw, mycotoxin and molds values for several months for a total of n = 9 replicate per sample (one per month) were taken over time to see the behavior of both batches.2.3.2. Bacterial Strains for Animal Feed ContaminationFive different Salmonella strains, including various serotypes (S. Typhi, S. Typhimurium, and S. Enteritidis), were used in this study. According to previous research in our laboratory, the selected strains demonstrated higher resistance to low water environments (not published). All the isolates are part of the bacterial collection of the Research Center for Tropical Diseases (CIET) from the University of Costa Rica (San José, Costa Rica), and they were maintained as glycerol stocks at −80 °C.2.3.3. Salmonella Growth in Animal Feed during StorageEach Salmonella strain was grown in Tryptic Soy Broth (TSB; Oxoid Ltd., Basingstoke, UK) and incubated at 35 °C for 24 h (final Salmonella population of approximately 9.0 log CFU mL−1). Then, a suspension for each strain was prepared by making decimal dilutions in PW to obtain a final Salmonella population of around 6.0 log CFU mL−1. Equal volumes of each strain suspension were mixed to obtain a Salmonella cocktail used to inoculate 10 g of the animal feed by adding 100 µL (final population in the product 3.0–4.0 log CFU g−1) of the bacterial suspension. A small volume of bacterial suspension was used in order to avoid significant alterations of the original water activity of the samples. The inoculated samples were thoroughly mixed by hand inside a sterile plastic bag, and they were dried out inside a biosafety cabinet for 15 min to remove excess humidity and promote bacterial attachment. Then, each sample was sealed in plastic bags and stored at 17 °C inside the LEBi warehouse. At different time intervals, samples were taken from storage to quantify the Salmonella population on solid media. Three repetitions were performed to build growth curves.2.3.4. Salmonella Survival in Animal Feed during Storage and UV ResistanceFor this experiment, the animal feed was inoculated as stated before but this time using the original Salmonella cocktail suspension with no dilutions (initial population in the sample between 6.0–7.0 log CFU g−1). Inoculated samples were stored at 17 °C inside the LEBi warehouse. A parallel set of samples was also prepared to validate the current UV radiation treatment on Salmonella present in animal feed; in this case, stored samples were exposed to continuous UV radiation (255 nm) using a UV lamp (Sylvania Germicide Lamp T8 30 W, San José, Costa Rica). Each sample bag was placed separately on a rack that was directly under the UV lamp (distance of 1.75 m). For both experiments, samples were removed from storage at different time intervals to quantify the Salmonella population as described (total sampling time of 64 days). Three repetitions were performed, and complete survival curves were constructed. A comparison was made between samples stored under normal conditions and those exposed to UV radiation.The inoculated feeds were storage at LEBi warehouse under the same conditions that lab feeds are normally stored there, including a temperature of 17 °C. This temperature is used to avoid rancidity and improve shelf life, as natural-ingredient diets should be stored at temperatures less than 21 °C (ca. 70 °F, and a RH of 50%) [1]. Under these conditions, dry laboratory animal diets stored properly can be used up to 6 months, or longer in some cases, after manufacture [1].2.4. Salmonella Thermal Resistance in Animal FeedAs described before, animal feed samples were prepared and inoculated with the Salmonella cocktail (initial Salmonella population of 7.0 log CFU g−1). The inoculated animal feed samples (10 g) were packaged in sterile plastic bags and immersed in a water bath with circulation capabilities (Isotemp 3013H, Thermo Fisher Scientific, Waltham, MA, USA) that was already set to the desired temperature. The time required to reach the target temperature (come-up time) was determined using non-inoculated product.The different time-temperature combinations used for this experiment can be seen in Table 1. After heating at the specific time, samples were removed and immersed in an ice-water bath. Then, Salmonella population after heating was determined by quantification on solid media as described before. Data was used to construct thermal death curves to calculate D-values with the inverse of the slope of the regression line using Excel software (2007, Microsoft, Redmond, WA); the D-values are expressed in minutes or hours. Similarly, the thermal destruction time (z-value) was calculated by plotting the temperature (x-axis) versus log D-value, and the data was also fitted by linear regression (2007, Microsoft).2.5. Statistical Analysis 2.5.1. For Mycotoxins AnalysisSpearman rank-order correlation tests were applied to determine the association between variables sampling date and aw between feed batches, aw and FB1 and FB2 concentration in a single sample over time. Simpson (D) and Shannon’s (H) Diversity Indexes were used to assess variation on mycotoxins among samples from different batches [33]. A t-student test was used to compare concentration means for the toxins shared between the supplier’s samples (i.e., ENNB, ENNB1, FB1, FB2, β-ZON) to assess whether they differed statistically. A similar test was used to determine if aw values differed among suppliers. For all statistical tests, a threshold value of α = 0.05 was used to consider differences between variables significant. 2.5.2. For Salmonella Behavior AnalysisAs represented by animal feed batch, Salmonella stock, and day of preparation, three independent replications were performed for each of the experiments. Data (log CFU per gram, D-values, and z-values) were compared using analysis of variance of the General Linear Model procedure of the Statistical Analysis System (SAS Institute Inc., Cary, NC, USA). Fisher’s least significant difference (p < 0.05) was used to separate the means.3. Results3.1. Water Activity between Different Feed BatchesTable 2 shows the water activity analysis for the samples and batches analyzed from suppliers 1 and 2. These values confirm that most samples from both suppliers can be classified as dry food (i.e., aw values < 0.60). Suppliers 1 and 2 showed n = 2 and n = 1 samples with aw values above 0.60, which would classify them as intermediate moisture foods. Finally, there does not appear to be a trend regarding the sampled batch of feed and its water activity (Spearman rho, p = 0.399).3.2. Safety and Quality of Animal Feed3.2.1. Mycotoxins of Animal FeedDuring the study, of the total samples analyzed (n = 16), only n = 2 of the samples were negative (representing 12.50%) for the 26 analytes tested. Both negative samples listed above correspond to supplier 1 (samples S1C and S1D).Regarding the rest of the samples analyzed from supplier 1, for n = 2 samples (S1B and S1E), only n = 1 toxin was observed, i.e., enniatin B1 and aflatoxin B1 with concentrations of (143.11 ± 17.49) μg kg−1 and (0.15 ± 0.02) μg kg−1, respectively (Table 3). The remaining samples exhibited n = 3 (11.54%), n = 9 (34.62%), n = 5 (19.23%), and n = 7/26 (26.92%) analytes respectively for samples S1A, S1F, S1G, S1H, and S1I (Figure 1A). For supplier 2, of the n = 7 samples tested from their food for laboratory animals, all of the samples exhibited a mixture of at least five or more toxins (Figure 1B).Unlike samples from supplier 1 (D = 8.67, H = 2.27), the toxins present in supplier 2′s feed are more varied (D = 12.47, H = 3.44). Enniatins were predominant in samples from supplier 1, but any toxin was in higher frequencies in supplier 2’s feeds (Figure 1A,B, Table 3). Enniatins and fumonisins are among the most frequent toxins found (with n = 4+ hits), and the samples from both suppliers share this trait. Average concentrations for both suppliers 1 and 2 are shown in Table 3. For supplier 1, higher levels of mycotoxin were observed for ENNB, ENNB1, and 15-ADON; for supplier 2, FB1 was present at higher values. For supplier 1 and supplier 2 samples, AFB1, and β-ZON were frequently found with concentrations of 3.30 ± 5.38 and 268.71 ± 527.85 μg kg−1, respectively on both accounts.3.2.2. Indicator Microorganisms and Pathogens in Animal FeedThe enumeration of yeast, molds, coliform bacteria, and E. coli were used as indicators of hygiene and fecal contamination of animal feed intended for laboratory animals. In all feed batches tested, none of the indicator microorganisms were found. Similarly, the analysis of the detection of Salmonella was used to assess the microbiological safety of animal feed. All feed samples tested were Salmonella negative.3.3. Microbiological and Mycotoxin Analysis of Lab Animal Feed during Storage3.3.1. Mycotoxin, Water Activity and Mold Analysis on a Feed Batch over TimeThere is no marked trend overtime on mycotoxin values in the samplings for supplier 1 (Figure 2A). However, for supplier 2, there is a notable trend for ENNB, FB1, and β-ZON, where higher concentrations were observed for the first three feed samplings; later, these concentrations decrease over time (Figure 2B).For supplier 1, the last sample tested exhibits a higher number of toxins n = 6 (FB1, AFG2, FB2, AFB1, ENNB, ENNB1) compared to those found in the initial sampling n = 3 (FB1, FB2, β-ZON). A similar situation is observed for the follow-up of supplier 2, a reduction in the number of analytes is observed, this is n = 7 (FB1, AFG2, FB2, AFB1, β-ZON, ENNB, ENNB1) versus n = 3 (FB1, FB2, β-ZON), respectively (Table 4).In general, no marked trend was observed for water activity or toxin levels (Figure 3B–D), except for the case of FB1 and the sample selected from supplier 1 (Figure 3A). However, there is no causality or association between variables (Spearman rho, p > 0.05). Finally, no mold contamination was detected during this storage period (i.e., <100 CFU g−1).3.3.2. Salmonella Behavior during Storage after InoculationBacterial GrowthThe five Salmonella spp. strain cocktail inoculated at low concentrations in the animal feed for rodents did not show any growth after 48 days of storage at 17 °C. The initial population in the animal feed was 3.4 ± 0.4 log CFU g−1 and after 48 days the final population was 1.7 ± 0.1 log CFU g−1 for a total decrease in 1.7 log CFU g−1.Bacterial SurvivalSalmonella survived for 64 days in animal feed stored at 17 °C (Figure 4). The initial population of the Salmonella spp. cocktail was 6.6 ± 0.2 log CFU g−1, and a final population of 3.9 ± 0.2 log CFU g−1 was observed after 64 days of storage at 17 °C for a total decrease in Salmonella spp. numbers of 2.7 ± 0.2 log CFU g−1. A significant decrease (p < 0.05) in Salmonella spp. population was observed between days 1 (6.6 ± 0.2 log CFU g−1) and 3 (5.27 ± 0.08 log CFU g−1). However, after day 3 of storage, no significant decrease in Salmonella population was observed, and the average reduction value obtained was just 0.04 log CFU g−1 per day. Similar results were observed in the case of samples stored under UV light (Figure 4) and no differences (p > 0.05) were determined between both conditions.3.4. Salmonella Resistance in Animal FeedSalmonella Thermal Resistance in Animal FeedFigure 5A–D show the thermal resistance of Salmonella spp. in animal feed at four different temperatures: 65 °C, 70 °C, 75 °C, and 80 °C. Survival and temperature share an inverse association with deltas of −2.58, −3.49, −3.86, and −4.11 log CFU g−1, for each condition, respectively. Time and temperature also share the same relationship with 150, 80, 60, and 24 min to reach a similar decrease in Salmonella population.The D-value was calculated for each temperature (Table 5) using the straight-line equation, which is the negative inverse of the slope.The log D-values for each temperature were used to build the thermoresistance curve of Salmonella in animal feed (Figure 6). Using the negative inverse of slope of the equation of the line (i.e., y = −0.1614 + 5.4303, R = 0.9275, standard error of estimate 0.7128) a z-value of 15.62 °C was calculated.4. Discussion4.1. Safety and Quality in Animal Feed4.1.1. Mycotoxin Analysis between Different Feed BatchesMycotoxins are secondary metabolites produced by filamentous fungi belonging mainly to the genera Aspergillus, Fusarium, Penicillium, and Alternaria [34]. Simultaneous presence of several toxins in animal feed has been described as having significant adverse effects, even at low concentrations, given additive or synergistic interactions [35,36]. In this sense, the analysis of the co-occurrence of several toxins simultaneously in an animal feed is of utmost importance [36]. The co-occurrence of mycotoxins in animal feed has already been described for products intended for other species, which has already shown that it can exert antagonistic, additive, or synergistic effects on animals [36]. In this regard, data reported herein are in line with those reported elsewhere for rodent feed [i.e., 0.3 (OTA) to 298 (DON) and 6.4 (ENNA1) to 303.6 (ZEA) μg kg−1] [12,13]. Besides toxins, other environmental contaminants have been described in rodent feed (e.g., PCBs, heavy metals, PCDD/Fs, and pesticides [4]. This will exert additional metabolic stress on the animal, and in the presence of xenobiotics, mycotoxins could potentiate their injuriousness. In addition, technological processes on raw materials, or balanced foods, have proven to be ineffective in post-harvest control of these contaminants [35]. This is relevant since rodent feed is processed through an extrusion process (i.e., subjected to temperatures ca. 150 °C under high pressure when forced through the die) [1]. By consensus, it has been described that the best strategy for the post-harvest control of mycotoxins is based on proper storage and handling of animal feed to prevent conditions that lead to fungal growth [37]. Temperature, water activity, and the presence of insects have been cited as the factors most associated with the formation of mycotoxins during storage [37].On the contrary, it is typical that the by-products of food materials used in human consumption (usually categorized in this context as waste) are reprocessed on occasion to be incorporated into the balanced animal feed. For example, rice husk is used as a vehicle to introduce balanced vitamin mixtures, but the outer part of this grain is the most prone, due to its exposure, to contamination by toxins [38]. Under these conditions, an 11.6% prevalence for aflatoxins in laboratory animal feeds has been reported for local products [38].Although the microbiological tests did not show the presence of fungi or yeasts in any of the samples tested, it should be noted that the mycotoxin contamination could have occurred at any point prior to its extrusion [39], or it could be a consequence of raw materials already contaminated [40]. Furthermore, mycotoxins can remain in the finished product even without the presence of the fungus that synthesized them since they are molecules that are highly resistant to industrial unit operations [21]. The presence of trichothecenes, produced by Fusarium sp., neosolaniol and fusarenone-X, indicates the use of raw materials from temperate climates [41].Mycotoxins, depending on the type, can cause different effects in animals usually used for experimentation, such as rats, mice, and rabbits. These effects include inhibition of the immune system, carcinogenic, teratogenic, hepatotoxic, nephrotoxic, endocrine and reproductive disorders, sterility, among others [24,25].Herein, we reported especially high, and frequent, rodent feed contamination with enniatins and fumonisins. Similar trends have been detailed previously for other feedstuffs in Costa Rica, with prevalence of contamination reaching, in some cases, up to 45–50% and with non-trivial concentrations [23,24].Specific adverse effects in vitro, and in vivo, have been reported for aflatoxin B1, enniatin, fumonisins, zearalenol-derivatives in animals, especially murine models. These effects include immunotoxicity, oxidative stress for AB1 and FB1 [42], and acute genotoxicity in hepatocytes for AFB1 [43,44]. Noteworthy, despite presence of AFG2, levels do not exceed the legal thresholds of 20 μg kg−1 [45,46].FB1 and DON have been demonstrated to alter the intestinal barrier, impair the immune response, and reduce feed intake and weight gain. Their presence in feed increases the translocation of bacteria; mycotoxins can also enhance the susceptibility to infectious diseases [47]. Additionally, blastocyst cell numbers and proportion of late blastocysts have been reported to decrease in mouse embryos in the presence of T-2 and FB1 [48].Beauvericin and both enniantins have demonstrated cytotoxicity of Caco-2 cells [49], neurotoxicity (can cross the brain-blood barrier, [50], genotoxicity [51], and intestinal toxicity) [52].Finally, ZEA and derivatives have demonstrated toxic effects in mice’s reproductive system [53]. In granulosa, cells from mice and domestic animals have shown impaired development and follicle steroidogenesis, reduced oocyte nest breakdown, damaged meiotic progression, poor fetal oocyte survival, accelerated primordial follicle activation, and enhanced follicle atresia [54].Data from this study indicate that laboratory animals could be chronically exposed to mycotoxin contamination from animal feed. Further studies are necessary to understand the clinical significance of this exposure, considering both the levels of mycotoxins reported in this study, the amount of feed that is consumed, and the regular lifespan of the rodents. However, considering a batch of food could last several weeks, or even months, a variation in the concentration of toxins during storage generates a more troublesome scenario by implying that some rodents would have a higher exposure to chemicals than others. The same situation arises if, for any reason, the feed were to be contaminated during storage.4.1.2. Microbiological Quality of SamplesIn line with the local legislation thresholds for animal feed (i.e., absence in 25 g) [45,46], data from this study confirm that the animal feed for rodents was not contaminated with Salmonella spp. or any indicator bacteria. The animal feed is an extruded product where most of the pathogenic microorganisms in the raw material are eliminated during processing [55]. Despite a high risk of post-processing contamination of this kind of food [56], Salmonella spp. or any microbial indicator were not detected, meaning that proper storage and handling conditions are being applied. At least this could be concluded for those batches analyzed in this study. In addition, some chemical or biological antimicrobial compounds (preservatives and antibiotics) present in the product could inhibit the growth of Salmonella spp. Although antibiotics are not allowed in animal feed according to applicable regulation [45,46], antimicrobial residues in compound feeds have been reported previously [23,57]. Then, it is possible that animal feed used by LEBi may also contain unintended antimicrobial compounds residues.On the other hand, it is important to clarify that the sub-samples used for the study are just a representation of larger product volumes. This means that only a small fraction of each batch was analyzed. It is possible that the presence of Salmonella could be underestimated because its distribution is not homogeneous within each bag of product, and it could have been present only in certain spots within small microenvironments [58]. Based on the experience at LEBi, it is challenging to obtain new batches every time the animal feed is bought. As LEBi requests only small amounts of animal feed, the food supplier may provide the same batch of animal feed for an extended period of time. This indicates that an exhaustive sampling analysis may be needed to discard the presence of Salmonella spp. Furthermore, indicator microorganisms, of fecal contamination or poor hygiene, were not found, which supports the effectiveness of the extrusion process and the adequate post-process handling. In addition, during storage, the growth of fungi and molds was not detected; this could be due to calcium propionate as an additive in the animal feed tested. Calcium propionate, in its acid form, is an inhibitor of fungal growth, causing the inactivation of enzymes. It also competes with essential amino acids, such as alanine, thus inhibiting microorganisms’ growth [59,60].Given that post-processing contamination with Salmonella spp. risks do exist for heat-treated feed and feed ingredients (e.g., prevalence of 3.37 and 26.74% for Costa Rican pet food and meat and bone meal, respectively [6,7]), it could be assumed that the safety of the animal feed may be compromised at some point. However, the results from microbiological analysis, of the animal feed at LEBi, support the idea that there is a small risk of microbiological contamination for rodents. This means that chances of affecting the experimental animals (at least with viable microorganisms) are low, and the product may remain stable throughout storage. 4.2. Water Activity between Different Feed BatchesIt should be mentioned that aw values above 0.85 are considered a threshold to favor the growth of pathogenic bacteria [61]. On the other hand, fungi and yeasts are capable of growing with these values [61]. Thus, the aw values also justify the general absence of pathogens or counts below the lower limit, in all the samples, for indicators such as total coliforms and Escherichia coli (<3 CFU g−1) and fungi and yeasts (<100 CFU g−1).4.3. Contaminants Behavior in Lab Animal Feed during Storage4.3.1. Water Activity and Toxin Behavior on a Feed Batch over TimeWhile at the end of the tests, both aw and mycotoxins present a downward trend, probably due to the effect that cold air has on the humidity of the product. Because the food is not hermetically sealed, there is gas transfer [62]. As in the case of Salmonella, it is possible to appreciate, in all cases, that sampling is essential to obtain accurate and accurate data [63]. There is inherent variability in the test as there are hot spots where the toxin may be since it is usually not distributed homogeneously in raw material or food [64]. Mycotoxin variation, within each sample, during storage might result from several factors, including abiotic and biotic-mediated hydrolysis of masked mycotoxins (e.g., glycated toxins) [65], microbial toxin metabolism [66], and toxin heterogeneity, within the feed, due to specific areas of elevated water activity [64]. Even though the feed samples tested can be considered microbiologically sound, based on general markers tested, it is relevant to remember that even feed samples that have been subjected to processing are not considered 100% sterile. Indeed, we have already reported Staphylococcus, Bacillus, or Lysinibacillus sp. in Costa Rican extruded feed samples [57].4.3.2. Salmonella Behavior during StorageAccording to our results, Salmonella spp. does not have the ability to grow in animal feed during storage. It seems that the low water activity levels were a significant factor for preventing the bacteria from multiplying. These results agree with previous reports in the literature where it is shown that pathogenic bacteria such as Salmonella spp. are not able to grow in aw values below 0.85 [67]. Similarly to our study, Beuchat reported that Salmonella spp. did not have the ability to grow in inoculated pecans (1.53 log CFU g−1) that have aw values between 0.43 and 0.51 [68]. Although Salmonella spp. can survive in foods with low aw values for periods as long as one year, the bacteria cannot grow until the moisture content of the product is increased [54]. This is the main reason why the product needs to be stored in a place with proper temperature and humidity conditions and protected from environmental contamination [69,70]. However, although Salmonella spp. present in the animal feed cannot grow, there is still a risk of contamination for the rodents. The infectious dose of Salmonella spp. is very low and can cause illness even though few viable cells are present [5].In the case in which contamination of animal feed did occur, Salmonella spp. can survive in high numbers for extended periods [11,28]. In Figure 4, it is observed that Salmonella spp. population can persist in the animal feed during 64 days of storage at 17 °C. These results agree with several studies related to the survival of Salmonella spp. in foods with low moisture content. For example, Santillana and coworkers demonstrated that Salmonella spp. can survive in protein powder for 168 days at 21 °C, and it can survive in the food matrix even when the moisture content values are further decreased [11]. Another study conducted by Beuchat demonstrated that Salmonella spp. can survive in pecans stored at −20, 4, 21, and 37 °C for 546 days and 364 weeks in pecans chunks [68]. Hence, the ability of Salmonella spp. to survive for extended periods could represent a high risk of contamination for rodents, especially because the product is stored for at least one month before use.4.4. UV and Thermal Resistance of SalmonellaSalmonella spp. inoculated in animal feed is not affected by UV light, which indicates that current inactivation protocol, applied by LEBi, is not serving the purpose of controlling microbial contamination of the animal feed. Similarly, other studies have reported no effect of UV light in food matrices with low aw content. For example, UV light is not adequate for the decontamination of spices in samples placed at a 10 cm distance from the light source; the Salmonella Typhimurium reduction was just 0.29 log CFU g−1. Another example is the minimal Salmonella spp. reduction reported in cumin seeds exposed to UV-C light during 60 min [71]. The low reduction in Salmonella from UV light could be attributed to the storage conditions of animal feed (used at LEBi) and the bacteria characteristics. The way the product is stored can affect the efficacy of the treatment as the animal feed is not directly exposed to the UV light and it is placed as far as 2 m from the primary source of radiation. It has been widely reported that bacterial inactivation is reduced when the product is far from the source of radiation and when the UV light is blocked [28,72]. Additionally, the animal feed is stacked vertically, which does not allow the radiation to be homogeneous throughout all the products. The way the animal feed is stacked does not allow the UV light to penetrate throughout the product, and the UV light cannot reach all the cells present [28,73]. This means that storage conditions do not follow the recommendations on the correct use of UV light. The UV light radiation may not affect Salmonella due to the protection mechanisms related to low aw environments [5]. It can be concluded that the disinfection mechanism with UV light does not have a significant effect on Salmonella spp. present in the product and that current ways of storing the animal feed could increase the risk of contamination for the rodents used for experimental analysis.Given the limitations posed by UV light decontamination, additional intervention strategies for animal feed may be considered. This is important, especially in those cases where contamination occurs, and the possibility to obtain new batches of animal feed may be compromised. Thermal treatment of the product could be considered as an additional intervention strategy to control Salmonella as it could be easily applied in settings with limited resources, such as LEBi and other similar vivariums, in developing countries. Still, scientific literature shows that Salmonella spp. can survive well in food products with low aw after heating [11,74,75]. When exposed to low-aw environments, Salmonella spp. activates protein adaptation mechanisms that result in stable cell structures that decrease the thermal denaturation [76]. Other thermoresistance mechanisms include osmoregulation, ribosomal RNA degradation, filamentous shapes, biofilms formation, and activation of viable, but not culturable, bacteria. These mechanisms improve the resistance of Salmonella to heat, thus prolonging its survival in the food [74]. For example, Lound and coworkers reported that Salmonella Enteritidis increased its resistance 60 times more in dehydrated egg albumin than in nutrient broth at 72 °C and 300 times at 82 °C [74]. However, our study demonstrates that Salmonella spp. can be eliminated from the animal feed by treating samples at different time-temperature combinations [75]. The D-values obtained from this study are useful tools, and they are comparable to those reported in previous research. The D- values of Salmonella Senftenberg 775W in meat and bone flour with a moisture content of 10 g/100 g were 65.47, 37.37, 19.38, 6.60, 4.15, 2.08, 1.16, and 0.36 min at temperatures of 60, 63, 66, 71, 74, 77, 79, and 85 °C, respectively, with a z-value of 20.16 °C [15]. This type of data could be used to design proper intervention strategies for the decontamination of animal feed, and the z-values become practical tools to predict the behavior of Salmonella in this product. This study indicates that the current methodologies for storage and handling animal feed should be revised. New preservation methods such as heating the product (using an oven or autoclave) could be considered. However, although these recommended alternatives could help kill pathogenic bacteria, they could also affect the original nutritional composition of the product. Therefore, the application of new intervention strategies, for in-house applications, must be validated to prevent a deleterious effect on the product’s final quality [77].5. ConclusionsExperimental animals may be at risk of contamination from the animal feed. Of special concern is the chronic consumption of mycotoxin-contaminated feed as this may undermine toxicological data (i.e., due to possible artifacts or confounding or undesired effects). Given the importance of vivariums, it is clear that keeping the well-being of the experimental animals should be in the priorities of different scientific units. This study provides input to generate recommendations to improve handling of the animal feed to prevent contamination issues and avoid compromising the quality and robustness of scientific studies. Monitoring of microbiological contamination could be improved by increasing the number of samples obtained per batch of animal feed. A robust sampling protocol could be applied considering the size of the batch and the distribution of the product within each sack. Additional testing should be incorporated to include the analysis of Salmonella and mycotoxins. Surveillance for pathogens, such as Salmonella spp. and mycotoxins, should be constant and rigorous. New data could be collected to determine the number of samples that must be necessary to assure proper monitoring of these variables. In terms of mycotoxins contamination, monitoring protocols could incorporate the analysis (with special emphasis on fumonisins and enniatins). Current strategy (e.g., UV light decontamination) is not serving its purpose and, according to our data, there is no use to maintain this handling protocol at the LEBi warehouse. Strengthening of monitoring protocols is the best way to compensate for the absence of any additional preventive measure. However, in case contamination occurs, D and z-values obtained from this study can be considered as an aid to design a post-processing decontamination protocol, especially in those cases where there is an urgent need of the animal feed and it is not possible to obtain additional product from the suppliers. Given the low water activity content of the product, heating protocols can be applied on product packaged in plastic bags suitable for autoclave use. This method is already applied by LEBi when sterilizing material that is sensitive to humidity. The information provided by this study, describes a reality that may be common to other countries where the access to adequate resources to perform experiments with animals could be limited. Therefore, other vivariums can use this information to take measures to improve the safety of the food they provide to their animals. Biomedical researchers may want to consider routinely testing feed for its quality and safety to, at least, foresee spurious effects caused by contaminants within the feed.

animals : an open access journal from mdpi

10.3390/ani11113310

PMC8614361

Cortisol is a key stress hormone in teleosts. Cortisol exerts its effects through genomic—and membrane-initiated mechanisms, however, the role of the latter in long-term stress responses is unknown. Here, we treated Sparus aurata with cortisol or cortisol-BSA (exclusive inductor to membrane-initiated effects) to emulate a long-term stress situation. We found that cortisol, but not cortisol-BSA, promotes energy substrate mobilization in the liver, together with the regulation of metabolism-related genes. We suggest that genomic cortisol actions exclusively participate in metabolic responses during prolonged treatment using cortisol in S. aurata. This study contributes to the current knowledge on cortisol’s involvement in stress responses in fish.

Cortisol is the main glucocorticoid hormone promoting compensatory metabolic responses of stress in teleosts. This hormone acts through genomic and membrane-initiated actions to exert its functions inside the cell. Experimental approaches, using exogenous cortisol administration, confirm the role of this hormone during short (minutes to hours)- and long-term (days to weeks) responses to stress. The role of membrane-initiated cortisol signaling during long-term responses has been recently explored. In this study, Sparus aurata were intraperitoneally injected with coconut oil alone or coconut oil containing cortisol, cortisol-BSA, or BSA. After 3 days of treatment, plasma, liver, and skeletal muscle were extracted. Plasma cortisol, as well as metabolic indicators in the plasma and tissues collected, and metabolism-related gene expression, were measured. Our results showed that artificially increased plasma cortisol levels in S. aurata enhanced plasma glucose and triacylglycerols values as well as hepatic substrate energy mobilization. Additionally, cortisol stimulated hepatic carbohydrates metabolism, as seen by the increased expression of metabolism-related genes. All of these responses, observed in cortisol-administered fish, were not detected by replicating the same protocol and instead using cortisol-BSA, which exclusively induces membrane-initiated effects. Therefore, we suggest that after three days of cortisol administration, only genomic actions are involved in the metabolic responses in S. aurata.

1. IntroductionIn recent decades, efforts have been made in aquaculture to improve management practices and the monitoring of animal welfare trough the evaluation of novel and/or classical stress indicators [1,2,3,4]. Cortisol is the main hormone that promotes compensatory metabolic response to stress in teleost fish [5,6]. Stressful events in natural and farming environments trigger plasma cortisol enhancement, which latently promotes the energetic substrate mobilization in liver and skeletal muscles, allowing for fish acclimatization and homeostasis recovery [7,8]. The relevance of cortisol as a key mediator of the stress response in fish has been extensively studied, including through in vivo experiments associated with exogenous hormone administration using saline and/or oil vehicles [9,10,11]. In gilthead seabream (Sparus aurata) long-term cortisol administration, using diet or slow-release implants, enhanced catabolism together with the regulation of several gluconeogenesis, glycogenolysis and proteolysis related genes [12,13,14].Overall, cortisol metabolic effects are associated with classical/genomic mechanisms involving interaction of intracellular glucocorticoid (GR) and mineralocorticoid receptors (MR), and the subsequent regulation of target genes [15]. Additionally, cortisol can also interact with plasma membrane-localized components activating rapid/intracellular GR-independent signaling pathways with novel features related to early compensatory responses to stress [16,17,18]. In this context, we recently determined that short-term cortisol treatment (1 to 6 h) triggers a metabolic response in the liver of gilthead seabream, characterized by glucose and lactate plasma levels enhancement, hepatic glucose, and glycogen mobilization, as well as the regulation of both glycolysis and gluconeogenesis-related genes [16]. All these metabolic effects were further validated using a specific impermeable-membrane cortisol analog, cortisol-BSA [16,17]. Nevertheless, as of yet, the contribution of long-term membrane-initiated cortisol action on metabolic responses in fish remains unknown.In this work, we evaluate the potential contribution of the prolonged effects of cortisol (three days of treatment) on metabolic and transcriptional responses of gilthead seabream by using cortisol or cortisol-BSA dissolved in coconut oil implanted intraperitoneally. For this purpose, S. aurata juveniles were treated with cortisol or with the membrane impermeable cortisol analog, cortisol-BSA, dissolved in coconut oil. After 72 h, plasma cortisol values and several metabolites in the plasma, liver, and skeletal muscle were determined. In addition, metabolic-related gene expression was also assessed at the hepatic and skeletal muscle levels. The results were discussed in relation to the possible contribution of membrane-initiated cortisol actions on the regulation of metabolic and transcriptional responses of S. aurata.2. Materials and Methods2.1. Experimental DesignImmature gilthead seabream (S. aurata) (24.74 ± 0.29 g body mass, mean ± SEM, n = 32) were provided by Servicios Centrales de Investigación en Cultivos Marinos (SCI-CM, CASEM, University of Cadiz, Puerto Real, Cádiz, Spain; Spanish Operational Code REGA ES11028000312). Fish were randomly distributed in eight 100-L tanks (~2.5 kg/m−3 density) and kept under a natural photoperiod (12:12 h LD) (march, 2018), constant temperature (18 °C) in a flow-through system. Fish were fed by hand twice per day (9:30 and 15:30 h; 2% of tank biomass per day) with commercial pellets. After ten days of acclimation, fish were anesthetized with 2-phenoxyethanol (0.3 mL/L) and treated with the following intraperitoneal implants (10 μL/g body weight): (i) coconut oil alone (Sigma-Aldrich, San Luis, MO, USA) (sham group), (ii) coconut oil containing cortisol (0.138 µmol per g of fish) (Sigma-Aldrich) (cortisol group), (iii) coconut oil containing cortisol-BSA (0.138 µmol per g of fish) (US biological, USA) (cortisol-BSA group), or iv) coconut oil containing BSA alone (0.001784 mg per g of fish) (Sigma-Aldrich) (BSA group). This latter group was included to consider the potential effects of BSA within the cortisol-BSA complex. The experiment was performed using duplicate tanks for each group and fish were not fed during the experiment.After 72 h of treatment, all fish were euthanized through an overdose of 2-phenoxyethanol (1 mL/L) and sampled. Plasma was obtained by centrifugation of the blood (3 min, 10,000× g, 4 °C), snap frozen in liquid nitrogen and stored at −80 °C until further analysis. Liver and skeletal muscle were excised, and the biopsies collected in microtubes were snap frozen in liquid nitrogen and stored at −80 °C until the assay of metabolites. Additionally, other liver and skeletal muscle samples were collected and placed into tubes with 10-volumes (v/w) of RNAlater™ Soln. (Invitrogen by Thermo Fisher Scientific, Waltham, MA, USA), held for 24 h at 4 °C and stored at −20 °C until total RNA isolation.2.2. Measurement of Plasma Cortisol and Tissue MetabolitesPlasma cortisol levels were measured by EIA kit (Arbor assays), which was previously validated in S. aurata [16,19]. Plasma glucose, lactate, and triacylglycerols (TAG) were quantified with the following Spinreact kits (Barcelona, Spain): HK Ref. 1001200, Lactate Ref. 1001330, and TAG Ref: 41030, respectively, adapted to 96-well microplates. Protein plasma levels were measure using a Pierce BCA Protein Assay Kit (Thermo Scientific, Hanover Park, IL, USA).For the analysis of tissue metabolites, skeletal muscle and liver were finely minced on an ice-cooled Petri dish and were homogenized by ultrasonic disruption in 7.5 volumes of ice-cold 0.6 N perchloric acid, neutralized using 1 M KCO3, centrifuged (30 min, 3220× g and 4 °C), after which the supernatant was stored until used for metabolites determination. Tissue lactate and triacylglycerols levels were determined spectrophotometrically with commercial kits (Spinreact, see before). The tissue glycogen concentration was assessed as described by Keppler and Decker [20], whereby tissue homogenates are incubated for 2 h at 37 °C with and without amyloglucosidase (Sigma-Aldrich A7420, San Luis, MO, USA) to break down glycogen molecules into glucose. The total glucose in both incubations was determined with a commercial kit (Spinreact, see above), and the glycogen content was expressed as glucose equivalents after free glucose subtraction [16,21].2.3. RNA Extraction and cDNA SynthesisThe total RNA of the S. aurata liver and skeletal muscle were obtained using the NucleoSpin RNA II kit (Macherey-Nagel) in accordance with the manufacturer’s instructions. The gDNA elimination step was performed using on-column RNase-free DNase digestion, following the manufacturer’s instructions. The RNA integrity was evaluated using 2100 Bioanalyzer using the RNA 6000 Nano Kit (Agilent Technologies, Santa Clara, CA, USA) and RNA quantification was estimated using Qubit® 2.0 Fluorometer (Life Technology, Carlsbad, CA, USA). Only RNA with RIN > 8.0 was used for cDNA synthesis. Retro transcription was carried out with the qScriptTM cDNA Synthesis Kit (Quanta BioSciences), using 500 ng from liver and skeletal muscle RNA as an input, following the manufacturer’s instructions.2.4. Real Time-PCRReal time PCR were performed using the Biorad qPCR system (Quanta BioSciences, Gaithersburg, MD, USA) in a final volume of 20 µL. Details of the qPCR reaction mixture are indicated in [16]. Several calibration plots with serial dilutions of input total RNA from liver and skeletal muscle had amplification efficiencies between 90.3–105.7%, and 95.1–103.8%, respectively.The PCR profile was as follows: 95 °C, 10 min; [95 °C, 20 s; 60 °C, 30 s] × 40 cycles; melting curve [60 °C to 95 °C, 20 min], 95 °C, 15 s. The results were normalized against beta actin (actb) and elongation factor 1a (ef1a) as housekeeping genes. Relative gene quantification was performed using the ΔΔCT method [22]. Candidate sequences corresponding to aldolase (aldo), phosphoenolpyruvate carboxykinase (pepck), glucose 6 phosphatase (g6pc), actb and ef1α were available in the NCBI database. Conversely, a sequence of phosphoglycerate mutase 1 (pgam1), enolase 3 (eno3), atrogin-1, and muscle RING-finger protein-1 (murf-1) was obtained from the available database belonging to the S. aurata sequencing project [23]. Primers were designed using the Primer 3 available tool (http://frodo.wi.mit.edu/primer3/; accessed on date 3 February 2018) and validated in NetPrimer (http://www.premierbiosoft.com/netprimer/; accessed on date 6 February 2018) and Oligo analyzer 3.1 (https://www.idtdna.com/calc/analyzer/; accessed on date 6 February 2018) available online tools. Finally, the primers sequence of the glucocorticoid receptor 1 (gr1), glucocorticoid receptor 2 (gr2) and mineralocorticoid (mr) were obtained from [24].2.5. Statistical AnalysisThe normality and homogeneity of variances were analyzed using the Kolmogorov-Smirnov’s and the Levene’s tests, respectively. All data were analyzed using a one-way ANOVA with treatment as the factor of variance, followed by a post-hoc Tukey’s honestly significant difference (HSD) test. All statistical analysis were performed using the Graph Prism 7.0 software (GraphPad Software, Inc., San Diego, CA, USA). Finally, differences were considered using a p value of <0.05.3. Results3.1. Plasma Indicators in S. aurata Administered with Cortisol or Cortisol-BSAPlasma cortisol levels significantly increased in the cortisol (100.37 ± 13.12 ng/mL) and cortisol-BSA (270.67 ± 17.17 ng/mL) injected groups compared to the sham (17.41 ± 5.13 ng/mL) and BSA (6.52 ± 1.75 ng/mL) groups, respectively (Figure 1).The increase of cortisol resulting from the administration of cortisol, but not cortisol-BSA, enhanced glucose (6.95 ± 0.39 mmol/L) and TAG (257.39 ± 24.64 mg/dL) plasma levels compared to the vehicle group (glucose: 4.07 ± 0.19 mmol/L and TAG: 92.64 ± 14.12 mg/dL) (Table 1). Finally, neither plasma proteins nor lactate displayed significant differences between the cortisol or cortisol-BSA groups against vehicle or BSA groups (Table 1).3.2. Energetic Metabolites in Liver and Skeletal Muscle of S. aurata Administered with Cortisol or Cortisol-BSAWith respect to liver energetic metabolites, the glycogen content increased in the cortisol group (0.79 ± 0.09 mg/g) but not in the cortisol-BSA administered group (0.23 ± 0.06 mg/g) (Table 2).According to the analyzed skeletal muscle energy metabolites, we only observed a non-significant tendency of increase in TAG for the cortisol-treated group (2.94 ± 1.17 mg/g) but not in the cortisol-BSA administered group (0.46 ± 0.15 mg/g) compared to the vehicle group (1.30 ± 0.55 mg/g) (Table 3). Finally, there were no changes in skeletal muscle glucose, glycogen, and lactate of S. aurata administered with cortisol or cortisol-BSA (Table 3).3.3. Corticosteroid Receptor-Related Genes Expression in Liver and Skeletal Muscle of S. aurata Administered with Cortisol or Cortisol-BSABy analyzing corticosteroid receptor-related gene expression, it was found that gr1 and gr2 mRNA levels in liver did not change between the cortisol or cortisol-BSA groups compared to in the vehicle group (Figure 2A,B). However, mr mRNA levels decreased in cortisol (0.57 ± 0.06 relative expression) but not for cortisol-BSA (0.95 ± 0.08 relative expression) administered fish compared to the vehicle group (0.95 ± 0.05 relative expression) (Figure 2C).On other hand, there were no changes observed for gr1 and mr expression in the skeletal muscle of the S. aurata administered with cortisol or cortisol-BSA (Figure 3A–C). However, gr2 mRNA levels increased in cortisol but not in cortisol-BSA injected fish (Figure 3B).3.4. Glucose Metabolism—And Atrophy-Related Genes in S. aurata Administered with Cortisol or Cortisol-BSAThe gluconeogenesis-related genes pepck and g6pc had an enhanced expression in the liver as a result of cortisol administration, but not for cortisol-BSA (Figure 4A,B). Additionally, we found a high correlation between both hepatic gluconeogenesis-related genes and plasma glucose levels (Supplementary Figure S1). With regard to glycolysis-related gene expression, eno3 increased its expression in liver for cortisol but not cortisol-BSA administered S. aurata (Figure 4E). On the other hand, neither hepatic aldol nor pgam1 expression is found to be modulated by cortisol or cortisol-BSA administration (Figure 4C,D).Otherwise, the glycolysis-related gene pgam1 increased its expression (0.74 ± 0.27 relative expression) in the skeletal muscle under cortisol but not cortisol-BSA treatment (Figure 5B), while there were no changes in the expression of eno3 or aldo in this tissue (Figure 5A–C).Finally, regarding the atrophy-related genes expression in skeletal muscle of S. aurata administered with cortisol or cortisol-BSA, atrogin-1 and murf1 had an enhanced expression (1.50 ± 0.22 and 1.34 ± 0.15 relative expression, respectively) the in cortisol administered group, but not in the cortisol-BSA administered group, compared to the vehicle group (0.62 ± 0.09 and 0.85 ± 0.10 relative expression, respectively) (Figure 6A,B).4. DiscussionCortisol is an essential glucocorticoid involved in the regulation of physiological and metabolic responses of teleost fish under basal and stressful conditions [5,12,25,26]. Cortisol-related effects on metabolism, growth, immunity, and other physiological processes have been evaluated through the exogenous administration of cortisol (in saline or oil as a vehicle) to determine its potential effects under acute or chronic stress [27,28,29]. Due to cortisol’s ability to activate intracellular and extracellular initiated actions, the use of a plasma membrane impermeable analog of cortisol, such as cortisol-BSA, seems useful to discriminate membrane-initiated cortisol effects during long-term stress responses [16,17,18]. In the present work, we found that, in S. aurata plasma and tissue, energy metabolites as well as hepatic and skeletal muscle expression of selected metabolism-related genes are mediated by cortisol but not cortisol-BSA treatment. We previously determined that cortisol-BSA is suitable to induce rapid specific membrane-initiated effects (1-6 h) in S. aurata [16], in agreement with the reports for other teleosts [30,31]. Additionally, the inability to cross the plasma membrane and its stability over an extended period has been proved in steroids attached to BSA molecules [6,17,32].Particularly, we determined that cortisol, but not cortisol-BSA, increased the hepatic glycogen content, which has been previously reported in S. aurata juveniles injected with slow-release cortisol implants [33]. Furthermore, the plasma cortisol levels that were reached in our in vivo experiment are similar to those observed in confined S. aurata for 24 to 72 h [34,35]. In addition, cortisol-administered fish enhanced plasma TAG values, similar to the levels found in this species during the recovery period after a stressful process [36]. Since cortisol-mediated stress has been linked to the regulation of the lipid metabolism, we suggest that genomic cortisol actions specifically mediate this response [33,37]. Taking into account these results, and the absence of metabolism-related processes regulated by cortisol-BSA, we postulate that metabolic reprogramming in S. aurata after three days of cortisol-mediated stress could be triggered exclusively by intracellular GR signaling. This is in agreement with previous studies conducted for other teleost fish species [38,39].Cortisol is capable of interacting with glucocorticoid (GR) and mineralocorticoid (MR) receptors to exert its effects in the cell [15]. Most fish, including S. aurata, demonstrate two or more GR with different expression patterns under stress/cortisol treatment [25,40,41]. We found that hepatic gr1 and gr2 mRNA levels remained unaltered after cortisol or cortisol-BSA treatments. Our results are therefore in agreement with previous studies that reported no variation in gr1 or gr2 expression levels in rainbow trout after five days of cortisol treatment, coinciding with the lack of variation in gr mRNA levels in S. aurata after seven days of cortisol treatment [13,41]. On the other hand, even though fish do not synthesize aldosterone (a mineralocorticoid), which represents the canonical ligand of MR, fish do express MR in different tissues including liver and skeletal muscle [16,42]. We found that cortisol, but not cortisol-BSA, decreased mr hepatic expression in S. aurata. Interestingly, similar approaches using slow-release cortisol implants in Cyprinus carpio induced an opposite behavior in this expression [43]. We speculate on the presence and differentiated control of species-specific mr expressed in teleosts. However, the role of MR signaling pathways in the liver of fish is poorly understood and requires further study [42]. Regarding the skeletal muscle of S. aurata, cortisol, but not cortisol-BSA, treatment enhanced gr2 expression levels after three days. This agrees with the reported results in the fine flounder (Paralichthys adspersus) for which 28 days of chronic stress increased plasma cortisol levels alongside gr2, but not gr1, mRNA levels in the skeletal muscle [44]. Additionally, gr1 and gr2 show differentiated expression patterns during an acute stress response in S. aurata [24]. Therefore, we postulate that GR2 regulates the cortisol effects in the skeletal muscle of this species. It is important to note that gr1 expression increased in in cortisol-BSA group compared to the sham-control group, but not in the BSA group. In this context we cannot rule out whether BSA or membrane-initiated cortisol actions mediate this effect. We also observed that both atrogin-1 and murf1 are regulated by cortisol but not by cortisol-BSA in this tissue. Both are critical atrophy-related genes that increase in expression under stress and/or glucocorticoid stimulus in fish [45,46]. Moreover, S. aurata subjected to stress-related conditions, such as fasting or exhaustive swimming, demonstrated an increased atrogin-1 and murf1 expression in the skeletal muscle [46,47]. We have recently reported the contribution of membrane-initiated cortisol actions in the early regulation of both atrogenes in isolated fish myotubes [48]. However, after prolonged cortisol treatment, atrophy-related processes are mediated by genomic cortisol signaling. This agrees with the reported muscular proteolytic action of cortisol as well as with its negative influence on growth during chronic stress situations [49,50]. It is well known that cortisol-mediated stress induces changes in the hepatic expression of gluconeogenesis-related genes in fish [13,51,52]. The observed hepatic pepck and g6pc expression enhancement in S. aurata administered with cortisol agree with this idea. The fact that cortisol-BSA does not modulate the expression of pepck and g6pc after 3 days of treatment leads us to hypothesize that both key gluconeogenesis-related genes are exclusive targets of cortisol action in teleosts. In particular, pepck presents a glucocorticoid response element (GRE), supporting the idea of its direct regulation by cortisol during the stress response [38]. On other hand, we have previously shown that S. aurata hepatic g6pc, but not pepck, expression is mediated by cortisol-BSA during early treatment (1 to 6 h) [16]. Overall, these data suggest that membrane-initiated cortisol actions participate in early-term gluconeogenesis regulation through g6pc induction, followed by cortisol genomic long-term actions trough the increase of both pepck and g6pc expression. Regarding those genes related to the glycolysis-metabolism via cortisol, the level of hepatic eno3 and skeletal muscle pgam1 expression enhancements are in agreement with previous reports in teleosts, including S. aurata, that are subjected to different stressors [53,54]. In this context, glycolysis, exclusively mediated by genomic cortisol actions, is critical for releasing energy in fish so as to overcome stress [5,6,25]. Finally, it is important to mention that all metabolic and transcriptional effects mediated by cortisol were observed using one experimental period (3 days). In this context, the potential effects of membrane or genomic-initiated cortisol actions during more prolonged treatments (weeks to months) presents a potential research avenue for future studies. 5. ConclusionsConsidering all of the metabolic and transcriptional results, we propose that a rapid cortisol signaling pathway plays a role in gluconeogenesis-glycolysis and atrophy-related processes. Its effects are evident at an early stage during the stress response in S. aurata, and is exclusively followed by the regulation of genomic cortisol pathways. Overall, our results suggest that genomic cortisol actions exclusively (potentially mediated by intracellular GR) participate in metabolic responses during prolonged treatment (3 days), as was observed in our study, using cortisol in S. aurata.

animals : an open access journal from mdpi

10.3390/ani13091550

PMC10177029

A salivary mucocele is the most common disorder affecting the salivary glands. Saliva leakage from the salivary gland parenchyma and/or from associated duct damage causes a chronic inflammatory process which can occasionally result in osseous metaplasia. Dogs having an ossified sialocele present with a hard, non-fluctuating mass containing a viscous, sticky liquid at centesis. Either radiography or computed tomography can confirm the thick bone-like wall pseudocyst. Surgical excision of both the pseudocyst and the affected salivary gland represents the therapeutic gold standard in these cases. The aim of this study was to report the clinical and diagnostic findings, surgical management and histopathologic report of three cases of cervical sialocele complicated by pseudocapsule osseous metaplasia.

Saliva is an irritant of the subcutaneous tissue, thus causing the development of a non-epithelial reactive pseudocapsule. Metaplastic ossification of the pseudocapsule is a condition rarely described in the veterinary literature. The main causes of calcification are trauma, tumours, various chronic inflammatory conditions and fibrodysplasia ossificans progressiva. The aim of the present case series was to describe three dogs affected by a calcified salivary mucocele. The medical records of dogs affected by a cervical sialocele were retrospectively evaluated, and three cases met the inclusion criteria. All the dogs in this study were referred to the Veterinary Teaching Hospital (VTH) of the Department of Veterinary Sciences of the University of Turin (Turin, Italy) for a large solid mass in the intermandibular region. The diagnosis of a mucocele was confirmed clinically by centesis and by radiography or CT. Complete excision of both the pseudocyst and the ipsilateral mandibular/monostomatic sublingual salivary gland was performed in all cases. The histological report showed large areas of bone metaplasia within the pseudocapsule and chronic sialadenitis. Based on this limited case series, complete excision of the pseudocyst and a concurrent sialoadenectomy provided an effective treatment for this rare salivary mucocele disorder.

1. IntroductionSalivary gland (SG) disorders have rarely been described in dogs and cats, with only a 0.3% overall prevalence of veterinary consultations [1]. Salivary gland disorders in dogs include neoplasm, sialadenitis, salivary mucocele (commonly called sialocele) and various degenerative or fibrotic lesions [1]. A sialocele is an accumulation of saliva in the subcutaneous tissue with a consequent inflammatory reaction to saliva [2]. A sialocele is lined by inflammatory reactive connective tissue and not by epithelial tissue which, however, does occur in a pharyngeal cyst. Young male dogs appear to be those most represented; the breeds most affected are Poodles and German Shepherds; however, all breeds can be affected [3,4,5]. Trauma is considered to be the primary cause of a sialocele, thus explaining its prevalence in younger dogs. Other suggested causes are foreign bodies, sialoliths, haematomas, neoplasia and recent oral surgery. However, in the majority of cases, the real cause remains unknown [6]. A sialocele is classified as sublingual, cervical, pharyngeal or zygomatic. The sublingual/mandibular gland and duct complex is the most commonly affected structure [3]. The diagnosis is based on clinical presentation, history of onset and the result of the centesis. Tumours and abscesses may appear clinically similar, but they are generally either firm or painful. A thyroglossal cyst, a pharyngeal cyst and lymphadenopathy should also be considered when making a differential diagnosis [7,8,9,10].The clinical presentation depends on the location: A cervical sialocele usually appears as a slowly enlarging nonpainful swelling in the intermandibular area. It can sometimes be difficult to differentiate the side of the affected gland involved in sialocele; if located on the midline, it may only slightly shift toward the originating side in a patient placed in dorsal recumbency. Centesis usually reveals a transparent viscous mucoid fluid, sometimes slightly bloody and/or cloudy. Periodic Acid–Schiff (PAS) staining can confirm this fluid as being saliva by revealing mucin [11,12].A sialocele can be investigated using radiography/sialography, ultrasound, magnetic resonance imaging (MRI) and computed tomography (CT) [13,14,15,16]. Sialography has demonstrated the rupture of the duct/gland complex in 55% of cases [17]. Radiography usually reveals a uniform soft tissue swelling at the level of the cervical region while a fluid attenuating, non-contrast enhancing lesion with a smooth, soft tissue attenuating, contrast-enhancing wall is more often observed with CT. The size, shape and degree of loculation are variable [18].Metaplastic ossification within the sialocele, for the most part at the level of its capsule, has previously been reported [19,20,21,22], and is likely due to chronic inflammation in the presence of granulation tissue.The pseudocapsule of a salivary mucocele is microscopically visible as granulation tissue enclosing mucin; it is made up of edematous, vascularised connective tissue containing macrophages, lymphocytes and other chronic inflammatory cells.In human medicine, an oral sialocele is a common SG disorder which appears more often on the lower lip mucosa or mandibular vestibule and is frequently caused by accidental lower lip biting. Bone metaplasia has also been reported in individuals with a caecal appendix [23,24] and a paranasal sinus mucocele [25,26].In the past, the treatment of sialocele with drainage alone was characterized by a high rate of recurrence [13]. Surgical excision is the treatment of choice and involves the removal of the entire SG and ductus complex in order to avoid recurrence. At present, the need for complete pseudocapsule resection is under debate [2,11]. Sialoadenectomy can be performed using either a lateral or ventral [6] surgical approach; both approaches have been proven to be effective [27]. The sublingual/mandibular SG and duct complex can be removed bilaterally without risking a dry mouth as the production of saliva is ensured by other major and minor salivary glands [13].No cases of human oral ossified sialocele have been reported to date. In veterinary medicine, only four case reports of sialocele-associated osseous metaplasia have been described [19,20,21,22].2. Materials and Methods2.1. Case SelectionThe medical records of dogs referred to the Veterinary Teaching Hospital (VTH) of the Department of Veterinary Sciences of the University of Turin (Turin, Italy) for sialocele from January 2000 to January 2023 were identified and reviewed.The inclusion criteria for this case series were as follows: (a) the presence of a mandibular/neck hard swelling; (b) a first or second level of diagnostic investigation to confirm the presence of a calcified lesion; (c) a sialoadenectomy, and complete or partial removal of the calcified lesion; (d) a complete histological report and (e) a minimum 30-day follow-up.Written consent was obtained from the owners to proceed with anaesthesia and all the diagnostic and therapeutic procedures.Perioperative standard-of-care management, including analgesia, was performed in all dogs. This study did not fall within the application areas of Italian Legislative Decree 26/2014, which governs the protection of animals used for scientific or educational purposes. Therefore, ethical approval was waived for this study. The animals were not treated as part of an experimental study; only the data were later selected and included in this study. No specific informed consent statement was obtained for inclusion in this retrospective study.2.2. Diagnostic ImagingThe head and neck were investigated using CT (16 MDCT unit, Siemens Somatom Emotion) or radiographically (latero-lateral and ventro-dorsal views). For the CT scan evaluation, the dogs were premedicated with an intramuscular combination of dexmedetomidine (2 mcg/kg) (Dexdomitor®, Orion Pharma, Turku, Finland) and methadone (0.2 mg/kg) (Synthadon®, Le Vet Beheer B.V., Oudewater, The Netherlands). General anaesthesia was induced using propofol (3 mg/kg, IV) (Proposure®, Merial Italia Spa, Noventa Padovana, Italy) and was maintained with isoflurane in oxygen (Isoflo; Esteve Spa, Barcelona, Spain). The CT images were acquired in basal condition and after the intravenous administration of contrast medium Iomeron 400 mg/mL (1.5 mL/kg IV) (Bracco S.p.A., Milan, Italy).2.3. SurgeryFor patients undergoing a CT scan, the surgery was performed separately from the diagnostic imaging procedure in order to shorten the anaesthesia time.After general anaesthesia, the dogs were clipped and moved to the operating theatre where they were positioned for the surgery and aseptically prepared. The intravenous administration of cefazolin (20 mg/kg) (Cefazolina Teva, Teva Italia SRL, Milano, Italy) was performed 20 min before skin incision and was repeated every 60 min until extubation. The affected SG was removed together with the entire ossified lesion. The suction of saliva was performed using a Poole suction tube. Both SGs and the pseudocapsule underwent histopathology; however, a portion of the pseudocapsule also underwent microbiological analysis. A passive drain was inserted at the surgeon’s discretion.2.4. HistopathologyBoth SGs and the pseudocapsule were fixed in 10% buffered formalin, decalcified for 48 h in Osteodec (Bio-Optica S.p.A., Milan, Italy), embedded in paraffin, sectioned at 4 μm and routinely stained with haematoxylin and eosin. A histological diagnosis was reached by the pathologists of the Department of Veterinary Science, University of Turin.2.5. Postoperative Care and Follow-UpPostoperative management included intravenous fluid (Lactated Ringer’s 2 mL/kg/h); analgesia with methadone (Semfortan®, Eurovet Animal Health B.V., Bladel, The Netherlands; 0.2 mg/kg IV q4h) or buprenorphine (Temgesic®, Schering-Plough, Segrate, Italy; 10 mcg/kg IV q8h); and meloxicam (Metacam®, Boehringer Ingelheim, Ingelheim/Rhein, Germany; 0.2 mg/kg IV the first day and 0.1 mg/kg IV thereafter). The dogs were discharged within 1–5 postoperative days with an Elizabethan collar and meloxicam (Metacam®, Boehringer Ingelheim, Ingelheim/Rhein, Germany; 0.1 mg/kg daily for seven days, PO), and were then re-evaluated weekly for a month after discharge. Complications were classified according to the guidelines of the “Veterinary Co-operative Oncology Group—Common Terminology Criteria for Adverse Events” (VCOG-CTCAE) [28]. The skin sutures were removed ten days post-operatively, and a first clinical evaluation was carried out. Additional follow-up information was acquired by periodic clinical evaluation or by direct communication with owners or the referring veterinarians.3. ResultsThree dogs presenting an ossified sialocele met the inclusion criteria and were included in this case series.3.1. Case 1A 3-year-old, 28 kg, intact female Boxer was referred to the VTH for a 6 cm painless firm swelling at the level of the right ventrolateral region of the neck. The swelling had been present for 3 months and had gradually grown. The dog showed no clinical signs of pain or discomfort except for drooling. Both the physical examination and laboratory profile (complete blood count, biochemical values and urinalysis) were within normal limits. There was no clinical evidence of regional lymphadenomegaly.Sterile needle centesis was performed and revealed the presence of a stringy, brown fluid. Cytology of this fluid was consistent with saliva, owing to the presence of mucin aggregates, with no evidence of infection. Radiographic examination of the head and neck showed a large soft tissue consolidation with a calcified wall between the right intermandibular and subauricular regions. A skull CT scan was performed to identify the origin of this consolidation and its anatomical relationship with the surrounding structures. The examination revealed an ovoid capsulate mass (3.6 × 4.7 × 4.3 cm) at the level of the right pharyngeal region (Figure 1). The capsule was thickened (4 mm) and mineralised in its dorsal, lateral and medial aspects; it was thin and contrast enhanced in its ventral part. The content was fluid attenuating (around 25 HU) with no contrast enhancement. The structure was dorsal to the ipsilateral mandibular SG that appeared moderately compressed and dislocated cranially. Its medial aspect was in contact with the hyoid apparatus and nasopharynx, extending dorsally and nearing the tympanic bulla, while its lateral aspect was close to the angle of the mandible. The right medial retropharyngeal lymph node was slightly enlarged (2.5 × 1.3 cm), showing a heterogeneous contrast enhancement. The right lateral retropharyngeal lymph node (1.3 × 1 cm) was dislocated laterally and ventrally, close to the mandibular lymph nodes.The dog was placed in left lateral recumbency, and a rolled towel was placed under the neck to elevate the surgical site. A lateral surgical approach to the right mandibular SG was performed. The cystic mineralised lesion was operated on using a 10 cm skin incision, from the ventral aspect of the vertical ear canal to the caudal portion of the ramus of the mandible, just over the swelling. The platysma and parotidoauricularis muscles were incised and reflected ventrally. The lateral portion of the calcified lesion was gently isolated from the surrounding tissue and completely excised (Figure 2).Part of the capsule and its contents were sent for microbiological analysis. The right medial and lateral retropharyngeal lymph nodes were excised.Subsequently, a mandibular/sublingual SG and duct sialoadenectomy was performed; the duct was ligated and transected as rostrally as possible at the level of the lingual nerve, which was spared. Using a combination of blunt and sharp dissection, the glandular tissue and its duct were dissected from the capsule and excised by tunnelling under the digastricus muscle. The surgical site was flushed with warm saline solution and a Penrose drain was placed within the subcutaneous space exiting ventral to the incision. The incision was closed routinely in three layers. Part of the capsule, the excised SG and lymph nodes were sent for histological evaluation.The dog recovered uneventfully within a week. The bacteriological analysis of the pseudocapsule was negative for any bacterial growth.Histologically, the SGs showed multifocal moderate lymphoplasmacytic chronic sialoadenitis with the interstitial spaces expanded by thick bundles of fibrocytes (fibrosis). The cystic lesion was lined by granulation tissue and small foci of an osseous metaplasia (Figure 3). The lymph nodes removed were characterised by hyperplastic lymphadenitis.There was no sign of recurrence more than five years after surgery.3.2. Case 2A 1.5-year-old, 38 kg, intact male Italian Cane Corso was admitted for an 8 cm hard swelling just caudal to the left mandibular ramus, which had been present for 6 months and significantly increased in size during the previous 2 weeks. The consistency appeared solid at physical examination. The macroscopic aspect of the fluid obtained by centesis was compatible with saliva; however, cytology was inconclusive. The latero-lateral radiographic view of both the skull and the cervical region highlighted an irregularly rounded area of bone-like radiopacity at the level of the pharyngeal region superimposed to the hyoid apparatus (Figure 4). Both the physical examination and laboratory profile (complete blood count, biochemical values and urinalysis) were within normal limits.Based on these findings, a suspicion of sialocele was formulated and a left mandibular/sublingual SG and duct excision was proposed. The patient was positioned in right lateral recumbency. A rolled towel was placed under the neck to elevate the surgical site. A vertical incision centred on the ossified lesion was performed caudally to the angular process of the mandible (Figure 5). The subcutis was undermined with blunt and sharp dissection, and the mandibular/sublingual SGs were identified between the linguofacial and maxillary veins. The capsule of the glands appeared to strictly adhere to the ventro-caudally located ossified lesion (Figure 5). The mandibular and the sublingual glands were dissected free from both the surrounding tissues and the sialocele. The salivary duct was followed rostrally and dissected; the lingual branch of the trigeminal nerve was identified and freed from the duct. The latter was then ligated and transected as rostrally as possible. Two ipsilateral mandibular lymph nodes were removed for histopathology as they were enlarged. The thick ossified capsule of the sialocele was then incised, revealing a brownish viscous content which underwent microbiological analysis together with a portion of the capsule. The sialocele was freed from the surrounding tissues and excised. The surgical site was flushed with warm saline solution and a Penrose drain exiting from a separate ventral incision was inserted. The incision was routinely closed in three layers. Part of the capsule, the excised SG and lymph nodes were submitted to histology. The drain was removed 3 days later. The dog healed without complications within 10 days and the sialocele never recurred and died 8 years later from causes not related to the sialocele. The bacteriological analysis was negative for any bacterial growth. Histological examination of the SGs showed a multifocal moderate lymphoplasmacytic chronic sialadenitis with moderate ductal hyperplasia and mild interstitial fibrosis. The cystic lesion surrounded an inner space containing amorphous eosinophilic material admixed with cellular debris (necrosis) and was composed of a fibrous capsule admixed with granulation tissue, foci of osseous metaplasia and aggregates and small mature lymphocytes. Histological examination of the mandibular lymph nodes revealed diffuse lymphoid follicular hyperplasia and foci of interstitial fibrosis.3.3. Case 3A 12-year-old, 8 kg, spayed female Dachshund was referred for a 5 cm painless swelling in the intermandibular region. The lesion had grown during the previous two weeks and appeared firmly attached to the surrounding tissues. The lesion had a bony/cartilaginous consistency; however, its ventral border appeared incomplete as several small openings were palpated. Due to the compression on the pharynx and the trachea, the dog presented with dyspnea on exertion. A small right sublingual sialocele was also present. Haematologic laboratory tests showed increased gamma-glutamyl transferase, bilirubin, total protein and hypoalbuminemia. These anomalies, associated with potbelly and bilateral and symmetrical back alopecia, were suggestive of hyperadrenocorticism, and an internal medicine specialist consultation was required. Low dose dexamethasone suppression test confirmed Cushing’s syndrome. Cytological analysis of the swelling was compatible with saliva; therefore, a calcified sialocele was clinically suspected.A CT scan revealed a complex cystic formation extending from the right mandibular region, between the tongue and the mandible, up to the ipsilateral laryngeal region. The swelling had dimensions of 4.5 cm (width) × 5.8 cm (height) × 6.1 cm (length) with multiple concamerations and partially mineralised walls (approximately 2 mm thick); the content was fluid attenuating with no contrast enhancement (average attenuation of 20 HU). The medial aspect of the lesion was in contact with the hyoid apparatus and the larynx, which were dislocated and compressed to the left, while the caudal margin was in contact with the right mandibular lymph node and the facial vein. The swelling extended dorsally in the pharyngeal region, reaching the right tympanic bulla and the external carotid artery. The right mandibular SG appeared adherent to the lateral wall of the swelling, resulting in its appearing dislocated and compressed laterally. The right retropharyngeal lymph node was enlarged (2.2 × 1.9 cm) with dishomogeneous contrast enhancement and was in contact with the caudal aspect of the swelling (Figure 6).The adrenal glands were enlarged bilaterally. The right measured 2.2 cm × 1.1 cm and the left 2.3 cm × 1 cm with dishomogenous contrast enhancement. No macroscopically visible pituitary gland abnormalities were found.Routine diagnostic tests confirmed the suspicion of primary hyperadrenocorticism, and trilostane (Vetoryl®, Dechra Veterinary Production S.r.l., Northwich, UK) therapy at 1 mg/kg q24h was started. On the day of surgery, the dog was positioned in dorsal recumbency, with a towel placed under the neck to slightly elevate it. A skin incision in the intermandibular region centring on the ossified lesion was performed. The platysma muscle was incised, and the ventral aspect of the lesion was separated from the surrounding tissues (Figure 7). The caudolateral aspect of the sialocele was freed from the mandibular SG capsule. A branch of the lingual vein, which tightly adhered to the capsule of the lesion, was ligated and transected in order to proceed with the dissection. The thick capsule was incised, and the brownish fluid was aspirated (Figure 7). Part of the pseudo capsule and part of its contents were sent for bacteriological analysis. The ventral portion of the sialocele was excised using Mayo scissors, revealing a multiloculated cavity. Two stay sutures were medially and laterally positioned on the remaining wall of the capsule in order to facilitate its separation from the most dorsal tissues, which were tightly adhered (but not infiltrated) to the thyroid cartilage. With a combination of blunt and sharp dissection, the ossified sialocele was completely excised. The right mandibular/sublingual SGs and duct were then surgically excised (Figure 7).The duct was tunnelled under the digastricus muscle and freed from the mylohyoideus muscle to get as close as possible to the sublingual caruncle. The enlarged right medial retropharyngeal lymph node was also excised. A Penrose drain exiting ventral to the surgical site from a small separate incision was inserted, and the wound was closed in three layers. The dog was discharged from the Intensive Care Unit of the VTH 3 days later after removal of the drain. Histological analyses were done on the pseudocapsule, the excised SG and lymph nodes.At the clinical check-up one week after surgery, the sublingual sialocele appeared to have been completely resolved. The bacteriological analysis carried out on the pseudocapsule and aspirated fluid was negative for any bacterial growth. The dog underwent periodic check-ups for the management of Cushing’s syndrome. At the time of writing, twelve months after surgery, there was no recurrence of the sialocele.The histological SG report showed a multifocal moderate granulomatous chronic sialadenitis. The pseudocapsule was constituted of fibroconnective tissue with abundant foci of osseous metaplasia while the medial retropharyngeal lymph node appeared to be hyperplastic.4. DiscussionSialocele is the most common SG disorder affecting the mandibular/sublingual gland or duct [29]. The diagnosis is relatively simple in the majority of cases as the aspiration of the non-painful swelling usually reveals a viscous, honey-coloured, clear or blood-tinged fluid consistent with saliva. Clinical suspicion is confirmed by cytology, which usually reveals a small to moderate number of non-degenerated nucleated cells and diffuse aggregates of mucin [2]. In the case series reported here, the clinical presentation was different from that usually found in sialoceles as palpation did not reveal a poorly defined and fluctuating swelling but a solid mass. On palpation, the calcified consistency was clearly perceptible at the level of the cranial cervical region or slightly more lateral. This condition widened the range of the differential diagnosis, thus including other head and neck lesions which could eventually calcify, such as tumours, inflammatory processes and congenital cysts. The differentiation between a congenital cyst and a sialocele can only be reached histologically by visualising a secretory epithelium in the inner layer of the capsule in the first case and the granulation tissue in the second one [10].Three subtypes of ectopic ossification may occur in the body: osseous choristoma, defined as the presence of normal bone cells in an atypical location; heterotopic ossification, defined as mineralisations due to systemic disease and osseous metaplasia, in which osteoblasts originate from other mesenchymal cells, such as fibroblasts. It has been hypothesised that this change occurs in order to replace more stress-sensitive cells with more resistant ones in an adverse environment, such as that created by chronic inflammation [24,25,30,31] and trauma [32]. A sort of genetic predisposition for osseous metaplasia has also been hypothesized [19]. It is probable that the chronic inflammation caused by the saliva in contact with the soft tissues is the underlying cause of the sialocele calcification [21]. Calcifications have been widely reported in the veterinary literature and appear to be related to tumours, endocrinopathies and inflammatory lesions [33,34]. Tumours which can typically present calcification spots are pilomatricomas, benign follicular tumours located on the neck and trunk [33,35,36]. Calcifications have also been reported in prostatic adenocarcinoma, cortical adrenal gland tumours [37,38] and metastatic lymph nodes [39,40]. Calcification within a thyroidal mass is not very common [41,42], and is often regarded as malignancy. Some endocrinopathies such as hyperadrenocorticism may also predispose to calcium deposition, although mainly affecting the dermis [34]. However, for dog #3, affected by hyperadrenocorticism, it cannot be excluded that the ossification process of the sialocele was accelerated by her endocrine condition. Calcinosis circumscripta is a rare syndrome generating calcified lesions, typically at the level of the subcutis and tongue. The pathogenesis of this syndrome is poorly understood; however, the outcome is usually positive after surgical excision, when possible [43]. Although not common, calcification is also widely documented in extraparenchymal prostatic cysts [44,45,46]. Bigliardi et al. [36] reported the removal of a 25 cm solid extraparenchymal prostatic cyst having a 1 to 3 mm thick wall, composed of vascularised fibrous connective tissue with multifocal osseous metaplasia and dense connective tissue infiltrated by a mixed population of neutrophils and lymphocytes. As these histological features were similar to those reported for calcified sialoceles, it could be hypothesised that the mechanism of occurrence was the same and was related to chemical insult.Any mass should first be sampled in an attempt to determine its nature [47]. In the cases presented herein, the calcification was not a limitation to the needle penetration as the metaplasia did not involve the entire pseudocapsule. Moreover, even in the case of complete calcification, a 2–3 mm capsule should be easily perforated by the needle. The macroscopic characteristics of the aspirated fluid were similar to those already described in previous studies regarding sialoceles [29,48,49].In the literature, there are only sporadic case reports of dogs affected by ossified sialoceles [19,20,21,22]. Although the presence of ossification foci within the sialocele is a common occurrence in veterinary medicine [18], complete pseudocapsule ossification is rare. In two of the cases reported here, calcification foci were also found within the sialocele, adherent to its capsule.Clinical presentation alone, similar to the unusual cases included in this case series, may not be enough to reach a diagnosis, identify the affected structures and plan the surgical procedure; in fact, diagnostic imaging is required. Radiography is able to confirm the calcification in the pseudocapsule but does not provide additional information, such as the origin of the lesion. Sialography is reported to have an accuracy of 66.7% in identifying the SG involved in the sialocele [50]; however, it has been almost completely replaced by more advanced imaging procedures which can also evaluate the extension of the lesion as well as its relationship with the surrounding tissues. A CT scan allows immediate and complete evaluation of the skull, neck and whole body when a complicated sialocele is suspected and a tumour and its metastasis have to be excluded.Furthermore, when a sialocele is suspected, its location (right or left side, or bilateral) in relation to the position of the SGs should be evaluated. The suspected SG is generally contiguous to the fluid collection, with the affected gland which may appear atrophied, normal or enlarged at CT. The attenuation of SGs at CT may be normal or altered, sometimes with fatty infiltration or fluid pockets. These findings, associated with the presence of local fat stranding, may help in identifying the SG involved in the sialocele. Moreover, the need for advanced imaging in a preoperative setting is confirmed by the anatomical variations of the mandibular SG reported in the recent literature [51]. Although in some cases it is not possible to define which gland is leaking saliva, in the majority of cases, the collection occurs contiguously with the capsule of the affected gland. The latter may also appear as contrast enhanced with respect to the contralateral gland. A CT scan was available for two of the three cases reported herein, and the findings were suggestive of sialocele in both cases. In case #1, the SG origin clearly appeared to be the right sublingual/mandibular complex while in case #3, the lesion was at the level of the median sagittal plane, and both the sublingual/mandibular SGs appeared clinically and tomographically normal. For case #2, the marked lateralisation of the sialocele, together with its rostral extension, allowed clinically defining the side of the affected sublingual/mandibular SGs, even without the help of a CT scan which, at that time, was not available at the authors’ VTH. In all the cases reported herein, the sialocele originated from the sublingual/mandibular SGs and the corresponding ducts. This is in line with the literature which reports these structures to be those most frequently affected [3,52].In a sialoadenectomy and duct excision regarding sublingual and mandibular SGs, both ventral and lateral approaches can be performed [2]. The ventral approach is preferred by some authors as it permits removal of the entire sublingual gland-duct complex [2]. It is associated with a lower risk of sialocele recurrence, and it is useful when a bilateral sialoadenectomy is planned. However, the ventral approach seems to be associated with an increased risk of wound-related complications [27]. Regarding the lateral approach, tunnelling under the digastricus muscle seems to increase salivary duct exposure, minimising the risk of leaving remnants of the rostral SG in situ [53]. In the cases reported herein, the choice of the surgical approach was, for the most part, based on the location of the saliva collection. In patient #3, a ventral approach was used to provide both a better visualisation of the sialocele and to excise as much of the salivary duct as possible, as a ranula was also present. On CT, both SGs appeared to be compressed by the lesion. The right sublingual/mandibular SG complex was excised as, after the positioning of the dog in dorsal recumbency, a right lateralisation of the sialocele was observed. For dogs #1 and #2, a lateral approach was used; however, the skin incision, instead of being horizontal and centred on the mandibular SG, was vertical and parallel to the vertical ramus of the mandible. The vertical incision allowed better visualisation of the pseudocapsule, which was separated from the surrounding tissues and was completely excised.In the literature, a clear recommendation regarding the complete excision of the pseudocapsule is lacking [11,52]. In the cases presented herein, the sialocele was large, severely displacing important vascular and nervous structures as well as the trachea, causing breathing discomfort. For this reason and to avoid leaving calcified material in place, the pseudocapsule was completely removed in all cases, and drains were inserted to avoid seroma formation due to the extensive tissue dissection.In all the dogs, the bacteriological analysis carried out on a sample of the pseudocapsule and its contents was negative, thus confirming the non-infectious origin of the lesion. Histological examination was performed in all cases, and it excluded a neoplastic process affecting the SG but, on the contrary, confirmed its inflammatory origin. Finally, the decision to perform an excisional biopsy of the tomographically or clinically enlarged lymph nodes can be a critical point; it was dictated by the fact that the unusual sialocele clinical presentation without the definitive histological diagnosis could not totally exclude a tumour.The lymph node excision did not require additional skin incisions. The excised lymph nodes were adjacent to the glandular structures removed; therefore, their dissection did not lead to lengthening the surgical time or to an increase in post-surgical complications [11]. Hyperplastic reactive lymphadenitis was the definitive expected histological diagnosis in this case series, reflecting the tissue condition caused by both the space-occupying mass effect and the local inflammation.The main limitations of this study were its retrospective nature and the small number of cases included, the latter reflecting the rarity of this condition. Another limitation was the different method of diagnostic imaging utilised as, for patient #2, a CT scan was not available.5. ConclusionsThe evidence of a huge and consolidated mass at the level of the neck should not immediately be considered highly malignant, and a complex sialocele should be considered in the differential diagnosis. Cytology, centesis and imaging are necessary to achieve the diagnosis and to plan the surgical approach. Ossified sialocele excision is associated with a favourable prognosis if the affected SGs are removed together with a part of or the entire pseudocapsule. Definitive diagnosis is provided by the histological examination.

animals : an open access journal from mdpi

10.3390/ani11123455

PMC8697938

The probiotic Bacillus amyloliquefaciens strain H57 (H57) may reinforce preferential feeding behaviour by changing ruminal fermentation parameters. Four rumen-fistulated steers were offered feedlot pellets, with (H57) or without (Control, C) the H57 probiotic. Half of the pellets were added to the rumen, at time zero, and half were offered for oral consumption over the next six hours, to make four feeding treatments. Each steer was offered each treatment over time. Each offering was over six days, with rumen fluid sampled over the last three days for a six-hour period per day. A five-minute preference test was performed at the end of each rumen sampling period by simultaneously offering the steers 4 kg of H57 and C pellets. The steers preferred the H57 over the C pellets but the route of offering (rumen versus oral) had no effect on preference. Ruminal pH and molar proportions of iso-butyrate and iso-valerate were higher and ammonia concentrations tended to be greater for H57 compared to C. However, since the route of offering had no effect on preference, the hypothesis, that ruminal fermentation changes take precedence over oral (taste) sensations in driving preference, was not supported.

This study examined whether the probiotic Bacillus amyloliquefaciens strain H57 (H57) affects ruminal fermentation parameters that exercise post-ingestive feedback appetite control mechanisms. A 4 × 4 Latin square design was used to separate pre- and post-ingestive effects of H57 in four rumen-fistulated steers. The steers were offered a set amount of feedlot pellets, inoculated with H57 or without H57 (control, C). Half of the total amount of pellets fed were introduced intra-ruminally (r), and then the remaining pellets were orally consumed (o) to make four feeding treatments: H57r/H57o, H57r/Co, Cr/H57o and Cr/Co. Rumen fluid was sampled at 2, 4 and 6 h after feeding. Preference behaviour was tested immediately after the 6 h rumen fluid sampling by simultaneously offering the steers 4 kg of each of H57 and C pellets in adjacent troughs for 5 min. Steers preferred the pellets with added H57 over the C pellets (56:44; p < 0.001) and their preferences were not affected by the treatment protocol imposed to separate post- from pre-ingestive effects (p > 0.05). Steers fed H57 pellets had higher ruminal pH, molar proportions of iso-butyrate and iso-valerate (p < 0.05) and tended to have greater ruminal ammonia concentrations compared to those fed C pellets (p < 0.1). However, post-ingestive signals did not affect diet preference more than pre-ingestive signals.

1. IntroductionReduced appetite and feed intake are common clinical signs of stress in cattle [1] and the amelioration thereof is increasingly a regulated requirement in the ruminant industries [2]. Probiotics can improve feed intake in ruminants, thus sustaining their performance and wellbeing during stressful situations as reviewed by [3,4,5,6]. Ramsing [7] and Yuan [8] suggested that supplementing fermented yeast culture obtained from Saccharomyces cerevisiae during the transition period increased the number of meals consumed by pre-partum cows, which ultimately led to an improvement in DMI. Likewise, the length of the daily meals and feeding times were greater for postpartum cows supplemented with fermented yeast culture compared with those that were not [7]. Diets containing the spores of Bacillus licheniformis and B. subtilis have stimulated the starter mixture intake of milk-fed calves [9]. Recently, H57 was shown to enhance the DMI of pregnant ewes and dairy calves fed pellets pre-inoculated with the H57 [10,11].H57 is a spore-forming bacterium that was isolated from lucerne leaves and developed as an inoculant in the hay-making process to reduce the risk of fungal spoilage [12]. When tested to indicate if ruminants would accept grass/clover hay containing H57, mature ewes preferred the H57-treated hay over the untreated hays [12]. Le [13] subsequently found the preference proportion of approximately 70:30 in weaned dairy calves for H57-inoculated calf pellets over un-inoculated pellets. This response was recorded within 6 h on the first day of offering the inoculated pellets after several days of ensuring that the calves had adjusted to the testing procedure. The 70:30 ratio remained the same for three consecutive days and on the fourth day, despite attempts to stimulate the calves’ preference for the un-inoculated pellets by adding a glucose sweetener, the preference for H57 pellets persisted. However, the mechanisms explaining the link between H57 supplementation, increased diet preference and intake of ruminants remain unclear.Diet preference and feed intake of ruminants are controlled by neurally mediated interactions between feed sensory characteristics and post-ingestive feedback [14]. Animals experienced sensory cues before the feed was swallowed and post-ingestive feedback, which includes digestive and metabolic signals experienced by the animals, after the feed was swallowed [15]. Favreau [16] and Baumont [17] highlighted how the sensory characteristics of a feed were used as a tool by ruminants to identify and choose between feeds. For example, sheep could use artificial flavours in feed as indicators to direct their preference and intake over relatively short-term periods [18]. In diet preference studies, post-ingestive feedback was considered predominant in the determination of feed intake in ruminants [14,15,19]. The post-ingestive feedback could include ruminal fermentation parameters such as pH, concentrations of VFAs, and its profile or NH3 concentrations [20]. Forbes [21] showed the connection between low ruminal pH and the regulation of diet preference as ruminal pH sent a feedback signal to the central nervous system to depress appetite and feed intake. Similarly, a high concentration of individual VFA, particularly propionate, can alter diet selection [21]. An elevation in the ruminal concentration of NH3 can also induce satiety and thus reduce feed intake [22,23].Despite the various aspects of feed intake regulation that have been elucidated in ruminants, we are not aware of any reports on how bacterial probiotics might regulate diet preference. There are indications that the addition of probiotics to diets can alter ruminal pH and concentrations of fermentation products (e.g., ammonia and total and individual VFAs) and that such changes may be associated with appetite regulation [10,11,24,25,26]. Thus, this study tested the hypothesis that H57 could improve diet preference in steers through changes in rumen fermentation parameters related to diet preference and feed intake regulation.2. Materials and Methods2.1. Animals and Experimental DesignFour mature Bos indicus cross, rumen cannulated, steers (mean liveweight (LW): 900 ± 37 kg, age 5 years old) were used. They were housed in individual pens (3 × 10 m) with rubberised flooring at the Queensland Animal Science Precinct, the University of Queensland (UQ) (Gatton, QLD, Australia). All experimental procedures involving animals were approved by the Animal Ethics Committee of the UQ (SVS/451/16).Throughout the experiment, a restricted amount of beef feedlot pellets inoculated with (H57) or without (Control, C) H57 were offered at 9:00 a.m. and at 3:15 p.m. Outside those times, chaffed Rhodes grass hay and water were provided ad libitum. The experiment was designed to dissociate the effects of pre-ingestive characteristics and post-ingestive feedback on diet preference in steers, based on a procedure similar to that used by Favreau [15]. The experiment was conducted over 64 days with four 16-day periods and five days washout between each period to minimise any carryover effects to the next period. The steers received either H57 or C pellets placed directly into the rumen (r), followed by either type of pellet presented in a trough so they could consume normally (orally, o). The four treatments were H57r/H57o, H57r/Co, Cr/H57o and Cr/Co. Treatments were allocated to the four steers in a complete Latin square design where each steer would receive each of the four treatments over the four treatment phases of 6 days with each treatment phase preceded by an adjustment phase of 10 days (Figure 1).On days 1–3 of each 10-day adjustment phase, about 1 kg of a 50:50 mix of H57 and C (500 g, as fed of each) was offered to familiarise the steers with both pellet types and reduce carryover effects between treatments. From day 4 to 10, for maintenance of LW, all steers were given a 50:50 mix of H57 and C pellet amounts equivalent to 0.35% DM of steer LW. For the subsequent 6-day treatment phase, the steers continued to be fed orally at the same rate of the pellet type they had been given in the previous 7-day (A4–A10) portion of the adjustment period. Additionally, at 8:00 a.m., either H57 or C pellets were introduced into the rumen via the fistula as per the experiment design at 50% of the weight of pellets that the steers consumed in the preceding A4 to A10 adjustment phase. These pellets were mixed 1:1 with artificial saliva (Kansas State Buffer) [27] immediately prior to introduction into the rumen. The remaining half of the total treatment pellet type was presented for oral consumption at 9:00 a.m. for 15 min and then chaffed Rhodes grass hay was provided ad libitum. Immediately after the 3:00 p.m. rumen sampling, a preference test was conducted to determine which type of pellet the steer preferred to assess the influence of the preceding treatment at 8:00 a.m. (feeding directly into rumen) and 9:00 a.m. (trough feeding). At the 3:15 p.m. test feeding, 4 kg (as fed) of each of the pellets was offered simultaneously in the two adjacent troughs, one containing H57 pellets and the other C pellets, and steers were allowed to eat freely from each for 5 min. Each day the positions of the troughs were reversed to prevent any bias. The relative preference for H57 pellets was calculated as (intake of H57 pellets)/ (intake of C plus H57 pellets) × 100 [28].2.2. Preparation of Beef Feedlot PelletsThe method used to prepare H57 inoculum for the cattle trial was as described by Schofield [29]. Briefly, the production of the H57 inoculum was performed in a series of batch cultures, then cultivated in a 100 L fermenter for the production of spores. The fermenter contents were centrifuged at 15,000 rpm to harvest the bacteria and suspended material from the fermenter. The paste from centrifuging the fermenter contents was mixed with sodium bentonite and freeze-dried. The product was then ground into a powder for subsequent use. The H57 bentonite powder (6.4 × 1010 cfu spores/g) was provided to inoculate 3 tons of feed at 1 × 106 cfu spores/g pellet. The bentonite powder was mixed with ground wheat in a feed mixer, and this diluted inoculum was incorporated into a final premix for steam pelleting in the Ridley Agriproducts Pty Ltd. feedmill at Toowoomba, QLD, Australia. The pellets were cooled and stored in 20 kg plastic bags. The bagged pellets were stored at 7 °C for 3 months and sufficient bags were removed to ambient conditions at the beginning of each period for feeding. The ingredients and chemical composition of the pellets are presented in Table 1.2.3. Sample Collection and MeasurementsApproximately 0.5 kg subsamples of feed offered to steers and refusals were collected daily and stored in a cool room at 7 °C. At the end of the experiment, these daily samples were pooled, mixed completely, then about 0.5 kg of hay and 1 kg of pellet subsamples were stored at −20 °C for chemical analysis. Samples of 100 g for the plate counts of the H57 populations were stored at 4 °C.Rumen contents were sampled at 8:00 a.m. (0 h) before introducing the pellets into the rumen and then at 11:00 a.m. (2 h), 1:00 p.m. (4 h) and 3:00 p.m. (6 h) on the last 3 days of the experimental periods. Rumen samples were collected through the fistula by first sliding into the rumen a PVC tube (40 mm ID; 60 cm length; Holman Industries HQ, Osborne Park, WA, Australia), with nylon stockings (Bonds, Melbourne, VIC, Australia) glued over the tube end inserted into the rumen to strain the rumen fluid. Then a smaller tube (10 mm ID; 80 cm length; Holman Industries HQ, Osborne Park, WA, Australia) was inserted inside the larger tube, which was attached to a 60 mL syringe (Terumo, Somerset, NJ, USA) that was used to withdraw about 40 mL of the rumen fluid. The rumen fluid was then placed into sterile 250 mL plastic containers (Thermo Fisher Scientific, Waltham, MA, USA). The ruminal fluid pH was measured within minutes of collection using a portable pH meter (Eutech pH 6+, Eutech Instruments, Ayer Rajah Crescent, Singapore). Two subsamples were transferred into 10 mL tubes (Thermo Fisher Scientific, Waltham, MA, USA) and placed in an ice bath. A 4 mL sample of the rumen fluid was preserved by adding 1 mL of 20% metaphosphoric acid for the analysis of total VFA and its profile. A separate 8 mL rumen filtrate was preserved by adding 2 mL of 0.5 M sulphuric acid for analysis of ammonia content. These rumen subsamples were stored at −20 °C for subsequent analysis.2.4. Cell Counting Procedure for H57 Spore and VegetativeThe viable count method of Harrigan and McCance [30] was modified and used to count H57 concentrations (spore and vegetative cells) in the pellet samples. The procedure is described in detail in Ridley-ARC [31]. Briefly, 1.0 g of powdered material was weighed into a sterilized 200 mL beaker and 100 mL sterilized chilled water added. The suspension was then mixed for 2 min at 24,000 rpm using a T25 digital Ultra-Turrax IKA homogenizer with a 25 mm dispersing tool (IKA, Staufen, Germany). Three independent 0.1 mL aliquots were taken from the feed suspension and mixed with 0.9 mL sterile water in sterile 1.5 mL Eppendorf tubes. The tubes were labelled as A, B and C samples. To count total cells, 0.1 mL aliquots (one replicate each from A, B and C tubes) was spread on nutrient agar and labelled appropriately. The Eppendorf tubes were then heated for 20 min at 80 °C in a heating block (1572VWR, VWRTM, Radnor, PA, USA) to kill vegetative cells, and a repeat 0.1 mL aliquot spread onto nutrient agar to count spores. Cells were grown overnight at 28–30 °C and the colonies on the plate were tabulated. Vegetative cells were determined as total cells minus spores. The C pellets had no H57 while the H57 pellets had cell counts that ranged from 0.6 to 0.9 × 106 cfu spores/g pellets).2.5. Chemical AnalysisThe feed samples were analysed for nutritive contents by Dairy One Forage Laboratory (Ithaca, NY, USA). All samples were analysed using the wet chemistry procedures for dry matter (DM) (method 930.15), crude protein (CP) (method 984.13), fat (method 920.39), minerals (method 985.01), starch (method 996.11) [32], acid detergent fibre (ADF) and neutral detergent fibre (NDF) [33].The frozen rumen fluid samples were thawed completely and then centrifuged at 2000× g for 20 min (Eppendorf® Minispin®, Hamburg, Germany) to separate the liquid phase. The clear supernatants were collected for VFA analysis. Separation of VFAs was determined by the gas-liquid chromatography method (Cottyn and Boucque, 1968) and performed on a GC-2010 plus (Shimadzu, Kyoto, Japan) using a polar capillary column (ZB-FFAP 30 m × 0.53 mm × 1.0 µm, Zebron®, Phenomenex, Inc., Torrance, CA, USA). The temperature program of the GC was set initially at 85 °C for four minutes then increased to 200 °C at a rate of 15 °C/min. The flow rate of carrier gas (high purity helium) was at 5.0 mL/min, 67 kPa for two minutes then from 1.8 kPa/min to 81 kPa. Individual VFAs were quantified at 210 °C by a Chromatopac C-R6A (Shimadzu, Kyoto, Japan) equipped with a flame-ionisation detector. Ammonia concentrations were determined by the distillation method [32] using the Buchi 321 distillation unit (Flawil, St Gallen, Switzerland) and then titrated against 0.01M HCl using the automatic titration unit (TitraLab 845, Hach, Loveland, CO, USA). The operating procedure followed the guidelines recommended by the manufacturers.2.6. Statistical AnalysisData were analysed using the mixed model [34] of the statistical software package SAS, Version 9.4 (AS Institute Inc., Cary, NC, USA) [35]. The results were presented as least-squares mean ± s.e.m. The differences between treatment effects were assessed by Tukey multiple comparisons. The differences were considered significant at p < 0.05 and statistical tendency was declared at p < 0.10.The following statistical models were adopted:Yijk = μ + Treatmenti + Steerj + Periodk + Treatmenti Periodk + eijk Yijkl = μ + Orali + Rumenj + Steerk + Periodl + Orali Rumenj + Orali Rumenj Periodl + eijkl where Yijkl is the variable response; µ, overall mean; Treatmenti (i = 1 to 4), fixed effect of treatment (H57r/H57o vs. H57r/Co vs. Cr/ H57o vs. Cr/Co); Steerj,k (j,k = 1 to 4), random effect of the individual steer within the square; Periodk,l (k,l = 1 to 4), fixed effect of the period within the square; Treatmenti Periodk, fixed effect of the interaction between treatment and period; Orali (i= 1 to 2), fixed effect of orally consuming H57 (H57r/H57o and H57r/H57o) vs. C pellets (H57r/Co and Cr/Co); Rumenj (j = 1 to 2), fixed effect of intra-ruminally consuming H57 (H57r/H57o and H57r/Co) vs. C pellets (Cr/H57o and Cr/Co); Orali Rumenj, fixed effect of the interaction between orally and intra-ruminally consuming pellet treatments; Orali Rumenj Periodk, fixed effect of the interaction among orally and intra-ruminally consuming pellet treatments and period; eijk and eijkl, are residual errors. All interactions were systematically removed from the model when they were non-significant, and a reduced model was used to determine treatment effects.A one-sample t-test was used to compare actual preference data, with 0.5 as the reference value. Data for ruminal pH, NH3 and total and individual VFA levels were analysed by sampling time using a repeated measure with compound symmetry.3. Results3.1. Daily Feed Intake and Short-Term Preference TestsThe route of pellets addition into the digestive system did not affect pellet DMI in a five-minute preference test (p = 0.87; Table 2). Over 24 h, the total pellet intake of the steers was similar between treatments (p = 0.90). Likewise, differences in total hay intake due to treatments were not detected (p = 0.52).Preference was not influenced by whether the pellet type had been previously introduced orally or intra-ruminally (p > 0.05; Figure 2). However, when considered separately from the morning feeding treatments, the steers’ preference for H57-inoculated pellets during the short-term preference test was consistent across treatments (p < 0.05). When the data were combined across all pellet treatments, the average preference of animals for the H57 pellets compared with their preference for the C pellets was approximately 56:44 (p < 0.001).3.2. Ruminal pHAcross the treatments, there was no difference in the pattern of ruminal pH during the 6 h period after feeding the pellets (p > 0.05; Figure 3). Ruminal pH declined to reach a minimum at 4 h post-feeding and began to recover towards the 6 h point but did not reach pre-feeding values (0 h) by that time. When the pellets were introduced directly into the rumen via the rumen fistula, ruminal pH at 4 h was higher for the H57 compared with the C pellets (6.20 vs. 6.12; p = 0.04) with a trend towards higher rumen pH appearing by 2 h (6.38 vs. 6.32; p = 0.08) and extending to 6 h (6.35 vs. 6.30; p = 0.06).3.3. Total Concentration and Molar Proportion of Individual VFAs in the RumenThere were no differences detected in the total VFA concentrations across the pellet treatments (p > 0.05; Figure 4). The total VFA concentrations increased to achieve the highest level at 4 h after the morning feeding and then reduced towards the end of the 6 h time point.The molar proportions of individual rumen VFAs including acetate, propionate, n-butyrate and n-valerate were generally similar among the pellet treatments (p > 0.05; Figure 5A–D). By contrast, at the 6 h time point, the molar proportions of iso-butyrate and iso-valerate were affected by the treatment (p < 0.05; Figure 5E,F). All treatments in which H57 was consumed orally or introduced directly into the rumen (H57r/H57o, Cr/H57o, H57r/Co), when averaged, had approximately 19% and 30% higher molar proportions of iso-butyrate and iso-valerate, respectively, than the Cr/Co treatment.3.4. Ruminal AmmoniaNo treatment effect was detected for ruminal NH3 concentration at 6 h (p > 0.05; Figure 6). However, at 2 h and 4 h, the NH3 concentrations for the treatment in which H57 pellets were consumed orally and introduced directly into the rumen (H57r/H57o) were 217.6 and 119 mg/L respectively, which tended to be greater than the remaining three treatments (p = 0.07 and p = 0.09, respectively).4. DiscussionImprovements in post-ingestive signals, such as changes in ruminal pH, NH3 and VFA concentration and profiles consequential to the inoculation of feed with H57, were expected to drive the preference decisions of the steers. However, no such effect was detected. Even though the steers consistently preferred the H57-inoculated over the un-inoculated pellets, and ruminal pH, NH3 and iso-acids were elevated in the short term due to the H57, these responses were unrelated to the treatment protocol, which was imposed to separate post- from pre-ingestive (i.e., taste and odour) effects. Interpretations could be that the post-ingestive signals were insufficiently extreme to initiate a preference response, or that H57 improves feed preference through a combination of pre- and post-ingestive signals.The protocol—imposed to separate the pre- and post-ingestive effects of a feed—was similar to that of Favreau [15] but with some differences. They used sheep, whereas this study used steers. They compared hays, whereas this study compared pellets. After a feed was given intra-ruminally (r) at 0 h, they offered hay ad libitum at trough (o) over several hours within a 6 h period, whereas this study offered a set amount of a given pellet after the ruminal introduction that was consumed within minutes. It was unable to offer the pellet ad libitum in the current study, as it would have put the steers at risk of ruminal acidosis and laminitis. Consequently, this study was unable to measure the rate of intake in the 6 h period. However, as did Favreau [15], this study applied one or the other feeds intra-ruminally at 0 h followed by one or the other feeds offered at trough and a five-minute preference test that was expected to be affected by the preceding four feeding treatment combinations. The logic was that there was a post-ingestive advantage afforded by H57 if ruminal parameters and preference for H57 (in the five-minute preference test) increased after H57r/H57o vs. Cr/H57o or H57r/Co vs. Cr/Co, whereas a pre-ingestive effect would have been determined if ruminal parameters and preference for H57 increased after Cr/H57o vs. H57r/Co. Since none of these comparisons were different, it is concluded that there was no evidence that H57 caused either a pre- or post-ingestive advantage.On the five minute preference test, steers consistently preferred H57-inoculated over the un-inoculated pellets. This was not influenced by whether the pellets were introduced orally, indicating no short-term (i.e., within the 6 h testing period) overriding effect of pre- or post-ingestive feedback on diet preference. The preference for H57 pellets observed in this study was consistent with Le [13], who found that the weaned dairy calves demonstrated a 70% preference for H57 pellets. A lower preference for H57 pellets recorded in the current study compared with that of Le [13] could be explained by the differences in the ages of the animals and methods to assess diet preference. While Le [13] used weaned dairy calves at 20 weeks of age, this current study used mature steers that were 5 years of age. Additionally, Le [13] conducted a pellet preference test that lasted for 6 h, whereas the current study assessed the pellet preference of steers for only 5 min. Miller-Cushon [36] suggested that diet preference tests conducted or repeated over a long period might induce results that were different to that of short-term preference studies. Le [13] also concluded that the sweet taste was not a driver of preference for H57 pellets. It is speculated H57 enhanced diet preference may be related to the ability to prevent feed degradation. H57 was developed as an inoculant to protect stored forages and grain-based pellets against spoilage by fungi and other feed-spoiling organisms [12,37]. Further, Schofield [38] reported that H57 had the capacity to produce several kinds of lipopeptides and polyketide compounds, which are biological control agents and thought to inhibit fungal and bacterial growth. If the H57 spores germinate and grow in the inoculated pellets during storage, then the vegetative cells may subsequently produce antimicrobial compounds that are effective at controlling microbial contamination. This may preserve the freshness of H57-inoculated pellets, thereby influencing diet preference.Ruminal pH was observed to be higher for the H57 than the C pellets at 4 h. Provenza [14] stated that the preference for a feed was adjusted according to its post-ingestive consequences, and on that basis, animals remembered to select or avoid that feed. This feedback could include ruminal pH, which is highly related to intake rate [39]. In the current study, although the steers that received H57-inoculated pellets had higher ruminal pH, these pH values were maintained above the critical rumen pH for fibre digestion—pH 6.0 [40]. It is speculated that an increase in ruminal pH due to H57 supplementation was possibly not sufficient to alter diet preference, and thus alteration of ruminal pH may not be an over-riding mechanism by which H57 improves preference. The advantageous effects of a probiotic supplement on ruminal pH have been shown in the studies of Nocek [41], Marden [24] and Bruno [42]. In the present study, although the rise in ruminal pH was much smaller than expected, the findings were consistent with the hypothesised ability of H57 to play a potential role in the stabilisation of rumen pH. This was supported by Le [10], who found that ruminal pH was higher for the ewes receiving H57 inoculated pellets (2.9 × 109 cfu/kg pellets), than that of those who were fed with un-inoculated pellets. In contrast, Sun [43] found ruminal pH of dairy cows decreased during the supplementation of B. subtilis natto (0.5 × 1011 cfu/cow/day or at 1.0 × 1011 cfu /cow/day). The observation that Bacillus probiotic supplementation had no effect on ruminal pH was recorded by Peng [44]. These authors indicated that there was a similarity in the ruminal pH between the control cows and those supplemented with a B. subtilis natto fermentation product containing 8.3 × 109 spores/g. The inconsistent effects of Bacillus probiotics on ruminal pH could be due to different strains of Bacillus and doses of probiotics used.H57 supplementation increased concentrations of iso-butyrate and iso-valerate in the rumen at the 6 h time point. However, this change may not directly explain the steers’ preference for H57 during the short-term preference test. Iso-acids can regulate diet preference and DM intake of ruminants [45] because of their ability to modify the ruminal fermentation by increasing ruminal concentrations of acetate and the ratio of acetate to propionate [46,47], or enhancing the growth of important cellulolytic strains of ruminal bacteria such as Butyrivibrio fibrisolvens, Ruminococcus albus, Ruminococcus flavefaciens and Fibrobacter succinogenes [48,49]. Another possible effect of iso-acids on feeding behaviour could be increased blood levels of growth hormone [47]. Although the advantages of increased concentrations of iso-acids on diet preference and DM intake were not observed in the current study, their roles in controlling feeding behaviour should be explored in further research. The increased iso-acid concentrations were in accordance with the higher levels of ruminal NH3 to be found in H57-fed steers. El-Shazly [50] suggested that an increase in the ruminal concentration of iso-acids was the result of the enhanced rate of protein degradation and amino acid metabolism by rumen microbes. Adding bacterial probiotics to the diet would lead to an increased iso-acids concentration, as previously recorded by Chiquette [51]. Iso-butyrate and iso-valerate productions in the rumen were increased by the probiotic Prevotella bryantii 25A supplemented in the dairy cow diet at the rate of 2 × 1011 cfu/animal/day.Ammonia concentrations tended to be higher in ruminal samples collected in steers receiving H57 inoculated pellets orally and intra-ruminally at 2 h and 4 h. Faverdin [22] suggested that changes in the NH3 concentrations of the rumen could influence feeding behaviour in ruminants. While excessive amounts of NH3 can temporarily decrease preference through the negative actions of NH3 [22], a rise in ruminal NH3 can increase preference through their contributions to the nitrogen metabolism of the rumen microbes [52]. In this current study, a higher ruminal NH3 concentration could not be a cue driving the preference responses of steers for H57 pellets. This might be because the increased NH3 concentrations were within the range generally required for fermentation and microbial protein synthesis in the rumen [45,53], and so were not high enough to drive a change in preference. The response of ruminal NH3 concentration to a Bacillus probiotic addition varied considerably in published studies, including no change [44], increased concentrations [43] and decreased concentrations [10,54]. Although not measured directly, a higher ruminal NH3 concentration due to the H57 supplementation recorded in this study may be associated with higher proteolytic and deamination activities by ruminal bacteria. Schofield [29] found that feeding H57 pellets to pregnant ewes caused an increase in Prevotella spp. populations, which are generally major contributors to protein and peptide degradation in the rumen [55].5. ConclusionsH57 increased the mature steers’ preference for H57-inoculated pellets. While it does appear that H57 supplementation can elevate ruminal pH, iso-acids and tend to increase NH3 concentrations, these proposed post-ingestive feedback signals were not directly linked to the steers’ preference for H57 pellets. The potential for increased ruminal iso-acids due to the H57 modulating feeding behaviour should be assessed through further experimentation. Determination of how feedback from the changes in the gut microbiota by H57 alters diet preference—potentially via microbial excretory products that can enter intermediary metabolic pathways—is required to provide more insight into the specific actions of H57 on dietary preference in ruminants.

animals : an open access journal from mdpi

10.3390/ani11040960

PMC8065711

The aim of this study was to determine the effect of partial and total replacement of protein from genetically modified soybean meal (GM-SBM) with protein from 00-rapeseed meal (00-RSM), alone or in combination with protein from low-tannin faba bean seeds (FB) or low-alkaloid yellow lupine seeds (YL) in grower-finisher diets on nutrient digestibility, nitrogen retention, and utilization, selected blood biochemical parameters and fattening performance of pigs. During two-phase fattening, hybrid Danbred growing-finishing pigs were fed grower diets where 50% of GM-SBM protein was replaced with 00-RSM protein, 00-RSM and FB protein or 00-RSM and YL protein, and finisher diets where 100% of GM-SBM protein was totally replaced with 00-RSM protein, and with 50% 00-RSM and FB protein or YL protein. It was found that GM-SBM protein can be partially (50% in grower diets) and totally (100% in finisher diets) replaced with 00-RSM protein (6%) combined with protein from low-tannin FB seeds (10%/12%) or low-alkaloid YL seeds (6%/7%) in pig diets. The evaluated diets contributed to high nutrient digestibility and N retention and improved fattening performance without compromising the health status of pigs.

The aim of this study was to determine the effect of partial and total replacement of protein from genetically modified soybean meal (GM-SBM) with protein from 00-rapeseed meal (00-RSM), alone or in combination with protein from low-tannin faba bean (Vicia faba L.) seeds (FB) or low-alkaloid yellow lupine (Lupinus luteus L.) seeds (YL) in grower-finisher diets on nutrient digestibility, nitrogen retention and utilization, selected blood biochemical parameters, fattening performance of pigs and carcass quality traits. Two digestibility-balance trials and one feeding trial were performed during two-phase fattening on male hybrid Danbred growing-finishing pigs were divided into four groups. The pigs were fed grower diets where 50% of GM-SBM protein (diet S-c) was replaced with 00-RSM protein (diet R), 00-RSM and FB protein (diet R + FB) or 00-RSM and YL protein (diet R + YL), and finisher diets where 100% of GM-SBM protein (diet S-c) was totally replaced with 00-RSM protein (diet R), and with 00-RSM and FB protein (diet R + FB) or YL protein (diet R + YL) in 50%. It was found that partial (50% in grower diets) and total (100% in finisher diets) replacement of GM-SBM protein with 00-RSM protein combined with FB or YL protein had no adverse effect on nutrient and energy digestibility, N balance, serum of blood carbohydrate and protein metabolism or the biochemical parameters of liver and kidney function. Protein from 00-RSM (6%) and FB seeds (10%/12%) contributed to high daily gains and high feed conversion efficiency. Protein from 00-RSM (6%) and YL seeds (6%/7%) in grower-finisher diets led to a further improvement in fattening performance. The analyzed vegetable protein sources had no negative influence on carcass quality. The results of the present study indicate that 00-RSM protein combined with protein from low-tannin FB or low-alkaloid YL seeds can be valuable high-protein feed ingredients alternative to GM-SBM in pig nutrition.

1. IntroductionModern pig fattening is based on complete diets with the nutritional value corresponding to the growth rate and protein deposition potential of animals. Pigs have high protein requirements; therefore, cereal-based diets must also contain high-protein components such as meals, including imported genetically modified soybean meal (GM-SBM) [1,2]. Recent years have witnessed considerable breeding progress in the genetic improvement of legumes. The seeds of new legume varieties are characterized by higher and more stable yields and a lower content of antinutritional factors [3]. Therefore, they can be fed to pigs to make up for the negative trade balance on high-protein ingredients used in monogastric nutrition.There has been an ongoing debate in Poland and in other EU countries over eliminating GM crops, in particular GM-SBM, from pig feedstuffs [4,5,6,7]. Due to the social pressure exerted on the Polish government’s policies, restrictions on the production and use of genetically modified organisms (GMOs) and trade in GMOs may be introduced on 1 January 2023 in Poland, including a ban on GM components in animal feeds [8]. Moreover, GMOs are prohibited in organic animal production, which is becoming increasingly popular. High fluctuations in the prices and supply of GM-SBM and the need to provide affordable protein for animal feeds have triggered a search for alternative, cheaper local sources of vegetable protein [4,9,10,11]. In Poland, 00-rapeseed meal (00-RSM), faba bean (Vicia faba L.) seeds of new low-tannin varieties (FB), and yellow lupine (Lupinus luteus L.) seeds of new low-alkaloid varieties (YL) can be valuable protein sources in pig diets.The results of previous experiments, though inconclusive, show that legume seeds should not be the sole protein source in pig diets [12,13]. Protein from FB and YL seeds has high lysine content and low concentrations of methionine and tryptophan. The ratio of lysine to methionine+cystine in these feedstuffs is 1:0.3–0.6 [14], whereas the ratio of the above amino acids in grower pig diets should be 1:0.65–0.70 [14]. Therefore, grain legumes cannot be the only source of supplemental protein in cereal-based pig diets. Protein from 00-RSM is characterized by high concentrations of methionine+cystine (2.1 and 2.5 g/16 g N, respectively) and a ratio of lysine to methionine+cystine of 1:0.83. It also has higher tryptophan content (1.3 g/16 g/N) than FB and YL protein. Therefore, 00-RSM protein can supplement FB and YL protein in diets for growing-finishing pigs.The aim of this study was to determine the effect of partial and total replacement of GM-SBM protein with 00-RSM protein combined with protein from low-tannin FB seeds or low-alkaloid YL seeds in grower-finisher diets on nutrient digestibility, nitrogen retention and utilization, selected blood biochemical parameters, fattening performance of pigs and carcass quality traits.2. Materials and MethodsThe animal protocol and the number of animals used in this study were consistent with regulations of the Local Institutional Animal Care and Use Committee (23/2013 Olsztyn, Poland), and the study was carried out in accordance with EU Directive 2010/63/EU on the protection of animals used for scientific purposes [15].2.1. High-Protein Vegetable Feed IngredientsThe chemical and amino acid composition of protein and the content of antinutritional factors in GM-SBM, 00-RSM, seeds of FB cv. Albus, and YL cv. Taper are presented in Table 1.Chemical analysis of protein quality in the analyzed high-protein vegetable feedstuffs was performed (Table 1), and the results were expressed as the essential amino acid index (EAAI) calculated using the method proposed by Oser [17] with chicken egg protein as the reference standard, and the EAAIp (essential amino acid index—ideal amino acid profile of protein for porkers) based on the concept of ideal protein for pigs proposed by Muller in 1999, as cited by Grela [18].2.2. Animals, Diets, and Experimental Procedures in Digestibility-Balance TrialsDigestibility-balance trials (experiments IA and IB) were performed using a simple balance method at the Animal Research Laboratory of the Department of Animal Nutrition and Feed Science, University of Warmia and Mazury in Olsztyn. A five-day experimental period proper was preceded by a seven-day pre-experimental (adaptation) period. The experimental materials comprised 20 male hybrid Danbred growing-finishing pigs divided into four groups of five animals for each experiment, IA (grower diets) and IB (finisher diets). The pigs were allocated to four groups using an analog method, according to the experimental design presented in Table 2. At the beginning of experiments IA (grower diets) and IB (finisher diets), the average body weight of pigs was 50 kg and 75 kg, respectively. The animals were housed in individual metabolism pens, with free access to water, and were fed crumbled feed that was offered wet (feed/water ratio of 1:1). The animals were fed complete grower (Table 3) and finisher diets (Table 3) in the grower and finisher phases of fattening, respectively. The composition of diets in digestibility-balance trials was identical to that in the feeding trial (experiment II). Daily feed allowance was determined based on the Feeding Guidelines for Pigs [15] at a target daily gain of 850 g.During the five-day experimental period of digestibility-balance trials (experiments IA and IB), feces and urine were collected quantitatively. Two samples of 5% each were collected from feces produced during 24 h. One sample was dried, and the other sample was frozen. The content of dry matter, crude ash, crude fat, crude fiber, and gross energy was determined in the average dried fecal sample. Nitrogen content was determined in the average frozen fecal sample. Urine was preserved with 20% sulfuric acid to maintain pH below 2.0, and 5% samples were collected to determine N content.The coefficients of apparent digestibility of dietary protein, fat, fiber, N-free extracts, and energy were calculated based on the chemical composition of feces and diets, and nutrient intake and output, with the use of the following equation: DC (%) = NI-NO/NI × 100; where: DC—digestibility coefficient (%), NI—nutrient intake (g), NO—fecal nutrient output (g), and NO/NI—nutrient digestibility (g).Nitrogen retention was determined based on dietary N intake and fecal and urinary N excretion for each fattening phase and different vegetable protein sources of the diets. Nitrogen utilization was calculated based on the apparent biological value of protein expressed as N retention relative to N digested.2.3. Animals, Diets, and Experimental Procedures in the Feeding TrialA feeding trial (experiment II) was conducted at the Animal Research Laboratory of the Department of Animal Nutrition and Feed Science, University of Warmia and Mazury in Olsztyn. The experimental materials comprised 28 male hybrid Danbred growing-finishing pigs. The animals were housed in individual pens with a surface area of 2.6 m2 (length 1.7 m × width 0.95 m × height 1 m) with a slatted floor equipped with nipple drinkers. The pigs were allocated to four groups using an analog method, according to the experimental design presented in Table 2. The initial body weight of pigs was 26 kg, and their final body weight was 104 kg. Fattening was divided into two phases (grower: 26–67 kg BW and finisher: 67–104 kg BW). The animals were fed complete grower (Table 3) and finisher diets (Table 3) in the grower and finisher phases of fattening, respectively. Pigs were fed crumbled feed that was offered wet (feed/water ratio of 1:1), and had free access to water. The composition and nutritional value of grower and finisher diets are presented in Table 3. The nutritional value of diets was determined based on the Feeding Guidelines for Pigs [14].Fattening performance was expressed as daily gains in the grower phase (26–67 kg BW), finisher phase (67–104 kg BW), and the entire fattening period (26–104 kg BW). The feed conversion ratio (FCR) was calculated based on feed, metabolizable energy, and digestible protein intake per kg of body weight gain in both fattening phases.At the completion of the feeding trial, carcass quality was evaluated. All 28 pigs subjected to the feeding trial were slaughtered at a body weight of around 104 kg in the “Warmia” Meat Processing Plant in Biskupiec. Immediately after slaughter, carcass dressing percentage was determined, and carcass conformation and fat cover were evaluated based on the EUROP system criteria, with the use of a CGM 100 ultrasonic device equipped with an optical probe. Measurements were performed at the level of the last thoracic vertebra, 7 cm from the dorsal midline.2.4. Selected Serum Biochemical ParametersBlood was sampled from five pigs per group to determine indicators of selected biochemical processes in pigs fed diets with different vegetable protein sources. Blood samples for analyses were collected from the anterior vena cava of live animals before the morning feeding at the end of the experimental period proper in both digestibility-balance trials (experiments IA and IB). Whole blood samples were collected into test tubes containing a chemically neutral clotting activator. Blood samples were allowed to clot at room temperature for 30 min, and they were centrifuged in the MPW-348 centrifuge at 2000 rpm for 10 min. The serum samples for testing were stored at −20°C. Carbohydrate metabolism was analyzed based on serum glucose concentrations. Protein metabolism was analyzed based on serum total protein and ammonium N levels. Liver function was evaluated based on the activities of aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyltransferase (GGT). Kidney function was evaluated based on serum creatinine concentrations.2.5. Laboratory AnalysesThe content of major nutrients in feedstuffs, diets and fecal samples, and urinary N excretion were determined by standard methods [19].Gross energy concentrations in feedstuffs, diets, and fecal samples were determined by adiabatic bomb calorimetry (IKA® C2000 basic, Staufen, Germany). Metabolizable energy concentrations in diets were determined based on digestible nutrients content, using the Rostock Feed Evaluation System (RFES) formula [14].The amino acid composition of protein and amino acid concentrations in the analyzed high-protein vegetable feed ingredients were determined with the use of an Amino Acid Analyzer AAA 400. The samples were hydrolyzed with 6M HCL at a temperature of 110 °C for 24 h. The concentrations of sulfur-containing amino acids were determined after sample oxidation with performic acid. Tryptophan content was determined in accordance with the Polish Standard [20].The content of crude fiber, including neutral detergent fiber (NDF), acid detergent fiber (ADF), and acid detergent lignin (ADL) in feedstuffs, was determined by the method proposed by Van Soest et al. [21] using the FOSS TECATOR Fibertec 2010 System.The concentrations of oligosaccharides in the seeds of YL cv. Taper and FB cv. Albus after extraction were determined by high-performance liquid chromatography (HPLC) using a refractive index RID-10A detector, Phenomenex Luna NH2 column, a mobile phase of 65% acetonitrile, a flow rate of 1.2 mL/min, and an injection volume of 20 μL.The concentrations of glucosinolates in 00-RSM samples were determined in the Shimadzu HPLC system using a Varian MetaCarb 67H column.The tannin content of FB seeds was determined by colorimetry [22], and the content of trypsin inhibitors in GM-SBM was determined by spectrophotometry [23].Serum glucose was estimated by the glucose oxidase method, serum ammonium N levels were determined using the kinetic method, total serum protein was analyzed using the biuret method, serum creatinine concentrations were determined by the Jaffe method, the activities of AST, ALT, and GGT were evaluated using the kinetic method using an ACCENT-200 automatic biochemistry analyzer and commercial Cormay kits. The results are expressed in SI units [24].2.6. Statistical AnalysisThe results of digestibility-balance and feeding trials (arithmetic means) were processed statistically by one-way analysis of variance (ANOVA) using licensed STATISTICA 12.0 software. The significance of differences between group means was estimated by Duncan’s multiple range test at significance levels of p ≤ 0.05 and p ≤ 0.01 [25].3. ResultsThe chemical composition of the analyzed high-protein vegetable feedstuffs is presented in Table 1. Similar values were reported by [12,26,27,28]. Protein from FB seeds cv. Albus had the highest lysine content (5.87 g/16 g N). Lysine concentrations were lower in protein from GM-SBM, 00-RSM, and YL seeds cv. Taper (5.20, 4.68, and 5.33 g/16 g N, respectively). Protein from legumes seeds had a low content of sulfur-containing amino acids (1.34 to 2.09 g/16 g N) and tryptophan (0.75 to 0.88 g/16 g N). The concentration of methionine+cystine was highest (4.28 g/16 g N) in 00-RSM.The EAAI of GM-SBM and 00-RSM protein was 71.0 and 70.8, respectively. The values of the EAAI were considerably lower in legume seeds (YL—65.1 and FB—68.2). The nutritional value of protein determined based on the EAAIp was higher in 00-RSM and (89.7) and GM-SBM (87.3), and lower in low-tannin FB seeds (81.7) and low-alkaloid YL (75.1).The data in Table 3 shows that complete grower and finisher diets were characterized by high nutritional value, consistent with Pig Nutrient Requirements [15]. Total protein concentrations ranged from 170 to 171.7 g/kg in grower diets and from 148 to 149.3 g/kg in finisher diets. The analyzed vegetable protein sources had a minor effect on the digestible protein content of diets due to differences in protein digestibility. Digestible protein concentrations were highest in grower controls (S-c) (150 g/kg) and R + YL (150 g/kg) diets and lowest in diet R (144 g/kg). A similar trend was noted in finisher diets, where digestible protein concentration was highest in diet S-c (129 g/kg), followed by diet R + YL (128 g/kg) and diet R (125 g/kg). The experimental factor had no significant effect on protein quality evaluated based on the concentrations of essential amino acids (lysine, methionine+cystine, threonine, and tryptophan). The tested vegetable protein sources induced minor differences in metabolizable energy concentrations, which were high in both grower (12.75 to 12.86 MJ/kg) and finisher (12.72 to 12.85 MJ/kg) diets.Grower diets where 50% of GM-SBM protein was replaced with 00-RSM protein significantly (p ≤ 0.05) decreased the digestibility of total protein and energy (Table 4), which was determined at 87.8% vs. 84.5% and 86.4% vs. 84.9%, in diet S-c vs. diet R, respectively.The replacement of GM-SBM protein with 00-RSM protein combined with protein from low-tannin FB seeds (diet R + FB) or low-alkaloid YL seeds (diet R + YL) had no significant negative effect on protein or energy digestibility compared to the S-c group. The obtained values of these indicators were high and amounted to 87.3% and 87.4% for protein, while for energy, they ranged from 86.3% to 86.8%.Crude fiber digestibility was significantly (p ≤ 0.05) higher in diets R + FB and R + YL than in diet R (46.0% and 47.1% vs. 43.3%, respectively).Daily N balance data for pigs fed grower diets is presented in Table 4. Fecal N excretion was significantly (p ≤ 0.05) higher in pigs fed diet R (11.0 g/day) than in the animals receiving the control diet S-c (8.6 g/day) and experimental diets R + FB and R + YL (9.0 g/day each). Urinary N excretion tended to be higher in the control group S-c and experimental groups R + FB and R + YL than in group R (27.5, 28.3, and 27.1 vs. 25.9 g/day, respectively). Fecal and urinary N losses had no negative effect on N retention or utilization. Nitrogen retention was high in all groups, ranging from 33.5 to 35.0 g/day (p = 0.104). Nitrogen retention as a percentage of N intake (47.2% to 49.2%, p = 0.182) and N digested (54.4% to 56.4%, p = 0.231) was also high in all groups. However, it should be stressed that N retention (35.0 g/day) and utilization (49.2% relative to N intake and 56.4% relative to N digested) were highest in growing pigs fed protein from 00-RSM and YL seeds (diet R + YL).Table 5 present the digestibility of nutrients from finisher pig diets where 100% of GM-SBM protein was replaced with 00-RSM protein, alone or in combination with protein from FB and YL seeds. No significant differences in total protein digestibility were found between groups (p = 0.694), but lower values were noted in groups R (84.4%) and R + FB (84.9%) than in groups R + YL (86.0%), and S-c (86.8%). The experimental factor had no significant positive effect on the digestibility of crude fiber, crude fat, N-free extractives, or energy.Daily N balance in pigs fed finisher diets is presented in Table 5. Fecal N excretion was highest in pigs fed diet R + FB (11.1 g/day), and this value was significantly (p ≤ 0.05) higher than in the animals receiving diets R + YL (9.7 g/day) and S-c (9.2 g/day). Urinary N excretion was significantly lower in pigs fed diets R (00-RSM), R + FB (00-RSM + FB seeds) and R + YL (00-RSM + YL seeds) than in those fed the control diet S-c, and the respective values were 24.7, 25.1 and 24.9 g/day vs. 26.2 g/day.Vegetable protein sources in finisher diets had no significant effect on N retention (p = 0.180) in pigs, which was high in all groups, ranging from 34.6 g/day in group S-c to 33.3 g/day in group R + FB. The experimental factor exerted no significant (p = 0.529) influence on the apparent biological value of protein expressed as N retention relative to N digested in pigs fed finisher diets with vegetable protein sources alternative to GM-SBM, which ranged from 56.3% to 57.8%.Selected serum biochemical parameters in pigs fed complete grower and finisher diets are presented in Table 4 and Table 5, respectively. Serum glucose concentrations were similar in all groups, ranging from 5.25 to 5.45 mmol/l (p = 0.121) in pigs fed grower diets and from 4.52 to 4.66 mmol/l (p = 0.432) in pigs fed finisher diets, which indicates that all diets had similar energy values. Total protein concentrations were highest in the control group S-c (63.78 and 67.52 g/l in grower and finisher diets, respectively) and lower in experimental groups R, R + FB, and R + YL, ranging from 61.58/65.68 to 62.72/67.18 g/l, but the noted differences were not statistically significant. Ammonium N levels tended to increase in the blood serum of pigs fed grower diets R + FB and R + YL (5.47 and 5.85 mmol/l, respectively) relative to those fed diet R (4.16 mmol/l). Similar trends were observed in pigs fed finisher diets. The above values point to lower utilization of protein from grower and finisher diets R + FB and R + YL. This fact was confirmed by N balance parameters (Table 4 and Table 5), showing that urinary N excretion was significantly higher in groups R + FB and R + YL (particularly in the grower phase) than in group R, characterized by the lowest urinary N excretion.Serum creatinine concentrations in pigs fed grower and finisher diets remained within the lower limit of the normal range (96.58–102.90/123.07–135.55 μmol/l, respectively) [24]. The above values show that the experimental factor had no negative influence on kidney function in growing-finishing pigs.The activities of AST, ALT, and GGT in the blood serum of pigs fed grower and finisher diets with different vegetable protein sources were similar in all groups. Serum AST levels remained within the reference ranges [24], whereas serum ALT and GGT levels slightly exceeded the upper limit of the normal ranges. The noted values indicate that the biochemical parameters of liver function were not adversely affected by the experimental factor.The fattening performance of pigs fed complete grower and finisher diets where 50% or 100% of GM-SBM protein was replaced with 00-RSM protein, alone or in combination with protein from FB and YL seeds, is presented in Table 6.At the beginning of the feeding trial, pigs had similar average initial body weights (26.21 kg to 26.57 kg). At the end of the grower phase of fattening, no significant (p = 0.509) differences in the average body weight of pigs were found between groups, and the noted values ranged from 65.54 kg in the control group S-c to 68.56 kg in group R + YL. The average final body weight of pigs did not differ significantly (p = 0.496) among groups.Complete grower diet R + YL (00-RSM + YL seeds) significantly (p ≤ 0.05) increased the average daily gains of pigs in the first phase of fattening (26–67 kg BW), compared with the control diet S-c (GM-SBM) and diet R (00-RSM), which reached 981 g vs. 915 g and 929 g. In the second phase of fattening (67–104 kg BW), finisher diets R + FB and R + YL tended to improve the growth rate of pigs (p = 0.074), relative to diet R. An analysis of average daily gains over the entire fattening period (26–104 kg BW) revealed that pigs fed diet R + YL were characterized by a significantly higher growth rate (1027 g/day) than the animals fed diets R (970 g/day, difference of 5.9%) and R + FB (986 g/day, difference of 4.3%).In the grower phase, FCR expressed as feed intake per kg of body weight gain was highly significantly (p ≤ 0.05) lower (2.46 kg/kg) in group R + YL than in the remaining groups (group S-c—2.68 kg/kg, group R—2.61 kg/kg, group R + FB—2.60 kg/kg). In the finisher phase, feed intake per kg of body weight gain was significantly (p ≤ 0.05) higher in groups R and R + FB than in groups S-c and R + YL. During the entire fattening period, the FCR was significantly (p ≤ 0.05) affected by the evaluated vegetable protein sources. The FCR ratio in group R + YL was 7.5% and 5.5% lower than in groups R and R + FB, respectively, whereas the difference relative to group S-c was not statistically significant.An analysis of metabolizable energy intake per kg of body weight gain revealed significant (p ≤ 0.05) differences between groups in the grower phase of fattening. The lowest value—31.36 MJ/kg was noted in pigs fed diet R + YL, compared with 33.56, 34.27, and 33.30 MJ/kg in groups S-c, R, and R + FB, respectively. In the finisher phase, metabolizable energy intake per kg of body weight gain tended (p ≤ 0.05) to be higher in groups R (39.17 MJ/kg) and R + FB (39.19 MJ/kg) than in group S-c (36.63 MJ/kg). During the entire fattening period, a significant tendency (p ≤ 0.05) to better utilization of metabolizable energy was noted in diet R + YL, compared with diets R and R + FB. The difference between diets R + YL and S-c was not statistically significant.An analysis of digestible protein utilization per kg of body weight gain revealed differences between groups. In the grower phase, the value noted in group R + YL (369 g/kg) was significantly (p ≤ 0.05) lower than those observed in groups S-c, R, and R + FB (392, 386, and 387 g/kg, respectively). In the finisher phase, digestible protein utilization tended (p = 0.081) to be worse in groups R, R + FB, and R + YL vs. group S-c. During the entire fattening period, digestible protein intake per kg of body weight gain was lowest (p = 0.073) in group R + YL (359 g/kg), compared with groups S-c (372 g/kg), R (375 g/kg) and R + FB (377 g/kg), and the noted differences reached 3.6%, 4.0%, and 5.0%, respectively.Carcass quality characteristics are presented in Table 6. No significant (p = 0.620) differences in dressing percentage were found between pigs fed diets with different vegetable protein sources. The average dressing percentage was lowest in the control group S-c (73.99%) and highest in group R + FB (75.19%). Pigs fed the R + FB diet were characterized by the highest average backfat thickness (17.00 mm). In the remaining groups, the values of this parameter ranged from 16.14 mm to 16.39 mm (p = 0.461). Carcass lean content ranged from 56.16% in group R to 56.74% in group R + YL (p = 0.287). Meat from pigs receiving diet R + FB was characterized by the highest value of loin eye height in the longissimus dorsi muscle (57.43 mm), whereas meat from pigs receiving diet R was characterized by the lowest value of this parameter (54.49 mm). The differences between group means were not statistically significant (p = 0.312).4. DiscussionPrevious research into the partial replacement of GM-SBM with 00-RSM, FB seeds or YL seeds in grower pig diets revealed differences in nutrient digestibility. In the present study, 00-RSM protein combined with protein from FB seeds or YL seeds as a substitute for 50% of GM-SBM protein in diets for growing pigs had a beneficial influence on nutrient digestibility. In the analyzed diets, the inclusion levels of 00-RSM, FB seeds, and YL seeds were 6%, 10%, and 6%, respectively. Such a combination contributed to a significant increase in the digestibility of protein and energy from grower diets, compared with the diet containing 12% of 00-RSM. It appears that the combined high-protein feedstuffs improved the quality of dietary protein due to the complementary effects exerted by nutritionally important, essential amino acids (methionine + cystine and lysine).Similar to the current study, Hanczakowska and Świątkiewicz [29] reported no significant differences in total protein digestibility between the control SBM-based diet and diets containing YL and FB seeds. Jezierny et al. [30] demonstrated that the standardized ileal digestibility of total protein from SBM and YL seeds did not differ significantly. However, the digestibility of FB protein was highly significantly lower than the digestibility of SBM protein (76% vs. 87%). Purwin and Stanek [31] analyzed the nutrient digestibility of legume seeds and found that total protein digestibility was lower in a diet containing 33% of FB seeds than in the control diet (67.2% vs. 75.5%). The inclusion of YL seeds at 18% improved protein digestibility from 73.8% to 74.4%. In a study by Mariscal-Landín et al. [32], the apparent ileal digestibility of total protein, including most amino acids, decreased significantly when RSM was added to the diet. Eklund et al. [33] and Torres-Pitarcha et al. [27] demonstrated that 00-RSM led to a decrease in total protein digestibility. The lower protein digestibility of RSM, compared with SBM, may be related to the high content of fiber, including lignin, in the seed coat of rapeseeds because amino acids absorbed by crude fiber are less available in the small intestine [27].In the present study, the coefficient of crude fiber digestibility was higher in groups R + FB and R + YL than in groups S-c and R. According to Hanczakowska and Świątkiewicz [29], the inclusion of FB and YL seeds in pig diets significantly increases crude fiber digestibility, from 29.1% to 34.5% and 39.2%, respectively. In a study of growing-finishing pigs fed legume-based diets, Stanek et al. [34] found that lower YL content (11.5%) increased, but higher YL content (18%) decreased crude fiber digestibility. Faba bean seeds led to a significant decrease in energy digestibility, whereas YL seeds had no significant effect on energy digestibility. Torres-Pitarch et al. [27] demonstrated that 00-RSM decreased the crude fiber digestibility of fattening pig diets.An analysis of the nutrient digestibility of finisher pig diets revealed a non-significant positive effect of 00-RSM combined with YL seeds on the digestibility of protein, fat, and energy, compared with 00-RSM applied alone. Hanczakowska and Świątkiewicz [29] found no differences in the digestibility of total protein, fat, or crude fiber between pig diets containing SBM, FB, or YL seeds. However, the digestibility coefficient of N-free extracts was highly significantly higher in the group fed FB seeds compared with the remaining groups. Purwin and Stanek [34] reported that FB seeds used as a complete substitute for SBM in finisher pig diets at 35% increased crude fat digestibility (59.0% vs. 62.0%) but decreased energy digestibility (82.2% vs. 79.7%). Crude fiber digestibility was improved (from 37.6% to 46.8%) by replacing SBM with YL seeds at 18%.In the present study, partial (50% in grower diets) and total (100% in finisher diets) replacement of GM-SBM protein with protein from 00-RSM, FB, and YL seeds had no adverse effect on the indicators of carbohydrate and protein metabolism or the biochemical parameters of liver and kidney function. However, 00-RSM used as a partial (50%) or total (100%) substitute for GM-SBM had a negative influence on fattening performance. Complete grower and finisher diets contained 12% and 13% of 00-RSM, respectively. Nevertheless, the average daily gain for the entire fattening period was high, at 970 g/day, and feed intake per kg body weight gain was 2.79 kg. The values of the above parameters were lower than in the control group fed SBM-based diets, but the noted differences were not statistically significant. The present results corroborate the findings of Sobotka et al. [4], McDonnell et al. [35], and Xie et al. [36]. According to McDonnell et al. [35], the results of previously published studies investigating 00-RSM as a substitute for SBM in diets for growing-finishing pigs are inconclusive. Some studies have demonstrated that 00-RSM can be included in pig diets at up to 20% without compromising growth performance, whereas others have shown that RSM inclusion levels of 10% to 20% may adversely affect fattening efficiency. These contradictory findings could result from differences in glucosinolate concentrations in rapeseeds. According to Torres-Pitarch et al. [27], a high nutrient content in pig diets is not always associated with better performance. The nutritional value of RSM may vary depending on processing technology, including temperature during toasting [36]. The negative impact of RSM on the growth performance of pigs, reported by Torres-Pitarch et al. [27], could result from its high glucosinolate content (15.58 μmol/g), which was more than two-fold higher than that determined in the present study.In this experiment, protein from 00-RSM combined with low-tannin FB seeds had a non-significant positive effect on the analyzed parameters. Grower and finisher diets containing 10% and 12% FB seeds, respectively, combined with 6% 00-RSM, had no significant negative influence on fattening performance compared with SMB-based diets. A combination of FB seeds and 00-RSM increased the growth rate of pigs by 4.5%, particularly in the grower phase when the average daily gain reached 956 g in group R + FB, compared with 915 g in group R. In the finisher phase, the average daily gain was similar in both groups, at 994 g and 986 g, respectively. The present results are consistent with the findings of other authors [29,37,38,39] who evaluated the effects exerted by the above dietary protein sources when applied alone.In the current study, partial (50%) and total (100%) replacement of GM-SBM protein with protein from 00-RSM and YL seeds had a significant positive influence on fattening performance. In grower and finisher diets, 6% of 00-RSM was combined with 6% and 7% of YL seeds, respectively. In group R + YL, the average daily gain was high in the grower (981 g) and finisher (1030 g) phases and over the entire fattening period (1027 g). Similar to the growth rate of pigs, FCR was also significantly better in group R + YL than in groups R and R + FB.Roth-Maier et al. [40] evaluated the effect of grower diets containing 20% of sweet lupine as a substitute for SMB on the growth performance and carcass characteristics of growing-finishing pigs and reported higher daily gains and body weights and a better FCR in the growing period in the experimental group, which corresponds to our findings. In the work of Hanczakowska and Świątkiewicz [29], YL seeds included at 8% and 12% in grower and finisher diets as a complete substitute for SBM had no significant positive effect on fattening performance. Similar observations were made by other authors [41,42], who found that YL could be used as a sole protein source in pig diets without compromising growth performance.In the current study, alternative vegetable protein sources partially and totally replacing SBM in grower and finisher diets had no significant effect on carcass quality traits. Similar results were reported by other authors who investigated 00-RSM [4,34,43], YL seeds [30,42], and FB seeds [39,40,41]. However, Degola [44] found that fat deposition increased significantly in the carcasses of pigs fed diets containing 20% of FB seeds.5. ConclusionsThe results of this study indicate that partial (50% in grower diets) and total (100% in finisher diets) replacement of GM-SBM protein with 00-RSM protein (6%) combined with protein from FB seeds or YL seeds in pig diets has no adverse effect on nutrient and energy digestibility, nitrogen balance, serum carbohydrate and protein metabolism or the biochemical parameters of liver and kidney function. Protein from 00-RSM (6%) and FB seeds (10/12%) contributed to high daily gains and high feed conversion efficiency. Protein from 00-RSM (6%) and YL seeds (6/7%) in grower/finisher diets led to a further improvement in fattening performance. The analyzed vegetable protein sources had no negative influence on carcass quality. It can be concluded that 00-RSM protein combined with protein from low-tannin FB or low-alkaloid YL seeds can be valuable high-protein feed ingredients alternative to GM-SBM in pig nutrition.

animals : an open access journal from mdpi

10.3390/ani11092550

PMC8464814

This study investigated the effects of Pennisetum purpureum waste mushroom compost (PWMC) supplementation on microbiota, as well as its effects on the antioxidant capacities and inflammatory response characteristics of broiler chickens. Results showed that a 5% replacement of a soybean meal via PWMC feeds could enhance the health of chickens by maintaining intestinal microbiota balance, improving antioxidant capacities, and decreasing inflammatory response. Supplementation also further increased the appetite of broilers, thereby improving their growth performances. Furthermore, the number of Lactobacillus also increased in the intestinal tracts. High-fiber mushroom waste compost effectively increased the mRNA expression of appetite-related genes in broilers. The broilers’ gut barrier function also increased, while the number of Turicibacter in the cecum decreased. It was concluded that a 5% replacement of a soybean meal via PWMC could enhance intestinal health; therefore, it is recommended for the broiler chickens’ diet.

This study investigated the effects of using mushroom waste compost as the residue medium for Pleurotus eryngii planting, which was used as a feed replacement; its consequent influence on broiler chickens’ intestinal microbiota, anti-inflammatory responses, and anti-oxidative status was likewise studied. A total of 240 male broilers were used and allocated to four treatment groups: the basal diet—control group (corn–soybean); 5% replacement of a soybean meal via PWMC (Pennisetum purpureum Schum No. 2 waste mushroom compost); 5% replacement of a soybean meal via FPW (Saccharomyces cerevisiae fermented PWMC); 5% replacement of a soybean meal via PP (Pennisetum purpureum Schum No. 2). Each treatment had three replicates and 20 birds per pen. The levels of glutathione peroxidase and superoxide dismutase mRNA as well as protein increased in the liver and serum in chickens, respectively; mRNA levels of inflammation-related genes were also suppressed 2 to 10 times in all treatments as compared to those in the control group. The tight junction and mucin were enhanced 2 to 10 times in all treatment groups as compared to those in the control, especially in the PWMC group. Nevertheless, the appetite-related mRNA levels were increased in the PWMC and FPW groups by at least two times. In ileum and cecum, the Firmicutes/Bacteroidetes ratios in broilers were decreased in the PWMC, FPW, and PP groups. The Lactobacillaceae in the ileum were increased mainly in the PWMC and control groups. Overall, high-fiber feeds (PWMC, FPW, and PP) could enhance the broilers’ health by improving their antioxidant capacities and decreasing their inflammatory response as compared to the control. Based on the results, a 5% replacement of the soybean meal via PWMC is recommended in the broiler chickens’ diet.

1. IntroductionAgricultural by-products such as wheat bran, crop hull, straw, etc., are inevitable substances in the agricultural process, but they are often not effectively used [1,2]. On the other hand, feed accounts for at least half of costs in the production of animal-related products [1,2]. However, with the increasing awareness of environmental protection, agricultural wastes or by-products are required to be properly treated [2,3,4]. In order to effectively reduce animal production costs and increase the value of agricultural by-products, previous studies have suggested the use of by-products such as mushroom waste compost since it is rich in mycelium and contains abundant functional secondary metabolites, which can improve antioxidant capacities, and regulate inflammatory response as much as animal feed ingredients [3,4,5].Carbohydrates, one of the main sources of energy for animals, can be roughly divided into two broad categories. The first category is the source of energy for animals such as starch, glucose, and sucrose, which are decomposed by enzymes produced by the animals. The second category is the dietary fiber that cannot be digested by the animal itself [6]. A distinction can also be made between hard-to-digest fibers such as cellulose, chitin, and other fibers, depending on whether the microorganism can be digested [5,6]. However, high-fiber agricultural by-products may reduce the performance of monogastric animals in traditional concepts [6], so this proposal has not been favored by animal producers for a long time. Nevertheless, there are increasingly more studies that have used agricultural by-products as the source of animal feed ingredients, and results have shown no reduction in the production performance of animals and that these could even help improve animal health [7,8]. It has been suggested that high-fiber feed ingredients could help to strengthen the intestinal barrier and antioxidant capacity of animals, and also reduce inflammation [4]. The additional antioxidant capacity can help in conditions of sudden oxidative stress, whether the negative factors come from the environment or pathogenic bacteria [9]. High-fiber feed ingredients also help with muscle formation and reduce adipose accumulation, thereby improving animal body composition [10]. In addition, the use of probiotics or enzymes to initially degrade agricultural by-products in vitro further increases the utilization of agricultural by-products [11].When animals are subjected to environmental stress such as pathogen infection or heat stress, stimulation causes an increase in oxidative stress and inflammation [12,13]. When stimulated by pathogenic bacteria, the animal initiates an immune response, which leads to a cytokine storm. However, excessive inflammation can reduce animal performance and even lead to death [14,15]. On the other hand, when the oxidative pressure is high, animals are not able to eliminate the damage caused by free radicals to cells and/or organs [16]. Moreover, the antioxidant system is mainly regulated by the liver; thus, the antioxidant capacity is also related to liver performance in animals [16,17,18]. Fiber could reduce the inflammatory response and increase the antioxidant capacity of the serum and the liver, as well as enhance the growth of intestinal villi and the overall health of animals [11]. Therefore, in the recent decade, more attention has been paid to the influence of gut microbiota composition on animal health [19]. Previous studies have pointed out that gut microbiota composition is not only related to animal health, anti-inflammatory levels, and antioxidant capacity, but also affected animal appetite, the circadian clock, and mood [19]. The intestinal microbial composition can be changed rapidly and directionally by altering dietary patterns [7]. Under the combined effects of microorganisms and diet, the metabolome of animals may be changed [20]. Studies of mammals have shown that the increased concentration of dietary fiber can promote the growth of Bacteroides, which reduce the number of Firmicutes in the gut [21]. Bacteroides are related to animals’ body maintenance, while Firmicutes are positively related to the degree of animal obesity [21]. In addition, increasing the dietary fiber concentration of animals also helps the growth of Bifidobacterium, which has a positive effect on animal health [21].Although considerable research has been conducted on the effects of gut microbiota composition on mammalian health, only a few studies have been performed on poultry. Therefore, the purpose of this study was to investigate the effects of high-fiber agricultural by-products on the broilers’ microbial activity under different treatments and to evaluate its anti-inflammatory and anti-oxidative effects in broilers.2. Materials and MethodsThe entire protocol for animal feeding was in accordance with rules established by the Animal Care and Use Committee at the NCHU (IACUC: 108−049). In order to evaluate the effects of PWMC (PP waste mushroom compost), FPW (Saccharomyces cerevisiae fermented PWMC), and PP (Pennisetum purpureum Schum No.2) on broilers’ antioxidant and anti-inflammatory capacities, intestinal tight junction expression, gut microbiota, and morphology were measured at the National Chung Hsing University (NCHU), Taiwan.2.1. Collection and Characteristics of Pennisetum Purpureum Schum No.2 (PP), PP Waste Mushroom Compost (PWMC), and Saccharomyces Cerevisiae Fermented PWMC (FPW)The PWMC is a kind of waste medium, which contains at least 70% PP after Pleurotus eryngii planting. Mature Pleurotus eryngii was removed and the medium was collected, dried, and stored at 4 °C in a refrigerator before use. The PWMC came from the Taiwan Agricultural Research Institute Council of Agriculture, Executive Yuan.The production of FPW was accomplished according to the method described by Chuang et al. [11], which was slightly modified for this study. PWMC was collected by the methods mentioned above; 1 mL of 108 CFU/mL S. cerevisiae was added to 100 g PWMC, and the moisture was adjusted to 60%. After culturing at 30 °C for 3 days, the fermented product was dried in an oven at 50 °C for 1 day and stored at −20 °C in a refrigerator before use. Before the test, we confirmed that the FPW contained at least 108 CFU/g of S. cerevisiae.The chemical composition (dry matter (DM), crude protein (CP), ether extract (EE), crush ash, neutral detergent fiber (NDF), and acid detergent fiber (ADF)) as well as the functional compounds (crude polysaccharide and volatile fatty acids) in PWMC, FPW, and PP were analyzed according to the methods described in AOAC (Assoc. Offic. Anal. Chem) [22] or the methods described below.2.1.1. Crude Polysaccharide MeasurementThe methods of pentose (D(+)-Xylose, Wako, Osaka, Japan) and hexose (D-(+)-Glucose, Sigma-Aldrich, St. Louis, MO, USA) measurement were modified based on the phenol-sulfuric acid method of Nielsen [23]. Briefly, the sample was extracted by deionized water at 95 °C for 30 min and cooled down to room temperature before use. After that, 1 mL 5% phenol solution and 5 mL sulfuric acid were added to PWMC, FPW, and PP water extracts. After incubation for 15 min, the absorbance was measured at 480 nm with U-2900 Spectrophotometer (Hitachi, Tokyo, Japan) for the standard curveof xylose and glucose2.1.2. Phenolic ComponentsThe PWMC, FPW, and FF water extracts (extracted at 95 °C for 30 min with deionized water) were filtrated by a 0.22 μm filter before use for the detection of catechins. The phenolic component content in the water extracts of PWMC, FPW, and FF were detected by HPLC (High Performance Liquid Chromatography) (Hitachi, Tokyo, Japan) with a pump (5110), a column (C18-AR, 250 × 4.8 mm, maintained at 40 °C by the column oven (5310)), an autosampler L-2200 (Hitachi, Tokyo, Japan), and a computer system with HPLC D-2000 Elite (Hitachi, Tokyo, Japan). The conditions of the mobile phase were (A) 0.05% v/v H3PO4 and (B) 3:2 v/v CH3OH/CH3CN solution, 1.0 mL/min, and UV detection at 280 nm. The known concentration (0.01–1.5 mg/mL) of gallocatechin (GC), epigallocatechin (EGC), catechin (CC), epicatechin (EC), gallic acid (GA), epigallocatechin gallate (EGCG), epicatechin gallate (ECG), catechin gallate (CG), and caffeic acid were also measured by the methods described above for the standard curve.2.1.3. Free Radical Scavenging Ability of 2,2-. Diphenyl-1-Picrylhydrazyl (DPPH)The assay for DPPH (2,2-diphenyl-1-picrylhydrazyl) scavenging capacity was conducted and slightly modified according to a previous report [24]. To deionize water, 1 g sample was added and soaked at 95 °C for 1 h in Tempette Junior TE-8J Heating Water Bath (Techne, Staffordshire, UK). After cooling, it was centrifuged at 3000 rpm× g for 10 min with Heraeus Megafuge 8 Centrifuge (Thermo Fisher Scientific Inc., Waltham, MA, USA). The supernatant was added to 0.1 mM DPPH and after incubation for 30 min, the absorbance was measured at 517 nm with a U-2900 Spectrophotometer (Hitachi, Tokyo, Japan). Butylated hydroxytoluene (BHT) was taken as the positive control. DPPH scavenging capacity (%) = (1 − sample 517 nm OD/control 517 nm OD) × 100%).2.1.4. Reducing PowderThe measurement of the reducing powder was performed as described by the Oyaizu [25]. Briefly, the sample was extracted in deionized water at 95 °C for 30 min and cooled down to room temperature before use. Four milliliters of sample solution, 1 mL 0.2 M PBS (pH 6.6), and 1 mL 1% potassium ferricyanide were mixed and incubated at 50 °C for 20 min in Tempette Junior TE-8J Heating Water Bath (Techne, Staffordshire, UK). The incubated mixture was cooled down with ice and 1 mL trichloroacetic acid was added. The new mixture was centrifuged at 3000 rpm× g for 10 min with Heraeus Megafuge 8 Centrifuge (Thermo Fisher Scientific Inc., Waltham, MA, USA) and 5 mL supernatant was moved to a new tube. The supernatant was mixed with 1 mL 0.1% ferric chloride and kept at room temperature for 10 min. The absorbance of the sample at 700 nm was measured with U-2900 Spectrophotometer (Hitachi, Tokyo, Japan) and compared to the results of BHT.2.1.5. Malondialdehyde ProductionThe method used for malondialdehyde (MDA) measurement was based on the use of thiobarbituric acid reactive substances (TBARS) described by Zeb and Ullah [26]. Briefly, the sample was extracted in deionized water at 95 °C for 30 min and cooled down to room temperature before use. One milliliter of the sample solution was mixed with 4 mL 1% lecithin, 20 mM Na2HPO4-NaH2PO4 buffer, 0.2 mL 25 mM FeCl3, and 2.5 mL 20 mM Na2HPO4-NaH2PO4 buffer (pH 7.4). The mixture was incubated at 37 °C for 2 h in Tempette Junior TE-8J Heating Water Bath (Techne, Staffordshire, UK) and mixed with 2 mL 20 mg/mL BHT, 4 mL 1% thiobarbituric acid, and 2 mL 2.8% TCA to stop the reaction. The absorbance of the mixture was measured at 532 nm with U-2900 Spectrophotometer (Hitachi, Tokyo, Japan) and the rate of inhibition of MDA production was calculated.2.1.6. Ferrous Chelating CapacityThe ferrous chelating capacity assay was performed according to the methods described by Yuris and Siow [27]. Briefly, the sample was extracted in deionized water at 95 °C for 30 min and cooled down to room temperature before use. One milliliter of the sample solution was mixed with 0.1 2 mM ferrous chloride solution and 3 mL deionized water. The mixture was kept at room temperature and protected from light for 10 min. Then, 0.2 mL ferrozine was added to the mixture and the absorbance was measured at 562 nm with U-2900 Spectrophotometer (Hitachi, Tokyo, Japan). The following formula was used to calculate the ferrous chelating capacity:Ferrous chelating capacity (%) = [(A0 − A1)/A0]/100(1) where A0 is the absorbance of the positive control (Ethylenediaminetetraacetic acid), and A1 is the sample absorbance.2.2. Animal Experiment DesignThe animal experiment design was slightly modified and followed by the procedure described by Chuang et al. [11]. Briefly, a total of 240 male broilers (Ross 308) were used and allocated to 4 different treatments: the corn–soybean basal diet, 5% replacement of a soybean meal via PWMC, FPW, and PP. Each treatment had three replicates (20 broilers for each pens). The initial weight of 1-day-old chicks for each treatment was similar (48.0 ± 0.7 g/bird). The room temperature was controlled throughout the experiment (33 °C for 1-day-old chicken, and decreased slowly to 22 °C after 30 days). Each group had similar crude protein and gross energy content in both starter (from 1 to 21-day-old) and finisher (from 22 to 35-day-old) feed; furthermore, the nutrient concentration of the broilers’ diet was achieved or exceeded nutrient requirement according to NRC 1994 (Table 1). Feed for the different treatments was recalculated and the approximate composition (energy equivalent to 3050 vs. 3175 kcal/kg DM and crude protein at 23 vs. 21% DM for the starter and finisher periods, respectively) were analyzed according to the methods described in AOAC [22] and shown in Table 1. The serum, chyme in ileum and cecum, liver, spleen, hypothalamus, ileum, and jejunum were collected from 35-day-old broilers for the follow-up analysis.2.3. Serum CharacteristicsSerum from broilers was collected from a total of 36 chickens (three from each pen, nine for each treatment) following the methods described by Chuang et al. [11]. Briefly, chicken blood was collected and stored at 4 °C in a refrigerator R2551HS (TECO Electric and Machinery Co., Ltd., Taipei, Taiwan) for 4–5 h. Blood samples were centrifuged at 3000 rpm× g for 10 min in order to separate the blood cells and the serum, and the serum was stored at −20 °C in a refrigerator before being analyzed. The concentration of tumor necrosis factor-alpha (TNF-α), glutathione peroxidase (Gpx), and superoxidase dismutase (SOD) in broiler serum were measured (TNF-α, Gpx, and SOD were from Cayman Chemical Co., Ltd. Ann Arbor, MI, USA). All the methods for analyzing these compounds followed the manufacturer’s protocol. Other serum characteristics were measured with an automatic biochemical analyzer (Hitachi, 7150 auto-analyzer, Tokyo, Japan).2.4. DNA Extraction and 16S rRNA Gene SequencingIleal and cecal digesta in 35-day-old broilers were collected from 6 broilers in each of the four (control, PWMC, FPW, and PP) groups. Samples were immediately isolated for their genomic DNA by using Quick-DNA Fecal/Soil Microbe Miniprep Kits (Zymo, Irvine, CA, USA); procedures were accomplished by following manufacturer’s protocol. Extracted DNA qualities, including purity and concentration, were analyzed through the NanoDrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). Briefly, ileum or cecum chyme from 35-day-old broilers was collected in collection tubes provided by the manufacturer and stored at −80 °C in a refrigerator before being analyzed. After a series of dissolution, filtration, and extractions, pure microbial DNA from broiler ileum or cecum chyme was isolated. The concentration and purity of purified DNA were measured with the absorbance ratio of 260/280 nm. For the identification of microbial species, the V3–V4 hypervariable regions of 16S rRNA were sequenced. The Qubit® 2.0 Fluorometer (Thermo Scientific) and an Agilent Bioanalyzer 2100 system were used for the library quality and the final library was sequenced by the IlluminaHiSeq2500 platform with 250 bp paired-end reads. After sequencing, the UCHIME algorithm was used to detect chimera sequences that were then removed to obtain effective tags. Uparse (Uparse v7.0.1090; http://drive5.com/usearch/, accessed on 11 June 2021) software was used for the sequence analysis. Microbes with operational taxonomic units (OTUs) at 97% similarity were clustered using USEARCH (version 7.0.1090). The representative sequence of the OTUs was selected for further annotation. The difference of alpha-diversity (Shannon index, Observed species, Simpson, and Abundance-based coverage estimator (ACE)) were analyzed by QIIME (Version 1.9.1) and displayed by R software (v.3.3.1) using the alpha function in the microbiome R package. The relative abundance of the OTU in every treatment was evaluated by metagenomeSeq.2.5. Total RNA Isolation, qPCR, and SequencingThe mRNA of 35-day-old broilers was collected from 6 broilers from each of the four groups and isolated from the liver (for antioxidant capacities measurement), spleen (for anti-inflammatory capacities measurement), ileum (for tight junction expression measurement), and hypothalamus (for appetite regulation measurement), respectively. Among these, the mRNA collected from the hypothalamus was removed from broilers before (H1) and after (H2) fasting for 24 h. During fasting, the birds were provided with water spray and ventilation. The methods of mRNA isolation and qPCR measurement were both described by Chuang et al. [11]. Briefly, total mRNA from each tissue (approximately 100 mg for individual tissues) was isolated following the manufacturer’s protocol (SuperScript™ FirstStrand Synthesis System reagent, Invitrogen, Carlsbad, CA, USA) and purity was determined with the absorbance ratio of 260/280 nm with Epoch Microplate Spectrophotometer (BioTek Instruments, Winooski, VT, USA). The PCR conditions were as follows: initial denaturation at 95 °C for 20 s, 40 cycles of 95 °C for 3 s, 55 °C for 30 s, and 72 °C for 30 s, and then final extension at 72 °C for 5 min using StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific Inc., Waltham, Massachusetts, USA). After synthesizing cDNA from mRNA, cDNA was mixed with 2× SYBR GREEN PCR Master Mix-ROX (Gunster Biotech, Co., LTD, New Taipei City, TW), deionized water, and each primer in the ratio of 25:6:9:5. StepOnePlus™ Real-Time PCR System (Thermo Fisher Scientific Inc., Waltham, MA, USA) was used to assess the qRT-PCR performance. After these, 2-ΔΔCt were measured in each group, and the relative expression level of each mRNA was calculated and compared to that of the control group. The methodology was as described by Schmittgen and Livak [28]. ß-actin was used as a housekeeping gene for normalization. All primes were designed according to the genes of Gallus (chicken) from GenBank (Table 2). After sequencing, whole tags were assembled using the UCHIME algorithm to detect chimera sequences; the chimera sequences were removed before the effective tags were obtained. Sequence analysis was performed using the Uparse software (Uparse 135 v7.0.1090; Available online: http://drive5.com/usearch/ (accessed on 15 January 2020)). Sequences with ≥ 97% similarity were assigned to the same operational taxonomic units (OTUs). A representative sequence of each OTU was selected for further annotation. Alpha diversity was applied to analyze the complexity of the species diversity of a sample by using six indices: Observed OTUs, Shannon, Simpson, Chao1, ACE, and PD whole tree. All the indices of our samples were calculated using Quantitative Insights Into Microbial Ecology (QIIME, v1.9.1). To evaluate differences in samples with respect to species complexity, beta diversity analysis on weighted unifrac was conducted using the QIIME software (v1.9.1). A principal coordinates analysis (PCoA), using the R software (v.3.3.1) with the ade4 and ggplots packages, was performed at the genus level. LEfSe (Linear discriminant analysis effect size) was performed to detect differential abundant taxa across groups using the default parameters.2.6. Statistical AnalysisThe data collected were statistically analyzed using the general linear models procedure of the SAS software (SAS® 9.4, 2018) following a Completely Randomized Design (CRD). Variables of serum, intestine, and gut microbiota were tested for normal distribution and homogeneity. Data on the dietary treatments were subjected to analysis of variance using the Statistical Analysis System Institute Package (SAS) and the mean values were compared using Duncan’s range test, with the significant level at p < 0.05.The mathematic model was as follows:Yij = µ + Di + εij where Yij is the average of birds in pen j and their dietary treatment i; µ is the overall mean; Di is the fixed effect of the dietary treatment (1: control, 2: PWMC, 3: FPW, and 4: PP); εij is residual error when the pen was regarded as an experimental unit, (0, σ2ε).3. Results3.1. The Characteristics of PWMC, FPW, and PPThere were several functional components in PWMC, FPW, and PP (Table 3). All PWMC, FPW, and PP had a high content of polysaccharide (85, 83, and 88 glucose mg/g DM, respectively), total phenol content (1.8, 1.8, and 2.1 GAE mg/g DM, respectively, data not shown), and total flavonoids (1.2, 1.3, and 1.7 QE mg/g DM, respectively, data not shown). The phenol content in PWMC, FPW, and PP, including GC, EGC, CC, EC, GA, EGCG, ECG, CG, and caffeic acid, were estimated. Among them, there was at least 0.1% EGC and GC content in PWMC, FPW, and PP, and about 0.02% EC and EGCG in PWMC, FPW, and PP. Particularly, after being fermented by S. cerevisiae, the phenol-like chemical increased by about 1.2 to 2 times in PWMC. Some natural organic acids were also present in PWMC, FPW, and PP. However, after fermentation by Pleurotus eryngii or S. cerevisiae, the volatile fatty acids increased much more in PWMC and FPW than in PP (31, 28, and 12 μM/g, respectively). Fermentation by S. cerevisiae could also increase the hemicellulose content from about 11% to 20% in FPW as compared to that in PWMC (data not shown, calculated by the difference between NDF and ADF).Both phenols and flavonoids were found to have high antioxidant capacities. The antioxidant capacities are shown in Figure 1, including the ferrous ion chelation rate (Figure 1A), Di (phenyl)-(2,4,6-trinitrophenyl) iminoazanium (DPPH) scavenging capacities (Figure 1B), total reducing capacities under 700 nm absorbance (Figure 1C), and malondialdehyde production inhibition rate (Figure 1D). Among them, PWMC, FPW, and PP had similar and high antioxidant capacities in all experiments.3.2. Serum Characteristic of BroilersThe data of the serum characteristics of 35-day-old broilers showed that an increase in the fiber content of broiler feed could significantly decrease the glucose concentration in the broilers’ blood as compared to that in the control group (from 276 mg/dL to 236, 240, and 227 mg/dL; p < 0.05). Nevertheless, the triglyceride concentration in the broilers’ blood also decreased by about 25%, in both the control group and in the other treatments (55, 41, 38, and 40 mg/dL; p < 0.001). As the classic antioxidant enzyme in animal, the concentration of SOD in the broilers’ serum was increased in the FPW and PP groups (p < 0.001), but not in the PWMC group, as compared to that in the control group. The concentration of TNF-α in the serum of the broilers also decreased significantly (p < 0.05) compared to that in the control group (Table 4); however, the interleukin 1 beta (IL-1β) was underdetermined in each treatment (not shown in data).3.3. Gut Microbiota Composition and Function3.3.1. Microbiota Composition in the Ileum of BroilersAccording to Figure 2A,B, comparing with other groups, the FPW group has the highest (p < 0.05) relative abundance of species in the broilers’ ileum, followed by the PP and PWMC groups. Among them, only three different microbes were detected in the control group, while FPW had 70 species (Figure 2B). The Firmicutes/Bacteroidetes ratios decreased (p < 0.05) in the PWMC, FPW, and PP groups, but the FPW group showed a higher (p < 0.05) standard error range (Figure 2C). At the phylum level, the relative abundance of Proteobacteria seemed hardly changed; however, the Epsilonbacteria and Bacteroidetes increased (p < 0.05) in PWMC, FPW, and PP. Among them, the relative abundance of Cyanobacteria increased (p < 0.05) only in the PP group (Figure 2D). At the genus level, the abundance of Romboutsia seemed similar to that of the control, FPW, and PP groups, but decreased (p < 0.05) in the PWMC group. The relative abundance of Campylobacter increased in broilers after a high-fiber treatment (PWMC, FPW, and PP), especially in the FPW group. Among them, the Eimeria praecox increased only in the PP group. The relative abundance of Candidatus arthromitus was 10 times higher (p < 0.05) in PWMC, and especially in the FPW and PP groups, as compared to that in the control group. Furthermore, Turicibacter decreased (p < 0.05) mostly in PP, followed by FPW and PWMC, as compared to that in the control group (Figure 2E,H). Particularly, at the order and family levels, the relative abundance of Lactobacillales and Lactobacillaceae, respectively, increased (p < 0.05) in the control and PWMC groups (Figure 2F,G). According to the alpha diversities shown in Figure 2I–L, the PWMC, FPW, and PP groups showen higher (p < 0.05) amounts of observed species and ACE index compared to those of the control group; however, the difference was not significant. Each group had a similar score in the Shannon and Simpson indices.3.3.2. Microbiota Composition in the Cecum of BroilersThe microbial relative abundance in the cecum was similar in each group (Figure 3A,B), which was in contrast to the data shown in the ileum (Figure 2A). However, the Firmicutes/Bacteroidetes ratio still decreased (p < 0.05) in the high-fiber groups (Figure 3C). According to the relative abundance shown in Figure 3D, we found that the percentage of Bacteroidetes and Verrucomicrobia increased (p < 0.05). At the genus level, almost 50% of the microbes occurred in all groups, in the control as well as in the other treatments. The relative abundance of Akkermansia increased (p < 0.05) in the high-fiber treatments, especially the FPW. Other microbes that increased (p < 0.05) in relative abundance included Ruminococcaceae_UCG_014, Lactobacillus, and Barnesiella. However, the Faecalibacterium decreased (p < 0.05) only slightly in the high-fiber treatments as compared to the control group (Figure 3E). According to the alpha diversities shown in Figure 3I–L, the treatment groups still had the higher number of observed species and ACE index (p < 0.05), but the standard error in the FPW group was much higher than in the other groups (p < 0.05) (Figure 3I,J).3.4. mRNA Expression in Liver, Spleen, and Ileum of BroilersConcerning mRNA expression in the spleen of 35-day-old broilers, the mRNAs of most of the inflammation-related proteins decreased significantly (p < 0.05). Among them, the IL-1β, inducible nitric oxide synthases (iNOS), and nuclear factor kappa B (NFκB) decreased (p < 0.05) by over 10 times after each treatment as compared to those in the control group (iNOS mRNA expression did not differ significantly in only PWMC and the control). Furthermore, the level of interferon-gamma (IFN-γ) mRNA decreased (p < 0.05) only in the FPW group (Figure 4).The nuclear factor erythroid 2-related factor 2 mRNA expression in the broilers’ liver increased 4–5-fold in the high-fiber treatment groups as compared to the control group; therefore, the glutamate-cysteine ligase catalytic (GCLC) and heme oxygenase-1 (HO-1) mRNA expressions also increased. Among them, GCLC increased significantly only in the FPW group, and HO-1 increased the most in the PP group. Furthermore, the GPx and SOD mRNA expressions in the broilers’ liver also increased significantly (p = 0.003 and < 0.001). High-fiber feed replacement could significantly increase the tight junction and mucus, including claudin 1, mucin 2 (MUC2), as well as occludin and zonula occludens 1 (ZO-1) mRNA expressions in broilers (p < 0.001, < 0.01, < 0.001, and < 0.005, respectively) (Figure 4).Among the expressions of appetite regulation-related mRNAs, there was no significant difference in the leptin (LEP) expression in the hypothalamus of the broilers before fasting; however, after fasting for 24 h, the LEP decreased by over 2-fold in the PWMC and FPW groups as compared to the control group, but not in the PP group (p < 0.001). Furthermore, the neuropeptide Y (NPY) expressions in PWMC and FPW were higher (p < 0.05) than in the control group, both before and after fasting; however, in the PP group, NPY increased only after fasting for 24 h. The pro-opiomelanocortin (POMC) and cocaine- and amphetamine-regulated transcript (CART) mRNA expressions in all treatments were lower (p < 0.05) than those in the control group. However, among the PWMC, FPW and PP treatment, the expression of POMC and CART mRNA in PP was still higher (p < 0.05) than in PWMC and FPW groups both before and after fasting (Figure 4).4. DiscussionThere are plenty of edible mushrooms worldwide; however, only a few mushrooms are suitable for planting and can be widely accepted by consumers [4]. Among them, Pleurotus eryngii is one of the most popular edible mushrooms [4]. Previous studies have shown that Pleurotus eryngii contains functional compounds such as triterpenes and phenol, which could increase the antioxidant capacities of animals and improve animal health [28]. When planting mushrooms, the cost and medium composition should be considered [2]. As a cheap, high-fiber content and several functional compounds, Pennisetum purpureum Schum No.2 was considered a good substance in which to culture mushrooms [2]. PWMC was the medium in which Pleurotus eryngii had been planted; thus, it contained a high amount of myelin in several functional compounds (Table 3). However, in order to maintain the growth of Pleurotus eryngii, the residue medium must contain higher cellulose and lignin (Table 3). This situation would be ameliorated by S. cerevisiae fermentation. FPW showed higher hemicellulose content, and more extractable EGCG-like compounds that were released from PWMC. Earlier studies showed that the concentration of extractable functional compounds (phenol and phosphorus, especially) would increase on fermented prebiotics [29]. Phenol and flavonoids have been shown to enhance antioxidant capacities both in vitro and in vivo [30]. By chelating ions, phenol could decrease the lipid oxidation rate, thereby decreasing the MDA content [9,31]. Furthermore, PWMC, FPW, and PP were also good at decreasing the free radicals from the DPPH. These results show that PP contains phenol-like compounds, which has slightly increased level after culturing with Pleurotus eryngii and S. cerevisiae fermentation.In Table 4 and Figure 4, we observed that the antioxidant-related enzyme and mRNA levels were increased, and a decrease in inflammation. In animals, the antioxidant system can be divided into two categories: enzyme and non-enzyme systems [9]. Among the enzymes, catalase could decrease the accumulation of hydrogen peroxide; the non-enzymatic category includes glutathione (GSH), which is well-known to reduce free radicals. Both of these help in reducing reactive oxygen species (ROS) production and in enhancing animal health [24]. Nrf2 is a common transcription factor upstream mRNA that can enhance the production of a series of antioxidant proteins such as GCLC and HO-1 by binding to the antioxidant responsive element (ARE) in the nucleus [32]. HO-1 could decrease the oxidant stress in animals, and the amount of GCLC is positively correlated with GSH content. GSH could bind to electrons on ROS and become GSSH to reduce the damage due to ROS. Gpx is the key enzyme that can oxidize GSSH in order to form GSH [31]. Although the Gpx protein increased only slightly in the broilers’ serum, the Gpx mRNA in the broilers’ liver increased significantly in the high-fiber treatment groups. Furthermore, both SOD protein and mRNA increased in broilers on the high-fiber treatment, especially in the FPW and PP groups, which could protect the broilers from ROS by transferring O- to H2O2 and further provide the substrate for the catalase [31]. In addition, inflammation is a double-edged sword—it does not only help animals to face a pathogen but it may also cause further cell damage [9]. Based on this, it can thus be understood that when an animal finds itself in a normal situation, oversensitive inflammatory reactions are not desired [9]. Therefore, under normal conditions, a decrease in inflammation levels in animals should be considered. PWMC, FPW, and PP could decrease both the protein and mRNA levels of TNF expression. Furthermore, the main upstream inflammatory signals of mRNAs, IL-1β, and NFκB all decreased significantly under a high-fiber treatment, and further decreased the iNOS mRNA expression [9]. It is worth noting that IFN-γ only decreased in the FPW group. This could be due to the glucan and mannan content in S. cerevisiae, which can help to encase lipopolysaccharides in the environment. Similar results were shown by Chuang et al. [11]; the addition of probiotic (106 CFU/kg feed of S. cerevisiae) fermentation could decrease the IFN-γ by over 10 times. Similar results in the broilers’ gut barrier function were shown; among them, FPW had the highest claudin-1 mRNA expression. Enhanced gut barrier function can help in animal defense against infection from pathogens in the gut and decrease the level of inflammation [7]. With no inflammation, the broilers’ feed intake increased. POMC is one of the main appetite regulation peptides in animals and is negatively correlated with animal feed intake by inducing CART and MSH family production [33,34]. Furthermore, LEP is well-known for its anti-obese function in mammals and was considered to be present in poultry until 2014 [35]. By improving satiety and accelerating muscle formation in animals, LEP could effectively decrease animal body weight. In contrast to these, NPY is the main neuropeptide that can enhance animal feed intake [33]. Our results showed that before fasting, the POMC and CART mRNA expressions decreased and that of NPY increased mainly in the PWMC and FPW groups. However, when fasting for 24 h, the LEP further decreased in the PWMC and FPW groups, and the NPY mRNA expression also increased in the PP group. These data reveal that PWMC and FPW could enhance appetite-inducing mRNA expression, while PP appears to maintain normal appetite levels in broilers. The possible reason may be the regulation of animal appetite not only through endocrines but also at the nutrient level and other physiological levels [36].Compared to the ileum, the cecum is the main site for fermentation and contains a large number of microbes [37]. Therefore, it seems that the high-fiber feed mainly affects the relative abundance of microbes in the ileum (Figure 2 and Figure 3). As expected, the Bacteroidetes increased in the high-fiber groups as compared to the control group in both ileum and cecum. Previous studies have shown that the dietary fiber in food is positively correlated with the abundance of Bacteroidetes, and increasing the ratio of Bacteroidetes/Firmicutes further enhances animal health and downregulates the inflammatory response [7,21]. Furthermore, the abundance of Bacteroidetes is also correlated with lower body weight [21], which probably because of an increase in the rate of lipolysis. In fact, the accumulation of subcutaneous fat in PWMC, FPW, and PP decreased by about 30%, and the percentage of the inner breast in the broilers’ carcass increased slightly (not shown in data). Turicibacter is a common bacterium in the animal gut, which is positively related to colitis [38]. However, the number of Turicibacter in the ileum of broilers decreased in the high-fiber groups, which could have consequently decreased the inflammatory level in broilers. Desai et al. [7] also indicated that an increase dietary fiber could decrease the level of colitis.Lactobacillus increased in relative abundance in the ileum of the control and PWMC groups, and in the cecum of the PWMC and PP groups. The possible reason is that S. cerevisiae competes with Lactobacillus because both of them are fiber- prefer microbes [39,40]. Furthermore, PWMC could provide more growing substrates for Lactobacillus than PP could, thus explaining the growth of Lactobacillus in the ileum of the PWMC group. Interestingly, accelerated lipolysis, decreased inflammation levels, and the increase in the number of Lactobacillus could be directed to promote animal health and effectively reduce fat accumulation. Overall, it seems that high-fiber feed could enhance animal antioxidant capacities, decrease inflammatory levels, and alter gut microbes both in the ileum and the cecum, thus leading to improv animal health. However, the types and sources of fibers also limit the availability for products. For example the fibers with too high a molecular weight of polymerization may be difficult digested and utilized in a short time of animals in diet. Moreover, many soluble non-starch polysaccharides (NSP) may lead to excessive fermentation by gut microorganisms and reduce growth/production performances in animals [15,21]. Therefore, in the study when considering the commercial aspect, PWMC—on which Pleurotus eryngii was first cultured—did not need further fermentation, and yielded similar antioxidant capacities and slightly higher growth performance in 35-day-old broilers. Therefore, a 5% replacement of a soybean meal via PWMC is recommended in broilers’ feed as opposed to other groups.5. ConclusionsOur results show that high-fiber feeds (PWMC, FPW, and PP) could enhance animal health by inducing antioxidant capacities and decreasing the inflammatory response in broilers as compared to the control group. Furthermore, the number of Lactobacillus (ileum) increases in the FPW and PP groups, also increases (cecum) in the PWMC and FPW groups. Additionally, this study showed that in order to enhance the broilers’ health, a 5% replacement of a soybean meal via PWMC is recommended for the broilers’ diet.

animals : an open access journal from mdpi

10.3390/ani12070914

PMC8997066

Laboratory tests are an integral part of the overall diagnostic procedures for both physiological and pathological conditions. Ovariohysterectomy (OHE) in bitches is an invasive procedure, producing moderate to severe pain. The response to surgical stress is characterized by a series of neuroendocrine and metabolic changes that can be represented on hematological and biochemical profiles. The objective of this study was to evaluate the physiological response to surgical stress, predict the presence of possible post-surgical complications, and where appropriate, establish the appropriate treatment based on the hematochemical changes during and after OHE in healthy bitches. OHE in this study induced transient changes in certain hematological and biochemical parameters. Knowledge of the hematochemical changes in response to stress and trauma induced by OHE in healthy dogs would allow to predict possible post-surgical complications.

The aim of this study was to monitor hematochemical changes during and after OHE in bitches. Twenty-four females were anesthetized with alfaxalone, midazolam, morphine and sevoflurane. Blood samples were taken before anesthesia (T0), at 30 (T1), and 60 min (T2), at 3 (T3), 6 (T4), 12 (T5), and 24 h (T6), and at 3 (T7) and 7 days (T8) from the start of surgery. Red blood cells (RBC) and packed cell volume (PCV) decreased significantly from T1 to T5 and hemoglobin (HB) concentration from T4 to T6. Both the white blood cell (WBC) and neutrophil (NFS) count increased significantly from T3 to T6, monocyte (MON) from T2 to T5, and eosinophil (EOS) at T5. Platelet (PLT) and plateletcrit (PCT) significantly decreased at T5 and increased from T6 to T8; platelet distribution width (PDW) increased significantly from T3 to T6. Creatine kinase (CK) activity increased significantly from T5 to T7. Glucose (GLU) concentrations increased significantly at T2 and P from T2 to T3. TG levels decreased from T2 to T4 and blood urea nitrogen (BUN) levels from T1 to T7, subsequently increasing until T8. Changes possibly resulting from stress and surgical trauma, as well as hemodilution and splenic storage, are due to anesthesia and surgery. In healthy bitches, these changes tend to gradually stabilize after the ending of OHE. A post-operative follow-up is essential to detect possible post-operative complications.

1. IntroductionOvariohysterectomy (OHE) is one of the most common surgical contraceptive techniques used for population control in bitches. Other potential benefits of sterilization include prevention of inherited diseases, elimination of undesirable behaviors associated with hormonal cycling and if performed before 2½ years of age, a reduction in the risk of ovarian, uterine, and mammary neoplasia [1]. Other pathologies that justify OHE include uterine and ovarian tumors, pyometra and glandular cystic uterine hyperplasia with secondary infection leading to chronic metritis [2,3]. For these clinical purposes, various surgical techniques, including traditional midline, lateral flank and laparoscopic OHE are used [4].Surgical stress response during and after surgery involves a physiological reaction of the organism to restore homeostasis or counteract the stimulus [5]. Somatic and autonomic afferent nerve impulses generated at the injury site activate the endocrine response mediated by the stimulation of the sympathoadrenal system and the hypothalamic-pituitary-adrenal axis [6]. At the same time, inflammatory and immune responses begin to develop, mediated by cytokine secretion products of activated leukocytes, fibroblasts and endothelial cells. These compensatory mechanisms prevent secondary damage and increase the availability of substrates required by essential organs and healing tissues [7,8] and implicate hematologic, metabolic, and immunomodulatory changes [8]. Several studies have evaluated the changes of hematochemical parameters in OHE bitches following various surgical techniques, and using various anesthetic protocols, providing different results [9,10,11,12].Surgical anesthesia is characterized by immobility, absence of consciousness, muscle relaxation and lack of pain [13]. It has been suggested that the application of any anesthetic can potentially modify the physiological response to the surgical procedure due to cardiovascular, respiratory, digestive, and neuroendocrine alterations, thus, causing homeostatic, metabolic, and immunological changes [8,14]. Nevertheless, there is controversy about this since a single anesthetic administration by itself has shown to cause scarce or no response to stress [15]. Therefore, most metabolic changes appear to be caused by the surgical procedure itself [16].Alfaxalone (3-alpha-hydroxy-5-alpha-pregnane-11, 20-dione) is a synthetic neuroactive steroid molecule that modulates the gamma-aminobutyric acid A (GABAA) receptor and causes neuro-depression, muscular relaxation, chemical restraint and/or general anesthesia [17]. It can be used for sedation, induction of anesthesia, and for total intravenous anesthesia (TIVA). It causes a minimal change in cardiac output or blood pressure when clinically relevant doses are administered to healthy patients [17]; it has a high therapeutic index, is short-acting, and is noncumulative [18,19]. These characteristics make alfaxalone ideal for its use as an induction agent or injectable anesthetic. Alfaxalone is used for OHE in bitches [20,21,22,23]. Midazolam, administered before anesthetic induction, provides good muscle relaxation, reducing the alfaxalone-related hyperkinesia [24,25]. Sevoflurane, due to its low blood:gas partition coefficient and rapid onset of action, allows easy control of the anesthetic depth [26]. Finally, for its efficacy in treating intraoperative and post-operative pain, morphine continues to be used in veterinary medicine [26], even though other drugs such as methadone and fentanyl are also being widely used in many countries.At present, the physiological importance of blood findings in bitches undergoing OHE in both clinical and experimental studies is not well defined. Although the incidence of complications in OHE procedures is relatively low in healthy animals, it still involves risks during and after surgery [27]. Accordingly, the aim of the present study was to monitor hematologic and biochemical changes during and after OHE in healthy bitches following the administration of alfaxalone, midazolam, and morphine as premedications, alfaxalone as induction, and sevoflurane as maintenance. The exploration of dynamic changes in normal hematochemical tests could help to better understand possible complications during and after OHE and to establish the most effective treatment as soon as possible.2. Materials and Methods2.1. AnimalsThis study was conducted in the Veterinary Teaching Hospital of CEU-Cardenal Herrera University (Valencia, Spain). A total of 24 client-owned adults, healthy, mixed-breed female dogs were included. A complete history of the animals, including breed, age, parity and stage of the oestrous cycle, was recorded. After being informed of the fundamentals of the study, the owners signed a consent form for their pets to be included. All animals underwent a pre-anesthetic evaluation to confirm that they were healthy. This evaluation included physical examination, complete hematological and biochemical analysis, thoracic X-rays, and electrocardiogram. In addition, all animals were subjected to a commercially available SNAP® 4Dx® Plus test for the detection of antigen to Dirofilaria immitis, antibodies to Borrelia burgdorferi, Anaplasma phagocytophilum, Anaplasma platys, Ehrlichia canis and Ehrlichia ewingii (IDEXX SNAP®), yielding negative results in all cases. Only animals classified as ASA I were included in the study. None of the animals received medical treatment during the last 6 months prior to the study, nor were they diagnosed with active disease. During the study period, any animal received medication, and no animals had to be excluded.2.2. Anesthetic and Surgical ProcedureAnimals were received and placed in individual accommodation 24 h before surgery with water and food ad libitum. The fasting time was 8 h of solids and 2 h of liquids. No animals received any medications prior to surgery. All the anesthetic procedures were performed by the same veterinary anesthesiologist. After placement of a cephalic 18 gauge 1” catheter, the anesthetic protocol consisted of premedication with 5 mg/kg IM alfaxalone (Alfaxan® 10 mg/mL; Dechra Veterinary Products SLU, Barcelona, Spain), applied at two different points to avoid patient discomfort, in combination with 0.1 mg/kg IM midazolam (Midazolam Hospira® 5 mg/mL, Roche, Madrid, Spain) and 0.2 mg/kg IM morphine (Morphine Chloride 2%®, B-Braun, Melsungen, Germany). Induction consisted of alfaxalone administered through the cephalic catheter and titrated to effect until tracheal intubation was possible with an appropriately sized cuffed endotracheal tube. Maintenance was conducted with sevoflurane (Sevoflo®, Esteve Veterinary, Barcelona, Spain) vaporized in a mix of 60% oxygen and 40% air, as needed. During the surgical procedure, all animals received IV fluid therapy with Ringer-Lactate solution (5 mL/kg/h, Braun®, Barcelona, Spain). Intraoperative monitoring consisted of body temperature, electrocardiogram, capnography, pulse oximetry, noninvasive blood pressure and oxygen and vapor concentrations. Intermittent positive pressure ventilation (VPPI) was applied when necessary to ensure normocapnia, and the volume was regulated to keep the end-tidal CO2 at normal levels (35–45 mmHg) [28]. Post-operative care consisted of hospitalization during the first 24 h and analgesia receiving morphine (0.2 mg/kg/6 h for 24 h).All surgeries were performed by the same surgeon with the help of an assistant. Thirty minutes before starting the surgery, a single dose of cefazolin was administered to all animals (20 mg/kg/IV, Cefazolina Normon®, Normon laboratories, Madrid, Spain) [29]. The surgical area was clipped and scrubbed. The distance from xiphoid to pubis was measured to calculate the size of the incision (20% of this length) and to standardize the surgical wound. Surgery was performed by a horizontal right flank approach, according to a published procedure [30]. A single dorsoventral incision in the skin was made on the right flank of the animal, starting caudal to the midpoint between the last rib and iliac crest. The subcutaneous tissue was dissected and the fibers from the muscle bellies of the external abdominal oblique, internal abdominal oblique and transverse abdominal were opened separately, according to their fiber direction, as described in the literature. The suspensory ligament was sectioned, the ovarian artery and vein and the uterine artery and vein were double ligated, and the uterus excised. For closure, muscles were sutured in two layers with the transverse and internal oblique muscle together, and the external oblique muscle as a second layer, as described in the literature [30]. Subcutaneous tissue was sutured with a simple continuous pattern, and the skin was closed with an intradermal pattern. For all layers, an absorbable monofilament glyconate suture was used (Monosyn®, B. Braun VetCare SA, Barcelona, Spain). No blood loss was observed in any of the animals and no post-operative complications occurred.2.3. Blood SamplingBlood samples were obtained from patients at different study times: before anesthesia (T0), at 30 min (T1), 60 min (T2), 3 h (T3), 6 h (T4), 12 h (T5), 24 h (T6), 3 days (T7) and 7 days (T8) after the anesthetic and surgical procedure. A total volume of 1.5 mL of blood was extracted from each animal, of which 0.5 mL was anticoagulated in EDTA (Tapval®, Barcelona, Spain) and 1.0 mL in lithium heparin (Tapval®, Barcelona, Spain) for hematological and biochemical analysis, respectively. The anticoagulated samples in lithium heparin were centrifuged at 3000 rpm for 10 min (Centronic model centrifuge, Barcelona, Spain) to obtain plasma. All samples were analyzed immediately after extraction.2.4. Analytical MethodsBoth hematological and biochemical analyses were performed at different study times. The hematological parameters included were: red blood cell count (RBC × 1012/L), hemoglobin concentration (HB; g/dL), hematocrit (PCV; %), medium corpuscular volume (MCV; fL), mean corpuscular hemoglobin (MCH; pg), mean corpuscular hemoglobin concentration (CMCH; g/dL), erythrocyte distribution width (RDW; %), reticulocytes (RET;/µL); white blood cells (WBC; ×103/L), neutrophils (NFS; ×103/L), lymphocytes (LYMPH; ×103/L), eosinophils (EOS; ×103/L), monocytes (MON; ×103/L), and platelet counts (PLT; ×103/L), plateletcrit (PCT; %), mean platelet volume (MPV; fL) and platelet distribution width (PDW; %). The parameters studied in the serum biochemistry were: total plasma proteins (TPP; g/dL), albumin (ALB; g/dL), globulins (GLOB; g/dL) and glucose concentration (GLU; mg/dL); alanine aminotransferase (ALT; UI/L), alkaline phosphatase (ALP; UI/L), aspartate aminotransferase (AST; UI/L), gamma glutamyl transferase (GGT; UI/L) and creatine kinase (CK; UI/L) activities; total bilirubin (TBIL; mg/dL), triglycerides (TG; mg/dL), cholesterol (CHOL; mg/dL), calcium (Ca; mg/dL), phosphorus (P; mg/dL) and sodium (Na; mEq/L), chloride (Cl; mEq/L), potassium (K; mEq/L) as electrolytes concentrations. Hematological determinations were carried out with the cell counter CELL-DYN Emerald (Abbot, Hanover, Germany) and for the analysis of biochemical parameters, the automatic system Spin 200E was used (Spinreact SAU, Barcelona, Spain).2.5. Statistical AnalysesStatistical analyses were conducted using SPSS (IBM SPSS Statistics for Windows, Version 23.0, IBM Corp, Armonk, NY, USA) statistical software. To analyze the existence of significant differences between the times considered (T0, T1, T2, T3, T4, T5, T6, T7 and T8), the Kruskal–Wallis test was used. By means of the Mann–Whitney test, significant differences among times were assessed. Results are given as mean ± SD. A p-value < 0.05 is considered significant.3. ResultsThe mean age, mean weight and mean body condition of the animals included in the study were 5 ± 2 years, 20.3 ± 7.2 kg, and 4–5 (according to the 9-point scale), respectively [31]. The required dose of alfaxalone was 3.1 ± 0.9 mg/kg. Compared to baseline levels, the RBC count and PCV decreased significantly from T1 to T6 (p < 0.05) and the HB concentration decreased from T4 to T6 (p < 0.05; Table 1). The WBC and NFS count increased significantly from T3 to T6 (p < 0.05), and the MON count increased significantly from T2 to T5 (p < 0.05). In addition, a significant increase in EOS count was seen at T5 (p < 0.05; Table 1). Compared to T0, the PLT count and PCT significantly decreased at T5, followed by a significant increase from T6 to T8 (p < 0.05), and the PDW significantly increased from T3 to T6 (p < 0.05; Table 1). No significant differences were observed in MCV, MCH, MCHC, RDW, RET, LYMPH and MPV, ranging between 61.1–73.8 fL, 18.2–22.3 pg, 29.0–33.0 g/dL, 11–15%, 0.4–8.9/µL, 1.10–7.40 ×103/L and 3.80–9.10 fL, respectively.Regarding biochemical parameters, GLU concentration increased significantly at T2, remaining elevated until T4. A progressive and significant decrease in BUN levels from T1 to T7 was observed, followed by a progressive increase until the end of the study, the time when BUN reached similar values to the baseline. The TG concentration decreased significantly from T2 to T4 (p < 0.05; Table 2), without subsequent changes. Plasma CK activity increased significantly after T5, remaining elevated until T7 (p < 0.05; Table 2). Compared to baseline values, p concentration increased significantly from T2 to T3 (p < 0.05; Table 2). No differences were found in TPP, ALB, GLOB and the ALB/GLOB ratio, CHOL, CREAT, ALP, GPT, TBIL, Ca and electrolytes (Na, K and Cl). TPP, ALB, GLOB concentrations and the ALB/GLB ratio ranged between 5.24–8.48 g/dL, 2.63–4.90 g/dL, 2.15–6.43 g/dL and 0.28–2.12, respectively. CHOL and CREAT concentrations, ALT and ALP enzyme activities and TBIL concentrations fluctuated from 79–291 mg/dL, 0.49–1.61 mg/dL, 20–82 IU/L and 38–202 IU/L and 0.1–0.9 mg/dL, respectively. The concentrations of Na, K, Cl and Ca varied between 139–157 mmol/L, 3.0–5.6 mmol/L, 106.0–120.0 mmol/L and 7.20–11.6 mg/dL, respectively.4. DiscussionIn this study, RBC and PCV levels decreased from 30 min to 24 h, also the HB concentration decreased from 6 to 12 h after surgery, followed by an increase to baseline levels. Similar results in relation to RBC and PCV were obtained in other studies [9,10,11,32] during intra and post-surgical times when different surgical and anesthetic protocols were used. The similarity between results suggests that erythrocyte parameters respond in a similar way despite the anesthetic and surgical protocols used in bitches that underwent OHE.Decreased RBC and PCV can have different origins. First, it could be due to the accumulation of blood in the spleen after the administration of anesthetic agents [33]. Although the application protocol differs considerably from that used in this study, the administration of a single IV bolus of alfaxalone resulted in a significant increase in canine splenic volume from 0.17 L at baseline to 0.24 L after 15 min, and 0.23 L after 30 min. In the same study, there was a decrease of PCV from 46.3% at baseline to 40.6% after 15 min and 41.7% after 30 min [34]. Splenomegaly, due to alfaxalone administration in dogs, could be secondary to changes in smooth muscle tone and changes in systemic blood pressure and cardiac output, which alters the blood flow and blood loss [35] as well as to hemodilution in response to fluid therapy to preserve the blood flow in vital organs (brain, heart, liver and kidney) at the expense of other organs such as the skin and pancreas [10,36]. Second, the decline of PCV could also be the result of the sequestration of RBC in nonsplenic sites [37]. It is important to note that many events that commonly occur during surgery, such as blood loss, tissue hypoperfusion/ischemia, hypoxia and intraoperative fluid therapy, among others, can alter hematology, and should not be ignored. On the contrary, other studies showed no significant differences among erythrocyte parameters along OHE in bitches [12].A temporary increase in WBC and NFS from 3 to 24 h, and in the MON count from 60 min to 12 h after the beginning of the OHE, without modifications in LYMPH counts, was observed in this study. These results partially confirm those obtained by other researchers using different anesthetic protocols and the same surgical technique. One study reported similar intraoperative increases in WBC and NFS, with LYMPH decrease [32]. The intraoperative and post-surgical neutrophilia and lymphopenia might have been the result of the stress caused by the anesthetic drugs, surgical trauma and the subsequent stimulation of the adrenal cortex [38]. It is well documented that OHE inflicts pain and stress as a result of tissue trauma, organ manipulation and inflammation [32]. Hancock et al. [39] reported a significantly higher peak of plasma cortisol levels in dogs 2 h after OHE. In response to the stress hormone, NFS shifts from the marginated to the circulating NFS pool. This neutrophilia might also be enhanced by the release of NFS from the bone marrow storage pool and the decreased migration of NFS to the tissues [40]. Although the transient mature neutrophilia in this study could be related to stress, no lymphopenia and eosinopenia were observed at any study time.Contrary to our results, other studies revealed no significant decrease in WBC [12] or a slight decrease for a short time during anesthesia, after OHE in bitches [11]. It is known that the administration of α2-agonists suppresses the circulating catecholamines, which exerts a modulatory effect on leukocyte subpopulations [41]. In addition, dissociation agents also reduce leukocyte counts [42]. For example, ketamine and butorphanol produce an analgesic effect, reducing the response to stress by reducing cortisol and adrenaline and, consequently, the number of leukocytes. Reports of a decreased WBC post-operatively compared to the baseline, with negligible changes in NFS counts, are also available. These insignificant changes in NFS can be attributed to the effects of dexmedetomidine, which directly (by inhibiting the neuroendocrine response) or indirectly (through sedation and analgesia) obstruct the stress response when administered systemically. However, decreased WBC has also been related to hemodilution [10].PLT count and PCT decreased at 12 h, and increased from 24 h to 7 days, whereas PDW increased from 3 to 24 h, without changes in MPV. The investigations regarding perioperative PLT counts in small animals are scarce and controversial. In some studies, there are higher PLT counts after OHE than pre-OHE in bitches [9]. PLTs are an acute-phase reactant that increases in response to systemic inflammatory processes because of increased pro-inflammatory cytokines, such as interleukins (IL-1, IL-6 and IL-11). Thus, the increase in PLTs and PDW observed in this study could be related to the systemic response and bleeding secondary to OHE. In addition, during the surgical inflammatory process, the release of activated platelets increases [43]. On the other hand, after sedation with acepromazine and atropine, a significant reduction in the PLT count and PLT aggregation capacity in response to ADP has been observed, without clinical signs of hemorrhage [44]. The decrease in the PLT count has been related to hemodilution during anesthetic procedures, due to vascular pooling as a consequence of vasodilatation or sequestration of blood cells in the spleen and lungs during the procedure [45]. Finally, one study reports no changes in PLT between pre- and post-operative counts in bitches undergoing OHE [46].In this study, plasma GLU increased at 60 min, remained elevated until 3 days, and decreased thereafter. Similar results have been documented in dogs undergoing OHE [12,47]. The maximum GLU levels in the latter studies were observed 1 h after anesthetic premedication, and levels were back to baseline after 24 h post-anesthesia. In other studies, using a protocol consisting of medetomidine, carprofen, propofol and isoflurane for OHE, a significant increase in GLU levels was found at 1 and 6 h, and a decrease in GLU values at 24 h [48]. Interestingly, stabilization of GLU levels in the present study occurred later than in previous studies. However, in bitches anesthetized with dexmedetomidine and subjected to conventional OHE through the ventral midline, GLU levels increased significantly over the baseline values in the post-operative period [36].It is documented that most anesthetic protocols induce a hyperglycemic state [49,50,51,52,53]. This hyperglycemia varies according to the mechanisms of action of the different anesthetic drugs. Thus, xylazine and α2-agonists (medetomidine and dexmedetomidine) cause insulin suppression and/or stimulation of glucagon release [54]; propofol induces decreased GLU transport and GLU utilization by tissues, impaired insulin, and increased adrenocortical hormones in dogs [55]. In addition, there is a failure of the usual cellular response to insulin (insulin resistance), which occurs in the perioperative period [49].OHE is considered a moderate to severely painful and invasive procedure [56,57]. Surgical trauma leads to a stress response and increases muscular activity caused by damage to superficial nerve endings [58], stimulating the corticotropic releasing hormone (CRH). Following the stimulation of the hypothalamus, CRH stimulates adrenocorticotropin hormone (ACTH) and induces the adrenal gland to release cortisol. Furthermore, both the surgical stress and the anesthetic protocol lead to the altered endocrine secretion of insulin antagonists such as growth hormone [51] The latter, together with cortisol, trigger glycogenolysis in muscles and liver [59]. Mild-to-moderate stress hyperglycemia is protective because it provides a source of fuel for the immune system and the brain at a time of stress [60]. In addition, the activation of the sympathoadrenal system releases adrenaline, noradrenaline and glucagon. Delayed GLU metabolism and utilization, gluconeogenesis, lipolysis and insulin resistance culminate in an elevated plasma GLU concentration [8]. Intraoperative and post-operative hyperglycemia, in this study, could be due to the combined effect of the anesthetic and the surgical procedures. It is possible that alfaxalone may help induce a state of hyperglycemia. However, in other clinical procedures, no changes in blood GLU were detected when the same anesthetic was used [61].CREAT values did not increase in the initial intervals after the administration of the agents, although BUN concentrations decreased from 30 min to 3 days [10]. Similar results for CREAT were observed for the same surgical procedure in the literature [32]. BUN levels progressively decreased from 30 min to 3 days, increasing progressively until the end of the study. The early decrease in BUN in these animals could be related to a lower protein intake prior to surgery or increased IV fluids. Taking into account that neither CREAT levels nor electrolytes (Na, K, Ca and Cl) concentrations change in the present study, any sign of renal dysfunction was ruled out. However, compared to the basal level, P concentration increased significantly from 60 min to 3 h, although none of the values were above the reference range. Hyperphosphatemia may result from reduced renal excretion secondary to decreased glomerular filtration rate (GFR) and is seen in cases of azotemia, which was not the case in this study [62].TG concentration decreased from 60 min to 6 h, without modifications in CHOL. Ketamine-xylazine administration in dogs has been shown to induce a decrease in TG before and during anesthesia, related to falls in cholinesterase activity [63]. On the other hand, hepatic TG synthesis decreases by 50% in rats during oophorectomy [64]. Conversely, Bilen [65] and Chagas et al. [66] showed an increase in TG in dogs after OHE without modifications in CHOL. This elevation of TG is considered a physiological and transient condition that decreases at 7–12 h and could be due to food restriction prior to surgery.CK activity increased almost 5-fold from pre-surgery values from 12 h to 3 days, subsequently decreasing to the baseline. CK is an enzyme found predominantly in skeletal muscle, and significantly elevated serum activity is largely associated with muscle damage. The marked increase in plasma CK in this study could be explained by the muscle damage produced during the flank approach. Earlier increases of this enzyme have been reported from 4 to 12 h [67] and, also at 48 h [68] post-OHE, and are probably proportional to the extent of muscle injury. The sustained increase in CK to 3 days in this study leads the authors to believe that CK is indeed a factor that can help measure stress and pain in these animals, despite the fact that OHE through the linea alba has been shown to cause minimal muscle trauma [67].TPP, ALB, GLOB and the ratio ALB/GLOB did not change through the study time Similar results were obtained by Del Romero et al. [28] in dogs submitted to an ovariectomy (OVE) using a similar anesthetic protocol when studying ALB levels at 1, 24, 72 and 168 h post-intervention. However, hypoproteinemia and hypoalbuminemia have been reported to be the most notable changes in response to surgery. Indeed, the type and duration of surgical trauma [69,70] are related to increased protein catabolism in response to inflammation and/or infection. Both phenomena led to the production of immunoglobulins and acute-phase reaction proteins [71]. Furthermore, a decreased production of ALB by the liver and an increased vascular permeability may affect ALB levels [72]. The stability of TPP, ALB and GLOB could be explained by the good hydration status of the animals during surgery. In the same way, the absence of changes in the enzyme profiles of ALT, AST and TBIL determined, rules out the presence of possible hepatobiliary damage. The fact that the ALP levels did not change could be explained by the absence of bone or kidney damage due to the analgesic and anesthetic protocols used in this study.A strength of this study was the fact that samples were obtained from healthy bitches with similar clinical histories, following the same anesthetic protocol and without post-operative complications, which allowed the authors to obtain representative and consistent results. However, a limitation of the study could be that some events that can occur during OHE and that are difficult to assess (such as the quality of tissue trauma, short stages of local hypoperfusion/ ischemia, or the release of some catecholamines, among others), could potentially affect some of the hematochemical results reported here.5. ConclusionsIn conclusion, conventional OHE in bitches using an alfaxalone-midazolam-morphine-sevoflurane protocol induces limited changes in certain hematological and biochemical parameters. Hematochemical analysis is a sensitive tool used to monitor the health status and surgical stress of the patient, as well as to detect any possible intra- or post-surgical complications.

animals : an open access journal from mdpi

10.3390/ani12060776

PMC8944486

Cattle are large animals that can cause serious injuries to humans. Humans may encounter cattle through working on farms, living on a farm, or traversing fields with cattle. A systematic review was carried out to assess the factors which may lead to a dangerous interaction with cattle. A literature search was carried out to find papers that included the criteria ‘Bovine’, ‘Handling’, ‘Behaviour’ and ‘Safety’, or terms therein. The search returned 17 papers, and after collation, six themes were identified: actions of humans; human demographics, attitude, and experience; facilities and the environment; the animal involved; under-reporting and poor records; and mitigation of dangerous interactions. Exploration of these themes shows that more accurate recording of interactions before an injury is required. Furthermore, targeted, tailored education for anyone who may come into contact with cattle could reduce cattle-induced injuries.

Cattle production necessitates potentially dangerous human–animal interactions. Cattle are physically strong, large animals that can inflict injuries on humans accidentally or through aggressive behaviour. This study provides a systematic review of literature relating to farm management practices (including humans involved, facilities, and the individual animal) associated with cattle temperament and human’s on-farm safety. The Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) was used to frame the review. Population, Exposure, and Outcomes (PEO) components of the research question are defined as “Bovine” (population), “Handling” (exposure), and outcomes of “Behaviour”, and “Safety”. The review included 17 papers and identified six main themes: actions of humans; human demographics, attitude, and experience; facilities and the environment; the animal involved; under-reporting and poor records; and mitigation of dangerous interactions. Cattle-related incidents were found to be underreported, with contradictory advice to prevent injury. The introduction of standardised reporting and recording of incidents to clearly identify the behaviours and facilities which increase injuries could inform policy to reduce injuries. Global differences in management systems and animal types mean that it would be impractical to impose global methods of best practice to reduce the chance of injury. Thus, any recommendations should be regionally specific, easily accessible, and practicable.

1. IntroductionWorking with cattle is widely reported as a major cause of human injury on the farm [1,2,3]. This is a global problem, with cattle cited as the cause of injury in 127 hospitalisations in the midland region of New Zealand over a five-year period [4] and 221 cattle-induced injuries reported over a seven-year period in the United States [5]. In England, one hospital reported sixty-seven patients that were admitted over a five-year period following cattle-related accidents [6]. However, it is difficult to quantify actual numbers of cattle-induced injuries, and hospital admissions under-report as cattle-induced injuries are often treated by a General Practitioner [4]. Cattle-related injuries can be severe, with 24 people reported to have died from injuries sustained from cattle between 2015 and 2020 in Great Britain [7]. In the Republic of Ireland, cattle were in the top five triggers for fatal accidents at work, accounting for seven fatal accidents [8]. Collating data on injurious interactions with cattle is difficult because official statistics often group injuries or fatalities, and as a result data is low resolution and can be difficult to interpret. Official statistics report numbers of fatalities or reported injuries, but do not specify the behaviour of the animal or human before the injurious event [7,8]. In order to prevent injuries and potential fatalities due to livestock, it is important to identify contributing factors that may result in cattle reacting in a dangerous manner.Livestock handling is a complex interaction between people, animals, and the environment [9]. Research into the person involved, the handler, is limited and disparate [10]. The person involved can induce fear reactions in animals, in turn reducing productivity and welfare [11,12]. It is reported that animals that are handled negatively may react adversely and sustain injuries [11]. Negative consequences of fear induced by bad handling may reinforce the handlers’ negative attitudes and actions towards animals [13]. The human’s attitude affects the interaction and how the animal is handled [14], with handlers who had a better attitude behaving in a manner which positively influenced animal behaviour [15]. In addition to the human element, the animals’ behaviour, personality, or temperament will affect human–animal interactions. The animals’ general behaviour and reaction to humans can be defined as their temperament [16]. Temperament can be affected by environmental factors, such as previous handling experiences [17] and animal genetics [18,19]. Although temperament is a subjective trait that is difficult to measure, farmers are reported to be aware of differences in temperament between individual animals [20]. The third aspect is the environment, with some cattle handling injuries directly related to farm infrastructure [21]. Improving farm infrastructure can improve profitability and safety [22] and reduce stress in animals [23]. However, an American survey found that farmers can be reluctant to upgrade handling facilities due to the large investments of time and money required [21].Although the relationships between animal genetics with animal behaviour and farm management with animal behaviour are well researched, there are few studies on the specific human actions which may reduce human safety. This paper aims to identify the common factors in incidents where cattle have had an injurious or potentially injurious interaction with a human. Knowledge of the factors leading to injury can be used to direct policy and aid decision makers on farm to develop protocols to reduce the number of injuries to humans involving livestock. The objective of this paper is to conduct a systematic review of the literature with the aim of identifying specific human–cattle interactions and management processes that may increase the likelihood of unsafe interactions with humans on farms. This review is designed to address on-farm safety and includes all potential animal encounters on-farm (including visitors and on-farm workers); it will not include animals bred for sport or fighting. To ensure this review can fulfil the aim of informing policy, papers will be limited to those published since the year 2000. The occurrence of fatalities involving animals have shown a marked increase in the legislative area (Northern Ireland, United Kingdom) in recent years [24] so only technologies relevant to modern farm managers and modern animal handling environments will be included. Additionally, the cattle will be limited to Bos taurus species. Bos indicus and their crosses are reported to be more reactive than Bos taurus cattle [25]; this review focuses on Bos taurus as this is the most populous breed in the target legislative area. The outcomes of this review will be used to develop guidelines and identify where improvements in animal handling need to be made.2. Materials and Methods2.1. Eligibility CriteriaIn order to undertake a systematic review, it is necessary to identify the Population, Exposure, and Outcomes (PEO) components of the research question [26]. Expert opinion from authors was considered, and a list of keywords relevant to the topic was developed. Preliminary searches of the Web of Science database using the keywords were carried out to assess the number and quality of papers returned. Search terms were discussed and agreed upon by all the authors and subsequently grouped into search categories. The population was defined as “Bovine”, with an exposure of “Handling” and outcomes of “Behaviour” and “Safety”. The search could not be pre-registered as the outcomes did not directly impact human health and so the search was not within scope for a designated public repository.2.2. Information SourcesThe procedure for undertaking this review was designed following the PRISMA framework [27]. Three databases were identified as suitable for searching: PubAg, Web of Science, and the ‘EBSCO Academic search elite’ option within the research database of EBSCO. Searches were limited to papers published between the year 2000 and the date of search (June 2021).2.3. Search StrategyThe following algorithm of keywords was designed, where results must have at least one search term from each PEO component: (cattle or cow or “steer” or heifer or bull or dairy or beef or herd or “Bos taurus”) AND (hand* or attitude or manage* or farmer or human or automat*) AND (temperament or excit* or aggressi* or fear* or docility or flight or “exit score” or “exit time” or “exit speed” or “exit velocity” or “chute score” or “strain gauge” or “movement measuring device” or personality or “coping style” or boldness or proactive or reactive) AND (safe* or injur*). An asterisk denotes a truncated term with ‘wildcard’ which may represent one or multiple characters, and inverted commas instruct the search engine to find an exact match to the term. Searches were conducted by a single author and validated through an independent replication of the search by another author.2.4. Selection ProcessIn total, 694 references were returned and exported as .csv files. Fifty-two duplicate texts were removed, then the process of title and abstract screening was carried out independently by two authors. Two articles were conference abstracts that could not be obtained through the institution’s current subscriptions, through interlibrary loans, or after contacting the original author and thus were discarded. General narrative reviews were discarded but used as sources of potentially relevant studies. Eligibility criteria was set whereby the studies must fulfil the following: (1) have Bos taurus as subject; (2) include farm management/human interaction; (3) have an animal response that can potentially negatively impact handler safety, i.e., a reaction which is identified as dangerous, threatening or risky; (4) be written in English (5); have the full text available. Papers were limited to those published in English as this was the working language of the authors. Once the screening sift was complete, any differences in opinion between the two reviewers were discussed and resolved. The number of full texts for eligibility assessment was 48. Of these, 17 relevant papers were included in the review. A full breakdown of the sifting process is shown in Figure 1.2.5. Synthesis of ResultsOnce suitable papers were identified, they were tabulated to compare study type, design, animals used, and the outcomes (Supplementary Table S1). The journal ranking of each publication was not considered an appropriate measure of credibility as this reflects the quality of the journal rather than the articles therein [28]. Due to the publishing dates of the papers included (ranging from 2009 to 2021), it would not be possible to accurately compare citation count. As a result, papers were evaluated by two authors independently using the Critical Appraisal Skills Programme (CASP) checklists [29]. It was not possible to collate results for meta-analysis as the data were heterogeneous and information from different papers could not be directly compared. Data were summarised by author, year, geographical location, animal breed type, sample size, measurement of exposure, study aim, findings and recommendations, and any limitations of the study. Once data was collated, a thematic analysis was carried out which identified similar themes and topics within the study aims, measurements of exposure, and findings.3. Results and Discussion3.1. Study SelectionThe principal reason for discarding papers was that the subject of the study was not Bos taurus, this included studies pertaining to Bos indicus or other species of animal which used the noun “Cow” for females. A large number of food supply chain studies were returned in the search due to the inclusion of the terms “dairy”, “beef”, and “safety” in the protocol. Papers which recorded either injury or handling but did not record any interaction between the two. Two studies were removed because they were not studies of farmed cattle, the subjects of these studies were cattle bred for rodeo or fighting.3.2. Study CharacteristicsThe 17 papers identified as meeting the criteria were appraised and considered suitable for inclusion. The papers predominantly assessed dairy breeds (n = 8), with four papers assessing beef cattle and five papers were undefined breeds within the Bos taurus species. Three non-narrative review studies were included which assessed media reports [30], online records, and on-farm injury reports [31] or hospital records [32]. These reviews were considered suitable for inclusion as they were non-narrative reviews that collated public records and followed a search protocol. Five papers were qualitative assessments of opinions gathered through interviews or focus groups. Participants included farmers who had participated in safety campaigns [33], used organic dairy systems [34], had a specific breed of cattle [35], veterinarians who carried out castration procedures [36], and immigrant workers who worked with dairy cattle [37]. Four studies were observational studies that monitored maternal temperament of beef cattle [38], habituation of primiparous cows to milking procedures [39], environmental effects such as restraint [40], and outdoor exercise [41] and their effects on the human–animal relationship. The remaining five studies combined observations of animal behaviour and questionnaires or surveys of farmer opinions [42,43,44,45,46]. Two of the articles were by the same author and on examination of the data were found to be two different analyses based on the same study, which included a small number of dairy farms [44,46]. The studies had a wide geographical spread, but were predominantly European (n = 10), with four North American studies and one Indian study. Two studies included reviews of international data. All were published between 2009 and 2021.3.3. Summary of EvidenceThe 17 papers identified a range of management practices that could affect, either positively or negatively, the occurrence of risky animal interactions and explored how both human actions and attitude can affect the occurrence of cattle attacking. An important aspect was the human involved and their behaviour through acting calmly, having a positive attitude to both cattle and risk, and assigning adequate time and facilities to undertake tasks on the farm. An animal’s propensity to attack could be affected by the animals’ previous experiences, inherent traits for disposition, and previous experiences, with human–animal interactions that were positive and not stressful for the animal. These were organised into six themes: actions of humans; human attitude and experience; facilities and the environment; the individual animal; underreporting; and current forms of mitigation.3.4. Themes3.4.1. Actions of HumansFarmers feel that they can positively impact animal behaviour through their own actions. Farmers reported that staying calm, using personal experience, and talking to the animals could facilitate cattle handling [43], and provisioning enough time for a task can reduce the likelihood of injury [37]. Positive experiences for the animal during human interactions, for example, the human providing concentrates directly to the animal rather than using a machine, resulted in significantly calmer cows measured by Qualitative Behaviour Assessment [42]. When working with animals, it is important to consider how they will react in a procedure that may be averse. A survey found that 37% of US veterinarians considered the risk of injury to the operator critically important when performing castration, a painful procedure [36]. An increase in cows slipping, falling, or cow behaviours indicating fear (collectively referred to as risk situations) occurred more when dairy cows were moved to hoof trimming than to milking [44,46], which may be because hoof trimming is a more aversive procedure. However, it can be difficult for farmers to judge how aversive an animal may find a procedure. When farmers were asked to rate their agreement (on a scale of 1 to 5; where 1 is “totally disagree” and 5 is “totally agree”) to the statement “animals experience physical pain as humans do” there was a wide variation in responses [45].Despite the causal relationship not being clear, quiet handling and minimal talking are recommended to reduce risk situations with cattle [46]. Forceful tactile interactions using an object should be avoided as they were found to elicit fear reactions from cattle and could be counterproductive, slowing down procedures by causing the animal to freeze [46]. Human interactions which led to potential injury included pulling the halter, which was positively correlated with being head-butted, and use of forceful interactions such as shouting, hitting, or tail twisting for an extended duration, which increased the likelihood of the handler getting kicked [46]. Time spent in the risk zone, i.e., within reaching distance of the cow, was correlated to the number of observed risk situations [46]. This increase was attributed to close proximity to the animal, leading to an increased risk of being injured by an unexpected response or reaction by the animal [44]. Conversely, farmers who made physical contact with cattle during monitoring tended to have herds with lower avoidance distances [43], suggesting that contact and close proximity to an animal can improve interactions.There was a paucity of literature on specific human actions which directly caused injurious interactions between cattle and humans. This is because the injury was reported by the injured party and generally did not clearly define the precise interactions which led to an injury [31,32] and may have created bias. Resultantly, many of the papers considered animal reactions that were potentially injurious rather than those which caused an injury to humans. Without records of every occurrence of each potentially dangerous behaviour, it is not possible to estimate how often these reactions lead to an injury, thus it is important that the events leading to animal inflicted injury are accurately reported and recorded. Collating this data can allow thorough analyses of specific actions leading to cattle-inflicted injuries.3.4.2. Human Demographics, Attitude, and ExperienceHuman actions will be dependent on a range of factors, including their demographics attitude, and experience. A questionnaire amongst Swedish handlers that identified the handler’s risk tolerance found that risk-accepting handlers encountered significantly fewer risk situations per minute when moving cows to hoof trimming than risk-averse handlers [44]. However, no correlation between the handler’s attitude towards cows and risk situations was found [44]. In contrast, cows in herds with managers that agreed more strongly with the importance of positive animal contacts were significantly calmer than cows in herds with managers who did not believe this [42]. More evidence of the human attitude adversely affecting cattle behaviour was expected. In the papers reviewed, farmer attitudes were often assessed through a survey or focus group—these are qualitative studies that are difficult to validate [47]. Some respondent groups were limited to certain breeds [35] or management [34] and the results may be less transferable to different management systems. In all cases, the focus groups/interviewees agreed to take part in the study, meaning respondents may have been more progressive than those who refused, or otherwise not representative of the general population.No study could identify a strong association between the handler demographic and handler safety. Although the majority of reported bull attack victims were middle-aged males, it was noted that this was consistent with the predominant demographic of US farmers [31]. It was reported that handler demographics did not affect interactions with cows, however, this study had a small cohort (12 handlers), with a high proportion (75%) of males [44]. Although no associations could be proven, it was suggested that youths should not work with bulls as they lack the necessary maturity and strength [31]. In a survey of Swedish farmers, it was found that older farmers thought they could mitigate risks by being more careful, however, it was found that their younger counterparts did not think that older farmers were careful enough [33]. This is a theme throughout the survey, with farmers fearing risks to third parties such as children, the elderly, and workers on the farm rather than themselves [33]. The issue of communication between workers, particularly migrant workers, was highlighted. Managers may have communication difficulties with migrant workers, meaning that they have limited control over staff [33], however, a survey of migrant workers blamed managers for not educating migrant workers about risks on the farm [37]. This highlights that farm safety is an issue that is not only pertinent to the individual, but the farm manager and risk assessments and education programmes must also reflect this.Most injuries recorded are from people who worked with cattle, such as farmers and vets [30]. This aligns with the finding that injury risk is associated with hours of exposure to an animal [31]. Conversely, people who are visiting a farm and unexpectedly come in contact with cattle may have limited knowledge of how to behave around them and may inadvertently provoke an attack [30]. There were 54 cases of walkers being attacked by cattle in Great Britain between January 1993 and May 2013 [30], demonstrating that cattle can be a risk to anyone who may come in contact with them.It has been reported that farmers could interpret their cattle’s facial expressions and posture to establish when an animal may become an immediate risk [35]. However, this may not be sufficient to reliably estimate the risk of an attack, as it has been reported that bull attacks can occur without any visual behaviour to communicate a warning [31]. This is supported by a study of 15 bull attacks, in which all victims reported that their attacks were unprovoked and that animals become aggressive suddenly [32]. Overconfidence on the part of the handler may also be a danger [31]. An example of this overconfidence may be in children who are overfamiliar with the farm (and animals/equipment therein) and may not perceive risks [33].Attitude and behaviour can be improved through extension services and education [48]. The range of people who may come in contact with cattle includes individuals traversing farmland, living on a farm, and individuals working on farms. This list is not exhaustive and demonstrates the wide range of extensions required to reach individuals. Clear signage on farms and areas where cattle may be encountered could help alert visitors who are less familiar with cattle to the dangers. The extension should be targeted to suit the individual demographic, for example, farm safety days for children [49]. Multifaceted approaches for farm managers which include an environmental change and safety audit have been shown to be effective [50]. Another method targeted at farmers is a “fear appeal” which motivates a farmer to adopt safer behaviours by exposing them to a hypothetical threat situation and then providing information on how to mitigate the threat [33]. A survey to assess the role of fear appeals in motivating farmers to adopt safer practices found that for a fear appeal to work, the farmer must be able to identify a threat and carry out the mitigating action. Additionally, the farmer must believe that the danger is real and that it could happen to them [33]. It was reported that farmers are more likely to act on simple threats and that more complex threats which are harder to identify and/or had multiple causes may be ignored. It was reported that farmers thought that written rules were unnecessary and valued their own experience more than training [33]. Furthermore, farmers felt extension officers were inadequate and so may have ignored dangers highlighted by advisors, and instead of adopting the mitigation recommendations, the farmer-controlled his or her fear reaction by denying the existence of the threat [33].3.4.3. Facilities and the EnvironmentThe design of facilities differs amongst farms [45], and there was geographical variation and a range of management systems employed in the studies included in this review. Dairy cattle require different facilities to carry out milking and they may be housed either continuously or for part of the year [51]. Many beef cattle outside Europe are raised on feedlots, however, the facilities in these lots may vary due to climatic conditions and rainfall [52]. In cool temperate areas, cattle may be housed during wet winters and graze outside during summer. Thus, farm design cannot be uniform and must be tailored to the management system employed, geographical topography, or regional variation. Poorly designed or inappropriate facilities are known to be significant contributors to cattle-related incidents [31]. The facilities dictate how the human interacts with the animal, how much time is spent in the risk zone, and thus how much scope there is for an incident to happen [44]. However, it is poorly understood which specific aspects of the facilities are most important. In-depth on-site investigations of a representative sample of cases of attacks have been suggested to be necessary to clarify the contribution of different aspects [31]. Minimizing personal contact time may be key, especially when working with bulls, as most bull-related incidents occurred when the person entered the bull’s territory [31]. Using suitable facilities rather than personal contact to perform part of the handling may aid in this. For instance, automatic manure scrapers can be used to “gently get cows to stand and move” [46], a task that would otherwise require a direct approach by the handler. Different facilities may be required for different animal types. For example, due to their size and strength, bulls are often held in a bullpen, and it is recommended to only enter a bullpen when the bull is restrained [31]. However, to improve the human–animal relationship it is recommended that interactions with dairy cows take place while the animal is unrestrained [40].Apart from facilities directly designed to handle the animals, the comprehensive design of the farm and its management will also contribute to the risk of injury. Pastured dairy cows have been found to show reduced reactivity compared with cows in tie-stalls [41], potentially reducing the risk of injury. Conversely, beef animals that are grazed in an extensive system with little human contact are known to be more aggressive and evasive than cattle in more intensive systems [35,45].In addition to the physical arrangement of the facilities, other external factors, such as the presence of a dog, can impact the occurrence of injury. Cattle may perceive dogs as more threatening than humans [30]. The presence of a dog was identified as a risk factor and recorded in 64.8% (n = 35) of attacks on walkers traversing farmland [30]. The effect of how the human interacted with the dog was less clear, with attacks occurring in situations both where the human had picked up the dog or released it from its lead. Other studies mentioned that the presence of a dog was recorded, but did not include this in the subsequent analysis [43].3.4.4. The Individual AnimalThe animal breed was not often reported, however, bull attacks were predominantly inflicted by animals classified as dairy breeds [31]. Eight studies focussed on dairy cows, however, it is possible that this could be attributed to more intensive management on dairy farms compared with beef farms, offering more opportunities to observe handling. A focus group discussion found that farmers who bred Pyrenean cattle could not only distinguish the temperament of Pyrenean cattle from other breeds but could also identify variation in temperament between individual animals of the Pyrenean breed. Farmers who took part in the study attributed these differences within the breed to genetics, age, and sex [35]. A review of media sources and published literature reported that bulls caused more injuries than cows [30], and one study focussed exclusively on bulls because these are considered to be more likely to attack [31]. However, in addition to bulls, primiparous and freshly calved cows are perceived as very dangerous to work with by the staff of American dairies [37]. Whilst one author suggested that injuries caused by cows were more likely to be due to ‘nonintentional’ behaviour than targeted attacks [31], cows with calves are likely to exhibit maternal aggression when they perceive a threat to their calf [30]. Although maternal aggression was identified as a risk factor, it is seen as a desirable trait in extensive management systems where a cow may have to defend its young from predators [35].Although there were no clear associations between sex and aggression, age was reported to affect animal aggression, with fewer attacks recorded from younger bulls [31]. Conversely, cows seem to become more docile with increasing age [35,37]. Despite this perception, maternal defensiveness of breeding beef cows when handled was reported to be a repeatable trait and remained consistent throughout different parities [38]. In addition to the inherent traits of the animal, the animal’s emotional state or how it has been previously treated can also affect the likelihood of an incident. The deviation from baseline heart rate was significantly greater when moving cows to hoof trimming than to milking, implying that it is more stressful for the animal [46], and more risk situations were observed during this apparently more stressful procedure [44]. Together, this suggests that stressed cows are potentially more dangerous. Hoof trimming is a less regular occurrence which cows will be less accustomed to than milking. Training cows to cope with aversive procedures can reduce fear levels. This can reduce stress for the animal [46], thus reducing the likelihood of behaviours that may result in injury. For example, heifers that had been introduced to the milking parlour 10 days prior to calving had a tendency to kick less often than ones that had not been introduced during their first post-partum milking [39]. Management or experiences during early life may impact an animal’s adult temperament. Certain actions in young bulls, such as rough play or teasing, can encourage aggression [31]. Similarly, offspring can inherit behavioural traits from cows through post-partum experiences and genetics [35].The identification and slaughter of aggressive bulls have been recommended [31]. This would not only remove the threat from the dangerous bull but also prevent further dissemination of any genetic predisposition to attack humans. However, survey data shows that at least some farmers continue to breed with potentially dangerous bulls [33]. None of the papers examined genetic effects, however, farmers who claimed to include behaviour in their selection criteria had cattle with lower avoidance distances [43], indicating cattle were less reactive towards humans. This took the form of a semi-structured interview and did not state the specific behaviour selected for [43].3.4.5. Under-Reporting and Poor RecordsDespite the aforementioned importance of handling facilities and environment, the location of incidents is often reported as “Farm”, making it difficult to identify the specific facilities or location involved in a bull attack [31]. Additionally, the accuracy of any existing records depends on incidents being reported. Interviews with immigrant workers on US dairy farms found they were often afraid to report incidents as they may lose their jobs and risk deportation [37]. This, coupled with the one-sided reporting where the interaction is not observed but reported retrospectively, may lead to sensationalism. This is exemplified by media reports portraying the human involved as blameless and cattle depicted as unprovoked instigators of the attacks [30]. Many incidents are believed to be unreported if not fatal [30,31]. In a review of 287 incidences, it was found that 57% of the reported attacks resulted in a fatality, the rest had varying degrees of injury, and few had no injury reported [31]. To calculate the risk of an attack when encountering cattle, both the number of attacks and the total number of interactions are required, but full records for both are usually unavailable [30], with similar poor records of on-farm encounters [29]. In the absence of this information, the creation of a central database to encourage self-reporting of cattle attacks is recommended [30]. It is important to note the under-reporting may have led to publication bias, whereby interactions that led to smaller injuries are not included in the reviews. Furthermore, there is disparity between an injurious event and the reporting, with injury and fatality data often not providing information on why livestock-handling injuries occur [53]. Linking injury to an event on-farm is made more difficult as there is no internationally accepted reporting method [2].The review papers are only representative of the records they reviewed, and both stated that records available for review may not have reported all incidents [30,31]. Media will be more likely to publish incidents where an injury has occurred, with ‘near misses’ not reported on. There is also the likelihood that incidents reported retrospectively by the human may not be correct as the human involved will not want to admit they behaved irresponsibly. The main focus of a review of hospital admissions was the outcome of the attack [32]. In all cases, patients reported that the bull attacks were unprovoked, but there is no evidence to support or oppose this. It is possible the injured party did not accurately describe the interaction leading to injury, or they may have been unable to interpret the cattle’s behaviour as threatening.The papers in this review were limited as they depended upon retrospective reports, farmer opinions, or small studies. To truly understand all factors which lead to cattle-induced injuries, it would be necessary to undertake multiple surveillance studies which include the observation of multiple human–animal interactions under various management systems and using different facilities. To undertake this surveillance at the necessary scale to record all potential human–animal interactions under all potential environmental conditions, the costs and manpower required would prove prohibitive.3.4.6. Mitigation of Dangerous InteractionsThere is a range of recommendations to mitigate bull attacks during the handling of cattle: ensure the bull wears a ring in its nose, have appropriate facilities, do not handle a bull older than one year old alone, ensure beef bulls are not crowded, and make sure to be wary of where a bull is at pasture [31]. These recommendations were taken from a single extension article published in 1981 by the College of Agricultural and Life Sciences, University of Wisconsin. Twenty-four separate web pages, which had guidelines for safe behaviour when encountering cattle when walking through fields of cattle, were reviewed [30]. These were collated and gave some contradictory advice such as “keep dog under control” and “let dog off the lead”; the National Farmers’ Union of Wales advised “Be bold and walk straight through the cattle if they come towards you”, whereas several other sources recommended skirting around the edge of the herd. In addition to the contradictory advice, some of the advice is less than practical, such as twisting the ring in a bull’s nose, as this would require being in very close proximity to the animal and not taking evasive action.Farmers were reported to feel that the presence of horns makes an animal more dangerous even if the individual animal is not aggressive [35]. Dehorning (‘polling’) cattle can reduce the damage resulting from the interaction. However, herds with a higher proportion of cows that had undergone polling were more fearful of humans with a higher avoidance distance and a higher proportion of fearful cows during tactile interaction [42]. A survey undertaken found that the main reason for dehorning was reported as fear of injuries to other animals (95%), farmers (84%) family members (48%), visitors (8%), or employees (5%) [34]. Twelve farmers reported injuries caused by cows with horns (five of which were injuries inflicted on humans, four of which were inflicted on other animals, three of which were encounters with dead animals).Cows are generally docile but are known to be dangerous when agitated [46]. Despite this, focus groups discussions found that only 50% of farmers agreed with the statement cattle are dangerous [43]. Risk management procedures recommended for the protection of third parties often concentrate on controlling cattle, with likely detrimental effects on the farmer. Suggestions to reduce injury to walkers included removing herds from publicly accessed fields, or euthanasia of reactive cattle with subsequent cost and management implications for farmers [30]. The review of cattle attacks on people traversing farmland recommends that bulls that attack should be killed to prevent further attacks [30]; this would incur a cost for the producer. When hazards were reported by migrant workers, the manager often did not act [37]. Health and safety regulations are in place for working on the farm, however, some farmers fear the introduction of additional regulations as they may lead to unnecessary costs [33]. Farmers were reported to be more afraid of risks to others, but more likely to take action on threats that were personal (i.e., affected the farmer themselves), and to blame others for being irresponsible [33]. Although the measures are taken to mitigate risks to farmers, their staff and third parties on the farm may be seen as costly to the farmer. It is necessary to manage cattle to prevent the risk of injury for the handler’s safety, but also the potential compensation costs if staff are injured [37].3.5. Limitations of This ReviewThis study concentrated on human interactions and management which led to behaviours in cattle that could potentially cause injury. The search was targeted at papers published between 2000 and 2021 and concentrated on Bos taurus animals. Although broadening the search would have increased the number of papers for review, it would not have provided contemporary information for the farm systems and cattle relevant to the production systems in the legislative area targeted by the policy the paper aims to inform. Four papers were discarded as they were not published in English or the full text could not be accessed, however, the abstracts were available. Of the two non-English texts, one was a German study that carried out a questionnaire with 317 farmers, of whom 25% reported injuries caused by bulls [54], however, the information could not be assimilated into this review in the absence of the full text. The other paper discarded due to language may not have been relevant as it surveyed workers on the practices of dairy farms in Brazil [55], but the contents of the survey could not be assessed as to whether it would involve animal interactions or not. The two texts which could not be found were not full articles but conference texts. This review was not limited to full papers and included conference texts as it can reduce publication bias [56], but they may be less reliable as they are not subject to peer review [56]. The inclusion of the discarded papers may have added value to this review, however, it is likely that they may have added to the discussion within one of the six themes rather than introducing a new theme. The six themes identified in this review were interlinked. It is not possible to quantify the importance of each theme, thus none of these themes can be considered in isolation and must be accommodated in the management system employed when working with cattle.4. ConclusionsA systematic review was developed searching for studies that identified how human actions and management systems could impact animal behaviour and increase the risk of injury to a person. Seventeen papers were identified which met the criteria. Six main themes were identified: actions of humans; human demographics, attitude, and experience; facilities and the environment; animal involved; under-reporting and poor records; and mitigation. The under-reporting of incidents, variation in management methods, and contradictory mitigation advice make it difficult to develop international standards of best practice. To develop appropriate policy to reduce injuries caused when working with livestock, it is recommended that a better understanding of causes of injury is attained either through observational studies or surveys which identify risk factors. To reduce cattle behaviours that can potentially induce injuries, it is recommended that standardised reporting and recording of incidents be introduced. Collation of injury data and the human actions resulting in injury will clearly identify the behaviours and facilities which increase injuries and so can inform choices to mitigate further incidents. Guidance for those working with cattle must be practicable, easily accessed, and there must be a consensus amongst providers to ensure that contradictory guidelines are avoided. This could lead to a specialized training programme for those working with cattle. Guidance must also be provided to anyone who may come into contact with cattle. This could be implemented through appropriate signage for those traversing farmland or visiting farms or extension through schools to target children living on the farm. It is further recommended that facilities that improve farm safety are installed and adequately maintained. These recommendations should be supported by including temperament and behaviour traits in breeding goals and the systematic culling of dangerous animals.

animals : an open access journal from mdpi

10.3390/ani11082394

PMC8388618

Campylobacter spp. has been the leading cause of human diarrhea in EU since 2005. Although poultry and poultry meat are considered as the primary source of transmission of campylobacteriosis to humans, pigs can be a significant reservoir of the pathogen, as well. Moreover, the increase of antibiotic resistance in the specific pathogen, especially against fluroquinolones and macrolides is considered a significant threat for public health. The purpose of the current study was to evaluate and molecularly characterize the antimicrobial resistance of Campylobacter infection in pig farms in Greece at both phenotypic and molecular level.

The purpose of this research was to characterize the antibiotic resistance patterns of Campylobacter spp. isolated from commercial farrow to finish farms in Greece, and analyze the relevant molecular resistance mechanisms among the resistant Campylobacter isolates. Susceptibility testing to five different classes of antibiotics was performed in 100 C. coli and 100 C. jejuni, previously isolated and identified. All isolates were found susceptible to meropenem. Very high rates of resistance were recorded for tetracyclines (84.5%), medium rates of resistance were recorded regarding quinolones (23%), and low and very low rates of resistance were identified for macrolides such as erythromycin and aminoglycosides (12% and 4%, respectively). Only 12.5% of the Campylobacter isolates displayed MDR. Regarding the molecular mechanisms of resistance, all ciprofloxacin resistant isolates hosted the mutant type Thr-86-Ile region of the quinolone resistance-determining region (QRDR) of the gyrA gene. In all erythromycin resistant isolates, the transitional mutations A2075G and A2074C in the 23S rRNA gene were only amplified. Molecular screening of tetracycline resistance genes indicated that the vast majority of Campylobacter isolates (92.3%) were positive for the tet(O) gene. In summary, these findings and especially the very high and medium rates of resistance for tetracyclines and fluroquinolones, respectively recommend that a continuous monitoring of Campylobacter isolates susceptibility in combination with the proper use of antimicrobials in livestock production is of great importance for public health.

1. IntroductionCampylobacter spp. are common pathogenic bacteria of both veterinary and human public health importance. They constitute the most common human gastrointestinal pathogens reported in EU since 2005 [1]. In 2018 in EU, the number of laboratory confirmed cases of human campylobacteriosis was 246,571, corresponding to a notification rate of 64.1 per 100,000 population. Within the same year, 524 campylobacteriosis outbreaks in total have been recorded in the 28 EU Member States, 522 of which were food borne whereas the remaining two were waterborne [1]. In USA, it is estimated that 2.1–2.4 million cases of human campylobacteriosis occur every year [2]. The most common sources of Campylobacter transmission are raw milk and chicken meat.Transmission occurs via the fecal-oral route after ingestion of contaminated food and water. The disease symptoms vary from a self-limiting watery diarrhea to a severe inflammatory diarrhea with abdominal pain and fever. Not infrequently, Campylobacter infections can be burdened with complications that can lead to chronic health problems. The main recognized sequelae after Campylobacter infection that can be triggered include Guillain-Barré syndrome (GBS) [3,4], reactive arthritis (REA) [5], and irritable bowel syndrome (IBS) [6].Management of human campylobacteriosis is based on fluid therapy, which is generally considered the therapeutic corner stone. Antimicrobial treatment is only required for patients presenting more severe disease clinical signs, as well as for those who are immunocompromised. The most common antimicrobial agents implicated in the treatment of Campylobacter infections are macrolides, such as erythromycin, and fluoroquinolones, such as ciprofloxacin [7]. Tetracyclines have been suggested as an alternative choice for the treatment of clinical campylobacteriosis, but are rarely used in clinical practice [7].During the past two decades, an increasing number of Campylobacter strains have developed resistance to fluoroquinolones and other antimicrobials such as macrolides. Moreover, the World Health Organization (WHO) identified Campylobacter as one of the high priority antimicrobial resistant pathogens, regarding its resistance to fluoroquinolones [7]. The resistance to both macrolides and fluoroquinolones is of major public health concern as it narrows therapeutic options for Campylobacter infections. Keeping this in mind, the EU continuously monitors the Campylobacter spp. prevalence and their resistance rates, in humans, animals, and food products. Therefore, it is considered as a public health priority.Members of the Campylobacter genus exhibit optimal growth when cultured at 42 °C. They are generally isolated from the intestines of cattle, sheep, swine, and the poultry caecum. Due to a higher body temperature, poultry as well as other avian species are among the most common edible animals hosting Campylobacter spp., representing the main source of infection for humans [8].Globally, Campylobacter jejuni is more prevalent in poultry, whereas Campylobacter coli is more common in pigs [9]. C. jejuni may co-exist with C. coli in pigs, but is typically detected in 10–100-fold lower levels than C. coli [10,11]. Pigs are considered as a natural reservoir of Campylobacter, exhibiting a prevalence of 50% and 100% with excretion levels ranging from 102 to 107 cfu/gr of feces [12,13]. In a consensus study conducted in Canada, amongst 1200 faecal samples examined, originating from 80 pig farms, 1.194 were positive for Campylobacter species. The prevalence of C. coli, C. lari, and C. jejuni were 99.2%, 0.6%, and 0.2%, respectively [14].Sows have been identified as the major source for piglet contamination. Piglets are usually infected within the first days after their birth and genotypic analysis has provided evidence that sows and piglets share similar profiles [15]. In an experimental study conducted by Young et al., newborn piglets exhibited an average incidence 57.8% of Campylobacter within the first 24 h after birth [16]. In another study by Alter et al. [13] in 15 pig farms, none of the 1-day old newborn piglets were positive for Campylobacter but the average prevalence in piglets increased within the first days of life to 32.8%. In the third week of age, the prevalence of Campylobacter positive piglets reached 41%, increasing to 56.6% at 4 weeks of age and reaching 66.8% at 24 weeks after birth [13].In pigs, the association of diarrhea with Campylobacter infection was firstly reported in 1948 by Doyle [17,18], who was able to reproduce a clinical syndrome, the so-called “pig dysentery”, in healthy pigs through experimental inoculation. Dysentery was also observed when C. coli was inoculated in gnotobiotic piglets by the oral route. Sala et al. [19] observed diarrhea, bacteremia, and bacterial distribution in many other organs such as lungs, kidneys, and liver of pigs after experimental inoculation with Campylobacter.At the farm level, infections with Campylobacter are associated mainly with lactating piglets and include fever and mild to moderate diarrhea. Dehydration and loss of appetite may also occur. In sows, nursery, and fattening piglets, clinical signs are rarely observed.The purpose of the current study was to evaluate and molecularly characterize the antimicrobial resistance of Campylobacter isolates from pig farms in Greece.2. Materials and Methods2.1. Samples and ProcessingA total of 200 Campylobacter isolates, originating from 16 commercial pig farms from six different Greek regions were examined. The median size of the farms was 550 breeding sows (230–1600). Among these isolates, 100 have been previously identified as C. jejuni and 100 as C. coli. Speciation of Campylobacter isolates has been made according to a multiplex PCR protocol [20]. All samples were cryopreserved at −70 °C, in Brucella broth (OXOID, UK) supplemented with 5% lysed horse blood (OXOID, UK) and 15% glycerol.The cryopreserved Campylobacter samples were defrosted, revived, and inoculated on to modified charcoal cefoperazone deoxycholate agar (mCCDA-OXOID, UK) with the selective supplement SR0155 (OXOID, UK). The inoculated plates were incubated under microaerobic conditions (85% N2, 10% CO2, and 5% O2) at 41.5 °C for 48 h.2.2. DNA IsolationBacterial DNA for PCR was extracted using the conventional boiling method. Briefly, C. jejuni and C. coli colonies were suspended in 250 μL of ΤΕ (Tris-HCl [10 mM]: EDTA [1 mM]) buffer and were homogenized by vortexing. Suspensions were boiled at 100 °C for 10 min and were immediately placed in an iced bath for another 10 min. After centrifugation at 13,500× g for 10 min, the supernatant [100 μL] were collected and transferred to new tubes, and stored at −20 °C for molecular analysis to detect antibiotic-resistance genotypes by PCR.The quantification of extracted DNA was performed spectrophotometrically and the quality of the extracted DNA was estimated from the ratio of absorbance at 260/280 nm. A value range of 1.8–2.2 was considered to indicate DNA isolation of high purity.Antimicrobial susceptibility testing (AST).The agar dilution method was applied for the antimicrobial susceptibility testing. All Campylobacter isolates were examined for their susceptibility to five antimicrobials of five different antimicrobial classes. The antimicrobials tested included gentamicin (GEN), erythromycin (ERY), ciprofloxacin (CIP), tetracycline (TET), and meropenem (MER) (Sigma-Aldrich). The Muller Hinton agar supplemented with 5% mechanically defibrinated horse blood and 20 mg/L β-NAD was used for Campylobacter isolates susceptibility testing.The European Committee on Antimicrobial Susceptibility Testing (EUCAST) breakpoint tables version 11.0 for C. jejuni and C. coli were used for the interpretation of the results [21]. Regarding gentamicin and meropenem, tables of the same version for Enterobacteriaceae were used. Antibiotics exhibiting phenotypic resistance to more than three different classes were regarded as Multidrug Resistant [22]. The C. jejuni ATCC 33560 and C. coli ATCC 33559 were used for the AST, Quality Control.2.3. Genotypic Characterization of Fluoroquinolone ResistanceAll C. jejuni and C. coli isolates that were found resistant to ciprofloxacin, were examined for the presence of Thr-86 to Ile mutations (C-to-T transition) in the quinolone resistance-determining region (QRDR) of the gyrA gene [23,24]. Determination of the gyrA gene presence was performed by applying the mismatch amplification mutation assay PCR and using the FastGene Taq DNA PCR Kit (Nippon Genetics, Düren, Germany) following the manufacturer’s recommendations, with primers (Table 1) and conditions as in Zirnstein et al. [22,23].2.4. Genotypic Characterization of Macrolide ResistanceThe resistance in macrolides was explored by the examination of point mutations at positions 2075 and 2074 in the domain V of the 23S rRNA gene [25], and of the presence of the ribosomal RNA methylase gene, ermB, that was amplified as described by Qin et al. (2014) [26] using the PCR amplification kit.2.5. Genotypic Characterization of Tetracycline ResistanceFor the evaluation of resistance to tetracyclines, three genes, i.e., tet(O), tet(A), and tet(B) were analyzed among the Campylobacter isolates. PCR amplification of these genes was performed using the FastGene Taq DNA PCR Kit following the manufacturer’s recommendations, with primers (Table 1) and reaction conditions as described by Abdi-Hachesoo et al. [27].animals-11-02394-t001_Table 1Table 1Primer sequences used for species identification and detection of resistance genes and mutations.Target GeneSequence (5′-3′)Amplicon Size (bp)Reference tet(O) F:AACTTAGGCATTCTGGCTCAC515[27]R:TCCCACTGTTCCATATCGTCA tet(A) F: GTGAAACCCAACATACCCC888R: GAAGGCAAGCAGGATGTAG tet(B) F: CCTTATCATGCCAGTCTTGC774R: ACTGCCGTTTTTTCGCC cmeB F:GACGTAATGAAGGAGAGCCA1166[28]R:CTGATCCACTCCAGCTATG gyrA Thr-86-Ile mutations (C. jejuni) F: TATGAGCGTTATTATCGGTC265[24]R: TAAGGCATCGTAAACAGCCA gyrA Thr-86-Ile mutations (C. coli) F:TATGAGCGTTATTATCGGTC192[24]R:TAAGGCATCGTAAACAGCCA 23S rRNA at position 2074 F:TTAGCTAATGTTGCCCGTACCG485[25]R: AGTAAAGGTCCACGGGGTCTCG 23S rRNA at position 2075 F:TTAGCTAATGTTGCCCGTACCG486[25]R:TAGTAAAGGTCCACGGGGTCGC ermB F:TGAAAAAGTACTCAACCAAAT692[26]R:TCCTCCCGTTAAATAATAGAT2.6. Genotypic Characterization of Efflux PumpsFinally, the cmeB gene was analysed molecularly in all Campylobacter spp. strains for the presence of the multidrug efflux pumps, using primers and the PCR amplification procedure suggested by Pumbwe et al. [28].In all analyses, PCR products were visualized by electrophoresis using an agarose gel stained with ethidium bromide.2.7. Statistical AnalysisComparisons between rates were performed after the preparation of contingency tables (chi-squared tests), as provided in the IBM® SPSS® version 25 statistical software (IBM Corp., Armonk, NY, USA). The level of significance was set at 5% (α = 0.05).3. Results3.1. Antimicrobial Susceptibility Testing (AST)From the totally 200 examined Campylobacter isolates, 23 (11.5%), nine (9%) C. coli, and 14 (14%) C. jejuni isolates were found susceptible to all five antibiotic classes tested, while 25 isolates (12.5%), 15 C. coli, and 10 C. jejuni were classified as multidrug resistant by showing resistance to three different classes of antibiotics.All clinical isolates were susceptible to meropenem. Very high rates of resistance were recorded for tetracycline, i.e., 169 Campylobacter isolates (84.5%), were resistant. More specifically, 88 (88%) and 81 (81%) of C. coli and C. jejuni, respectively, presented resistance in this class of antibiotics. Medium rates of resistance were recorded regarding fluroquinolones as 46 isolates (23%), 22 (22%) and 24 (24%) of C. coli and C. jejuni, respectively were resistant in ciprofloxacin. Finally, low and very low rates of resistance were identified for macrolides and aminoglycosides. For erythromycin, 24 isolates (12%), 13 (13%) and 11 (11%), C. coli and C. jejuni, respectively, were exhibiting resistance, while only eight isolates (4%), five C. coli and three C. jejuni were found resistant to gentamicin. It is noteworthy that all erythromycin resistant Campylobacter isolates, presented high level resistance against the selected antibiotic with MIC ≥ 32 mg/L. Statistical analysis did not reveal significant differences (p > 0.05) concerning the rates of resistance to any of the investigated antibiotics between C. coli and C. jejuni isolates. However, a clear pattern was discerned within both bacterial species, with resistance to ciprofloxacin being significantly (p < 0.001) less frequent than the resistance to tetracycline, and significantly more frequent that the resistance to erythromycin, gentamicin, and, self-evidently, meropenem.Among all Campylobacter isolates, three MDR phenotypes were determined. The resistance CipEryTet phenotype was the most common, as it was present in 17 Campylobacter spp. isolates (8.5%). The phenotype CipGenTet followed, as it was identified in seven isolates (3.5%). In one isolate (0.5%), we observed resistance in erythromycin, gentamicin, and tetracycline. Moreover, three Campylobacter spp. isolates (1.5%), were resistant to both erythromycin and ciprofloxacin, drugs of choice for the treatment of invasive human campylobacteriosis. All nine resistance phenotypes recorded in the current study are shown in Table 2 and resistance rates of Campylobacter isolates in Table 3.3.2. Antibiotic Resistance Genes (Molecular Mechanisms of Resistance)Molecular screening of tetracycline resistance genes indicated that 92.3% of Campylobacter isolates (156/166) were positive for tet(O). Particularly, the tet(O) genetic locus was detected in 94.3% of C. coli isolates (83/88) and 90.1% of C. jejuni isolates (73/81). The tet(A) locus was only found in 6.5% of Campylobacter spp. isolates (11/169). More specifically, eight C. jejuni and only three C. coli isolates were harboring the tet(A) gene. It should also be noted that we found two C. coli isolates positive for both tet(O) and tet(A) genes. None of the Campylobacter isolates was found positive for the tet(B) resistance gene.Concerning the ciprofloxacin resistant isolates (n = 46), they all hosted the mutant type Thr-86-Ile region of the quinolone resistance-determining region (QRDR) of the gyrA gene.In all erythromycin resistant isolates (n = 24), the transitional mutations A2075G and A2074C in the 23S rRNA gene were only amplified. The ermB gene was not identified in any isolate.In a significant number of the isolates tested for the presence of efflux pumps as a resistance mechanism, the cmeB gene was amplified. More specifically, 23.1% of the C. coli (n = 21) and 4.6% of C. jejuni (n = 4) that were characterized as resistant in at least one of the antibiotic classes, were found to harbor the cmeB gene.No significant differences were revealed between C. coli and C. jejuni in all but one of the above percentages. More specifically, the presence of the cmeB gene was more frequently (p < 0.05) amplified in the resistant C. coli compared with the C. jejuni isolates.The frequencies of common mutations and genes conferring resistance to fluoroquinolones, tetracyclines, macrolides, and influx pumps are presented in Table 4.4. DiscussionThe AMR rates recorded in our study and more specifically the low and very low rates of erythromycin and gentamicin resistance, are in line with the results reported in the EFSA-ECDC summary report on antimicrobial resistance for 2017. The median EU resistance rates for tetracycline were 51.5%, and for erythromycin and gentamicin, 15.6% and 7.7%, respectively, in 979 C. coli isolates from fattening pigs. In regards to fluoroquinolones such as ciprofloxacin, we have determined a significant lower rate of resistance compared to 52.3% of the EFSA report [29]. Moreover, for erythromycin resistance, the low rates of resistance that we have recorded are in great difference with the data from China [30], where Tang et al. found that all 23 (100%) C. coli isolates from pigs and overall 75.3% of Campylobacter isolates from poultry and pigs were resistant to erythromycin. These differences can be attributed to the fact that macrolides, including erythromycin, are the only antibiotics authorized by the Chinese government for use as feed additives [31].For tetracycline, the median EU resistance rate for 2017 was 51.5%. Spain has recorded in the EFSA-ECDC summary report on antimicrobial resistance for 2017, a 65.3% of resistance to tetracycline from Campylobacter isolates in fattening pigs. On the contrary, we recorded extremely high AMR rates for the specific antibiotic, exceeding 84% of the total isolates. Higher than the median EU rate but not as high as in our study rates, tetracycline resistance from pigs investigated in China were 64% of the Campylobacter isolates resistant to tetracycline [30]. In accordance with our findings, Padungtod et al. (2006) [32] reported 88% tetracylcin resistance in Campylobacter isolates from pigs in Northern Thailand.The EU median for MDR C. coli according to the EFSA report [29] was 21.2%. We recorded that MDR Campylobacter spp. isolates were relatively lower (12.5%), in particular 15% for C. coli and 10% for C. jejuni. The most common MDR phenotypes identified in the current study (CipEryTet and CipGmTet) are those recognized by the EFSA report.According to the third joint inter-agency report on integrated analysis of antimicrobial agent consumption and occurrence of antimicrobial resistance in bacteria from humans and food-producing animals in the EU/EEA, titled “Antimicrobial consumption and resistance in bacteria from humans and animals” (ECDC, EFSA, EMA, 2021), consumption of tetracyclines in food-producing animals in Greece for the year 2018 (48.9 mg/PCU) far exceeded the mean value of the 31 countries included (31.7 mg/PCU), ranking sixth overall [33]. The extensive use of tetracyclines in livestock production may be related with the high rates of resistance.On the contrary, consumption of macrolides (4.1 mg/PCU) significantly lagged behind the European mean value (8.0 mg/PCU), whereas consumption of both fluoroquinolones and aminoglycosides (2.2 and 6.5 mg/PCU, respectively) was very close to the overall means (2.5 and 6.4 mg/PCU, respectively) [33]. It is noted that meropenem is not authorized for use in food-producing animals, at least within the European Union, and since it belongs to Category A (Avoid), according to the EMA/CVMP/CHMP classification (2019), its use is exceptionally allowed in companion animals only. Antibiotic consumption data that indicate limited to moderate use of macrolides, aminoglycosides, and fluroquinolones in livestock production could explain the moderate and low resistance rates in the specific classes of antibiotics. Moreover, the susceptibility of all Campylobacter isolates to meropenem is related with the fact that it is not authorized for use in livestock production.The principle molecular mechanism for ciprofloxacin resistance of Campylobacter is the alteration of codon 86 from threonine to isoleucine in the gyrA genomic region [34]. All the ciprofloxacin phenotypically resistant Campylobacter isolates in our study shared the same mechanism. Our findings are similar to Woźniak-Biel et al. (2016) [35] and El-Adawy et al. (2012) [36], who revealed the same mutation in all ciprofloxacin-resistant Campylobacter strains from broilers and turkeys, respectively. Moreover, Tang et al. [30] characterized the T86I amino acid substitution as the sole mutation recorded in quinolones resistant Campylobacter isolates from poultry and pigs.Concerning the macrolides molecular mechanism of action for resistance, we did not detect the presence of the ermB gene in any of the isolates tested. Erythromycin resistance in all Campylobacter isolates was determined by detecting point mutations at position 2075 and 2074 in the V 23S rRNA gene. The high resistance levels observed in our study (MIC ≥ 32 mg/L) are in agreement with the correlation of the specific resistance mechanism with high resistance levels [37]. On the other hand, our results are in divergence with those revealed by Tang et al. [30], where 52.7% of Campylobacter isolates, mainly from poultry, were found positive for the ermB gene. Furthermore, more than half of the ermB-positive isolates also demonstrated the A2075G 23S rRNA mutation. Only three Campylobacter isolates from pigs were found to carry the ermB gene.In 92.3% of all Campylobacter isolates (n = 156), the specific tet(O) gene was amplified providing evidence that resistance against tetracycline was mediated mainly through the gene, whereas in only 6.5% (n = 11) of all isolates, the tet(A) gene was identified. These results are in accordance with several previous studies [27,30,38,39] referring to the tet(O) gene as the principal mechanism of tetracycline resistance in Campylobacter isolates from different sources (animal, human, food).The cmeB efflux pump is not only related with fluoroquinolone resistance, but with resistance to multiple antibiotics (macrolides, chloramphenicol, tetracycline), dyes (acridine orange), and disinfectants, as well [28,40,41]. The cmeB gene was amplified in 14.1% (n = 25) of all Campylobacter isolates. These findings are similar to those reported previously from Campylobacter isolated from turkeys [34]. Significant differences were recorded between the two Campylobacter species concerning the presence of the cmeB gene. The specific gene was amplified in 21 C. coli (23.1%) isolates and only in four C. jejuni isolates (4.7%).5. ConclusionsThis study was designed to assess the phenotypic and molecular patterns of resistance of C. coli and C. jejuni isolates from commercial farrow-finisher pig farms in Greece.Fluoroquinolones and macrolides have been classified as category I and category II antimicrobials, respectively and are characterized as “first line” antibiotics for campylobacteriosis treatment. The high rates of resistance in tetracyclines and the moderate rates of fluoroquinolones resistance highlight the necessity for a continuous and systematic monitoring and surveillance of Campylobacter isolates from pig farms, regarding their phenotypic and molecular resistance patterns. Monitoring Campylobacter isolates susceptibility and the proper use of antimicrobials in livestock production are considered of great importance in order to tackle antimicrobial resistance and the spread of antimicrobial resistance pathogens and resistance genes.

animals : an open access journal from mdpi

10.3390/ani13081345

PMC10135201

Cultivated fish face multiple stressors that impact their viability, growth, and health; additionally, environmental stressors provoke the transmission of bacterial diseases in fish populations. It is widely recognized that feed supplements of plant origin can alleviate stress in fish via the stimulation of non-specific defense responses. We tested the effects of a dietary mix consisting of an antioxidant, dihydroquercetin, and a prebiotic, arabinogalactan, on growth and biochemical indices in farmed rainbow trout, Oncorhynchus mykiss. The supplement has been shown to maintain the viability of fish affected by a natural bacterial infection as well as alter the fatty acid composition and decrease oxidative damage to the liver. Thus, plant-origin substances could be readily available and safe alternatives to pharmacotherapy and used as a way to improve the health of fish and promote their ability to tolerate stressors under intensive production.

Natural feed supplements have been shown to improve fish viability, health, and growth, and the ability to withstand multiple stressors related to intensive cultivation. We assumed that a dietary mix of plant-origin substances, such as dihydroquercetin, a flavonoid with antioxidative, anti-inflammatory, and antimicrobial properties, and arabinogalactan, a polysaccharide with immunomodulating activity, would promote fish stress resistance and expected it to have a protective effect against infectious diseases. Farmed rainbow trout fish, Oncorhynchus mykiss, received either a standard diet or a diet supplemented with 25 mg/kg of dihydroquercetin and 50 mg/kg of arabinogalactan during a feeding season, from June to November. The fish in the control and experimental groups were sampled twice a month (eight samplings in total) for growth variable estimations and tissue sampling. The hepatic antioxidant status was assessed via the quantification of molecular antioxidants, such as reduced glutathione and alpha-tocopherol rates, as well as the enzyme activity rates of peroxidase, catalase, and glutathione-S-transferase. The lipid and fatty acid compositions of the feed and fish liver were analyzed using thin-layer and high-performance liquid chromatography. The viability, size, and biochemical indices of the fish responded to the growth physiology, environmental variables such as the dissolved oxygen content and water temperature, and sporadic factors. Due to an outbreak of a natural bacterial infection in the fish stock followed by antibiotic treatment, a higher mortality rate was observed in the fish that received a standard diet compared to those fed supplemented feed. In the postinfection period, reduced dietary 18:2n-6 and 18:3n-3 fatty acid assimilation contents were detected in the fish that received the standard diet in contrast to the supplemented diet. By the end of the feeding season, an impaired antioxidant response, including reduced glutathione S-transferase activity and glutathione content, and a shift in the composition of membrane lipids, such as sterols, 18:1n-7 fatty acid, and phospholipids, were also revealed in fish fed the standard diet. Dietary supplementation with plant-origin substances, such as dihydroquercetin and arabinogalactan, decreases lethality in fish stocks, presumably though the stimulation of natural resistance in farmed fish, thereby increasing the economic efficacy during fish production. From the sustainable aquaculture perspective, natural additives also diminish the anthropogenic transformation of aquaculture-bearing water bodies and their ecosystems.

1. IntroductionRainbow trout, Oncorhynchus mykiss, is one of the most cage-cultured fish species in high-latitude countries due to its fast growth in low-temperature environments [1,2]. Unlike wild conspecifics, intensively farmed fish are known for having an altered physiological state and increased susceptibility to infections [3]. In order to sustain farm ecosystems and improve fish welfare under environmental or human-derived stressors, a number of biologically active compounds is proposed to be introduced into the diet of farmed fish [4,5,6]. Some biologically active substances of plant origin, such as dihydroquercetin (also known as taxifolin) and arabinogalactan, are predicted to promote fish stress resistance and reduce the need for antimicrobial treatments under intensive cultivation. Dihydroquercetin is a natural polyphenol that is extracted from larch (Larix gmelinii and L. sibirica) and is structurally similar to other bioactive flavonoids such as quercetin, hesperidin, or rutin. It is widely used in the medical and food industries due to its antioxidative, anti-inflammatory, antimicrobial, and antitumor activities [7,8,9,10]. There have also been attempts to use dietary dihydroquercetin in veterinary medicine, which have had either beneficial effects by increasing the immune status of gilthead seabream [11] and suppressing Cd toxicity in zebrafish embryos [12] or no impact on growth performance or any of the studied physiological variables [13]. Arabinogalactan, a plant polysaccharide, was reported to have immunomodulator activity through protecting symbiont intestinal microbiota [14,15,16,17]. Because dihydroquercetin was established to have strong antimicrobial effects in vivo [18,19,20] and arabinogalactan was reported to eliminate bacterial pathogens [15,21,22,23,24], their combination is expected to have a protective effect against infectious diseases. Both substances are promoted by their producers to fish farms as feed supplements; however, the experimental data on their benefits in the fish-rearing industry are limited [11,25,26,27,28].The aim of the study was to estimate some physiological and biochemical indices in farmed rainbow trout (Oncorhynchus mykiss) that received either a standard diet or a diet supplemented with dihydroquercetin and arabinogalactan throughout a growing season for five months. Due to an episode of a natural bacterial infection that occurred during the observation period, the antioxidant and immunomodulating properties of a natural additive were able to be verified.2. Materials and Methods2.1. Fish Maintenance and FeedingThe experiment was carried out from June to November on a trout aquafarm in Ladmozero Lake in Northwest Russia. Healthy rainbow trout (Oncorhynchus mykiss) yearlings (average weight 100.1 ± 2.3 g, age 1+) were placed in four cages with an initial stocking density of 2.1 kg/m3; the total biomass of the fish in each cage was around 886 kg. The fish were fed one of two diets (in duplicate), either the commercial diet BioMar (Denmark; for composition, see Table S1) without any supplements (control diet) or the same commercial diet supplemented with 25 mg/kg of dihydroquercetin and 50 mg/kg of arabinogalactan in accordance with the producer’s recommendations (quality and safety certificate no. 396-08.17, Ametis, Russia). The feed was delivered to the fish manually twice a day. The amount of food rations was calculated daily by considering the current biomass in each cage, water temperature, and dissolved oxygen content. Food fractions (3.0 or 4.5 mm) were chosen based on fish size; the compositions of both fractions of feed were similar (Table S1). The lipid composition of the commercial feed pellets was analyzed and the results are presented in Table S2. The supplement (dihydroquercetin and arabinogalactan) was top-dressed as feed granules by the staff of the farm directly on the day of feeding. For this, a portion of the supplement was dissolved in water at 50 °C in a 10 L tank and the solution was sprayed onto feed pellets while mixing the feed manually. During the five-month experiment, five courses of supplementation lasting for 10 to 14 days (shown in pink in the figures) were given. Feeding the supplement in short, repeated courses made it possible to achieve the desired biological effect at reduced economic cost.The water temperature and dissolved oxygen content values measured daily with S9 Seven2Go pro (Mettler Toledo, Switzerland) throughout the study period were between 6 and 18 °C and 5 and 11 mg/L, respectively (Figure 1). The water quality variables, both directly in the cages and within 500 m from the fish farm, were assessed twice during the observation period on 29 June and 4 October (Table S3), and all key parameters were found to be satisfactory.During the observation period, a sporadic bacterial disease occurred in the fish stock, with the initial manifestation on 12 July in rainbow trout yearlings in both groups. The infected fish swam on their sides, refused food, and had hepatic abnormalities and hemorrhages in the liver. The infection was identified as a bacterial hemorrhagic septicemia-type disease caused by an association of Pseudomonas putida and Cytophaga psychrophile. Two weeks after the disease manifestations, the fish in all cages were treated with an antibiotic, enrofloxacin (25 mg/kg of fish weight), using a six-day regimen.2.2. Fish SamplingAt the indicated sampling dates, eight fish from each cage were euthanized with an overdose of clove oil (250 mg L⁻1). The liver samples were dissected with scissors, then frozen in liquid nitrogen and stored at −80 °C until the enzymatic assay and evaluation of reduced glutathione and vitamin contents. For the lipid analysis, hepatic samples were fixed in 97% ethanol (with the addition of 0.001% butylated oxytoluene as an antioxidant) and stored at +4 °C.2.3. Biometric Indices and MortalityFish mortality data were provided by the fish farm staff as the number of fish deaths on the water surface in the cages per month. Twice during the observation period (26 July, 16 November), the cages were lifted from the depths to the count dead fish accumulated at the bottom of the cages. Thus, the November mortality values included the number of unnoticed dead fish accumulated at the bottom of the cages since August (Table S4). The total wet mass of the fish was measured using a portable Ohaus Scout SPX Series balance instrument (Figure 2).2.4. Biochemical AssayBiochemical analyses of hepatic antioxidants and lipid profile in the rainbow trout O. mykiss were conducted in the Equipment Sharing Center of the Karelian Research Center of the Russian Academy of Sciences (Petrozavodsk, Russia).2.4.1. Antioxidant AnalysesThe frozen 0.1–0.4 g liver samples were homogenized with a Digital Disruptor Genie unit (Scientific Industries, Bohemia, NY, USA) in 5.0 mmol Tris-HCl buffer, pH 7.5. The homogenate was centrifuged at 60,000× g for 1 h at 4 °C in an Allegra 64R centrifuge (Beckman Coulter, Brea, USA). The resultant supernatant was used for a biochemical assay of the enzyme activities and GSH level.The glutathione S-transferase (GST) activity was determined from the rate of reduced glutathione (GSH) binding with the 1-chloro-2,4-dinitrobenzene (CDNB) substrate [29]. A microplate well was injected with 0.225 µL of reaction mixture containing 1 mmol of CDNB and 1 mmol of GSH in 0.125 M of PBS at pH 6.5. The reaction was initiated by adding 0.025 µL of the homogenate solution, and the following increase in the solution’s optical density was recorded at 340 nm at 25 °C with a CLARIOstar microplate reader (BMG Labtech, Ortenberg, Germany). The relative activity of the enzyme in the fish tissue was expressed in µmol of the reaction product generated within a minute, calculated per mg of soluble protein in the tissue (µmol/mg protein/min).The catalase (CAT) activity was determined using a technique by Beers and Sizer [30]. The activity of the enzyme was measured by decomposing 25 mmol of hydrogen peroxide in 50 mmol of PBS at pH 7.4 and 25 °C. The optical density of the resulting solution at 240 nm was registered. The relative CAT activity was expressed in hydrogen peroxide µmol decomposed within a minute and calculated per mg of soluble protein in the tissue, and is reported in µmol/mg protein/min.The peroxidase (Px) activity was determined using a technique used by Maehly and Chance [31] in a reaction mixture containing 10 mmol of guaiacol and 25 mmol of hydrogen peroxide in 50 mmol of PBS at pH 7.4 and 25 °C. The peroxidase activity was determined by measuring the absorbance of the reaction product at 470 nm and is reported in µmol/mg protein/min.The concentration of reduced glutathione (GSH) was determined based on the Cohn and Lyle method [32] modified by Hissin and Hilf [33] as described below. Briefly, soluble proteins were precipitated from the homogenate using 5% trichloroacetic acid and separated by centrifuging at 2500× g for 15 min. The supernatant was adjusted to pH 8.5 with 5 mol of NaOH and then mixed with 0.4 mol of Tris-HCl at pH 8.5, containing 5 mmol of ethylene diamine tetraacetic acid (EDTA). Then, 0.01% ortho-phthalaldehyde in methanol, prepared immediately before use, was added to the reaction mixture. The mixture was stirred and incubated at room temperature for 15 min, after which its fluorescence was measured (em 420 nm, ex 350 nm wavelengths). The reduced glutathione concentration was calculated according to the calibration curve plotted using series of solutions with different GSH concentrations in 0.4 mol of Tris-HCl at pH 8.5, containing 5 of mmol EDTA. The relative glutathione concentration was expressed in GSH µg per mg protein.The soluble protein concentration in the supernatant was measured spectrophotometrically through peptide bond absorbance at 220 nm at 26 °C [34].The concentration of alpha-tocopherol (vitamin E) in the fish liver was determined using high-performance liquid chromatography (HPLC). The samples of tissues were homogenized in 0.25 mol of sucrose solution at pH 7.4 and mixed with an equal volume of 0.125 mg/mL of butylated oxytoluene in ethanol to precipitate the proteins. Next, each sample was diluted twice with n-hexane. The mixture was vortexed for 5 min for the extraction of vitamins and centrifuged at 3000× g for 10 min. The hexane layer was injected onto the HPLC system [35]. Chromatographic separation was carried out on a 0.4–0.6 m 250–300 µL silica gel Analytical Chromatographic Column 6 using a Milichrom 6 chromatograph (Nauchpribor, Russia). A mixture of n-hexane and isopropanol (98.5:1.5, v/v) was applied as the mobile phase. The eluate was monitored at 292 nm, and the vitamin was identified by the retention time compared with the pure standard (MP Biomedicals, Irvine, CA, USA).2.4.2. Lipid Composition AnalysisLipids were extracted with chloroform/methanol (2:1, v/v) according to Folch et al. [36]. The extracted lipids were spotted onto TLC Silica gel 60 F254 thin-layer chromatography plates (Merck, Darmstadt, Germany) and separated into different fractions of lipid classes using petroleum ether/diethyl ether/acetic acid (90:10:1, v/v) as the mobile phase. The identification of the fractions was performed using standards such as the phospholipid mixture (P3817, Supelco, St. Louis, USA), cholesterol (C8667, St. Louis, Sigma-Aldrich, USA), glyceryl trioleate (92860, Sigma-Aldrich, St. Louis, USA), and cholesteryl palmitate (C78607, Sigma-Aldrich, St. Louis, USA). The quantitative composition of the fractions was measured at 540 nm for phospholipids, triacylglycerols, and cholesterol esters and at 550 nm for the sterol fraction using an SF-2000 UV/Vis spectrophotometer (Spektr, Saint Petersburg, Russia) [37,38].The composition of the phospholipid fractions, including phosphatidylinositol (PI), phosphatidylserine (PS), phosphatidylethanolamine (PE), and phosphatidylcholine (PC), was determined by high-performance liquid chromatography using the method of Arduini et al. [39] on a Nucleosil 100-7 column (Elsiko, Moscow, Russia) with the acetonitrile/hexane/methanol/phosphorus acid mixture as a liquid phase (918:30:30:17.5, v/v) and a UV-spectrophotometer at 206 nm using a Stayer liquid chromatography system (Aquilon, Moscow, Russia). Peaks were identified by referencing the retention times of the authentic standard phospholipid mixture (P3817, Supelco, St. Louis, USA) and phosphatidylserine (P7769, Sigma-Aldrich, St. Louis, USA).Fatty acid methyl esters (FAME) from the total lipids were prepared using methanol and acetyl chloride. The FAME were separated on an Agilent 7890A gas–liquid chromatograph (Agilent Technologies, Santa Clara, USA) with a flame ionization detector using DB-23 columns (60 m–0.25 mm) (Agilent Technologies, Santa Clara, USA) and nitrogen as the mobile phase. The FAME from fish total lipids were identified by comparison with standard mixes (Supelco, St. Louis, USA). Minor fatty acids constituting less than 1% of the total fatty acids were not identified.2.5. Statistical AnalysesThe statistical analyses were performed using R Statistical Software [40]. Differences between samples were estimated using the nonparametric Mann–Whitney U test. Correlations between the investigated parameters were analyzed using Spearman’s rank correlation coefficient with p-values adjusted using the Holm method. The weight gain rates in the two groups were compared using an ANCOVA. The differences were considered significant at p ≤ 0.05. The values of the studied biochemical parameters are presented as the median values ± half of the interquartile range.3. Results3.1. Fish Survival and GrowthDuring the observation period, a significant difference in fish mortality was observed between the studied fish groups (Table S4). The total number of fish deaths was much lower (about 2.5 times) in the group that received dihydroquercetin and arabinogalactan than in the supplement-free group (2.5 vs. 6.5% per season). In the experimental group, the highest mortality rate (0.95% of the total number of fish per month) was observed in August, at the third week since the manifestation of the sporadic bacterial infection at the fish farm. After disease curation, mortality in the supplement-fed group slowed down, with a consistently low rise after October. In the control group that received commercial feed only, the increase in mortality started around the same time (in early August), rising from 0.96% per month to 2% per month by September, and then remained high through to the end of the experiment.The weight gain in the farmed trout of either group was linear (correlation coef. r = 0.98; Figure 2) and equal in both groups (ANCOVA) during the season. There were no differences between the weight gain rates of the control and experimental groups.3.2. Hepatic Antioxidant ComponentsThe glutathione-mediated antioxidant responses varied within the period of observation, mostly depending on the infectious status of the fish (Figure 3, Figures S1–S3), with maximum glutathione-S-transferase (GST) activity in August. Additionally, a significant positive correlation of the GSH level with the water temperature (r = 0.6; Figure S4) was found. No prominent correlations of fish biometrics with AOS parameters were found. The GSH concentrations in fish tissues revealed no significant diet-related differences in any sampling date except those in November, when an increase (about 1.5-fold) in the hepatic GSH pool was observed in the supplement-free group (Figure 3A). Similarly, the hepatic GST activity in the standard-fed fish was higher (about 2-fold) in November (Figure 3B). The levels of other studied antioxidants, such as hepatic α-tocopherol (vitamin E), and the catalase and peroxidase activity levels were similar in both studied groups throughout the observation period (Figures S1–S3).3.3. Hepatic Lipid and Fatty Acid ProfileIn farmed rainbow trout, the fish mass gain was found to be correlated with phosphatidylethanolamine and phosphatidylserine (r = −0.6 both), as well as with sterols and vaccenic 18:1n-7 and palmitoleic 16:1n-7 acids (r = 0.6, 0.7, 0.5, respectively) (Figure 4A), indicating the structural function of these lipids. The water temperature was shown to be negatively correlated with the contents of eicosapentaenoic 20:5n-3 and arachidonic 20:4n-6 fatty acids, the constituents of membrane phospholipids (r = −0.7, −0.6, respectively). In general, the hepatic fatty acid composition was shown to be very similar to that of commercial feed (Figure 4B). Expectedly, abundant components of the fish feed, such as oleic 18:1n-9, linoleic 18:2n-6, alpha-linolenic 18:3n-3, gamma-linolenic 18:3n-6, eicosapentaenoic 20:5n-3, and palmitoleic 16:1n-7 acids, are positively correlated with each other in fish livers, likely due their common exogenous origin (Figure 4A).In both studied fish groups, the TAGs were intensely accumulated in summer in accordance with the total fat deposition (negative correlation with water temperature, r = −0.5) (Figure 4A and Figure 5) and positively correlated with oleic 18:1n-9 and palmitoleic 16:1n-7 acids (r = 0.5, 0.6, respectively). Although it was unlikely due to the fish feed, greater portions of docosahexaenoic 22:6n-3 and arachidonic 20:4n-6 acids were found in fish livers (Figure 4B). The correlation network analysis indicated the conversion of alpha-linolenic 18:3n-3 to docosahexaenoic acid and linoleic 18:2n-6 to arachidonic acid (r = −0.6 and −0.8, respectively). Similarly, saturated 18:0 fatty acids were transformed to monounsaturated oleic 18:1n-9 and vaccenic 18:1n-7 acids (r = −0.6 and −0.5, respectively) in the livers of the fish (Figure 4A).The dietary supplementation with arabinogalactan and dihydroquercetin for just a week resulted in an increase in the 15:0 fatty acid and a decrease in the 18:0 fatty acid contents in fish livers (Figure 6A,B). There were no significant intergroup differences in hepatic lipid content while the infection progressed (through August) (Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8, Figures S5–S13). In samples collected on 12 August after antibiotic treatment and the second two-week course of supplement administration, significant decreases in hepatic 18:0 and 16:0 fatty acids were found in fish fed an enriched diet (Figure 6B,C), while decreases in linoleic 18:2n-6 and alpha-linolenic 18:3n-3 were detected in supplement-free trout for the same dates (Figure 7A,B). In the postinfection period, no prominent difference in lipid composition between the fish groups was found; in contrast, by the time the water temperature dropped to 6 °C (November 16), the contents of hepatic sterols and vaccenic acid 18:1n-7 increased and of phosphatidylethanolamine decreased in fish receiving an enriched but not a standard diet (Figure 7C and Figure 8A,B). Similarly, vaccenic acid 18:1n-7 as a component of phospholipids positively correlated with fish weight (Figure 4A). In infected supplement-free fish, the content of vaccenic acid 18:1n-7 decreased while the contents of more easily oxidizable substrates, polyunsaturated docosahexaenoic 22:6n-3 and arachidonic 20:4n-6 acids, did not differ between the groups (Figures S11 and S13).4. Discussion4.1. Infections and MortalityDihydroquercetin- and arabinogalactan-containing feed had a beneficial effect on fish viability, significantly reducing the cumulative mortality rate within the growing season. The impaired welfare observed in the supplement-free fish group was likely a result of a more severe infection, lower susceptibility to the antibiotic treatment, re-infestation, or other causes. The similar growth rates mean that regardless of the diet, the fish consumed food and effectively converted the obtained energy into body mass. The unperturbed growth but lower viability in fish that were fed a standard diet possibly indicate the sudden death of individuals due to sporadic acute infection rather than the chronic effect of any undetermined hazardous agents on caged fish. We assume that the tested supplement mixture may have a beneficial long-lasting effect on fish viability, probably via increasing the resistance to both infections and re-infections. Our data on the influence of the studied plant-origin substances on the survival and welfare of rainbow trout correspond with data in the literature describing the effects of other related compounds on the welfare and mortality of fish. For example, medicinal plant feed additives enhanced the growth and survival of cultured Clarias gariepinus [41]. Dietary medicinal plant extracts improved the growth, immune activity, and survival of tilapia, Oreochromis mossambicus [42].4.2. Antioxidant DefenseA number of stress factors (e.g., temperature, bacterial infection, medications) initiating reactive oxygen species (ROS) generation and lipid peroxidation (LPO), and consequently providing oxidative stress, affect the survival and growth of trout [43,44,45,46]. The antioxidant system (AOS) protects an organism from the oxidative stress at the molecular level [43,45]. The AOS includes (1) low-molecular-weight antioxidants (e.g., reduced glutathione, tocopherol, and others) directly interacting with reactive oxygen species and (2) antioxidant enzymes, such as catalase, peroxidase, glutathione-S-transferase, catalyzing the deactivation and elimination of adverse molecules [43,47,48]. The conjugation of pro-oxidants with GSH, either spontaneously or via being catalyzed by glutathione-S-transferase, and the subsequent excretion of the complexes from the organism is one of the main pathways of elimination of ROS and their derivatives [49,50]. Interestingly, dihydroquercetin was reported to directly upregulate glutathione-S-transferases in human cells through the antioxidant-responsive element (ARE) binding the promoter regions of GST genes [51]; however, we did not observe a similar effect in trout. The increased GST activity and GSH content found in fish of the supplement-free group in November suggested enhanced ROS production in their tissues in response to low temperatures. Catalase and peroxidase decompose organic and inorganic peroxides causing membrane LPO and α-tocopherol captures free radicals in cell membranes [43,47,48]. As these antioxidants were unaffected by the infection in both studied fish groups, we can assume that the ROS generation and lipid peroxidation were moderate and did not exceed the antioxidant capacity of the low-molecular-weight antioxidant pool members, such as glutathione. Similarly, dietary dihydroquercetin at a dosage 500 mg (20 times more than what the trout received) produced no effect on hepatic α-tocopherol and peroxidase in broiler chickens [13]. In the supplement-free fish group, glutathione and GST activation in response to ambient temperature but not to the infection is probably sufficient to prevent LPO progression.4.3. Lipid and Fatty Acid Response to StressorsBeing a key component of various metabolic pathways, lipids play a crucial role in the survival and tolerance of an organism to environmental stressors. Cell lipids contain polyunsaturated fatty acids enriched by double bonds within their hydrocarbon chains, which are primary targeted by pro-oxidants and can indicate LPO progression. The main lipid components of the O. mykiss liver are membrane phospholipids. Together with sterols (mainly cholesterol), they comprise cell membranes and affect membrane fluidity and the activity of membrane-bound enzymes, ion channels, and receptors [52,53,54]. Membrane phospholipids are also a source of bioactive compounds and messengers [55,56]. Phospholipids include a hydrophilic head, which may include substances such as choline, ethanolamine, or serine and hydrophobic long-chain fatty acids.The water temperature determines the dissolved oxygen content and feed consumption rate of the fish and eventually their metabolic rate [28], and directly affects the cell membrane stability. Our data on the negative correlation of the eicosapentaenoic 20:5n-3 and arachidonic 20:4n-6 fatty acid contents and ambient water temperature support the observations on their contribution to the maintenance of membrane fluidity under temperature response [52,53,56].Triacylglycerols (TAGs) are high-energy lipid molecules stored during intense feeding, readily covering the energy costs needed to maintain homeostasis at different life stages or in different environments [54,56]. In both studied fish groups, the TAGs were intensely accumulated in summer in accordance with the amount of supplied feed and total fat deposition, and their content was positively correlated with oleic 18:1n-9 and palmitoleic 16:1n-7 acids, probably indicating their alimentary uptake and deposition as TAG constituents in hepatic fat.According to previous observations, the fatty acid composition of the fish liver is very similar to that of commercial feed, with an equal ratio of oleic 18:1n-9, linoleic 18:2n-6, alpha-linolenic 18:3n-3, gamma-linolenic 18:3n-6, eicosapentaenoic 20:5n-3, and palmitoleic 16:1n-7 acids, suggesting their common exogenous origin [57,58,59,60]. This is especially true for linoleic 18:2n-6 and alpha-linolenic 18:3n-3 acids, which are not produced by fish and can only be of food origin. Compared to the fish feed composition, greater portions of docosahexaenoic 22:6n-3 and arachidonic 20:4n-6 acids were found in the fish livers, indicating that they are at least partly of endogenous origin. We found supporting evidence of alpha-linolenic 18:3n-3 conversion to docosahexaenoic acid, linoleic 18:2n-6 to arachidonic acid, and saturated 18:0 to monounsaturated oleic 18:1n-9 and vaccenic 18:1n-7 acids, which physiologically occurs in the liver of freshwater fish [59].The accumulation of 15:0 fatty acid (mostly of bacterial origin) and a decrease in 18:0 coincided with the initial stage of infection, and this response may arise from exposure to stress or pathogens. Although 15:0 acid is known to be of bacterial origin, it still unclear whether it was sourced from pathogenic or symbiotic microbiota or if these components were unequally assimilated from food by the fish in both studied groups [61].Since linoleic and alpha-linolenic acids are both essential and food-derived, these results can be interpreted as an impaired digestion sign in supplement-free fish. The beneficial effect of dietary arabinogalactan on symbiotic intestinal microflora (bifidobacteria and lactobacilli) as well as the boosting of immunological characteristics were reported previously in mammals [15]. The accelerated postinfection recovery and enhanced assimilation of nutrients observed in fish fed an experimental diet presumably resulted from the antioxidant and immunostimulant activity of the supplement. We can predict diet-related differences in fatty acid utilization pathways in fish; thus, the 18:0 and 16:0 acids could be more readily utilized in monounsaturated acid synthesis in fish fed an enriched diet, whereas linoleic and alpha-linolenic acids could be more intensively used to synthesize polyunsaturated acids in fish receiving a standard diet. The preferred synthesis of unsaturated fatty acids in fish in the control group may have been due to the need to stabilize cell membranes in animals more perturbed by the infection, because in response to this pathogen cells produce ROS both possessing antimicrobial activity and initiating LPO [43,44,45]. A possible explanation for the vaccenic acid 18:1n-7 decrease detected in the supplement-free fish is its oxidation by ROS. However, the contents of polyunsaturated docosahexaenoic 22:6n-3 and arachidonic 20:4n-6 acids, which are more easily oxidizable substrates in cell [56,59], did not differ between the groups, excluding LPO progression and severe oxidative stress.Sterols (mainly consisting of cholesterol) and phosphatidylethanolamine both comprise important indexes of membrane fluidity, such as the ratios of phospholipids to cholesterol and phosphatidylcholine to phosphatidylethanolamine [62]. The regulation of membrane fluidity is essential for adaptation to temperature variations; therefore, it can be assumed that fish from the control group having low cholesterol and increased phosphatidylethanolamine may be less adapted to wintering. Cell membrane stabilization and the regulation of membrane permeability were proposed earlier [63] as mechanisms of action of the plant flavonoid extract Legalon® (including dihydroquercetin), a licensed drug against liver cirrhosis [64]. It should also be noted that a depressing effect of dihydroquercetin on liver cholesterol through the inhibition of its hepatic biosynthesis and secretion of the cholesterol carrier apolipoprotein (apoB) was reported previously for human cells [65,66]. However, this effect was not observed during our study on trout, and the opposite effect was observed in the November samples.Due to the poor response of the antioxidant system and moderate loss of lipid substrates for oxidation during the bacterial infection, the protective effect of dihydroquercetin and arabinogalactan against severe oxidative stress [43,44,45] in fish fed supplements remains unproven. However, certain differences in cell membrane organization and antioxidant response induced by low ambient temperatures and associated food shortages were revealed between both fish groups.5. ConclusionsSome beneficial effects of the dietary mix of dihydroquercetin and arabinogalactan on physiology responses of farmed rainbow trout, O. mykiss, were revealed. The welfare of the fish in both groups was impacted by a sporadic bacterial infection, including an increased mortality rate and impaired assimilation of nutrients, particularly the malabsorption of essential linoleic 18:2n-6 and alpha-linolenic 18:3n-3 fatty acids. The feed supplementation partially reduced both the lethality of the fish stock and the impairment of their digestion during the infection and at the postinfection period. Therefore, the enhanced resistance to the infection and an accelerated postinfection recovery rate observed in fish fed a supplemented diet could be the results of the biological activity of plant substances. Presumably, plant-origin substances stimulate vitality and natural resistance in farmed fish though increasing the efficacy of fish production. Although the benefits of these feed supplements in fish production are indisputable, further experiments under controlled conditions are required to study the mechanism of the immunomodulating activity of the dietary mix in detail.

animals : an open access journal from mdpi

10.3390/ani13071185

PMC10093030

Previous study showed that moderate amounts of CPC in place of fishmeal can improve the growth performance of Litopenaeus vannamei, but there have been limited investigations of the mechanism of the immunological response to CPC substitution for fishmeal. In this study, high-throughput sequencing analysis was employed to investigate the changes at the gene transcription level, aiming to clarify the mechanism by which CPC substitution for fishmeal affects the immunological response in L. vannamei. The results showed that moderate amounts of CPC in the diet significantly improved the non-specific immune activity and expression of L. vannamei, which may be due to the fact that CPC increases the expression of AMP genes by inhibiting the expression of cactus genes, which may ultimately improve the immunity of L. vannamei. The results of this study contribute to understanding the mechanism of CPC as an alternative to fishmeal and provide useful information for the development of novel protein sources in shrimp feed.

Cottonseed protein concentrate (CPC) is a new non-food protein source with high crude protein, low price, and abundant resources, making it an ideal substitute for fishmeal. In this study, we investigated the effects of CPC re placing fishmeal on the immune response of Litopenaeus vannamei using transcriptome sequencing. L. vannamei (initial body weight: 0.42 ± 0.01 g) were fed four isonitrogenous and isolipid feeds for eight weeks, with CPC replacing fishmeal at 0% (control, FM), 15% (CPC15), 30% (CPC30), and 45% (CPC45), respectively. At the end of the feeding trial, the changes of the activities and expression of immune-related enzymes were consistent in L. vannamei in the CPC-containing group when compared with the FM group. Among them, the activities of ACP, PO, and LZM in the group whose diet was CPC30 were significantly higher than those in the FM group. Moreover, the activities of AKP, SOD, and CAT were significantly higher in the group containing CPC than in the FM group. Furthermore, all CPC groups had considerably lower MDA levels than the FM group. This suggests that the substitution of fishmeal with CPC leads to a significant immune response in L. vannamei. Compared with the FM group, transcriptome analysis identified 805 differentially expressed genes (DEGs) (484 down and 321 up), 694 (266 down and 383 up), and 902 (434 down and 468 up) in CPC15, CPC30, and CPC45, respectively. Among all DEGs, 121 DEGs were shared among different CPC-containing groups compared with the FM group. Most of these differential genes are involved in immune-related signaling pathways. The top 20 signaling pathways enriched for differential genes contained toxoplasmosis, pathogenic Escherichia coli infection, insulin resistance, and Toll and immune deficiency (IMD) pathways, in which NF-kappa-B inhibitor Cactus were involved. In addition, trend analysis comparison of the DEGs shared by the group with CPC in the diet and the FM group showed that Cactus genes were significantly down-regulated in the group with CPC in the diet and were lowest in the CPC30 group. Consistently, the expression of antimicrobial peptide genes was significantly higher in both diet-containing CPC groups than in the FM group. In conclusion, the moderate amount of CPC substituted for fishmeal may improve the immunity of L. vannamei by suppressing the expression of Cactus genes, thereby increasing the expression of antimicrobial peptide (AMP) genes.

1. IntroductionLitopenaeus vannamei, commonly known as the South American white shrimp, belongs to the Arthropoda, Crustacea, Decapoda, Penaeid, and Penaeus [1]. It is native to the eastern Pacific Ocean in northern Peru and Sonora, Mexico, and was introduced to China in 1988. Shrimp is currently one of China’s most important breeding species because of its delicious meat and high yield per unit [2]. With intensive farming of L. vannamei, compound feed gradually became the main production cost of aquaculture [3,4]. Fishmeal is considered the greatest protein source in aquatic animals because of its rich and balanced nutritional components and good palatability as a high-quality protein source for compound feeds [5]. Due to the impact of overfishing and environmental degradation, fishmeal production has declined yearly [6]. Many researchers have been committed to identifying alternative high-quality protein sources for fishmeal, and the promotion and use of sustainable high-quality protein sources have become an inevitable trend in the development of the aquaculture industry [7,8,9,10,11,12].Cottonseed protein concentrate (CPC) is made from high quality cottonseed, with the production process of high temperature steaming and frying replaced by the technology of shell–kernel separation, which could remove some of the water-soluble non-starch polysaccharides, cotton phenols, tannins, phytic acid, and other anti-nutritional factors, minimize the degree of protein heat denaturation, and enhance the protein content and nutritional value [7,13].Numerous studies have shown that CPC acts as a novel protein source with different effects on growth performance and immunity and intestinal flora of different aquatic animals [14,15]. In studies on golden pompano (Trachinotus ovatus) and rainbow trout (Oncorhynchus mykiss), it was found that replacing moderate amounts of fishmeal with CPC did not negatively affect the growth performance and immune response [10,16]. Similar studies have been reported in ♀Epinephelus Fuscoguttatus × ♂Epinephelus Lanceolatu via [17]. In addition, Wang et al. showed that replacing moderate amounts of fishmeal with CPC improved the growth performance, immunological response, and digestibility of L. vannamei [18]. However, excessive replacement of fishmeal with CPC will adversely affect the intestinal health of aquatic animals, causing intestinal inflammation and affecting immune enzyme activity, thus affecting growth performance [14,19]. In addition, Yin et al. showed that excessive replacement of fishmeal with CPC resulted in ♀E. Fuscoguttatus × ♂E. Lanceolatu via of liver in grouper, thus reducing the antioxidant capacity [15]. A similar study found that excessive replacement of fishmeal with CPC negatively affected the growth, body composition, hemolymph indicators, and blood enzyme activity of L. vannamei [20]. Overall, the mechanism of CPC as a novel plant protein source to replace fishmeal in L. vannamei has been less studied up to now.High-throughput sequencing has become a routine experimental technique used in the field of life science [21]. Transcriptome sequencing (RNA-SEQ) is a new technology that uses high-throughput sequencing to study gene transcription, which can comprehensively and quickly obtain the sequence and expression information of almost all transcripts of a specific cell or tissue in a certain state [22,23]. It can accurately analyze gene expression differences, structural variation, molecular markers, and other important problems in life science. High-throughput sequencing has also been widely used in transcriptome analysis to study the mechanisms of growth, development, immunity, and other related life activities. Hou et al. used transcriptome sequencing analysis to find that dietary TWS119 may improve growth performance and immune resistance in L. vannamei by activating the Wnt/β-catenin pathway and inhibiting the activity of LvGSK3β [21]. Yin et al. employed transcriptome sequencing analysis to discover that lipids from L. vannamei are a source of nutrition for Vibrio vulnificus, and they are identified by regulating lipid homeostasis to avoid being infected [22].Although our previous study showed that moderate amounts of CPC in place of fishmeal can improve the growth performance of L. vannamei [18], there have been limited investigations of the mechanism of the immunological response to CPC substitution for fishmeal. In this study, high-throughput sequencing analysis was employed to investigate the changes at gene transcription level, aiming to clarify the mechanism by which CPC substitution for fishmeal affects the immunological response in L. vannamei. The results can provide a theoretical basis for the shrimp feed industry, as it seeks new protein sources to replace fishmeal.2. Materials and Method2.1. Experimental DietsFour isonitrogenous and isoenergetic diets containing different levels of CPC are shown in Table 1. Fishmeal, soybean meal, and peanut meal were used as intact protein sources; fish oil and lecithin were used as the lipid sources; and flour was used as the carbohydrate source. The groups were recorded as FM (0 replacement level, 0%), CPC15 (low replacement level, 15%), CPC30 (middle replacement level, 30%), and CPC45 (high replacement level, 45%). The ingredients were ground into a fine powder, sieved through an 80-mesh size, and precisely weighed according to the formula. The micro constituents were mixed homogenously by the sequential expansion method. Then, the lipids and deionized water were added and thoroughly mixed to obtain a homogenous mixture. Subsequently, the dough was passed through the pelletizer with 1.0 mm and 1.5 mm diameters and dried at 60 °C in a ventilated oven for 0.5 h. Dry pellets were placed in plastic bags and stored at −20 °C until needed.CPC products were provided by Xinjiang Jinlan Co., Ltd. (Xinjiang, China). The amino acid composition of the experimental diets is shown in Table 2.2.2. Feeding Trial and Experimental ConditionsL. vannamei were purchased from Hengxing 863 Fisheries Science and Technology and temporarily reared in a 4.5 m × 4.9 m × 1.8 m outdoor cement pond for two weeks to adapt to the test conditions. The experiment was conducted in a fiberglass tank (0.3 m3). Healthy and strong individuals, with an average body mass of 0.42 ± 0.01 g and without visible injuries, were selected and divided into four groups of three replicates, each with 40 shrimps per replicate and cultured for eight weeks. Shrimps were fed the experimental diets to apparent satiation four times daily (07:00, 11:00, 17:00, and 21:00) for eight weeks. During the experimental period, water quality parameters were measured daily at 7:00 and 21:00 to ensure the temperature was 29.0–30.0 °C, the salinity was 27–30 g∙L−1, the dissolved oxygen level was at least 6.0 mg∙L−1, the pH value was 7.7–8.0, and the ammonia nitrogen level was lower than 0.05 mg∙L−1.2.3. Sample CollectionThis study protocol was approved by the ethics review board of Guangdong Ocean University. All procedures were performed in accordance with the Declaration of Helsinki and relevant policies in China. At the end of the eight-week period, shrimp were fasted for 24 h before collecting samples. Hemolymph collected from three shrimp (per tank) was used as a sample, which was centrifuged (4000× g) at 4 °C for 15 min after storage at 4 °C for 12 h, and the serum supernatant was obtained for the analysis of enzyme activeties. Hemolymph from another three shrimp randomly selected from each tank was withdrawn into modified ACD anticoagulant solution and centrifuged for 5 min (3000× g at 4 °C) to isolate hemocytes, using as a sample. The hemocytes were stored at −80 °C for RNA extraction and used for the detection of gene expression and transcriptomic analysis, as our previous study showed [21].2.4. Evaluation of Non-Specific Immune IndicesThe phosphatase (ACP), alkaline phosphatase (AKP), superoxide dismutase (SOD), phenol oxidase (PO), lysozyme (LZM), hydrogen peroxidase (CAT), alanine aminotransferase (ALT), aspartate transaminase (AST), and malondialdehyde (MDA) were measured using the ACP and AKP assay kit (Cat. No. A060-2), SOD assay kit (WST-1 method) (Cat. No. A001-3), PO kit (Cat. No. H247), LZM assay kit (Cat. No. A059-2), CAT assay kit (visible light) (Cat. No. A007-1), ALT assay kit (Cat. No. A022-2), and AST kit (Cat. No. A247), respectively. The content of MDA was determined using an MDA assay kit (TBA method) (Cat. No. A003-1). All these assay kits were purchased from the Nanjing Jian Cheng Bioengineering Institute (Nanjing, China) and used as per the manufacturer’s instructions.2.5. Gene Expression AnalysisTotal RNA was extracted from hemocytes samples using TransZol Up Plus RNA kits (TransGen, China) following the manufacturer’s protocol, and the quality and concentration were assessed by spectrophotometric analysis (Nanodrop 2000). Total RNA (1 μg) was used for cDNA synthesis by PrimeScriptTM RT reagent kit with gDNA Eraser (AG, China) according to the manufacturer’s instructions. Using qPCR, the gene expressions of PO, copper superoxide dismutase (SOD1), manganese superoxide dismutase (SOD2), CAT, LZM, AKP, ACP, ALT, and AST were determined. The real-time PCR for the target genes was performed on a Light Cycler 480 with a SYBR® Green Premix Pro Taq HS qPCR Kit II by the following program: 1 cycle at 95 °C for 30 s, 40 cycles at 95 °C for 5 s, 57 °C for 30 s, and 78 °C for 5 s. Elongation factor 1α (EF1α, GenBank accession No. GU136229) was used as the internal control. Three replicated qPCRs were performed per sample. The primer sequences are listed in Table 3.2.6. Transcriptome Sequencing and Analysis2.6.1. RNA Extraction and Transcriptome SequencingTotal RNA of hemolymph of L. vannamei from FM, CPC15, CPC30, and CPC45 groups was extracted by the Trizol method. The extracted RNA was treated with RNase-free DNase I to avoid residual genomic DNA contamination. The quality of the isolated RNA was checked by gel electrophoresis on a 1.5% agarose gel and NanoDrop 200 spectrophotometer, USA, and 5 µg total RNA was collected for each group. The library was constructed using the Illumina TruSeq RNA Sample Preparation Kit (Illumina, USA). The mRNA was enriched by magnetic beads, which contained Oligo (dT) (Tiangen, China). Library construction and sequencing were carried out by Guangzhou Gene Denovo Biotech Co., Ltd. (Guangzhou, China). Briefly, the mRNA was enriched with magnetic beads with Oligo (dT), and the fragmentation buffer was added to break the mRNA into small fragments. Using the post-fragment mRNA as a template, random hexamers were used to synthesis the first strand of cDNA. dNTPs, buffer, DNA polymerase I, and RNase H were then used to synthesise the second strand of cDNA using Agencourt AMPure XP Beads reagent. The cassette was purified, and EB buffer eluted with end-repair, base A, and sequencing linker, and the target fragment was recovered by agarose gel electrophoresis. Finally, PCR was used to complete the entire library preparation work. The PCR product was purified to create the final cDNA libraries. Lastly, the library preparations were sequenced on an Illumina HiSeqTM platform that generated ~200 bp pair–end raw reads.2.6.2. Data AnalysisBefore the filtered reads were mapped to the L. vannamei reference genome (NCBI Genome database ID: No. PRJNA438566), using HISAT software [21], raw reads generated by Illumina Hiseq 2000 from the twelve libraries were cleaned using SeqPrep (https://github.com/jstjohn/SeqPrep; accessed on 1 October 2021) and Sickle (https://github.com/najoshi/sickle; accessed on 1 October 2021) software by removing reads with adaptors, reads with more than 10% Q < 20 bases (those with a base quality less than 20), and low-quality sequences (reads with ambiguous bases ‘N’).2.4.3. Serum Biochemical Parameters2.6.3. Differential Expression Analysis and Functional AnnotationAs shown in a previous report, the abundance of all genes was normalized and calculated using uniquely mapped reads by RPKM (read per kilobase of exon model per million mapped reads) [23]. Log2(FC) was used as an indicator of the transcriptomic differences among the FM, CPC15, CPC30, and CPC45 groups. A false discovery rate (FDR) <0.001 was used as the threshold of the p-value in multiple tests to determine the significance of gene expression differences. Genes were considered differentially expressed when the FDR ≤ 0.001 and a greater than twofold change (absolute value of log2 ratio > 1) in expression across libraries was observed. The uniqueness was DEGs if the absolute value of log2(FC) was greater than 1 and the FDR was less than 0.05. DEGs were analyzed using tools in the GO (http://www.geneontology.org/ accessed on 1 October 2021) and KEGG (https://www.genome.jp/kegg accessed on 1 October 2021) databases. The Series Test of Cluster was performed using STEM software with log2 normalization for gene expression preprocessing. The p-value was used to measure the number of genes within the module in relation to the expected value of the random distribution. The lesser the p-value, the more significant the gene set.2.7. Validation of DEGs by qPCRTo validate the RNA-Seq data, RNA samples for transcriptomes were measured using qPCR. Sixteen DEGs and eight AMP genes were selected, and cDNA samples were prepared from transcriptome-sequenced samples using the PrimeScript™ RT kit and the gDNA Eraser (Perfect Real Time) kit (Takara, Japan). Primers were designed using Primer Premier 5 software (Table 3), and their relative expression was quantified using the qPCR method described above.2.8. Statistical AnalysisAll results were subjected to one-way analysis of variance (ANOVA) followed by Duncan’s multiple range test to determine significant differences among treatment groups. Statistical analysis was performed using SPSS version 22.0 (SPSS Inc, Chicago, IL, USA), and all data were expressed as mean ± standard deviation (SD), with p < 0.05 indicating significant differences.3. Results3.1. Non-Specific Immune Indices in the SerumNon-specific immune indices in the serum are shown in Table 4. The activities of ACP and LZM in the serum of L. vannamei showed a trend of increasing and then decreasing with the increase of the CPC substitution ratio. Moreover, the values of ACP and LZM were significantly higher in the CPC30 group than in the FM group. (p < 0.05). In addition, the values of PO were not significantly different among the groups. However, the values of AKP, SOD, and CAT were significantly higher in the CPC15, CPC30, and CPC45 groups than in the FM group. (p < 0.05). Further, the AST and ALT values in the CPC30 group were significantly lower than those in the FM group (p < 0.05). In addition, the levels of MDA were considerably lower in all the CPC-added groups than in the FM group (p < 0.05).3.2. Expression of Non-Specific Immune Genes in HemocytesOverall, the expression of immune genes in the hemocytes could be induced to varying degrees in the CPC-containing group compared with the FM group, and the expression of some genes was significantly higher in several groups (Figure 1). Specifically, the expression of PO and LZM was substantially higher in the CPC30 group compared with the FM group, and there was no significant difference in the expressions of the remaining groups (p < 0.05). However, the ACP expression in the CPC45 group was significantly lower than that in the FM group (p < 0.05). Gene expressions of SOD1, AKP, and AST were significantly higher in most CPC-containing groups than in the FM group (p < 0.05). The highest expression of AKP and AST was found in the CPC45 group, while SOD1 expression was highest in the CPC15 group (p < 0.05). The expressions of SOD2 and ACT peaked in the CPC15 group compared with the FM group and were significantly higher than in the rest of the groups (p < 0.05). In addition, there was no significant difference in the expression of ALT among the groups.3.3. Assembly and Sequence Alignment AnalysisThe raw data of this study have been deposited in the SRA database with the accession number PRJNA808042. Four cDNA libraries from the mRNA extracted from the hemocytes of the L.vannamei in FM, CPC15, CPC30, and CPC45 groups were sequenced. Sequencing was performed using the Illumina platform, and the total number of clean data averaged 85,783,958,047 (bp) after removing articulators and filtering low-quality sequences, as shown in Table 5. Filtering from the raw reads yielded an average of 576,681,882 clean reads. Among them, the FM group generated 140,046,034 high-quality average reads from 140,467,610 raw average reads, and the CPC15 group generated 139,776,820 high-quality average reads from 140,162,764 average raw reads, with Q20% greater than 97% and Q30% greater than 92% of the mean for each group. There were 546,587,036 valid data for the total mean, and the mapping rate was greater than 62% when comparing the analysis statistics with the reference genome data. The assembly results indicate that the sequencing was of good quality and could be used for transcriptome analysis.3.4. Identification of DEGsCompared with L. vannamei in the FM group, 769 (448 up, 321 down), 649 (266 up, 383 down), and 902 (434 up, 468 down) DEGs were identified in the CPC15, CPC30, and CPC45 groups, respectively (Figure 2). As shown in Figure 3, 121 of these DEGs are shared. In addition, CPC15, CPC30, and CPC45 had 367, 295, and 521 unique DEGs, respectively, compared with the FM group.3.5. GO Enrichment Analysis of the DEGsAll DEGs were mapped to terms in the GO database to evaluate their functions. As shown in Figure 4, 51 significantly (p < 0.05) enriched terms were identified and classified into three major functional classes, including biological processes, molecular functions, and cellular components. There were 24, 10, and 17 significantly enriched terms classified into biological processes, molecular functions, and cellular components, respectively. In the category of biological processes, DEGs are annotated mainly to single-organism processes (109 up, 106 down), cellular processes (120 up, 100 down), and metabolic processes (104 up, 80 down). Among the molecular functions, catalytic activity and binding dominate (112 up, 75 down). In addition, the highest content of cellular fractions was classified as cell (130 up, 83 down) and cell parts (130 up, 83 down).3.6. KEGG Enrichment Analysis of the DEGsOn the basis of KEGG enrichment analysis, all DEGs were annotated to cellular processes, environmental information processing, genetic information processing, metabolism, human diseases, and Organismal Systems. Moreover, these annotated DEGs were further classified into 42 subclasses (Figure 5). Consistently, the largest group in these subcategories is the global overview map. However, the top 20 pathways were not identical among the CPC-contained groups and FM group. As shown in Figure 6, only the top 20 pathways of CPC45 compared with the FM group were functionally associated with five KEGG classes, with the top five signalling pathways being aldosterone-regulated sodium reabsorption (ko04960), platinum drug resistance (ko01524), toxoplasmosis (ko05145), small cell lung cancer (ko05222), and pathogenic Escherichia coli infection (ko05130; Figure 6C). Among the top 20 signalling pathways that were significantly enriched in KEGG in the CPC15 group compared with the FM group, seven were related to Organismal Systems, four to metabolism and six to human diseases, with the top five signalling pathways being IL-17 signalling pathway (ko04657), Toll and immune deficiency (IMD) pathways (ko04624), apoptosis (ko04210), carbohydrate digestion and absorption (ko04973), and beta-Alanine metabolism (ko00410; Figure 6A). However, nine of the top 20 pathways significantly enriched in KEGG in the CPC30 group were related to body systems, three to metabolism, and five to human diseases. The first five of these signalling pathways were caffeine metabolism (ko00232), platinum drug resistance (ko01524), steroid biosynthesis (ko00100), Th17 cell differentiation (ko04659), and toxoplasmosis (ko05145; Figure 6B). In addition, compared with the FM group, analysis of DEGs containing the CPC group revealed that among the common differential pathways, platinum drug resistance, toxoplasmosis, pathogenic Escherichia coli infection, steroid biosynthesis, measles, carbohydrate digestion and absorption, insulin resistance, and Toll and IMD pathways were significantly enriched (p < 0.05; Figure 6). However, among the 20 most affected signalling pathways, the NF-kappa-B inhibitor cactus (Cactus) gene is of particular interest because of its involvement in several different signalling pathways, including the toxoplasmosis, pathogenic Escherichia coli infection, insulin resistance, and Toll and IMD pathway (Table 6).3.7. Series Test of ClusterSeries Test of Cluster is shown in Figure 7, and there are four significant trends, with 54, 32, 14, and 9 DEGs, respectively. In addition, Cactus gene expression decreased and then increased with an increasing substitution ratio (Figure 7B). Cactus gene expression was considerably lower in all CPC-containing groups than in the FM group, with the lowest value observed in the CPC30 group (p < 0.05; Figure 7C).3.8. Validation of qPCRSixteen DEGs were randomly selected for validation, including seven up-regulated expression genes and nine down-regulated expression genes. The qPCR results showed the same expression trend as the high-throughput sequencing data. Thus, the qPCR analysis results confirmed the expression of DEGs detected in the high-throughput sequencing analysis (Figure 8).3.9. Expression of AMP Genes in HemocytesAs shown in Figure 9, the expression of Pen2, Pen3, Pen4, and Cru3 genes was significantly higher in the group with CPC with the diet than in the FM group (p < 0.05). In addition, the expression of Cru1, ALF1, and ALF3 genes showed a trend of increasing and then decreasing with the increase in the CPC substitution ratio, and the expression in the CPC30 group was substantially higher than that of the FM group (p < 0.05). However, the expression of the ALF2 gene peaked in the CPC45 group and was significantly higher than that in the FM group (p < 0.05).4. DiscussionA series of studies have confirmed that replacing fishmeal with a certain amount of CPC can improve the immune capacity of aquatic animals [15,24]. However, there have been fewer studies on how CPC improves immune mechanisms in animals. In addition, studies using high-throughput sequencing to assess CPC on gene transcription levels are uncommon. In this study, CPC influenced the immune response of L. vannamei, and the immune response was studied by detecting immune enzymes and analyzing gene expression and transcriptome sequencing. These studies indicate that CPC increased non-specific immune enzyme activity and gene expression in L. vannamei and may induce the expression of AMP genes by regulating Cactus genes, ultimately improving the immunity of L. vannamei.Dietary sources of plant protein have been shown to influence growth performance and innate immune responses in animals [25]. As a type of plant protein, CPC has been proved to improve the growth performance of L. vannamei when used as a replacement of fishmeal [18]. L. vannamei is an invertebrate that relies on its innate immune system to combat disease [18,26,27,28]. Therefore, various non-specific immune indices are often chosen to assess the impact on immunity [29]. Commonly used indices include enzyme activities such as ACP, AKP, SOD, PO, LZM, CAT, and the level of MDA. To determine the effect of the CPC substitution of fishmeal on the immunity of L. vannamei, we analyzed non-specific immunity indicators. In the present study, the activities of AKP, SOD, and CAT were significantly higher in all CPC-containing groups, while the activities of MDA were considerably lower in all CPC-containing groups (p < 0.05). A similar study showed that replacing 36% FM with CPC significantly increased SOD and GSH-PX activities while decreasing MDA levels [10]. In addition, it has also been shown that replacing no more than 24% of fishmeal with CPC can significantly increase the activity of SOD and CAT and decrease the MDA content of groupers [20]. In shrimp, LZM is an important immune enzyme that can lyse microbial cell walls and kill microorganisms and is a direct indicator of immune competence [30]. Ye et al. showed that replacing up to 20% of fishmeal with CPC increased LZM activity in grouper, while replacing no more than 60% resulted in no significant change in LZM activity [15]. In addition, it has been suggested that replacing 45% of fishmeal with CPC can improve the activity of LZM [18]. Consistently, LZM activity was significantly higher in the CPC30 diet group than in the FM group in this study, with no significant difference among the remaining groups. This indicates that replacing fishmeal with a moderate amount of CPC increases LZM activity. ALT and AST are the most important transaminases involved in amino acid metabolism and are direct indicators of health [31]. In the present study, the activity of AST showed an increasing trend with an increasing substitution ratio, and the activity of AST and ALT was significantly higher in the CPC45 group than in the other groups. Similar studies on L. vannamei showed an increasing trend in ALT and AST activities as the proportion of CPC substituted fishmeal increased [20]. In addition, studies have pointed out that many enzymes are dimeric or tetrameric and that each of these subunits is determined by a gene [32,33]. Thus, it is not entirely accurate that the expression level of a gene is consistent with the activity of the corresponding enzyme. The expression levels of PO enzyme and PO gene were not completely consistent in this study. This may be a result of the enzyme activity being influenced by a number of factors. Nonetheless, CPC could have immune-boosting effects in L. vannamei, as evidenced by gene expression of several immunological enzymes. The probable reason is that CPC is a high-quality plant protein source with increased content of essential amino acids, such as Arginine. In addition, reasonable processing has reduced the degree of thermal denaturation of its protein and made it easier for animals to digest and absorb [13].Transcriptome sequencing, a common tool in biology research at present, is also extensively used to analyses the mechanisms associated with protein source substitution in diets on the growth, metabolism and immunity of organisms [21,34,35]. In the present study, a comparative transcriptome analysis of L. vannamei containing the CPC and FM groups was performed to investigate the mechanism of the effect of the CPC substitution for fishmeal. GO and KEGG analysis showed that the distribution of DEG enrichment functions and the classification of DEG-related signaling pathways were essentially the same among the different CPC-containing and FM groups. Therefore, it can be seen that the effects of different levels of CPC-substituted fishmeal on L. vannamei at the transcriptome level are approximately the same. In the DEG enrichment analysis, it was found that Cactus genes affect many different pathways of interest. Examples include toxoplasmosis, pathogenic Escherichia coli infection, insulin resistance, and Toll and IMD pathway. These pathways are all related to the immune pathway, suggesting that CPC substitution for fishmeal affects the immune response of shrimp. The NF-κB pathway, as is well known, is often considered the main regulatory pathway of the shrimp immune response [36]. Bacterial disease infections can be recognized by Toll and IMD pathways, which then activate two NF-κB transcription factors, Dorsal and Relish, respectively, ultimately promoting the expression of various AMP genes [37,38,39]. Cactus functions as an inhibitor of Dorsal and can negatively regulate the NF-κB pathway [40,41]. It contains an N-terminal regulatory region responsible for ubiquitin recognition and proteasomal degradation and has an adjacent Ankyrin repeat sequence that is capable of binding to the Rel-homology region [42]. The Rel-homology region and a destabilizing C-terminal PEST domain are required to inhibit DNA binding [42]. In the present study, Cactus expression was considerably decreased in the CPC-containing group compared with the FM group, reaching its lowest point in the CPC30 group. Consistently, the expression of AMP genes was also differentially induced to elevated levels in the CPC-containing group of the diet. In conclusion, the CPC-containing group in the diet may maintain the normal regulation of the NF-κB pathway in animals by regulating the expression of the Cactus gene, promoting the expression of various AMP genes, and ultimately improving the disease-fighting immunity of L. vannamei. This was further verified by the gene expression of these AMP genes. However, the exact mechanism needs further validation.5. ConclusionsIn the present study, dietary moderate amounts of CPC significantly improved the non-specific immunize activity and expression of L. vannamei, which may be due to the fact that CPC increases the expression of AMP genes by suppressing the expression of Cactus genes, thereby possibly resulting in ultimately improving the immunity of L. vannamei. The results of this study contribute to understanding the mechanism of CPC as an alternative to fishmeal and provide useful information for the development of novel protein sources in shrimp feed.

animals : an open access journal from mdpi

10.3390/ani11041039

PMC8067852

With the increasing pressure to address the problems of bacterial resistance and drug residues, medicinal herbs are gradually taking a more important role in animal production. Scutellaria baicalensis is a common and widely used Chinese medicinal herb. The main bioactive compounds in the plant are baicalein and baicalin. These compounds have many biological functions including anti-oxidation, antipyretic, analgesic, anti-inflammatory, antiallergic, antimicrobial, immunomodulatory, and antitumor effects. S. baicalensis and its extracts can effectively promote animal growth, improve the production performance of dairy cows, reduce the stress and inflammatory response, and have effective therapeutic effects on diseases caused by bacteria, viruses, and other pathogenic microorganisms. This paper summarizes the biological function of S. baicalensis and its application in sustainable animal production to provide a reference for future application of S. baicalensis and other medicinal herbs in animal production and disease treatment.

Drugs have been widely adopted in animal production. However, drug residues and bacterial resistance are a worldwide issue, and thus the most important organizations (FAO, USDA, EU, and EFSA) have limited or banned the use of some drugs and the use of antibiotics as growth promoters. Natural products such as medicinal herbs are unlikely to cause bacterial resistance and have no chemical residues. With these advantages, medicinal herbs have long been used to treat animal diseases and improve animal performance. In recent years, there has been an increasing interest in the study of medicinal herbs. S. baicalensis is a herb with a high medicinal value. The main active compounds are baicalin and baicalein. They may act as antipyretic, analgesic, anti-inflammatory, antiallergenic, antimicrobial, and antitumor agents. They also possess characteristics of being safe, purely natural, and not prone to drug resistance. S. baicalensis and its extracts can effectively promote the production performance of livestock and treat many animal diseases, such as mastitis. In this review, we summarize the active compounds, biological functions, and applications of S. baicalensis in the production of livestock and provide a guideline for the application of natural medicines in the production and treatment of diseases.

1. IntroductionFood from animal sources provides the human body with protein, fat, minerals, vitamin A, B vitamins, and other nutrients, thus giving animal products an important role in the food supply worldwide. Improving the health and production of food animals would also benefit human health. Manufactured chemical and antibiotic feed additives are widely used in animal husbandry, resulting in growing problems with antibiotic resistance and chemical residues, which are becoming limiting factors in animal husbandry [1,2,3,4]. Concerns over these issues have prompted various organizations and countries in the world to introduce policies for banning or restricting the use of chemicals and antibiotics as feed additives [5]. The “one health” approach is used worldwide [6,7]. Future actions and the World Health Organization action plan against antimicrobial resistance are based on best practices in implementing and monitoring the “one health” plans, supporting novel solutions to prevent and treat infections, thereby increasing the efforts in terms of combating antimicrobial resistance and related risks worldwide [8]. Medicinal herbs, with their unique advantages as natural products, might be one of the solutions to these issues in animal production.Medicinal herbs and their extracts are natural, efficient with few side effects, and have minimal risk of inducing bacterial resistance [9]. At present, the application of medicinal herbs and their extracts in animal production is still at the emerging stage. It has been proven that medicinal herbs can effectively improve the utilization of feed protein, and the performance of growth and reproduction in animals [10,11], modulate rumen fermentation towards high efficiency [12,13,14], increase feed intake and digestibility by affecting feed intake behaviors and the secretion of digestive enzymes [14], alleviate the effects of stress by enhancing antioxidant ability [15,16], and consequently improve the health status of animals. Furthermore, the antimicrobial effects of medicinal herbs can also be used for the prevention and treatment of diseases. Thus, medicinal herbs as alternative natural feed additives to replace chemicals and antibiotics have attracted much attention [17].Scutellaria baicalensis is an herb that has roots of high medicinal value. It has the functions of clearing heat and detoxification, purging fire and drying dampness, improving fertility, and hemostasis. It is also widely used in clinical practice. This paper reviews the research progress on the use of S. baicalensis and its extracts in the sustainable production and health of animals. The biological functions of S. baicalensis and its extracts and their applications in animal production and disease prevention are briefly described to form a reference for the future research and application of herbal preparations in animal production.2. Scutellaria baicalensis and Its Active ComponentsScutellaria baicalensis, also known as Chinese skullcap or Baikal skullcap, is a perennial herb of the family Lamiaceae. It mainly grows at an altitude of 60–2000 m in sandy soil on sunny slope land. The plant is natively distributed in East Asia and widely cultivated in European and American countries. China is the main producer of S. baicalensis for medicinal use [18]. The dry roots of the plant are often used as medicine. The chemical components of S. baicalensis roots are mainly flavonoids, anthraquinones, lignin, organic acids, volatile oils, and other compounds. Hereafter, S. baicalensis roots are sometimes referred to as S. baicalensis. The characteristic components of S. baicalensis are baicalin, baicalein, scutellarin, and scutellarin. Baicalin is a monomer active component extracted from the roots of S. baicalensis. Baicalein is the aglycon form of baicalin, which is a typical component of S. baicalensis [19].Baicalein and baicalin are the main active compounds of S. baicalensis, which are flavonoids. Baicalin is the glucuronide of baicalein [19]. The molecular formula of baicalein is C15H10O5 with a relative molecular mass of 270.24, and baicalein can be crystallized as yellow needles and is easily soluble in alcohol, acetone, and slightly soluble in chloroform. The chemical structure is shown in Figure 1. The molecular formula of baicalin is C21H18O11, with a relative molecular weight of 446.36. Pure baicalin exists as light yellow crystal needles and is soluble in hot acetic acid, insoluble in acetone, methanol, and ethanol, and almost insoluble in water. The molecular structure is shown in Figure 2.3. Biological Functions of S. baicalensisS. baicalensis has a long history as a medicinal herb due to its extensive biological and pharmacological activities. Its use in treating lung and liver diseases was first recorded in Sheng Nong’s herbal classic during the Han Dynasty in China. S. baicalensis has a cooling effect and is bitter in taste. It has antimicrobial, anti-inflammatory, antipyretic, and analgesic effects and improves animal growth performance (Table 1). It plays an important role in livestock and poultry production [20]. Baicalin, as the most abundant flavonoid in the roots of S. baicalensis, is stable in acidic environments and organic solvents but unstable in alkaline environments and in the plant. The absorption rate of baicalin is low via oral administration. However, baicalein formed from enzymatic hydrolysis of baicalin in the intestinal tract is easily absorbed into the blood and then is chemically converted into baicalin and other metabolites in the liver showing biological activities [21,22].3.1. AntioxidationFlavonoids have significant antioxidant activity. There are two antioxidant mechanisms of flavonoids. One is a direct hydrogen pumping reaction mechanism. The antioxidant molecule loses the phenolic hydroxyl hydrogen atom to generate phenolic oxygen radicals, and its antioxidant activity depends on the ease of phenolic hydroxyl O-H bond breaking. The other one is a single electron transfer mechanism. The antioxidant transfers one electron to the reactive oxygen radicals to generate the flavonoid radical cation [67]. The A-ring of baicalin contains an o-diphenol structure, whereas the molecular structure of baicalein contains three hydroxyl groups [68]. The key to baicalin’s antioxidant mechanism is the regulation of nuclear factor-erythroid 2-p45 derived factor 2 (Nrf2). Nrf2 is the core transcription factor that regulates the cellular oxidative response. It can induce the activity of the antioxidative reactive protein (Ares) and thus inhibit the formation of oxidative free radicals (ROS), consequently reducing oxidative stress to keep the stability of the intracellular environment [25]. Baicalein can scavenge free radicals by decarboxylation in the body to regulate oxidative stress. Baicalein has a strong scavenging effect on free radicals such as alkane peroxide and superoxide anion [45].Gao et al. [23] detected free radicals using the electron spin resonance (ESR) technology and found that baicalin and baicalein have a strong scavenging effect on hydroxyl radicals, 2,2-diphenyl-1-picrylhydrazyl (DPPH) radicals, and alkyl peroxide radicals. In addition, baicalein can inhibit xanthine oxidase activity and thus inhibit the production of oxygen free radicals during its metabolism [19]. Peroxides (such as hydrogen peroxide) can induce neuronal cell injury. Baicalin and baicalein have significant protective effects on the oxidative injury of human neuronal cells induced by H2O2, and baicalein has stronger antioxidant activity than baicalin [43]. The free radical scavenging and antioxidant effects of baicalin and baicalein can be used to effectively treat free radical and oxidative stress-related diseases.To protect the heart, baicalein can concentration-dependently reduce hypoxia reoxygenation induced myocardial death and apoptosis. Further studies revealed that the cardioprotective effects of baicalein were mediated through μ- and δ-, but not κ-opioid receptors and their associated signal transduction pathways, such as protein kinase C and ATP-sensitive potassium (KATP) channels [44]. Baicalin can effectively attenuate oxidative stress and apoptosis by activating the Nrf2 signaling pathway, thereby protecting the thymus from structural and functional damage mediated by mycoplasma infection [24].3.2. Antipyretic and Analgesic EffectsBaicalin has a significant antipyretic effect (Figure 3), and the mechanism underlying the effect may be related to the reduction of the contents of tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β) in a significant dose-dependent manner [27]. The antipyretic effect of baicalin is via inhibiting the hydroxyl radical pathway of the N-methyl-D-aspartate receptor in the hypothalamus and accumulating TNF-α during fever. Studies have shown that baicalin can play an antipyretic role by reducing the concentration of prostaglandin E2 (PGE2) and cyclic adenosine monophosphate (cAMP) in the hypothalamus [28]. Other studies have found that baicalin reduces heat-stress-induced apoptosis by regulating the Fas/FasL pathway and upregulating heat shock protein 72 (HSP72) expression in bovine testicular Sertoli cells [29]. In addition to its antipyretic effect, baicalin also has a substantial analgesic effect [30]. It is found that baicalin has anti-inflammatory and analgesic effects in the studies on the evaluation of its effects on inflammation, pain, and edema [31]. Pain-related animal models were used to evaluate the analgesic activity of UP446 (a standard bioflavonoid component of baicalin and catechol). It was found that the pain sensitivity of hyperalgesic animals induced with carrageenan pretreatment reduced by 39.5% after an oral administration at 150 mg UP446/kg body weight. In the writhing test and formalin test, a single dose of UP446 orally administered at 100 mg/kg body weight showed 58% and 72% inhibition of pain sensitivity, respectively [32].3.3. Anti-Inflammatory and Antiallergic EffectsS. baicalensis and its extracts have a strong anti-inflammatory function, participate in the regulation of a variety of inflammatory factors, and also have cartilage protection effects [52]. Baicalin and baicalein treat skin diseases by regulating enzymes and pro-inflammatory factors such as interleukin-6, oxygenase-1, TNF-α, and cell adhesion factors [33,46,53]. In vitro experiments showed that baicalin has a protective effect against lipopolysaccharide (LPS)-induced inflammatory injury in mammary epithelial cells and can play an anti-inflammatory role by inhibiting nuclear factor kappa-B (NF-κB) activation and mitogen-activated protein kinase (p38) phosphorylation, and upregulating HSP72 to alleviate LPS-induced inflammation and cell apoptosis in mammary epithelial cells [69,70]. A study on rats found that baicalein can inhibit the release of interleukin-8 and the synthesis of cyclooxygenase-2 and increase the formation of heat shock protein 70 to improve the anti-inflammatory ability of the body, consequently blocking the inflammatory injury caused by inflammatory factors [71]. Baicalin can inhibit the production of inflammatory factors TNF-α, IL-1β, Interleukin-6 (IL-6), Interleukin-17 (IL-17), matrix metalloprotein-9 (MMP-9), and regulate NF-κB signaling pathway to have anti-inflammatory effects [72].Baicalein and baicalin can also inhibit vascular permeability, the adsorption and migration ability of leukocytes, and inhibit the production and release of inflammatory mediators, providing a reference and theoretical basis for the development of drugs to treat cardiovascular-related diseases [73]. In addition, baicalin can reduce inflammation and edema as a dual inhibitor of cyclooxygenase and 5-lipoxygenase [34]. S. baicalensis exerts antiallergic effects mainly by inhibiting the mast cell degranulation process and inhibiting the release of the histamine slow-reacting substance of anaphylaxis (SRS-A). This may alleviate the itching, gastrointestinal contraction, and other symptoms caused by type I, II, and IV allergic reactions in animals, without apparent side-effects [59,60,61,62,63,64,65].3.4. Antimicrobial EffectThe plant extract of S. baicalensis has a broad-spectrum inhibitory effect on the growth of bacteria, including mycoplasma and spirochetes type bacteria, fungi, and viruses. Baicalin and baicalein are the active compounds that inhibit the growth of bacteria by destroying the nucleic acid formation of bacteria and altering the energy metabolism of bacteria, as well as inhibiting the formation of biofilms of bacteria such as Klebsiella pneumoniae and Pseudomonas aeruginosa [74]. In experiments evaluating the inhibitory effect of baicalin on milk-derived Escherichia coli, it was found that baicalin had an inhibitory effect on E. coli in vitro [35]. After the use of baicalin, the sensitivity of most strains to other antimicrobial agents was enhanced [36].In addition, S. baicalensis and its extract have a bactericidal effect on Helicobacter pylori, Staphylococcus aureus, and other pathogenic bacteria [55,56]. Baicalin in conjunction with penicillin and ciprofloxacin has a synergistic effect on the treatment of penicillin-resistant S. aureus and methicillin-resistant S. aureus (MRSA) [37,38]. Baicalein can prevent the formation of bacterial biofilms and disrupt the biofilms and consequently reduce the production of staphylococcal enterotoxin A and α-hemolysin, thereby inhibiting the growth of S. aureus [47]. In addition, baicalein can reduce the pathogenic ability of bacteria such as S. aureus and E. coli by disrupting their cell wall integrity, reducing bacterial enzymatic activities, and inhibiting bacterial energy production and nucleotide synthesis [48]. It was found that the extract of S. baicalensis has significant antifungal effects, such as Candida albicans, Aspergillus fumigatus, Hydramycetes, etc. Baicalein and wogonin have strong antifungal activities, which may induce programmed apoptosis of the fungal cells via excessive production of reactive oxygen species [49]. In addition, the combination of baicalein with fluconazole and other antifungal drugs can more effectively treat fungal infections such as Candida [50]. The extract of S. baicalensis has an in vitro inhibitory effect on tick-borne encephalitis virus via direct inhibition of the adsorption and intracellular replication of tick-borne encephalitis virus [57]. S. baicalensis also has marked therapeutic effects on viral diseases such as influenza, infectious bronchitis, and viral diarrhea [54,58]. Relevant studies have shown that baicalin has a polyphenolic hydroxyl structure and prevents glyoxal-induced cystatin aggregation by affecting its aggregation process, which opens a new way for the treatment of protein misfolding diseases [75]. In addition, baicalin extract has potential applications in the fields of anti-fibrosis, anti-cancer, anti-aging, anti-depression, and immune regulation [76,77,78,79,80].3.5. Antitumor ActivityBaicalin is able to induce the apoptosis of tumor cells. Baicalin could promote apoptosis in pancreatic cancer cells (SW1990 cell line) via dose-dependent upregulation of the expression of mitochondrial Bax (Bcl2-associated Xprotein) and cleavage-type enzymes caspase-3, and p53, consequently significantly decreasing B-cell lymphoma/lymphoma 2 (Bcl-2) protein levels, possibly activating c-Jun N-terminal kinase/forkhead box protein O1/Bcl-2 interacting mediator of cell death (JNK/Foxo1/BIM) pathway [39]. Baicalin can induce apoptosis in hepatoma cell lines HepG2 and SMMC-7721 cells [40]. It can also induce colon cancer cell apoptosis through microRNA-217/dickkopf (miR-217/DKK1)-mediated inhibition of the wingless-related integration (Wnt) signaling pathway [41]. Moreover, the combination of baicalin with drugs such as hexamethylene bis-acetamide (HMBA) showed promising efficacy against leukemia [42].Baicalin can inhibit tumor invasion and metastasis. Baicalin effectively inhibits the invasion, migration, and adhesion abilities of multiple tumors by inhibiting the expression levels and activities of mitochondrial membrane potential 2 (MMP2) and mitochondrial membrane potential 9 (MMP9) [81]. Phosphorylation of AMP-activated protein kinase (AMPK) leads to mitochondrial membrane potential (MMP) expression and promotes tumor invasion and metastasis [82].Baicalin inhibits non-small cell lung cancer migration and invasion by down-regulating MMP expression through activation of the mammalian target of rapamycin (mTOR) and silencing information regulator1/AMP-activated protein kinase (SIRT1/AMPK) signaling pathways [83]. In addition, baicalin also has the ability to induce cancer cell cycle arrest [84,85], overcome the drug resistance of tumor cells [86], and modulate tumor-associated inflammatory microenvironment [70,87].Baicalein inhibits tumor cell development by modulating different metabolic signaling pathways, as well as decreasing tumor growth and metastasis rates, significantly decreasing CD31 (an endothelial cell marker) and α-smooth muscle actin (α-SMA, a parietal cell marker) expression and inducing cell death in tumor tissues [51].Angiogenesis is the key process to promote cancer, and S. baicalensis has anti-angiogenesis activity in vitro. Liu et al. [66] evaluated the potential of baicalin and baicalein as antiangiogenic agents through the analysis of chicken chorioallantoic membrane (CAM) and the culture of human umbilical vein endothelial cells (HUVEC) in vitro. The study showed that baicalin and baicalein had the potential of antiangiogenesis and inhibited the migration of endothelial cells and the differentiation of endothelial cells into tubular branch networks in a dose-dependent manner [66]. Wang et al. [88] also determined the activity of angiogenesis through the proliferation of blood vessels on the chicken CAM model and cultured bovine aortic endothelial cells (BAEC). The study indicated that S. baicalensis could significantly inhibit the activity of angiogenesis.Wogonoside is an active component of S. baicalensis. It can inhibit the migration and angiogenesis of HUVEC stimulated by LPS, as well as the microvascular sprouting of rat aortic rings in vitro. It may also have potential therapeutic value for diseases related to inflammation and angiogenesis [89]. Oroxylin A (one of the active components of S. baicalensis and one of the metabolites of baicalin in vivo) can also inhibit LPS-induced angiogenesis and may affect the lipopolysaccharide/Toll-like receptor 4 (LPS/TLR4) signaling pathway [90]. Wogonin also inhibits H2O2-induced angiogenesis by inhibiting the phosphatidylinositol 3-kinase/protein kinase B (PI3K/Akt) and NF-κB signaling pathway [91].4. Application of S. baicalensis in Sustainable Animal Production for Better PerformanceMedicinal herb extracts have been used as feed additives, veterinary medicines, and environment-friendly disinfectants [92] with support from the public due to their natural, safe, and effective characteristics. As an important medicinal herb, S. baicalensis can enhance the immunity of the body, reduce allergic reactions, protect the liver, and treat the body’s peroxidative reactions [19]. S. baicalensis could play an important role in the improvement of animal growth and production performance. Table 2 summarizes the main effects of S. baicalensis in animal production.4.1. PoultryZhou et al. [93] evaluated the effects of baicalein on growth performance, immunity, and antioxidant activity of broilers at doses of 100 and 200 mg/kg diet. Compared with the basal diet, the baicalein-supplemented diet had no significant effect on the average daily feed intake but significantly increased body weight, average daily gain, and feed conversion efficiency of broilers at the age of 21–42 days and 7–42 days. The best growth performance was observed at the dose of 200 mg/kg diet. Compared with the control group, baicalein significantly increased CD3+/CD4+ and CD3+/CD8+ ratio, interferon γ (IFN-γ) concentration, anti-IB antibody titer, and spleen index. Total cholesterol, triglyceride, and low-density lipoprotein cholesterol were significantly decreased after baicalein intake compared with basal diet, while superoxide dismutase (SOD), glutathione peroxidase (GSH-Px), and catalase (CAT) activities in serum were increased with baicalein supplementation, and total antioxidant capacity (T-AOC) activity, total superoxide dismutase (T-SOD), and GSH-Px levels in liver tissues were significantly increased, while the intake of baicalein was significantly decreased. Malondialdehyde levels in serum and meat tissues were reduced. Baicalein can be used as an effective natural feed additive in broiler diets, and a 100–200 mg/kg diet is evaluated as the optimal dose. Króliczewska et al. [94] evaluated the effect of the roots of S. baicalensis on the production performance of broilers and found that the addition of S. baicalensis roots to the diet (5–15 mg/kg diet) was able to increase broiler body weight and feed conversion efficiency but had no effects on the quality and chemical composition of broiler leg muscles.The addition of S. baicalensis extracts to the drinking water of turkeys (0.009, 0.018, and 0.036 mL/kg bodyweight) was able to change the concentrations of sodium, potassium, calcium, magnesium, copper, zinc, and iron in plasma, showing an upward trend in the concentrations of calcium and magnesium and a downward trend in the concentration of sodium, potassium, copper, zinc, and iron [95]. In addition, the dietary supplementation of S. baicalensis extracts to laying hens at a dose of 5 g/kg diet effectively increased the weight of the eggs, decreased the microbial content in the cecum, reduced the amount of propylene glycol, and delayed lipid oxidation in the eggs [96]. Fermented medicinal plants (Gynura procumbens, Rehmannia glutinosa, and S. baicalensis) at doses of 0.5–2 g/kg diet can be used as an alternative to reduce the use of antimicrobial agents in broilers for improved production performance [108].Medicinal herbs are often applied in the form of a composite preparation after the assessment of compatibility. Varmuzova et al. [97] found that Curcuma (Curcuma longa) extract alone was not enough to reduce intestinal inflammation caused by heat stress. However, the mixture of C. longa and S. baicalensis plant extracts as feed additives reduced intestinal inflammation caused by a high air temperature or by Salmonella enteritidis, reduced the counts of Salmonella in cecum-midgut, and had no negative effects on body weight or humoral immune response. Using the 16S rRNA sequencing technique, it was found that the dietary supplementation of the two plant extracts had no effects on microbial diversity. However, if the plant extract supplements are provided to chickens infected with S. enteritidis, Enterococcus faecalis, and Lactobacillus spp., the bacterial genera with known positive effects on intestinal health are actively selected. Therefore, the supplementation of chicken feed with Curcuma and Scutellaria plant extracts can be used in poultry production to effectively reduce intestinal inflammation and improve chicken production performance [97]. Liu and Kim [109] found that the addition of S. baicalensis and Lonicera japonica extracts to the feed at a dose of 0.25–0.5 g/kg diet alleviated the detrimental effects of seasonal heat stress on the production performance of laying hens. Lv et al. [98] found that the dietary supplementation of S. baicalensis and L. japonica extracts could improve the growth performance of broilers, promote the development of immune organs, and improve the antioxidant function. Their results showed that the addition of 500 mg plant extracts/kg of diet could increase the average daily weight gain at the age of 21–42 days and 1–42 days, increase feed conversion efficiency, increase the thymus and bursa index of 42-day-old broilers, increase the activity of serum catalase, and decrease the level of malondialdehyde in 21- and 42-day-old broilers.Cheng et al. [110] studied the anti-inflammatory effect of baicalin on LPS-induced chicken liver inflammation and its molecular mechanism. Histopathological changes, serum biochemical analysis, nitric oxide (NO) level, and myeloperoxidase activity showed that baicalin pretreatment alleviated LPS-induced liver inflammation. ELISA and qPCR analysis showed that baicalin dose-dependently inhibited the formation of IL-1β, IL-6, and TNF-α. In addition, baicalin significantly decreased the mRNA expression of inducible nitric oxide synthase (iNOS) and cyclooxygenase-2 (COX-2). Molecular studies showed that baicalin pretreatment inhibited TLR4 expression and the activation of the NF-kB signaling pathway. Baicalin pretreatment down-regulated TLR4 expression and inhibited NF-kB activation, thus had a protective effect on chicken liver from LPS-induced inflammation. Ishfaq et al. [111] also showed that baicalin treatment could effectively prevent Mycoplasma gallinarum-induced inflammation, apoptosis, and energy metabolism dysfunction, providing a basis for new therapeutic targets to control M. gallinarum infection. However, Króliczewska et al. [112] found that excessive supplementation of S. baicalensis roots (5–15 g/kg diet) may cause immunosuppression and may have a negative impact on the development of immune organs. S. baicalensis roots inhibited the formation of radially segmented nuclei, showing anti-metastatic properties and phagocytosis of chicken heterophils.Dietary supplementation of S. baicalensis extracts improves growth performance, blood parameters, nutrient digestibility, and meat quality of broilers [113]. The combined use of S. baicalensis extract and Zn in the feed not only improved the quality of broilers but also showed antioxidant capacity [114]. Liang et al. [99] examined the effects of baicalin on the growth performance and intestinal bacterial community of broiler chickens at different doses supplemented to the basal diet and found that the average body weight of broilers in all treatment groups was increased compared with the control group. Among different treatments, the dose of 10 mg baicalein/kg diet had the highest body weight and the supplementation of 5 mg/kg baicalein to the diet increased the average daily feed intake. Compared with the control group, the number of E. coli and Salmonellas were decreased and the number of Lactobacillus and Bifidobacterium were increased in all experimental groups. The addition of an appropriate dose (10 mg/kg diet) of baicalein in the diet could promote the growth of broilers and modulate the intestinal microbial community. Wang et al. [115] studied the potential alleviative effect of combined plant extracts of L. japonica and S. baicalensis (active ingredients are chlorogenic acid and baicalin) on intestinal damage and bacterial dysbiosis caused by Salmonella pullorum. It was found that this preparation could effectively alleviate the intestinal damage and the loss of animal performance caused by S. pullorum, for which the regulation of intestinal microbial composition by the plant extracts is an important mechanism of action.4.2. SwineLi and Diao [100] found that the addition of fermented S. baicalensis to the diet could enhance the appetite of weaned piglets at the age of 28 days, increase the average daily intake, reduce the feed-to-weight ratio and diarrhea rate, and improve the feed conversion rate. Among the different doses, a dose of 1.5 g/kg diet had the best effect on daily intake. With this dosage, feed intake and daily weight gain were increased by 13% and 33%, and the feed-to-weight ratio and diarrhea rate decreased by 15% and 31% compared with the control group, respectively.S. baicalensis and its extracts can effectively improve the growth performance and manipulate the intestinal microflora of pigs. Zhao et al. [101] evaluated the dietary supplementation of fermented medicinal plants (FMP) consisting of Gynura procumbens, R. glutinosa and S. baicalensis in weaned piglets. It was found that FMP additives significantly increased the average daily weight gain, average daily feed intake, the ratio of body weight to feed, and the apparent total tract digestibility of dry matter, nitrogen, and gross energy, and decreased the concentrations of ammonia, total thiol, and hydrogen sulfide. In addition, the diarrhea rate of piglets was reduced with the addition of FMP. Thus, FMP can be used to improve growth performance and nutrient digestibility, mitigate harmful gas emissions from feces, and reduce the early diarrhea rate of weaned piglets.The addition of FMP containing S. baicalensis to the diet of fattening pigs was able to improve body weight gain, feed conversion efficiency, and digestibility and mitigate harmful gas emissions [116,117]. The combination of S. baicalensis, Gardenia jasminoides, and lactic acid bacteria can accelerate the clearance time of bacteria in feces, improve the immunity of infected pigs, and regulate the enzymatic activity of intestinal microorganisms in order to convert herbal compounds into active compounds [102]. Therefore, this mixture as a feed additive can be a potential preventive agent for Salmonella infection.The addition of S. baicalensis extract to the diet at 1000 mg/kg diet can effectively alleviate diarrhea in weaned piglets and may reduce the expression of inflammatory cytokines by inhibiting the nuclear factor kappa-B/mitogen activated protein kinase (NF-κB/p38) signaling pathway, thereby improving intestinal health [103]. Baicalin in combination with aluminum can reduce piglet diarrhea. Genomic analysis indicated that the baicalin aluminum complex preparation could modulate the gut microbiota of diarrheal piglets and was associated with flagellar assembly, bacterial chemotaxis, lipopolysaccharide biosynthesis, ATP binding cassette transporters, biosynthesis of amino acids, and phosphotransferase systems [104]. The addition of mixed medicinal herbs containing S. baicalensis to the diet of pregnant and lactating sows can lead to decreased weight loss after delivery and can improve litter performance [107].In addition, baicalin improves parthenogenetically activated and in vitro fertilized pig embryos by inhibiting the production and apoptosis of reactive oxygen species, regulating mitochondrial activity, and activating sonic hedgehog signaling [118].4.3. RuminantsS. baicalensis extracts can alter the microbial flora in the rumen of ruminants, thereby promoting the fermentation of forage in the rumen [119]. S. baicalensis has evident antipyretic effects and has been incorporated into many medicinal herb preparations [120], and thus has great potential for reducing the impact of heat stress in intensive and large-scale ruminants farming. Baicalin can be used as an anti-apoptotic agent to alleviate heat stress-induced apoptosis. Guo et al. [29] examined the effects of baicalin on heat stress-induced apoptosis of bovine Sertoli cells using flow cytometry and found that pretreatment with baicalin at 1, 10, and 20 μg/mL significantly reduced apoptosis induced by heat stress at 43 °C for 1 h. There was a dose-dependent relationship between baicalin concentration and the rate of cell survival.Systemic inflammation is more common in early lactation dairy cows and can result in a decline in milk production. S. baicalensis contains flavonoids with anti-inflammatory and antioxidant effects. Olagaray et al. [26] incorporated S. baicalensis extracts into pelleted feed and provided them to lactating Holstein dairy cows at a dose of 10 g extracts/d (containing 3.3 g of flavonoid baicalin/d), which resulted in a reduced incidence of mastitis. A long-term supplementation (60 days) significantly improved milk yield in the whole lactation period, while short-term (5 days) administration did not change the milk production. The mechanisms underlying the improvement of lactation performance warrant further study.5. Application of S. baicalensis in Disease Prevention and TreatmentMedicinal herbal preparations are mainly natural plant extracts, which can not only be used as feed ingredients but also have significant anti-parasitic, antibacterial, anti-viral, and other effects due to the effective bioactive compounds. Thus, they are of great value in the prevention and treatment of animal diseases.Liao et al. [105] studied the therapeutic effect of baicalein injection on artificially infected swine edema disease in weaned piglets aged 28–35 days. Intraperitoneal injection of Vibrio cholerae O139 bacteria suspension was used to establish the swine edema disease model of artificially infected pigs, and different doses of S. baicalensis extract were injected for the treatment of the disease. It was found that the doses of 0.2 and 0.4 mL S. baicalensis extract/kg body weight significantly increased the weight gain rate of the experimental animals. The weight gain rate and total effective rate were 96% and 90%, respectively. This indicates that the injection of S. baicalensis extract could effectively treat the edema disease of artificially infected pigs.Baicalin can effectively inhibit Mycoplasma gallisepticum induced immune deficiency and attenuate inflammatory response and apoptosis in the chicken bursa of Fabricius. Baicalin attenuates the levels of pro-inflammatory factors, and suppresses NF-κB expression at the protein and miRNA levels, alleviating reduction in M. gallisepticum induced CD8+ cells and bacterial burden in the bursa [121]. In addition, treatment with baicalin (450 mg/kg BW for 5 days) could effectively alleviate the extent of lesions in lung and tracheal tissues, alveolar space, and mucosal layer thickening were restored, cilia were gradually restored, and the IL-17 signaling pathway-related genes were significantly reduced. The activities of cytokines and chemokines (CXCL1, CXCL2, MMP1, GMCSF, and MUC5AC) were decreased significantly. Baicalin was able to effectively treat co-infection caused by M. gallinarum and E. coli [122]. Baicalin can resolve intestinal dysbacteriosis caused by H9N2 subtype avian influenza infection by modulating lactic acid bacteria, thus preventing the high mortality caused by the secondary infection of Escherichia. Baicalin had beneficial effects for clinical prevention and control of H9N2 Subtype Avian infection and secondary bacterial infections and inflammation by inhibiting the loss of intestinal structure and improving antioxidant capacity, affecting blood biochemical indexes, and inhibiting the production of inflammatory factors [123]. Baicalin effectively treats polyserositis caused by Glaesserella parasitis infection by alleviating the downregulation of mRNA for tight junctions, preventing the abnormalities, and maintaining the integrity of tight junctions, and is a potential natural medicine for the prevention and treatment of G. parasitis [124].Baicalin showed significant antibacterial activity against E. coli in vitro. In a study by Zhao et al. [36], the minimum inhibitory concentration of baicalin against E. coli isolated from mastitis in dairy cattle was 4000 µg/mL, and antimicrobials such as streptomycin, ciprofloxacin, and ampicillin had synergistic effects in combination with baicalin. The combinations could significantly increase the susceptibility to E. coli. Therefore, baicalin may be used as an antimicrobial agent alone or in combination with antibiotics to treat E. coli caused mastitis in dairy cows. It was found that panton-valentine leucocidin (PVL)-killing interleukin produced by S. aureus was one of the causes of dairy cow mastitis. Apoptosis and necrosis of bovine mammary epithelial cells were associated with PVL, while baicalin could inhibit interleukin-producing S. aureus and consequently reduce apoptosis. Thus, baicalin has great potential for use in the treatment of mastitis caused by S. aureus [125]. High-yielding dairy cows are prone to oxidative reactions which are closely associated with inflammation. Baicalin has considerable anti-inflammatory and antioxidant effects on LPS-induced inflammatory damage of mammary epithelial cells in dairy cows. It can be used to prevent oxidative metabolic disorders in dairy cows and has been effectively applied in clinical practice [69,126].Baicalin is widely used to treat viral diseases such as bovine viral diarrhea and enteritis [127]. Previous studies have shown that the cure rate of the decoction of S. baicalensis in conjunction with Fructus gardeniae on bovine diarrhea was 95%, and the cure rate in vivo was 75% [128]. In addition, the decoction of S. baicalensis in conjunction with F. gardeniae can directly inhibit the bovine viral diarrhea virus in vitro [129]. Lv et al. [106] showed that oral administration of baicalin at medium and low doses (850 and 425 mg/d for 5 days) could effectively treat piglet diarrhea and found that baicalin could inhibit inflammatory exudation and reduce the release of inflammatory cytokines, which might be one of the mechanisms of its therapeutic effect. Baicalin showed potent activity against Newcastle disease virus and direct killing effect against Newcastle disease, capable of inhibiting the infection of chicken embryo fibroblasts and blocking the intracellular Newcastle disease virus, and has the potential to be used as a pharmaceutical ingredient [130].Baicalin had inhibitory effects on duck hepatitis virus (duck hepatitis A virus type 1, DHAV-1). An in vitro mechanistic study revealed that baicalin inhibited the propagation of DHAV-1 by interfering with the viral replication and release of the virus. An in vivo mechanistic study showed that the antioxidant and immunological enhancement functions of baicalin played a crucial role in its therapeutic effects against duck viral hepatitis. Baicalin may serve as a potential agent for the treatment of duck viral hepatitis [131].6. ConclusionsS. baicalensis contains a variety of active compounds with multiple biological functions, such as antioxidation and bacteriostatic and anti-inflammatory effects. Thus, it has extensive potential in the production and health of food animals. At present, S. baicalensis is sometimes used in animal husbandry production in the form of composite preparations. The chemical composition of the composite preparations is complicated, which makes it challenging to elucidate mechanisms of actions and results in unstable effects. Therefore, it is particularly important to study monomer compounds present in medicinal herbs. Currently, most studies on the biological activity and pharmacological effects of baicalin, baicalein, and other extracts of S. baicalensis are conducted with laboratory animals and few studies have been performed in animal production and clinical application. There is an urgent need to strengthen the research and application of baicalin and other herbal active compounds in clinical practice. The substitution of antibiotics by natural herbs and their active compounds has become a trend in the future research and development of animal feed, with great potential in practice.

animals : an open access journal from mdpi

10.3390/ani13050824

PMC10000244

Interactions between free-roaming dogs and humans influence the quality of life and behavior of both species. Understanding the spatial distribution of free-roaming dogs is essential in designing policies to control zoonoses and improve canine well-being. In the present study, by means of photographic captures and recaptures and geospatial position recordings of 554 dogs, we demonstrated that the location of the animals in an urban environment was influenced by the direct supply of food offered by the human population. Free-roaming dogs stayed closer to community feeders than to commercial food outlets. We know that community feeders are essential to improve the quality of life of free-roaming dogs. However, they should be in areas with reduced movement of people/vehicles. Our results may be representative of different areas of Brazil and of other parts of the world. They expand the understanding of canine ecology and behavior in the urban environment and highlight the importance of human contributions to the maintenance and distribution of free-roaming dogs.

Understanding the distribution of dogs in the environment is relevant for establishing human and animal health actions. In the present study, we analyzed the influence of community feeders and commercial food outlets on the spatial distribution of free-roaming dogs in an urban area of a municipality in Southeast Brazil. The dogs were identified via photographic capture and recapture performed over five sampling efforts. The spatial densities of dogs were determined using the Kernel method. Spatial correlations between the distribution of free-roaming dogs and the locations of community feeders and commercial food outlets were analyzed using the K function. During the study, 1207 captures/recaptures were performed encompassing 554 dogs, the majority (62.6%) of which were males. Agglomerations of male and female dogs were observed in the areas where food was present. Positive spatial autocorrelations were detected between the distribution of dogs and food sources. The median distances between dogs and community feeders or commercial food outlets were 1.2 and 1.4 km, respectively, and the difference between these two was statistically significant. The presence of community feeders and food outlets demonstrates the influence of human activity, on the spatial distribution of free-roaming dogs. These results will be useful for developing strategies aimed at the improvement of animal welfare and the prevention of zoonoses.

1. IntroductionIn Brazil, dogs are part of the urban ecosystem in both small country towns and large city conurbations [1]. The canine population within the country is estimated to number around 54.2 million animals [2], with a large proportion being free-roaming dogs that roam the streets freely without the direct supervision of humans [3,4].Although dogs provide numerous benefits to humans, such as companionship, stress reduction, physical activity, improved mental health, assistance for people with disabilities, protection, and safety, free-roaming animals cause accidents and spread infectious diseases [5,6,7]. Moreover, such free-roaming dogs may have a shorter lifespan in comparison with domiciled animals [8,9], considering that the former receive minimal or no veterinary care [10] and have limited access to adequate nutrition. For these reasons, the management of free-roaming dog populations is gaining increased research attention and is considered a relevant issue from the viewpoint of public health and animal welfare [11].While interactions between restricted dogs and humans exhibit many different forms [3], the density of the free-roaming canine population tends to increase concurrently with that of the human population. The maintenance of these animals in urban environments depends on direct sources of food and support (community feeders), and on indirect sources such as commercial food outlets (stores and restaurants) and garbage collection points [12,13]. The availability of food influences the spatial distribution of dogs, while the search for food motivates them to travel longer distances within their environs [14,15,16]. The population dynamics of free-roaming dogs are influenced substantially by the supply of food by householders since a small number of community feeders can support a large canine population [10,13,17].Despite its likely relevance to the ecology and spatial distribution of free-roaming dogs within the environment, the direct supply of food by the human population has been poorly analyzed in the literature. A better understanding of this issue may be pertinent for population control, improvement of animal welfare and the prevention of zoonoses [18,19]. In a study carried out previously in the same area as the present work, agglomerations of unrestricted dogs were identified in places close to food stores [19]. This study extends the investigation of the influence of the human community on the ecology of the canine population. In the present investigation, we set out to analyze the influence of community feeders and commercial food outlets on the ecology and spatial distribution of free-roaming canine populations in an urban setting by carrying out photographic captures and recaptures.2. Materials and Methods2.1. Site of StudyThe investigation was performed in Divinópolis, the largest municipality in the mid-west of the state of Minas Gerais, Brazil, with a population of approximately 242,505 inhabitants [20]. The specific study site encompassed eight neighborhoods in the municipality comprising some 7600 residents [9,21].2.2. Data CollectionPhotographic captures and recaptures of free-roaming dogs (defined as animals roaming the streets not accompanied by an owner) were performed over five sampling efforts (campaigns) conducted between September 2018 and September 2019 inclusive. Samplings were carried out in the mornings of three consecutive days during the first two weeks of the month and were repeated every three months. The vehicle used during each sampling always followed the same route and covered all of the streets in the designated areas at a speed of 20 km h−1. The team comprised the driver together with two researchers who were responsible for photographing and recording the general characteristics of the photographed animals, including size, color, natural marks and sex, all of which were useful for later identification. A Canon PowerShot SX60HS camera was used to photograph the dogs at a maximum distance of 10 m in order to obtain good-quality pictures while taking care not to frighten the animals away. At least three photographs of each dog were taken from different positions. The geographical coordinates of each capture/recapture point (the place where the dog was sighted) were recorded together with those of community feeders (food such as dry dog food and food scraps provided by households either in containers or directly on the ground) and commercial food outlets (fast-foods, grocery stores, bars, restaurants, bakeries, butchers and food markets).Data concerning dogs captured during the three days of sampling were aggregated in order to assemble the database from the five campaigns. The aggregation of data in campaigns was defined to increase the robustness and power of comparisons and, as well as to compare our results with that of past literature.Dogs that had been photographed in one campaign but captured again in a new campaign were considered recaptured. Each animal received an identification code and its details (pictures, description and position) were recorded in a spreadsheet. Data were analyzed and classified independently by two researchers. Using the Kappa index, a high (0.86) and significant (p < 0.01) agreement was recorded between them in identifying the 1207 photocaptures. Any discrepancies were resolved by consensus. Finally, a database was created containing information about all of the dogs captured and/or recaptured during the five campaigns.2.3. Statistical AnalysisThe nearest-neighbor distance approach was used to identify the existence of regions with agglomerations of dogs in the study area [22]. Kernel maps were prepared for the entire study period, and separately for each capture/recapture event, and stratified by sex of the dogs. The search radius was set at 100 m and the quantile normalization method was adopted as the most appropriate statistical tool for visualization of the different concentrations of dogs in the geographic space [23]. Possible spatial correlations between the distribution of free-roaming dogs and the locations of community feeders and commercial food outlets were investigated using Ripley’s bivariate K function [19,24]. The linear distance matrix function was used to estimate the distances of the meeting points of the dogs in relation to the community feeders and commercial food outlets, with the values expressed as medians and interquartile ranges (IQR). Median distances were compared using the Mann–Whitney test with the significance level set at 5%. Spatial analyses were performed using QGIS software version 3.16.16 while statistical analyses were carried out with the aid of R software version 4.2.0.3. ResultsA total of 1207 photographic captures/recaptures involving 554 different free-roaming dogs were accomplished during the study period (September 2018 and September 2019 inclusive). The proportion of captured free-roaming male dogs was typically twice that of females, with a similar profile observed for recaptured dogs. (Table 1). Most dogs in all campaigns were small, as compared to medium and large animals. In addition, black dogs were predominant (Table 2). The number of dogs captured for the first time decreased as capture efforts progressed. The number of recaptured individuals tended to increase over the study period, although there was a decrease in the last effort. (Figure 1).The numbers of community feeders and commercial food outlets increased slightly from the first to the third sampling effort but decreased thereafter. There were more community feeders than commercial food outlets (Table 3).According to nearest neighbor distance analysis, the animals were distributed in a clustered manner, that is, they were not randomly distributed in geographic space. The distribution patterns of agglomeration were statistically significant (p < 0.01). As shown by the Kernel map (Figure 2), animal clusters were located in the vicinity of community feeders and food outlets. Although the sites of food sources varied over the study period, canine agglomerations always remained close to these locations (Figure S1) regardless of the sex of the animals (Figures S1 and S2).The K function revealed the existence of positive spatial autocorrelations up to 500 m between the distribution of free-roaming dogs and food sources (Figure 3). The median distance between dogs and community feeders was 1.2 km (IQR = 0.5–2.6 km), whereas the median distance between dogs and commercial food outlets was significantly higher at 1.4 km (IQR = 0.5–2.7 km; p < 0.01).4. DiscussionThe results demonstrate that the interaction between humans and dogs plays a relevant role in the distribution of the animals in the study area with the occurrence of clusters being positively correlated with the presence of food sources, particularly community feeders.The greater number of male dogs and the agglomeration of males and females in the vicinity of food outlets in the study area has been established previously by Melo et al. (2020) [19] with distance values similar to those reported herein. The predominance of males occurs due to behavioral factors of the dogs and cultural aspects related to the care of animals in the region [19]. The consistency of the results related to the agglomeration may be explained by the absence of interventions aimed at the dispersion of the animals and the lack of control of the free-roaming canine population [25]. In addition, the number of new free-roaming dogs captured decreased as the number of dogs recaptured increased. The existence of new free-roaming dogs (captured) in the region after a year of continuous captures, may demonstrate the lack of responsible custody by owners (canine abandonment or escape) and, to a lesser extent, the existence of canine reproduction [9].The novel finding of dog clusters in areas close to community feeders reinforces the significance of human-dog interactions in the maintenance of free-roaming dogs in urban settings. It is of note that during the study, a small number of dog clusters were observed in locations that had no feeders or food outlets. This finding can be explained by the behavior of some guardians who allow free access of their domiciled dogs to the streets and even feed them outside their properties [26]. Such results show that in the absence of interventions that consider the responsible ownership of animals and the home ranges of dogs, there will be no reduction in the stray dog population.Dogs, unlike wolves from which they evolved, depend primarily on humans for food [27,28]. Thus, direct support by community feeders through the provision of both food and care for the free-roaming animals favors the continuation and augmentation of the free-roaming canine population in the urban environment [9]. Thus, human support is essential for understanding the ecology of unrestrained dogs in urban environments. Families who regularly feed free-roaming dogs are responsible for sustaining these animals, so only a small fraction of households can accommodate large, tolerant, unrestrained dogs [10,13].In the present study, free-roaming dogs gathered together closer to community feeders than to food outlets. Most of the time there were no humans near the feeders, a situation that contrasts with that established for food outlets. Kittisiam et al. (2021) analyzed the contact network of free-roaming dogs in a university campus in Thailand and demonstrated that the average number of contacts for the weekend network was significantly higher than that for the weekday network, indicating that dogs tended to cluster more intensely in the absence of humans.Our study showed that community feeders might be more advantageous to the animals than food outlets because they afford a continuous and possibly more stable nutritional source. As part of an educational project (‘Projeto AlimentaCão’) aimed at alerting schoolchildren to the problem of abandonment of domestic animals, [29] and installed feeders at points on the campus of the Universidade Tecnolólogica Federal do Paraná that were some distance away from the university restaurant. These researchers reported that the number of free-roaming dogs that congregated in the surroundings of the restaurant diminished whilst the population using the feeder points stabilized and became healthier and more docile. An alternative approach of restricting the amount of food would also be effective in controlling a population of free-roaming dogs at a particular location [30]. However, such a measure would be ethically questionable and would simply lead to the dispersion of animals as they move to new areas in search of food [9,25]. The more advantageous solution would be to relocate community feeders to defined areas with less movement of people and vehicles. Although the control of dog habitats and their ranges may not result in a decrease in the canine population, it may be effective in reducing accidents, the transmission of diseases, bites and pollution in busier locations.The availability of food resources was constant throughout the period of the present study, a situation that may be representative of different areas of Brazil and of other parts of the world. In India, for example, 37% of individuals reportedly feed free-roaming dogs [17]. Feeding free-roaming dogs can create and strengthen the bonds between humans and dogs [31], and allow the animals to move, socialize and express their natural behavior [32]. On the other hand, the support offered by community feeders can also be considered a public health problem because it sustains canine populations that act as reservoirs of zoonotic diseases [33,34], cause accidents and pollute the environment. Moreover, from the viewpoint of animal welfare, the quality of life and the longevity of free-roaming dogs tend to be inferior to those of domiciled animals [35,36].While various studies support the thesis that the search for food is a key factor in determining canine mobility and agglomeration [19,37,38], it should be noted that other factors influence the distribution of free-roaming dogs in urban areas including the growth of cities, climate, reproductive status and the search for partners [39,40,41]. Therefore, controlling the habitat and movement of free-roaming dogs is a challenging task that must be adopted together with other strategies. Sterilization can contribute to the reduction in birth rates, but in the region of the present study, the abandonment of dogs is the main reason for the increase in the unrestricted population [9].In order to reduce or eliminate the population of free-roaming dogs, it will be necessary to adopt broad and effective measures of responsible animal guardianship, which may be difficult to implement in the short term depending on the socioeconomic and cultural contexts [26,42]. In this sense, the provision of food and the management of appropriate care may be relevant actions. Installing feeders in favorable environments can reduce the risk of accidents and bites, minimize the risk of disease transmission and improve animal welfare [43]. In addition, if combined with other actions such as vaccination, disease prevention and installation of shelters, the provision of feeders would certainly improve the quality of life of the animals [44,45]. However, it is important to emphasize that such measures are palliative and that the problems associated with free-roaming dogs will only be solved through the awareness and accountability of guardians concerning the social and economic consequences caused by the abandonment of their animal companions [46], along with the implementation of responsible adoption policies [26,42].Strategies to estimate animal populations share as basic principles, uncertainties regarding the detection of all individuals that pertain to the target population in a given area and heterogeneities in the individual encounter probabilities [47]. The validity of the methods based on counting or the capture and recapture of dogs is only achieved when the animals are correctly identified in all capture activities [48]. Obtaining accurate information about free-roaming dogs can be a challenging process, and inadequate methods can result in biased information [26,47]. Among the various methods for estimating population, photographic capture and recapture is advantageous because it is a safe, fast, and cost-effective option. Animals do not need to be physically captured, reducing their exposure to risks and adverse effects. Moreover, the photographic capture technique enables the individual identification of animals based on natural markings, and analysis of the previous capture history can determine whether an animal has been recaptured [38]. Combining this method with the Geographical Coordinate System (GPS) to record the location of free-roaming dogs allows for expanded analysis and a better understanding of the ecology of unrestricted dogs in urban areas [21]. On the other hand, the low quality and lack of detail in the photograph can make the process of identifying the dog more difficult. However, in the present study, the comparison of photos by independent researchers, combined with quality training, can optimize the identification process.The present study was subject to some limitations that are intrinsic to the photographic capture method. The majority of free-roaming dogs move around, which makes it difficult to photograph them, and this may give rise to the loss of some information. However, during the search process, all streets in the study area were covered and this minimized sample deficiencies because when a dog could not be found in one street, it was almost certain that it could be found in another. An additional limitation was that the sampling efforts took place only in the morning and not throughout the day. However, dogs are more active in the morning and tend to rest as the temperature rises [49]; therefore, sampling efforts were focused in the period of highest activity.5. ConclusionsAgglomerations of free-roaming dogs were positively associated with the locations of commercial food outlets and, more especially, with those of community feeders. The provision of food by community feeders was essential in maintaining the concentration of dogs in the vicinity of food sources. The results obtained not only add to our understanding of the ecology and behavior of free-roaming dogs in urban areas but also highlight the need for policies to reduce the risks associated with the presence of free-roaming animals. Moreover, appropriate management of food and water resources should be implemented in order to improve the quality of life of these animals such as the allocation of specific and supervised points far from busy roads and residential/commercial areas. We hope that the outlined recommendations can help improve future interventions by providing more appropriate strategies. Additionally, we suggest that the spatial distribution of free-roaming dogs be more widely studied in the literature, especially through intervention studies, assessments in urban and rural areas, and evaluations of residents’ perceptions of supporting these animals.

animals : an open access journal from mdpi

10.3390/ani11061580

PMC8230262

Changes in skin mucus production and composition offer a new means to study how fish cope with changes in the environment. We explored the utility of skin mucus as an indicator of physiological responses and energy use in a reference fish species, the European sea bass. We evaluated the exudation volume of skin mucus and the main stress- and osmoregulation-related biomarkers in both mucus and plasma. We demonstrate the viability to study the exuded volume of skin mucus composition and its parameters as an informative tool of the fish energy waste at different environmental salinities. This study is of great interest for both aquaculture and ecological studies.

The European sea bass (Dicentrarchus labrax) is a euryhaline marine teleost that can often be found in brackish and freshwater or even in hypersaline environments. Here, we exposed sea bass juveniles to sustained salinity challenges for 15 days, simulating one hypoosmotic (3‰), one isosmotic (12‰) and one hyperosmotic (50‰) environment, in addition to control (35‰). We analyzed parameters of skin mucus exudation and mucus biomarkers, as a minimally invasive tool, and plasma biomarkers. Additionally, Na+/K+-ATPase activity was measured, as well as the gill mucous cell distribution, type and shape. The volume of exuded mucus increased significantly under all the salinity challenges, increasing by 130% at 50‰ condition. Significantly greater amounts of soluble protein (3.9 ± 0.6 mg at 50‰ vs. 1.1 ± 0.2 mg at 35‰, p < 0.05) and lactate (4.0 ± 1.0 µg at 50‰ vs. 1.2 ± 0.3 µg at 35‰, p < 0.05) were released, with clear energy expenditure. Gill ATPase activity was significantly higher at the extreme salinities, and the gill mucous cell distribution was rearranged, with more acid and neutral mucin mucous cells at 50‰. Skin mucus osmolality suggested an osmoregulatory function as an ion-trap layer in hypoosmotic conditions, retaining osmosis-related ions. Overall, when sea bass cope with different salinities, the hyperosmotic condition (50‰) demanded more energy than the extreme hypoosmotic condition.

1. IntroductionWild European sea bass (Dicentrarchus labrax) moves seasonally from seawater to freshwater environments and vice versa, including coastal areas, lagoons, estuaries and other parts of rivers [1,2,3,4]. Despite this haline plasticity, water salinity can affect sea bass growth in extreme conditions below 10‰ and over 50‰, as already reported by Dendrinos and Thorpe [5] and Eroldogan et al. [6] who found better growth performance at lower salinities (10‰, 20‰, 25‰ and 30‰) than control (33‰). Varsamos et al. [1,2] measured blood osmolality at larval and juvenile stages, while Jensen et al. [7] studied the effect of salinity on osmoregulation and branchial Na+/K+-ATPase. In those studies, the authors suggested that the acclimation process was completed in four to eight days. Similar responses were found when analyzing growth performance, osmoregulatory and metabolism as part of the acclimation process in other marine species, such as gilthead sea bream (Sparus aurata) [8,9,10,11,12], shi drum (Umbrina cirrosa) [13] and red porgy (Pagrus pagrus) [14]. Within the first few days, a classic pattern develops known as “crisis and regulation”, which consists of an initial phase of blood metabolic and osmotic changes, followed by a phase of regulation, when osmoregulatory and metabolic parameters achieve a steady “normalized” state [3,7,9,13]. The first evidence of this is a variation of blood osmolality and the main osmosis-related ions (Na+, Cl− and K+) [7,11,12,14,15,16]. Gills are markedly affected by these osmotic changes, which modify Na+/K+-ATPase activity and chloride cell dynamics processes which are mediated by cortisol [17,18,19,20]. This eventually restructures gill energy metabolism and requirements [8,10,11,21,22]. Meanwhile, blood metabolic changes are mainly related to a decrease in plasma glucose, triglycerides and cholesterol [3,7,9,13]. After a condition has been sustained for weeks, energy metabolism is reorganized towards an increased energy expenditure, reallocation of resources and depletion of carbohydrate reserves in several tissues, such as liver, gills, kidney and brain [8,9,10,12,14].Although blood analysis is a non-lethal method to measure stress, the procedure can result in injuries on blood vessels, leading to hemorrhage, in fish skin, which may increase the risk of infection. Therefore, in recent years, growing interest has been shown in the use of minimally-invasive methods to assess fish physiological status and welfare, such as fish skin mucus analysis [23,24,25,26,27]. It has also been reported that endogenous and exogenous factors, such as fish developmental stage, sex, infections or environmental changes, can modify fish skin mucus composition [23,24,25,26,28,29,30,31,32]. Moreover, it has been observed that the components of exuded mucus are modified in response to stressors [33,34,35,36,37], including acute salinity challenges [27]. Indeed, measurement of some stress indicators found in mucus, such as cortisol, glucose and lactate, has been proposed as a feasible, non-invasive, analysis of stress biomarkers [23,24,25,38,39,40]. Nonetheless, to date, most experiments have analyzed short-term stress, with few long-term studies. Recently, Fernández-Montero et al. [41] studied the effect of different stressors, such as temperature, stock density and handling, on cortisol release and mucins expression in the skin of the greater amberjack Seriola dumerili. They reported an increase in muc-2 expression, which encodes a secreted glycoprotein forming part of the insoluble mucous barrier, in high stock densities and in response to handling protocols, indicating a possible increase in mucus exudation in relation to stress. In addition, skin mucus has many relevant biological and ecological roles, and among them osmoregulation [42,43]. Previously Roberts and Powell [44] evaluated skin mucus modifications under different salinities over three months in salmonids, finding that skin mucus was hyperosmotic with regard to hypoosmotic surrounding water, and that gill mucous cells shifted from neutral to acid when fish were moved from freshwater to seawater. Recently, we used skin mucus biomarkers to evaluate the response of sea bass to several acute osmotic challenges [27] and observed, under hypersalinity conditions, the production of a very large volume of skin mucus, with the highest total contents of cortisol, glucose, and protein. Thus, although that study only focused on the acute response, this could be an undesirable condition if the salinity condition becomes chronic.Considering the above-mentioned studies, we believe that changes in skin mucus production and composition offer a new means to study how fish cope with changes in the surrounding water. Despite a number of papers focused on fish osmoregulatory responses there is less information on the metabolic impacts of salinity adaptation, and little research has considered skin mucus as a target or indicator for the study of osmotic response. Thus, our main aim here was to study skin mucus biomarkers, together with plasma and gill parameters, in the response of juvenile sea bass to sustained osmotic challenges. To this end, we exposed seawater (35‰) acclimated fish for 15 days to a hypersaline condition (50‰) and to two hyposaline environments: An almost freshwater condition (3‰) and a mid-estuary condition (12‰), which is practically isoosmotic to the fish internal milieu. We explored the usefulness of mucus as an indicator of physiological responses and energy usage by evaluating the volume of mucus exuded and the main stress-related and osmoregulation-related biomarkers in mucus and plasma. Moreover, we also analyzed gill energy usage (Na+/K+-ATPase activity) and gill mucous cell classes and shapes. All these findings contribute to increased knowledge about the acclimation responses of European sea bass to environmental salinity and the repercussions in the energy expenditure required to maintain homeostasis under a chronic condition, which could be useful for conservation biology and aquaculture.2. Materials and Methods2.1. The Animals and Experimental ProceduresEuropean sea bass juveniles were obtained from a commercial source (Mariscos de Esteros, SA, Huelva, Spain) and acclimated indoors at the CCMAR Ramalhete marine station (Faro, Portugal). Fish were reared for two months in open-system fiberglass tanks (1000 L), at seawater salinity (34.9‰ ± 0.1‰) pumped from the marine environment at the naturally occurring temperature (15.7 °C ± 0.2 °C ) under the natural photoperiod (April to May, 2019), and fed a commercial diet twice a day (2.5% w/w). For the assay, fish (129.2 ± 3.6 g) were randomly allocated into 500 L tanks (n = 10/tank at 2–3 Kg/m3) in 4 semi-closed systems. After one week, fish were acclimated to experimental conditions aiming at water salinities of 3‰, 12‰, 35‰ and 50‰. The transition between 35‰ and each experimental condition was carried out during three days by increasing freshwater flow in the systems for the 3‰ and 12‰ conditions, and by adding highly concentrated seawater (prepared with commercial sea salt) in the system for the 50‰ condition. Once the experimental salinities were achieved, the fish were kept in the experimental tanks for a period of 15 days and fed normally until 24 h prior to sampling. To maintain low ammonia levels and assure good water quality, the systems were not completely closed. In control and low salinities, the salinities were achieved by balancing the flow of seawater, SW, with that of freshwater, FW (so 100% SW for 35‰, 35% SW + 65% FW for 12‰ and 8.5% SW + 91.5% FW for 3‰). These mixtures were prepared in a head tank and flowed into the fish tanks. For the high salinity, 50‰, sea salt was added and mixed in the head tank and flowed into the fish tank and recycled into the head tank. Renewals of the water in the head tank were performed every 48 h. Oxygenation and adequate mixing were provided by strong aeration in the head tanks and oxygenation was also guaranteed by aeration in the fish tanks. Water salinity, oxygen saturation and ammonia levels were controlled throughout this period. Measured salinities varied by 0.5 to 1 ppt above and below the target values; oxygen saturation values were always above 85% and ammonia levels were always below 2.5 mg/L. The 15-day exposure period was selected in accordance with the reported effects of osmotic challenges on sea bass osmoregulation [7,8,9,11,12,45,46].After the 15-day period, the animals from each condition were sampled for skin mucus, blood and gills. All fish from each tank were sampled before moving to the next condition in a rapid procedure that did not expand more than 30 min per tank. Individual mucus samples were collected once the fish was anaesthetized with 2-phenoxyethanol (1:250, Sigma-Aldrich, Madrid, Spain) as described in Fernández-Alacid et al. [23]. The fish removal from the tank and its onset of sedation took less than 1 min, and no fish remained more than 5 min under sedation before the mucus sampling. A sterile glass slide was used to carefully remove mucus from the over-lateral line, starting from the front and sliding in the caudal direction. The glass was gently slid along both sides of the animal, avoiding the non-desirable operculum, ventral-anal and caudal fin areas, and the skin mucus was carefully collected into a sterile tube (1.5 mL), snap-frozen in dry ice and stored at −80 °C until analysis. Thereafter, each fish was laterally (all on the left side) photographed with a Nikon D3000 camera (Nikon, Tokyo, Japan), weighed and measured. Blood was subsequently obtained from the caudal vein with a 1 mL heparinized syringe with a 23G needle. Plasma was separated from whole blood by centrifugation at 10,000× g for 5 min, aliquoted, immediately frozen and stored at −80 °C. The animals were then killed by severing the spinal cord and gill filaments from the second gill arch was collected and placed in a tube with 100 μL of ice-cold SEI (sucrose-EDTA-imidazole) buffer (150 mM sucrose, 10 mM EDTA, 50 mM imidazole, pH 7.3) [47], snap-frozen and stored at −80 °C. An additional section from the same gill arch was severed and placed in a 2 mL tube with Bouin-Holland solution.The research was approved by the Centre for Marine Sciences (CCMAR)-Universidade do Algarve animal welfare body (ORBEA) and the Direção-Geral de Alimentação e Veterinária (DGAV), Permit 2019-06-04-009758, in accordance with the requirements imposed by Directive 2010/63/EU of the European Parliament and of the Council of 22 September 2010 on the protection of animals used for scientific purposes.2.2. Metabolite Biomarkers and Cortisol Levels in Mucus and PlasmaThe soluble components of skin mucus samples were previously obtained from the mucus homogenization, using a sterile Teflon pestle and centrifugation at 14,000× g for 15 min [23]. Enzymatic colorimetric tests for glucose and lactate (LO-POD glucose and LO-POD lactate, SPINREACT, Sant Esteve de Bas, Spain) adapted to 96-well microplates and fish mucus and plasma samples were used. Following the manufacturer’s instructions, mucus and plasma samples and standard dilutions were mixed with working reagents in triplicate. The OD was determined at 505 nm with a microplate reader (Infinity Pro200 spectrophotometer, Tecan, Barcelona, Spain). The glucose and lactate values were expressed as mg·dL−1 for plasma and μg·mL−1 for skin mucus.Cortisol levels were measured using an ELISA kit (IBL International, Hamburg, Germany). Briefly, an unknown amount of antigen present in the sample competed with a fixed amount of enzyme-labelled antigen for the binding sites of the antibodies coated onto wells. After incubation, the wells were washed to halt the competition reaction. Therefore, after the substrate reaction, the intensity of the color was inversely proportional to the amount of antigen in the sample. Following the manufacturer’s instructions and adaptations for fish mucus and plasma [24,25], the samples and standard dilutions (from 0 to 3 μg·dL−1) were mixed with enzyme conjugate and incubated for 2 h at room temperature. The substrate solution was added after rinsing the wells with a wash solution and incubated for 30 min. The reaction was halted by adding stop solution and the OD was determined at 450 nm with a microplate reader (Infinity Pro200 spectrophotometer, Tecan). The cortisol values were expressed as ng cortisol·mL−1 of plasma or skin mucus.Mucus and plasma protein concentrations, as well as protein in gill homogenates (for Na+/K+-ATPase activity, see ahead), were determined using the Bradford assay [48] with bovine serum albumin (BSA) as the standard. Bradford reagent was mixed with the samples in triplicate and incubated for 5 min at room temperature. The OD was determined at 595 nm with a microplate reader (Infinity Pro200 spectrophotometer, Tecan or MultiScanGo, ThremoFisher Scientific, Rochester, NY, USA). The protein values in plasma or skin mucus were expressed as mg protein·mL−1.2.3. Osmolality and Ion Quantification of Plasma and Skin MucusPlasma osmolality was measured with a vapor pressure osmometer (Wescor Vapro 5520, ELITechGroup, Norwood, MA, USA) and was expressed as mOsmol·kg−1. Plasma Na+ and K+ levels were measured using a Flame Photometer (BWB XP, BWB Technologies, Berkshire, UK) and expressed as mmol·L−1. Plasma chloride concentration was measured using a colorimetric test (SPINREACT, Spain) adapted to microplates, and the OD was determined in a microplate reader (MultiScanGo, ThremoFisher Scientific); measured values were expressed as mmol·L−1. Mucus osmolality and ion concentrations (Na+, K+ and Cl−) were measured using an ion analyzer (ISElyte X9, Tecil, Barcelona, Spain). Osmolality values were expressed as mOsmol·kg−1 and ion concentrations as mmol·L−1.2.4. Mucus Exudation ValuesTo determine the effects of the osmotic challenge, total mucus exudation was obtained by measuring the volume of mucus collected (in μL), related to the skin area (in cm2) and to fish weight (in g). Skin area was obtained using the ImageJ program (US National Institutes of Health, Bethesda, MD, USA). The area was manually marked as an approximation of the area actually scrapped, avoiding the dorsal and the lateral fins, and over the lateral line. This was then measured using the software included in the program. Furthermore, soluble collected mucus (μL) was referred to the sampling area and to fish weight, to calculate collected mucus per area (μL·cm−2) and collected mucus per unit weight (μL·g−1).2.5. Gill Na+/K+-ATPase ActivityGill Na+/K+-ATPase activity was determined using the method developed by McCormick [47] adapted for microplate assay [49]. Gill tissue was homogenized in 125 µL of SEI buffer with 0.1% of deoxycholic acid and centrifuged at 2000× g for 30 s. Samples were mixed with the assay buffer with or without 0.5 mM ouabain and a decrease in absorbance was measured at 340 nm for 15 min at 25 °C (MultiScanGo, Thermo Scientific, Rochester, NY, USA). An enzymatic coupling of ATP dephosphorylation to NADH oxidation was used to detect ouabain-sensitive ATPase activity, and Na+/K+-ATPase activity was expressed as μmol ADP·mg protein−1·h−1. Protein in gill homogenates was determined as indicated previously.2.6. Gill Histology and Histological AnalysisAfter 24 h in Bouin solution at RT, the gill arches were rinsed several times in 70% ethanol and stored at 4 °C. The tissues were cleared in graded xylene and were later embedded in paraffin (Paraplast Plus; Sherwood Medical, St Louis, MO, USA) and sectioned at 6 μm. After dewaxing and rehydration, the sections were placed on slides with APES treatment (Aminopropyltriethoxysilane, Sigma, Madrid, Spain). The slides were stained using a periodic acid-Schiff (PAS) and Alcian Blue (AB) staining protocol. For histological analysis, the slides were photographed using a light microscope (BX61; Olympus, Tokyo, Japan) connected to a digital camera (DP70; Olympus, Tokyo, Japan) at a magnification of ×20. Goblet cells were counted on 6 no-consecutive lamellae sections of 250 µm, from the second branchial arch, of 5 fish per condition. The counts were performed in blind condition with sections from all samplings mixed and counted by one person. Goblet cells were counted using ImageJ (US National Institutes of Health, Bethesda, MD, USA), while cell counting, frequency (cell·mm−2), size (μm2), perimeter (μm) and shape were calculated, with acid mucins (purple) and neutral mucins (magenta) differentiation.2.7. Statistical AnalysesThe experiment was performed without tank replicates and a priori statistical evaluation was performed to obtain the minimal number of fish required for a one-way ANOVA, based on fixed effects. To compare the data obtained for stress-related biomarkers, osmotic parameters and Na+/K+-ATPase activities for the different salinity challenges, we used one-way ANOVA. Additionally, Student’s t-test was used to compare osmotic parameters between plasma and mucus. For all our statistical analysis, a priory study for homogeneity of variance was performed using Levene’s test. When homogeneity existed, Tuckey’s test was applied; if homogeneity did not exist, then the T3-Dunnet test was applied. All statistical analysis was undertaken using SPSS Statistics for Windows, Version 22.0 (IBM Corp, Armonk, NY, USA) and all differences were considered statistically significant at p < 0.05.3. ResultsBody weight, body length and condition factor were obtained, and no significant differences were observed in response to the osmotic challenges (Table 1). To determine the effects of sustained salinity challenge on osmoregulatory homeostasis, plasma osmolality together with osmotic-related ions (Na+, Cl− and K+) were measured (Table 2). Although plasma osmolality strongly buffered the changes in water salinity, maintenance for 15 days at 3‰ provoked a slight but significant reduction in plasma osmolality (322 ± 3 mOsmol·kg−1, p < 0.05) while the maintenance at 50‰ significantly increased plasma osmolality by 5% (358 ± 5 mOsmol·kg−1, p < 0.05). The sum of the two major main osmosis-related ions, Na+-Cl−, represented around 90% of plasma osmolality, irrespective of the challenge condition, even if levels at 50‰ (but not at low salinities) were significantly elevated in relation to control. Interestingly, plasma potassium was lower (p < 0.05) than control at both hypo- and hyperosmotic challenges.3.1. Gill Na+/K+-ATPase Activity and Gill Mucous CellsAt 50‰, Na+/K+-ATPase activity was significantly increased (by around 60%, p < 0.05) in relation to control values at 35‰ (Figure 1). Although Na+/K+-ATPase activity at the lower salinities did not differ significantly from control, values in fish at 12‰ (a near isosmotic condition to plasma osmolality) were the lowest and significantly below those recorded for fish at 3‰ (p < 0.05). The distribution of gill mucous cells (Figure 2) was markedly affected by the salinity challenges, which impacted cell frequency, size and shape (where the shape value = 1 corresponded to perfect circular cell). Cell counts showed that acclimation to hypersalinity induced a significant proliferation of mucous cells. Higher number of cells per mm2 of gill filament were detected (p < 0.05) at 50‰ than at 35‰ (Figure 2A), showing greater abundance of both types of mucous cells, acid (white arrows) and neutrals (black arrows) (Figure 2D,E), but with no differences in cell size (Figure 2B) or shape (Figure 2C). At lower salinities; however, acclimation to 12‰ condition did not modify any mucous cell parameters (with regard to control values), exposure to 3‰ resulted in the smaller number of cells per filament and significant changes in cell size (increased area) and shape (more elongated). Overall, gill acid mucin cells were far more numerous than neutral mucin cells, but the ratio between acid mucin cells and neutral mucins cells, which is indicative of gill mucus composition, was modified by the extreme conditions: for fish acclimated to the intermediate salinities this ratio was approximately 47:1 and 45:1 at 35‰ and 12‰ respectively, for fish in either hypo-osmotic or hyperosmotic salinities the ratio was significantly reduced to 12:1 at 3‰ (p < 0.05) and 17:1 at 50‰ (p < 0.05).3.2. Mucus Exudation ParametersThe skin mucus volume was recorded (Figure 3) and the exuded volume per area of collection and per unit of body weight were calculated (included in Figure 3). The amount of skin mucus increased in all conditions in relation to control (p < 0.05). In both hyposaline environments, 3‰ and 12‰, the volume of mucus exuded was increased by 50% and 80% in relation to those at 35‰, at Day 15 of exposition, and this trend remained when the exuded mucus was expressed per area of collection or per unit of body weight. The hyperosmotic challenge at 50‰ provoked the greatest skin mucus over-exudation, 130% (p < 0.05) of the control values at 35‰, which meant that the mucus exudation per cm2 of the calculated fish surface increased from 3.03 ± 0.37 µL at 35‰ to 5.73 ± 0.62 µL at 50‰ (p < 0.05).3.3. Plasma and Mucus OsmoregulationCompared with plasma, mucus osmolality did not buffer water salinity, and was strongly correlated with environmental osmolality (Figure 4A). Mucus osmolality at 35‰ was 944 ± 24 mOsmol·kg−1, almost identical to the values measured in water and quite different from the plasma control values (339 ± 5 mOsmol·kg−1). A similar situation was observed at 50‰ but in lower salinities mucus osmolality was maintained at values 40% and 100% above water osmolality (for fish at 12‰ and 3‰ respectively), which may be related to the presence of organic molecules as osmolytes or to increased ion retention capacity of mucus at low salinities. Ions (Na+, Cl− and K+) were measured in skin mucus (Figure 4B–D). Interestingly, the sum of main the osmosis-related ions, Na+ and Cl−, in mucus proved to be a better approach to water osmolality than the whole mucus osmolality (101.5 ± 5.6 mmol·L−1 for 3‰ and 225.2 ± 6.9 mmol·L−1 for 12‰). In fact, the Na+-Cl− sum explained 36.6% ± 2.5% and 50.2% ± 1.4%, respectively, of the total measured mucus osmolality (Figure 4). Consequently, other mucus components must contribute greatly to mucus osmolality. On the contrary, the 35‰ and 50‰ conditions showed closer osmolarities between mucus and water, and the Na+-Cl− sum covered around 64.02% ± 1.94% and 70.14% ± 1.19%, respectively, but far from the 90% covered in plasma. Interestingly, mucus K+ concentration, although not participating greatly in total osmolality values, did not show the same proportions for the conditions and mucus potassium showed the higher values at 50‰ (p < 0.05).3.4. Physiological Biomarkers in Mucus and PlasmaMain skin mucus biomarkers, such as soluble protein, glucose and lactate, as well as cortisol exuded levels exhibited different responses to the osmotic challenges (Table 3). In response to the extreme salinities of 3‰ and 50‰, significant increases in protein exudation were recorded: from 6.12 ± 0.61 mg·mL−1 for control mucus to 9.44 ± 0.85 mg·mL−1 (> 50% higher, p < 0.05) and to 10.81 ± 1.05 mg·mL−1 (> 75% higher, p < 0.05) for the 3‰ and 50‰ challenges, respectively (Figure 5A). Cortisol, one of the main indicators of acute stress response, appeared to be exuded in greater amounts (p < 0.05) in the lowest salinity, 3‰, while for both 12‰ and 50‰, mucus cortisol did not differ from control values. However, mucus glucose did not change significantly among the different salinities while mucus lactate was two-fold over-exuded at 50‰ compared to control. The same biomarkers were also measured in plasma. Protein levels in plasma showed the same trend as in skin mucus, with the lowest values observed in the 35‰ condition. However, significant differences in plasma protein occurred only in the 3‰ fish (which was 66.7% higher, p < 0.05). Glucose levels did not show any significant differences between the osmotic conditions, as reported in mucus. A significant increase in plasma lactate concentrations was observed in all challenging conditions in relation to control: Lactate was 80% higher in the 3‰ condition (p < 0.05) and 30% higher in both 12‰ and 50‰ conditions (p < 0.05). In contrast, plasma cortisol concentration was highest at 35‰, being twofold (3‰ and 12‰ conditions) and threefold (50‰ condition) higher than at other salinities; however, it was only significantly different from the 50‰ condition (p < 0.05). No significant correlation, by the Pearson’s coefficients analyses, was found between skin mucus and plasma biomarkers.3.5. Energy Expenditure by Mucus ExudationAs the individual volumes of mucus exuded were recorded, the total amount of each biomarker in mucus was calculated and showed in Figure 5, to evaluate their loss as future waste and energy expenditure. Thus, both hypoosmotic conditions caused significant over-exudation of protein (Figure 5A), glucose (Figure 5B) and lactate (Figure 5C) with respect to the control condition. The hyperosmotic condition generated the greatest and significantly highest release of the same metabolites, two-fold greater (p < 0.05) than the release at hypoosmotic conditions and five-fold with respect to control values (p < 0.05). This represents a sustained energy expenditure when animals are maintained at this salinity. With regard to mucus cortisol levels (Figure 5D), the 3‰ challenge resulted in a significant three-fold increase of total exuded cortisol, compared to control values. Lastly, the ratio of glucose:lactate exuded is also represented in Figure 5E, as an indicator of changes in aerobic metabolism in response to the salinity challenges. After 15 days at the experimental salinities, the mucus glucose:lactate ratio was significantly higher (p < 0.05) at 3‰ than at 50‰, evidencing different aerobic metabolism responses.4. DiscussionEuropean sea bass is an euryhaline marine teleost species that withstands different salinity conditions, hypoosmotic and hyperosmotic. Here we focused on evaluating the volume of mucus exuded as well as the amount of several different biomarkers contained in that mucus in response to sustained exposure to low (3‰ and 12‰) and high (50‰) salinity. We show that, in addition to osmoregulatory and metabolic changes, complete acclimation to such conditions has important impacts on mucus production and composition. Moreover, gill activity and modifications in gill mucous cell class, distribution and shape were recorded. All these data, together with parameters related to osmosis, allowed us to relate chronic salinity challenges with fish energy expenditure and waste via skin mucus over-exudation.The total energy consumed by osmoregulatory adaptation was estimated to be between 20%–68% in different species [50], but these values must include not only the actual cost of ion transport but consider the energy used by other metabolic processes which respond to changes in salinity [51]. Salinity-triggered hormones affect different pathways of energetic metabolism and other non-osmoregulatory organs such as the liver and brain also show changes in energetic metabolism [9,12,14,52,53,54,55]. Nonetheless, it is generally expected that a reduced gradient between the environment and the internal milieu lowers the need for osmoregulatory activity allowing more energy to be used for other functions [50] but this appears to be a more complex, and multifactorial process. In fact, previous reports have indicated the optimal conditions for sea bass growth to be between the isosmotic salinities 12‰ and 15‰ [7,8,56,57,58], but additional studies also show that 28‰ to 30‰ were better growth conditions for sea bass [5,59]. In our study, no significant differences were found in morphometric parameters (fish growth, length or condition factor) between the experimental conditions after two weeks. In a previous study, we already noted the relevance of measuring exuded mucus volume under an acute salinity challenge for sea bass [27], which could be further related to energy costs to maintain increased mucus exudation over time. In the current experiment, we recorded the volume of mucus exuded in response to the proposed challenges. For all the conditions, the amount of skin mucus increased with respect to the control condition and was 130% higher for the hyperosmotic condition. It is well established that fish exude more mucus in stressful situations, as described in acute stress experiments [23,24,25,27]. Several studies have reported an apparent increase in mucus production when fish transition both from freshwater to seawater [60,61,62,63,64] and from seawater to freshwater [43,65,66,67]. Nonetheless, those empirical findings were not tested, as mucus exudation was not measured. Thus, for the first time, here we have provided data on the exacerbated and continued skin mucus exudation in response to salinity challenges in sea bass. In consequence, fish energy status could be affected due to the need to exude the many mucus components, such as the gel-forming mucins, which are heavy and large glycoproteins [41], or the large numbers of soluble proteins [37] and the energy metabolites, like glucose and lactate [23].4.1. Plasma Changes, Gill Activity and Stress Impacts of Osmoregulatory ResponsesPlasma osmolality has been used as a physiological indicator when measuring the effects of salinity on fish physiology [11,68,69,70,71]. Adult euryhaline teleosts maintain plasma osmolality between 300 and 350 mOsm·kg−1 under tolerable salinities [11,72,73]. However, this adjustment is not immediate, and unbalances may occur in the first hours or days post transfer indicating incomplete osmoregulation in plasma [7,74,75,76]. For instance, after abruptly transferring European sea bass to freshwater, Jensen et al. [7] recorded plasma hemodilution, with plasma osmolality at 240 mOsm·kg−1 and being tolerated for at least 10 days. In the present case, 15 days after the initial salinity change, European sea bass showed a slight but significant decrease in the hypoosmotic conditions and a slight but significant increase in the hyperosmotic condition, compared to the control condition. However, those changes in the average plasma osmolality were minimal and variance within each group were quite low which suggests that a new equilibrium was achieved (i.e., a compromise for favorable osmoregulation in each salinity). It seems therefore, from the current data, that sea bass acclimated well to a 15-day exposure to altered salinity, maintaining plasma osmolality within narrow homeostatic limits (11% variation range limited by a 6% increase and 5% decrease in average osmolality levels at the highest and lowest salinity conditions, respectively).Changes in plasma osmolality were paralleled by variations in the concentrations of the major ions sodium and chloride, which, together with potassium, account for around 90% of the total osmolality. Plasma Na+ and Cl− followed the trends in osmolality while K+ actually dropped in hypo- and hyper-osmotic condition in relation to fish at 35‰. The concentration of these ions in plasma is mostly controlled by the action of the branchial Na+/K+-ATPase, which has a paramount role in salinity acclimation favoring the excretion of Na+ and Cl− in hyperosmotic conditions and being responsible for ion uptake in hypoosmotic conditions [18,77,78]. In the present study, Na+/K+-ATPase activity showed a typically “U-shape”, as reported before, in European sea bass [7,11,13], with the lowest activity being under the isosmotic condition (12‰) [2,7,11,13,50]. This shape of Na+/K+-ATPase activity has been associated to relevant changes in gill energy metabolism during adaptation to extreme salinity conditions, as 3‰ and 50‰, and could represent an increase in consumption of alternative energetic substrates such as lactate [9].There is no doubt that salinity challenges, despite well accommodated, represent an energy-demanding stress. The most commonly used physiological stress indicators in fish are plasma hormones, namely cortisol, and metabolites [79,80]. Cortisol is the principal glucocorticoid secreted under conditions of acute stress, via stimulation of the neuroendocrine system hypothalamus-hypophysis-interrenal (HPI), and a posterior cascade of metabolic and physiological changes occurs, making glucose and lactate readily available to the tissues [81,82]. In most fish, both metabolites and cortisol reach their highest circulating concentrations within a few hours, with plasma levels being stressor dependent and species specific, and with greater discrepancies and controversy when the stressor is chronic [83,84]. In the current chronic challenges, it was unexpected to observe higher cortisol levels in the control group, but that had no correspondence with the downstream metabolites. Cortisol is also associated with the onset of hyposmoregulation, namely in salmonids and euryhaline fish [77,85,86], but this is again not clear in our data as fish in 50‰ showed the lowest cortisol values. As for other prolonged stressful situations, it is possible that the response of the HPI may be less sensitive. Further studies will be necessary to assess this concern in fish exposed to the more challenging situations [87].When related to chronic salinity acclimation, plasma lactate appeared to be the most affected indicator. Usually plasma lactate tends to increase in situations of increased activity or reduced oxygen availability [82,87], but it has been observed in both hypoosmotic and hyperosmotic acclimated fish [88,89]. It could be considered that lactate becomes an important source of energy during osmotic acclimation as it can supply energy to different tissues, such as gills, kidney and brain as it was previously reported in several fish species [9,11,90,91,92]. For its part, plasma glucose concentration registered no significant differences due to salinity challenges, becoming a less discriminating metabolite during long-term osmotic acclimation, as also reported for other stressors [8,9,12,13]. No clear function for plasma protein has been suggested yet in long-term osmotic acclimations, but Sangiao-Alvarellos et al. [8] hypothesized that plasma protein functionality could be related to secondary metabolic reallocation of energy resources, once carbohydrate storage has been mobilized and exhausted. In our study, soluble protein is significantly changed only for the 3‰ condition, showing an increase. While some authors indicate that plasma protein increases as salinity changes [8], other authors report no changes [93] or even a reduction [94].4.2. Gill Mucus Cells and Skin Mucus Changes in Response to SalinityIn this study, the impact of salinity on gill remodeling was analyzed histologically by total mucus cell counts and by measuring cell parameters, such as mucus cell class, frequency and shape. Overall, the frequency of mucus cells in gill filaments increased in fish exposed to higher salinity. Although we have not measured production, it is likely these cells were active, thus increasing the amount of mucus secretion in the hyperosmotic environment in relation to other conditions. Gill mucus was suggested to provide a selective semi-permeable layer during salinity change [95] and in the Japanese eel (Anguilla japonica), gill mucus cells were activated after SW transfer, producing Na-binding molecule-containing mucus layers that protect against the high osmolality of SW [96], while in the red sea bream (Pagrus major) the amount of secreted mucus showed a decreased trend when transferred to low salinity [97]. These previous observations are in agreement with our cell frequency data.The nature of the mucins produced is also modified, and the mucus cells of shi drum (Umbrina cirrosa) reared at full seawater contained a mixture of acid and neutral mucins, whereas in fish adapted to hypo-osmotic environment only neutral mucins were observed [13]. In fact higher cell counts of neutral mucin mucus cells in freshwater and acidic mucin mucus cells in seawater have been recorded in different species, such as gilthead sea bream, rainbow trout [98], and shi drum [13]. In agreement with those results, we recorded a gradual increase in acid mucin mucus cells in response to salinity with the highest frequency at 50‰, which almost doubled the 3‰ cell frequency. Concomitantly, the frequency of neutral mucin mucus cells, which were much lower than for acid mucin mucus cells, increased in the lowest salinity. However, interestingly, the highest salinity and cell frequency was also over threefold greater in both extreme conditions than at 35‰.With regard to the role of skin mucus, it is known to be involved in fish osmoregulation. In this line, discussing the role of mucus in freshwater fish, Shephard [42] mentioned the impermeability function of skin mucus in water and ion flux diffusion but suggested this function would only reduce water diffusion by 10%. Mucus (skin and gill) were also previously suggested to support active ion uptake by concentrating cations from an ion-deficient environment [99]. In the same way, skin mucus as a polyanionic gel, has the potential to trap cations and allow anion diffusion [100,101], thus reducing both ion gradients to the plasma and ion transport processes and thereby allowing for a reduction in the cost of ion transport [102]. In our observations, skin mucus maintained osmolarity identical to seawater at high salinities but above environmental water osmolality at low salinities, as previously reported for freshwater salmonids [44], thereby reinforcing its possible role as an ion capture mechanism [99]. In addition, we found that the mucus concentration of Na+ and Cl− the main osmosis-related ions, represented a very low proportion of the total osmolality in the 3‰ and 12‰ conditions and that these ions were also slightly underrepresented in mucus at other salinity conditions in relation to seawater ion concentrations. This difference indicates that osmoregulation might be mediated by other molecules exuded in mucus at higher proportions at low salinities than under control conditions or high salinities, thereby preserving osmosis-related ions inside the body of the fish. Nevertheless, despite significant decreases in absolute values, mucus K+ concentration was relatively conserved at values similar to those of 35‰ seawater when fish were transferred to 12‰ or 3‰, but increased by over 100% in the mucus of fish at 50‰, suggesting that it is retained and that it may have a relevant role, perhaps for the activity of immune-related enzymes. To the best of our knowledge, there is no information in the literature regarding alteration of the secretion of skin mucus components, such as ion-binding proteins, during salinity acclimation, but salinity-induced changes in mucus lysozyme activity have been reported in salmonids [103,104].The changes in ions and osmolality observed in mucus also indicate that the osmotic pressure of the mucus layer may not only buffer the chronic entry of water and loss of ions across the skin at low salinities but that at high salinities, the high concentration of ions in the skin mucus may increase water exudation from the skin which would contribute to increased mucus volume and skin or muscle dehydration. Indeed we showed that fish transferred from 35‰ to 50‰ showed the highest increase in mucus production; however, skin mucus production, either in absolute volume exuded or in relation to fish size, increased when fish were transferred to both hypoosmotic or hyperosmotic conditions. This seems to indicate that the volume of mucus exuded is not only related to simple osmotic gradients but may be regulated. On the contrary, the osmotic parameters of the mucus after 15 days of acclimation to the different salinity conditions were similar to those observed in our previous studies using acute osmotic stress at 3 h [27], suggesting that there are no specific long-term mechanisms to further control skin mucus osmolality. Further studies in protein-osmotic components of skin mucus and how these respond to the hormonal changes that take place during hypo-or hyper-osmotic acclimation will be of great interest to understand better the characteristics, properties and role of skin mucus in fish subjected to salinity challenges. It also remains to be determined if salinity changes the nature of secreted mucins and skin mucus cells, as it happens in gill mucus cells.4.3. Energy Waste by Mucus ExudationAlthough skin mucus biomarkers have been considered a powerful tool for determining fish welfare and physiological status via less invasive methods [23,24,25,26,30,37,39,105], there is little work on the use of mucus analysis for the study of long-term chronic stress in fish. In addition to modifying circulating protein concentration, salinity challenges also resulted in important changes in soluble protein exudation into mucus, which significantly increased for the extreme conditions 3‰ and 50‰. This increase begins immediately after transfer, as we previously reported [27], and values are even higher after 15 days of exposure. The reasons for this increase are elusive, since the proteins accounting for it could not yet be identified. It is possible that those may be related to enhanced protection due to the relevance of skin protein components in innate immune defences [31], to higher mucus viscosity [23,25] or to the osmoregulatory properties of skin mucus which may provide a barrier to the movement of ions and water. Lactate exuded in mucus was also greatly affected by salinity, and its concentration was over two-fold higher in the 50‰ condition than in the control group. Although no similar data yet exists in the literature, we recorded the individual volumes of mucus exudation and we were able to calculate the total amount of each biomarker in the mucus. These data could be used to evaluate the energy expenditure and waste required to maintain the considerable amounts of exuded mucus in response to salinity challenges. For the first time, we have demonstrated that these increased amounts of mucus exude large amounts of soluble proteins, lactate and even cortisol, with the 50‰ condition once again causing the greatest effects. Sangiao-Alvarellos et al. [9] reported that the acclimation period is composed of an initial stage of increased energy use (i.e., increases in glucose and lactate) and reorganization of tissue energy metabolism, both in osmoregulatory (gills and kidney) and non-osmoregulatory (liver and brain) tissues. This is followed by the second stage of homeostasis in osmoregulatory parameters and a return to normality of metabolic parameters. Here, we found that energy modification varied mainly in lactate metabolism but also with an important release of protein to mucus, while glucose metabolism was not affected. Thus, we calculated the new steady-stage of sea bass in the hypersaline condition, to be a high energy loss stage compared to control. This is also true for the lowest salinity group, although with a lesser impact, both showed that despite the species euryhalinity there is an important allostatic load to be considered during hyper and hypo-osmotic acclimation. This evidence matches the observations of Boeuf and Payan [50], who measured less energy expenditure for osmoregulation in isosmotic conditions, sparing resources for other physiological processes, such as growth. In addition, mucus exudation showed the same production volume at 50‰ salinity as that of acute stress for 3 h [27], which would mean that fish status was not improved from that of the initial stress being a chronic and putatively harmful condition for the animal. Following the same reasoning, fish maintenance under the 3‰ and 12‰ conditions increased exuded mucus volume similar to those from acute stress [27]. This emphasizes the relevance of skin mucus for coping with salinity acclimation, but also suggest that the initial tolerance of freshwater conditions [27] would result in a chronic challenge in terms of energy expenditure as greater amounts of exuded soluble protein and lactate demonstrate. Whether such fast and long-lasting changes in skin mucus properties are specific for sea bass or differ in species with other degrees of euryhalinity remains to be seen.5. ConclusionsIn short, we compared the acclimation of European sea bass one hypoosmotic (3‰), one isosmotic (12‰)and one hyperosmotic (50‰) salinity conditions after 15 days, by measuring morphometric parameters, skin mucus and plasma stress biomarkers, and osmoregulation parameters, together with gill energetic and structural remodeling. Growth was not significantly affected, but a tendency towards decreased growth was noted in extreme conditions (3‰ and 50‰). The volume of skin mucus exuded proved to be an informative parameter: an exacerbated expenditure of energy was recorded in the hypersaline condition, and to a lesser extent in hyposaline conditions, with regard to control values. Gill Na+/K+-ATPase activity showed a typical “U-shaped” pattern, while gill remodeling resulted in a shift from neutral to acidic mucin mucus cells when moving from hyposaline to hypersaline conditions, together with a decrease in size and an increase in frequency. Skin mucus osmoregulation shifted from facilitating ion capture and ion transport at low salinities to retaining water at high salinities. Herein, we demonstrate the usefulness of skin mucus as a minimally invasive tool to analyze chronic situations, like salinity changes, and the need for further studies of the functions of mucus metabolites.

animals : an open access journal from mdpi

10.3390/ani11030838

PMC8002347

Cattle are fed a high-concentrate diet to improve their productivity; however, it alters the rumen ecosystem due to high structural carbohydrates level, resulting in ruminal acidosis. This study investigated the effect of changing diet on ruminal fermentation parameters, bacterial community composition, and expressed genes of Holstein Friesian cows, with changes induced by transition from a high-forage to two succeeding high-concentrate diets, and then returned to a high-forage diet. Ruminal pH drastically decreased; however, ammonia nitrogen, and individual and total volatile fatty acid (VFA) concentrations increased during the high-concentrate diet period. High-concentrate diet also reduced rumen bacterial richness and diversity. Gene expression in rumen epithelia was affected and altered by changing diet through the obtained differentially expressed genes.

Effects of changing diet on rumen fermentation parameters, bacterial community composition, and transcriptome profiles were determined in three rumen-cannulated Holstein Friesian cows using a 3 × 4 cross-over design. Treatments include HF-1 (first high-forage diet), HC-1 (first high-concentrate diet), HC-2 (succeeding high-concentrate diet), and HF-2 (second high-forage diet as a recovery period). Animal diets contained Klein grass and concentrate at ratios of 8:2, 2:8, 2:8, and 8:2 (two weeks each), respectively. Ammonia-nitrogen and individual and total volatile fatty acid concentrations were increased significantly during HC-1 and HC-2. Rumen species richness significantly increased for HF-1 and HF-2. Bacteroidetes were dominant for all treatments, while phylum Firmicutes significantly increased during the HC period. Prevotella, Erysipelothrix, and Galbibacter significantly differed between HF and HC diet periods. Ruminococcus abundance was lower during HF feeding and tended to increase during successive HC feeding periods. Prevotella ruminicola was the predominant species for all diets. The RNA sequence analysis revealed the keratin gene as differentially expressed during the HF diet, while carbonic-anhydrase I and S100 calcium-binding protein were expressed in the HC diet. Most of these genes were highly expressed for HC-1 and HC-2. These results suggested that ruminal bacterial community composition, transcriptome profile, and rumen fermentation characteristics were altered by the diet transitions in dairy cows.

1. IntroductionDairy cattle feeding patterns have been changed to provide the required energy and nutrients by feeding them concentrate feeds instead of fiber-rich forages [1]. Dairy cattle frequently undergo dietary transitions to meet the energy requirements for milk production around the start of their lactation period [2]. Such dietary transitions have supported the increase in milk yields; however, they raise concerns about rumen function in these cattle [1,3,4]. These dietary transitions affect chewing behavior and rumen buffering, which may lead to accumulation of large amounts of volatile fatty acids (VFA) in the rumen fluid [5]. Furthermore, the transition of the diet from high forage to high grain results in greater VFA concentrations and accumulation of lactate in the rumen, thus increasing the risk of subacute ruminal acidosis (SARA) [6,7,8,9,10]. SARA has been associated with metabolic and microbial alterations, and imbalances in the rumen, which are involved in metabolic health disorders in dairy cattle [11,12]. Among its various manifestations, it reduces cellulolytic bacteria counts, leading to impairment of bacterial activity due to undesirable ruminal pH, which may result in low fiber degradation [13,14,15,16]. Energy and essential nutrients are mainly obtained by ruminants through a complex symbiotic relationship with the rumen microbiome [17]. The health and efficiency of the host ruminants are significantly affected by changes in the bacterial community in the rumen [18]. While a high-forage diet is usually switched to a high-concentrate diet to improve the productivity of ruminants, it can alter the rumen ecosystem [17].The rumen is a complicated environment of microbes and hosts numerous amounts of bacteria that constitute an efficient mutual relationship of host animal and microorganisms [19,20,21]. Ruminants rely on ruminal microbes for the degradation of structural carbohydrates, and VFA and microbial protein synthesis served as the main sources of protein and energy [22]. The combined activities of a wide range of bacterial community in the rumen metabolizes the carbohydrates from feed [23]. Normally, various groups of bacteria are used to specialize for the utilization of specific types of polysaccharides, such as starch or cellulose, which are known to be digested by saccharolytic and cellulolytic bacteria, respectively [24]. Starch is one of the most abundant polysaccharides found in high-grain diets, which are mainly used to increase animal performance and are more favorable to saccharolytic bacteria [25,26,27,28]. However, the use of starch-rich feedstuffs was associated with the reduction of bacterial richness and diversity in the rumen and large intestines [23], which resulted in a decrease in relative abundance of Bacteroidetes and an increase in case of Firmicutes [29,30,31]. Ruminal saccharolytic bacteria can be considered as potential initiators of SARA, as higher starch basically results in greater production rates of short-chain fatty acids [23]. A previous study investigated the impact of SARA on the rumen microbial community using high-throughput 16S rRNA sequencing and found that there was a major shift in the three dominating phyla [1]. Another next-generation sequencing (NGS) technique is the RNA sequencing technology, which is an indispensable tool for meta-transcriptome analysis of different gene expression [32]. Meta-transcriptome analysis is more powerful compared to other NGS methods because of the information it can provide about microbial populations that are transcriptionally active [33], as well as the in-depth analysis of functional gene activity and metabolic pathways [34]. However, meta-transcriptome analysis in rumen with regards to SARA is limited [35]. In addition, despite adaptation and recovery of rumen bacteria during SARA challenge through high grain-feeding, the effect of changing diet in dairy cattle is largely unclear. This study was, therefore, undertaken to investigate the effect of changing diet on ruminal fermentation parameters, bacterial community composition, and transcriptome profile of Holstein Friesian cows, with changes induced by transition from a high-forage diet to two succeeding high-concentrate diets and then back to a high-forage diet. The findings provide plausible information that diet transitions would induce changes in ruminal pH and fermentation ability, consequently altering rumen bacterial community composition and transcriptome profile in dairy cows.2. Materials and Methods2.1. Animal CareThis study was conducted at the Sunchon National University (SCNU) animal farm and in the Ruminant Nutrition and Anaerobe Laboratory, the Department of Animal Science and Technology, SCNU, Suncheon, Jeonnam, Korea. The animals used and all experimental protocols were reviewed and approved by the Sunchon National University Animal Research Ethics Committee (SCNU IACUC, approval number: SCNU IACUC-2018-01).2.2. Animals, Feeding, and Experimental DesignThree rumen-cannulated Holstein Friesian cows, with average body weights of 600 ± 47 kg, were used in this study. A 3 × 4 cross-over design was used for this experiment to evaluate the effects of changing diet on rumen fermentation, bacterial composition, and transcriptome profile of the animals. This changing diet challenge was defined as the high-forage period followed by two succeeding high-concentrate periods, and then a return to the high-forage period to investigate the changes in ruminal pH and the adaptive capability of the rumen microbiome. Prior to the experiment, cows were fed with mixed Klein grass hay and concentrate at a ratio (kg) of 8:2. During the experimental period, all cows were fed twice a day continuously with the same diet ratio for 2 weeks. Then, after the first period, diet was changed to the ratios (forage:concentrate) as follow: 2:8 (2 weeks), 2:8 (2 weeks), and 8:2 (2 weeks). These feeding ratios served as treatments (HF-1 (first high-forage diet), HC-1 (first high-concentrate diet), HC-2 (succeeding high-concentrate diet, and HF-2 (second high-forage diet, served as a recovery period)). The HF-1 consisted of dairy cows fed a high-forage diet for 2 weeks. Then, it was followed by HC-1, which involved the same group of cows fed for 2 weeks with high-concentrate right after the HF-1 period prior to acidosis. Afterwards, a succeeding 2 weeks of feeding with high-concentrate diet were allotted for HC-2 in the same group of cows, followed by a transition to a high-forage diet (HF-2), which served as a recovery period for the animals. Animals were housed in individual stalls and had free access to water. The concentrate given to dairy cows was supplied by Purina® Cargill, Korea. The chemical compositions of the forage and concentrate fed to the dairy cows are shown in Table 1.2.3. Rumen Fluid Collection and Analysis of Ruminal Fermentation ParametersRuminal contents were collected from the three rumen-cannulated cows before morning feeding on the last day of each period before transitioning to the next diet period. The fluid samples were squeezed and strained through a four-layer surgical gauze and placed in an amber bottle, which was subsequently capped after collection. The collected fluid samples were maintained at 39 °C and immediately transported to the laboratory for analysis of the ruminal fermentation parameters. An aliquot of rumen fermenta was separated from the sample bottles, transferred to two 1.5 mL microcentrifuge tubes, and stored at −80 °C, prior to ammonia-nitrogen (NH3-N) and VFA analyses. Following this, the samples were thawed at room temperature and were centrifuged for 10 min at 13,000 rpm at 4 °C using a Micro 17TR centrifuge (Hanil Science Industrial, Korea). The obtained supernatant was used for NH3-N and VFA concentration analyses. The NH3-N concentration was measured according to the colorimetric method developed by Chaney and Marbach [36] using a Libra S22 spectrophotometer (Biochrom Ltd., CB40FJ, Cambridge, UK) at an absorbance of 630 nm. The analysis of the VFA concentration was done using high-performance liquid chromatography (Agilent Technologies 1200 series, Tokyo, Japan) with an ultraviolet (UV) detector set at 210 and 220 nm. Samples were isocratically eluted with 0.0085N H2SO4 at a flow rate of 0.6 mL/min and a column temperature of 35 °C [37].2.4. Metataxonomic and Transcriptomics AnalysesRumen fluid samples for metataxonomic analysis and rumen tissue biopsies for transcriptomics analysis were obtained from each treatment, and were sent to Macrogen (Seoul, Korea) for sample preparation, library construction, sequencing, and data analysis.For metataxonomic analysis, after quality assessment of extracted DNA from the samples, a survey of bacterial community composition was performed. Libraries were constructed by targeting the V3–V4 variable regions of the 16S ribosomal RNA (rRNA) gene using Herculase II Fusion DNA Polymerase Nextera XT Index Kit V2 following the Illumina’s 16S Metagenomic sequencing library preparation. After assessing the quality of the constructed libraries, sequencing was done using the Illumina sequencing platform (Miseq). Then, the generated paired-end (PE) sequencing raw reads from the base call binary data obtained by real-time analysis were converted to fastq format. Filtering of PE raw reads from overhang adapter sequences were done, then the filtered PE reads were merged. Quality filtering, trimming of short and extra-long reads, and removal of duplicate reads were done, then the filtered reads were clustered at 100% identity using CD-HIT-OTU [38]. After identifying chimeric reads, noise filtering was done, and the remaining clusters were binned to operational taxonomic units (OTU) with a cut-off value of 97% species level identity using the same program. Afterwards, taxonomic assignment of representative sequences from observed OTUs were matched by BLAST on the NCBI 16S rRNA database using Quantitative Insights Into Microbial Ecology (QIIME-UCLUST) [39]. The bacterial diversity was graphically presented using Metagenomics Core Microbiome Exploration Tool (MetaCOMET) [40] by uploading a biological observation matrix to the web server [41], generated using Mothur [42].For transcriptomics analysis of the host’s response to challenge diet, total RNA were extracted from the rumen epithelial tissue biopsy. DNA contamination was eliminated using DNAse. After quality assessment of extracted RNA from the samples, libraries were prepared using the TruSeq Stranded Total RNA LT Sample Prep Kit (Illumina, San Diego, CA, USA) with Ribo-Zero following the manufacturer’s Prep Guide. Sequencing was carried out using the TruSeq 3000 4000 SBS Kit v3 in an Illumina HiSeq 4000 sequencer (Illumina, San Diego, CA, USA). The sequenced raw reads were subjected to quality control by calculating the overall reads’ quality, total bases, total reads, and guanine-cytosine (GC) percentage. Then, sequencing artifacts like low-quality reads and adapter sequences were filtered out using Trimmomatic [43]. The remaining trimmed reads were then mapped against cow reference genome (Bos_taurus_UMD_3.1.1/bosTau8) retrieved from the University of California Santa Cruz genome database website (genome.ucsc.edu), using hierarchical indexing for spliced alignment of transcripts 2 (HISAT2) [44] by handling spliced reads’ mapping using Bowtie2 aligner [45]. Transcripts were assembled using StringTie [46] with aligned reads to provide information of known, novel, and alternative splicing transcripts based on the reference genome model. Afterwards, the level of abundance of transcripts in fragments per kilobase of transcript per million mapped reads (FPKM) was normalized. The FPKM value of known genes was used to sort the differentially expressed genes among samples. Three comparison pairs were conducted, where sample HF-1 served as the point of comparison of the three samples to determine whether a certain known gene was up- or down-regulated between diet types. During this analysis, low-quality transcripts were filtered, then the FPKM normalization was performed. Statistical analysis was performed using Fold Change per comparison pair (fc), and significant genes were selected if the fold change per comparison pair was greater than or equal to two. Afterwards, hierarchical clustering analysis was performed among the significant genes in order to group the genes and samples based on similarity of expression patterns.2.5. Statistical AnalysisAll analyses were carried out using Statistical Analysis Systems (SAS) software version 9.4 (SAS Institute 2012) (SAS Institute Inc., Cary, NC, USA). Data were statistically evaluated using Proc Glimmix for a complete randomized design. The experiment was done twice, and the treatments were conducted in triplicate. Least square means was used to identify differences among treatments. A p < 0.05 was considered indicative of significant differences.3. Results3.1. The Effects of Treatments on Rumen Fermentation ParametersThe ruminal pH was significantly higher (p = 0.001) during the high-forage diets. As the feeding switched to a high-concentrate diet, a lower ruminal pH was observed. The rumen pH of HC-1 and HC-2 was significantly (p = 0.001) lower than the HF periods due to the high concentrate ratio in the diet. The NH3-N concentrations for HF-1 and HF-2 were significantly lower (p = 0.001) and had an opposite result to the cows receiving the high-concentrate diet. Furthermore, in the HC diet group, the individual and total VFA concentrations were significantly higher (p = 0.001) than the HF group. The acetate to propionate ratio was significantly lower (p = 0.001) during HC-1 and HC-2, then during the HF feeding period, an opposite result was observed. Meanwhile, a significantly lower (p = 0.001) VFA concentration was observed during the HF diet periods (Table 2).3.2. General Ruminal Bacterial Community CompositionThe bacterial community of the rumen microbiome after diet transition is presented as a Venn diagram in Figure 1a. Out of 353 overall representative species of observed OTUs, a total of 125 (35.4%) were present across all communities (core), 142 (40.23%) were observed as shared by 2 or 3 communities, and a total of 86 (24.36%) were distributed uniquely in the four communities. This figure also shows that as the diet changes from high forage to high concentrate, the number of bacterial species also increases. However, at the point when the diet was reverted back to high forage (HF-2), the number of bacterial species decreased, but was still of higher count compared to HF-1. It can also be noted that these bacteria uniquely shared between two communities were higher between the same diet (HF-1 and HF-2: 27, HC-1 and HC-2: 32) compared to shared species between different diet types (HF-1 and HC-1: 1, HF-1 and HC-2: 6, HC-1 and HF-2: 5), except between HC-2 and HF-2, for which the number of species observed was higher than ten. The unweighted unifrac diversity principal coordinate analysis (PCoA) plots showed a similarity within the diet changing stage (Figure 1b). The high-forage diet periods (HF-1 and HF-2) had similar bacterial communities with non-distinct clusters and no spatial separation among populations. A portion of the bacterial community of HC-2 was also related to HF-2 due to the transition of the diet from high-concentrate to high-forage. There was also a tendency toward more similar bacterial communities between the high-concentrate diet periods (HC-1 and HC-2).A total of 15 bacterial phyla were identified within the rumen bacteria communities. The most abundant phyla included Bacteroidetes, Firmicutes, and Actinobacteria. With these phyla, Bacteroidetes was dominant in all treatments. However, its relative abundance decreased as cattle received a high-concentrate diet and a notable increase in abundance when the diet changed to high forage. This was contrary to the relative abundances of Firmicutes. As the rumen was subjected to the high-concentrate diet condition, the abundance of Firmicutes significantly increased (p = 0.039). A similar response was observed with the phylum Actinobacteria. Its abundance increased as the high-concentrate diet condition was prolonged (Figure 2a).The analysis of genus level composition revealed 143 identified genera, of which 19 were dominant (>1% of the relative abundance) (Figure 2b). The relative abundance of Prevotella, Erysipelothrix, and Galbibacter significantly differed between the high-forage and the high-concentrate diet periods (p < 0.05). Prevotella had a higher relative abundance during the first high-concentrate diet period. However, it significantly decreased (p = 0.024) as the high-concentrate diet was continuously supplied to cows. Meanwhile, Erysipelothrix and Galbibacter both showed similar responses to the treatments. These genera significantly decreased their abundances as the diet transitioned from high forage to high concentrate (p < 0.05), and gradually increased as the rumen recovered with the return to the high-forage diet. Ruminococcus had lower abundances under the high-forage diet period and tended to increase in abundance, though not significantly, during the two subsequent high-concentrate diet periods (p = 0.086). A similar pattern as with Ruminococcus was observed in the case of Ethanoligenes and Treponema, which showed high abundance during the periods of high-concentrate diet and the least in both high-forage diets.At the species level, Prevotella ruminicola was the most dominant species for all treatments, with relative abundances of 12.61%, 9.57%, 11.70%, and 14.67%, respectively (Figure 2c). Its abundance decreased as the diet transitioned from the high-forage diet to the high-concentrate diet. Meanwhile, as the high-concentrate diet period continued, a notable increase in abundance was observed. A similar scenario was observed with Paraprevotella clara, Prevotella brevis, and Capnocytophaga cynodegmi. The transition of the diet from high forage to high concentrate increased the abundances of Ethanoligenens harbinense, Ruminococcus bromii, Prevotella histicola, Prevotella jejuni, Olsenella umbonata, Anaerobacterium chartisolvens, and Pseudoramibacter alactolyticus. There was also a continual increase in their relative abundances when the high-concentrate diet was supplied consecutively. Furthermore, Prevotella oralis and Olivibacter sitiensis had a drastic reduction of abundance after changing the diet from a high-forage to a high-concentrate diet; however, these particular species recovered when the diet returned back to a high-forage one. In case of Prevotella buccalis and Prevotella marshii, the contrary was observed when the diet was shifted from high-forage to high-concentrate. The relative abundance radically increased, but as the high-concentrate diet continued, a remarkable decrease in abundance was noticed. A statistical comparison of single species showed a significant effect of treatments only in the case of Galbibacter mesophilus and Erysipelothrix larvae. These species were both significantly more abundant for treatments with the high-forage diet (p = 0.006; p = 0.010). A heatmap was also generated showing the relationship of each bacterial species on the pH values of treatments clustered based on Pearson correlation (Figure S1). The normalized data presented in Figure 3 shows the clustering based on similarity of relative abundance of families and treatments. The four communities after the diet transitions were clustered based on diet type, indicating close similarity of family abundance in each diet type. On the other hand, families were clustered based on their relative abundance, distinguishable by color based on their normalized value, by which two major clusters were formed: cluster of families with low relative abundance (lower cluster), and families with higher relative abundance (upper cluster). Distinctly observed is a subcluster containing the highest relative abundance, and among the families in this cluster, Ruminococcaceae, Flavobacteriaceae, Erysiopelotrichaceae, and Rikenellaceae displayed significant differences (p < 0.05) between treatments. Family Ruminococcaceae was initially lower during the HF diet, but significantly increased as the diet was altered to HC, and dropped when the diet was reverted to HF. This pattern was oppositely observed in Family Flavobacteriaceae, wherein abundance had decreased when changed to HC, then increased significantly when diet was reverted to HF. At the same time, Families Erysipelotrichaceae and Rikenellaceae were observed to be higher in HF1 and HF2, respectively.The boxplot representation of alpha diversity indices is shown in Figure 4. Alpha diversity indices are composite indices that reflect abundance and consistency. The Chao1, which estimates species richness, significantly increased during the high-forage diet (p = 0.025), while the opposite was observed when cattle received the high-concentrate diet (Figure 4a). Shannon index, which reflects the diversity of the OTU in samples, presented high-forage diet groups as the most diverse (p = 0.013) among treatments, and high-concentrate diet groups as the least (Figure 4b). Moreover, Figure 4c shows the boxplot of OTUs of observed species from the samples. The number of OTUs in both high-forage diet groups were higher (though not significant; p = 0.259) than the other groups. The diversity index is used to analyze the temporal and spatial changes in species composition, which reflects whether bacterial communities between groups have differences.3.3. Rumen Epithelial Differentially Expressed Genes between Treatments Assigned as the High-Concentrate (Treated Group) and High-Forage (Control Group) DietsThe sequence reads generated per sample ranged from 56,321,752 to 111,981,472 (Table S1). The number of mapped reads also ranged from 65.07% to 88.32%. Based on the mapped data statistics, the high-concentrate diet group had a greater number of reads than the high-forage diet group. There were 2266 (1120 upregulated and 1146 downregulated), 1494 (769 upregulated and 725 downregulated), and 1719 (899 upregulated and 820 downregulated) differentially expressed genes (DEG) when comparing HC-1, HC-2, and HF-2 groups with the HF-1 (control), respectively (Figure 5). The expression volume was defined as the geometric mean of two groups’ expression level. In order to confirm the genes that showed a higher expression difference compared to the control according to expression volume, a volume plot was drawn (Figure 6). It showed a summary of the top expressed genes obtained from the samples between the groups. There were nine top genes obtained after analysis, namely carbonic anhydrase 1, the RNA component of mitochondrial RNA processing endoribonuclease, S100 calcium binding protein A12, keratin 5, keratin 6A, keratin 14, keratin 15, basigin (Ok blood group), and peroxiredoxin 6. Figure 6a shows the top five expressed genes between the diet groups. Carbonic anhydrase I, the RNA component of mitochondrial RNA processing endoribonuclease, and S100 calcium binding protein A12, were highly expressed in HC-1; however, Keratin 15 and 5 were highly expressed in HF-1 (Table S2). Figure 6b shows that the RNA component of the mitochondrial RNA processing endoribonuclease, keratin 14 and 6A, and basigin (Ok blood group) were all highly expressed genes in the concentrate-treated group, except keratin 15, which was found to be highly expressed in the control group (Table S3). Meanwhile, the S100 calcium binding protein A12 and peroxiredoxin 6 were both highly expressed in the treated group, while keratin 1, keratin 14, and keratin 15 were highly expressed in the control group (Table S4, Figure 6c). The results of hierarchical clustering are presented in Figure 7, showing the clustering of genes and samples based on expression level (normalized value) from the significant list. The hierarchical clustering between samples showed that HF-1 was of a different cluster, separating HC-1, as well as HC-2 and HF-2, based on gene expression pattern. This evidently showed that transitioning of diet from high-forage to high-concentrate has an almost completely reversed pattern of expressed genes (HF-1 vs. HC-1). However, during HC-2, a different gene expression pattern was observed compared to the previous period of the same diet. Also, even when the diet was reverted back to high-forage, gene expression was different, which presented a more closely related expression pattern with HC-2.4. DiscussionA sudden decline in ruminal pH is currently one of the major health issues in dairy farming that causes feed intake reduction, digestion problems, and production losses. The prevalence of this mainly affects cattle health and drastically increases management costs. In all ruminants, the rumen is a complex microbial ecosystem and is inhabited by a large density of microbiota, bacteria, anaerobic fungi, archaea, and ciliate protozoa [50]. Rumen microbes have vital roles in the degradation of feedstuffs, and the production of VFAs, lipids, amino acids, and hydrogen, which are essential for the maintenance, growth, and production performance of ruminants [51]. These microbes also supply VFAs, proteins, and vitamins to their ruminant hosts through degrading and fermenting feed materials [17]. A better understanding of the rumen microbiome under extensive feeding conditions is essential due to the complexity of the rumen ecosystem, as the manipulation of ruminal microbiota can improve feed efficiency and optimize rumen function [52,53]. Previous studies have demonstrated that changes in rumen microbial communities are affected by several factors, such as ruminant species, age, health, season, geographical location, feed additives, and diet [18,54,55,56]. Specifically, a diet switch from high forage to high concentrate will definitely cause an enormous change in the rumen bacterial community, which can negatively affect productivity and has the potential to develop metabolic disorders in ruminants [57]. In the present study, we investigated the effect of changing diet on ruminal fermentation characteristics, bacterial community composition, and transcriptome profile of Holstein Friesian cows, which was induced by the transition from a high-forage to a high-concentrate diet, and then returned to a high-forage diet.Ruminal pH and its daily fluctuations are considered to be a major factor in the occurrence of SARA and the regulation of microbial activity [15,58,59]. Based on the proposed pH thresholds that define ruminal acidosis, it seemed that the Holstein cows in the present study were in SARA condition during the two successive high-concentrate diet periods. A similar result was obtained from the study conducted by Lee et al. [17], wherein the ruminal pH of a Holstein cow drastically decreased upon receiving a high-concentrate diet. The chemical composition of the rumen fluid during the transition from hay to high-concentrate diet can be associated with the development of subacute acidosis [57,60]. A ruminant diet with 35% concentrate reduced the rumen pH below 5.6 for more than 180 min/day, which was an indication of acidosis [15,61,62]. Generally, high-concentrate diets increase both lactic acid producers and utilizers, while decreasing the number of fiber-degrading bacteria. The high level of non-structural carbohydrates from a high-concentrate diet resulted in a drastic decrease in pH of the rumen [63]. Reduction in pH is due to the rapid fermentation of non-structural carbohydrates, and volatile fatty acid accumulation in the rumen [60]. Consequently, reduced pH was also a result of the accumulation of lactic acid in the rumen [17,64]. Moreover, Tajima et al. [64] added that once ruminal pH decreased below 5.6, acid-resistant microbes became dominant in the rumen, which can cause metabolic disorders.A high concentration of VFA was identified during the high-concentrate diet periods of the present study, which is in accordance with the results of Bevans et al. [65], Sato [66], and Nagata et al. [2]. Lee et al. [17] also stated that a reduction in ruminal pH was directly related to the concentration of total VFA. Previous studies demonstrated that an increase in the VFA concentration, rather than lactate, primarily affected ruminal pH in response to high-concentrate-feeding in cattle [67]. The result of the study by Tajima et al. [57] showed the total VFA concentration tended to increase when cattle were fed a high-concentrate diet. Moreover, Li et al. [15] added that the increase in the concentration of total VFA and the molar proportion of propionate was an indication of a large amount of starch, which usually occurs during the SARA condition. A typical response to an acidotic-like condition in the rumen is changes in the relative proportion of VFA [29,60]. The high concentration of butyrate in the high-concentrate diet period of the current study is also in agreement with the result of Wang et al. [68] and was consistent with that of Ribeiro et al. [69]. Similar to the present study, high propionate concentration in the rumen of cows fed a high-concentrate compared with a high-forage diet has been reported in the literature [70]. Greater propionate production might trigger gluconeogenesis and milk production response [71]. Meanwhile, acetate and butyrate can be converted into each other in the rumen [68]. According to a previous study, 28% of acetate is not absorbed by the rumen in the form of acetate [72] and microorganisms can use this to produce butyrate by acetyl-Coenzyme A transferase and/or butyryl-Coenzyme A transferase [73]. Rumen microbes can promote the metabolism of microorganisms through the energy dissipation process and convert acetate to butyrate continuously during a high-concentrate diet period [73]. Additionally, the present study also supports the claim of Nagata et al. [2], who identified a concurrent increase in the total VFA and NH3-N concentrations during the high-concentrate feeding period. Lana et al. [74] reported that forage- and concentrate-fed cattle had various populations of ammonia-producing bacteria in rumen; thus, the optimal ammonia concentration in rumen fluid resulted in the production of microbial protein and the maximum fermentation rate. An NH3-N concentration higher than 5 mg/L is the minimum requirement for maximal microbial growth and indicates that a high-concentrate diet can produce a large amount of ruminal microbial crude protein for utilization by the animals [75]. Moreover, the increased NH3-N concentration is indicative of an increased rate of proteolysis and amino acid metabolism in the animals [76].In concordance with many previous studies, the transition of a diet from high forage to high concentrate resulted in various changes in the condition and substrate availabilities in the rumen, which further reduced the richness and diversity of ruminal fluid microbiota [77,78,79,80]. Reports from many studies showed that concentrate-feeding and concentrate-induced SARA alter microbial community structures in the digestive tract [29,30,79]. Ruminant dietary changes contributed significant impacts on rumen bacterial communities [64,81,82]. A previous study demonstrated that dietary forage to concentrate ratio increased from 80:20 to 20:80, and distinctly changed the rumen bacterial population structure [79]. Both studies of Ogata et al. [83,84] presented that low ruminal pH decreased bacterial richness and diversity. During acidosis challenge induced by feeding a concentrate diet, the low ruminal pH could lead to the death and lysis of bacteria, resulting in low relative abundances [84]. The metagenomics survey of bacterial community composition in the present study agreed with previous research showing Firmicutes and Bacteroidetes as the most common phyla in the rumen [29,77,78]. Furthermore, those studies stated that the high-concentrate diet of cattle increased the relative abundance of Firmicutes, while decreasing the abundance of Bacteroidetes. Petri et al. [77] and Henderson et al. [54] also stated that Firmicutes and Bacteroidetes were both identified as the core rumen microbiome. This statement suggests that these two phyla may be less affected by rumen environment changes due to acidic challenges or dietary changes. Excessive grain feeding reduces richness and diversity of rumen microbiota, resulting in a relative abundance reduction of Bacteroidetes and an increase of Firmicutes in the rumen [77,78,85]. Studies have shown that the low proportion of Bacteroidetes during the high-concentrate diet period was due to high acidity in the rumen [66,78]. Meanwhile, Kaoutari et al. [86] concluded that Bacteroidetes was more efficient at degrading structural carbohydrates than Firmicutes. Moreover, the utilization of dietary fiber leads to an increased abundance of Firmicutes rather than Bacteroidetes. Henderson et al. [54] stated that the most dominant taxa in the rumen were the genera Prevotella, a well-known degrader of starch, β glycans, protein, pectin, and hemicellulose, and Ruminococcus, a cellulose degrader. This statement on the abundance of Prevotella was in accordance with the microbiome result of the present study. The ability of this genus to use a variety of substrates allows it to dominate in the rumen under a range of diets [87]. Furthermore, Prevotella species are prominent ruminal proteolytic bacteria which produce a variety of extracellular degradative enzymes [88], while Ruminococcus are specialized amylolytic bacteria known for the degradation of celluloses in the rumen [89]. Prevotella ruminicola appeared to be the predominating species in the present study. It is one of the most numerous groups recovered from the rumen and plays a role in degradation of polysaccharides [90,91,92]. The increase in abundance of Prevotella marshii and Prevotella jejuni during the high-concentrate diet condition might be due to its cell function as saccharolytic and its ability to ferment glucose [93]. On the other hand, the high prevalence of Prevotella oralis under the high-forage condition was likely due to their utilizing function on a wide variety of polysaccharides, and they are thought to be essential contributors to the degradation of xylan [94,95,96,97,98]. Whitehead [99] emphasized that Prevotella species in the rumen could contribute to cell wall degradation through synergistic interactions with various species of cellulolytic bacteria. Moreover, the research of Ntougias et al. [100] revealed that Olivibacter sitiensis is a xylanolytic bacterium which is involved in the cleavage of β-1,4-xylosic bonds in hemicelluloses. This statement supported the abundance of O. sitiensis in the present study, which is also observed during a high-forage diet condition. The significant increase in abundance of Galbibacter mesophilus and Erysipelothrix larvae in the high-forage diet was likely due to its capability to utilize celluloses, which is an important role in regulating the host’s metabolism, thus promoting efficient degradation of polysaccharides [101,102]. Although Ruminococcus is a well-known cellulolytic bacterium, several species, such as R. bromii, are capable of fermenting starch [103,104]. This may explain the abundance of R. bromii in the present study. Succiniclasticum are known to ferment succinate and convert it to propionate [51,79,105]. The abundance of Succiniclasticum ruminis in the high-concentrate diet periods might explain the high propionate concentration in the rumen. There are also some bacterial species which were found with lower abundance (not presented in figures) that have been reported to exhibit an essential role in pivotal rumen function as rumen homeostasis index due to their contribution in reducing sulfate and its metabolic flexibility [106]. Published studies have shown that this species has the ability to ferment and convert succinate to propionate, which is an essential precursor of glucose in ruminants [105].In the present study, sufficient information was provided on the differential gene expression pattern in rumen epithelial tissue’s response to the adverse effect of diet transition from a two-week high-forage diet, to a four-week acidosis challenge, back to a two-week high-forage diet. Apart from short-chain fatty acid metabolism, nutrient absorption and transportation are critical functions of the ruminal epithelium [35,107,108,109]. Rumen epithelial structure development is attributed to VFA absorption; thus, the excessive amount of VFA can lead to a reduction in pH that can damage the ruminal epithelium [108,110,111]. Moreover, a high-starch diet in ruminants can cause SARA, which may lead to the destruction of the rumen epithelial tissue in the long term [61,107]. Steele et al. [12] demonstrated that a significant reduction of the total epithelium depth occurred when using a high-concentrate diet-induced acidosis model. In a recent study of Li et al. [35], where cattle were subjected to rumen acidosis by feeding a high-starch diet, the rumen epithelial transcriptome showed a high number of genes that were differentially expressed, impacting biological pathways, specifically genes responsible for cell signaling and morphogenesis. The two assigned groups in the present study had nine major expressed genes, which were confirmed by the high expression difference through the expression volume plot. The high expression of carbonic anhydrase I (CAI) in the high-concentrate-feeding period was in agreement with the results of previous studies [112,113]. They stated that CAI catalyzed the rapid hydration and dehydration of CO2 and H2CO3, respectively. It also played essential roles in physiological systems, such as acid-based balance, respiration, bone resorption, ion transport, ureagenesis, lipogenesis, and gluconeogenesis. Moreover, it secreted parotid saliva containing a HCO3−-rich alkaline solution, which helped maintain the rumen pH in the range of 6–7 [112]. Meanwhile, the S100 calcium binding protein A12 (S100A12) is an important mediator in various cellular functions which involve apoptosis, cell proliferation, inflammation, and immunity, and is known to be associated with innate immune responses [114,115,116]. This may explain its high expression in the high-concentrate-feeding period, given the fact that these cattle were under SARA conditions, which has been associated with the inflammation of different tissues and organs of dairy cattle [58,110,117]. The high expression volume of the gene responsible for the release of keratin in the high-forage feeding period was supported by the results of previous research [118,119,120]. Keratin is known as the epithelial-specific member of the intermediate filament genes and proteins family, which is responsible for structural support and the regulation of metabolic processes [118]. Fiber-diet-fed ruminants were associated with hard keratins in their epithelium [119]. In a previous report on gene expression, Xiang et al. [120] stated that sheep fed different quality fibrous diets had full-thickness rumen wall tissue due to the keratin produced. The stratified squamous epithelium of the rumen surface was cornified and keratinized to protect it from physical damage caused by ingested plant material [121]. Basigin (BSG), also called CD147 or extracellular matrix metalloproteinase inducer (EMMPRIN), is a transmembrane protein that belongs to the immunoglobulin superfamily and is involved in various pathological and physiological phenomena. It is associated with several proteins, including monocarboxylate transporters (MCTs), which catalyze the proto-linked transport of monocarboxylates such as lactate, pyruvate, and ketone bodies, across the plasma membrane [122]. This protein also facilitates the transport of metabolites from the rumen epithelium to the blood [123,124] and possesses an essential role in rumen pH regulation [125,126,127,128]. These data on BSG support the results of the present study. Furthermore, the high expression volume of peroxiredoxin 6 (PRDX6) in our study was in accordance with previous research [129,130]. Hollmann et al. [129] observed that the expression of PRDX6 was downregulated in high-energy diet-fed animals. Meanwhile, Bondzio et al. [130] described PRDX6 as an important antioxidant enzyme that protects cells against oxidative injury by the reduction of endogenous levels of peroxides. This enzyme might also be involved in maintaining cellular homeostasis in the rumen epithelium during concentrate-supplemented diet adaptation. The comparative analysis of the transcriptome profiles revealed that changing the diet can alter rumen epithelia gene expression [131]. The gene expression pattern in epithelial tissue of the rumen was drastically affected as a result of diet transition. Although, the diet was reverted back to high-forage, the hierarchical clustering of HF-2 based on DEG pattern was more related to the patterns during the acidosis period, which could imply that these harsh changes on the transcriptome of rumen epithelium would require more time to recover to the normal gene expression pattern. Hence, analysis of DEG using RNA sequencing in an extended period of high-forage diet after acidosis challenge is recommended to assess how long would it take for the animals to recover their normal gene expression pattern.5. ConclusionsRuminal fermentation characteristics, the rumen bacterial community structure, and differentially expressed genes were affected by the changing diet and were induced by transition from a high-forage to a high-concentrate diet. Ruminal pH drastically decreased during the high-concentrate diet period, while a greater increase in concentrations of NH3-N and individual and total VFAs were observed. This study was able to confirm the changes in the rumen bacterial community and structure. Among the diets, the high-concentrate diet reduced the richness and diversity of the rumen microbiota. The metagenomics survey on bacterial abundance revealed that Bacteroidetes dominated all the treatments at the phylum level. Prevotella abundance significantly differed between the high-forage and high-concentrate diet periods and had a higher relative abundance among the microbial genera. Besides degrading starch, β glycans, protein, pectin, and hemicellulose, they also have the ability to use a variety of substrates, allowing them to dominate the rumen under a range of diets. Meanwhile, there were nine top expressed genes that satisfied the two-fold change based on the expression volume analysis. The differentially expressed genes’ analysis revealed that changing diet can alter gene expression in the rumen epithelia.

animals : an open access journal from mdpi

10.3390/ani13081428

PMC10135321

The aim of this study was to evaluate the effect of Ovopel on the reproductive effectiveness (measured as weight of eggs, egg quality, and ovulation rate) of females from two strains (Polish line 6 and Lithuanian line B) of common carp (Cyprinus carpio L.) and the release of luteinizing hormone (LH) and 17α,20β-dihydroxyprogesterone (17α,20β-DHP) during ovulation induction. After Ovopel treatment, the weight of eggs was higher in line 6, but the egg quality and ovulation rate were higher in line B. The observed differences in LH and 17α,20β-DHP levels between the lines were non-significant. LH levels between ovulated and non-ovulated fish did not differ within the lines. Steroid levels at 24 h after the priming dose were significantly higher in ovulated fish only in line 6. Summing up the results of comparing reproduction effects in two breeding lines of carp revealed higher reproduction effectiveness in B. Furthermore, the obtained results indicate that levels of tested hormones 12 h after the application of the resolving dose of Ovopel were higher in fish from that line which displayed higher reproduction effectiveness.

The study evaluates the impact of Ovopel on the reproductive effectiveness of carp from Polish line 6 and Lithuanian line B and the release of luteinizing hormone (LH) and 17α,20β-dihydroxyprogesterone (17α,20β-DHP) in females from these lines during ovulation induction. The levels of both hormones were determined in blood plasma samples taken just before the priming injection of Ovopel (0 h), at the time of administering the resolving dose of Ovopel (12 h), and after the next 12 h (24 h). Following Ovopel treatment, the mean egg weight obtained for line 6 was higher, but not statistically different, than that observed for line B. Egg quality, on the other hand, was significantly higher in line B. Female provenance did not significantly affect the number of eggs and living embryos after 70 h incubation. However, the total egg number for line 6 was higher. The mean number of living embryos (70 h) was similar for both lines. LH concentrations at 0, 12, and 24 h were not statistically different between the lines. A comparison of LH concentrations between ovulated and non-ovulated females at different sampling times revealed no significant differences either within or between the lines. Statistically significant differences in LH levels were found for both ovulated and non-ovulated females from a given line between the sampling times. The results for 17α,20β-DHP were similar, with only one difference: 24 h after the priming dose of Ovopel, 17α,20β-DHP levels in ovulated fish were significantly higher compared with non-ovulated females, but only in line 6.

1. IntroductionThe common carp (Cyprinus carpio L.) is a fish species of significant importance for the aquaculture industries of many countries. In 2019, global carp production reached 3,821,611 tonnes [1]. In 2020, carp production in Poland was estimated at around 21,000 tonnes [2].Even though carp can easily reproduce in captivity, spawning is not synchronized and usually occurs late in the growing season. This leaves only a short period for larval and fry training in their first year. In order to obtain as many fry as possible, carp are mainly bred under controlled conditions, not only in temperate countries. Such controlled reproduction involves the induction of ovulation using mainly carp pituitary homogenate/extract (CPH/CPE), e.g., [3,4,5,6,7,8] or mammalian/salmon gonadotropin-releasing hormone analog (GnRH-a) administered in combination with a dopamine antagonist, e.g., [9,10,11,12,13,14,15,16,17,18].Mammalian gonadotropins such as human chorionic gonadotropin (hCG) and pregnant mare serum gonadotropin (PMSG), have not found wide application in the controlled breeding of carp, even though satisfactory propagation results can be achieved in this species with the administration of those heterologous gonadotropins to females (either alone or in combination with CPH) [19,20,21,22,23].A series of studies on carp carried out at the Gołysz Institute (Poland) demonstrated that one cannot expect similar results of controlled reproduction in fish from different breeding lines after the induction of ovulation using preparations of natural origin (CPH, BPH, hCG, PMSG) and after the induction of ovulation with the use of synthetic preparations (GnRH-a), e.g., [6,13,14,15,21,24,25,26]. Differences in breeding results are observed between carp of different provenance even when the females used for breeding are of the same age and similar weight and are kept under the same conditions prior to the induction of ovulation, and when ovulation in females of various origins is induced using the same preparation and at the same time point [27]. Differences are usually observed between different lines in the ovulation ratio, weight, total number and quality of eggs, and, consequently, the number of living embryos and fry. The results of long-running studies on the stage of maturity of females from different breeding lines (as assessed by determining the degree of maturity of the oocytes of the oldest generation, sampled in vivo immediately prior to the administration of a spawning-inducing agent [28]) showed no significant difference in the stage of maturity between females of different provenance. In post-vitellogenic females, luteinizing hormone (LH—until 1999 called GtH) (whose synthesis and release from the pituitary gland is stimulated by the gonadotropin-releasing hormone—GnRH) stimulates the secretion from ovarian follicles of 17α,20β-dihydroxy4pregnen3 one (17α,20β-DHP), a maturation-inducing hormone—MIH [29,30,31,32]. 17α,20β-DPH induced the formation in oocytes of maturation-promoting factor (MPF), a complex of cdc kinase and cyclin B [33], which promotes the resumption of meiosis and subsequent ovulation. Thus, LH and 17α,20β-DPH are the main endocrine factors that regulate reproduction in post-vitellogenic females during the maturation of oocytes and ovulation [34,35].The authors of the present study wondered whether the levels of those hormones present in the serum of female carp of different provenance at the time when ovulation induction begins and after the administration of a spawning-inducing agent may have a significant impact on the results of controlled reproduction. In order to answer the question, we conducted an experiment on carp from two genetically distant breeding lines (Polish line 6 and Lithuanian line B) forming part of the “live gene bank” of the Gołysz Institute of Ichthyobiology and Aquaculture, where we applied Ovopel as an ovulation stimulator, which is a complex preparation successfully used for the induction of ovulation and spermiation in carp and other important fish species, e.g., [36,37,38,39,40]. The aim of the present study was to characterize and compare the reproductive effectiveness of these two breeding lines, identify any significant differences in the LH and 17α,20β-DHP levels in the samples collected from females before and after Ovopel treatment between and within each of the lines studied, and determine whether differences in the levels of LH and 17α,20β-DHP between ovulated and non-ovulated females from the same line at a particular sampling time were significant.2. Materials and Methods2.1. EthicsThe experiment was conducted in accordance with the principles of the Ethical Committee for the Protection of Research Animals at the Polish Academy of Sciences. The study was approved by the II Local Institutional Animal Care and Use Committee (IACUC) in Kraków, Poland (resolution No. 83/2019). 2.2. Handling of Spawners before the Induction of Ovulation and Treatments The experiment was conducted at the Institute of Ichthyobiology and Aquaculture in Gołysz (Polish Academy of Sciences) during the natural spawning period of carp (in May). The study included 28 females aged 10–11 years, weighing 7.1–11.1 kg. These included 13 females from Lithuanian line B and 15 females from Polish line 6 [41]. The fish were sampled based on external morphological signs of maturity (well-rounded soft abdomen and swollen genital opening [7,42]) from a large population harvested from the pond (in spring, when water temperature reached 18 °C). The females were transported to the hatchery and placed in seven 3 m3 tanks (4 females per tank). After a 2-day period of acclimatization, during which the water temperature was gradually increased from 18–19 °C to 20–21 °C, the females received a priming dose (1/5 pellet kg−1 BW) of Ovopel [D-Ala6, Pro9NEt-mGnRH-a) + metoclopramide] (Interfish, Budapest, Hungary). After 12 h, a resolving dose of Ovopel of 1 pellet kg−1 of the female’s BW was administered to the females. One pellet of Ovopel contains 18–20 µg of D-Ala6, Pro9NEt-mGnRH-a, and 8–10 mg of metoclopramide. Ovopel pellets, just like dried pituitary glands, were ground in a mortar and then used to prepare a suspension in saline [43]. Both doses were administered by intraperitoneal injection. In order to increase the yield and quality of the sperm used for fertilization, males from line B and line 6 received an intraperitoneal dose of Ovopel of 1 pellet kg−1 BW.A heparinized 21-gauge needle with a 1 mL syringe was used to collect serial blood samples (1000 μL) from the females by caudal venipuncture. The first sample was collected prior to the first injection of Ovopel (time 0 h), the second at the time of the second injection of Ovopel (time 12 h), and the third sample was collected at the time of fish ovulation checked at approximately 12 h after the second injection of Ovopel (time 24 h). The blood samples were centrifuged at 4000× g for 5 min at 8 °C, and plasma was stored at −80 °C until analysis.2.3. Handling of Eggs ObtainedFemales were checked for ovulation every 1 h as from 12 h post the resolving dose of Ovopel until 20 h after this dose. Eggs released in response to gentle abdominal pressure were stripped into a dry plastic container. The eggs obtained from each female were weighed and fertilized with pooled sperm collected from 3–4 males from the same breeding line as the females that had released the eggs. After the elimination of stickiness using the method described by Woynarovich and Woynarovich [44], 300 g of eggs from each female were incubated in separate Weiss glasses with a capacity of 7 L in water at a temperature of 21 ± 1 °C. After 24, 48, and 70 h of incubation, the percentage of living embryos was calculated for each female. In addition, the total number of eggs, taking into consideration the weight of 1 carp egg depending on the female BW class (as described by Cejko and Brzuska [45]) and the number of living embryos after 70 h of incubation were calculated for each fish. The percentage of ovulating females was also calculated for the lines studied and the latent period.2.4. Hormone AnalysisLH and 17α,20β-DHP concentrations in plasma samples were assayed using the ELISA method: LH as described by Kah et al. [46] and 17α,20β-DHP as described by Szczerbik et al. [47]. LH and 17α,20β-DHP concentrations were determined by measuring the absorbance using a 96-well plate reader (BIO-TECH INSTRUMENTS, EL 311) at 490 and 450 nm, respectively.2.5. Statistical Analysis2.5.1. Statistical Analysis Relating to the Reproduction EffectsBasic statistical tests of the data were carried out using standard procedures included in the Statistica 10 software (StatSoft Polska Sp. Z o.o.). For data expressed as percentages, transformations were made using the arcsine function before the analysis. Data on the breeding were analyzed using the least-squares analysis of variance [48] in order to determine the impact of the females’ provenance on the variables analyzed. The variables included the weight of eggs in grams, the weight of eggs expressed as a percentage of the female’s body weight, the percentage of living embryos after 24, 48, and 70 h of incubation, the total number of eggs, and the number of living embryos after 70 h of incubation. The analysis of variance was performed according to the following linear model:Yij = α + ti + bWij + eij where Yij is observation… j, α is the theoretical overall mean with the assumption that Wij = 0, ti is the effect of the female’s provenance, i = 1, 2…, b is partial regression in the female’s body weight, Wij is the female’s body weight, and eij is the random error associated with observation j.The analysis allowed for estimating least-squares means for the variables within each breeding line. The significance of the impact of females’ provenance on the variables determining reproductive effectiveness was tested using the F-test.As not all females from lines B and 6 released eggs at the same time following the administration of the resolving dose of Ovopel, the question arose whether the reproduction performance of fish that had released eggs earlier differed from the reproduction performance of those fish that had released eggs later. In order to answer this question, a least-squares analysis of variance was performed for each line, in which the classification factor was the latent period. The analyses were performed according to the following linear model:Yij = α + li + bWij + eij where Yij is observation… j, α is the theoretical overall mean with the assumption that Wij = 0, li is the effect of the latent period, I = 1, 2…, b is partial regression on the female’s body weight, Wij is the female’s body weight, and eij is the random error associated with observation j.The analysis allowed for estimating least-squares means for the variables determining reproduction effectiveness when the latent period differed both within line B and within line 6.The authors of the present study also examined whether the difference in reproduction effects between line B and line 6 was significant when the eggs in both lines were obtained at the same latent period. In order to answer this question, the least squares analysis was performed according to the linear model in which the main classification factor was the line. The significance of the impact of the latent period within line B and line 6 and of the line at the same latent period on the variables determining propagation effects was tested using the F-test. An X2 (chi-square) test was used to verify the dependence of distribution on classification (lines B and 6; ovulated and non-ovulated females) [49].A multiple regression equation was formulated for each line to predict the number of living embryos after 70 h of incubation. In this equation, the weight of females, the weight of eggs in grams, and the weight of eggs expressed as a percentage of the female’s BW were used as independent variables.2.5.2. Statistical Analysis Relating to Hormone AnalysisThe resulting LH and 17α,20β-DHP concentrations were analyzed using GraphPad Prism statistical software (version 5, 2007, GraphPad Software: San Diego, CA, USA). All data are presented as arithmetical means ± standard error of means (SEM). A non-parametric two-tailed Mann–Whitney U-test was performed for the comparison of hormone concentrations in fish from different lines. In the case of hormone levels measured over time within a given line of fish, the Friedman test followed by Dunn’s test was performed. The differences between the means were considered significant for p ≤ 0.05.3. Results3.1. Reproductive Performance 3.1.1. Female Ovulation after Ovopel Treatment and Latent PeriodIn females from Lithuanian line B and those from Polish line 6, ovulation occurred at two time points. In 30.7% of females from line B, ovulation occurred 13 h after administration of the resolving dose of Ovopel, whereas in 46.2% of females from this line, ovulation occurred after a further 2 h. In turn, 33.3% and 20% of females from line 6 released eggs 15 h and 18 h after the second injection of Ovopel, respectively.The ovulation ratio for line B was 10/13 and for line 6 8/15. The X2 test performed for the classification adopted and the distribution obtained did not reveal any statistically significant differences.3.1.2. Effect of the Provenance of Females on the Weight and Quality of Eggs ObtainedThe provenance of females did not significantly affect the weight of eggs obtained, either when expressed in grams or as a percentage of the female’s body weight. However, the least-squares means estimated for those variables clearly show that the weight of eggs obtained from females from Polish line 6 was higher by 303 g and 2.95%, respectively, compared with the weight of eggs released by females from line B (Table 1). The provenance of females significantly determined (p ≤ 0.05) the percentage of living embryos after 24, 48, and 70 h of incubation. The means estimated for those variables were higher for line B in the case of all three periods analyzed (Table 1). The percentage of living embryos after 70 h of incubation for females from line B was higher by as much as 14.5% than that calculated for line 6 (Table 1).3.1.3. Effect of the Provenance of Females on the Total Number of Eggs and the Number of Living Embryos after 70 h of IncubationThe effect of the line of the females was statistically insignificant in respect of the total number of eggs and the number of living embryos after 70 h of incubation. However, the least-squares means estimated for the variable clearly show that the number of eggs obtained from females from line 6 was higher by 209,000. The number of eggs for this line was higher by 104,000 than the mean value for the entire set (Table 1). The mean number of living embryos (70 h) was similar for both the lines studied (Table 1).3.1.4. Latent Period and the Weight and Quality of Eggs ObtainedOvulation time did not significantly determine the weight of eggs obtained from females from line B. However, it should be stressed that the weight of eggs released by females 13 h after administration of the resolving dose of Ovopel was higher than the weight of eggs obtained from those females from this line that ovulated 2 h later (Table 2). The percentage of living embryos after 24, 48, and 70 h of incubation was similar for both latent periods in females from line B (Table 2).Similarly, no statistically significant association was found between the latent period and the weight of eggs obtained from females from line 6. However, it was observed that the mean weight of eggs obtained from these females 15 h after administration of the second dose of Ovopel was higher by 351 g compared with the mean estimate for the weight of eggs released by the fish which ovulated 3 h later. The latent period was a significant determinant of the quality of eggs produced by females from line 6. The percentage of living embryos after 24, 48, and 70 h of incubation was higher for the latent period of 15 h (Table 2). It should be stressed that both after 48 h and after 70 h of incubation, the percentage of living embryos in eggs obtained from females from this line after longer latency was lower by as much as 34% (Table 2).A comparison of the mean weight of eggs (g) released by females from line B and females from line 6 15 h after the second injection of Ovopel showed that the mean calculated for this variable was significantly higher for line 6. The mean weight of eggs expressed as a percentage of the female’s body weight was also significantly higher for line 6 (Table 2). However, the quality of eggs from females from this line after 48 h and after 70 h of incubation was substantially poorer (Table 2).3.1.5. Latent Period and the Total Number of Eggs and the Number of Living Embryos after 70 h of IncubationA significant effect of ovulation time on the total number of eggs and the number of living embryos (70 h) was noted only for line B (Table 2). The least-squares mean estimated for the total number of eggs was higher for the latent period of 13 h. The number of eggs obtained from females from line 6 15 h after administration of the resolving dose of Ovopel was lower compared with the number of eggs obtained from those females that ovulated 3 h later. However, the number of living embryos (70 h) was higher by 90,000 in the case of the latent period of 15 h (Table 2).Analyzing the least-squares mean values for the number of eggs and the number of living embryos (70 h) in the case of the latent period of 15 h revealed that the mean for those variables was significantly higher for line 6 (Table 2).3.2. Regression PredictionsTable 3 shows multiple regression equations for the number of living embryos after 70 h of incubation of eggs obtained from females from line B and females from line 6, as well as the values of the coefficient of determination (R2). The data presented show that the predictability of this trait was more satisfactory for line B (R2 = 0.98).3.3. Results of Hormone Level Analysis3.3.1. LH Levels in Fish from Both Lines Sampled at 0, 12, and 24 h In each line of fish, the levels of LH increased significantly over time (Figure 1): in line 6 from 2.798 (0 h) to 200.3 ng mL−1 (24 h) (p ≤ 0.05) and in line B from 2.669 (0 h) to 242.7 ng mL−1 (24 h) (p ≤ 0.05). The comparison of LH concentrations between the lines within each sampling time did not reveal any significant differences. However, the mean LH level at 24 h sampling time in females from line B was 42.4 ng mL−1 higher than in fish from line 6.3.3.2. LH Levels in Ovulated or Non-Ovulated Fish from Both Lines Sampled at 0, 12, and 24 hThere were no statistically significant differences in LH concentrations between ovulated and non-ovulated fish within the lines and sampling times (Figure 2). However, at 12 and 24 h after the priming dose of Ovopel LH concentrations in ovulated fish from line 6 were lower than in non-ovulated ones, which stands in contrast to line B, where LH levels in ovulated females were higher than in non-ovulated fish. The same data presented over time (Figure 3A,B) showed that within each line the levels of LH increased significantly for both ovulated and non-ovulated fish: in ovulated females from line 6 from an average of 2.65 ng mL−1 before the first injection of Ovopel to 191.94 ng mL−1 at 24 h post-injection (p ≤ 0.05); in line B from 3.045 ng mL−1 before injection to 245.46 ng mL−1 post-injection (p ≤ 0.05) (Figure 3A). In females which did not ovulate (Figure 3B), LH levels also increased significantly over time: in fish from line 6 from 2.95 ng mL−1 (before treatment) to 209.78 ng mL−1 at 24 h post the first Ovopel injection (p ≤ 0.05). In non-ovulated females from line B, LH concentrations were significantly higher (p ≤ 0.05) at 24 h after the priming dose of Ovopel (an increase from 1.41 ng mL−1 at 0 h to 233.29 ng mL−1).3.3.3. 17α,20β-DHP Levels in Fish from Both Lines Sampled at 0, 12, and 24 hIn each line of fish, the levels of 17α,20β-DHP increased significantly over time (Figure 4): in fish from line 6 from 0.1918 ng mL−1 before the first injection of Ovopel (0 h) to 0.5481 ng mL−1 24 h later and in line B from 0.189 ng mL−1 to 1.359 ng mL−1 (p ≤ 0.05). The comparison of 17α,20β-DHP concentrations between the lines within each sampling time did not show statistically significant differences. However, the mean 17α,20β-DHP level at 24 h sampling time in females from line B was 0.811 ng mL−1 higher than in females from line 6 (Figure 4).3.3.4. 17α,20β-DHP Levels in Ovulated or Non-Ovulated Fish from Both Lines Sampled at 0, 12, and 24 hThere were no statistically significant differences in 17α,20β-DHP concentrations between ovulated and non-ovulated fish within the lines before the first injection of Ovopel and 12 h post-injection (Figure 5). At 24 h sampling time, the level of steroid concentrations in ovulated fish from line 6 (0.664 ng mL−1) was significantly higher than in non-ovulated ones (0.416 ng mL−1) (p ≤ 0.05). At the same sampling time in fish from line B, steroid levels were 1.639 ng mL−1 and 0.423 ng mL−1 in ovulated and non-ovulated females, respectively, but this difference was not statistically significant.The same data presented over time (Figure 6A,B) showed that within both lines of ovulated females, the levels of 17α,20β-DHP increased significantly: from 0.1958 ng mL−1 at 0 h to 0.6641 ng mL−1 at 24 h in fish from line 6 (p ≤ 0.05) and from 0.199 ng mL−1 to 1.639 ng mL−1 in fish from line B (p ≤ 0.05) (Figure 6A). In ovulated fish from line B, a statistically significant (p ≤ 0.05) difference between the levels of 17α,20β-DHP at sampling times 0 h and 12 h was noted. In non-ovulated fish (Figure 6B) from line 6, a significant (p ≤ 0.05) increase in the steroid level was found at 24 h after the first injection of Ovopel as compared with the initial level (0 h). In fish from line B, there were no differences in steroid levels between sampling times.4. DiscussionBased on the results obtained in the present study, it can be noted that the reproduction effectiveness differed between the carp lines investigated. Better reproduction effects were demonstrated for line B, even though the weight and the total number of eggs released by females from line 6 were higher.The percentage of ovulated females from line B was higher as compared to all the fish of this origin undergoing hormonal ovulation induction. Furthermore, within this line, there were no such substantial losses during the egg incubation time as were observed in line 6. The number of living embryos (70 h) was similar for both lines (slightly above 500,000) but the predictability of this variable was more precise for line B. From the breeding practice perspective, it is important that the costs incurred for inducing ovulation and obtaining a specific number of living embryos should be significantly lower in the case of line B. The latent period observed for the two lines also differed. Females from both line B and line 6 released eggs at two time points, but both these latent periods (earlier and later) were shorter for line B than those for line 6. This indicated that females from the Lithuanian line matured faster after the administration of the resolving dose of Ovopel. Within line B, eggs were stripped from a larger percentage of ovulated females and the number of living embryos (70 h) was higher after the longer latent period (15 h) as compared to the corresponding values after the latent period shorter by 2 h. On the other hand, within line 6 the percentage of ovulated females and the number of living embryos (70 h) were higher in females which released eggs after the shorter latent period. Higher reproduction effectiveness in line B compared to line 6 was also noted in studies reported by Brzuska [24] and Cejko & Brzuska [26], not only after Ovopel application, but also after twofold hypophysation. To date, research conducted at the Gołysz Institute on the effectiveness of reproduction in carp from different breeding lines has not taken into account the profile of hormonal changes in spawners during the ovulation induction period. Data about changes in the levels of such hormones as gonadotropins or steroids in response to the use of ovulation-stimulating agents are significant for knowledge-building reasons, but they can also have some importance in the context of reproduction outcomes in this important fish species.The subject literature includes numerous works demonstrating the effects of carp pituitary extract or GnRH-a on plasma gonadotropin release and/or on concentrations of gonadal steroids associated with oocyte maturation during spawning induction in the carp, e.g., [29,50,51,52,53,54,55]. However, the authors of the present study are not aware of any publications addressing the problem of the association between the levels of these hormones measured during ovulation induction in common carp and the effects of controlled reproduction characterized by numerous traits. It is worth noting that changes in GtH levels in parallel with the progress of oocyte maturation were reported in common carp treated with LHRH by Sokołowska [56], with LHRH or LHRH-a by Billard [51], with CCPE by Levavi–Zermonsky and Yaron [57], and in fish from this species treated with an sGnRH superactive analog combined with/without dopamine receptor antagonist metoclopramide or with CCPE by Drori et al. [58].The study described in the present paper demonstrated that the average LH concentrations in groups that included all females within the line and those observed after dividing the fish into ovulated and non-ovulated did not differ statistically significantly between the females from the lines investigated. The only difference was observed at 24 h sampling time (i.e., 12 h after the application of the resolving dose of Ovopel, which contained 18–20 µg of D-Ala6,Pro9NEt-mGnRH-a and 8–10 mg of metoclopramide), with LH levels in females from line B clearly, however, not statistically, higher than in fish from line 6. It is interesting to note that within line 6, the mean concentrations of LH in samples collected both 12 h and 24 h after the application of the priming dose of Ovopel were higher in non-ovulated fish, whereas within line B higher LH concentrations at both these time points were observed in ovulated fish.Billard et al. [51] reported the plasma GtH profile in ovulated and non-ovulated carp females receiving two injections of LHRH-A (5 µg and 50 µg of female’s BW). These authors note that the plasma GtH levels in ovulated fish ranged from 20 to 58 ng mL−1 between 6 h and 18 h after the first injection of LHRH-A, but the mean levels in the ovulated females were not significantly different from the levels in non-ovulated fish treated with LHRH-A.Our study did not demonstrate a significant difference in the basal levels of LH between the lines. However, these levels were low for both lines (exceeding 2.5 ng mL−1, but not reaching 3 ng mL−1) when compared with corresponding data from experiments on carp conducted by other authors, e.g., [54,55,58,59].Comparing classical hypophysation with LHRH application in females of two strains of carp (strain Z and strain S), Weil et al. [60] found that even though the GtH level prior to the first injection of CPE was significantly different for these fish strains (Z—6.36 and S—3.24 ng mL−1), the sensitivity of fish of different provenance to hypophysation was similar—the same pattern of changes in the GtH level was observed for each of the lines.The analysis of the LH levels during our experiment showed a similar hormonal pattern of LH secretion for both lines and for ovulated and non-ovulated females. At 12 h after the first injection, the rise in LH concentrations averaged 40 ng mL−1 and was not significantly different in relation to the pre-injection values. During the next 12 h after the first injection (24 h of the experiment), the average LH level was about 220 ng mL−1, which means an increase by an average of 180 ng mL−1. This increase was statistically significant in relation to the initial concentrations (levels measured before the first Ovopel administration) but not significant in comparison to the levels found 12 h later.Similar to the LH levels, the concentrations of 17α,20β-DHP did not show significant differences between the lines of fish when a comparison was made between ovulated and non-ovulated females. When comparing the levels of this steroid between ovulated and non-ovulated groups of both lines, a statistically significant difference was observed between ovulated and non-ovulated females but only within 24 h of the first Ovopel injection and only in line 6. The level of this steroid was significantly higher in the group of ovulated females.The analysis of 17α,20β-DHP levels during our experiment showed that 12 h after the first injection of Ovopel, the steroid level increased by an average of 0.33 ng mL−1 (a similar increase in both lines and similar for ovulated and non-ovulated fish). This increase was statistically significant compared to the baseline value only for line B in ovulated fish. At 24 h after the first Ovopel injection in ovulated females of both lines and non-ovulated ones from line 6, the steroid levels were significantly higher in comparison with its initial concentrations.Weil et al. [61] determined the level of 17α,20β-DHP depending on the occurrence or lack of ovulation in carp submitted to hypophysation and found that in non-ovulated fish the level of this steroid was in most cases undetectable (<0.5 ng mL−1). In partially ovulated females, there was a slow increase in the level of this hormone between 18 h and 21 h after the CPE injection, which was followed by a decrease in the level of this steroid after 24 h. In ovulated fish, the level of 17α,20β-DHP was high (7 ng mL−1), though not as high as that (111 ng mL−1) reported by Levavi-Zermonsky and Yaron [57]. In our study, 24 h after administering the priming dose of Ovopel, the level of 17α,20β-DHP remained significantly higher than the concentration of this hormone prior to the injection in ovulated fish from both lines and in non-ovulated fish from line 6. A somewhat different profile of this steroid was observed by Peter et al. [62] in a study on goldfish. These authors demonstrated a significant increase in the levels of both gonadotropin and 17α,20β-DHP as soon as 6 h after the application of a GnRH-a and pimozide. After a further 14 h, the level of gonadotropin remained significantly higher as compared to the control, whereas the level of 17α,20β-DHP decreased dramatically, down to the concentrations observed in the control group. In their experiment involving the application of sGnRH-a with metoclopramide to female carp, Drori et al. [58] reported that the level of 17α,20β-DHP started increasing 7 h after the injection of this analog, reaching 23.9 ng mL−1 after another 4 h. Twenty-six hours after administering sGnRH-a + metoclopramide, the level of this steroid decreased down to the baseline value.Based on the results of our research, it may be concluded that the profiles of the two hormones investigated in the blood serum of females from line 6 and line B sampled before and after Ovopel administration generally do not deviate from the results obtained by other authors in similar experiments on carp [34,60]. Nevertheless, there are some differences: in our experiment, 24 h after the application of the priming dose of Ovopel, the levels of both LH and 17α,20β-DHP did not show a downward trend, as was the case in the above-cited studies. Furthermore, the levels of 17α,20β-DHP determined in our study are relatively low, with the highest values not exceeding 2 ng mL−1. It seems justified to mention that when analyzing the data obtained in the course of the study, an association emerged between the levels of LH and 17α,20β-DHP during ovulation induction and selected features describing the effectiveness of reproduction in the breeding lines studied. Our data on the concentration of these hormones 12 h after the application of the resolving dose of Ovopel show that the levels of both hormones were higher for line B. Eggs were obtained from a higher percentage of females from this line and their quality was considerably higher. The quality of eggs, measure as the ability to be fertilized and subsequently develop into a normal embryo is the most important issue in aquaculture [63,64]. Further research, also at the molecular level (the eventual alterations in the transcript of reproductive-related genes), is needed to explain the observed differences in reproductive effectiveness between studied strains of common carp.5. ConclusionsThe results of comparing reproduction effects in two breeding lines of carp, i.e., line 6 and line B, revealed higher reproduction effectiveness in B. The differences in LH and 17α,20β-DHP serum levels in these lines were not statistically significant, either in samples collected just before administering the priming dose of Ovopel, 12 h after administering the priming dose (i.e., at resolving dose application) or 12 h after administering the resolving dose. However, the results obtained clearly indicate that concentrations of each of these hormones 12 h after the application of the resolving dose of Ovopel were higher in fish from that line which displayed higher reproduction effectiveness. Further studies are needed, e.g., measured by changes in mRNA transcript abundance of genes essential for reproduction, to explain these differences in reproductive effectiveness between female carp of different provenance.

animals : an open access journal from mdpi

10.3390/ani11041085

PMC8068923

We pose based on a fundamental science examination that events that occur around the time of slaughter have the potential to intensify the pain response, through a process called sensitisation, or an exaggerated response to painful stimuli. Health conditions which result in inflammation, injuries arising from transport and handling and exaggerated fear responses may all be present at the slaughterhouse. Whilst there is limited evidence of a direct effect of these on the processes of sensitisation in animals at slaughter, by analogy with the human neurobiology literature the connection seems plausible. In this review we outline the biology of such a response, and the rationale for suggestion of a possible linkage between events at slaughter and a heightened animal pain response.

We pose, based on a neurobiological examination, that events that occur around the time of slaughter have the potential to intensify the pain response, through the processes of sensitisation and enhanced transmission. Sensitisation, or an enhanced response to painful stimuli, is a well-discussed phenomenon in the human medical literature, which can arise from previous injury to an area, inflammatory reactions, or previous overstimulation of the stress axes. A number of events that occur prior to arrival at, or in the slaughterhouse, may lead to presence of these factors. This includes previous on-farm pathology, injuries arising from transport and handling and lack of habituation to humans. Whilst there is limited evidence of a direct effect of these on the processes of sensitisation in animals at slaughter, by analogy with the human neurobiology literature the connection seems plausible. In this review a neurobiological approach is taken to discuss this hypothesis in the light of basic science, and extrapolations from existing literature on the slaughter of ruminants. To confirm the postulated link between events at slaughter, and processes of hypersensitisation, further dedicated study is required.

1. IntroductionRecently the International Association for the Study of Pain (IASP) revised the definition of pain to “an unpleasant sensory and emotional experience associated with or resembling that associated with, actual or potential tissue damage” [1]. Pain experienced by animals at abattoirs has long been a topic of discussion and may arise at multiple stages of the slaughter process through transport, lairaging as well as slaughter itself. In ruminants the common method of causing death is through sticking. This involves the cutting of the soft tissues of the neck, carotid arteries and jugular veins for bleeding. It is imperative that all major blood vessels are severed to ensure time to death is reduced. The sticking process causes significant nociceptor stimulation [2], in addition to causing distress through method of restraint, as well as through aspiration of blood [3,4]. Other factors associated with pain prior to death include the use of electric drivers or prods, hot-iron branding and trauma caused by sticks or tubes that can inflict painful lesions and produce over-stimulation of nociceptors [5,6,7]. These events can also result in central and peripheral sensitisation that causes animals to perceive pain more intensely [5,6,7].Antemortem stunning is designed to minimise pain and distress by rendering animals insensitive prior to sticking. Stunning techniques prevent correct functioning of neurons in the brain leading to a state of unconsciousness which may be reversible or irreversible [8]. However, whilst the purpose of stunning is to reduce welfare impact, stunning methods have also been a source of controversy with regard to impact on welfare [9]. Mechanical methods like penetrating captive bolt guns, are prone to technical problems that can affect the quality of stunning [10], for example air pressure in pneumatic stunning devices must be sufficient to propel the bolt at an appropriate velocity to cause enough damage to the brain to cause insensibility.Therefore, the aim of this article is to review the neurobiology of antemortem pain in ruminants, the procedures that cause pain during pre-slaughter handling and those that can mitigate it, and the consequences for the welfare of the animal. A particular focus of the review is on how events occurring at slaughter could impact on hyperalgesia or sensitisation, and the consequences of this for the welfare of the animal. The concerns of an inadequate desensitisation and slaughter without stunning are also discussed.2. ConsciousnessA key consideration in relation to pain at slaughter is the point at which consciousness is lost since this loss prevents the animal experiencing pain and fear [8]. Two events cause this unconsciousness: stunning and/or hypovolemic shock caused by blood loss from sticking.Consciousness is generally considered to consist of two components; the content, represented by awareness of the environment and oneself, and the level (wakefulness) [11]. The former requires functioning of the cerebral cortex and connecting networks to subcortical structures such as the thalamus [11]. In the context of slaughter, perception of the environment is particularly relevant since information from the senses may all contribute to a fear response, for example smell of blood or fear pheromones, stress vocalisations from other animals and visual threats. Due to the lack of specific studies any relationship between these factors and any effect on stunning efficiency is only speculative. The wakefulness component depends on the ascending reticular system in the upper brainstem tegmentum and thalamus [11]. Projections in this system travel to the cortex and enable cortical functioning. Therefore, lesions to large areas of the cerebral cortex, the reticular formation itself or the ascending projections, created during stunning or slaughter lead to unconsciousness [8]. As a consequence, it is these brain regions that are targeted by various stunning methods [12,13].The processes of nociception described below are ’pre-conscious’. However, pain perception or the emotional experience of pain is described as a ‘conscious sensation’ since it requires a functioning cerebral cortex [14]. Therefore, it is generally assumed that an unconscious state implies that no pain is ‘felt’ [8]. There is however some contrary evidence from human studies suggesting that a low level of residual pain perception may exist in an unconscious state [15].3. The Neurobiology of Pain The nociceptive pathways and neurophysiological mechanisms associated with pain are similar across mammalian species since the neuronal processes involved in the stages of conduction, modulation and integration develop through a chain of neurons with three pathways [16,17,18]. The following provides a brief overview of the important features of the pain pathway and its modulation as it relates to events at slaughter. Figure 1 summarises the pain pathway described in the following sections.3.1. Transduction and TransmissionTransduction is the first element in the pain pathway. This refers to the transformation of harmful stimuli, such as those of thermal, chemical, or mechanical origin into electrical impulses by nociceptors [6]. Nociceptors are sensory neurons found in the skin, muscles, bones and viscera [5]. During sticking or handling the animal with sharp tools or electrical devices (goading), tissue damage occurs creating a harmful stimulus and consequent nociceptor activation. Opening of ionic Ca2+, K+ or Na2+ channels follow to produce electrical impulses that transmit the nociceptive signal along the neuronal axons to the spinal cord, brainstem, thalamus and cortex [19]. Nociceptors can be differentiated based on their expression of a series of channels that confer a type of sensitivity to various stimuli, such as heat or cold, and acid or chemical irritants. The channels most often observed are transient receptor potential vanilloid channel (TRPV1), transient receptor potential melastatin (TRPM8) and TRPA1. These distinct classes of nociceptors are associated with specific functions and stimuli, although variation in pain intensity may also depend on the tissues and nerve fibers involved. The nociceptors activated at this stage are called mechanoreceptors or mechanonociceptors that are activated by touch and pressure. They consist primarily of myelinated Aδ or type C unmyelinated fibers [5]. Usually silenced, once nociceptors reach their threshold of activation, they will respond in proportion to the pressure applied. Transmission is the next stage in pain signalling. This occurs when the electrical signal generated by the nociceptors is conducted along the axons of first-order neurons towards synapses with second-order neurons. Second-order neurons form synapses in the dorsal horn of the spinal cord using the so-called “Lissauer tract”, which consists mainly of a series of propriospinal fibres in which Aδ and C fibres predominate. The Aδ nerve fibres described as non-nociceptive participate in nociceptive phenomena when there is sensitisation (see later) [20]. Transmission of the nervous impulse occurs along high-velocity, myelinated conduction fibres (12–30 m/sec) which are associated with “first pain”, also described as throbbing pain. The signals from the silent, polymodal nociceptors transported by demyelinated, slow conductivity C fibres (0.5–2 m/sec) are responsible for “second pain”, which is characterised as penetrating, visceral pain that has been described as “burning”. A third class of Aβ nerve fibres, such as those of touch (50 m/sec), may be activated at low stimulus thresholds, but contribute to transmitting pain signals only in cases of peripheral sensitisation [20].3.2. HyperalgesiaSince nociceptors are heterogeneous and respond to multiple types of stimuli, it is possible for a stimulus acting via one modality, such as pressure, to alter the response of the nociceptor to other modalities, such as temperature [21]. Furthermore, tissue injury can cause an increased sensitivity of the afferent nerves to stimulation, so-called peripheral sensitisation. This “primary hyperalgesia” results from the production of neuropeptides by the nociceptors and upregulation of existing local receptors rendering them more responsive. In contrast, central sensitisation may occur in tissue adjacent to the injured area. Increased responsiveness in this area implies that this hypersensitivity arises due to a central mechanism acting via the dorsal horn and this is termed secondary hyperalgesia [21]. Both primary and secondary hyperalgesia (Figure 2) are key elements of concern during slaughter due to the potential for a heightened sensitivity, or exaggerated response to pain, arising as a result of events occurring at slaughter.Released glutamate bonds to α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors, generating post-synaptic excitatory potentials involving C fibre peripheral nociceptors that repetitively stimulate neurons in the dorsal horn. This results in mass depolarisation of neurons, the consequence of which is the release of neurotransmitters like substance P and brain-derived neurotrophic factor. These substances produce slow excitatory potentials. When substance P is released by high-threshold fibres, CGRP is simultaneously released and acts to extend the zone of the spinal cord capable of releasing CGRP. This phenomenon contributes to the excitability and activation of glutamate receptors like NMDA [22]. The activation of these receptors eliminates Mg2+, increasing intracellular Ca2+ and promoting production of prostanoids and nitrous oxide. This is accompanied by activation of wide dynamic range neurons (WDR) and hyper-excitability of the specific nociceptive neurons (NNSO). Excitation thresholds are lowered and two events may occur: (1) an exaggerated response to painful stimuli (hyperalgesia) or (2) pain perception after exposure to stimuli that are not normally painful (allodynia) [22].Lesions or trauma also stimulate a neuroendocrine reaction due to the intricate connections between the two systems. This reaction has three components: (1) the sympathetic-adrenomedullary (SAM) response mediated by catecholamines such as dopamine, noradrenaline and adrenaline, (2) Corticotropin-releasing hormone (CRH) activation of the hypothalamic-pituitary-adrenal (HPA) axis leading to production of glucocorticoids and (3) activation of the locus coeruleus and the noradrenergic limbic brain [2,23,24]. Limbic structures, including the hippocampus, amygdala and pre-frontal cortex have an inhibitory influence by resetting the HPA axis following a stressful event [25]. Activation of these systems does not directly initiate pain signalling but has an indirect influence via effects on neuroimmune cells. A key action is via mast cells which become activated and infiltrate peripheral injured tissues in response to CRH. This leads to enhanced peripheral nociceptor interaction, increased stimulation or ‘priming’, with subsequent increased ascendant pain signalling and reduced descendant inhibition. Sensitisation occurs as a result [25]. Tissue injury also leads to chemical release from non-neuronal cells, such as neutrophils, and from sensory terminals of primary afferent fibres. Mediators such as histamine, bradykinin, acetylcholine, serotonin and substance P are generated as part of this inflammatory process. Whilst these chemicals may not activate nociceptors directly, they are thought to enhance pain sensation by increasing rate of action potential firing and reducing the firing threshold in the nociceptive neurons [21,26]. Frequency of firing corresponds to pain intensity [27], hence increased firing enhances pain and mediates sensitisation. Meanwhile, serotonin released by mastocytes and platelets participates in the first phases of nociception and acute inflammation. This release, accompanied by activation of the sympathetic nervous system, triggers the release of noradrenaline that accelerates sensitisation of the nociceptors and the subsequent release of the CGRP-related peptide and substance P. These peptides generate degranulation of mastocytes, vasodilatation and oedema in tissues around the damaged area [22]. These events in turn lead to sensitisation of nearby sensory nerve endings as described previously [25]. It is therefore important to control these substances in order to reduce pain perception [28,29]. 3.3. Modulation and ProjectionModulation, or integration, occurs when the stimulus reaches the dorsal horn of the spinal cord. This process attenuates the transmission in a phenomenon that can be produced at distinct levels. Transmission continues along the ascending (spinal-encephalic) pathways of the spinal cord to the neuronal relay centres that are involved in certain aspects of the experience of pain and stress, including the rostro-ventral medulla nucleus (pain, temperature information), bridge nucleus (pressure) and mesencephalic nucleus (proprioception). These superior centres are responsible for creating aspects of the pain experience through their action on a feedback circuit called the descending pain modulatory system [20]. During modulation, excitatory and inhibitory mechanisms alter the transmission of the nervous impulse [20]. Sensory information is then transported to the brain through projection neurons that begin in the laminae of the dorsal horn. Two supraspinal structures called the spinothalamic and spinoreticular tracts are of key importance in projection to higher brain regions [20]. 3.4. PerceptionThe final stage is perception, which refers to the processing and integration of the information from several areas of the brain, including the cerebral cortex, to define the sensory characteristics of the painful stimulus: namely origin, localisation and type of stimulus [20]. It is important to consider that pain is multi-dimensional, with not only nociceptive and nocifensive (motor action) components, but cognitive and emotional components [30]. ‘Pain’ messages being relayed up the spinal cord are received at the thalamic nuclei. Pain information is then sent to cortical and subcortical regions, including the amygdala, hypothalamus, periaqueductal grey and areas of the cerebral cortex. The subjective experience or perception of pain is likely associated with activation of these cortical regions, but the experience can be modulated by these thalamic and limbic system connections [31]. For example, the amygdala has an important role in the emotional-affective component [30] but can also influence cognitive aspects such as decision-making in relation to pain [5,32,33,34]. The pain perception stage is of particular interest due to the effect that central sensitisation particularly has on modifying the response of the nervous fibres. This response is characterised by cascades of excitatory neurotransmitters such as glutamate, aspartate, catecholamines, prostaglandins, prostacyclins, bradykinin, substance P, interleukins and leukotrienes which are produced within the second order neurone. These substances diffuse back to the primary afferent fibre and generate hyperalgesic or even allodynic responses through actions at NMDA, AMPA and kainate receptors [5]. Hyperalgesia triggers stimulation of brain regions such as the cerebral cortex and reticular formation, to cause heightened pain perception [6,8,23,35]. A priority for mitigating pain around the time of slaughter is therefore to control the events and factors that lead to this ramping up of the pain response. The consequences of this are prolonged pain, and time to death, due to the perception of high-intensity harmful stimuli. The following sections specifically consider the impact of slaughter and associated events on the pain pathway.4. Hyperalgesia and Events at Slaughter 4.1. Hyperalgesia and Transport and HandlingDuring loading, transport and unloading, animals may be injured or exacerbate existing lesions resulting in inflammation [36]. Evidence for this as a potential concern in the context of slaughter comes from a number of studies. Herzberg et al., 2020 [37] reported elevated concentrations of cytokines including tumour necrosis factor (TNF-α), IL-1, IL-13, IFN-α and IFN-γ in Holstein-Friesian dairy cows with lameness compared to sound cows. Furthermore, it has been shown that the presence of previous pathologies or chronic injuries, such as metritis or pneumonia (associated with cytokine release), modifies the action potential of the nerve endings and triggers harmful responses of higher intensity [38]. However, in a recent study [39], the health status of 237 cull cows was evaluated at slaughter, along with specific slaughter-focussed welfare assessment indices. An indirect relationship between health status and carcass bruising was demonstrated; mammary issues and lameness predisposed to low body condition with the latter being a risk factor for bruising. Bruising might be assumed to be associated with inflammatory cytokine production and therefore lead to peripheral sensitisation, affecting behaviour at stunning. A potential outcome of this could be ineffective stun or need for second stun due to movement. However, haematoma number and severity was not associated statistically with welfare-specific variables recorded during the stunning process, such as requirement for a second shot, or increased time between stun and stick. Therefore, further large-scale epidemiological research is needed to elucidate the actual effects of inflammation on pain at slaughter.The conditions under which animals are transported to the abattoir can also affect their behaviour often manifesting in dominant aggressive behaviours [40]. This predisposes to skin lesions such as haematomas and bruising. A study of water buffaloes transported under conditions of acute nociceptive stress found a total of 244 contusions or concussions, distributed in the following categories of bruises: 9.8% small, 59% small but deep, 19.3% medium, 6.1% medium but deep and 5.7% large. The areas of the body most often affected were the hind limbs, abdomen, shoulders, neck, back and perineum [41]. Similarly, a study by Ahsan et al. (2014) [42] found that 89% of cows and buffaloes presented at slaughter with lesions on the back, in the ventral area, at the base of the tail, and in the lumbar, thoracic and scapular regions. These were determined to be caused by other animals and/or contact with the transport vehicle and were prevalent when animals were forced to travel at densities much higher than recommended levels. These findings are of concern since such injuries will lead to nociceptor stimulation in the skin, and the underlying tissues depending on the depth of injury potentially leading to peripheral and central sensitisation, as described earlier. Risks of enhanced sensitisation are exacerbated by prolonged trip length due to the increased likelihood of animals being injured by others, by contact with the vehicle, or by movement.Another important consideration in relation to injury and sensitisation, as the central topic of this paper, is that location of injury and degree of trauma influences the sensitisation response (Figure 3). A grade 1 trauma which only damages the skin stimulates mechanoreceptors and thermoreceptors without producing a peripheral sensitisation response or hyperalgesia. However, in cases of grade 2 trauma where damage to muscle tissue occurs, polymodal receptors are stimulated in addition to nociceptors. This leads to a modified transduction and transmission process, whereby a mechanical stimulus causes a chemical response, mediated by chemoreceptors sensitive to the release of excitatory neurotransmitters [5,39,43]. More severe or grade 3 trauma, characterised by bone damage typically caused during transport or pre-slaughter handling, leads to both peripheral and central sensitisation of the nociceptive arch. Stress and blows might not end with transport. Animals may be struck by the guillotine door at the entrance to the stunning box and can be subject to poor handling practices and inadequate behaviour of some workers [44], including shouting, having their tails twisted to encourage movement, receiving blows to thoracic limbs and hindquarters [45], and excessive use of electric drivers [46]. Previous studies have identified that the most frequent animal behaviours in the stunning box are struggling (38.3%) and falling (9.5%) [47]. Since, these reactions can lead to pain an important recommendation is to perform desensitisation within the first 5 s after restraint is applied [47]. Animals that receive more positive human contact on farm during their early life have been shown to perform fewer resistance movements in the stunning box, as well as exhibiting greater forward movement [48]. In this regard, animal fear reactions to humans are a significant contributor to poor welfare at slaughter [49]. Another consideration in regard to early life experience is that experiences during development [25], such as chronic poor welfare or mistreatment, may lead to altered functioning of the HPA axis, affecting both its regulation and output. [25] This dysregulation leads to both central and peripheral sensitisation as described earlier. Therefore, in improving welfare at slaughter, consideration should not just focus on the abattoir environment but prior on-farm environment and the nature of human-animal interactions created there.The OIE prohibits the use of electrical instruments to move small pigs, sheep and horses [48]. In cases where they are required, the OIE only allows battery-powered instruments with a maximum voltage of 30 V [50]. Drive use typically occurs when animals refuse to move even though there is space for them to do so, when they lie on the floor, obstruct the circulation of other animals, or when they fall during transport and must be made to stand up to prevent injuries caused by trampling [51]. Intense nociceptive stimulation may result from the application of electric prods on certain areas of the body, such as an animal’s face or genitals [47,52]. The occurrence of this event likely differs in frequency based on geographic region and local factors. Tissue damage generally correlates with time in contact with the drive, based on exposure time to electrical current. This contact time will influence the nature of the sensitisation processes occurring through the mechanisms discussed earlier, and dependent on location of injury. Electricity can also cause electrothermal wounds at the point of contact and in internal organs [53]. 4.2. Hyperalgesia and StunningIt has been suggested that following sticking in cattle, consciousness may last for 60 s or more [54]. During this period there is potential for animals to experience both pain and distress due to transection of a series of nociceptive fibers, that when activated generate a stream of sensory impulses that make pain perception inevitable [2]. In most jurisdictions, regulatory and guidance documents require that steps are put in place to mitigate the experience of pain associated with sticking. This is usually done by requiring pre-stick stunning [55,56,57]. The stunning process, or desensitisation, when performed correctly, renders animals unconsciousness until death supervenes. During this period of unconsciousness nociceptors continue to transduce harmful stimuli, but the brain can neither receive the stimulus nor perceive the pain generated because ascendent nervous transmission is inhibited [58,59]. An inadequately applied stunning method or rapid return of sensitivity can trigger a repetitive stimulation of the nociceptors due to cellular damage or extensive inflammation that lowers the threshold required for their activation. During this harmful phenomenon, poorly-stunned animals may develop primary hyperalgesia characterised by an exaggerated, prolonged response to a harmful stimulus that impacts the periphery. This is visualised clinically by an increase in the perception of, and sensitivity to, pain in the injured zone. Secondary hyperalgesia development may follow [22]. For this reason, is important to check for signs of recovery of consciousness in stunned animals and to perform a re-stun immediately when these are found. A significant number of cross-species studies have examined the causes and impact of inefficient stunning on animal welfare. It is beyond the scope of this review to examine all of this literature and the reader is referred to several excellent reviews on the topic [60,61,62]. Some pertinent points are presented here.Gibson et al. (2015) [22] assessed the mechanical factors that impact stunning efficacy with a penetrating captive bolt gun using goat and cattle manikins. Kinetic energy values needed for effective stun are higher as animal size increases and ability to achieve this varies with make of captive bolt gun used. They also determined that repeated firing caused heating to 88.8ºC, which reduced stun effectiveness by decreasing the depth of penetration. This evidence highlights the need to assess the following aspects in consideration of stun efficacy: the amount of kinetic energy delivered to the animal’s head, depth of penetration, species and relevant anatomical considerations, and the time between stunning and sticking. Gregory et al. (2007) [63] examined in greater detail the depth of penetration after captive bolt stunning. It was found that soft sound shots with 4.5-g cartridges were associated with shallow cerebral concussions that resulted in outward signs of inefficient stun such as failure to achieve physical collapse, presence of the corneal reflex, and post-collapse vocalising and nystagmus (Figure 4). Absence of tongue protrusion was determined to be an unreliable indicator of the depth of concussion. Animal species anatomy should be considered in relation to selected stunning technique, since certain methods may be superior. Collins et al. (2017) [64] compared the soft tissue and bone damage caused by applying a penetrating vs. a non-penetrating captive bolt gun by testing them on the cadavers of 12 crossbred female goats. Effectiveness was evaluated by magnetic resonance and computed tomography to examine the effects of the shots on the bony and soft tissues of the head. Both methods produced fractures in the occipital and interparietal regions, with structural damage to the soft tissues adjacent to the shot site being apparent. This damage was prevalent in the cerebellum, mesencephalon, myelencephalon, diencephalon and occipital lobes, and was accompanied by haemorrhaging in the underlying soft tissues. Penetration must reach the thalamic region to produce unconsciousness or insensibility to pain [65]. This may not be achievable using commonly used bovine captive bolt gun in buffalo. Schwenk et al. (2016) [65] evaluated brain injury caused by the captive bolt method in water buffaloes, finding that the depth of penetration, measured by magnetic resonance, was not sufficient to reach the thalamus. Furthermore, sex may influence risk of ineffective stun with an increased risk of failing to induce motor paralysis being found in bulls, regardless of the actual level of deviation of bolt entry from the ideal position [66]. This finding is postulated to arise due to the relatively thicker bone mass on the forehead in bulls, which provides increased resistance to the kinetic energy delivered by the bolt [66]. Other studies have similarly reported an increased chance of ineffective stun in bulls [63,67]. However, in contrast Gouveia et al. 2009 reported the opposite with a higher proportion of unsuccessful stuns in females [68]. These findings may however have been biased due to differences in age distribution of animals in their examined population, with a relatively greater proportion of older females. Electrical stunning is designed to disrupt the normal rhythmic electrical activity in the brain produced by cerebral neurons [8]. Correct application leads to synchronous de- or hyperpolarization of neurons leading to an epileptiform seizure. This seizure activity originates in the interconnections between the thalamus and cortex, and the brainstem, and then extends out to other nearby structures [69]. Reversibility and duration of the seizure is dependent on the current passed and the brain areas affected. Akin to mechanical stunning, an advantage of electrical stunning is the possibility that unconsciousness might be achieved almost instantaneously. However, failures arise due to insufficient electric current reaching the brain as a result of inadequate equipment maintenance or poor electrode placement [70]. In the case of improper stunning, events related to sensitisation can occur in the absence of unconsciousness. For example, adrenaline is released in large quantities contributing to a modified pain pathway [71]. Furthermore, head–only or head-back stunning may, dependent on species, cause intense muscle contractions which can lead to bone fractures, tearing and bruising of the muscle tissues, as well as blood pressure increases which serves to increase bruising [72,73]. These events trigger peripheral sensitisation processes and subsequent secondary hyperalgesia, heightening pain responses until death supervenes or a second stun is received. 4.3. Hyperalgesia and its Relationship with Slaughter Performed without Prior StunningReligious slaughter requires a single cut to the neck severing the jugular veins and carotid arteries bilaterally, to exsanguinate the animal and induce loss of consciousness due to cerebral ischemia. This is done in accordance with rules laid down in the holy texts (the Koran or the Torah) [15]. Religious slaughter without stunning is allowed in many countries, but this may not be the only reason for slaughter without stunning. For example, emergency or backyard slaughter may occur without pre-stun [74].Issues of concern specific to slaughter without stun are: 1) the cutting of the neck, blood vessels and surrounding tissues, without pre-stun and 2) the immediate post-cutting period [75]. Pain and distress arise due to the incision itself, which activates various nociceptors, but also due to sustained activation of the SAM system due to hypotension caused by bleeding. This causes tachycardia, tachypnea, hyperthermia and a redistribution of the blood [2,23].Imlan et al. (2020) [2] evaluated the effects of the knife edge on biochemical parameters, plasma catecholamines and electroencephalographic responses (EEG) of Brahman crossbreed steers after neck-cutting. Their procedure entailed cutting the carotid supply routes, jugular veins, trachea and throat during Halal religious slaughter, where prior stunning was not used. A significant increase in post-slaughter adrenaline, glucose, kinase creatinine and lactate dehydrogenase levels occurred when a commercial knife was used compared to a knife sharpened and then tested by a method designed to ensure optimum sharpness. The EEGs showed significant increases in median frequency and total power when the commercial knife was used, indicating that animals experienced greater pain and stress under that condition. This confirms that neck-cutting leads to the harmful stimuli being perceived integrally in both the cerebral cortex and reticular formation [23]. In goats, electroencephalographic changes similarly demonstrate that sticking causes noxious stimulation, regardless of whether animals are conscious or unconscious at the time of stick [76]. Effective stunning, with quick disgorgement prevents pain perception occurring in the higher brain regions, evidenced experimentally by lack of EEG response [77]. Stunning may also prevent elevations in stress hormones [78], which can contribute to sensitisation via various mechanisms as described earlier. Bozzo et al. (2018) [79] compared the plasma cortisol and catecholamine concentrations of 60 Charolais male beef cattle slaughtered following either traditional method (with stunning prior to neck-cutting) or the Kosher method. They evaluated three stages: on- farm, post-transport and exsanguination, with the finding that whilst on farm and post-transport cortisol and catecholamine levels were low, during exsanguination the levels of both substances were 50% higher in the animals slaughtered by the Kosher method than in those stunned prior to slaughter. Alternately, no significant differences in catecholamines were found between goats either slaughtered conscious or after minimal anaesthesia [76]. Care also has to be taken in interpreting such findings since these changes may not necessarily equate with pain [78]. Pain represents a source of stress which will activate the physiological stress systems. However, stress at slaughter does not only arise from a physical cause, such as a lesion creating pain [8]. Stress may also arise from fear reactions created by mixing with unfamiliar animals and human contact, as well as those created by the unknown environment and events of slaughter [80]. During sticking the animal experiences prolonged nociceptive stress associated with the time taken to onset of unconsciousness. Moreover, blood may also be aspirated into the airways causing respiratory distress [3,10]. Gregory et al. (2009) [81] examined the airways of cattle slaughtered by religious slaughter and slaughter with prior stun using penetrating captive bolt. The animals which were not stunned continued to breathe during the first part of the bleeding process, in contrast to the ones that were stunned which decreased their rate and depth of breathing during exsanguination. In that study, 19% of the animals slaughtered by the Shechita method, 58% of those slaughtered by the Halal technique and 21% of those that were stunned had accumulations of blood in the trachea. In addition, 36%, 69% and 31%, respectively, presented blood in the bronchial tract, and 10%, 19% and 0% respectively, had fresh blood in the trachea. On the basis of these findings, the researchers concluded that slaughter without stunning caused respiratory distress [81]. The entrance of liquid –in this case blood– into the bronchial tract and alveolar space triggers a reaction to release pro-inflammatory cytokines, TNF-α and interleukins such as IL-1 and IL-10. This inflammatory soup may then contribute to sensitisation, although time course of this may be prolonged and therefore of less relevance in the context of effects at sticking. The literature is controversial on effects of blood aspiration and slaughter distress. While some authors claim that aspiring blood through the upper airways and lungs during slaughter causes intense suffering if the animal has not been stunned, others advance that no suffering occurs because the afferent signals activated by the presence of irritants in the lungs are regulated by neurons of the vagus nerve which have been severed by the cut [82]. However, in conscious animals with intact vagus nerves, the presence of liquid in the respiratory tract causes irritation of nociceptors in the airways, especially those of the glottis, larynx and trachea, to provoke coughing or activate reflexes to expel the foreign substance [83]. Therefore, risk of distress is increased if the vagus nerves are not completely transected. Level of the transection may play a role in distress caused by fluid entry to the respiratory tract. The laryngeal reflex is regulated through neuronal connections in the C2 cervical vertebra, but in the Halal and Shechita methods the neck is cut between C3 and C5 [83], so this reflex continues to be present and active in animals that are conscious due to lack of stunning. Cut level is then potentially relevant to suffering in these animals and presents an opportunity for method refinement to improve welfare. The time to collapse after slaughter is another consideration relevant in comparisons of pre-stun and non-stun methods. Gregory et al. (2010) [3] reported that 14% of non-stunned cattle collapsed but stood up again before collapsing definitively, whilst 8% took 60–75 s to collapse. These results are suggestive that slaughtering animals without stunning increases the risk for a prolonged period of pain and distress. Zulkifli et al. (2014) performed a comparative analysis of the effects of penetrative stunning, non-penetrative stunning and post-slaughter stunning on biochemistry and the EEG in cattle [84]. Plasma noradrenaline concentrations increased in all animals. In addition, post-slaughter plasma ACTH concentrations in the animals stunned by the captive bolt method, followed by Halal slaughter were significantly higher than in the other study groups, likely indicating a physiological response to stress. Based on the EEG results obtained, penetrative stunning maximised the possibility of desensitisation post-stunning, while the animals stunned after the Halal cut showed t increases in encephalographic activity consistent with the presence of harmful post-slaughter stimuli associated with sectioning and lesioning of tissues. The researchers concluded that penetrative stunning was the most reliable method for ensuring insensitivity by minimising transduction, transmission, projection and perception of pain.5. ConclusionsThe use of poor transport and handling methods prior to slaughter, as well as the application of unsuitable stunning methods, may result in processes of sensitisation –central or peripheral– which intensify pain perception. In addition, animals destined for slaughter may be suffering from previous pathologies, the sequelae of which could contribute to sensitisation. Even where suitable stunning methods are utilised, deficiencies in handling prior to application of an effective stun, poor equipment maintenance and/or inadequate personnel training could also result in nociceptor stimulation, and trigger processes of peripheral sensitisation and pain perception. Further dedicated study is needed to confirm the hypothesis of a link between events at slaughter and processes of hypersensitisation.

animals : an open access journal from mdpi

10.3390/ani11092744

PMC8464926

Crimes against companion animals are universal and represent a major problem in human/animal interaction. This study characterizes forensic cases received at the Laboratory of Pathology of the National Institute of Agrarian and Veterinary Investigation (Vairão, Portugal) since the enforcement of the Portuguese law that criminalizes the mistreatment and abandonment of companion animals. Based on the consult of 160 data files of forensic necropsies analyzed for this study, the suspicion of prior crime against companion animal was confirmed in 38 cases (24%). Most of the assaulted animals were medium-size (57%) male (58%) dogs (87%) of crossbreed (55%), whose death was related to blunt force trauma (31%), firearms (27%), poisoning (27%), and asphyxiation (15%). However, in cats, death was related to blunt force trauma only (100%). In Portugal, violence against animals is a reality and the complaint of these crimes is gradually increasing due to the population’s raising awareness about animal rights.

Animal crimes are a widespread phenomenon with serious implications for animal welfare, individual well-being and for society in general. These crimes are universal and represent a major problem in human/animal interaction. In Portugal, current law 69/2014 criminalizes the mistreatment and abandonment of companion animals. This study characterizes forensic cases received at the Laboratory of Pathology of the National Institute of Agrarian and Veterinary Investigation (Vairão) since the enforcement of the aforementioned legislation. A retrospective study was carried out based on the consult of 160 data files of forensic necropsies from 127 dogs and 33 cats. Necropsies confirmed prior crime suspicion in 38 cases (24%), from which 33 were dogs and five were cats. Among confirmed cases, most of assaulted animals were medium-size (57%), crossbreed (55%) male (58%) dogs (87%), which were the victims of blunt force trauma (31%), firearms (27%), poisoning (27%) and asphyxiation (15%). In cats, most of the assaulted animals were juvenile (60%) females (60%) of unknown breed (40%), which suffered blunt force trauma (100%) as the only cause of death. The present study shows that violence against animals is a reality, and complaints about these crimes are gradually increasing due to the population’s raising awareness about animal rights. Greater communication and coordination between clinicians, veterinary pathologists, and law enforcement officers are essential to validate and legally support these cases and subject them to trial.

1. IntroductionAnimals are endowed with conscience and remain the most vulnerable group of all sentient beings, as they totally depend on humans to survive. Animals need care in the same way as children do, but unlike them, they do not have a voice, and will never gain independence, remaining vulnerable throughout their lives and hence subject to abuse and cruelty [1].Animal crimes are a widespread phenomenon with serious implications for animal welfare, individual well-being, and for society in general [2]. Crimes against companion animals are universal and represent a major problem in human/animal interaction, since the main threat to animals arises from human behavior [3].In recent decades, scientific studies have proven that animal abuse and interpersonal violence can occur simultaneously. It is now recognized that crimes against animals and domestic violence can co-exist and that animal abuse can be indicative of intrafamiliar problems [4]. The concerns and fears that influence the society’s perception on animal suffering work as a cultural filter on people’s views about cruelty [5].Animal cruelty occurs within a social context, and it is the responsibility of the community members working in the field of animal health and safety to find resources to put an end to animal victimization, namely by promoting greater understanding of the human/animal relationship. Studies in this area also provide new perspectives on the subject of violence [6].Investigations and prosecutions of crimes against companion animals are common occurrences and attract widespread attention. Countless animal crimes are committed every day and anywhere but rarely by disturbed individuals. The highlight of these events by the media has increased awareness in the general population about animal abuse [7].Despite recent advances on this topic, the number and most frequent causes of animal abuse in Portugal are still unknown. The aim of the present study is to characterize canine and feline necropsy cases received at the Laboratory of Pathology of the National Institute of Agrarian and Veterinary Investigation (INIAV, Vairão, Portugal) since the implementation of the legislation criminalizing companion animal abuse in the country, namely between 29 August 2014 and 31 March 2020, and for which a suspicion of crime against companion animals already existed.2. Materials and Methods2.1. Sample Population and Cases SelectionA retrospective study was carried out based on the consult of the data files from the canine and feline forensic necropsies performed at the Laboratory of Pathology of the National Institute of Agrarian and Veterinary Investigation (INIAV, Vairão). The INIAV laboratories are national reference laboratories for the diagnosis of animal diseases and the Vairão subunit, which is located in the north of Portugal, mainly covers the needs of the northern coastal region despite being able to receive cases from the whole country. The laboratory performs a total of 500 necropsies per year, and its social area of intervention is mainly urban.The cases received between 29 August 2014 (date of enforcement of the Portuguese law that criminalizes the mistreatment and abandonment of companion animals) and 31 March 2020 were selected for this study and detailed information was extracted from them regarding (1) the victim: species, age, breed, sex, clinical history/suspicion, results of forensic necropsy, and complementary diagnostic tests such as toxicology research and X-rays; (2) the crime: country region of occurrence, public location or private property, and the type of weapon used (whenever applicable); and (3) gender of complainants.Both animal species were classified according to their age groups (as juvenile, adult, or senior), and dogs in particular were characterized according to their size (as small, medium, or large breeds), as proposed by Fred L. Metzger [8].2.2. Statistical AnalysisDescriptive statistical analysis was performed using GraphPad Prism software (version 5.04).3. ResultsIn the period under study, a total of 2981 post-mortem examinations were performed, among which 160 (5.4%) were classified as forensic necropsies and included 127 dogs (79.4%) (56 females, 69 males and two undetermined specimens due to advanced cadeveric decomposition) and 33 cats (20.6%) (14 females and 19 males).The forensic necropsies confirmed the prior crime suspicion in 38 cases (24.0%) (Table 1 and Table 2) (Figure 1).In the remaining 122 cases (76.0%), death did not occur violently: 63 (39.0%) deaths resulted from natural causes or pathological conditions and in 59 cases the necropsies led to inconclusive results (37.0%). In dogs, most cases lacking a final verdict after a forensic necropsy was performed were related to suspected poisoning, negligence, or abandonment (which are complex issues themselves and are difficult to prove without resorting to other types of evidence or eyewitnesses).Similar results were obtained in cats, for which approximately one third (36.0%) of suspected poisoning cases could not be confirmed.3.1. VictimAmong the cases in which post-mortem analysis was compatible with violent death, 33 were related to dogs (87.0%) and five to cats (13.0%) (Table 1 and Table 2).With regard to dogs, 18 animals were of crossbreed (55.0%), 15 presented a defined breed (45.0%), and only one specimen was considered as a potentially dangerous breed (XRottweiler). From these 33 dogs, 19 were male (58.0%), and 14 were female (42.0%). For the 23 dogs with known age group, nine were adults (39.0%), seven were juveniles (30.5%), and seven were seniors (30.5%); the youngest was a one-month old female drowning victim of undetermined breed; the eldest was a 10 year old crossbred male that was strangled. From the 33 dogs, it was possible to determine the size for 30 of them (91.0%); 17 were medium-sized (57.0%), nine were large-sized (30.0%), and four were small-sized (13.0%). Size could not be determined for three puppies due to their young age (1–2 months old). The average weight of dogs was 16 kg, which falls into the medium size category. For the 33 dogs whose cause of death was determined as violent, 10 were injured through blunt force trauma (31%); nine with firearms (27%); nine others through poisoning (27.0%); and five through asphyxiation (15.0%).For cats, three out of five were female (60.0%), and two were male (40.0%) with different breeds: one was a Common European (20%), one was a Siamese (20.0%), another was an XSiamese (20.0%), and the breed could not be determined of two of them (40.0%). Three cats were juvenile (60%), one was an adult (20%), and the age group of the last one could not be determined (20%). The youngest was a three-day old male feline and the eldest an 8 year old female XSiamese. The cause of death for the five cats was related to trauma.Amongst the animals victimized through blunt force trauma, three dogs (30.0%) and two cats (40.0%) suffered traumatic brain injuries (Figure 2A,B).Additionally, four dogs (40.0%) and one cat (20.0%) suffered motor vehicle or traffic road accidents presenting signs of polytrauma consisting of bone fractures and osteoarticular injuries, organs fractures (spleen, liver, and kidneys), internal hemorrhages, and subcutaneous and muscular bruises. As for the remaining cases of blunt force trauma, one dog (10.0%) also presented multiple fractures and hemorrhages (Figure 3A–D).Perforating ante-mortem lesions were also observed on the right forearm and trunk of a canine apparently sick and unable to move, which was found in a ditch. Due to its extreme weakness, the animal was dragged into this location with the help of a rug, which corroborates the suspicion of abuse and abandonment. Subsequently, the animal was euthanized by the veterinarian clinician in a high suffering state.In one dog (10.0%), death occurred due to closed chest trauma with pulmonary fractures and haemothorax, compatible with history of kicks allegedly perpetrated by an individual who later attacked another dog that survived with severe injuries and lost locomotor capacity.As for cats, two (40.0%) suffered thoraco-abdominal trauma with subcutaneous haemorrhagic infiltrations and pulmonary and hepatic tears, compatible with previous testimonies of fall or defenestration.In this investigation, injuries caused by firearms reached the thoraco-abdominal region in four cases (45.0%) and the thoracic region only in three cases (33.0%) (Figure 4A,B), causing, in both regions, internal bleeding, bone fractures, and injuries to various organs, namely to the heart, lungs, liver, spleen, and intestine. In one out of nine cases (11.0%), only the abdominal region was affected, with internal bleeding occurring through perforation of the abdominal aorta. The head region was also affected in one case (11.0%) with fracture of the temporal and parietal cranial bones, fracture near the atlanto-occipital joint, and brainstem and cervical cord laceration of the neuroparenchyma by lead spheres. The projectiles from these cases were collected for ballistic examination (Figure 4C,D) and five of these cases were also subjected to imaging examination, namely X-ray, before the necropsy (56.0%) (Figure 4E).In the present study, canine poisonings occurred due to carbamates (45%), followed by rodenticides (33.0%), organophosphates (11.0%), and cyanide (11.0%). In five out of nine of the poisoning cases, the macroscopic findings corresponded to generalized congestive-hemorrhagic conditions (56.0%) (Figure 5A,B).In two of these cases, the cause of death was not determined at necropsy (22.0%), and in the remaining two cases, the necropsy result was inconclusive due to advanced cadaveric decomposition (22.0%). However, in all of them, the toxicological tests confirmed poisoning.Regarding the 15.0% of asphyxiation cases, four were due to strangulation (80.0%) (Figure 6A,B). In one of these animals, necropsy findings were compatible with hanging through incomplete body suspension (Figure 6C,D). This animal exhibited congestion of the head area, exophthalmos, mark of the incomplete groove in the mandibular coat, hemorrhages in the soft tissues of the neck, larynx, and peritraqueal, blackened blood coloring, tracheobronchial foam, and pericardium hemorrhages.Another asphyxiated animal was found in a sewer pipe in the streets, presenting signs of drowning (20.0%). Macroscopically, the lungs were enlarged, inflated in aspect, crackling on palpation with multifocal to coalescent and miliary subpleural bright red lesions, which were more extensive at the periphery of the lung edges, with blood and foam exudation at cross section.3.2. Crime Geographical Location and ContextFrom the 38 confirmed cases, 30 occurred in the northern region of Portugal (79%) and eight in the central region (21%) (Figure 7).Among the total crimes committed with firearms, three occurred in a public location (33.5%), two inside a private property (22.0%), three during hunting activities (33.5%), and this information is unknown for the remaining cases (11%).3.3. Complainant and ComplaintsIn 22 out of 26 (85.0%) reports of suspected crimes against companion animals, it was possible to retrieve information about the gender of the complainant: 11 were men (50.0%) and 11 were women (50%).In 2014, there were no reports of these crimes, and in 2015, there was only one criminal complaint. However, a growing number of complaints was noticed from 2016, which culminated in 20 criminal complaints in 2019. Despite the increase in reported cases, the average number of violent death cases confirmed through forensic necropsy remained stable per year.4. DiscussionDuring the period under study, 160 suspected cases of crimes against companion animals were admitted to the Laboratory of Pathology of INIAV Vairão, but the forensic necropsies only confirmed the suspicion of violent death in less than a quarter.Similarly to previous studies, dogs are the most affected species, followed by cats [9]. The tendency to commit crimes against a specific animal species can be explained by several factors, namely the availability and behavior of the animal species, the social attitude towards them, and their physical characteristics determining adequacy for abuse. Cats, in comparison to dogs, are more independent, have more free access, and spend more time outside and, most often, do not have a known owner. On the other hand, some dogs are more empathetic and submissive animals, more dependent on human interaction, and maintain these behaviors even when they are the victims of abuse [10].In the present study, the small number of formalized cases against cats can be explained by the inherent behavior of the species; when cats are injured or frightened, they tend to isolate and hide themselves and avoid human interaction, making it difficult to be detected as victims of abuse. In addition, the owner may delay the search for his missing cat, thinking that it will eventually return or find a new home. Negative social attitudes towards cats can potentially influence public opinion making criminal acts against this species significantly underreported [11].Although the number of males is higher than females, no major differences were found in the present study. According to Intarapanich et al. (2016), male animals were more frequently the victims of non-accidental injuries and motor vehicle accidents; however, this difference was not statistically significant [10]. In another report, male dogs were probably more affected by violent acts, due to the fact that they can behave more aggressively and less controllably than females, or because aggressors tend to prefer male animals [12].The age distribution of animals suffering violent death follows a normal distribution, showing no evidence of a certain age group preference. However, several authors claim that animal victims tend to be young [10,12,13]. Young animals are more immature, restless, and difficult to control, and hence probably disturb and irritate their owners, triggering aggressive behaviors from them. Young animals are also more fragile and less able to defend themselves or escape [10,12,13]. Additionally, younger animals tend to explore new environments, being more susceptible to criminal acts by neighbors, for example [12].Most affected dogs are medium-sized, which are normally easier to control and manipulate by humans, followed by large-sized animals. Some studies show that large-sized dogs are more affected since they are more likely to be kept outside the house and remain more exposed to the observation of the general population who may denounce a suspicion of mistreatment [14,15].Blunt force injuries are frequent findings in veterinary forensics. Lesions in companion animals are comparable with those described in humans and crucial information can be obtained through macroscopic and microscopic examinations in order to provide evidence to court cases [16]. In the present study, blunt force trauma was found to be a major cause of violent death in dogs and cats. Violent death can arise from accidental or non-accidental causes and vary from: motor vehicle accidents; falls from heights; activity injuries; and injuries resulting from physical aggression. It is essential to examine the entire animal in order to ascertain the incident and events sequency; there may be signs of repetitive abuse, such as bruises, scars, or multiple fractures revealed at different stages of convalescence and indicative of occurrence at different time points [17].The importance of the injury depends on its anatomical location and on the size and nature of the traumatic pathological process. In the present study, 30% of dogs and 40% of cats presented traumatic brain injuries. Lesions affecting the CNS are more likely to culminate in death [18]. Animals wounded on public roads can be run over or suffer accidents by motor vehicles [13], and dogs seem to be more involved in these accidents than cats [10].In the present study, violent deaths provoked by projectiles victimized only dogs. Capak et al. (2016) reported that, annually, approximately 42 dogs and 28 cats suffer projectile injuries. This discrepancy can be explained by the fact that dogs are the most represented species in small animal veterinary clinics, emphasizing the clinical relevance of this type of lesions, especially in this species [19].In the present study, the majority of the cases associated with firearm occurred on the public road or inside a property and a minor number during hunting activities. Similar observations were noted by Capak et al. (2016), where hunting accidents represented only 13% of projectile injuries, concluding that such lesions are a relevant cause of trauma in animals that are not used for hunting [19].Poisonings are a serious cause of mortality in companion animals [20,21]. In the present study, cases of poisoning were found in dogs, which corroborates previous descriptions from Berny et al. (2010) in five European countries, namely in Belgium, France, Greece, Italy, and Spain. According to the Laboratory of Toxicology of Ghent University, dogs accounted for 20% of all cases of intoxication, followed by cats (11%). Similarly, in France, dogs account for 35% of the cases registered annually [22], and this value is consistent with the numbers recorded in Germany, Italy, Spain, and Austria [23,24,25,26,27]. In the USA, the Animal Poison Control Center (APCC) receives thousands of suspected cases annually. Between 2002 and 2010, 76% of the incidents received affected dogs and 13% cats [28], and this trend continued in following years as canids accounted for 65% and 63% of the cases, respectively, in 2016 and 2017 [29]. In Brazil, intoxications follow the same pattern, amounting to 86% in dogs and only 13% in cats [30]. This difference may be related to the fact that cats are more selective in their food since they tend to absolutely reject any food that presents a smell or taste that they do not find pleasant. Dogs are usually curious animals that eat and play with many things they find and have a voracious appetite [21,30,31].The most common substances involved in intoxications are insecticides, rodenticides, other pesticides, such as herbicides and fungicides, plants, human and veterinary drugs, metals, household products, toxins, and ingredients that comprise human food [22,26,28,29,32]. Most poisonings occurred through carbamates (45%), rodenticides (19–33%), and organophosphates (5–11%) [22,25,30]. According to Guzmán et al. (2002), the high rates of insecticide poisoning are due to their widespread use in agricultural practices [25]. In the present study, only one death occurred due to cyanide poisoning. This can be explained by the use of this product in the jewelry industry [33,34,35], since the animal in question came from an area of the country where this activity is prevalent.The main causes of poisoning in urban environments result from interpersonal conflicts. However, in rural areas, these cases arise in the course of human activities, such as hunting or agriculture [13,23,36,37]. The production of poisonous baits may not be difficult, since some of the agents used for this purpose are easily accessible on the market and the choice of a specific toxicant for intentional animal poisoning may arise from popular knowledge about its toxicity and commercial availability [38]. Intoxications depend on various available features in the animals’ environment: the tendency of exposure of the animal to the agent; the toxin amount and also the individual sensitivity of the animal to the toxic agent adverse effects [20]. Epidemiological studies are scarce, and the actual number of intoxication cases may be underestimated. This fact comes from the animals’ natural inclination to hide when they feel sick, making it difficult for their body to be recovered after death [23,25]. Another fundamental fact is the failure in clinical or analytical diagnosis, since tests are often not performed because there is no toxicological screening capable of detecting all known toxic agents. Thus, in addition to being expensive, tests can be unsuccessful [37]. These factors may have influenced the number of poisoning cases in the present study as the necropsy often leads to inconclusive diagnosis either due to cadaveric decomposition or to nonspecific congestive-hemorrhagic conditions and without toxicological tests being performed, the suspicion cannot be confirmed and cannot be excluded either.In the present study, 15% of deaths occurred due to asphyxiation and only dogs were the victims of this type of act [12]. It is plausible that an animal could die as a result of accidental hanging with a rope due to incomplete suspension [39]. It is known, however, that some people practice the execrable act of drowning new-born animals in order to get rid of them. The frequency of animal drowning is unknown, and the literature addressing this topic is sparse [40,41]. A drowning diagnosis in animals is even more complex than in humans due to the different anatomophysiological respiratory system specificities of the species, and it remains a challenge both for human and veterinary forensics [42]. There are no pathognomonic drowning necropsy findings; therefore, drowning remains an exclusion diagnosis based on the circumstances of death and nonspecific findings of the necropsy [43]. As in humans, in suspected cases, the presence of diatoms can be investigated [42], although this was not found necessary in the only case present in this study.The largest number of cases are located in the northern region of Portugal, but this does not necessarily mean that more criminal acts are committed there. This can simply indicate that the number of animals is higher, or that the population is more likely to report these cases in this geographical location. Moreover, the Vairão subunit of INIAV where this study took place is also based in this region. Therefore, most of the complaints were focused on northern coastal districts, which present marked rural and urban contrast. This geographical distinction, which only allowed us to study one location by country region, largely limits our interpretation of the real social context.No statistical differences were found between the complainants’ genders in this study. However, studies revealed a greater propensity of female individuals to report this type of crime, varying between 40% to 80%, while male individuals vary between 12% and 22% [9,14,15]. Previous studies demonstrated that men who witnessed acts of animal cruelty showed a more insensitive attitude, while women exhibited greater sensitivity and seemed to be more emotionally linked to companion animals [44,45].The annual evolution in the number of cases with a necropsy conclusion compatible with crime shows that before 2016 the number of these cases was inferior, at five per year. After that, this number approximately doubled. This may be related to the election of a specific Portuguese political party in 2015 with well-defined aims on animal welfare. The greater expression and verbalization of this political party in society may have given confidence to some citizens to report their suspicions of crimes against companion animals, which may also reflect the number of cases confirmed as a criminal act. This increase over the years does not necessarily mean that there is more violence against animals than in previous years. It can merely suggest that people have become more sensitive to the topic and therefore are more likely to report their suspicions.McEwen (2012) found that there is a gradual linear trend, which is consistent and statistically significant, in the submission of cases for forensic necropsy. Several reasons can justify this propensity, such as an increasing attention from the media, changes in the country’s justice system and legislation, the relationship between crimes against companion animals and domestic violence, as well as professional and public interest [46]. Even so, from all the suspected cases submitted in the period under analysis, approximately 37% proved to be inconclusive after the forensic necropsy, and the vast majority are related to the advanced decomposition state of the corpse, which makes it impossible to carry out most of the necessary tests, and to limitations related to the laboratory resources (as certain cases may require specific tests that are not included in the conventional routine of the official laboratory).The difficulty in producing clear evidence, due to lack of proper examination of the crime scene and scarce laboratory and economic resources, supported by the existence of unclear legislation that is difficult to interpret, mean that gathered facts are not sufficient to prove the crime or the author of the criminal act.5. ConclusionsThe present retrospective study has allowed the identification of suspected cases of crimes against companion animals by investigating the main causes of violent death through the characterization of animals/victims and to evaluate injuries found in forensic necropsies.Companion animals play an important role in the life of human beings. The reports of crimes against companion animals have clearly increased in recent years; however, this may not be directly indicative of a real increase in abuse acts. We must concede that the growing awareness of the society regarding animal rights translates into the civic duty of reporting abuses in order to guarantee animal protection. Still, we believe that there are many more crimes than those reported and only a small percentage reach the courts and are judged, and an even smaller percentage are condemned.In order to move safely towards an evolved and fair society, the training of professionals in this area is crucial. Law enforcement officers and veterinarians must have adequate training, share knowledge, and work together to obtain legally valid evidence. In Portugal, although greater sensitivity, interest, and media coverage of animal violence is noticeable, there is still a long way to go.

animals : an open access journal from mdpi

10.3390/ani13060955

PMC10044561

The rhinoceros iguana (Cyclura cornuta cornuta), like many other iguana species, has become severely endangered as a result of human activities, earning the highest level of protection under the CITES convention for protected species. Moreover, it has been classified as vulnerable by the International Union for Conservation of Nature (IUCN). The rhinoceros iguana belongs to the family Iguanidae and is native to the island of Hispaniola. Since its population has declined, the Canary Island Government has promoted an initiative for its recovery in a completely natural environment. The limited literature regarding the anatomy of this species motivated us to investigate its cranial structures by using modern imaging techniques such as computed tomography (CT) and magnetic resonance imaging (MRI) combined with anatomical dissection, to obtain helpful information on the structures that form the rhinoceros iguana’s head.

In this paper, we attempted to elaborate on an atlas of the head of the rhinoceros iguana, applying modern imaging techniques such as CT and MRI. Furthermore, by combining the images acquired through these techniques with macroscopic anatomical sections, we obtained an adequate description of the relevant structures that form the head of this species. This anatomical information could provide a valuable diagnostic tool for the clinical evaluation of different pathological processes in iguanas such as abscesses and osteodystrophy secondary to nutrient imbalances, skull malformations, fractures, and neoplasia.

1. IntroductionThe rhinoceros iguana (Cyclura cornuta cornuta), belonging to the Iguanidae family, is an endemic species of the island of Hispaniola, in the Caribbean Sea [1]. Its skin has rough epidermal scales, and its greyish-brown or olive coloring camouflages very well with the environment. The name comes from the bony protuberances or pseudo-horns dorsal to the snout. Males have larger body dimensions than females, and their horns are also larger. This iguana has a polygamous mating system and is oviparous (5–20 eggs per clutch), reaching maturity between 5 and 9 years of age [2]. This species is extremely territorial. To intimidate conspecifics or predators, these animals move their heads and neck, while rotating their bodies. These movements are also used in mating rituals. Like all iguanas, this species is heliothermic, which means it must organize its activities to use solar radiation to regulate its body temperature. This also means that it can only survive in tropical/subtropical climates. Its diet is herbivorous, although it may occasionally feed on insects, land crabs or carrion. Due to human encroachment and destruction of the environment, it has been severely displaced from its habitat (near the coast but with low rainfall). Limestone mining, predation, as well as deforestation by the wood industry and forest fires have forced this species to migrate or even become extinct, as is the case with the closely related subspecies Cyclura cornuta onchiopsis, native to Navassa Island [3].The great interspecific anatomical complexity between mammals and reptiles and the growing interest in reptiles as companion animals present a challenge to veterinary clinicians in diagnostic imaging studies interpretation [4]. Diagnostic imaging has brought a radical change in their clinical practice due to the facility to obtain information on the internal structures of the animal body [5]. Thanks to technological developments in this field, morphological and anatomical information can be obtained in a less invasive and considerably faster way, and this technique has become a fundamental tool for this practice, using conventional radiology [6] and ultrasound [7], as well as more advanced imaging techniques such as CT [8] or MRI [9], which offer certain advantages over conventional ones, such as overlapping structures avoidance, fast image acquisition and high contrast resolution, among others [10,11]. All this has meant that the veterinary clinician has to become acquainted not only with these techniques but also with the anatomy and physiology of reptiles [12].Some relevant literature on the anatomical, physiological and pathological study of these species is already widely available [8,9,10,11,12,13,14,15]. As far as we know, the anatomy of different reptile species has already been thoroughly described employing diagnostic imaging techniques [13,14,15]. Authors have reported atlases of green, loggerhead and leatherback sea turtles [15,16,17,18], komodo dragons [19,20] and green iguanas [12,15,21,22], snakes and lizards [21]. Concerning the rhinoceros iguana, most of the studies are focused on population genetics [23], evolution and historical biogeography [24], and on some pathologic descriptions such as osteopetrosis-like conditions [25]. However, to the authors’ knowledge, none have performed anatomic investigations on the head of the rhinoceros iguana. Therefore, this study aimed to describe the normal anatomy of the head of this species, using CT, MRI and those anatomic sections that are the more informative regions of gross anatomy and can help to understand all the structures that form the head of the rhinoceros iguana. The application of CT, MRI and macroscopic anatomical sections could provide essential information for anatomic descriptions in teaching and clinical practice.2. Materials and Methods2.1. AnimalsTwo carcasses of adult female rhinoceros iguanas (Cyclura cornuta cornuta) from the zoological park “Rancho Texas Lanzarote Park” (Lanzarote, Canary Islands, Spain) were collected. One female measured 94 cm, and the other was 91 cm long from snout to tail. They weighed 5 and 4.2 kg, respectively. The animals died due to natural causes. No abnormalities were found on physical examination.2.2. Anatomic EvaluationWe performed anatomical gross-sections to facilitate the identification of structures observed in the CT and MRI images. Immediately after the scanning procedures, these specimens were placed in a plastic isolation holder in ventral recumbency and successively stored in a freezer (−80 °C) until completely frozen. Subsequently, the two frozen carcasses were sectioned using an electric band saw to obtain sequential anatomical gross-sections. Contiguous 1 cm transverse slices were obtained starting at the snout and extending to the first cervical vertebra region. These slices were thicker than those for CT and MRI to preserve integrity and position of the anatomic structures in the sections. These sections were cleaned with water, numbered and photographed on the cranial and caudal surfaces.Afterwards, we selected those anatomic sections that better matched the CT and MRI images to identify the relevant structures of the rhinoceros iguana head. To help in this process, we also used anatomical texts and relevant references describing this species [4,21,26,27].2.3. CT TechniqueTransverse CT images were obtained at the Veterinary Hospital of Las Palmas de Gran Canaria University using a 16-slice helical CT scanner (Toshiba Astelion, Canon Medical System, Tokyo, Japan). The animals were placed in ventral recumbency as symmetrically as possible on the CT couch. A standard clinical protocol (100 kVp, 80 mA, 512 × 512 acquisition matrix, 1809 × 858 field of view, a spiral pitch factor of 0.94 and a gantry rotation of 1.5 s) was used to obtain sequential transverse CT images (1 mm thickness). The original transverse data were recorded and transferred to the CT workstation. No CT density or anatomic variations were detected in the head of the reptiles used in the investigation. In this study, we applied two CT windows by adjusting the window widths (WW) and window levels (WL) to appreciate the CT appearance of the head structures: a bone window setting (WW = 1500; WL = 300) and a soft tissue window setting (WW = 350; WL = 40). Moreover, dorsal and sagittal multiplanar reconstructed (MPR) images were also obtained to better visualize other iguana head structures.2.4. MRI TechniqueMRI images were obtained with a Canon Vantage Elan 1.5 T equipment, using T1W sequences in a transversal plane (TR: 634 ms, TE: 10, FOV: 1809 × 829, thickness 2 mm, matrix 192 × 160), T2W sequences in a transversal plane (TR: 4769 ms, TE: 120, FOV: 1809 × 829, thickness 2 mm, matrix 192 × 224), T2W sequences in the dorsal plane (TR: 5271 ms, TE: 120, FOV: 1809 × 829, thickness 2.5 mm, matrix 240 × 192), T2W sequences in the sagittal plane (TR: 4450 ms, TE: 120, FOV: 1809 × 829, thickness 2.9 mm, matrix 224 × 224); enhanced spin-echo sequences were performed in the dorsal, transverse and sagittal planes. The MRI images were obtained with a thickness of 2.7–3.5 mm.3. ResultsFigure 1 corresponds to a bone-window CT sagittal plane, in which each line and number (I–V) represents approximately the level of the following anatomical, CT and MRI transverse planes. Transversal sections revealing the relevant anatomical structures of the rhinoceros iguana head are presented (Figure 2, Figure 3, Figure 4, Figure 5 and Figure 6). Figure 2, Figure 3, Figure 4, Figure 5 and Figure 6 are composed of three images: (A) macroscopic, (B) bone window CT and (C) MRI. The images are presented in a rostrocaudal progression from the maxillary bone (Figure 2) to the brainstem levels (Figure 6). Figure 7 and Figure 8 are composed of two images: (A) bone-window CT and (B) MRI in T2W, in a dorsal and sagittal plane, respectively. Figure 9 is a T2W sequence sagittal image showing the angulation concerning the horizontal axis of the myelencephalon (40°).3.1. Anatomical SectionsDifferent structures belonging to the central nervous system were visualized by anatomical gross-sections. Therefore, we identified the brain (telencephalon) and the two telencephalic hemispheres (Figure 4A and Figure 5A) separated by the fissura longitudinalis cerebri (Figure 5A), the diencephalon (thalamus) (Figure 4A and Figure 5A), the dorsal part of the mesencephalon with the two caudal colliculus (Figure 6A) and the ventral part of the cerebellum with the nodule, covering part of the fourth ventricle, as well as the ventral surface of the brainstem (myelencephalon) (Figure 6A). These sections were also helpful for the observation of the olfactory bulb that showed an extracranial location, ventral to the frontal bones, and coursed between the two eyeballs (Figure 3). In addition, these sections allowed the description of structures of the eyeball, identifying the cornea, the sclera, the retina, the vitreous humor, the lens, as well as associated structures such as the interorbital septum and the orbital sinus (Figure 3A and Figure 4A). In addition, these transverse sections allowed the identification of structures belonging to the oral cavity such as the tongue (Figure 2A) and other structures such as the larynx, with the corniculate tubercles of the arytenoid cartilage, the thyroid cartilage, and the laryngeal ventricle (Figure 3A), as well as of different structures of the nasal cavity such as the nasal glands that filled almost the entire cavity, the conchal grooves and the stammteil located laterally to the nasal septum (Figure 2A). Adjacent structures such as the trachea and the nasopharynx were also well identified (Figure 4A, Figure 5A and Figure 6A). Furthermore, most bony structures that form the neurocranium were observed, such as the pterygoid, frontal, postfrontal-postorbital, parietal, supraoccipital, basioccipital, exoccipital, otoccipital, parabasisphenoid and sphenoid bones (Figure 3A, Figure 4A, Figure 5A and Figure 6A), as well as those that form the splanchnocranium such as the nasal, vomer, palatine and maxillary bones (Figure 2A and Figure 3A) and also the medial horn located dorsal to the nasal bone (Figure 2A), the mandible, with the dentary bone (Figure 3A), and the hyoid apparatus, visualizing the central body of the hyoid arch, between the lateral branches of the dentary bone (Figure 3A). Rostromedially to the two dentary bones, we identified different muscles related to the hyoid apparatus such as the musculus genihyioideus, hyoglossus and intermandibularis (Figure 2A and Figure 3A). In the following sections, we also observed the muscle groups corresponding to the mandibular musculature (we were not able to dissect them, and therefore they were treated as a group), including the pterygoideus, omohyoideus, sternohyoideus, ceratohyoideus, adductor mandibulae externus medialis and its homonym superficialis (Figure 4A and Figure 5A).3.2. Computed Tomography (CT)Regarding the neurocranium, the CT images allowed us to distinguish bone structures such as the prefrontal, frontal, postfrontal-postorbital, parietal, squamosal, quadrate, epipterygoid, pterygoid, basioccipital, exoccipital, otoccipital and parabasisphenoid bones (Figure 3B, Figure 4B, Figure 5B and Figure 6B); related to the splanchnocranium, we observed the nasal, premaxilla, maxilla, septomaxilla, vomer, jugal and palatine bones, and the nasal septum (Figure 2B, Figure 3B, Figure 4B, Figure 7B and Figure 8B); the mandible structures such as the dentary, angular, surangular, coronoid and articular bones (Figure 2B, Figure 3B, Figure 4B, Figure 5B and Figure 6B) and the hyoid apparatus (Figure 2B, Figure 3B, Figure 4B, Figure 5B and Figure 6B). CT scanning and post-processing transverse images showed the relation between the different bones that form the head of the rhinoceros iguana, the junction of the nasal and the prefrontal bone, as well as that of the parietal bone with the postfrontal-postorbital bone (Figure 2B and Figure 4B), and the palatine and quadrate processes of pterygoid bone (Figure 5B and Figure 6B respectively). The prominent medial horn was identified dorsal to the nasal bone with soft-tissue attenuation and a thin, lamellar-shaped mineral structure bordering it regularly on its most external aspect (Figure 2B).Concerning the nasal cavity, the transverse CT image showed the nasal glands as symmetrical bilateral structures, with regular and well-defined margins, located on both sides of the nasal cavity and with soft tissue attenuation (Figure 2B). Moreover, those structures with intraluminal gas content such as the nasal conchal recess (Figure 2B), oral cavity, nasopharyngeal duct, trachea, adductor fossa (Figure 4B) and the otic cavity (Figure 6B) were identified with this technique, appearing with a vacuum effect. In addition, there were areas of soft tissue attenuation medial to the mandible and bilateral to the hyoid apparatus, compatible with the intermandibularis, genihyioideus and hyoglossus muscles (1 in Figure 2B and Figure 3B), the pterygoideus, omohyoideus, sternohyoideus and ceratohyoideus muscles (2 in Figure 4B), the adductor mandibulae externus medialis and superficialis muscles located, respectively, dorsomedially and ventrolaterally to the adductor fossa (3 and 4 in Figure 4B and Figure 5B).In addition, we distinguished with adequate resolution different structures of the central nervous system such as the olfactory bulb, the brain (telencephalon and diencephalon), the cerebellum and the brainstem (Figure 3B, Figure 4B, Figure 5B and Figure 6B).3.3. Magnetic Resonance Imaging (MRI)The soft structures of the iguana’s head, such as the central nervous system as well as the eyeball’s structures (vitreous humor and lens), the oral cavity with the tongue and the masticatory muscles, showed an accurate visualization using MRI (Figure 2C, Figure 3C, Figure 4C, Figure 5C and Figure 6C). Therefore, an increased volume of both eyeballs in proportion to the size of the head was seen in all sequences (Figure 3C and Figure 7B). As in CT, the structures with gas content (Figure 2C and Figure 6C) appeared with a vacuum effect, being hypointense in all sequences. The nasal glands were bilaterally symmetric, with regular and well-defined margins, located on both sides of the nasal cavity, being iso/hyperintense in T1W and T2W sequences, compared to the encephalic grey matter (Figure 2C and Figure 4C). In contrast, the medial horn appeared hypo/isointense on T1W and T2W sequences concerning the white matter and with mild differentiation of the external bony cortex in relation to the white matter (Figure 2C).In contrast to the CT images, the bone junctions were not distinguishable on MRI, but those bones that formed the neurocranium, such as the frontal, postfrontal-postorbitary, parietal and supraoccipital bones were identified (Figure 3C, Figure 4C, Figure 5C, Figure 6C and Figure 8B). The cranial musculature was found isointense concerning the thalamus in T2W. This technique enabled a better resolution to identify the muscle groups already mentioned (Figure 2C, Figure 3C, Figure 4C and Figure 5C). In the transverse planes of the encephalon, the cerebral cortex was observed slightly more hyperintense than the white matter, which was more hypointense in T2W sequences (Figure 4C). The diencephalic region (Figure 4C, Figure 5C and Figure 8B) was hypointense (T2W) compared to the cerebral cortex (Figure 5C), showing the thalamus and hypothalamus (Figure 8B). The brainstem appeared hypo/isointense in T2W compared to the cerebral cortex, as well as presenting a markedly tortuous horizontal alignment (Figure 8B). Moreover, the caudal colliculus and the fourth ventricle were also displayed in excellent detail. In the rostral aspect of the telencephalon, we distinguished the dorsal pallium rostral part with its lateral and medial portions (Figure 4C). Interestingly, the dorsal MRI image was quite helpful to identify the olfactory bulb located extracranially, which extended rostromedially between the eyeballs (Figure 3C, Figure 7B and Figure 8B).4. DiscussionTechnological developments in imaging techniques have improved the anatomical knowledge and the diagnosis of several pathologies. From conventional imaging methods, such as radiography and ultrasound, to advanced ones such as CT and MRI, the level of resolution, the rapid acquisition of images and the absence of superimposition have meant innovation in research, daily clinical practice and academic purposes [11,12,28,29,30].According to other descriptive studies of different reptile species [1,3,12,14,16,17,20], the images obtained by CT, MRI and gross-sections were adequate to study the rhinoceros iguana head. Therefore, the gross-sections provided accurate anatomic characteristics of the head structures, mainly those related to the brain, the eyeball and the larynx. To the authors’ knowledge, a unique description using gross-section and conventional CT equipment was performed of the green iguana head [22], but only a few brain and laryngeal structures were labelled.Considering the green iguana [26], some differences were found between the bony structures of these species. Thus, we identified the presence of three horns on the dorsal aspect of the nasal region, which are absent in the green iguana. Moreover, the evaluation of the transverse CT images obtained in the bone window and the post-processing in MPR showed a lower prominence of the occipital ridge in the rhinoceros iguana compared to the green iguana. In addition, as described by other authors [22], the CT images of the rhinoceros iguana head displayed in excellent detail the bony structures compared to the anatomic gross-section. With MRI, the palatine, frontal, postfrontal-postorbital, parietal, supraoccipital bones and crest were distinguishable due to a gross-section and with both imaging techniques. In addition, the eyes were easily displayed with all the techniques used. Nonetheless, specific eyeball structures, such as the iris and the scleral ossifications, were better identified by CT and were hardly visible in anatomic gross-sections, where the lens, the retina, the sclera and the vitreous chamber could be visualized. Some of these structures are of scientific and morphological interest to perform further studies on the dimensions of the eyeball and vision ability [29,30].The use of MRI was valuable for visualizing organs located in the head. Thus, transverse MR images T2W facilitated the accurate identification of the main components of the brain, such as the telencephalon, diencephalon, mesencephalon, metencephalon and myelencephalon. Therefore, we identified specific structures such as the lateral ventricles, the anterior dorsal and posterior dorsal ventricular ridges, the cerebellum, the fourth ventricle, the dorsal pallium and the brain stem with the caudal colliculus. Interestingly, the dorsal MR T2W image showed the olfactory bulb isointense compared to the telencephalon, located in the rostral portion of the encephalon but in an extracranial situation, remaining in the rostromedial aspect of the eyeballs (extracranial structures). In contrast to other articles that excluded the olfactory bulb description, we present the shape, location and intensity of this structure [31]. This extracranial location was already observed in previous studies on the tawny dragon [31]. However, this finding had not been reported in other species of reptiles, such as the rhinoceros iguana. The gross-section images confirmed the presence of these structures and their rostral extension. Interestingly, in other species, such as galliform birds, the olfactory bulbs are separated from the telencephalic hemispheres but located intracranially, protected by the frontal bone [32].In addition, the MRI and CT findings showed a greater angulation concerning the horizontal axis of the encephalon, with an angulation of 40° between the central horizontal axis of the skull and the central axis of the encephalon (dorsal displacement concerning the encephalon of the dog) (Figure 9), differing from other studies where an angulation of 28° was described [31]. In galliform birds, the brain has a similar angulation to the skull axis [32]. More similar aspects to the latter were found, such as the large size of the eyeballs, which were almost as large as the whole encephalon [32]. Nonetheless, further studies with a large number of animals should be performed to confirm these findings.Finally, several studies carried out in reptiles have been performed with micro-CT equipment, as it has a higher resolution, allowing a better distinction of anatomical formations [33,34,35]. However, in our work, the combined application of different advanced techniques (CT, MRI) and macroscopic anatomical sections provided helpful information for anatomic descriptions in teaching and clinical practice.5. ConclusionsIn conclusion, CT and MRI appeared to be adequate tools, providing anatomical references of the different bone and soft tissue structures comprising the head of the rhinoceros iguana. Therefore, the findings obtained in this study are helpful for evaluating numerous processes involving the head of these animals, such as abscesses, metabolic bone diseases, fractures and neoplasia. Moreover, these two imaging techniques could contribute to veterinary anatomy learning by our students as these techniques allow the visualization of structures without overlapping, eliminating the difficulties of visualizing the extent of different types of lesions.

animals : an open access journal from mdpi

10.3390/ani11082235

PMC8388426

The aim of this study was to evaluate the proinflammatory (interleukin 12, GM-CSF, interferon-γ) to anti-inflammatory (interleukins 10, 4) cytokine ratios, oxidant level (malondialdehyde) and antioxidant enzyme (superoxide dismutase, glutathione peroxidase) activities in sera from symptomatic and asymptomatic cattle infected with blood parasites relative to clinically healthy cattle. Elevations in serum GM-CSF level was recorded in both symptomatic and asymptomatic cattle infected with blood parasites. Cattle naturally infected with C. M. haemobos and T. orientalis but without clinical signs had anti-inflammatory cytokine immune responses characterized by low proinflammatory to anti-inflammatory cytokine ratios which favor parasite persistence and resistance in the host. Serum activities of antioxidant enzymes (superoxide dismutase and glutathione peroxidase) and malondialdehyde concentration were elevated in all the symptomatic and asymptomatic blood parasite infected cattle groups. Therefore, a high GM-CSF:IL10 response alongside marked antioxidative enzyme responses were predominant findings in cattle with clinical signs of natural blood parasite infections.

The aim of this study was to measure the serum proinflammatory (IL-12, GM-CSF & IFN-γ) to anti-inflammatory (IL-10, IL-4) cytokine ratio, oxidant (MDA) level and antioxidant enzyme (SOD; GPx) activities after blood parasite infections. The blood and serum samples were obtained from 130 cattle and screened for identity of the infecting blood parasites by conventional PCR. The following blood parasite species were detected: Candidatus Mycoplasma haemobos (70/130); Theileria orientalis (65/130); Theileria sinensis (32/130); Anaplasma marginale (49/130); Anaplasma platys (7/130); and Trypanosoma evansi (4/130). The GM-CSF/IL-10 ratio showed significantly higher values in all the symptomatic blood parasite infected cattle groups except for symptomatic A. platys infected cattle groups. Anti-inflammatory cytokine immune responses were notable findings in symptomatic and asymptomatic cattle infected with C. M. haemobos and T. orientalis characterized by low serum IL-12:IL-10, IFN-γ:IL-10, IL-12:IL-4 and IFN-γ:IL-4 (p < 0.05). Therefore, high serum GM-CSF:IL:10 in the symptomatic blood parasite infected cattle, low serum IL-12:IL-10, IFN-γ:IL-10, IL-12:IL-4 and IFN-γ:IL-4 ratios in asymptomatic cattle, high MDA level, and increased antioxidant enzyme activities could be useful predictive tools for outcome of natural blood parasite infections in cattle.

1. IntroductionVector-borne diseases such as those caused by Anaplasma, Theileria, Trypanosoma, haemotropic Mycoplasma (hemoplasma) species are common in tropical and sub-tropical parts of the world. This is ascribed to the constant interaction between these haematophagous vectors and host, and a favourable hot and humid climate for breeding and survival [1,2]. The blood parasites invade or attach onto the erythrocytes thereby inducing haemolytic anaemias which may be mild, moderate or severe depending on the number of infecting pathogens, host immunity, nutrition, age and breed [3]. Invasion or attachment of the parasites onto blood cells usually occurs following evasion of the immune system by these parasites, through their respective evading mechanisms [4]. Recovered animals may become asymptomatic carriers and become sources of infection for naïve or immunocompromised cattle recently introduced into a herd. Clinical signs associated with blood parasite diseases include mild to severe anaemia, fever, weight loss, increased ocular discharge, decreased milk production in dairy herd, neurological signs such as convulsion and tremors, anorexia, depression, abortion in severe cases, coagulation disorders, jaundice, haemoglobinuria, oedema and reproductive abnormalities. Route of infection is mainly by the bite of ixodid ticks (Haemaphysalis, Boophilus, Rhipicephalus, Amblyomma) or biting flies (Stomoxys calcitrans) in the case of Trypanosoma evansi. Transplacental and/or colostral transmission have also been documented [5].The host immune response (innate or acquired) to invading pathogens usually results in inflammatory processes that are mediated by cytokines [6]. These immune responses (cellular and humoral) are also primarily based on the transmission of cytokine inhibitory or stimulatory signals [7]. Cytokines are genuine pleiotropic molecules that play a role in immune response via autocrine and paracrine signaling in order to regulate and physiologically influence specific target cells. They are synthesized or expressed by different cell types such as monocytes, macrophages, T lymphocytes (T Helper 1 and 2), monocytes, endothelial cells, fibroblasts, astrocytes and microglia cells following induction by pathogens [8]. These cytokines, also known as protein hormones, act as messengers between the local site of injury and the liver cells, which produce the acute phase proteins, leading to induction of acute phase response in early inflammatory processes (Reviewed in [9]). Inflammatory processes usually involve increased vascular permeability and blood flow which is accompanied by leukocytic infiltration (recruitment of inflammatory cells) and secretion of pro- and anti-inflammatory cytokines. Sometimes, inflammation may go unnoticed or its not properly displayed with obvious clinical signs resulting in subclinical infection which can affect performance in food animals [9,10]. Therefore, in order to establish a potent immune response, there must be a dynamic balance between the pro- and anti-inflammatory cytokine responses that allow for restoration and maintenance of immune homeostasis, by acting as chemical mediators to bridge the innate and adaptive immunity in inflammatory conditions associated with infections [11]. Increase in pro-inflammatory cytokine levels such as interleukin-12 aids in blood parasite clearance from peripheral blood, while IL-10 plays a central role by suppressing the inflammatory process. It inhibits cytokine synthesis by T helper 1 cells and inhibits the cytotoxic activities of monocytes and macrophages as well as synthesis of proinflammatory cytokines and acute phase proteins and thus, has been regarded as the major cytokine that suppress or regulate the inflammatory process and immune response [12]. The aforementioned properties of IL-10, therefore, encourages blood parasite persistence in the host, and consequent carrier state in animals. Cytokines such as interleukin-12, IFN-γ, TNF-α are produced during the acute stages of infection. Due to increase in different types of cytokines at different stages of infections, cytokines are proposed as potential biomarkers of infection. They also contribute to the pathology observed in certain diseases. A shift from the production of proinflammatory cytokines at the beginning of infection to anti-inflammatory cytokines in late or chronic infection is correlated with the ability of the host immune system to reduce parasitaemia and pathology, respectively [13]. Granulocyte-macrophage colony stimulating factor (GM-CSF), also known as colony stimulating factor 2 (CSF-2), is a monomeric glycoprotein and a product of cells of the innate immune system (macrophages, natural killer cells, mast cells, T cells, endothelial cells and fibroblasts), T cells, activated during the inflammatory or pathologic conditions upon receiving immune stimuli, and a cytokine with immunomodulatory activity ([14]; reviewed in [15]). It upregulates the adhesion molecules on fibroblast and endothelial cells. As a white blood cell growth factor, it is involved in the regulation of myelopoiesis by promoting myeloid cell development, maturation and survival [16]. It stimulates differentiated haematopoietic precursor cells into monocytes and granulocytes (neutrophils, eosinophil and basophil) [17], and plays an active role in the proliferation and differentiation of monocytes into macrophages and dendritic cells [18]. Amongst IL-4 and IL-13, GM-CSF is the major cytokine that stimulates the in vitro production of dendritic cells [18]. It regulates T cell, dendritic cell functions and antigen presenting cells activation [19]. GM-CSF has an effect on mature cells of the innate immune system. For instance, GM-CSF enhances neutrophil migration into inflammatory sites and alters the receptors expressed on their surface [20]. Therefore, it is part of the immune/inflammatory cascade, by which activation of a few macrophages can rapidly lead to an increase in their numbers, a mechanism that is important for fighting infection and enhancing their functions in host defense. Macrophages and granulocytes are stimulated by GM-CSF to secrete plasminogen and produce various pro-inflammatory cytokines such as TNF-α, IL-1 and IFN-γ, thereby enhancing their microbicidal and phagocytic activities [18]. Goldstein et al. reported its use as therapy for treatment of innate immunity disorders such as inflammatory bowel diseases. It is also approved for treating bone marrow transplant graft delay or neutropenia in patients after induction of chemotherapy in cancer patients (reviewed in [21]). Its use as an adjuvant in vaccine stimulates cellular and humoral immune response with remarkable T and B cell proliferation and differentiation. The imbalance between proinflammatory and anti-inflammatory may contribute to the development of certain pathologies such as development of multiple organ dysfunction in humans [22].Reactive oxygen species (ROS) and free radicals such as superoxide radical (O2−), hydrogen peroxide (H2O2) and hydroxyl radical, are constantly generated during metabolic processes that occur in cells/tissues. Under physiological processes, there is a balance in the cellular ROS generation and activity of antioxidant enzymes and non-enzymes, and other redox molecules. However, a disturbance in the oxidant/antioxidant equilibrium, seen in blood parasite infections (e.g., theileriosis) can initiate lipid peroxidation which in turn leads to free radical chain reaction that causes cell injury and death [23]. Antioxidant enzymes such as superoxide dismutase (SOD) and glutathione peroxidase (GPx) are the first line of defense against reactive oxygen species mediated injury [24,25]. Superoxide dismutase catalyzes the dismutation of superoxide anion free radical into molecular oxygen and hydrogen peroxide and depletes the level of superoxide radical (O2−), whose increased concentration can cause excessive damage to the cells. Several studies have revealed the therapeutic properties of SOD, and its role as an anti-inflammatory agent and treatment of cancer (reviewed in [26]. Glutathione peroxidase detoxifies endogenous and exogenous reactive oxygen species and eliminates xenobiotics in cells. In addition, its function is to prevent oxidative damage of cellular constituents, thus its deficiency promotes oxidation of protein and deoxyribonucleic acid (DNA) (reviewed in [25]). Malondialdehyde concentration (MDA) is used as a marker of lipid peroxidation, an indicator of oxidative stress and it is derived from polyunsaturated fatty acids oxygenation [27]. To date, only very few available literatures on the alterations in cytokine profile and oxidant/antioxidant levels following natural blood parasite infection in cattle [6,23,24]. Therefore, this present study compares the serum concentrations of proinflammatory (IL-12, GM-CSF and IFN-γ) and anti-inflammatory (IL-10, IL-4) cytokines, proinflammatory to anti-inflammatory cytokine ratios, oxidant (malondialdehyde) and antioxidant enzymes (superoxide dismutase; glutathione peroxidase) from symptomatic and asymptomatic cattle with blood parasite infections.2. Materials and Methods2.1. Ethics StatementThis study was conducted according to the guidelines of the Animal Care and Use Committee, Universiti Putra Malaysia, Animal Welfare Act, with approval number, UPM/IACUC/AUP-R008/2020.The cattle owners consented to this study through written consent forms.2.2. Animals and Blood Sample CollectionThe selection of cattle for evaluation of serum concentration of cytokines (IL-12, GM-CSF, IFN-γ, IL-10, IL-4) and oxidant/antioxidants (MDA, SOD and GPx) was performed on the basis of physical examination and a PCR analysis of blood samples for parasite detection. The blood samples (6 mL) were obtained from randomly selected 90 crossbred Kedah-Kelantan x Brahman (private farm) and 40 Bali (government farm) cattle breeds of varying ages (1–13 years old) and sex (male and female) via coccygeal venipuncture. These cattle were sampled from private and government beef cattle farms in Muadzam Pahang, Malaysia. The animals were reared in semi-intensive farming system and forages supplemented with minerals. All cattle breeds employed in this study had previous history of blood parasite infections in their respective farms. The clinically healthy cattle were selected based on absence of clinical signs of haemoparasitism (anaemia, icterus, weakness, anorexia, ocular discharge) and absence of tick infestation. Their health status was further confirmed by negative PCR detection of blood parasites, absence of blood parasites on Giemsa-stained blood smear, and normal haematology and serum biochemistry parameters. Blood samples for cytokine and oxidant/antioxidant determinations were placed into plain blood vacutainers while blood for genomic DNA extraction were collected into K2 EDTA blood vacutainers. The blood samples were then transported in an ice box to the haematology and clinical biochemistry laboratory, Universiti Putra Malaysia for processing. Sera was harvested from the clotted blood into microcentrifuge tubes after centrifuging at 2250× g for 10 min using portable table centrifuge (EBA 20 Hettich Zentrifugen, Tuttlingen, Germany), and stored at −80 °C until use.2.3. DNA ExtractionGenomic DNA was extracted from K2 EDTA whole blood stored at −80 °C using the DNeasy® Blood and Tissue kit (Qiagen, Hilden, Germany) following the manufacturers’ protocol. The eluted DNA was stored at −20 °C until use. DNA concentration and purity were measured with a Nanodrop spectrophotometer (Tecan Infinite M200®, Grodig/Salzburg, Austria GmbH). DNA samples with A260/A280 ratios between 1.8–2.2 were further analysed.2.4. Detection of Blood Parasites by Conventional PCRPolymerase chain reaction test was done to amplify the partial gene fragments of MPSP gene of Theileria sinensis, MPSP gene of Theileria orientalis, MSP4 gene of Anaplasma marginale, 16S ribosomal RNA gene of Candidatus Mycoplasma haemobos and RoTaT 1.2 VSG of Trypanosoma evansi using previously reported species-specific primers. Primers and thermocyclic conditions are presented in Table 1. Forward and reverse primers for Theileria, sinensis, MPSP gene of Theileria orientalis, MSP4 gene of Anaplasma marginale, 16S ribosomal RNA gene of Candidatus Mycoplasma haemobos and RoTaT 1.2 VSG Trypanosoma evansi were as specified by Bai et al. [28], Shkap et al. [29], Ota et al. [30], Urakawa et al. [31] and Su et al. [32] respectively. Each PCR run was performed in a final reaction volume of 25 µL in 0.2 mL PCR reaction tubes, comprising 5× Green GoTaq® flexi buffer (Promega, Madison WI, USA), 25 mM MgCl2, 10 mM dNTP mix (dATP, dCTP, dGTP, dTTP), 5 U GoTaq®G2 flexi DNA polymerase (Promega), 10 µM of each primer, molecular grade water (Millipore Corporation, Billerica MA, USA) and 100 ng of template DNA. The positive control DNA used in the PCR run consisted of a field isolate confirmed by PCR and sequencing. The field isolate was obtained from the Veterinary Parasitology Laboratory, Universiti Putra Malaysia. For negative control, molecular grade water (Millipore) was substituted for DNA template. PCR amplification was performed using the BioER Little Genius® LED thermal cycler (Hangzhou Bioer Technology, Zhejiang, China). The gel was stained with RedSafeTM (iNtRoN Biotechnology, Jungwon-Gu, Korea) for 10 min and PCR products were examined by 1.2% gel electrophoresis at 80 V for 80 min. The amplified products were visualized using a UV transilluminator (GeneDireXTM, Taoyuan, Taiwan).2.5. Quantification of Serum Inflammatory Cytokines in CattleCommercial bovine cytokine ELISA quantitation kits were employed to measure interleukin-4 (IL-4) (Bioassay Co., Ltd., Shanghai, China), interferon gamma (Fine Test, Wuhan Fine Biotech Co., Ltd., Wuhan, China), interleukin-10 (Bioassay Co., Ltd.), interleukin 12 (Bioassay Co., Ltd.) and Granulocyte Monocyte-Colony stimulating factor (GM-CSF) (Fine Test, Wuhan Fine Biotech Co., Ltd.), following the manufacturers’ instructions. The absorbance was read at 450 nm with a microplate reader (Infinite 200 pro Tecan, Grodig/Salzburg, Austria) and the serum IL-4, IL-10 and IL-12 values (in ng/L), and GM-CSF and IFN-γ values in (pg/mL) were extrapolated from their respective standard curves. 2.6. Oxidant and Antioxidant AssaysFor the oxidant assay, the product of lipid peroxidation (malondialdehyde) was measured in serum by the thiobarbituric acid reactive substances (TBARS) assay as described by Ohkhawa et al. [34] with a slight modification. To determine the serum malondialdehyde concentration, a standard curve was obtained following preparation of standard solutions using 1,1,3,3-tetraethoxypropane (TEP) at different concentrations (0, 0.5, 1, 2.5, 5, 10, 15, 20 MDA/µM). A mixture was obtained by adding 2.4 mL of 1/12 sulphuric acid (H2SO4) and 0.3 mL of 10% sodium tungstate dehydrate (Na2WO4) to 300 µl of plasma (or TEP standard or water blank). The mixture is incubated at room temperature for 10 min before being centrifuged using (EBA 20 Hettich Zentrifugen, Tuttlingen, Germany) at 2625× g for 10 min. The supernatant was thoroughly removed and 0.05 mL of distilled water, 3 mL of 0.05 M hydrochloric acid (HCL) and 2 µL of 1% thiobarbituric acid (TBA) was added to the sediment to make a reactive mixture. The reactive mixture was made up until 5 mL with distilled water before it was placed in a water bath at 95 °C for 60 min. The boiled mixture was kept at room temperature for 10 min before being centrifuged at 2625× g for 10 min. The absorbance of the supernatant was read at 532 nm with a microplate reader (Infinite 200 pro, Tecan) and the results of the serum MDA levels (in nM/mL) were extrapolated from the standard curve.Serum antioxidant levels were evaluated using commercial ELISA kits to measure superoxide dismutase (SOD) (Elabscience, Houston, TX, USA) by the WST-1 method, and glutathione peroxidase (GPx) (Bioassay Technology Laboratory, Shanghai Korain Biotech Co., Ltd., Shanghai, China) following methods described in the kits. The optical density was read at 450 nm with a microplate reader (Infinite 200 pro, Tecan), and results of serum SOD and GPx were extrapolated from their respective standard curves.2.7. Statistical AnalysisData analysis was performed using SPSS 25.0 (Chicago, IL, USA) following normal distribution of data (Shapiro-Wilk’s test). Data were expressed as mean ± standard error. Values for all the parameters measured for each cattle group were extrapolated from their standard curves respectively. 95% Confidence interval was calculated using proportion test https://epitools.ausvet.com.au/ciproportion (accessed on 12 November 2020). One way analysis of variance (ANOVA) followed by Duncan multiple post hoc comparison test was applied on all data generated from this study. Only the data generated from single blood parasite infected cattle and clinically healthy cattle were utilized in this study. Cattle infected with more than one blood parasite were not involved in this study. Pearson’s correlation coefficient was used for determining the association between factors (age, gender and clinical symptoms) and cytokine level in each of the blood parasite infected cattle. p < 0.05 was considered statistically significant. Determination of the effect of age and farm location on the oxidative stress parameters were done using independent sample t-test. The differences in the response to infection and oxidative stress with respect to age and farm ownership were determined by independent sample t-test. p < 0.05 was considered as statistically significant.3. Results3.1. Molecular Detection Rate of Blood Parasites in the Symptomatic and Asymptomatic CattleThe following blood parasite species were detected by PCR in both symptomatic and asymptomatic cattle: Theileria orientalis, Theileria sinensis, Candidatus Mycoplasma haemobos, Anaplasma marginale and Anaplasma platys. Trypanosoma evansi was only detected in the crossbred Kedah-Kelantan x Brahman cattle. Non-specific clinical signs such as weakness, pallor of mucous membrane, inappetence, cachexia, lateral recumbency, icterus, dehydration, mild tick infestation was only observed in most of the symptomatic cattle.The detection rates of blood parasites in this study are presented in Table 2. Candidatus Mycoplasma haemobos was the blood parasite with the highest detection rate 70/130 (53.9%; 95% CI 45.3–62.2%), followed by Theileria orientalis 65/130 (50%; 95% CI 41.5–58.5%), A. marginale 49/130 (37.8%; 95% CI 29.8–46.3%), T. sinensis 32/130 (24.6%; 95% CI 18.0–32.3%), A. platys 7/130; (5.38%; 95% CI 2.63–10.7%) and Trypanosoma evansi 4/130 (3.08%; 95 CI 1.20–7.64%) (Table 1). About 47/130 (36.25%) cattle had single blood parasite infection while 63/130 (48.5%) were infected with more than one blood parasite species.3.2. Serum Proinflammatory, Antiinflammatory Cytokine and Oxidant/Antioxidant Levels in Symptomatic and Asymptomatic Cattle3.2.1. Serum Cytokine and Oxidant/Antioxidant Levels in Symptomatic and Asymptomatic Cattle Naturally Infected with Candidatus Mycoplasma haemobosThe results of the serum cytokine and oxidant/antioxidant levels in symptomatic and asymptomatic cattle naturally infected with Candidatus Mycoplasma haemobos were presented in Table 3. The serum level of IL-12 in C. M. haemobos-infected cattle (asymptomatic cattle) tended to be the highest when compared with that of the C. M. haemobos-infected cattle (symptomatic cattle) groups and clinically healthy cattle (p < 0.05), with 1.39 and 2.42-fold increases in serum IL-12 levels compared to C. M. haemobos-infected cattle (symptomatic cattle) and clinically healthy cattle (p < 0.05). Also, the mean IL-12 value of C. M. haemobos-infected cattle (symptomatic cattle) was significantly higher (p < 0.05) than that of clinically healthy cattle (Table 3). The serum IL-10 level of the C. M. haemobos-infected cattle (asymptomatic cattle) was higher than those of the other cattle groups (p < 0.05), with a 1.24-fold and 2.58-fold increase compared to C. M. haemobos infected cattle (symptomatic cattle) and clinically healthy cattle, respectively. Both symptomatic and asymptomatic Candidatus Mycoplasma haemobos-infected cattle had the highest mean GM-CSF value (p < 0.05) with 9.13 and 8.79-fold increases, respectively, in their serum GM-CSF level when compared to clinically healthy cattle. The level of IFN-γ in the serum of C. M. haemobos infected cattle (asymptomatic cattle) increased by 1.13 and 1.10-fold, relative to C. M. haemobos infected cattle (symptomatic cattle) and clinically healthy cattle (p < 0.05). A significant increase in the serum level of IL-4 was recorded in both symptomatic and asymptomatic C. M. haemobos-infected cattle, relative to clinically healthy cattle (p < 0.05). There was no correlation between other factors (age and gender) and cytokine level in the C. Mycoplasma haemobos (p < 0.05)-infected cattle. However, a negative correlation existed between clinical symptoms and cytokines: IL-12 (r = −0.685; p = 0.001), IL-10 (r = −0.618; p = 0.003), GM-CSF (r = −0.315; p = 0.006) and IL-4 (r = −0.409; p = 0.000) (Table S1).The MDA level in sera from both symptomatic and symptomatic C. M. haemobos-infected cattle was significantly higher than that from serum of clinically healthy cattle (p < 0.05). Also, high serum activities of SOD and GPx were recorded in both symptomatic and asymptomatic cattle infected with C. M. haemobos, and were significantly higher than that of clinically healthy cattle, with the mean SOD value of the C. M. haemobos-infected cattle (asymptomatic) being highest among the cattle groups (Table 3). In addition, the data showed a 1.8-fold increase in the serum SOD activity of C. M. haemobos-infected cattle (symptomatic), relative to the clinically healthy cattle, and a 1.03-fold increase compared to C. M. haemobos-infected cattle (asymptomatic). Increased serum GPx activity was found in both symptomatic and asymptomatic C. M. haemobos-infected cattle, with a 2.54-fold increase in their GPx value compared to clinically healthy cattle group (p < 0.05) (Table 3). The effect of farm ownership (private and government) and age on the oxidant and antioxidative stress responses was determined too and the serum SOD activity of C. M. haemobos cattle sampled from private farm (941.85 ± 15.52) was significantly higher than that from cattle sampled from government farm (876.49 ± 21.35), (p = 0.017). However, age had no significant effect on the oxidative and anti-oxidative stress responses in C. M. haemobos-infected cattle (p > 0.05) (Table S7).3.2.2. Serum Cytokine Levels and Oxidant/Antioxidant Levels in Symptomatic and Asymptomatic Cattle Naturally Infected by Theileria orientalisThe results from cattle infected with Theileria orientalis (symptomatic and asymptomatic) revealed high serum levels of IL-12 and IL-10 compared to clinically healthy cattle (p < 0.05) (Table 4). Specifically, 2.47-fold and 2.35-fold increases in mean IL-12 and IL-10 values were found in Theileria orientalis-infected cattle (symptomatic), relative to clinically healthy cattle (p < 0.05). However, no significant differences were found in the mean IL-12 and IL-10 values of symptomatic Theileria orientalis and asymptomatic Theileria orientalis- infected cattle (p > 0.05). Increased serum GM-CSF value were recorded in the symptomatic and asymptomatic Theileria orientalis-infected cattle l (p < 0.05), with a 5.57 and 5.80-fold increases in their GM-CSF value when compared to clinically healthy cattle. In addition, increases in serum IL-4 levels were found in both symptomatic (1.97-fold) and asymptomatic (2.02-fold) Theileria orientalis-infected cattle, relative to the clinically healthy cattle (p < 0.05). Significant increase in mean IFN-γ value was recorded for both symptomatic and asymptomatic T. orientalis-infected cattle, with 1.7 and 1.6-fold increases in their mean IFN-γ values compared to clinically healthy cattle (Table 4). A Pearson’s correlation coefficient test revealed a negative correlation between gender and IL-10 (r = −0.294; p = 0.022) in the T. orientalis-infected cattle. In addition, a negative correlation was found between clinical symptoms and cytokine level [IL-12 (r = −0.323; p = 0.007), IL-10 (r = −0.374; p = 0.002), GM-CSF (r = −0.279; p = 0.007), IL-4 (r = −0.406; p = 0.001) and IFN-γ (r = −0.337; p = 0.005)] of T. orientalis-infected cattle (Table S2).The mean MDA values of Theileria orientalis-infected cattle (symptomatic and asymptomatic) were significantly higher than those of the clinically healthy cattle (p < 0.05), with a 2.2−2.9—fold increases, relative to the latter. Also, mean SOD and GPx values of both the symptomatic and asymptomatic Theileria orientalis-infected cattle were higher than that of the clinically healthy cattle (p < 0.05), with the most significant increase in mean SOD value observed in the Theileria orientalis-infected cattle (asymptomatic) showing a 1.7-fold increase in its serum SOD activity when compared to the clinically healthy cattle. However, no significant difference in the mean GPx values was observed between the symptomatic and asymptomatic cattle infected with Theileria orientalis (p > 0.05), but the serum activity was significantly higher than that of clinically healthy cattle (p < 0.05) (Table 4). There were no significant differences in the oxidative and antioxidative stress responses from T. orientalis-infected cattle sampled from private and government farms (p > 0.05. Age was not computed because all the T. orientalis-infected cattle were adults (8–13 years) (Table S8).3.2.3. Serum Cytokine and Oxidant/Antioxidant Levels in Symptomatic and Asymptomatic Cattle Naturally Infected with Anaplasma marginaleIn symptomatic and asymptomatic cattle naturally infected with Anaplasma marginale, an increase in mean IL-12 and IL-10 values were recorded in comparison to clinically healthy cattle (p < 0.05). The asymptomatic A. marginale-infected cattle had the highest mean IL-12 and IL-10 values, with 1.9-fold and 15.8-fold increases in serum IL-12 level, and 8.7-fold and 17.5-fold increases in serum IL-10 level, relative to symptomatic A. marginale-infected cattle and clinically healthy cattle, respectively. A 1.0-fold and 6.3-fold increase in serum GM-CSF level were found in symptomatic A. marginale-infected cattle compared to asymptomatic cattle and clinically healthy cattle, respectively (p < 0.05). The highest mean IL-4 value was recorded for the asymptomatic A. marginale-infected cattle, whose IL-4 value was higher by 1.3-fold and 2.8-fold, relative to symptomatic A. marginale- infected cattle and clinically healthy cattle (p < 0.05). The level of IFN-γ was also higher in sera from symptomatic and asymptomatic cattle infected with A. marginale, compared with the clinically healthy cattle. In addition, a 1.3-fold increase in serum IFN-γ level was found in asymptomatic A. marginale-infected cattle compared to symptomatic A. marginale infected cattle, and 2.2-fold increase compared to clinically healthy cattle (Table 5). No correlation (p > 0.05) was found between gender and cytokine level, for the except IL-12 value (r = 0.300; p = 0.051) which was positively correlated with gender and nearly significant (Table S3). Also, a negative correlation existed between clinical symptoms and cytokine level [IL-12 (r = −0.574; p = 0.000), IL-10 (r = −0.507; p = 0.000), GM-CSF (r = −0.426; p = 0.002), IL-4 (r = −0.541; p = 0.000) and IFN-γ (r = −0.524; p = 0.000)] of A. marginale-infected cattle (Table S3).In both symptomatic and asymptomatic A. marginale-infected cattle, a significantly higher serum level of MDA, and higher serum activities of SOD and GPx were recorded, relative to the clinically healthy cattle (p < 0.05). However, there were no significant difference in the mean SOD value of both symptomatic and asymptomatic A. marginale-infected cattle. In addition, 2.3 and 1.7-fold increases in serum SOD activities were found in A. marginale-infected cattle (symptomatic) and A. marginale-infected cattle (asymptomatic) respectively, compared to the clinically healthy cattle (p < 0.05). A 2.7-fold and 5.1-fold increases in serum GPx activities were found in A. marginale-infected cattle (symptomatic) and A. marginale (asymptomatic) respectively, compared to the clinically healthy cattle (p < 0.05) (Table 5). A significantly higher serum SOD activity was recorded in the A. marginale-infected cattle sampled from private farm (989.61 ± 12.90), relative to those from the government farm (844.98 ± 16.28, p = 0.003) (Table S8). The effect of age was not computed because all A. marginale infected cattle were adults (8–13 years) (Table S9).3.2.4. Serum Cytokine Levels in Symptomatic and Asymptomatic Cattle Naturally Infected with Theileria sinensisThe results of the serum levels of cytokine and oxidant/antioxidant in cattle infected with Theileria sinensis are presented in Table 6. High IL-12 and IL-10 level were found in serum from symptomatic cattle infected with T. sinensis, relative to asymptomatic cattle infected with T. sinensis and clinically healthy cattle (p < 0.05), but, no significant difference in the mean IL-12 value was found in T. sinensis-infected cattle (asymptomatic) and clinically healthy cattle (p > 0.05). In addition, 5.1-fold and 1.3-fold increases in the serum IL-12 and IL-10 level, respectively, were found in symptomatic T. sinensis-infected cattle, relative to asymptomatic T. sinensis-infected cattle. Significant increases in the mean values of GM-CSF was recorded from the serum of both symptomatic (4.2-fold) and asymptomatic (3.9-fold) T. sinensis-infected cattle, compared to clinically healthy cattle. Also, when compared to the clinically healthy cattle, the symptomatic and asymptomatic T. sinensis- infected cattle had a higher serum level of IL-4 (p < 0.05). Specifically, a 2.8 and 2.0-fold increase in the mean IL-4 value was recorded for the symptomatic T. sinensis-infected cattle and asymptomatic T. sinensis-infected cattle, respectively, compared to the clinically healthy cattle. The IFN-γ level from the serum of symptomatic T. sinensis-infected cattle was higher than that of the asymptomatic T. sinensis-infected cattle, and their mean IFN-γ values were significantly higher relative to clinically healthy cattle (p < 0.05). Furthermore, the mean IFN-γ value increased significantly by 2.1-fold in the symptomatic T. sinensis-infected cattle and by 1.5-fold increase in the asymptomatic T. sinenis-infected cattle when compared to the clinically healthy cattle. The Pearson’s correlation coefficient revealed a negative correlation between clinical symptoms and cytokine level [(IL-12 (r = −0.444; p = 0.005), IL-10 (r = −0.433; p = 0.006), GM-CSF (−0.600; p = 0.000), IL-4 (r = −0.488; p = 0.002) and IFN-γ (r = −0.441; p = 0.005). However, IL-4 was the only cytokine found to be negatively correlated with age (r = −0.512; p = 0.003) There was no correlation (p > 0.05) between gender and cytokine level (Table S4).The mean MDA value was increased significantly in the asymptomatic T. sinensis-infected cattle by 2.9-fold and by 3.7-fold in the symptomatic T. sinensis-infected cattle when compared to the clinically healthy cattle. Both symptomatic and asymptomatic cattle infected with T. sinensis all showed significant increases in their mean SOD and GPx values when compared to the clinically healthy cattle (p < 0.05). Furthermore, a 1.8-fold increase in mean SOD values was recorded for both symptomatic and asymptomatic T. sinensis-infected cattle, respectively, in comparison to clinically healthy cattle. Also, when compared to clinically healthy cattle, a 4.5-fold increase and 4.4-fold increase in mean GPx value were also recorded for symptomatic T. sinensis and asymptomatic T. sinensis-infected cattle respectively (Table 6). There was no significant effect of farm ownership (private and government) and age (young and adult) on the oxidative and anti-oxidative stress responses of T. sinensis infected cattle (p > 0.05) (Table S10).3.2.5. Serum Cytokine and Oxidant/Antioxidant Levels in Symptomatic Cattle Naturally Infected with Trypanosoma evansiThe results of the serum cytokine and oxidant/antioxidant levels in cattle infected with Trypanosoma evansi are presented in Table 7. The mean IL-12 and IL-10 values increased, respectively, in the symptomatic Trypanosoma evansi-infected cattle by 3.8-fold and 3.6-fold, relative to clinically healthy cattle (p < 0.05). A similar trend was observed for the serum GM-CSF and IL-4 levels in the symptomatic T. evansi-infected cattle, with 6.8-fold and 1.7-fold increases in the serum GM-CSF and IL-4, respectively, compared to the clinically healthy cattle (p < 0.05). Also, the serum level of IFN-γ in the T. evansi-infected cattle was higher than that of the clinically healthy cattle by a 1.7-fold increase (p < 0.05). The Pearson’s correlation coefficient test revealed a negative correlation between clinical symptoms and cytokine level [ IL-12 (r = −0.984; p = 0.000), IL-10 (r = −0.996; p = 0.000), GM-CSF (r = −0.996; p = 0.000), IL-4 (r = −0.997; p = 0.000) and IFN-γ (r = −0.978; p = 0.000) in the T. evansi-infected cattle (Table S5). The MDA level was significantly higher in serum from symptomatic T. evansi, compared to serum from clinically healthy cattle (p < 0.05). Also, high serum activities of SOD and GPx were found in symptomatic T. evansi cattle when compared to the clinically healthy cattle. The data showed a 2.1-fold increase in the serum SOD activity of T. evansi -infected cattle relative to the clinically healthy cattle. Furthermore, 2.4-fold increase in serum GPx activity was found in T. evansi-infected cattle relative to clinically healthy cattle (p < 0.05).3.2.6. Serum Cytokine Levels and Oxidant/Antioxidant Levels in Symptomatic Cattle Naturally Infected with Anaplasma platysA 24.8-fold increase in serum IL-12 level was found in symptomatic A. platys-infected cattle compared to clinically healthy cattle (p < 0.05). Likewise, the serum IL-10 level was highest in the symptomatic A. platys-infected cattle, with a 35.1-fold increase compared to clinically healthy cattle (p < 0.05). Furthermore, GM-CSF value obtained from serum of symptomatic A. platys-infected cattle was significantly higher than GM-CSF value obtained from sera of clinically healthy cattle by a 5.2-fold increase (p < 0.05). The mean IL-4 value of symptomatic A. platys was 3.7-fold higher than that of clinically healthy cattle (p < 0.05). Also, the mean IFN-γ value of symptomatic A. platys-infected cattle was 2.1-fold higher than that of clinically healthy cattle (p < 0.05) (Table 8). There was no correlation between gender and cytokine level in the A. platys-infected cattle. However, a negative correlation existed between clinical symptoms and cytokine level [IL-12 (r = −0.785; p = 0.002), IL-10 (r = −0.670; p = 0.005), GM-CSF (r = −0.941; p = 0.000) (Table S6). The mean MDA value of the symptomatic A. platys-infected cattle was significantly higher than that of the clinically healthy cattle (p < 0.05), with a 4.3-fold increase. Also, serum SOD and GPx activities of symptomatic A. platys-infected cattle were higher than those of the clinically healthy cattle (p < 0.05), showing 2.9 and 6.3-fold increases in the serum SOD and GPx activities, respectively, relative to the clinically healthy cattle (Table 8). Anaplasma platys-infected cattle from the private farm (1561.46 ± 15.74) had a significantly higher serum SOD activity relative to those from the government farm (1029.26 ± 13.14) (p = 0.033). The effect of age is unknown as all the A. platys infected cattle were all adults (Table S11).3.3. Pro-Inflammatory and Anti-Inflammatory Cytokine Ratios in Symptomatic and Asymptomatic Cattle Naturally Infected with Blood Parasites3.3.1. Candidatus Mycoplasma haemobos-Infected CattleThe balance of pro- and anti-inflammatory cytokines reflected by the cytokine ratios were determined for the symptomatic and asymptomatic cattle naturally infected by C. M. haemobos, and the clinically healthy cattle group. A very low IL-12:IL-10 ratio was recorded for the symptomatic and asymptomatic C. M. haemobos-infected cattle, and their cytokine ratios did not differ significantly (p > 0.05) from those of the clinically healthy cattle (Figure 1). A high GM-CSF to IL-10 ratio was recorded for the symptomatic and asymptomatic C. M. haemobos-infected cattle, relative to clinically healthy cattle. A very low IFN-γ:IL-10 ratio was recorded for both symptomatic and asymptomatic C. M. haemobos-infected cattle. Also, a very low IL-12:IL-4 was recorded for both symptomatic and asymptomatic C. M. haemobos-infected cattle. However, there were no significant difference in the IL-12:IL-4 ratio of all the cattle groups (p > 0.05). Also, a slightly high GMCSF:IL-4 was recorded from the serum of symptomatic and asymptomatic C. M. haemobos-infected cattle, relative to the clinically healthy cattle. Likewise, a very low IFN-y:IL-4 ratio was recorded for the symptomatic and asymptomatic C. M. haemobos-infected cattle when compared with the ratio obtained from sera of clinically healthy cattle (p < 0.05) (Figure 1).3.3.2. Theileria orientalis-Infected CattleA low serum IL-12:IL-10 ratio was recorded from both the symptomatic and asymptomatic T. orientalis-infected cattle. (Figure 2). Also, a high GM-CSF:IL-10 ratio was recorded for the symptomatic and asymptomatic T. orientalis-infected cattle, compared to clinically healthy cattle (p < 0.05). The serum IFN-y:IL:10 and IL-12:IL-4 ratios were also low in the symptomatic and asymptomatic T. orientalis-infected cattle. A high GM-CSF:IL-4 was reported for both symptomatic and asymptomatic T. orientalis-infected cattle groups, relative to clinically healthy cattle (Figure 2). Furthermore, a low IFN-γ:IL-4 value was obtained for the symptomatic and asymptomatic T. orientalis-infected cattle, relative to the clinically healthy cattle (p < 0.05) (Figure 2).3.3.3. Anaplasma marginale-Infected CattleThe ratio of IL-12 to IL-10 obtained from the serum of asymptomatic A. marginale- infected cattle was high (p < 0.05), however no significant difference in the IL-12:IL-10 ratio between the symptomatic A. marginale infected cattle and clinically healthy cattle were recorded (p > 0.05). A very high serum GM-CSF:IL-10 ratio was also recorded for asymptomatic A. marginale infected cattle relative to symptomatic A. marginale infected cattle and clinically healthy cattle. However, there was no significant difference in the mean IFN-γ:IL-10 ratio of all cattle groups (p > 0.05). A very low IL-12:IL-4 ratio was obtained from asymptomatic A. marginale = infected cattle compared to clinically healthy cattle. Also, a high GM-CSF:IL-4 ratio was recorded for both symptomatic and asymptomatic A. marginale-infected cattle, relative to clinically healthy cattle (p < 0.05). However, IFN-γ:IL-4 value of the cattle groups did not differ from each other (p > 0.05) (Figure 3).3.3.4. Theileria sinensis-Infected CattleIn the symptomatic cattle infected with Theileria sinensis, high IL-12:IL-10 was recorded relative to asymptomatic Theileria sinensis-infected cattle and clinically healthy cattle (p < 0.05). A high GM-CSF:IL-10 was recorded for the symptomatic and asymptomatic T. sinensis-infected cattle. There was no significant difference in the IFN-γ:IL-10 ratio of both symptomatic and asymptomatic T. sinensis-infected cattle (p > 0.05), but their IFN-γ:IL-10 was lower than that of clinically healthy cattle. The serum IL-12:IL-4 ratio from the symptomatic T. sinensis-infected cattle was lower than that obtained from the sera of asymptomatic T. sinensis-infected cattle and clinically healthy cattle (p < 0.05). High GM-CSF:IL-4 ratio was recorded for both symptomatic and asymptomatic T. sinensis-infected cattle, relative to clinically healthy cattle. The serum IFN-γ:IL-4 ratio was low in all cattle groups but did not differ significantly from each other (p > 0.05) (Figure 4).3.3.5. Trypanosoma evansi-Infected CattleTo determine the balance of pro- and anti-inflammatory cytokines in clinical trypanosomosis, cytokine ratios were determined in the T. evansi-infected cattle and clinically healthy cattle. A low IL-12:IL-10 ratio was recorded for the T. evansi-infected cattle and clinically healthy cattle though not significant. High GM-CSF:IL-10 and GM-CSF:IL-4 and low IFN-γ:IL-10 were recorded in the T. evansi-infected cattle when compared to clinically healthy cattle. A slight increase in IL-12:IL-4 was also recorded in the T. evansi-infected cattle when compared to clinically healthy cattle. IFN-γ:IL-4 ratio in the serum from T. evansi-infected cattle was low but did not differ from that of clinically healthy cattle (Figure 5).3.3.6. Anaplasma platys-Infected CattleIn the symptomatic cattle infected with A. platys, a low IL-12:IL-10, GM-CSF:IL-10 and IFN-γ:IL-10, relative to clinically healthy cattle, were recorded. A high IL-12:IL-4 ratio was also recorded in the A. platys-infected cattle showing clinical signs of the disease. Likewise, a high serum GM-CSF:IL-4 ratio was found in symptomatic A. platys-infected cattle, and the IFN-γ:IL-4 value obtained from serum of symptomatic A. platys-infected cattle which did not differ significantly from that of clinically healthy cattle were also recorded (p > 0.05) (Figure 6).4. DiscussionThis study investigated immune and oxidant/antioxidant responses concerning pro- and anti-inflammatory cytokine levels and oxidant/antioxidant levels in cattle naturally infected by different blood parasites, respectively. In many diseases, a balance between pro-inflammatory/anti-inflammatory cytokines has been reported to be an important marker of infection and disease outcomes. The balance in these cytokine levels has been demonstrated in infectious [35] and non-infectious diseases [36], and their activities are highly dependent upon their serum concentration and duration of expression [23]. The present study showed that A. marginale, T. orientalis, T. evansi, A. platys and T. sinensis infected cattle had increased level of serum IL-12 and IFN-γ. Interleukin 12 and IFN- γ (proinflammatory cytokines) are involved in blood parasite clearance and cytokine mediated pathology. The over expression of these pro-inflammatory cytokines could have exacerbated the severity of these vector-borne diseases, but IL-10 and IL-4 inhibited the pro-inflammatory cytokine activity, evidenced by low IL-12/IL-10 ratio, low IFN-γ/IL-10 and low IFN-γ/IL-4 ratios reported in these cattle groups. However, it seems there is an immunomodulatory effect, as the massive production of IL-12 and IFN-γ was rapidly regulated by IL-10 and IL-4. The IL-10 and IL-4, therefore, might have attenuated or resolved the inflammatory response. The finding of a low IL-12/IL-10 cytokine ratio in T. evansi agreed with Bakari et al. [6], who reported a decreasing trend of IL-12 and a sustained IL-10 cytokine in cattle naturally infected by Trypanosoma species. Suppression of IL-12 and IFN-γ activities however, results in reduced immunopathology, tissue destruction and multiple organ failure observed in natural blood parasite infection in cattle [37,38]. Interferon-gamma was one of the cytokines detected in high amounts in sera from both symptomatic and asymptomatic cattle. Our findings suggest that IFN-γ might be one of the cytokines that persists longer in circulation after resolution of the primary infection or it is among the pro-inflammatory cytokines such as IL-6 that has the ability to persist longer in circulation in animals manifesting the clinical signs of blood parasite infection and asymptomatic carrier cattle [39].The finding of an anti-inflammatory immune response characterized by low IL-12:IL-10 ratio in symptomatic Theileria orientalis-infected cattle groups, was however, in contrast with Theileria annulata infection in Egyptian cattle, where pro-inflammatory cytokine response was observed [23]. This difference could be due to the stage of infection, as pro-inflammatory cytokine responses are usually prominent in the early stage of an infection [40] and we could not ascertain the time origin of infection because this study was a field one. The pro-inflammatory cytokine immune response in the asymptomatic T. orientalis cattle characterized by high IL-12/IL-10 and high IL-12/IL-4 ratios, and symptomatic A. marginale infected cattle characterized by high IL-12/IL-10 ratio thus signifies close contact of the susceptible host with their respective vectors in the farmhouse, persistent antigenic stimulation or the animals may be recovering from the disease.Granulocyte macrophage-colony stimulating factor was found to be present in high amounts in the serum of all blood parasite infected cattle groups with or without the clinical signs of the disease, but the predominant GM-CSF pro-inflammatory cytokine immune response characterized by high GM-CSF:IL-10 ratio were present in the symptomatic and asymptomatic T. orientalis and C. Mycoplasma haemobos, symptomatic T. evansi, and symptomatic and asymptomatic T. sinensis-infected cattle. This indicates a possible response to low white blood cell count, especially low neutrophil count, in the blood parasite-infected cattle reported in our previous study [4]. Leukopenia characterized by neutropenia is usually a consistent finding in blood parasite infections such as trypanosomosis [41]. This finding also suggested that these cattle maybe repeatedly infected with these blood pathogens as shown in the farm records. GM-CSF behave in a paracrine manner by recruiting mature circulating neutrophils, monocytes and lymphocytes to enhance their functions in host defense (reviewed in [14]). Since blood pathogens are associated primarily with anaemia (haemolytic) or have more effect on the erythrocytic parameters, we suggest that the increase in serum GM-CSF levels (pro-inflammatory response) and proinflammatory immune response of GM-CSF characterized by high GM-CSF:IL10 ratio in the aforementioned cattle groups was in response to the degenerative left shift (higher immature neutrophils to mature neutrophils) reported in our previous study [4]. Though we do not have an evidence of secondary bacterial infection, serum GM-CSF, as a white blood cell growth factor is known to increase in animals with infections or inflammatory conditions associated with rapid exhaustion of mature neutrophils and increase demand of neutrophils or monocytes in the peripheral circulation. This demand, however, causes a stimulation of the bone marrow to release immature neutrophils, resulting in either a degenerative or regenerative left shift. Furthermore, GM-CSF stimulates the increase in monocyte and macrophage numbers whose primary function is to phagocytose invading pathogens and resolve the inflammation afterwards.An anti-inflammatory cytokine immune response characterized by low IL-12:IL-10, IL-12:IL-4, IFN-γ:IL-10 and IFN-γ:IL-4 ratios predominated in the C. Mycoplasma haemobos and T. orientalis asymptomatic cattle. The anti-inflammatory cytokine response might have encouraged blood parasite persistence in the host due to inhibition of proinflammatory cytokines production by IL-10 and IL-4 activities as well as the cytotoxic effects of monocytes and macrophages [12].Free oxygen radicals cause lipid peroxidation whose end product is malondialdehyde. The moderate increase in the serum malondialdehyde level in all the blood parasite infected cattle groups suggest a high degree of oxidative stress response during the course of the infection. This agrees with El-Ashker et al. [24] in A. marginale infection, Shiono et al. [42] in T. sergenti infection in cattle, Mishra et al. [43] and Parashar et al. [44] in T. evansi infection in Holstein cattle and horses, respectively. The measurement of glutathione peroxidase and superoxide dismutase activities are indirect ways to evaluate the status of antioxidant defense of the body [45].The increase in the serum activities of glutathione peroxidase and superoxide dismutase (antioxidant enzymes) reported in this study is in contrast with Esmaeilnejad et al. [46] who reported decreases in GPx and SOD activities in crossbred Holstein cattle naturally infected with A. marginale, and in [47] Rezaei and Dalir-Naghadeh [47] and Hassanpour et al. [48] who reported decreased GPx and SOD activities in Holstein cattle with tropical theileriosis. The increase in serum SOD activity observed in A. platys infection from this study was different from that of Himalini et al. [49] who reported a decrease in serum SOD activity in dogs with mixed Babesia canis and Anaplasma platys/phagocytophilum infection. These contrasting findings could be attributed to the time of sampling and stage of infection. The increase in serum GPx and SOD recorded in this study was because these enzymes are the first line of defense and also responded to increase in MDA concentration. Their roles are to mop up free radicals and reactive oxygen radicals that accumulate during a disease process, and thereby maintain the oxidant-antioxidant balance. Furthermore, there is a possibility that the animals are in close contact with arthropod vectors in the farm, thus the host immune system, nutrition and treatment might have played roles in impeding the effect of the blood parasites and hence increase the antioxidant status in the host and eliminating the reactive oxygen species thereafter [50]. Compatible with our findings, Grewal et al. [51] reported an increase in GPx activity in cattle with natural tropical theileriosis, but with no significant change in serum SOD. The increase in GPx and SOD could be attributed to the fact that these enzymes constitute a first line antioxidant defense system which plays key roles in the total defense mechanisms in biological systems such as intracellular destruction of lipid peroxides [52].5. ConclusionsPro-inflammatory cytokine immune response characterized by high serum GM-CSF:IL-10 ratio was the predominant finding in all the symptomatic natural blood parasite infected cattle except for A. marginale. Anti-inflammatory cytokine immune responses were notable findings in asymptomatic cattle infected with C. M. haemobos and T. orientalis characterized by low serum IL-12:IL-10, IL-12:IL-4, IFN-γ:IL-10 and IFN-γ:IL-4 ratios. Also, our findings clearly suggested that natural blood parasite infections can induce marked oxidative stress responses. The increased serum levels of malondialdehyde induced an increase in antioxidative stress responses characterized by increase in serum activities of SOD and GPx. A negative correlation existed between clinical symptoms and cytokine level in all the infected cattle groups. Furthermore, a negative correlation existed between gender and IL-10 in the T. orientalis-infected cattle, and between age and IL-4 in the T. sinensis infected cattle. Also, the highest serum SOD response was found in the C. M. haemobos, A. marginale and A. platys infected cattle sampled from private farm. Therefore, high serum GM-CSF:IL-10 in symptomatic cattle, low serum IL-12:IL-10, IL-12:IL-4, IFN-γ:IL-10 and IFN-γ:IL-4 ratios in asymptomatic cattle, and high serum MDA level alongside increased SOD and GPx activities in both symptomatic and asymptomatic cattle could be useful predictive tools for outcome of natural blood parasite infections in cattle.

animals : an open access journal from mdpi

10.3390/ani12070937

PMC8996838

Toxicological studies of the effects of BPA on tropical clawed frog (Xenopus tropicalis) early embryos show that temporary exposure to BPA during early embryonic development can result in dramatic teratogenesis, DNA damage, and abnormal gene expression. The overall results of this study provide valuable insights for a more holistic assessment of the environmental risks related to BPA in aquatic ecosystems.

Bisphenol A (BPA), an environmental estrogen, is widely used and largely released into the hydrosphere, thus inducing adverse effects in aquatic organisms. Here, Xenopus tropicalis was used as an animal model to investigate the oxidative effects of BPA on early embryonic development. BPA exposure prevalently caused development delay and shortened body length. Furthermore, BPA exposure significantly increased the levels of reactive oxygen species (ROS) and DNA damage in embryos. Thus, the details of BPA interference with antioxidant regulatory pathways during frog early embryonic development should be further explored.

1. IntroductionBisphenol A (BPA) is currently a widely used plastic monomer and plasticizer. BPA levels were found to be increasing rapidly in the hydrosphere, especially in developing regions [1]. In addition, BPA can be released from a variety of products, including dental sealants [2], tin cans [3], and food contact items [4]. Moreover, alkalinity, incomplete polymerization, high temperature, and other environmental effects can promote the release of BPA into both small and large waterbodies. In large waterbodies, low levels of BPA are maintained due to dilution. For example, the concentration of BPA measured in the water column of the Yong River is 0.07–6.20 nmol/L [5]. From 2004 to 2016, the concentrations of BPA in the Elbe River were found to range from 0.02 to 6.57 nmol/L [6]. In small waterbodies, the main sources of many kinds of toxins are anthropogenic, such as the diffusion pollution of garbage dumps on rainy days and pollution due to chemical fertilizers and pesticides. However, in small waterbodies, such as small lakes, ponds, streams, springs, or ditches, the levels of the pollutants are usually much higher than in large waterbodies [7,8]. It is easily speculated but rarely reported that there are short-term high concentrations of BPA when it is released into the environment. These small waterbodies provide habitats to many species and are thus critical for maintaining freshwater biodiversity and ecosystem service [9]. Therefore, it is necessary to explore the impact of short-term high concentrations of BPA on organisms present in small waterbodies. BPA has been well studied as an endocrine disruptor that is structurally similar to diethylstilbestrol (DES) and interacts with estrogen receptors α and β (ERα and ERβ) [10]. In addition to being an endocrine-disrupting compound, BPA has been reported to have a plurality of molecular targets: estrogen-related receptors, G-coupled protein receptors, pregnane X receptors, androgen receptors, thyroid hormone receptors, glucocorticoid receptors, and PPARγ [11]. BPA can interfere with the endocrine system and adversely affect the reproductive, developmental, immune, and other systems of the organisms with which it comes into contact, inducing, for example, abnormal growth patterns and neurodevelopmental delays in children [12]. Previous studies have reported that exposure of Xenopus laevis to BPA caused deformities and failed metamorphosis [13,14,15]. In addition to these diverse effects, growing evidence suggests that the induction of reactive oxygen species (ROS) by BPA may contribute significantly to carcinogenesis [16] and reproductive toxicity [17]. Nrf2 is a transcription factor that is activated under oxidative and electrophile conditions, thereby transactivating the expression of antioxidant and cytoprotective genes [18]. In addition to its antioxidant roles, Nrf2 is also involved in cell proliferation and the determination of cell fate [19]. Intriguingly, Nrf2 is a maternal gene, and the mRNA levels of Nrf2 decrease up to stage 12 (the neurula stage) in Xenopus [20]. Is there a correlation between teratogenesis and ectopic expression of Nrf2, if any, by BPA exposure in early embryonic stages? We hypothesize that BPA exposure induces ROS, which further interfere with the expression pattern of Nrf2 and the following orchestrated embryonic development.Amphibian adults are usually found on sidewalks or in low-lying grass but normally spawn in small waterbodies such as temporary static ponds and puddles formed after heavy rains. Due to their unshelled eggs and highly permeable skin, their embryos and tadpoles are susceptible to contamination, especially in these small waterbodies. Therefore, in this study, we utilized the embryos of a model animal, i.e., X. tropicalis, to investigate the oxidative effects of BPA during early embryonic development (prior to the neurula stages), as well as any prolonged toxicity and effects during later development stages (tadpole stages). We found that BPA exposure increased ROS levels and interfered with the Nrf2 signaling pathway in embryonic development, suggesting an association with teratogenesis. These results can also serve as a reference for the effects of BPA on human and vertebrate embryonic development.2. Materials and Methods2.1. Exposure and SamplingX. tropicalis (Nigerian) embryos were provided by Xiao Huang, Zhejiang University, and were bred and maintained in our lab. The animal research procedures were conducted in compliance with the guidelines of the Wenzhou University Animal Care and Use Committee. According to the requirements for indoor tropical aquatic breeding, the frogs were maintained in a water circulation system, the temperature was controlled at 28 ± 0.5 °C, and the photoperiod was 12 h light/12 h dark. Thirty pairs of mature male and female frogs were induced to mate by injecting them with 150 U and 200 U of human chorionic gonadotropin (Ningbo Second Hormone Factory, Ningbo, China), respectively, in the dorsal lymph sac, thereby inducing the females to start spawning 4 h later. Instead of dissecting the testis to perform in vitro fertilization (IVF), we allowed them to participate in amplexus. To obtain synchronic embryos, we collected the embryos every 30 min, which differed by a half developmental stage. The embryos were slight dejellied using 2% cysteine (pH 8.0) until they could be separated. Then they were washed with 0.1 × Marc’s modified ringer solution (MMR) immediately to stop the dejelling and cultured in 0.1 × MMR. The developmental stages were assessed according to those previously described [21]. Because X. laevis did not survive even for 48 h when treated with 100 μmol/L BPA [14], high concentrations and 12 h exposure were set for further exploration of the toxicology effects of BPA. For the BPA exposure experiment, BPA (Sigma-Aldrich, St. Louis, MO, USA) was dissolved in dimethyl sulfoxide (DMSO, Sigma-Aldrich, St. Louis, MO, USA) to 4 mol/L for the stock solution. DMSO may be used as a solvent without complications at a maximum concentration of 0.01% [22]. The stock solution of BPA was diluted in 0.1 × MMR to 25, 50, and 100 μmol/L (0.000625%, 0.00125%, and 0.0025% DMSO, respectively). Each treatment contained three 15 cm glass dishes, and each dish consisted of 200 mL test solutions and 100 randomly selected embryos. Embryos at the four-cell stage were cultured in 0.1 × MMR containing different concentrations of BPA or 0.01% dimethyl sulfoxide (control group) for 12 h at 26 ± 1 °C, followed by another 60 h without BPA. The test solution was changed every 6 h for the first 12 h, and every 12 h thereafter. During the exposure, dead embryos were counted and removed, and abnormal embryos were counted every 12 h. Dead embryos could be recognized by the presence of gray or white patches on the embryo surface. Apparent malformations were recorded based on the Atlas of Malformations [23] and compared with those found in control embryos.2.2. Detection of Levels of Reactive Oxygen Species (ROS)Reactive oxygen species (ROS) were detected using 2,7-dichlorodihydrofluorescein diacetate (H2DCFDA, Sigma-Aldrich, St. Louis, MO, USA) following a standard protocol [24]. H2DCFDA was used to determine the hydrogen peroxide (H2O2) concentration within the embryos [25]. Fifteen normal live embryos were randomly selected from each dish for ROS detection and were pooled into one sample (three mixed samples per treatment group) at time points of 12 and 72 h. The fluorescence intensity of ROS in X. tropicalis embryos was detected using a fluorescence microplate reader (Cytation 3M, BioTek, Winooski, VT, USA) with an excitation wavelength of 485 nm and an emission wavelength of 520 nm.2.3. RNA Isolation and Quantitative Reverse Transcription PCR (RT-qPCR)At the 12 h time point, 15 embryos were randomly selected from each dish and were pooled into one sample (three samples per treatment group) for RNA extraction. Since RNA is easily degraded, embryos need to be frozen. Before freezing, 1g/L MS-222 was added as anesthesia, and the samples were then frozen using liquid nitrogen. Total RNA was extracted from frozen embryos using TRIzol Reagent (Life Technologies, Carlsbad, CA, USA), and the quality of RNA was established using gel electrophoresis and NanoDrop 2000 (ThermoFisher, Waltham, MA, USA). According to the procedure, 1 μg of total RNA was reverse-transcribed to cDNA using oligonucleotide (dT)-tailed primer and Reverse Transcriptase M-MLV (Takara, Shiga, Japan). Before RT-qPCR, gel electrophoresis was used to verify the specificity of the amplified products. In addition, the melting curves were assessed as another method for verification of specificity during the RT-qPCR process. RT-qPCR reactions were performed in triplicate for each sample, using FastStart Universal SYBR Green Master (Roche, Basel, Switzerland) in a CFX-Connect Real-Time System (BIO-RAD, Hercules, CA, USA). Based on the studies of X. laevis [26], β-actin was taken as the reference gene, and the gene-specific primer sequences are listed in Table 1. The efficiency of these genes ranges from 94.27% to 102.35%, which corresponds to an amplification efficiency between 90 and 110%. The relative mRNA expression level of each gene was normalized to that of the reference genes and calculated using the 2−ΔΔCt method.2.4. Whole-Mount In Situ Hybridization (WISH)Nrf2 sequences were retrieved from the NCBI and Xenbase (http://www.xenbase.org, accessed on 9 February 2022) databases. Following PCR amplification, fragments of Nrf2 were cloned into the PCS107 plasmid. The plasmids were then linearized using BamHI and the mMACHINE SP6 Transcription Kit (Thermo Fisher, Waltham, MA, USA) was used to synthesize the DIG-labeled antisense fragment. Three normal live embryos in the control and 25 μmol/L groups were randomly selected from each dish for WISH at stages 19, 32, and 38 (nine samples per treatment group per stage). Then, the embryos were fixed in MEMFA (0.1 mol/L MOPS, 2 mmol/L EGTA, 1 mmol/L MgSO4, 3.7% formaldehyde, pH 7.4) for 1–2 h at room temperature and completely dehydrated using different grades of absolute ethanol. WISH was performed according to a previously described procedure [27]. Alkaline phosphatase (AP)-coupled anti-DIG antibody was used to recognize the DIG-labeled probe. Staining was conducted in BM Purple (Roche) at 37 °C. The embryos were observed under a stereomicroscope (SteREO Discovery. V8, Carl Zeiss, Jena, Germany), and the images were taken using a digital camera (AxioCam MRc, Carl Zeiss, Jena, Germany). 2.5. Micronucleus TestEach treatment was conducted on 30 tadpoles (around Stage 47), and 10 tadpoles were kept per 10 L glass tank filled with 2 L of different BPA concentrations (25, 50, and 100 μmol/L) or 0.01% dimethyl sulfoxide (control group) for 6 h, before being moved to 0.1 × MMR for another 24 h. Three tadpoles per group were randomly selected and euthanized; then, blood was immediately collected from the tadpole heart to prepare a blood smear. Two smears per tadpole, fixed in methanol for 15 min, were then stained with 10% Giemsa for 30 min. Subsequently, the slides were rinsed with deionized water 3–5 times, before air-drying for microscopic examination. The number of cells that contained one or more micronuclei was determined in a total sample of 1000 blood cells per tadpole. Then, the micronucleus rate was calculated, taking the average (MCN‰). 2.6. Statistical AnalysesIn this study, descriptive data (mean ± SD) were generated for every dependent variable. The mortality, deformity, and micronucleus rates were compared pairwise by chi-squared testing according to a previous study [28]. Before one-way analysis of variance, Shapiro–Wilk was used to test the distribution normality, and Levene’s test was used to test the homogeneity of variance. The data for SOD2 and CAT were compared after conversion (SOD2 was converted by the ln function; CAT was converted by the sin function). The average difference between control and treatment was evaluated by one-way analysis of variance (ANOVA) and the least significant difference (LSD) post hoc test; a p-value of < 0.05 was considered statistically significant. 3. Results3.1. BPA Causes DNA Damage in the TadpoleResults showed that the micronucleus rates were 10‰ ± 1‰, 14‰ ± 2‰, 23‰ ± 2‰, and 33‰ ± 4.16‰ for the control and 25, 50, and 100 μmol/L BPA exposure groups, respectively (Figure 1). Exposure to 50 and 100 μmol/L BPA significantly increased the micronucleus rate (χ2 = 46.94, p < 0.05). 3.2. BPA Is Teratogenic to X. tropicalis EmbryosAll the embryos at the four-cell stage were exposed to different concentrations (25, 50, and 100 μmol/L) of BPA for 12 h. At this endpoint, the embryos arrived at the early gastrula stage (Figure 2(A1)). Treating with BPA was found to result in developmental delays (Figure 2(A2)), abnormal blastopore closure (Figure 2(A3)), and death (Figure 2(A4,A5)). Compared with the control group, all BPA treatment groups showed a significant increase in these factors (χ2 = 244.35, p < 0.05, Figure 2B). In addition, the mortality rate showed a significant dose-dependent increase (χ2 = 39.67, p < 0.05, Figure 2C). After 12 h of BPA exposure, the surviving embryos were transferred to fresh 0.1× MMR and allowed to develop for another 60 h. They were estimated to have arrived at the late tailbud stage in normal conditions (Figure 2(A6)). The embryos in the BPA treatment groups showed flexure of the tail (Figure 2(A7)), pericardial edema (Figure 2(A8,A10)), hypopigmentation (Figure 2(A7,A8,A10)), and a shorter tail (Figure 2(A10)). The most characteristic malformation was a shortened body length (Figure 2(A7,A10)). The malformation rate at the 72 h time point exhibited the same pattern as that at the 12 h time point (χ2 = 203.51, p < 0.05, Figure 2D). As shown in Figure 2E, the mortality rates were also significantly increased along with the dose-dependent increase in BPA concentrations (χ2 = 539.05, p < 0.05, Figure 2E).3.3. BPA Enhances Antioxidant Signaling during Embryonic DevelopmentROS levels were measured after the termination of BPA exposure (12 h) and after a further 60 h of development (72 h). As shown in Figure 3A, ROS levels in the embryos were significantly increased by BPA (F3,8 = 41.37, p < 0.05) in a dose-dependent manner. After removal of BPA, the ROS levels of the 25 and 50 μmol/L BPA treatment groups also showed a significant increase compared to the control (Figure 3B, F2,6 = 51.27, p < 0.05). At 12 h, significant upregulation of Nrf2 and NQO1 was observed in the 50 and 100 μmol/L BPA groups compared to their levels in the control (Figure 3C, F((Nrf2) (3,8)) =7.79, F((NQO1) (3,8)) =8.63, p < 0.05). SOD2 expression was significantly upregulated compared to the control in embryos exposed to 25, 50, and 100 μmol/L BPA (Figure 3C, F3,8 = 11.57, p < 0.05). Compared to the control, the mRNA level of CAT was not significantly changed in the 25, 50, and 100 μmol/L BPA groups.3.4. BPA Interferes with the Spatiotemporal Expression Patterns of Nrf2We further examined the spatiotemporal expression patterns of Nrf2 by WISH. During the neurula stages, the mRNA expression levels of Nrf2 in previous BPA-exposed embryos were significantly increased in the neural crest and neural plate, compared to the control embryos (Figure 4A,B). Following exposure to BPA, a significant upregulation of Nrf2 was observed in the tail and spinal cord, with curvature of the tail and trunk observed at stage 32 (Figure 4D–F). Later, at stage 38, Nrf2 was upregulated in the cloaca, tail, and spinal cord following previous BPA exposure (Figure 4G–I). This showed that BPA upregulated the mRNA expression levels of Nrf2 in the malformed parts of the embryos, suggesting that Nrf2 is linked with teratogenesis following BPA exposure during the early embryonic stages.4. DiscussionEmbryogenesis is an extremely complex but delicately controlled process. According to the spatiotemporal coordination of cell specification and cell migration, the fertilized egg develops into a shaped individual, which mainly occurs within a short period termed gastrulation [29]. Therefore, even a tiny error at the gastrulation stage will result in obvious adverse outcomes, such as malformation and death. The expression level of Nrf2 decreased from the blastula stage, and reached a minimum at stage 12 [20]. Therefore, BPA treatment of embryos from the four-cell stage to the early gastrulation stage (approximately 12 h in duration) was conducted to investigate whether there is a correlation between the teratogenicity caused by BPA exposure and ectopic Nrf2 expression in early embryos. We aimed to uncover the oxidative effects of BPA during embryogenesis, focusing on discerning the effect of BPA exposure during early embryonic stages. In this study, increased BPA exposure increased the deformity and mortality rates of X. tropicalis embryos. Several studies have found that exposure to high concentrations of BPA induces rapid and high mortality in X. laevis, zebrafish embryos, and Artemia nauplii [30,31,32]. However, our results show that the mortality rate of 100 μmol/L BPA treatment was 16.33% ± 2.08%. The different mortality rates can be explained, in part, by the difference in susceptibility between different species. In addition, differences in exposure time points and periods also result in different mortality rates. We found that treatment of X. tropicalis embryos with BPA resulted in developmental delays (Figure 2(A2)), abnormal blastopore closure (Figure 2(A3)), death (Figure 2(A4,A5)), flexure of the tail (Figure 2(A7)), pericardial edema (Figure 2(A8,A10)), hypopigmentation (Figure 2(A7,A8,A10)), and a shorter tail (Figure 2(A10)). Among these effects, the most typical deformity was a shortened body length (Figure 2(A7,A10)). The shortened body length may reduce the growth and survival rate and predation by natural enemies [33]. In addition, it was recently reported that BPA (1-50 μmol/L) treatment in X. laevis caused disrupted cleavage divisions, slowed cytokinesis and cellular dissociation, and resulted in subsequent teratogenesis [34]. The shorter tail and shortened body length observed in our study are consistent with these reported deformities, though we did not observe a curved spinal cord and craniofacial malformations. A relevant study has shown that BPA can cause X. laevis eye dysplasia by affecting Notch signaling [35]. BPA exposure significantly interferes with cell specification and cell migration during embryogenesis, though the underlying mechanisms are largely unclear. The micronucleus test is a simple, rapid, and effective screening method to detect the damage of the cell genetic material (i.e., chromosome) caused by environmental pollutants [36]. It has been recommended by many countries and international organizations as one of the genetic toxicology methods to detect carcinogens and mutagens [37]. The micronucleus rate in BPA treatment groups was found to increase with the increase in BPA concentration (Figure 1), indicating an increase in DNA damage following BPA treatment. The DNA damage caused by BPA was thought to be caused by the indirect effect of ROS [38].It is known that short-term exposure to micromolar doses of BPA can increase the level of oxidative stress [38,39], which could lead to lipid, protein, and polysaccharide oxidation and DNA damage, disrupting the process of apoptosis in the process of organ formation [40]. Even after removing exogenous BPA, increased ROS levels can persist [39]. Due to the fragile environment of the embryo, it is extremely vulnerable to damage by ROS and, thus, ROS damage affects embryo development via the impact on physiological and pathological processes in subsequent stages [41]. To avoid oxidative damage, antioxidant systems such as SOD and CAT play an important protective role. SOD can catalase the conversion of superoxide dismutase into oxygen and H2O2 [42], while CAT can convert H2O2 into water and oxygen, thereby reducing toxicity [43]. Previous studies have demonstrated that BPA can induce oxidative stress by altering the CAT and SOD activity or the transcription of antioxidant-related genes in mice and rats [16]. In the present study, H2O2 concentrations and SOD2 expression levels increased in the embryos after exposure to BPA, which is consistent with the previous study. This might indicate that BPA causes deformities and deaths by inducing oxidative damage in X. tropicalis embryos.Moreover, the Nrf2 pathway also plays a vital role in resisting redox stress [44,45,46]. In addition, it was found that BPA can activate the Nrf2 signaling pathway at high micromolar concentrations (>10 μmol/L) [47]. In the present study, Nrf2 expression was upregulated by 25–100 μmol/L BPA and the expression of NQO1 and SOD2 followed the tendency of Nrf2, which is consistent with the previous studies. However, the expression of CAT showed a weak relation to Nrf2 in embryos. We speculate that the mechanisms regulating CAT in response to BPA exposure are complex.Nrf2 is known to play an important role in cell fate specification [48] by affecting the Notch signaling pathway [49], which is involved in developmental processes. Therefore, Nrf2 may be a key molecule that affects stem cell renewal and cell fate in embryonic and adult tissues [49,50]. In the present study, BPA-induced Nrf2 overexpression may affect cell differentiation and migration during embryogenesis, suggesting that the overexpression of Nrf2 in the embryo interferes with cell migration, thus preventing the embryonic pores from being completely closed (Figure 2(A2,A3)).5. ConclusionsIn this study, we provided evidence that temporal exposure to BPA during early embryonic development stages results in dramatic teratogenesis. Furthermore, following BPA exposure, the expression of the Nrf2, NQO1, and SOD2 genes was upregulated in response to 25-100 μmol/L BPA treatments, suggesting that the Nrf2 signaling pathway and redox balance was disrupted, which may be associated with the teratogenic effects of BPA. It is suggested that organisms living in small waterbodies might be threatened by BPA. Therefore, more research should be carried out to evaluate the potential hazards of BPA to human beings and ecosystems.

animals : an open access journal from mdpi

10.3390/ani13111825

PMC10251992

The concept of flagship species plays a vital role in biodiversity conservation. In order to establish a technical route for selecting flagship species and strengthen the application of the concept of flagship species in biodiversity conservation at the local scale, we selected birds as a target group and conducted a study on selecting flagship species in Lishui, Zhejiang Province, China. Through the analytic hierarchy process-entropy weight method (AHP-EM) and the MaxEnt model, 10 bird species were selected as the flagship species of Lishui, and a further analysis showed that these 10 species were fully representative of birds of Lishui and that their distribution covered the main protected areas in Lishui. It is hoped that the selection method of flagship species detailed in this study can provide a reference and promote a biodiversity conservation level for other regions.

The concept of flagship species is widely used in conservation biology. Flagship birds play a key role in raising conservation funds, increasing awareness of biodiversity conservation, and maintaining ecosystem services. This study selected flagship bird species in Lishui, Zhejiang Province, China, and assessed their conservation effectiveness and ability to serve as umbrella species. A regional bird survey program from 2019–2022 recorded 361 bird species in Lishui. This study constructed a framework of flagship species selection based on social, ecological, economic, and cultural criteria. The analytic hierarchy process-entropy weight method (AHP-EM) was used to rank the score of 361 bird species, and the MaxEnt model was used to analyze the suitable distribution areas of these species. Finally, 10 species, which covered the distribution sites of all 361 bird species, were selected as the flagship species of Lishui. The distribution areas covered all the nature reserves and the priority areas of biodiversity of Lishui, in which these 10 species can also serve as umbrella species to protect local biodiversity. The methodology and ideas in this study could provide insights into the application of conservation concepts at the local level, as well as suggest possible recommendations for local governments to select flagship species for conservation.

1. IntroductionThe Earth is currently experiencing its sixth mass extinction event [1,2] due to habitat destruction, over-exploitation, environmental change, climate change, and other reasons [3]. One of the most important challenges facing humanity in the 21st century is how to prevent species extinction and natural hazards, and the conservation of biodiversity has become a topic of concern around the world [4]. Successful biodiversity conservation requires stable and reliable financial support [5]. The Resource Mobilization Expert Group at the 15th Conference of the Parties to the United Nations Convention on Biological Diversity (CBD) estimated that global funding requirements to meet the targets of the Kunming-Montreal global biodiversity framework range from USD 150 billion to USD 300 billion per year, with only around USD 78–91 billion currently spent annually on biodiversity worldwide [6]. In the case of limited capital supply and protection, it is essential to choose a cost-effective way to use all available resources [7]. To increase the amount of biodiversity that can be conserved with a limited budget, scientists have developed several concepts to try to educate the public on the problem of biodiversity loss and identify species of potential conservation value [8,9]. Among these approaches, the use of flagship species and umbrella species has been an effective and widely used concept for the restoration and conservation of biodiversity across the globe [10].In this context, the positive impacts of flagship species have been intensively discussed and applied [11]. However, the appropriateness and validity of the flagship species concept are highly debated [12]. The selection of flagship species appears to be based purely on their marketing value, with no correlation to their endangerment status or their status and value in the ecosystem [12,13]. For flagship species to better act as ‘ambassadors’ [14], many scientists have adjusted the selection methods and criteria for flagship species: flagship species should not only have strong spiritual, popular aesthetic, or social and cultural connotations based on the economic, cultural, historical, and social values of different species in different countries or regions [15], but they should also have broader ecological and economic conservation value [16]. The use of flagship species with charisma and appeal as an important conservation tool to mobilize positive emotions and attitudes linked to interest in conservation action can help to promote and control the management of large areas of habitat and protect many other lesser-known species [17,18]. Umbrella species can protect numerous co-occurring species, share similar habitat criteria (overlapping ecological niches) or interact with each other [19,20,21], and have an ecological role that distinguishes them from flagship species. Conservationists have proposed the concept of a flagship umbrella species, where a species functions as both a flagship species and an umbrella species. The giant panda (Ailuropoda melanoleuca) in China and the Bale monkey (Chlorocebus djamdjamensis) in Ethiopia are good examples [9,10,22,23].Birds are often used as flagship species to gain support for conservation, including in public relations, education, and fundraising. This is due to the variety of shapes, colors, and sounds that highlight the beauty of birds and the ease with which they can be found and distinguished [24]. Within the 27 Brazilian federal units, birds are most often used as regional flagship species, with all Brazilian states having at least one bird as their flagship species [15]; Hornbills (Bucerotidae) [25] and White-tailed Eagles (Haliaeetus albicilla) [26] are good examples of species that make important contributions to biodiversity conservation as flagship species. Birdwatching tourism [27,28,29], as a typical ecotourism type, is considered one of the fastest-growing nature-based tourism industries in the world [28].As a prefecture-level city in Zhejiang Province, China, Lishui is one of seventeen key areas of global significance for biodiversity conservation in mainland China. Lishui preserves a typical and intact central subtropical forest ecosystem in eastern China, with a rare and large area of broad-leaved evergreen forest zonal vegetation. The species diversity of Lishui is the highest in Zhejiang province, which is known as the biological kingdom of east China, containing endemic species such as Abies beshanzuensis (Pinaceae) [30] and the Baishanzu Horned Toad (Megophrys baishanzuensis) [31]. A recent botanical survey revealed the presence of 42 species of rare and endangered plants among the wild higher plants in Lishui, accounting for 76.4% of the total number of rare and endangered plants in the province. Although the public plays a vital role in biodiversity conservation, flora and fauna are unfamiliar to the public and it is not easy to raise public awareness and understanding of them in a short period, making the task challenging. Lishui is located in the Wuyi Mountains Priority Area, one of China’s priority areas for biodiversity conservation. With its rich bird diversity and good ecological environment, Lishui has strong natural resources for the development of birdwatching tourism [32].This study selected birds as the target species based on the results of a background survey on bird diversity in Lishui. In this study, we selected flagship bird species in Lishui by using the analytic hierarchy process (AHP) [33], the entropy weight method (EWM) [34], and the analytic hierarchy process-entropy weight method (AHP-EWM) [35], and we also testing the feasibility of using these species as umbrella species by using the MaxEnt model [36].2. Materials and Methods2.1. Study AreaLishui is located in the southwest of Zhejiang Province, China (118°41′–120°26′ E, 27°25′–28°57′ N), with a total area of 17,298 km2 [37] (Figure 1). Lishui is a prefecture-level city with the largest land area in Zhejiang Province. The region is in the subtropical monsoon climate zone, with a mild climate, warm winters and early springs, long frost-free periods, and abundant rainfall, with an annual average temperature of 17.8 °C. The terrain is dominated by mid-mountain and hilly landforms, sloping from the southwest to the northeast. Lishui is rich in bird biodiversity and provides habitat for nine species of Class I state key protected wild birds, including Cabot’s Tragopan (Tragopan caboti) and Elliot’s Pheasant (Syrmaticus ellioti), and 65 species of Class II state key protected wild birds. The region has the highest species diversity in Zhejiang Province and is important for biology study [38].2.2. Bird Distribution DataThe bird distribution data were derived from a background survey of bird diversity in Lishui. The survey was conducted from December 2019 to February 2022. A total of 258 line transects and 112 point counts were laid out into 183 grids, each 10 × 10 km in size. Each line transect was at least 1 km, with a total line transect length of 1238 km (Figure 1). Each transect and point was surveyed in each of the four seasons in a year. Grids with multiple line transects and point counts existed, and survey data were statistically categorized to record all bird species surveyed for that grid. The observed species were identified using A Field Guide to the Birds of China [39] and classified according to the List of Birds Classification and Distribution in China (third edition) [40].2.3. Environmental DataIn this study, twenty-seven environmental variables influencing the bird species distribution were selected for the initial simulation of the model from five aspects, i.e., climate, vegetation, topography, man-made interference, and land utilization (Table 1). The 27 environmental variables were cropped according to the study area boundaries and resampled to a spatial resolution of 30 × 30 m, and the coordinate system was standardized to WGS_1984_UTM_zone_51N, thus ensuring that the boundaries and ranks of each environmental factor were consistent. Finally, the processed raster variables were converted to ASCII format.2.4. Selection of Flagship Species2.4.1. Ranking of BirdsThe analytic hierarchy process-entropy weight method (AHP-EWM) was used to rank the birds of Lishui. To determine the weights accurately, this study used a combination of subjective and objective methods to determine the indicator weights in the flagship species evaluation system for birds. The indicator weights were first calculated separately using the AHP method and the EWM. The weights of the two methods were then combined using the principle of minimum information entropy, and the least-squares method was used to optimize the weight model [41].As each indicator had different units, dimensions, and orders of magnitude, to unify the standards, all evaluation indicators were first standardized and transformed into standard values with no dimensions or differences in order of magnitude. Scores were assigned in the range of 0–1 before being analyzed and evaluated. This study identified nine elements that influenced the selection of flagship species and specified the ranking and scoring criteria for each element (Table 2). These criteria and the determination of scores were based on the combined opinions of experts and authors in the relevant fields rather than social surveys. The individual and combined indicator weights with the value yj were synthesized from the dimensionless processing of the indicators for each target species to obtain the assessed values of the subjective and objective indicators and the combined assessed value Zj for each target species. The combined assessed values for the target species were then ranked. (1)zj=∑j=1mwjyj2.4.2. Potential Distribution of SpeciesThe MaxEnt model was used to predict the potential distribution of species. Many bioclimatic variables were spatially auto-correlated, which could lead to the over-fitting of model predictions. The selection and correlation testing of all environmental variables was required to improve the accuracy of the model predictions [42]. The jackknife procedure was used to analyze the relative degree of influence of the environmental variables on the potential distribution of the species [43]. The area under the curve (AUC) of the receiver operating characteristics (ROC) curve was used to verify the predictive accuracy of the model simulation results. The AUC was used as a test of the model’s prediction accuracy, with values closer to 1.0 indicating a higher degree of predictive accuracy. The evaluation criteria were as follows: poor at 0.6–0.7, fair at 0.7–0.8, good at 0.8–0.9, and excellent at 0.9–1.0 [44]. The maximum training sensitivity plus the specificity threshold was chosen to convert the 0–1 continuous species potential distribution predictions into a binary distribution of 0 or 1, where 1 represented the presence of a species and 0 indicated no distribution [45].2.4.3. Identifying the Flagship SpeciesThe results of the MaxEnt distribution model developed for the flagship species were superimposed on the distribution patterns of ‘all birds’ (the 361 species modelled) to determine whether the suitable range of the flagship species included other bird sites.2.5. Protection Effectiveness of Flagship SpeciesAn overlay analysis was used to assess the protection effectiveness of the flagship species. Suitable distribution areas for the flagship species were overlaid with nature reserves and priority areas for biodiversity conservation, and the overlaid distribution was then analyzed for vacancies with priority areas and reserves. This was also used to test whether the selected flagship species could also take on the function of umbrella species protection.3. Result3.1. Bird Species DiversityA total of 361 species were recorded in this study, and they belonged to 19 orders and 80 families. Of these, the largest number of species was recorded in the order Passeriformes, with 184 species in 39 families, accounting for 48.75% of the total number of families and 50.97% of the total number of species. The next largest number of families was nine families of Charadriiformes, accounting for 11.25% of the total number of families. Residents accounted for 38.50%, passage migrants accounted for 26.87%, winter birds accounted for 18.28%, and summer-breeding birds accounted for 16.34%.The survey recorded nine species of Class I state key protected wild bird, namely Cabot’s Tragopan, Elliot’s Pheasant, Baer’s Pochard (Aythya baeri), the Scaly-sided Merganser (Mergus squamatus), the Siberian Crane (Leucogeranus leucogeranus), the Oriental Stork (Ciconia boyciana), the Black-faced Spoonbill (Platalea minor), the White-eared Bight Heron (Gorsachius magnificus), and the Yellow-breasted Bunting (Emberiza aureola). Sixty-five species of the wild birds (18.01% of the total) identified are protected as Class II state key protected wild birds. According to China’s Red List of Biodiversity includes: vertebrates [46], three species of which were recorded at the critically endangered level; six species at the endangered level; eleven species at the vulnerable level; and forty-three species at the near threatened level.3.2. Selection of Flagship SpeciesThe scores of the 361 bird species were calculated using the AHP, EWM, and AHP-EWM methods. The target species were ranked according to the AHP-EWM scores. The top 50 species ranked according to the combined method are shown in Table 3.To achieve effective umbrella protection, potential species should have a sufficient habitat width to cover most of the space in the habitat of each target species within their range [21]. The potential species were predicted using the MaxEnt model to determine suitable ranges and perform an overlay analysis with the full range of birds.In terms of prediction accuracy, the MaxEnt model performed well. The AUC values of Yellow-Breasted bunting, Scaly-sided Merganser, Cabot’s Tragopan, Mandarin Duck, White-necklaced Partridge, and Elliot’s Pheasant were all >0.9, with excellent prediction accuracy being achieved. The study showed that the prediction results of all the models were highly reliable, and there was a good correlation between the environmental variables selected for each species and the prediction results, which could be used for the habitat suitability assessment of the species (Figure 2 and Table 4).The results showed that the suitable distribution area obtained by superimposing the potential distribution maps of the top 10 species covered the distribution sites of all birds. Therefore, this study selected the top 10 species as flagship species (Figure 3).In terms of ecotype, among the ten bird species selected using the above method, four species of land birds were represented by rare pheasants, there were two species each of songbirds and swimming birds, and one species each of raptors and climbing birds. In terms of the IUCN Red List of Threatened Species [47], there were six threatened species at the VU level and above. In terms of conservation status, all the species selected were key protected wild animals in China. In terms of residence type, seven species were mainly resident birds (Table 5).3.3. Overlap Analysis of the Protection AreasAs a prefecture-level city in Zhejiang Province, China, Lishui has the designated Baishanzu National Park, five nature reserves at or above the provincial level, and seventeen biodiversity conservation priority areas. The results showed that the potential distribution areas of the 10 flagship species covered 74.70% of the total area of Lishui, 84.62% of the biodiversity conservation priority areas, and 94.65% of the nature reserves (including the national park) (Figure 4).4. Discussion4.1. Improvement in Flagship Species Selection MethodIn recent years, a variety of approaches have been used to select flagship species. For example, Veríssimo emphasized the marketing role of flagship species and an interdisciplinary framework to improve flagship identification [12]. Roll developed a selection method for interest in reptiles using Wikipedia page viewing rates [48]. Qian established six criteria based on conservation biology, ecosystem function, and socioeconomic and cultural importance, and a candidate flagship species was identified if it satisfied four of the six indicators [18]. The flagship species selected using these methods all seem to contribute to local conservation activities, but it is worth noting that most of the selection methods are applicable to large-scale areas, lacking methodological references at smaller scales (municipal and county), as well as the in-depth confirmation of whether flagship species have any ecological value as umbrella species.4.1.1. Combined Subjective and Objective Weighting AnalysisIn the selection criteria for the weighting system, this study chose a variety of criteria for the selection of the flagship species, some of which were related to the priority given to the ecological value of the species (the conservation rank of the target species, endangerment rank, stability of population size and distribution, population size, and knowability of distribution) and the ecological uniqueness of the species, so that the flagship species selected would have a significant umbrella effect and would be effective in protecting other species in the ecosystem. Other criteria were related to the role of a species in society, the economy, and culture, which can provide a variety of services such as sources of inspiration, aesthetics, and education and play an important role in forming a sense of place, social values, and cultural identity [49], as well as representing the cultural heritage of each region or country [50].The previous selection method based on using the AHP alone to determine indicator weights was improved [51]. As a subjective weighting method, the AHP fully considers the knowledge, experience, and intention preferences of decision makers. Although the ranking of index weights often has a high rationality, it cannot overcome the shortcomings of subjective randomness. As an objective weighting rule, the EWM fully excavates the information contained in the original data itself and determines the index weight completely according to the attributes of the index value. However, the EWM cannot reflect the opinions of decision makers, and the weights obtained may not be consistent with the actual importance or can even contradict it [52]. The use of EWM and AHP alone cannot reflect both the actual importance and significance of the indicators. In the AHP-EWM method, the AHP is combined with the EWM, and least-squares combination optimization is performed [41]. This method not only reduces the subjective influence of the AHP but also weakens the degree of deviation between some conclusions of the EWM and the actual situation, making the results more scientific [53].In this study, although the order of the 10 flagship species in the EWM and AHP-EWM rankings differed, the same set of flagship species ranked in the top 10. However, in the AHP scoring ranking results, the common kestrel was ranked ninth, while the flagship species mandarin duck was ranked seventeenth. Comparing the scoring results of the two methods revealed a large difference between criterion 3 and criterion 9. For selection criterion 3, ‘Stability of population size and distribution’, the common kestrel population size and distribution in Lishui were relatively stable and common, scoring 1.0 point, while the mandarin duck population size and distribution were relatively stable and rare, scoring 0.6 points. For criterion 9, ‘Recognition’, the mandarin duck was well known to the public, scoring 1.0, while the common kestrel was almost exclusively known to professional bird researchers or bird enthusiasts, scoring 0.1. The weight of criterion 3 in the AHP was much greater than the weight of criterion 9; as such, the common kestrel ranked ahead of the mandarin duck. In the EWM, if the information entropy Ej of an indicator is smaller, the degree of variation in the value of the indicator is greater and more information is provided; moreover, the role that Ej can play in the comprehensive evaluation as well as the weight becomes greater [54]. The Ej = 0.892 for criterion 3 was greater than the Ej = 0.609 for criterion 9, and the EWM weight of 0.040 obtained by the kestrel was much less than the mandarin duck’s EWM weight of 0.147, thus placing the mandarin duck in 10th place and the common kestrel in 39th place.Finally, after optimization based on the least-squares combination, the mandarin duck ranked 10th and the kestrel ranked 19th. In general, species that are stable in their local ecosystem distribution and that can act as umbrella species should be given priority consideration as flagship species [55]. According to the statistics, of the nine flagship species other than the mandarin duck, 88.9% had more stable populations and distributions in Lishui each year, and to some extent could cover Lishui birds at different times of the year. However, only 44.4% of the nine flagship species and 6.1% of the three hundred and sixty-one birds were well known to the public. Flagship species are those that capture the imagination of the public and induce support for conservation action, and the level of public familiarity with the species greatly influences conservation efforts [56,57]. A comprehensive comparison therefore suggests that the mandarin duck is more suitable than the common kestrel as a flagship species in Lishui.4.1.2. Species Suitability Distribution Area Range PredictionAlthough a variety of species distribution models have been developed to predict distributions, most species distribution models have certain requirements for data sources [58] and have major limitations. MaxEnt better bridges this gap. Studies have shown that the MaxEnt model outperforms other models in terms of prediction accuracy, especially when the amount of species distribution data is limited and only the presence point data of species can be obtained [59,60]. The MaxEnt model was thus chosen to predict the potential distribution of the species in this study [61]. The current research using the MaxEnt model mainly considers the influence of natural factors such as climate and altitude, and this method is not suitable for areas that are increasingly affected by human activities [62,63]. However, in recent years, human activities have seriously affected the distribution habitats of species. Birds, as intermediate and high-level consumers in the food chain, are highly sensitive to environmental changes. Their number and distribution characteristics play an important role in indicating the status of other members of the ecosystem [64]. Therefore, based on the background survey data of birds in Lishui, this paper used the MaxEnt model to simulate the spatial distribution patterns of birds in Lishui considering both natural and human factors. The results showed that the MaxEnt model did perform well, and the results were easy to understand. The findings indicated that the MaxEnt model could be used to provide a scientific basis for biodiversity conservation planning.4.1.3. The Shortcomings of the MethodThe rating criteria and determined scores formulated in the AHP-EWM method mainly summarize the opinions of some experts and publications, however, the opinions from public was missing. Whether a flagship species can fully effectived depends largely on public support [65]. Moreover, the scoring results of this method are based on existing conditions and have a certain timeliness. Therefore, in the future, the scoring results of species may change with time, and it is necessary to regularly detect whether the selected flagship species are still applicable.The composition and structure of biological communities are largely influenced by the environmental factors shaping the habitat [66]. The habitat and viability of species will change over time [67]. Temporal changes in suitable habitats for different species may occur at different rates [68]. Future work should regularly monitor the role of these selected flagship species in biodiversity conservation in Lishui and adjust the number and species of flagship species that are not suitable for biodiversity conservation on a regular basis.4.2. Effectiveness of Flagship Species in Playing a Conservation Leadership RoleFlagship species are one of the most common marketing tools for biodiversity conservation [69]. Natural resource conservationists often use flagship species to raise conservation funds, stimulate positive conservation attitudes toward the species, and raise awareness to reduce biodiversity loss [12]. Few marketing tools have been as effective in rallying support for conservation as those based on flagship species, where each species is the focus of conservation marketing campaigns based on the characteristics it possesses that appeal to its target audience [70]. The public preference for charismatic bird and mammal species is reflected in a greater willingness to pay for their conservation [71]. Flagship species can play the expected role of target species in conservation planning for wildlife reserves in China [23].Birds are one of the most well-known and popular groups of animals [72,73,74]. The 10 flagship species selected in this study were not only protected species but also had the characteristics of gorgeous plumage, pleasant calls, and lovely appearances [75]. These species covered land birds, swimming birds, raptors, songbirds and climbing birds. The use of multi-species strategies can usually expand the biodiversity coverage under the umbrell and flagship species conservation method, thereby better representing the habitat requirements of sympatric species [23]. From a single point of view, the suitable distribution area of any of the 10 selected flagship species could cover 30.47–78.67% of Lishui’s 361 birds, but the total suitable distribution areas of the 10 flagship species could cover all of the 361 birds surveyed in Lishui. In addition, their adaptive distribution areas covered most of the key areas of biodiversity distribution in Lishui. Therefore, habitat protection for the suitable distribution areas of these 10 flagship species is also conducive to the protection of all of the birds in the city, and protecting these species could serve as umbrella protection. In addition to representing different aspects of biodiversity, the identification of a single umbrella species is inevitably limited by factors such as randomness, demography, phenology, and sampling efforts. Thus, biodiversity hotspots may not be occupied by a single umbrella species, thus providing an additional argument for the selection of multiple umbrella flagship species [76].Increasing the protection of flagship species can effectively promote the protection of other threatened non-flagship species and their habitats and ecosystems [77]. Some species are closely related to local cultures. The publicity and promotion of these charming flagship species are conducive to promoting scientific education on biodiversity conservation and the importance of biodiversity conservation to humans and natural ecosystems. In addition, the promotion of these flagship species could help to attract the attention of the government, social groups, and the public to the cause of biodiversity conservation, increase the influence of urban biodiversity, and stimulate public awareness and action for conservation [78,79]. Governments, businesses, and individuals should be encouraged to raise more financial support for flagship species populations, habitats, and other conservation projects [15,80], thereby increasing capital investment in urban biodiversity conservation. The promotion of flagship species is also conducive to the continuous improvement of biodiversity data management and monitoring platforms and the use of policy advantages and competitive incomes to attract more talented people with expertise to join local biodiversity conservation efforts [81]. Finally, these efforts will result in the holistic and comprehensive protection of local biodiversity [5,82] and achieve the construction of efficient protection management systems.4.3. Challenges and Recommendations for Local Governments in Using Flagship SpeciesThe challenges faced by local governments in flagship species protection are multifaceted, including the requirements for funds, professional knowledge, and skills, as well as interference caused by human activities [83,84]. Local governments need to strengthen the promotion and publicity of flagship species and tap into their cultural and ornamental values. The advertising effects of flagship species is continuously strengthened through commercial or public interest promotion [85]. Flagship species can also provide product endorsements for local food, cultural and creative events and enterprises, and for commercial companies [86], enriching product content and enhancing brand images, or even through the registration of the flagship species image as a trademark [87]. Biodiversity products such as art, literature, and music could also be developed around flagship species to raise awareness and expand the city’s influence from a cultural perspective. Local governments also can develop ecotourism led by flagship species. Flagship species publicity and promotion are conducive to improving public awareness of flagship species and encouraging people to participate in relevant ecotourism, thereby improving local finance and local incomes.5. ConclusionsBased on the survey results of bird diversity in Lishui, this study used a combination of AHP, EWM, AHP-EWM, and MaxEnt to select out 10 bird species that can both serve as flagship species and umbrella species. It is hoped that the flagship species of Lishui will play a leading role in the protection of Lishui’s biodiversity and provide a reference for the selecting of flagship species at the municipal local level.

animals : an open access journal from mdpi

10.3390/ani11071993

PMC8300349

In order to obtain accurate infrared predictions, a large number of training animals are needed, aiming to increase the predictive ability of Fourier-transform infrared (FTIR) predictions. In this study, we compared different validation scenarios that involved combining specialized and dual-purpose dairy breeds in the training population FTIR predictions for three different phenotypes in the major cattle breed, i.e., Holstein cattle. Results show that the design of the training population is an important factor in improving predictive ability in the Holstein breed with potential implications also for the minor breeds. However, this improvement is limited by the phenotypic variability of traits of concern and spectral variability between the training and validation sets and the number of animals in the training population.

In general, Fourier-transform infrared (FTIR) predictions are developed using a single-breed population split into a training and a validation set. However, using populations formed of different breeds is an attractive way to design cross-validation scenarios aimed at increasing prediction for difficult-to-measure traits in the dairy industry. This study aimed to evaluate the potential of FTIR prediction using training set combining specialized and dual-purpose dairy breeds to predict different phenotypes divergent in terms of biological meaning, variability, and heritability, such as body condition score (BCS), serum β-hydroxybutyrate (BHB), and kappa casein (k-CN) in the major cattle breed, i.e., Holstein-Friesian. Data were obtained from specialized dairy breeds: Holstein (468 cows) and Brown Swiss (657 cows), and dual-purpose breeds: Simmental (157 cows), Alpine Grey (75 cows), and Rendena (104 cows), giving a total of 1461 cows from 41 multi-breed dairy herds. The FTIR prediction model was developed using a gradient boosting machine (GBM), and predictive ability for the target phenotype in Holstein cows was assessed using different cross-validation (CV) strategies: a within-breed scenario using 10-fold cross-validation, for which the Holstein population was randomly split into 10 folds, one for validation and the remaining nine for training (10-fold_HO); an across-breed scenario (BS_HO) where the Brown Swiss cows were used as the training set and the Holstein cows as the validation set; a specialized multi-breed scenario (BS+HO_10-fold), where the entire Brown Swiss and Holstein populations were combined then split into 10 folds, and a multi-breed scenario (Multi-breed), where the training set comprised specialized (Holstein and Brown Swiss) and dual-purpose (Simmental, Alpine Grey, and Rendena) dairy cows, combined with nine folds of the Holstein cows. Lastly a Multi-breed CV2 scenario was implemented, assuming the same number of records as the reference scenario and using the same proportions as the multi-breed. Within-Holstein, FTIR predictions had a predictive ability of 0.63 for BCS, 0.81 for BHB, and 0.80 for k-CN. Using a specific breed (Brown Swiss) as the training set for prediction in the Holstein population reduced the prediction accuracy by 10% for BCS, 7% for BHB, and 11% for k-CN. Notably, the combination of Holstein and Brown Swiss cows in the training set increased the predictive ability of the model by 6%, which was 0.66 for BCS, 0.85 for BHB, and 0.87 for k-CN. Using multiple specialized and dual-purpose animals in the training set outperforms the 10-fold_HO (standard) approach, with an increase in predictive ability of 8% for BCS, 7% for BHB, and 10% for k-CN. When the Multi-breed CV2 was implemented, no improvement was observed. Our findings suggest that FTIR prediction of different phenotypes in the Holstein breed can be improved by including different specialized and dual-purpose breeds in the training population. Our study also shows that predictive ability is enhanced when the size of the training population and the phenotypic variability are increased.

1. IntroductionFourier-transform infrared spectroscopy (FTIR) technique is used to obtain the infrared spectra of absorption, emission, and photoconductivity of solids, liquids, and gases. It measures the vibration and rotation of molecules determined by infrared radiation at a specific wavelength [1]. In the animal breeding context, the increasing availability of genomic information has pushed the practice towards the implementation of high-throughput phenotyping techniques such as FTIR, which is able to generate real-time, non-invasive, accurate phenotypic predictions at the population level. In dairy cattle, FTIR spectroscopy is extensively applied to the milk matrix for the prediction of standard milk composition, but it has been also proven to be a useful tool for predicting phenotypes that are difficult or expensive to measure such as milk fatty acids [2], milk proteins [3], methane emission [4], fat, and animals’ metabolic and production efficiency [5]. In the context of genomic selection, the use of a multi-breed framework was investigated with the purpose to increase prediction accuracy for difficult to measure traits in dairy cattle [6].Phenotypic prediction using FTIR milk spectra requires the construction of calibration and validation sets, generally from a small dataset, which may affect the predictive ability, especially for complex phenotypes. A solution to this problem might be to combine information from different breeds and/or populations in the training set. Indeed, some studies in the field of genomic selection have done so, and this has increased the available information for developing the calibration equations, resulting in improved model predictive ability and robustness [7,8,9,10]. However, in FTIR prediction, different cross-validation (CV) strategies have not yet been evaluated. In particular, integrating information from different breeds to increase sample size and phenotypic variability could be useful to improve predictive ability [7].Using multi-breed information in a training dataset is an attractive tool to numerically increase the training set and obtain accurate predictions. A critical step towards implementing FTIR prediction using different breeds is the assembly of a training population to exploit phenotypic variability and improve the prediction accuracy. Multi-breed predictions are more complex than single-breed predictions, and the use of a CV strategy that exploits the potential of different breeds improved the accuracy of genomic prediction [7,8]. We hypothesized that this approach could be potentially transposed at the phenotypic level to improve the accuracy of FTIR predictions. Therefore, the aim of this study was to investigate the effect of combining phenotypic information from different specialized and dual-purpose breeds in the training population to predict difficult-to-measure phenotypes not directly measured from milk (body condition score [BCS] and serum β-hydroxybutyrate [BHB]) and those directly measured in milk, (kappa casein—k-CN), from FTIR spectra to maximize the accuracy in the major cattle breed, i.e., Holstein.2. Materials and Methods2.1. Ethics ApprovalThis study did not require any specific ethics permit. The cows sampled belonged to private commercial herds and were not experimentally manipulated. Milk samples and blood samples were collected during routine milk recording carried out by technicians from the Breeders Federation of Trento Province (FPA, Trento, Italy), and were therefore authorized by a local authority.2.2. DataPhenotypic records were obtained from 1461 cows (specialized and dual-purpose breeds) belonging to 41 multi-breed dairy farms located in the province of Trentino (northeastern Italy) as part of a broader project (Cowplus project), described by Cecchinato et al. [11], investigating cattle farming in mountain areas. All the cows were enrolled in the milk-recording program of the Provincial Federation of Breeders (FPA, Trento, Italy) and monitored for milk production. The multi-breed dairy farms operate according to different production systems: small, traditional farms in the mountainous areas, and farms with larger, modern operations [12]. The 41 selected farms comprised from 1 to 5 breeds, which could be divided into two groups: (i) specialized dairy breeds—Holstein (31 herds, 468 cows) and Brown Swiss (35 herds, 657 cows); and (ii) dual-purpose breeds—Simmental (20 herds, 157 cows), Alpine Grey (14 herds, 75 cows), and Rendena (9 herds, 104 cows). Milk production was recorded by the official milk recording system.Milk samples and phenotypic records were collected from one herd per day. Cow health status was determined based on rectal temperature, heart rate, respiratory profile, appetite, and fecal consistency. Only cows that were clinically healthy at the time of the visit were included in the study. Milk samples (50 mL) were collected from each cow during the evening milking and either (i) maintained at 4 °C (without preservative) until processing (within 24 h) at the Department of Agronomy, Food, Natural Resources, Animals, and Environment (DAFNAE) of the University of Padua, or (ii) stored at −80 °C until chromatographic analyses at DAFNAE’s Central Chemical Laboratory (LaCHI). To quantify BHB, blood samples were collected by a veterinarian via jugular venepuncture using vacutainer tubes. On the same day as milk sampling, BCS was assessed by technicians from the Breeders Federation of Trento Province (FPA, Trento, Italy). Milk FTIR spectra were stored during the milk sampling by the Breeders Federation of Trento Province (Trento, Italy).Prediction in Holstein cows of three different phenotypes with different biological meanings and variabilities, i.e., BCS, serum BHB, and k-CN in % N, were assessed using an ensemble method (gradient boosting machine, GBM), with alternative cross-validation designs (within-breed, across-breed, and multi-breed).Body condition score (BCS) was measured by a trained operator on the same day as milk sampling in cows, defined as days in milk (DIM), ranging from 10 to 380 days, based on Edmonson et al.’s [13] methodology, which classifies cows on a scale ranging from 1 (emaciated) to 5 (extremely fat).β-Hydroxybutyrate (BHB, mmol/L) was measured in individual blood samples collected by jugular venepuncture using vacutainer tubes without anticoagulant. The blood samples were centrifuged at 1780× g for 10 min at 4 °C and stored at −20 °C until analysis at the laboratory of the Department of Animal Medicine, Production, and Health of the University of Padua (Padua, Italy). Blood BHB level was determined by the Ranbut RX Monza test (Randox, Crumlin, UK) on a Cobas C-501 analyzer (Roche Diagnostics, Mannheim, Germany).Kappa casein (k-CN, % N) was measured in individual milk samples collected during the evening milking and stored without preservative at −80 °C until protein fraction analysis at DAFNAE, University of Padua. The separation of milk proteins was performed using validated reversed-phase high-performance liquid chromatography (RP-HPLC), as proposed by Maurmayr et al. [14], and expressed as a percentage of the total milk nitrogen content (% N).Phenotypic quality control was performed for each breed separately by removing observations outside the interval between 3.0 standard deviations below and above the mean. After quality control, normal distribution for each phenotype was checked and at least five animals of each breed in a herd were required for inclusion in the analysis.2.3. Infrared Milk SpectraSpectra were obtained from the milk of all cows using a MilkoScan FT6000 (Foss A/S, Hillerød, Denmark) in the laboratory of the Breeders Federation of Trento Province (northeastern Italian Alps). Two milk spectra per animal were collected, covering the region from the short-wavelength infrared (SWIR) to the long-wavelength infrared (LWIR), with a total of 1060 spectral points in the region from 5011 to 925 cm−1 [15]. Milk spectra transmittance (T) was transformed to absorbance (A) using the equation A = log(1/T), and the two spectra from animals within each breed were then averaged before data analysis (Figure 1A). A principal component analysis integrating Mahalanobis distance was performed on the FTIR spectral data to remove possible outliers, at probability level < 0.01, according to Shah and Gemperline [16]. After quality control for phenotypic information and milk spectra through Mahalanobis distance, data from 460 Holstein, 646 Brown Swiss, 155 Simmental, 99 Rendena, and 73 Alpine Grey cows were included in the subsequent analyses (Figure 1).2.4. Design of the Cross-Validation (CV) PopulationsPrediction of the three target phenotypes in Holstein cows was assessed using within-breed (10-fold_HO), across-breed (BS_HO), and multi-breed (BS+HO_10-fold and Multi-breed) CV strategies.Within-breed (10-fold_HO): The Holstein dataset was randomly split into 10 non-overlapping folds of approximately equal size. The training population consisted of 9 folds (414 cows for BCS, 407 for BHB, and 358 for k-CN), and the testing population was made up of the remaining fold (46 cows for BCS, 45 for BHB, and 35 for k-CN). To evaluate the reliability of the model, the cross-validation process was repeated ten times with each fold used as the testing population only once. We calculated predictive ability in each repetition, as well as the average across 100 estimates (10 folds and 10 replications) and the standard deviation.Across-breed (BS_HO): Brown Swiss cows were assigned to the training dataset in order to develop the calibration equations for BCS (646 cows), BHB (620 cows), and k-CN (520 cows), and these equations were then used to predict the three target phenotypes in the whole Holstein population for BCS (460 cows), for BHB (468 cows), and for k-CN (393 cows). This CV scenario was used to evaluate the possibility of using a prediction equation from a specialized breed in Holstein cattle.Breed combination: this CV design included three scenarios:(i) BS+HO_10-fold, using only specialized dairy breeds (Brown Swiss and Holstein): the training population consisted of all the available Brown Swiss cows (646 cows for BCS, 620 cows for BHB, and 520 cows for k-CN) and 9 folds of the Holstein population (414 cows for BCS, 423 for BHB, and 358 for k-CN), while the testing population consisted of the remaining fold of the Holstein population (46 cows for BCS, 45 for BHB, and 35 for k-CN). We calculated predictive ability in each repetition, as well as the average across 100 predictions (10 folds and 10 replications), and the standard deviation.(ii) Multi-breed, where both specialized and dual-purpose dairy breeds were assigned to the training population. In this scenario, Brown Swiss (646 cows for BCS, 620 cows for BHB, and 520 cows for k-CN), Simmental (154 cows for BCS, 155 cows for BHB, and 93 cows for k-CN), Rendena (99 cows for BCS, 97 cows for BHB, and 95 cows for k-CN), Alpine Grey (73 cows for BCS, 73 cows for BHB, and 68 cows for k-CN) and 9 folds of the Holstein cows (414 cows for BCS, 423 for BHB, and 358 for k-CN) comprised the training population, while the testing population was comprised of the remaining fold of the Holstein cows (46 cows for BCS, 45 for BHB, and 35 for k-CN). As previously described, each Holstein fold was used as the testing population only once, the entire analysis was repeated ten times, and the average predictive ability and its standard deviation were calculated from 100 replications.(iii) Multi-breed CV2, in which both specialized and dual-purpose dairy breeds were assigned to the training population, and we fixed the number of animals in the training population to be similar to that of 10-fold CV (417 cows for BCS, 407 for BHB, and 358 for k-CN). To this aim, the training population from the Multi-breed CV was split into 3 folds, considering a total of 417 cows for BCS (Brown Swiss—189 cows, Holstein—116 cows, Simmental—51 cows, Alpine Grey—25 cows, and Rendena—33 cows), 423 for BHB (Brown Swiss—198 cows, Holstein—116 cows, Simmental—51 cows, Alpine Grey—25 cows, and Rendena—33 cows), and 358 for k-CN (Brown Swiss—167 cows, Holstein—103 cows, Simmental—33 cows, Alpine Grey—23 cows, and Rendena—32 cows). Each training fold was used to predict the phenotypes in the testing population as in the 10-fold HO CV (46 cows for BCS, 45 cows for BHB, and 35 cows for k-CN); this process was repeated 10 times for the testing population and three times for the training set. The average predictive ability and its standard deviation were calculated from 900 replications.2.5. Statistical MethodPhenotypic prediction using FTIR spectra was assessed using GBM, a forward learning ensemble method that converts weak learners into strong learners by combining different predictors in a sequential way to reduce both bias and variance [17,18,19]. GBM sequentially builds regression trees with some shrinkage and variable selection, and each new model is added to the previous model with the aim of reducing the predictive error of the prior tree model considering dependencies among the trees [17,20,21]. The GBM model can be represented as follow:y^=∑m=1Mβmb(x,γm) where y^ is the target phenotype (BCS, BHB, or k-CN); M is the number of iterations (expansion coefficients); βm are the function increments, also called “boosts”; and b(x,γm) are the base learners, simple functions of the multivariate argument x with a set of parameters γm={γ1, γ2,…,γm}. In GBM, expansions of the coefficients {βm}1M and parameters {γm}1M are used to map associations between FTIR predictor variables (x) and the target phenotype (y), considering the joint distribution of all values (y,x) that minimise the loss function L{yi,F(x)}, given [y,Fm−1(xi)+h(yi;xi,pm)], where pm is the FTIR (only 1 FTIR spectrum is selected at each iteration) that minimises the ∑i=1nL[y,Fm−1(xi)+h(yi;xi,pm)]. GBM uses the algorithm specified by Hastie et al. [17]. GBM analyses were performed using the h2o R package (https://cran.r-project.org/web/packages/h2o, accessed on 7 May 2021).The predictive performance of the GBM method depends on four parameters to minimize the error of predictions on the validation subset. These parameters are: (1) the number of trees (Ntree represents the total number of trees in the sequence used in the model), (2) learning rate (determines the contribution of each tree to the final model and performs shrinkage to avoid variable overfitting), (3) maximum tree depth (establishes the level of complex interactions between predictors), and (4) minimum samples considered in each leaf (controls the complexity of each tree). The Ntree values used in the random search ranged from 10 to 8000 in intervals of 10, the learn rate ranged from 0.001 to 1 in intervals of 0.001, maximum tree depth was determined using the values from 1 to 80 in intervals of 1, minimum samples per leaf was determined from 1 to 100 in intervals of 5. We performed a random grid search of hyperparameters using the h2o.grid function in the h2o R package (https://cran.r-project.org/web/packages/h2o, accessed on 7 May 2021) in order to select the optimal hyperparameters combination that minimizes the predictive loss function (i.e., prediction error) of the model for each trait (Figure 2). The random grid search was performed using the training set from each CV design (10-fold_HO, BS_HO, BS+HO_10-fold, Multi-breed and Multi-breed CV2) for each trait, splitting it into a 5-fold CV [22]. Thus, 4 folds were assigned to hyperparameter optimization, aiming to find the best combination of the main hyperparameters for GBM approach, while the 1 remaining fold was used to evaluate the model performance based on the loss function (root mean square error—RMSE) and prediction accuracy (r-square—r2) [19]. After finding the best-trained model with the lowest root mean square error (RMSE) and highest prediction accuracy (r2), it was applied to a disjointed testing population for each CV scenario (10-fold_HO, BS_HO, BS+HO_10-fold, Multi-breed, and Multi-breed-CV, previously described in Section 2.4 Design of the Cross-Validation) to obtain the final prediction parameters (Figure 2).2.6. Assessment of Model PerformanceThe predictive ability of the GBM across the CV scenarios was assessed by Pearson’s correlation (rp) between the observed and predicted phenotypes, root mean square error (RMSE), and the bias of prediction on the testing dataset. To assess the differences in predictive ability (rp) across CV scenarios, evaluation was performed using a Hotelling–Williams t-test [23]. The unbiasedness of the prediction was given by the slope of the linear regression of the observed and predicted values in each cross-validation design. The mean percentage error (MPE%) was used as another model bias parameter: (MPE(%)=∑i=1n(y¯i,obs−yi,predyi,obs)×100n), where y¯i,obs is the average value of the observed phenotypic information, yi,pred is the predicted value in the testing population, and n is the number of animals with phenotypes predicted in the testing population. We also evaluated the relative difference (RD) in predictive ability, calculated as RD=(rm−rC)rC×100, where rm represents model performance using the BS_HO, BS+HO_10-fold, and Multi-breed CV scenarios, and rC represents model performance using 10-fold_HO CV for the training scenario. Fisher’s Z-transform test based on Zou [24] was used to determine the significance level of the differences in predictive ability (Pearson’s correlation) between the BS_HO, BS+HO_10-fold, Multi-breed, and Multi-breed CV2 CV scenarios and the 10-fold_HO.3. Results3.1. Phenotypic and FTIR Spectra InformationThe across-breed descriptive statistics for BCS, BHB, and k-CN are shown in Table 1. Specialized breeds (Holstein and Brown Swiss) had lower BCS than dual-purpose breeds (Simmental, Rendena, and Alpine Grey), with the Holstein breed differing significantly from the dual-purpose breeds (p < 0.0005) and the Brown Swiss differing significantly from the Simmental (p < 0.0035). Simmental cows had the highest value for serum BHB (0.62), statistically different from the other breeds (p < 0.005), while the Brown Swiss and Alpine Greys had the highest values for k-CN proportions (16.13% N and 15.34% N, respectively), which were also statistically different from the other breeds (p < 0.01), and Holstein cows had the lowest phenotypic values (13.73% N; Table 1). Principal component analysis (PCA) was applied to the milk FTIR spectra to visualize the differences across breeds (Figure 1B). The first two principal components (PCs) accounted for 25.80% and 12.25% of the FTIR spectra variability, respectively. We observed no differences in FTIR spectra among breeds, indicating similarity across the specialized and dual-purpose breeds (Figure 1B). Comparing the mean values for the principal components (i.e., the big dots in Figure 1B), we observed a greater similarity between the Simmental and the Holstein breeds against the other breeds. However, the distance existing between Holstein and Brown Swiss was comparable to the distance observed between Holstein and dual-purpose breeds, except for Simmental breed.3.2. Cross-Validation ScenariosThe within-breed (10-fold_HO), across-breed (BS_HO), and multi-breed (BS+HO_10-fold, Multi-breed, and Multi-breed CV2) cross-validations were compared on the basis of the model fit parameters (Table 2). The sizes of both the training and validation sets of each CV scenario are reported in Supplementary Figure S1. The accuracies of the FTIR predictions for BCS, BHB, and k-CN using the different CV strategies are shown in Table 2. Prediction accuracies obtained with BS_HO were 7.5% lower than 10-fold_HO (Table 2 and Figure 3). Interestingly, the use of multi-breed training sets increased phenotypic prediction accuracy by around 6.5% (BS+HO_10-fold) and 8.5% (Multi-breed) (Table 2). To demonstrate the effect of the CV scenarios on predictive ability, the relative difference (RD) was assessed comparing the alternatives CV against the 10-fold Holstein. Including different breeds in the training set led to an increase in model predictive performance, which was more evident with the multi-breed scenario (8% for BCS, 7% for BHB, and 10% for k-CN). On the other hand, splitting the multi-breed training population to consider an equal number of animals as in the 10-fold_HO training population led to a slight relative difference among the 10-fold_HO and Multi-breed CV2 scenarios (−1.23% for BHB, −0.47% for BCS, and 2.47% for k-CN; Table 2).When we investigated the significance of these relative differences using the Hotelling–Williams t-test, we found the RDs were significant for all traits (p < 0.05), except for BCS with the BS+HO_10-fold CV, although here the increase was 4.76% with Pearson’s correlation (Table 2). The relative gains observed with the multi-breed scenario were significantly higher than with all the other CV scenarios (p < 0.005).3.3. Bias and Predictive Error Parameters of the Cross-Validation ScenariosThe coefficient of regression (slope) of the observed values on the predicted values was calculated as a measure of the bias of each CV scenario. A value of bias equal to one is ideal, indicating unbiased predictions [25]. For all traits, the slopes of all the CV scenarios were not significantly different from one, indicating no significant bias in the predictions (Table 2). Nonetheless, the slope value of the BS_HO CV scenario was slightly higher than 1.1 for BCS and k-CN traits (average across traits 1.18), and lower than 0.95 for BHB. Notably, with the larger training populations, i.e., the Multi-breed CV scenario, there were more unbiased predictions (i.e., closer to 1) than with the other cross-validation designs.Predictive error parameters (MPE and RMSE) showed that FTIR predictions with the BS+HO_10-fold and multi-breed scenarios led to lower residual parameters (Table 2). Model fit assessment using RMSE indicated that the multi-breed CV considerably reduced the predictive error, from 8% to 22% for Multi-breed and from 0.1% to 11% for BS+HO_10-fold. Both the BS+HO_10-fold and Multi-breed CV scenarios had higher probabilities of lower residual values compared with 10-fold_HO and BS_HO scenarios (Figure 3). The use of a predictive equation developed in a specific breed, i.e., the BS_HO CV, resulted in a higher number of extreme residual values, suggesting more biased predictions (Figure 3). Using the BS_HO CV scenario, we observed the highest MPE estimation, mainly for k-CN trait with −21.09%, indicating that this scenario CV led to a higher overestimation of the prediction, leading to a slope of 1.21 (Table 2).4. DiscussionAssembling sufficiently large training populations to make accurate FTIR predictions is a major challenge for high-throughput phenotyping in dairy cattle, especially for traits that are difficult or time-consuming to measure. To overcome this problem, we investigated the feasibility of using combinations of different populations and/or different breeds in the training set to obtain greater variability in the available information, and hence improve the prediction accuracy of FTIR spectra. Specifically, we evaluated the performance of within-breed, across-breed, and multi-breed FTIR training sets for phenotypic prediction in the Holstein breed. The results show that combining different breeds in the training set greatly increased FTIR prediction accuracy. Moreover, the phenotypic variability and composition of the training population had a large impact on prediction performance.Using a pure Brown Swiss training population to predict phenotypes in Holstein resulted in lower accuracies compared with using Holsteins in both the training and validation sets. Although, the predictive equation that was developed exhibited poor predictive ability, there were no differences in FTIR variability between these two specialized dairy breeds. This means that the major factor affecting model predictive performance is probably related to phenotypic differences between training and validation individuals (Supplementary Figure S2). These observed differences in phenotypes across breeds were an important factor in the reductions in predictive ability of 10% for BCS, followed by 6.19% for BHB and 6.18% for k-CN when a specific breed was used as the training set (BS_HO scenario) for predictions in the Holstein breed (Table 2). Attaining high predictive ability (rp > 0.55) for FTIR prediction is of interest in the context of genetic selection for the target traits in dairy breeding programs due to the fact that its high accuracy is associated with the highest genetic relationship between measured and predicted traits [26]. The predictive accuracy observed was affected by differences in the number of observations and the extent of phenotypic variability in the training population used to develop the calibration equation (Table 2). In this framework, McParland et al. [27] and Maurice-Van Eijndhoven [28] observed that use of the calibration equation was similarly limited in phenotype and spectra variability among the training and validation subsets. The authors reported that inclusion of different Holstein populations [27] and different breeds [28] in the training population used to create the calibration equations for phenotype prediction is of paramount importance in obtaining more accurate predictions. Differences in milk composition across different breeds is an additional important factor, as the FTIR wavelengths extract interpretable information linking the complex presence of specific chemical bonds in milk and the target phenotype. In this case, the calibration equations were less accurate in linking either the complex trait or the spectral data from the training data with the validation data set.The BS_HO design, where the training population was Brown Swiss and the validation population Holstein, was not useful as it led to poor FTIR predictions, caused by the phenotypic differences between the breeds (Table 2). Studies comparing different CV strategies using a single breed, where the training and validation sets overlapped to a greater (k-folds) or lesser (leave-one-herd-out) extent, confirmed the hypothesis that where there is less overlap prediction accuracy decreases, the residual parameters increase (MPE and RMSE), and there are more biased predictions [4,29]. In principle, FTIR prediction performance could be affected by CV strategies and the inherent nature of the target phenotype, as well as by the populations assigned to the training and validation sets [17]. Major factors accounting for the reduction in predictive ability with the BS_HO CV design could be the number of animals in the training and testing populations (Supplementary Figure S1) and phenotypic differences between the Brown Swiss (training set) and Holstein cows (testing population; Table 1).The prediction ability of FTIR is directly associated with the number of animals in the training population (Supplementary Figure S1) and similarity in terms of phenotypic variation between the training and validation populations. Here, when the equations were calibrated using the target Holstein cows (9 folds) jointly with the Brown Swiss cows (BS+HO_10-fold) or using Brown Swiss plus dual-purpose dairy breeds (Multi-breed), with the aim of increasing the training population, the predictions were more accurate and robust, which shows that better performance is obtained by combining breeds in the training set than by using single (10-fold_HO) or specific breeds (BS_HO). However, when the CV scenario was designed to combine the specialized and dual-purpose breed, maintaining a similar number of animals in the training population as 10-fold_HO, a slight difference was observed, which evidenced that not only the phenotypic variability but also the size of the training set is essential to build robust predictions (Table 2). This could be explained by the greater number of animals in the training population and by the closely related phenotypic values in the training and testing populations, which could be a key factor in increasing predictive ability and obtaining unbiased FTIR predictions [2]. It seems, therefore, that differences in FTIR prediction accuracies are linked to differences in phenotypic and spectral variability [28]. Increasing the number of training cows and the phenotypic variability seems to be important for more accurately determining the contribution of FITR to phenotypic variability and also for distinguishing their effects from random noise regions. Compared with the 10-fold_HO, FTIR predictive ability with the BS+HO_10-fold and Multi-breed CV scenarios increased by 6% and 9%, respectively.This increase in prediction accuracy when using an admixture of breeds in the training population could be a useful tool to predict phenotypes with economic importance in dairy cattle. Our results show that training populations consisting of different breeds, including the target breed (Holstein), represent an efficient way of increasing the accuracies of predictions (Table 2). For multi-breed predictions, it is crucial that milk spectra and phenotypic information exhibit the same variability.The design of the training population using different breeds had a strong influence on the prediction accuracy (Figure 1B). It was observed that using a single specialized dairy breed (i.e., Brown Swiss) BS_HO provided lower predictive ability against 10-fold_HO, which was confirmed by the difference in PCA average (Figure 1B). The observed differences between Holstein and Brown Swiss breeds may be due to differences in milk composition, which are captured by FTIR spectroscopy. This led to a general decrease in accuracy by around 10% for BCS and 6% for BHB and k-CN, and an increase in prediction error by 18% for RMSE with an overestimation of around 168% for BCS and k-CN and underestimation of 139% for BHB based on the MPE parameter (Table 2), which could be due to the FTIR relationship between breeds (Figure 1B). We have also shown that increasing the number of animals in the training population (BS+HO10-folds and Multi-breed) increased prediction accuracy and reduced prediction errors (MPE and RMSE), leading to a less unbiased prediction (MPE) with a slope slightly different from 1 (Table 2). On the other hand, the Multi-Breed CV2, which included different breeds in the training set, keeping the same number of animals as the 10-folds_HO scenario, did not improve accuracy, indicating that the variability and size of the training set represent the main factors to improve the prediction accuracy. Overall, the effectiveness of FTIR predictions in small training populations can be improved by increasing the variability and size of the training set and target population. These results suggest that FTIR spectra (when prediction accuracy is moderate to high) may represent a valid alternative to “standard” phenotyping and can be exploited in dairy breeding programs for traits that are expensive and difficult to measure, achieving a similar or slightly inferior genetic response to the measured traits [30,31]. Furthermore, the use of multi-breed CV scenarios seemed to improve prediction accuracy, explaining a greater proportion of the phenotypic variation of the target trait.5. ConclusionsOur findings confirm that accurate Fourier-transform infrared-based predictions in dairy cattle can be achieved by increasing the size and the phenotypic variability of the training population. Our comparison of different validation strategies showed that phenotypic prediction of Holstein records using a pure Brown Swiss training set resulted in the worst performance, while the best performance was obtained with Multi-breed training populations that included Holstein animals, which increased predictive ability by 6% to 8%. Overall, these results indicate that validation scenarios using combinations of different dairy breeds constitute a promising strategy to improve Fourier-transform infrared-based phenotypic prediction. Moreover, they open the possibility of using a similar approach for improving the phenotypic prediction accuracy also in minor cattle breeds.

animals : an open access journal from mdpi

10.3390/ani11030743

PMC8002136

Physical exercise induces various stress responses leading to a disturbance of homeostasis and a number of regulatory systems are called upon to return the body to a new level of equilibrium. The aim of this study was to evaluate the peripheral modulators of serotoninergic function and neurohumoral factors’ changes in athletic horses during an official jumping competition, and to evaluate their relationship with the physical performance of competing horses. The findings obtained in the current survey showed that jumping exercise influenced the levels of tryptophan, leucine, valine, non-esterified fatty acids (NEFAs), dopamine, and prolactin and that these changes are related to the physical performance of competing horses. These findings suggest that the serotoninergic system may be involved in fatigue during jumping exercise.

The current study aimed to investigate whether peripheral modulators of serotoninergic function and neurohumoral factors’ changes in athletic horses during an official jumping competition, and to evaluate their relationship with the physical performance of competing horses. From 7 Italian Saddle mares (6–9 years; mean body weight 440 ± 15 kg), performing the same standardized warm-up and jumping course during an official class, heart rate (HR) was monitored throughout the competition. Rectal temperature (RT) measurement, blood lactate and glucose concentration, serum tryptophan, leucine, valine, the tryptophan/branched-chain amino-acids ratio (Try/BCAAs), dopamine, prolactin, and non-esterified fatty acids (NEFAs) were assessed before the exercise event (T0), at the end of the competition stage (5 min ± 10 s following the cessation of the exercise, TPOST5), and 30 min after the end of competition (TPOST30). Highest HR values were recorded during the course and at the outbound (p < 0.0001); blood lactate concentration and RT increased after exercise with respect to the rest condition (p < 0.0001). Lower leucine and valine levels (p < 0.01), and higher tryptophan, Try/BCAAs ratio, and NEFAs values were found at TPOST5 and TPOST30 with respect to T0 (p < 0.0001). A higher prolactin concentration was found at TPOST5 and TPOST30 compared to T0 (p < 0.0001), whereas dopamine showed decreased values after exercise compared to rest (p < 0.0001). Statistically significant correlations among the peripheral indices of serotoninergic function, neurohumoral factors, and athletic performance parameters were found throughout the monitoring period. The findings provide indirect evidence that the serotoninergic system may be involved in fatigue during jumper exercise under a stressful situation, such as competition, in which, in addition to physical effort, athletic horses exhibit more passive behavior.

1. IntroductionAnimals competing in equestrian disciplines, particularly in show jumping, are required to have high technical skills. Among the parameters considered for the evaluation of athletic performance in horses, heart rate and blood lactate concentration are the most studied as these parameters are good indices of the fitness level and of the workload effort. Cardiovascular, respiratory, and metabolic responses and musculoskeletal adaptations to exercise have been well studied in equine species, allowing specific exercise testing for racehorses [1]. Nowadays, new markers of performance are used in racehorses, including the evaluation of peripheral blood mononuclear cell proliferation and activity or cytokines’ mRNA expression. Changes in immune cell proliferation, lymphocyte populations, and monocyte functionality have been described in trained and untrained racehorses after exercise, confirming the creation of an anti-inflammatory environment in well-trained horses [2]. It has been suggested that long-distance endurance rides involve strenuous effort, which induces numerous changes in the horse’s body, including the exercise-induced acute-phase response [2,3]. Moreover, regular physical activity results in a decrease of proinflammatory states [3]. Although studies have dealt with exercise testing for event [4], dressage [5], and jumper horses [6], a paucity of information is available on the physical performance and fatigue signals arising in athletic horses during official competition.Despite its current popularity, the cause(s) of fatigue and its underlying mechanism(s) have only recently begun to be explored in the athlete horse. Fatigue is an important factor affecting exercise and sporting performances. It is defined physiologically as the inability to maintain power output [7,8], and the organism uses it as a defense mechanism to avoid irreversible damage due to excessive exertion. Fatigue during exercise has traditionally been considered to be the result of events localized within skeletal muscle [9]. However, fatigue cannot always be ascribed to events residing within skeletal muscle; indeed, it is a complex multifactorial element with peripheral and central components. Central fatigue develops in the central nervous system and involves brain serotonin levels [10]. Although it seems naive to assume that the activity of serotoninergic neurons should be determined by substrate availability alone, the central fatigue hypothesis proposes that this situation may arise during periods of stress, such as those occurring during exercise and/or competition. The serotonergic system is associated with numerous brain functions that can positively or negatively affect endurance [11] Accordingly, the synthesis and metabolism of serotonin in the brain increases in response to exercise [12]. Furthermore, the rise of brain serotonin concentration is associated with markers of central fatigue, such as decreased motivation, lethargy, tiredness, and loss of motor coordination [11]. An increase of serotonin synthesis in the brain is correlated with high levels of blood-borne tryptophan, the amino acid precursor to serotonin. The rate-limiting step in the synthesis of serotonin is the transport of tryptophan across the blood–brain barrier into the brain [13].Besides tryptophan, other compounds in the bloodstream, such as branched-chain amino acids (BCAAs), lipids, hormones, and neurotransmitters, can orchestrate the fine mechanism of the onset of fatigue. During and after exercise, BCAAs intervene in muscle protein synthesis by stimulating mRNA translation [14] and prevent muscle proteolysis by inhibiting mechanisms that involve the mammalian target of rapamycin [15,16]. The increased uptake of BCAAs and their oxidation by skeletal muscle will reduce the concentration, available for competition, with tryptophan for transport across the blood–brain barrier, resulting in an increase in the uptake of tryptophan to the brain, and therefore the synthesis and release of serotonin by the brain [17]. Studies carried out on human species [18,19,20,21] have shown that, depending on the intensity and duration, exercise stimulates the release of the hormone prolactin as well as dopamine synthesis, resulting in behavioral and physiological changes. Indeed, emotional or physical stress can temporarily increase prolactin levels. Prolactin also interacts with the dopaminergic system and has been linked to anxiety in several species. The influence of neurotransmitters on fatigue has also been proposed, and dopamine seems to be linked to the ‘central’ component of fatigue for its well-known role on motivation and motor behavior and it is therefore thought to have an enhancing effect on performance [22,23].Although show jumping is part of the Olympic equestrian disciplines, and despite the need to monitor fitness and workload, there is a paucity of information about field exercise testing in jumper horses during official competition [6,24] and none of these tests have assessed the indices of serotoninergic function and their relationship with exercise performance.In view of these considerations, the current study aimed to investigate the effects of jumping exercise on some peripheral modulators of serotoninergic function (e.g., tryptophan, tryptophan/BCAAs ratio, NEFAs) and neurohumoral factors (e.g., dopamine, prolactin) as well as their relationship with physical performance in high-level jumper horses performing the same standardized warm-up and jumping course during an official competition.2. Materials and Methods2.1. Animals and Experimental DesignThis study was carried out on 7 Italian Saddle mares (6–9 years; mean body weight 448 ± 15 kg), after the informed consent of the owners. All horses were managed equally at the same horse training center in Sicily, Italy (latitude 38°10′35′′ N; longitude 13°18′14′′ E), housed in individual boxes (3.5 × 3.5 m), under natural photoperiod and environmental conditions (mean temperature of 26 ± 5 °C and mean relative humidity of 66 ± 4%). All horses were trained and ridden by the same trainer and rider, respectively. The diet composition for the horses was formulated according to their training requirements. In particular, animals were fed twice a day (7.00 a.m.; 5.00 p.m.), with a total food amount of about 2.5% dry-mater of horse body weight (forage: concentrate ratio 70:30), and water was available ad libitum. All animals were clinically healthy (based on a thorough clinical examination) and free from internal and external parasites. Horses enrolled in the current study had the same level of training, similar fitness, and the same experience of jumping competition.The horses took part in an outdoor jumping competition. The session was preceded by a warm-up on the flat consisting of walk 1 (1 min), trot (3 min), walk 2 (1 min), canter (2 min), and walk 3 (1 min). The next stage of standardized warm-up included 4 vertical and 4 oxer jumps of increasing height (height: from 100 to 140 cm). The trainer timed each warm-up stage and used the same sequence of jumps for all subjects (alternating verticals and oxers with the same eight). After the warm-up, horses competed in the same jumping course with the following technical specifications: total length, 500 m; obstacle height, 140 cm; and total efforts, 15 (9 verticals, 6 oxers, 1 double combination, 1 triple combination). The competition stage was made up of four phases, including: inbound (waiting time inside the arena before competing), course (time of the jumping phase), outbound (time lapse between the end of the course and the exit from the arena), and end (time lapse between the exit and the arrival to the stable).2.2. Rectal Temperature (RT) and Heart Rate (HR) Measurement The RT, taken as being representative of the body temperature, was measured by means of a digital thermometer (HI92704, Hanna Instruments, Leighton Buzzard, Bedfordshire, UK), inserted 15 cm in the rectum before the exercise event (T0), at the end of competition stage (5 min ± 10 s following the cessation of the exercise, TPOST5), and 30 min after the end of the competition (TPOST30).To evaluate the workload during the competition, each horse was equipped with equine HR monitors (Polar Horse Trainer, S 610, Polar Electro Europe BV, Fleurier Branch, Avenue Daniel-Jeanrichard, Switzerland) to record HR during each step of the warm-up and competition. Two electrodes were placed against the horse’s wet coat: the positive electrode was first placed under the saddlepad, the negative electrode was then fixed to the saddle girth on the left side of the thorax, and finally the electrodes were connected to a transmitter (T51H, Polar Electro Oy, Kempele, Finland), fixed to a breast strap, that transmitted data to a watch-type data logger (Polar S-610I, Polar Electro Oy, Kempele, Finland), placed near the electrodes. Recorded data were then downloaded on a personal computer, by using the Polar Equine 4.0 software (Kempele, Finland), to be analyzed. In particular, HR was logged every 5 s during the whole exercise. The values of HR measured before the exercise event at rest (T0); during each step of warm-up, including walk, trot, canter, and jumps 1–8; and competition, including inbound, course, outbound, and end of the competition stage (5 min ± 10 s) following the cessation of the exercise, TPOST5); and 30 min after the end of the competition (TPOST30).2.3. Blood Sampling Procedures and Laboratory AnalysisBlood collection was performed by the same operator before the exercise event (T0), at the end of the competition stage (5 min ± 10 s following the cessation of the exercise, TPOST5), and 30 min after the end of the competition (TPOST30). The samples were collected by jugular venipuncture into EDTA vacuum test tubes (Terumo Corporation, Tokyo, Japan) to assess blood lactate and glucose concentration immediately after collection using a small handheld meter (Accutrend Plus, Roche Diagnostics, Deutschland, Germany). Moreover, from each subject, blood samples were collected into three vacuum tubes with clot activator (Terumo Corporation, Tokyo, Japan), which were placed in refrigerated bags and transported to the laboratory for analysis. One tube with clot activator was centrifuged at 1300g for 10 min and the individual serum samples obtained were deproteinized with 5-sulfosalicylic acid and leucine, valine, and tryptophan concentrations were assessed by the high-performance liquid chromatography method. The ratio between tryptophan and BCAAs levels (Try/BCAAs) was then calculated for each sample. On serum samples, the concentration of serum total proteins (biuret method) and NEFAs was determined by means of commercially available kits (total proteins, Byosistems, Reagents and Instruments, Barcelona, Spain; NEFAs, Randox, Crumlin, UK) by means of an automated analyzer ultraviolet-visible spectrophotometer (model Slim SEAC, Florence, Italy).The second tube with cloth activator was centrifuged at 1000× g for 20 min for the assessment of serum dopamine concentration, whereas for the measurement of serum prolactin concentration, the third tube with cloth activator was centrifuged at 1000× g for 15 min. Serum dopamine concentration was assessed using a commercially available kit by the quantitative Sandwich ELISA method with a sensitivity of 1.0 pg/mL, a detection range of 6.25–200 pg/mL, and both an intra-assay and inter-assay coefficient of variability of less than 15%. Serum prolactin concentration was assessed using a commercially available kit by the quantitative Sandwich ELISA method with a sensitivity of 3.1 ng/mL, a detection range of 6.25–400 ng/mL, and an intra-assay coefficient of variability of less than 8%, and inter-assay coefficient of variability of less than 12%.2.4. Statistical AnalysisBefore statistical analysis, all data were tested for normality of distribution using the Kolmogorov–Smirnov test. All data were normally distributed (p > 0.05) and parametric statistical analysis was performed. One-way repeated measures analysis of variance (ANOVA) was applied to determine statistically significant differences in the values of RT, HR, blood lactate and glucose concentration, serum tryptophan, leucine, valine, Try/BCAAs, total proteins, NEFAs, dopamine, and prolactin obtained from each horse during the monitoring period. The Bonferroni multiple comparison test was applied for post hoc comparison.The Pearson’s correlation test was performed to assess whether the peripheral modulators of serotorinergic function (e.g., tryptophan, leucine, valine, Try/BCCAs, NEFAs) correlated with themselves, and with neurohumoral factors (e.g., dopamine and prolactin), as well as with indices of athletic performance in jumper horses throughout the monitoring period.p-Values < 0.05 were considered statistically significant. Data were analyzed using statistical software Prism v. 4.00 (Graphpad Software Ldt, San Diego, CA, USA, 2003).3. ResultsAll the results are expressed as mean values ± standard deviation (±SD).As shown in Figure 1, higher HR values were found at TPOST5 compared to T0 and TPOST30 at each step of the competition (p < 0.0001). Moreover, the highest HR values were recorded during the course and at the outbound. The statistical analysis of data of HR recorded during the jumping stage of the warm-up showed a significant increase of HR values when shifting from 100- to 125-cm jumps (p < 0.001). As shown in Figure 2, blood lactate concentration showed higher values at TPOST5 with respect to T0 and TPOST30 (p < 0.001), and RT also increased after exercise (TPOST5 and TPOST30) with respect to the rest condition (p < 0.0001), whereas the total protein and glucose concentration was unchanged after exercise compared to the rest condition (p > 0.05). According to the findings gathered from the application of one-way ANOVA and as shown in Figure 3, leucine and valine showed lower levels (p < 0.01) after exercise (TPOST5 and TPOST30) compared to baseline values measured at rest (T0), whereas higher tryptophan, Try/BCAAs ratio, and NEFAs values were found at TPOST5 and TPOST30 with respect to T0 (p < 0.0001). A higher prolactin concentration was found at TPOST5 and TPOST30 compared to T0 (p < 0.0001), whereas dopamine showed lower values after exercise compared to rest (p < 0.0001).Statistically significant correlations among the peripheral indices of serotoninergic function, neurohumoral factors, and athletic performance parameters were found throughout the monitoring period (Table 1).4. DiscussionPhysical exercise, depending on the type and intensity, induces various stress responses leading to a disturbance of homeostasis and a number of regulatory systems are called upon to return the body to a new level of equilibrium [5]. The nervous and endocrine systems work in concert to initiate and control movement, and all involved physiological processes. According to the findings obtained in the current study, the main indices of horses’ fitness showed dynamic changes throughout the monitoring period. Specifically, in agreement with previous studies investigating horse performance during official jumping competition [24,25], the results of the current study confirmed the significant contribution of anaerobic metabolism during jumping exercise [1], showing an increase of the blood lactate concentration 5 min after completing the exercise followed by a decline 30 min after the end of competition. As a matter of fact, although within a few minutes lactate release started to decrease because the removal from the circulation exceeds the release rate from the muscles [26], it is known that the clearance of lactic acid from the muscles and bloodstream takes 30 min after exercise in well-trained horses [27]. The HR of each subject was influenced by each step of both the warm-up and jumping course of the competition, showing the highest values during and immediately after the jumping course. Noteworthy, the high HR values registered during the inbound to the show jumping arena highlighted that the official competition itself represented a contributing factor to HR increase as mental stress [28] and emotionality [29]. As a matter of fact, the stress response after training sessions and competitions in horses is different and HR may indicate the animal’s internal physiological state; indeed, the rise in HR is not only caused by an increase in physical activity, but also reflects heightened emotional reactivity [29]. The trend of HR registered in jumper horses of the current survey agrees with the observations of a previous study carried out on Belgian Saddle breeds competing in the Belgian Championship [24] in which it was suggested that horses experienced a need for acceleration near the finish of the competition, causing HR increase due to the onset of fatigue. Furthermore, the results gathered in the present study highlighted that the obstacle height caused an increase of HR values in competing horses. The rise in HR could be due to sympathetic nervous activity, which is known to increase with increasing exercise intensity, leading to a consistent rise of catecholamine levels. It is well established that the increase of the catecholamine concentration enhances the force of cardiac contraction and thus the cardiac output and the increase of HR. The results obtained in this study showed a significant increase of RT after physical exercise and this was not surprising. Indeed, during exercise, horses’ muscles work and a conversion of stored chemical energy into mechanical energy and thermal energy occurs [30,31]. The heat produced by muscle during exercise raises the core temperature. Therefore, heat dissipation mechanisms should be activated, and nowadays, the most developed techniques, such as infrared thermography, which measures the muscle temperature changes and correlates with lactate concentration, are often used in race training [32]. The primary physiologic mechanism leading to heat dissipation is represented by the increase in cardiac output that results in heat transfer from the core to the cutaneous blood circulation by eliciting reflex neurogenic vasodilatation in the skin, resulting in increased skin temperature [30].Body temperature is controlled by the opposing effects of adrenergic and serotoninergic neuronal activity in the hypothalamus. A rise in body temperature may be due to augmented serotoninergic activity combined with inhibition of adrenergic activity, or purely due to increased serotoninergic activity [33]. Noteworthy, the results found herein seem to suggest that the potential peripheral modulators of serotoninergic function measured in athletic horses during a jumping competition tended to respond as predicted in the “central fatigue hypothesis” by Newsholme et al. [10]. The levels of leucine, valine, and dopamine decreased after exercise in competing horses, whereas tryptophan, Try/BCAAs ratio, prolactin, and NEFAs showed increased concentrations following jumping competition exercise. It could be hypothesized that the BCAAs are used by muscles, mainly active skeletal muscle, during short-term intense exercise as a source of energy for muscle contraction [8]. The increase of tryptophan found after exercise could be related to the rise of the lipid content in the blood, which is likely to occur during exercise, as confirmed by the positive correlation found between the values of this amino acid and NEFAs throughout the monitoring period. During physical exercise, free fatty acids are mobilized from adipose tissue and are transported via the blood to the muscle, where they serve as fuel for muscle contraction [34]. Both free fatty acids and tryptophan compete for binding to albumin and thus, the increase of the NEFAs concentration during exercise, resulting from an elevated adrenergic tone that stimulates subcutaneous lipolysis [35], could limit the binding of tryptophan and albumin, increasing the concentration of serum free tryptophan. The uptake of BCAAs by the active muscle as a source of energy yields an increase of the tryptophan/BCAA ratio after exercise [36]. The decrease in the BCAAs blood serum levels and the increase of the tryptophan level after jumper exercise could increase the rate of passage of tryptophan across the blood–brain barrier because BCAAs and tryptophan compete for the same transporter [37,38,39]. Thus, enhanced entry of tryptophan into the brain would lead to increased serotonin levels, which ultimately leads to anticipated central fatigue [36]. The changes in the values of peripheral modulators of serotoninergic function found in competing horses after exercise showed a relationship with the considered indices of exercise performance, as emphasized by the positive correlation found between the serum levels of tryptophan, Try/BCCAs, and NEFAs, and the values of blood lactate, HR, and RT. On the other hand, the amino acids leucine and valine showed a negative correlation with the values of blood lactate, HR, and RT measured in competing horses affect exercise. Particularly noteworthy, the peripheral modulators of serotoninergic function were negatively correlated with dopamine and positively correlated with prolactin, suggesting a possible linkage among the central fatigue and the neurohumoral factors. Dopamine, acting on specific brain areas, is related to the motivation for locomotion [23]. Exercise may elevate the global dopamine level in the brain without affecting cerebral release via the jugular venous blood, because polar catecholamines do not readily penetrate the blood–brain barrier [40,41], and maybe only dopamine released into the hypophysial portal blood will appear in the jugular blood [42]. However, in healthy subjects, the brain demonstrates noradrenaline spillover into both the major and the minor jugular vein [43], and, although the passage of monoamines from the bloodstream to the brain is restricted by the blood–brain barrier, the existence of a barrier to movement in the opposite direction is less certain [44]. It has been suggested that the main source of dopamine in circulating blood could be the activity of the adrenal medulla [45]. Although what occurs in the central nervous system is known, little information is available from the literature about the effect of exercise on dopamine levels at the peripheral level. The decreased dopamine values found in the present study after the competition conclusion could be due to the transition of circulating dopamine in the spleen, likely to occur during a stressful condition [46]. In the spleen, the noradrenergic sympathetic fibers can take up circulating dopamine, which could then be released on sympathetic activation to be taken up in turn by leukocytes through active transport [47]. The finding of the current study showed an increase of prolactin levels after jumping competition that, hypothetically, could be due to an immediate antagonism of hypothalamic dopamine input to the pituitary, or to some other stimulation factors related to the hypothalamic–hypophysial portal system [18,48]. Thus, it could be assumed that the higher prolactin values recorded after competition are related to the superimposition of stimulatory input of prolactin-releasing factors, such as thyrotrophin-releasing hormone, arginine vasopressin, vasoactive intestinal peptide, and serotonin. Particularly noteworthy, the values of prolactin were positively correlated with the fitness and stress indices, including HR, blood lactate, and RT, recorded from competing horses; however, whether this hormone is involved in the increased perception of effort and the earlier onset of fatigue during exercise remains to be determined. It has been proposed that exercise-induced hyperprolactinemia was related to changes in the peripheral modulators of serotonergic function in human species [49]. However, studies on the possible relationship of prolactin secretion and peripheral modulators of serotoninergic function during exercise showed conflicting results. Fischer et al. [50] found an increase in prolactin concentration during exercise in proportion to the rise in the plasma free tryptophan concentration. On the other hand, prolactin secretion during exercise has subsequently been reported to be unrelated to the peripheral modulators of serotoninergic function [51].5. ConclusionsTo the best of the authors’ knowledge, this is the first study investigating the effect of jumping exercise on peripheral modulators and indices of serotoninergic function and their relationship to exercise performance in horses competing in an official jumping class. The findings provide indirect evidence that the serotoninergic system may be involved in fatigue during jumping exercise under a stressful situation, such as competition, in which, in addition to physical effort, athletic horses experience emotional reactivity and mental stress. Although a relationship among peripheral modulators, neurohormones, and indices of fitness seems to exist during jumping exercise in horses, further studies are demanded to determine whether this linkage is responsible for the increased perception of effort and the earlier onset of fatigue during both training and competition in athletic horses. This would contribute to an improvement of the knowledge on horse’s response to jumping effort, and allow a better understanding of how a horse is coping with its training exercise or passive behavior during a competition, which will in turn contribute to improving animal welfare and athletic performance.

animals : an open access journal from mdpi

10.3390/ani13101666

PMC10215638

Blended learning strategies are becoming more popular in health sciences teaching, including veterinary medicine. However, the use of these methods in practical classes is less commonly described. This paper outlines a blended learning approach using flipped classrooms, collaborative learning, and gamification in gross anatomy practicals for first-year veterinary students. The practicals were restructured to include video pre-viewing, quizzes, collaborative group work, and a card game. Results showed a small but significant improvement in practical anatomy exam scores for the locomotor apparatus, with no significant difference in organic systems exams. Student satisfaction was high, with positive feedback regarding motivation, stimulation, and learning with peers. This study demonstrates that blended learning in anatomy practicals, including a flipped classroom and gamification, can enhance the student learning experience and optimize the limited time available for practicals.

The use of blended learning strategies is increasingly common in health sciences, including veterinary medicine; however, there are very few descriptions of these methods being applied to practicals. We describe here the application of blended learning based on the implementation of flipped classrooms with collaborative learning and gamification to the 2020–2021 veterinary medicine gross anatomy practicals at CEU Cardenal Herrera University (Spain). Students prepared for the sessions by pre-viewing videos and taking a quiz before the start. The sessions were conducted in small groups where students learned through collaborative work and reviewed their learning with a card game. A small but significant increase was observed when comparing the scores of practical exams of the locomotor apparatus with those of 2018–2019 (6.79 ± 2.22 vs. 6.38 ± 2.24, p < 0.05), while the scores were similar (7.76 ± 1.99 vs. 7.64 ± 1.92) for the organ system exams. Students’ responses in a satisfaction survey were mostly positive (>80%) regarding the motivating and learning-facilitating effect of this educational method. Our work shows that the application of blended learning in anatomy practicals based on a flipped classroom and with elements of gamification and collaborative work can be an effective way to improve the learning experience of students.

1. IntroductionHigher education in health sciences disciplines is constantly evolving. Advances in technology and the application of new pedagogical approaches are modifying university teaching. In recent years, the emergence of COVID-19 has led to a rethinking of the ways of teaching and to the development of new pedagogical approaches seeking to facilitate the improvement of students’ learning, giving rise to experiences that could then be applied in the post-COVID era [1]. In this context, blended learning (BL) has emerged as a strategy that uses different pedagogical approaches with the aim of modifying some aspects of how students learn to improve their learning experience [2,3,4,5]. The use of combined strategies in anatomy teaching has increased in recent years, with some studies showing the positive effects of using different methodologies [6]. This use of a diversity of e-learning resources, such as quizzes, anatomical demonstrations, computational 3D models, medical imaging, and dissection activities, has been shown to be effective in consolidating anatomical knowledge [7].It is generally accepted that student-centered approaches and dynamic teaching methods facilitate learning. The flipped classroom (FC), collaborative learning, and gamification have emerged as three methodologies based on these principles, which facilitate active learning and student engagement and ultimately improve student learning [8].In the FC, the ordinary roles in the learning process are reversed: the teacher acts only as a guide, and the student is the main actor. Flipped teaching is a student-centered approach where information is introduced outside the classroom, allowing more time during class to process the information and practice the content with a variety of active learning strategies [9,10]. The FC provides an opportunity for students to watch and study the course materials on their own time prior to the actual class; once they arrive, they are ready to apply the new knowledge [2,11].In addition to the FC, peer-assisted or collaborative learning, in which students help each other and learn by teaching, has proven effective in undergraduate medical programmes [12]. More specifically, in anatomy courses, collaborative learning promotes engagement and improves the study and group-working skills of students, moving them from passive consumers to active participants [13].Gamification experiences, or the use of game design elements in non-game environments to enhance academic performance, have also been used in different health sciences disciplines, reporting generally positive results and improving learning behaviours and learning outcomes [14].Different BL approaches have been used in anatomy courses, and there is considerable debate about to what extent and how these new methods should be applied and can help to improve students’ learning [7,15,16]. A recent systematic review concludes that the FC is often perceived positively by students and has some advantages, such as increased self-confidence, academic performance, interest, learning activity, interaction, satisfaction, and overall performance in the anatomy curriculum [17]. In relation to collaborative learning, several team learning approaches have been proposed for teaching anatomy [18], and different types of gamification experiences have also been described [19]. In our own experience, the development of a simple card game as a means to consolidate content has been well received by students, who valued the game as enjoyable, challenging, useful for improving their knowledge and understanding of clinical anatomy, and effective for anatomy exam preparation [20].In this work, we describe a BL teaching experience that combines FC, collaborative learning, and gamification applied to practical sessions of gross anatomy for veterinary students. Similar to in the medical sciences, the use of active learning methods is increasing in veterinary education. A recent international study involving universities in the USA, United Kingdom, and Australia showed that 95% of participants were familiar with the FC technique, although fewer (64%) used it in their teaching [21]. Several studies have published different models of active or BL in veterinary medicine, most of them focusing on clinical or preclinical subjects and generally reporting positive results [22,23]; however, little has been reported in relation to practicals in veterinary anatomy.With this background, our hypothesis was that introducing elements of flipped learning (so that students had to prepare the content of the practice beforehand), collaborative learning (so that the development of the sessions was cooperative), and gamification (applying a card game that would allow them to review the fundamental aspects of the sessions) to the practicals of veterinary gross anatomy would have positive effects on the learning experience of the students.2. Materials and Methods2.1. Anatomy in the Veterinary Medicine CurriculumThe teaching of anatomy subjects at the CEU Cardenal Herrera University (Valencia, Spain) is divided into three modules, called Structure and Function I, II, and III, during the first three semesters of the Veterinary Degree. The students involved in this study are those enrolled in Structure and Function (SF) I (locomotor apparatus) and II (organ systems), both in the first academic year. The practical sessions of gross anatomy in SF I in the 2020–2021 academic year included 4 sessions of comparative osteology of domestic mammals and 3 sessions of prosections on dog cadavers to study the locomotor apparatus. In SF II, 2 sessions of prosections on dog cadavers (one for the organs of the thoracic cavity and one for the abdominal and pelvic cavities) and 3 sessions of comparative anatomy of organs were carried out. The duration of each session was 2 h for both subjects. A practical exam consisting of identifying a series of anatomical structures was held at the end of each course of SF I (in December) and II (in May).2.2. Participants and ProcedureStudents enrolled in the 1st year of veterinary medicine during the academic year 2020/2021 were included in this study. There were 348 students in SF I and 325 in SF II. The students were divided into permanent and stable groups of 12–14 people for each practical. Exclusion criteria: students who did not take the practical examination were excluded from this study.The study was approved by the Vice-Rectorate for Academic Organisation and Teaching Staff of the CEU Cardenal Herrera University, Ref: PI40A-VV-20.2.3. VideosVideos showing the anatomical structures to be studied in each session were recorded by the instructors and made available to the students as a basis for preparing the practicals. The videos ranged in length from 3 min to 12′ 30″. They included a description of the main anatomical details in the case of bones and viscera and explained the regional anatomy of the dog on prosections from previously dissected animals. The number of videos included in each practical was variable (between 2 and 5) depending on the content to be covered in each session. Videos were provided for all SF I and SEF II practicals. Students were encouraged to watch the videos prior to the practical as part of their learning process.2.4. Pre-Practical TestTwelve self-assessment tests were designed (one for each practical session, i.e., seven for SF I and five for SF II). The pre-practical tests were in multiple choice question format (with 4 possible answers, only 1 correct) and contained 5 questions on pictures of the structures explained in the videos (see supplementary material S1 as an example). They were provided via the Internet approximately one week before each practical. Students answered the questionnaire before doing the practical, and they had feedback on their answers immediately after taking the test. Access to the practical was also allowed if the student had not carried out the activity or watched the video.2.5. Anatomy Cards GameAt the end of the practice, the students used a card game with different questions about the practical. Basically, the game consisted of answering questions related to the day’s practical. These questions were organised in 4 groups and appeared on a series of previously prepared cards [20].2.6. Teaching and Learning Approach2.6.1. Before the PracticalAt least one week in advance of each practical, the guide and the videos related to the structures to be identified were made available to the students via the Internet. Students had to study the guide and watch the videos, and then, they had to complete a five-question test via the Internet before the start of the practical. Prior to the implementation of this methodology in 2020–2021, students were provided, as support material, with the guide and a collection of videos that only partially covered the practical content, and no pre-practical quizzes were made available to them. The completion of these tests was taken into account as a component in the assessment of participation in the proposed activities (10% of the total value in SFI and 5% in SF2). The students were allowed to enter the practical regardless of whether or not they had carried out the activity and watched the videos.2.6.2. During the PracticalIn each practical, the students of each stable group (consisting of 12–14 students) were freely subgrouped into small groups of 3–4 students per dissection table to carry out the session (identification and recognition in situ of the corresponding structures). Each session lasted 2 h. Before the application of this methodology in 2020–2021, the teacher explained the structures to be identified for approximately 1 h, and the students used the remaining hour to work on the cadaver or on the models (bone pieces, viscera). In this new approach, students work on the cadaver or models from the beginning, using approximately 1 h 30 min to identify the structures and 30 min to play the self-assessment and knowledge consolidation card game. If the students were not able to identify the structures or had any doubts, an instructor was available to answer their questions. From time to time, this instructor also asked questions directly to the students to verify that the learning was being carried out correctly.2.7. Practical ExaminationAt the end of each SF I (in December) and II (in May), an oral practical examination was held to assess the knowledge acquired. In the examination session, which was invigilated, the student had to identify 15 anatomical structures. There were 15 sample exams which were randomly distributed among the students at the time of the test. The student, having identified the structure, showed it to the examiner. They had a maximum time of 15 min to complete the test. The exams were similar in the academic years analyzed (2020–2021 and 2018–2019) and carried out by the same examiners in both years.2.8. QuestionnaireAt the end of the academic year, a satisfaction survey was presented to the students to find out their level of satisfaction with this teaching system. The purpose of the survey was explained to the students, and they were informed that the data would be used for academic and research purposes only. Participants were also informed that answering the questionnaire was voluntary and that responses were collected anonymously. The main objective of the survey was to determine the overall impact of this methodology on the attitude and learning experience of the students. The questions were designed in such a way that students could assess the effectiveness of the organisation of this course in their learning process. Students were asked about the effectiveness of the pre-recorded videos, the help of other students, the effectiveness in preparing for the practical exam, and about the general consideration of anatomy. For each question, students could choose their level of satisfaction between “strongly agree”, “agree”, “indifferent”, “disagree”, and “strongly disagree”. Students were also given the possibility to express their opinion in qualitative responses. No demographic data were collected. This questionnaire has a Cronbach’s α of 0.77.The instructors (n = 7) also completed a questionnaire asking about the effect of this methodology on improving students’ accountability, the appropriateness of following up on students’ previous work, improving students’ learning, the importance of videos, and the effect on student–teacher interaction. This questionnaire has a Cronbach’s α of 0.60.2.9. Outcomes ComparisonThe results of the practical exams obtained by the students were compared with those obtained in the same examinations by students in the 2018/19 academic year to analyse the academic outcomes of this experience. Teaching and examinations were conducted by the same instructors in 2018–2019 and 2020–2021. The academic year 2019/20 was not considered as it was affected by the special measures for COVID-19.2.10. StatisticsAnalysis of data included variable scale parameters as mean and standard deviation. Data followed a normal distribution as calculated by Kolmogorov–Smirnov and Shapiro–Wilk normality tests and were compared using Student’s t-test. Pearson’s correlation coefficient was calculated to determine the statistical relationship between variables. p < 0.05 was considered statistically significant. All analyses were performed using R statistical software, version 4.2.3 for Windows (R Core Team, 2023).3. Results3.1. Students’ Engagement and Academic Results in Locomotor Apparatus PracticalsWhen analysing the students’ completion of quizzes prior to the practical, results showed that almost all students completed the questionnaires. In total, 74% of the 1803 proposed quizzes were completed. The number of students who completed the questionnaires, which was fairly constant (between 75% and 80%) for the first ones, decreased slightly towards the middle of the course and markedly in the last quiz proposed, which was completed by 61% of them (Figure 1).Almost half (n = 163, 47%) of the students took all the quizzes and 20% (n = 68) only left one undone. On the opposite side, 48 students (14%) did none (Table 1).We then examined the correlation between having completed the quizzes and the practical exam marks. The students’ final score in the anatomy practical exam was 6.79 ± 2.22 out of 10. We found a direct and discreet relationship (0.29, p < 0.001) between the number of quizzes completed by the students and their marks in the practical exam. We then analysed whether the opposite groups of students, i.e., those who had completed all the quizzes and those who had completed none of them, differed in their marks in the practical exam. Students who had completed all seven quizzes obtained a mean score of 7.67 ± 1.98 compared to those who had completed none, who scored 5.98 ± 1.98. This difference was very highly significant (p < 0.001).Comparing these results with those obtained in the academic year 2018–2019 (the 2019–2020 academic year was not considered due to the special conditions of its teaching by the appearance of COVID-19), we found that academic results in 2020–2021 (6.79 ± 2.22) were slightly but significantly higher than in 2018–2019 (6.38 ± 2.24; n = 353, p < 0.05).3.2. Students’ Engagement and Academic Results in Organ System PracticalsIn total, 78% of the 1273 questionnaires related to organ systems were completed by students. The first two questionnaires were completed by more than 80% of the students, and thereafter there was a very slight but steady decline in the number of students taking the tests (Figure 2).The number of students completing the questionnaires in the organ systems practicals was slightly higher than in the locomotor system practicals, with more than half of the students (n = 178, 55%) completing all the questionnaires and only 23 (7%) not completing any (Table 2).The students’ score in the final organ systems practical examination was 7.76 ± 1.99. A positive correlation (0.33, p < 0.001) was observed between the number of quizzes completed by the students and their marks in the practical exam. Comparing the scores of students who had completed all quizzes (8.31 ± 1.81) with those who had completed none of them (6.23 ± 2.04), a very highly significant difference was again found (p < 0.001). However, no difference was found, in this case, when comparing the marks for the organ systems practical exam in the year of this study (7.76 ± 1.99) and in 2018–2019, wherein students obtained 7.64 ± 1.92 (n = 372).We next compared the percentage of students with low performance (those obtaining less than 5 out of 10 in the practical exam) in both SF I and SF II. The percentage of students with low performance was reduced in SF I (19% in 2020–2021 vs. 29% in 2018–2019), while in SF II, there was no change (10% in 2020–2021 vs. 8% in 2018–2019).3.3. Students’ and Instructors’ Satisfaction SurveyA satisfaction survey was provided to the students at the end of the course, which was completed by 167 students. The majority of students answered positively to the proposed questions, as more than 80% of respondents agree or strongly agree with most of the proposed questions, with “My anatomy interest has increased after practical sessions” being the highest percentage (95%) and “Videos helped improving my anatomy learning and knowledge” the lowest (72%) (Figure 3).The students also had the opportunity to give their opinion, and 78 of them expressed their qualitative comments on the system. Most of them were positive, but in some cases, they suggested modifications in the method or in the quality of the materials provided. Some examples of the students’ opinions are as follows:“It would be nice to have access to the videos and tests from the beginning of the course and have the ability to do the tests multiple times”.“I think videos are great bonus tools, as well as doing a test before each practical. But i prefer very much to be taught by anatomy teachers during the practical itself. So basically, watch vids, do the quizzes and then be taught by the teachers during practicals to avoid confusion. Even though they’ve always been available in case of doubts;)”.“I would like to say that without the videos, I would never have obtained the grade I had at the anatomy practical exam. I really appreciate the videos but I can say that some of them didn’t really follow the guide. The cards are really good to train ourselves during anatomy sessions. It would be great that they last longer and students helping us during the practicals is also very useful.”Additionally, a questionnaire was also completed by the teaching staff (n = 7). The results indicate that, in the opinion of the staff, this methodology is particularly positive for improving the personal responsibility of the students and also highlights the convenience of checking the students’ previous work with questionnaires before doing the practicals. On the other hand, although mostly positive, there was a diversity of opinions as to whether the videos can replace the direct explanation of the content by the teacher and whether the system increases teacher–student interaction (Figure 4).4. Discussion4.1. The Usefulness of Blended Learning in the Teaching of Anatomy and the Academic Outcomes of the StudentsMany different approaches are used to teach anatomy, and no single teaching tool has been found that meets all curriculum requirements [24,25,26]. In this regard, the application of new teaching methodologies has been described for both medical and veterinary anatomy, and ultimately, all the different changes that have occurred in recent years, including those resulting from the emergence of COVID-19, have promoted the need for adaptation and the development of new learning models [27,28,29,30].Among the multiple forms of training, some authors consider that the best way to teach anatomy is by combining multiple pedagogical resources that complement each other, maximising interactive learning, and using time-saving materials that help students learn gross anatomy effectively and efficiently [29,30,31].We describe here the transformation of traditional hands-on teaching into a BL approach in veterinary anatomy practicals. This change basically consisted of transforming sessions where a large part of the time was occupied by teacher explanation into active learning sessions, integrating flipped teaching, collaborative learning, and gamification. The results indicate that this switch was highly appreciated by the students and had some beneficial effect on academic results.Different methods can make up a BL experience in anatomy. In our study, the FC is the central element, which is closely linked to students’ cooperative work throughout the sessions and to a gamification element consisting of a card game designed to self-assess and reinforce knowledge [20]. Our results are in line with those of other authors who have reported the beneficial effects of BL in human anatomy teaching and show that active learning strategies are appreciated by students. Medical students preferred the hybrid, interactive, and student-centred learning method for practicals as a more effective system for understanding anatomy than face-to-face or online methods [32]. In another work, students appreciated the pre-class activities of this kind of learning, as they facilitated their engagement and concentration during class, thus increasing their academic performance and engagement [33]. In this regard, it has also been shown that veterinary anatomy students also favour active learning classes compared to traditional classes [34].4.2. The Implementation and Usefulness of Flipped ClassroomPre-class work is essential in FC pedagogies. Preparing courses by studying the material provided allows for better use of class time, creating time for integrative group activities during the session [35]. In this model, students have more responsibility for each laboratory experience, which can lead to more effort, but also to more satisfaction and confidence. While most FC approaches are applied to lecture-based courses, our results show that FC could also be applied to practicals. In our study, most of the students adequately prepared for the practicals before attending the sessions. In the preparation of the sessions, the use of videos is very important. Video as a support material for anatomy lectures is considered a very useful resource in a condensed curriculum scenario [36,37], and we have shown that their use can help decrease anxiety and enhance students’ learning experience [38]. In our study, students considered these videos as important elements in the preparation of the practicals, although, in some cases, students requested more detailed documents. This would explain why the degree of satisfaction was not as high as in other questions of the survey. Although we do not have data on the video visualizations, the fact that many students completed the quizzes before the practicals suggests that most of the students prepared themselves thoroughly before attending the practicals. The number of completed quizzes exceeded 75% in the first sessions, but there was a 15–20% reduction in the number of tests completed towards the end of the semester. This decrease is consistent with Greene’s [37] study, which showed a significant drop over the year in the total viewing of videos provided to review structures identified in lab sessions throughout an organ systems course. Likely, the decrease in completion of the quizzes was due to the feeling that time was not sufficient at the end of each semester, as reported in other studies [36,37,39]. At the end of the semester, there is generally a greater workload for the student; obviously, this depends on the organisation of the curriculum, but it is common for these dates to coincide with imminent exams, the preparation of which requires extra time and creates stress for students that can cause them to drop some activities. It is interesting to note that there is a correlation, though not very high, between having carried out these activities and academic outcomes. However, if we compare the results obtained by the students who completed all the tests with those who did not do any, the differences are highly significant in both topics, indicating the effectiveness of active involvement in this type of teaching.Different results have been published on the effect of FC on academic scores in several disciplines of health sciences. Most of the studies report a slight improvement in test scores, while others show neutral results. Tune et al. [40] stated improvements of around 11–12% in the results of cardiovascular, respiratory, and renal physiology tests. Street and colleagues (2015) described a slight improvement, at the limits of the statistical significance, in a preclinical physiology course [41]. Better results after the implementation of FC have also been reported in practical examinations of dental students [42], in a pharmacotherapy course for pharmacy students [43], in an advanced physiology course for pharmacy students [44], and in physiology courses for health sciences students [45].In other cases, the development of FC did not produce improvements in the performance of the students. No differences in grades or level of satisfaction were shown in an FC for neuroanatomy [46], and learning gains were not significantly different between a flipped biochemistry course and a traditional lecture [47]. However, in this last report, student perception of learning gains did differ and indicated a higher level of satisfaction with the flipped lecture format [47]. Indeed, students’ perceptions about the effectiveness of FC are not necessarily accompanied by improvement in academic performance. This was also shown in a study performed in an introductory epidemiology class on masters-level graduate students [2].A partial improvement has also been reported. Pharmacy students’ performance in a flipped teaching pilot on cardiac arrhythmias improved in two of the three classes [48], and in a preclinical course (foundations on animal health) of veterinary medicine, FC cohorts performed well in the final exam but not in multiple-choice tests [23].Similar results have been reported in studies specifically related to anatomy. A recent review shows that most FC experiences in anatomy report better exam scores than traditional courses, although there is also some work reporting a neutral effect [17]. Elzainy and Sadik [49] reported that FC medical students scored higher than traditional classroom students in anatomy theory classes and had lower absenteeism. Moreover, a small but significant improvement has been demonstrated in the histology exam scores of FC students compared to the control group [50], and nursing students instructed with the FC method on anatomy and physiology of the respiratory system scored higher on the final exam than the control group [51]. Finally, results reported by Yang and colleagues [52] and Elzainy and Sadik [49] on different human anatomy experiences for medical students pointed in the same direction.As noted above, some anatomy-specific studies also describe partial improvements or neutral results from the application of FC. Fleagle et al. reported that there was no change in mean scores on the first and second lab exams, but higher scores on the final lab exam for dental gross anatomy students [35], and in another study, FC students also performed better than lecture hall students on analysis items, but there was no difference in performance between students in the two groups for knowledge [53]. Our results are close to the latter works, as the students in the experience we present here obtained a slight but statistically significant improvement in the practical examination of the anatomy of the locomotor system, while there was no effect in the examination of the anatomy of organ systems and body cavities.4.3. The Role of Collaborative Learning and Gamification in a Model of FCFC and peer learning are closely related since FC provides more opportunities for students to interact effectively and learn from their peers [2]. It has been shown that cooperative learning deepens student understanding but also offers them an opportunity to practice the generic skills in veterinary anatomy dissection courses, enhancing students’ comprehension of the subject matter and encouraging them to share responsibility for their learning [54]. Peer-assisted learning also had positive results in a study on medical students, which appreciated most aspects of this method and obtained higher scores in osteology [55]. In this sense, it has been shown that combining flipped teaching and team-based learning results in positive outcomes in a physiology course [44]. On the basis of a different experimental design, Laakkonen and Muukkonnen [56] also pointed out the importance of collaboration between students to support mutual learning in gross anatomy courses of veterinary medicine. Among the diversity of peer-assisted learning methods, our model is based on collaborative work between students, where, organised into small groups, they received help from other students who solved their doubts in the first instance and in reviewing the content. In fact, this feature was one of the most valued aspects by the students, emphasizing the importance of collaborative work in this type of teaching. Interestingly, a positive outcome has been reported when students work cooperatively rather than when peer interaction was absent [9]; it has also been shown that the combination of e-learning, peer teaching, and FC significantly improved the average score in a physiology laboratory course for nursing students [57].The development of the card game that ends each practical session in our model was also carried out in a collaborative environment [20] in a manner wherein everyone learns, regardless of the outcome of the game. In this sense, positive student outcomes have been reported in different gamification experiences in anatomy and related disciplines [58,59], reinforcing the positive effects of cooperative learning. Similar to other active approaches, the use of game-based methods is increasing in anatomy, and although the research concerning game-based anatomy learning is relatively limited, gamification seems to have a positive impact on different aspects of students’ learning [16].4.4. The Students and Staff Perceptions of Active LearningBeyond the academic results, one of the advantages of the different BL models is that they are appreciated by the students as they improve their learning experience. In our work, as previously indicated, students rated the interaction with other students very positively and found the system motivating and challenging, and it increased their interest in anatomy. In the flipped model, students have more responsibility, which can increase effort but also satisfaction and confidence [35]. Active learning in anatomy demonstrations has been shown to be a preferred method by students [32], and perception surveys revealed students’ enthusiasm for pre-class activities, leading to better performance in class with increased engagement with peers and teachers [60]. The positive perception of students toward active learning has been highlighted in numerous works ([60] and references therein). These results are also consistent with those shown in a systematic review showing that FC is generally perceived positively by students, increasing self-confidence, academic performance, interest, learning activity, interaction, satisfaction, and overall performance in the anatomy curriculum [17]. It is possible that this new system would entail a little more workload. This aspect was not explored, but the overall high degree of student satisfaction suggests that it was not an excessive effort.Regarding the opinion of the staff, most of them were positive about the benefits of this method, since they all considered that it improves personal responsibility and ultimately student learning, but it is also important to point out the convenience of monitoring the work as a means of control of the preparation of the practical. This type of opinion is not always generalizable, and it has been described that, sometimes, teachers are more reluctant to implement this type of teaching than students. In this regard, it has been described that teachers sometimes consider that students do not prefer FC to traditional teaching and are not sure of the students’ response to FC, in contrast to the positive views about this method in the student satisfaction survey [33].4.5. Limitations of the StudyThis teaching system was installed in the anatomy practicals for all students as a consensual teaching decision accepted by the University, so there is no control group in the same year. For this reason, the comparisons in the academic results have been made with those of the 2018–2019 academic year, the last one in which the presence of COVID-19 did not alter the teaching strategy.On the other hand, the organisation of the teaching did not make it possible to follow the viewing of the videos.Finally, a demographic study was not carried out, so it is not possible to know whether such factors influenced the results.5. ConclusionsOur study shows that the development of an FC system in veterinary anatomy practices can have a beneficial effect on academic results. Moreover, it is well appreciated by students who positively value cooperation with other students to solve problems and consider this a highly motivating methodology. While most FC experiences are applied to lectures, our approach is aimed at practical gross anatomy classes. According to our experience, the application of this type of teaching organisation allows for better use of time, which is increasingly reduced in many anatomy curricula, and provides a good teaching experience for students and staff. Learning based on FC with a strong cooperative component among the students and completed with some gaming element, such as the work that we present here, can be applied, with the necessary adaptations, in a general way, both in veterinary medicine as well as in the different disciplines of health sciences that involve the practice of veterinary and human anatomy.

animals : an open access journal from mdpi

10.3390/ani11113014

PMC8614311

We first report the prevalence of Enterocytozoon bieneusi in Holstein Cows and dairy buffalo in Yunnan province of China, with a percentage of positivity of 0.59% (5/841). Two novel zoonotic potential genotypes of E. bieneusi were found. We analyzed the different influencing factors (region, season, variety, breeding mode, gender, age), but the differences were not statistically significant.

Enterocytozoon bieneusi is a fungus-like protist parasite that can cause diarrhea and enteric diseases. The infection of E. bieneusi has been reported in many host species, including cattle and humans. However, information on prevalence and genotype distribution of E. bieneusi in dairy cattle in Yunnan province in China is still absent. In this study, 490 Holstein Cows and 351 dairy buffalo fecal samples were collected from three regions in Yunnan province, China. By using nest-PCR that targets the internal transcribed spacer (ITS), we found that the prevalence of E. bieneusi was 0.59% (5/841). DNA sequence analysis showed that five E. bieneusi genotypes were identified in this study, including two novel genotypes, YNDCEB-90 and YNDCEB-174, and three known genotypes (I, J, BEB4). Phylogenetic analysis revealed that two novel genotypes, YNDCEB-90 and YNDCEB-174, were clustered into Group 1, representing the zoonotic potential. The remaining genotypes I, J, and BEB4, which are the most frequent genotypes of E. bieneusi infection in cattle and lead to E. bieneusi infection in humans, belonged to Group 2. Although the lower prevalence of E. bieneusi was detected in dairy cattle in Yunnan province, it indicates that dairy cattle should be considered to be one of the potential hosts for transmitting E. bieneusi to humans. These findings are important for the development of effective prevention strategies for microsporidiosis.

1. IntroductionEnterocytozoon bieneusi belongs to microsporidial species. It is an enteric unicellular microsporidian parasite that can infect invertebrate and vertebrate hosts worldwide [1]. The phylum microsporidia consists of more than 200 microsporidian genera with more than 1500 species [2,3]. E. bieneusi is the most prevalent pathogen in human beings and various mammals [4,5,6,7], causing symptomatic and asymptomatic intestinal infections through accidental ingestion of food or water that has been contaminated with the viable spores of E. bieneusi. E. bieneusi is also deemed as one of the riskiest opportunistic pathogens for patients with HIV/AIDS [8,9,10].It is difficult to distinguish E. bieneusi from other microsporidia species by conventional staining methods due to the extremely small size of the spore. Currently, the most efficient method to characterize E. bieneusi is using PCR and DNA sequencing of the ribosomal internal transcribed spacer (ITS) [11]. The first cases of E. bieneusi infection in cattle were found in Germany, and two genotypes, I and J, were identified [12]. The prevalence of E. bieneusi in cattle ranged from 1.13% to 46.8% in China, the United States, Argentina, South Korea, Germany, Portugal, South Africa, Brazil, and the Czech Republic [12,13,14,15,16,17,18,19,20]. Although most of E. bieneusi genotypes are detected in cattle and belong to Group 2, these genotypes have also been identified in humans, animals, drinking water sources, and municipal sewage, indicating the possibility of cross-species or zoonotic E. bieneusi transmission [21,22].Yunnan province has unique climatic characteristics, which include subtropical and tropical monsoon climates, and the minimum monthly average temperature is above 0 °C. This environment is suitable for the reproduction of intestinal protozoa, e.g., E. bieneusi. E. bieneusi affects not only the development of the livestock breeding industry, leading to the decline in production performance and loss of economic benefits, but it also pollutes water, soil, and food, causing food and waterborne outbreaks [23]. However, the information on prevalence and genotype distribution of E. bieneusi in dairy cattle in Yunnan province is still absent. In the present study, we use PCR and DNA sequencing to screen E. bieneusi infection in dairy cattle of Yunnan province and assess the risk factors for prevalence of E. bieneusi, such as geographic region, season, variety, breeding mode, gender, and age. The results of this study revealed that the infected animals can be identified as potential sources of E. bieneusi infection between dairy cattle and humans.2. Materials and Methods2.1. Collection of SpecimensEight hundred forty-one fecal samples of dairy cattle were collected from June 2019 to August 2020 on 15 farms in Kunming city (n = 248), Dali city (n = 357), and Tengchong county (n = 236) in Yunnan province, China. The dairy cattle were divided into six groups: region group, season group, species group, breeding mode group, gender group, and age group. All the fecal samples were collected individually using sterile gloves and stored in a refrigerator at 4 °C until DNA extraction.2.2. DNA Extraction and PCR AmplificationThe 10 g of fresh fecal specimens of each dairy cattle was washed twice with distilled water to remove the impurities. Genomic DNA was extracted from 200 mg of fecal samples using an EZNAR stool DNA Kit (OMEGA, Biotek Inc. Norcross, GA, USA) according to the manufacturer’s recommendations. E. bieneusi infection was screened by nested PCR amplification of the ribosomal internal transcribed spacer (ITS) [24] and amplification of the loci MS1, MS3, MS4, and MS7 multilocus sequences to analyze the genotypes of E. bieneusi. Positive samples were further characterized by multilocus sequence typing analyses using loci MS1, MS3, MS4, and MS7 [25], and the amplification products were subjected to electrophoresis on 2% agarose gel to be observed.2.3. Statistical AnalysisThe P value, odds ratios (ORs), and their 95% confidence intervals (95%CIs) of variables, including region, season, variety, breeding model, gender, and age, were calculated using SPSS20.0 (IBM Corporation, Armonk, NY, USA) and SAS9.1 (SAS Institute Inc., Cary, NC, USA).2.4. Sequencing and PhylogenyThe positive-PCR products were sequenced by Sangon Biotech (Kunming, China). The sequencing accuracy was confirmed by using two-directional sequencing. All the obtained sequences were aligned to the reference sequences available from the GenBank database using the Basic Local Alignment Search Tool (BLAST) (http://www.ncbi.nlm.nih.gov/BLAST accessed on 17 January 2021), and computer program DNAMAN 6.0 (Lynnon Biosoft, San Ramon, CA, USA) was used to identify the genotypes of E. bieneusi. The phylogenetic tree was constructed for assessing the genetic relationship between the E. bieneusi genotypes using the Neighbor-joining (NJ) method with MEGA5 (MEGA, Auckland City, New Zealand). The reliability analysis of the evolutionary tree was estimated by the Bootstrap test, repeated 1000 times, and more than 95% was the threshold of significance. The novel ITS sequences of E. bieneusi isolates were submitted to GenBank with the accession numbers MZ229914.1 for YNDCEB-90 and MZ229915.1 for YNDCEB-174.3. ResultsIn this study, the global positive rate of E. bieneusi infection was 0.59% (5/841). As shown in Table 1, the E. bieneusi positive rates of the fecal specimens of dairy cattle in Kunming, Dali, and Tengchong were 0.81% (2/248), 0.56% (2/357), and 0.42% (1/236), respectively. There was no significant difference among the three regions (p = 0.856 > 0.05). The global infection rate of E. bieneusi in female cattle was 0.67% (5/751, 95%CI = 0.08–1.25), while no infection was found in male cattle. We further investigated the infection ratio as a seasonal variation. As shown in Table 1, the infection rate was 1.04% (1/96, 95%CI = 0.00–3.07) in autumn, 0.67% (4/599, 95%CI = 0.02–1.32) in summer, and no E. bieneusi infection was detected in winter. No significant difference was found in both the gender group (p = 0.774 >0.05) and season group (p = 0.535 > 0.05). Of the 841 analyzed samples, Holstein cows (positive ratio 0.61%, 3/490) showed higher infection rates than dairy buffalo (positive ratio 0.57%, 2/351), and there was no significant difference between the two varieties. To investigate whether the infection was associated with breeding modes, the cattle were classed into captivity group and grazing group, respectively. As shown in Table 1, the cattle raised in captivity showed an infection ratio of 1.44% (2/139, 95%CI = 0.00–3.42) while the cattle raised in grazing showed an infection ratio of 0.43% (3/702, 95%CI = 0.00–0.91). Both of the two breeding modes showed no significant difference (p = 0.156 > 0.05).The ITS ranges of positive samples were consistent with three known genotypes (I, J, BEB4) and two novel genotypes (YNDCEB-90, YNDCEB-174) which have 99.23% and 98.46% similarity with MK841506 (genotype J), respectively (Figure 1). Genotypes I and J were identified in Kunming. Genotypes BEB4 and YNDCEB-90 were identified in Dali. Genotype YNDCEB-174 was identified in Tengchong. The three known genotypes (I, J, and BEB4) were clustered into Group 2, and the two novel genotypes (YNDCEB-90, YNDCEB-174) were clustered into Group 1 (Figure 2). We identified three, zero, three, and zero types at the MS1, MS3, MS4, and MS7 loci, respectively.4. DiscussionE. bieneusi is an important pathogen of neonatal calf diarrhea (NCD) that can cause high morbidity and mortality in dairy cattle [26]. While there is no vaccine commercially available to prevent E. bieneusi infection in humans and animals, effective measures to control E. bieneusi prevent the transmission of these parasite spores that pollute food, drinking water, and soil.In this study, the overall prevalence of E. bieneusi in dairy cattle was 0.59% (5/841), which is lower than that of other provinces in China, such as Ningxia Hui autonomous region [13], Heilongjiang province [22], Henan province [27], Shaanxi province [28], Jiangsu province [29], Shandong province, Guangdong province, and Gansu province [30]. The different breeds, detection methods, geographical differences, and sample sizes are many factors that may contribute to the varying prevalence. Moreover, E. bieneusi was more likely to be found in autumn (odds ratio, OR = 1.56 [0.17, 14.16]) compared to summer, although the difference was not statistically significant (p > 0.05) (Table 1). Between the two breeds of cattle, the infection rate of E. bieneusi in Holstein cows was 0.61% (3/490), which was slightly higher than that of dairy buffalo 0.57% (2/351). The prevalence of E. bieneusi in captive cattle and free-ranging cattle did not differ significantly (p > 0.05); however, it was frequently higher (OR = 3.40 [0.56, 20.55]) in free-ranging cattle (Table 1). This result suggested that we should pay attention to strengthening the prevention of free-ranging cattle. Interestingly, in the present study, the pre- and post-weaned cattle were not found to be infected by E. bieneusi. The result was different in previous studies. We speculated that the difference could result from the better care that was provided to pre- and post-weaned cattle in our investigated area.In this study, five ITS genotypes of E. bieneusi were identified in dairy cattle in Yunnan province, including three known genotypes (I, J, BEB4) and two novel genotypes (YNDCEB-90, YNDCEB-174). The genotypes I (n = 1) and J (n = 1) were most prevalent in dairy cattle in the present study, which were similar to results reported in Ningxia Hui autonomous region and Henan province [13], Heilongjiang province [21], northeastern China [22], Shaanxi province [27], Jiangsu province [29], Shandong province [31], and Jilin province [32] in China. The prevalence of E. bieneusi in Yunnan province was also similar to that in the United States [14], Argentina [15], and other countries. These findings revealed that genotypes I and J are prevalent worldwide. In this study, we found that YNDCEB-90 has three single-nucleotide substitutions (SNPs), and YNDCEB-174 has SNPs at the ITS region (243 bp) in Figure 1. The ITS sequences of the two novel genotypes (YNDCEB-90, YNDCEB-174) have 99.23% and 98.46% similarity with MK841506 (genotype J), respectively. The detection of two novel genotypes suggested the possibility of genetic variations in E. bieneusi in dairy cattle in Yunnan province. The Phylogenetic tree analysis showed that two novel genotypes were classed into the category previously described as a zoonotic Group 1, and the three known genotypes (I, J, and BEB4) were clustered into Group 2 (Figure 2). All of those genotypes have been detected in humans [21]. Therefore, the dairy cattle may be potential hosts of human-pathogenic E. bieneusi. In this study, we only identified three, zero, three, and zero types at the MS1, MS3, MS4, and MS7 loci, respectively. The result may be that there was only a small amount of E. bieneusi in the fecal specimen, and the different PCR primers have different amplification efficiencies, leading to the identification of zero types at MS3 and MS7 loci.5. ConclusionsThe present study is the first report of the prevalence (0.59%, 5/841) of E. bieneusi infection in dairy cattle in Yunnan province, China. We demonstrated the three known ITS genotypes (I, J, BEB4) and two novel genotypes (YNDCEB-90, YNDCEB-174) that have zoonotic potential. The different influencing factors (such as region, season, variety, breeding mode, gender, and age) were not statistically significant. Since there is a deficiency of effective prophylactic and therapeutic strategies for E. bieneusi, and due to the existing threat posed by the fecal–oral transmission of E. bieneusi, molecular epidemiological investigation is required to be conducted in more animal hosts.

animals : an open access journal from mdpi

10.3390/ani13071273

PMC10093556

In this systematic review, we assessed studies on automatic monitoring of respiratory disease in livestock. This can help to understand if precision livestock farming (PLF) technologies are able to fulfill their purpose and can provide insights into the potential of commercially available PLF options. Few PLF technologies had good performance measures with tests under field conditions and using a reliable reference test, indicating that technologies still fall short for monitoring clinical signs or the onset of respiratory diseases in bovine, swine, and poultry productions. This review assessed previously published development and validation articles of PLF technologies to monitor respiratory health, and it highlights that more than just performance measures should be assessed when discussing the potential and pitfalls of PLF technologies.

Respiratory diseases commonly affect livestock species, negatively impacting animal’s productivity and welfare. The use of precision livestock farming (PLF) applied in respiratory disease detection has been developed for several species. The aim of this systematic review was to evaluate if PLF technologies can reliably monitor clinical signs or detect cases of respiratory diseases. A technology was considered reliable if high performance was achieved (sensitivity > 90% and specificity or precision > 90%) under field conditions and using a reliable reference test. Risk of bias was assessed, and only technologies tested in studies with low risk of bias were considered reliable. From 23 studies included—swine (13), poultry (6), and bovine (4) —only three complied with our reliability criteria; however, two of these were considered to have a high risk of bias. Thus, only one swine technology fully fit our criteria. Future studies should include field tests and use previously validated reference tests to assess technology’s performance. In conclusion, relying completely on PLF for monitoring respiratory diseases is still a challenge, though several technologies are promising, having high performance in field tests.

1. IntroductionWorldwide livestock respiratory diseases are highly prevalent, reducing productivity and increasing death risk [1]. In poultry production, animal-level prevalence ranges from 3% [2] to 49.3% [3] while flock-level respiratory diseases prevalence often reaches over 80% [4]. In swine, 38.5% [5] of animals are affected by this malady, and in cattle the prevalence ranges from 0.5% to 61% [1]. Mortality due to respiratory diseases is relatively high in livestock, reaching 47% in poultry [6], 4% in swine [5], and 19% in bovine [7,8].Many pathogens are responsible for causing respiratory diseases in livestock [9,10,11], but the clinical manifestation is similar in most cases. The main clinical signs are coughing, sneezing, lethargy, and nasal and ocular discharge [9,11].Precision Livestock Farming (PLF) is defined as automated continuous animal monitoring to aid with farm management [12]. PLF technologies are based on the use of sensors to monitor animals and have been developed for monitoring respiratory health in dairy calves [13], poultry [14], and swine [15]. These technologies can monitor respiratory diseases by monitoring signs, such as cough, sneezing, and fever or detecting specific types of diseases such as Newcastle disease and porcine wasting diseases. These technologies can detect the presence of diseases or clinical signs at the group level [15,16] or individual level [16].To assess if a technology has external validity (i.e., is effective to be implemented in commercial farms), several points should be observed: high performance measures and the validation conditions must be similar to the conditions in commercial farms [17].Performance measures include specificity (Sp), sensitivity (Se), precision, accuracy, positive predictive value (PPV), and negative predictive value (NPV). According to Dominiak and Kristensen [18], if the performance measures are not satisfactory there is no need to further assess the time window and validation conditions. The performance measurements ultimately arise from the results of a reference test (usually called a gold standard), which is used as a reference to assess if the developed technology was efficient in achieving its objectives [19]. Thus, when evaluating PLF development, it is critical to assess the adequacy of the chosen reference test.The study condition may determine if the performance is replicable in a real-life situation (i.e., commercial setting); therefore, the efficiency from technologies validated with field condition tests is likely to be similar to the efficiency on commercial farms [18].Performance measures may also be influenced by the type of data collected and the type of sensors applied. For example, sensors that are only capable of recording data related to nonspecific signs of disease, such as elevated body temperature, are not indicated for detecting a specific type of disease [20]. In addition, performance measures that consider the total number of cases (i.e., error rate, number of false positives for every true positive prediction) are biased by the true disease prevalence, with rarer diseases having higher error rates compared to diseases that are highly prevalent [18]. Therefore, critically combining all this information with performance measures allows one to assess technology readiness.There have been recent systematic reviews about PLF on disease monitoring [20,21,22,23]. However, these reviews focused on technologies that target many welfare-related issues at once (e.g., lameness, respiratory diseases, diarrhea). A more profound investigation of PLF devices developed exclusively for monitoring respiratory health status is needed to enhance the availability of information on the efficiency of these technologies. Understanding the effectiveness of a technology is much more complex than only looking at its performance. To support that the technology is reliable and able to perform well in real-life conditions, the study should present the study’s condition, reference test, and a full description of how the technology was tested.The objectives of this systematic review were to assess: (1) the advances of PLF for monitoring respiratory diseases at group or animal level in different livestock productions, (2) the quality of studies in reporting the effectiveness of the technology, and (3) if it is possible to rely on PLF to automatically monitor respiratory diseases in livestock species.2. Materials and Methods2.1. Literature SearchA systematic literature review was conducted to gather data on validated PLF technologies for the detection of respiratory diseases in livestock species.The PRISMA guideline [24] was used for the systematic literature search. The search was conducted in the Scopus, Web of Science, and IEEE databases on 21 September 2022. The search field in Web of Science was “topic”, and in Scopus was “article title, abstract, keywords”. Search terms are presented in Table 1.These terms were sorted as “species” terms, “technology” terms and “type of condition” terms. The search string consisted of a combination of terms separated by the Boolean operator “AND” and “OR”; for example: “Dairy Cow*” OR “Cow” AND “Precision Livestock Farming” OR “Non-invasive Technology” AND “Respiratory Disease*” OR “Cough”. To filter the search, in Web of Science and Scopus platform the option “Articles” and “Conference paper” was selected in “Document Types”.The inclusion criteria were original research studies published at any time about the development or validation of PLF technologies for monitoring respiratory disease in production animals (this included technologies aimed at monitoring clinical signs and technologies developed for disease detection). The exclusion criteria were studies that did not develop, validate, or apply PLF technologies to exclusively detect respiratory diseases or respiratory diseases clinical signs (i.e., research papers on general health monitoring or other disease were not included); this included research articles about the development or improvement of a specific technology’s feature, for example, localization of the cough event or differentiation between types of coughs were excluded. Studies of technologies developed for other species that are not present in the “species” terms, were also excluded.Two researchers were involved in the initial search and screening process (L.F.C. Garrido and S.T.M. Sato). The researchers worked independently on the search using the same terms and databases. The inclusion and exclusion of articles were also conducted independently by each researcher. The results of the screening process were compared after the search and inclusion or exclusion of articles. This step was conducted to ensure that all relevant studies were included. The screening started with the exclusion of duplicate studies. Secondly, studies that were not related to PLF were excluded by title and abstract. Then studies were screened following the inclusion and exclusion criteria after analyzing the title, abstract, and keywords [25]. The remaining studies were included or excluded after analyzing the entire content of the article. Conflicted studies were analyzed together by screeners and were included or excluded from the review after agreement if the study complied with the inclusion or exclusion criteria. Information about the number of excluded and included studies and the reason for excluding studies is presented in Figure 1.Three articles were excluded because we were unable to find the full text. One article was just available in Mandarin and therefore excluded to prevent mistakes due to inaccurate translation.All articles included in the review were used for assessing the advances of PLF technologies to monitor respiratory diseases.2.2. Data Gathered from ArticlesAfter the screening process, the following data were extracted from included articles: species the technology was applied in, conditions in which the study was conducted (laboratory or field condition: a study was considered performed in laboratory conditions when performed in a university or research center or conducted in a controlled environment not similar to a commercial farm; field conditions were considered when the study was conducted in a commercial farm, or field conditions were simulated in a controlled environment), the type of sensor/device used (sound-based, image-based, or any other sensor found in the literature), performance measures (Se, Sp, precision, accuracy, PPV, PNV, or any performance measure presented by the study, if recall was presented we considered it as Se), and the reference test used for validating the technology.The animal production stage in which these technologies are tested was also assessed. For swine production, we determined the production stage based on animals’ age or weight when the articles did not clarify the stage. Animals were considered in the fattening stage when animals were ten to sixteen weeks old and weighed between 30 and 60 kg. Animals with age above sixteen weeks and that weighed more than 60 kg were considered in the finishing stage. For bovines, the stage was considered the type of animal used (e.g., calves, steers) and poultry the age.Information regarding the production system and the aim of the production (e.g., egg, broiler, dairy, or meat production) was also gathered when presented by the article. For studies conducted in laboratory conditions, the production system was considered as an experimental setup.2.3. Risk of BiasWe assessed the risk of bias in studies included in this review. This process was held to understand the consistency of the results based on the study design [26]. A study is considered to have a low risk of bias when all information that could bias the results is presented. Similar to Stygar et al. [23] and Hendriks et al. [27], we considered that the article had to provide the following information to prevent bias: conditions in which the study was held (laboratory or field conditions), housing, type of sensor, and how it was installed (manufacturer of the sensor used, how many meters from the ground, where was it installed), the software used (what type of software, what machine learning method was used). In addition, we added three more pieces of information that had to be presented to assess bias; namely, the population description (age, weight or production stage), number of animals and raw data used to calculate performance measures. A study was categorized to have a high risk of bias if one or more essential information was not presented.2.4. Technology ReliabilityIn this review, the word “reliable” refers to the sole use of PLF technologies to detect respiratory diseases or to monitor the clinical signs of disease. Performance, study conditions, and the reference test used were assessed to decide if a technology was considered reliable. Only studies with a low risk of bias were assessed for reliability.For dairy cows, the International Organization for Standardization (ISO 20966, 2007) [28] presents a minimum performance required to support a technology efficiency (Se > 80% and Sp > 99%). These minimum requirements are supported because the reported success that farmers have identifying clinical mastitis is 80%, and therefore a technology with a performance above this threshold is able to outperform the human [18]. To assess performance of PLF technologies for monitoring respiratory diseases, it would be ideal to use a similar rationale to that of the ISO 20966, though to our knowledge there are no specific standards for monitoring respiratory diseases in livestock. Therefore, similarly to Stygar et al. [23], we assessed the validated technology’s performance by considering high performance when Se > 90% and Sp or precision was >90%. Thus, both Se and Sp had to be higher than 90% for the technology presented. If Sp was not presented but the article presented precision, both Se and precision had to be higher than 90%. Articles that did not present Se and Sp or precision were not included in the discussion of reliable technologies. When a study compared various methods for developing the technology (e.g., when multiple machine learning methods were used), we collected the performance of the technology that achieved the highest performance.The reference test was considered reliable when the study presents evidence that it is effective to assess the clinical sign that is being monitored. For an article to prove reference test reliability, tests should be presented demonstrating that the used reference test was able to detect the disease or clinical sign. Another way of considering reliability was the use of a well-established reference test, given that the authors provided peer-reviewed references for it. For sound-based technologies, remote audio labeling sounds from audio recordings were not considered reliable due to misclassifications that may occur in the labeling process [29].In sum, the criteria for considering a technology reliable were high performance, use of a reliable reference test, tests conducted in field conditions, and the study having low risk of bias.3. ResultsA total of 23 articles were included in this review. Most of the articles were on PLF technologies for swine (13/23), followed by poultry (6/23) and bovine production (4/23). No articles were found about PLF technologies for small ruminants (goats and sheep).Most technologies used sound-based devices (21/23). Other types of sensors found were image-based (1/23) and accelerometer (1/23). Table 2 summarizes the data gathered from the studies included in this review.The production stage, production system, and aim of the production of all species are presented in the Supplementary Table S2 available at [https://doi.org/10.6084/m9.figshare.21758543.v2, accessed on 14 March 2023].3.1. Studies’ Condition and Reference TestsIn this section, we report all reference tests found in studies included in this review. The reference test varied in relation to the species and the technology applied. Reference tests found were: blood analysis (1/23; used the number of neutrophils), clinical assessment (2/23; assessment of clinical signs presented by the animals), clinical assessment and blood analysis (1/23; number of neutrophils and clinically assessment for respiratory diseases), remote audio labeling (11/23; manually label sounds from audio files), live audio labeling (3/23), PCR (2/23), remote audio labeling and blood analysis (1/23; label sounds and collected blood samples to identify the disease), video labeling and blood analysis (1/23; used both image of the animals and audio recordings to label cough sounds and collected blood samples to identify the disease), video labeling, and PCR (1/23).3.2. Risk of BiasA total of eight studies were considered to have a high risk of bias: five swine production studies [32,33,37,38,41], two poultry production studies [43,44], and one for bovine production study [47]. Table 3 shows the specific information that was presented or not presented in each study.3.3. Respiratory Disease PLF Technologies for Poultry ProductionAll the technologies found for poultry production were sound-based (6/6). However, depending on the methodology, the monitored sounds could be cough and snore (1/6), vocalization (3/6), sneeze (1/6), or rale sounds (1/6). The performance measures of the technologies for poultry production are presented in Table 4.From the studies included, Liu et al. [45] achieved good performance measures (Se, precision, accuracy, and F1-score) and was developed in field conditions.Banakar et al. [42] and Cuan et al. [44,46] developed technologies that aimed to detect and diagnose different types of respiratory diseases in poultry. The technologies were developed based on vocalization before and after virus inoculation under laboratory conditions and used PCR tests as reference test. Cuan et al. [44] achieved a high accuracy for the detection of Avian Influenza but did not present other performance measures (Se, Sp, and precision). Banakar et al. [42] achieved a high performance (Se and Sp) for detecting Avian Influenza but failed to achieve a high Se for Bronchitis Virus and Newcastle Disease. Cuan et al. [46] achieved high performance for all performance measures presented, with a technology aimed at detecting Newcastle disease.All other technologies were based on specific sounds for monitoring respiratory disease. Rizwan et al. [43] monitored rale sounds and achieved high accuracy (97.6%) but was unable to achieve high Se and precision. Carpentier et al. [14] developed a technology that detects sneeze sounds; however, the technology Se was low and therefore ineffective to identify true positive cases of the disease. Both studies were conducted under laboratory conditions.3.4. Respiratory Disease PLF Technologies for Bovine ProductionSeveral types of sensors were used in bovine studies, sound-based technology (2/4), image-based (1/4), and accelerometer (1/4). The performance measures of the technologies for bovine production are presented in Table 5.Almost all technologies developed for bovine production aimed to detect respiratory diseases in dairy calves [13,16,48]. This is in line with the period of highest risk for respiratory diseases in bovines [49].The sound-based technologies [13,48] were developed and validated in field conditions; however, both technologies were inefficient to correctly identify true positive cases of the disease.Another study used an accelerometer embedded in a necklace to detect respiratory diseases in dairy calves [16]. The accelerometer monitors feeding and activity behavior. This study achieved a high Sp; however, like the technologies from Vandermeulen et al. [13] and Carpentier et al. [48], the Se was low.3.5. Respiratory Disease PLF Technologies for Swine ProductionAll studies assessed were sound-based (14/13). The performance measures of the technologies for swine production are presented in Table 6. Most technologies were developed for the fattening stage (9/13), followed by finishing stage (3/13), and in one study no information was provided about the production stage.Many technologies for swine production presented good performance measures [15,37,39,40]. Swine respiratory disease automatic detection has been researched for over two decades [30,31,32,33], enabling a myriad of studies on this topic.Most technologies aimed at detecting coughs to monitor respiratory health and thus were not developed to detect a specific respiratory disease [30,31,32,33,34,35,36,38,39,40]. The performance mostly assessed by these studies was the cough detection rate, which refers to the accuracy of the technology in detecting cough sounds. Many studies just presented this value to “prove” technology’s efficiency; therefore, we were incapable to evaluate if the technology could be considered reliable by the standards used in this study.The technologies developed by Shen et al. [39,41] and Yin et al. [40], achieved high performance in the detection of cough sounds. However, the studies used audio labeling recordings as the reference test, perhaps resulting in a less reliable dataset [29].Another set of studies from the same research group [15,37] developed a technology that aimed to identify specific swine respiratory diseases (Mycoplasma Hyopneumoniae, Postweaning Multisystemic Wasting Syndrome and Porcine Reproductive and Respiratory Syndrome virus) by monitoring cough sounds. Both studies achieved an overall performance > 90% for Se, precision, and F1-score. Because their technology aimed to detect the disease, the reference test used was serological analysis of suspected pigs to detect the disease. The cough detection rate was also >90% for both studies. Different from the reference test of Shen et al. [39,41] and Yin et al. [40], the audio was labeled with audio and video recordings and therefore could lead to a more reliable dataset due to the video footage that could help conclude if an event is a cough sound.4. DiscussionWe found promising results regarding PLF technologies developed for monitoring respiratory diseases in livestock production. Some studies were able to achieve a high performance; however, when assessing key points of the validation process (reference test and study condition), many technologies were not considered reliable for automatic detection of livestock respiratory diseases. Nine studies did not perform the validation process in field conditions. We highlight that tests in laboratory conditions are important for an initial analysis of technology’s potential, but tests in field conditions are required to assess technology performance in real-life situations.Most of the technologies applied for respiratory disease detection are sound-based. The most common respiratory diseases’ clinical signs lead animals to emit sounds [9,14,50], and thus it is possible to monitor with the use of microphones. Diagnosis of respiratory disease is more complex than monitoring only one clinical sign and requires assessment of multiple signs (e.g., cough, fever, nasal discharge, lethargy, lung consolidation). Two articles excluded from this review monitored coughing with PLF technologies to assess air quality in swine productions [51,52]. Albeit air quality is a major risk factor for respiratory diseases, an alert based on air quality does not necessarily mean that the animals are sick.Monitoring more than just one clinical sign of respiratory disease is very important to create a technology that could lead to a more reliable diagnosis; for example, monitoring cough and fever together could be useful for developing a technology able to better diagnose respiratory diseases. Other data, not collected by the technology, could also be used to overcome false positives, for example: season, feeding time (feeding may increase fine particles in the air causing irritation and thus coughing when inhaled), and animal handling. One article used a camera to monitor animals’ temperature [47]. Even though this study presented information that confirms the reliability of temperature as a clinical sign that can be monitored for detecting respiratory diseases, many other diseases might be responsible for elevated body temperature. Non-specific, clinical signs are useful for veterinarians to support decisions on the health status of an animal; however, when developing technology, monitoring specific clinical signs is preferred for detecting targeted diseases [20].Due to the current concerns about antimicrobial resistance [53], methods to identify diseased individuals could be beneficial for reducing the overall antibiotic use on the whole group as only a single animal would be treated. None of the reviewed studies on poultry and swine attempted to identify the individual with the ailment, and only one study on dairy calves was able to detect the individual because of the type of sensor used: an accelerometer attached to the calf [16]. Current management practices make it hard for the development of PLF technologies for monitoring respiratory health that can detect sick individual using bioacoustic sensors. Perhaps a technology able to identify the individual through voice recognition could be developed for identifying the individual affected by a disease. We understand that the development of such a solution would require much research to overcome usability and technological challenges.4.1. Performance MeasuresOverall, Se was the most presented performance measure; this test shows the capacity that the algorithm must correctly diagnose diseased individuals [54].Compared to Se, Sp was presented in fewer articles. Sp indicates the capacity of the technology to correctly identify individuals that are not diseased [54]. A technology with low Sp could lead to an increasing number of false alarms, which in turn may reduce system’s usage (e.g., farmer stop paying attention to the alerts) or antibiotic overuse.In assessing the system’s validity, Se and Sp may give different but complementary insights. Evaluating these performance measures together demonstrates how reliable an alarm given by a technology is, so we hypothesize that high Se and Sp can potentially increase customer’s confidence in the alarms given.The equation for Se is presented as “Se = true positive/true positive + false negative”. Technologies with low Se will fail to identify diseased animals or monitor clinical signs because too many false negatives will be detected by the technology.Precision (i.e., number of true positives in all positive alarms) is another performance measure that can be useful for understanding the validity of PLF technology. The equation for precision is presented as “Precision = true positive/true positives + false positives” and Sp is presented as “Sp = true negatives/true negatives + false positives”. While Sp will indicate the number of false positives considering true negatives (i.e., of all healthy animals or when clinical signs are absent, how many were mistakenly detected as disease animal or clinical sign), precision will indicate the number of false positives considering true positives (i.e., how many clinical signs or diseased animals were mistakenly detected by the technology). Although they are not the same, in cases where the number of true negatives does not matter (e.g., when treatment is applied to the entire group of animals regardless of whether there is a percentage of animals that are healthy or that do not present any clinical signs), precision may be used instead of Sp. This could justify why many technologies for swine or poultry did not present Sp, while for bovine all studies presented Sp. In swine production, pigs are grown in group pens and all animals are treated when there is a case of a disease [55]. However, in productions where the specific individual is treated (e.g., bovine production) Sp should be presented.In this review, we considered that just Se and Sp, or Se and precision, could be presented by articles to support performance. However, the decision whether which performance measure is more important will vary according to the objectives of the user. If Se has little impact and Sp is very impactful in a specific situation, the user should focus on Sp instead of Se.Analyzing performance measures can give the reader full insight into how helpful a technology will be to monitor a disease. Therefore, we find that the articles should present all relevant performance measures (Se, Sp, precision, accuracy, F1-score, PPV and PNV) to support technologies’ validity.The nomenclature used by different studies for some performance measures may hinder interpretability. Exadaktylos et al. [34] presented a performance measure named “overall performance”, which is the same as “total accuracy” [36] and “overall accuracy” [37]. A way to overcome this would be to present the equation for the performance measure, which could help the reader to understand exactly what is being presented. However, this is not ideal since using many nomenclatures to describe the same thing could lead the reader to confusion. For example, the performance measure “Cough detection rate” is likely to be the same as Se, since many articles describe it as the ability that the technology had to detect truthful cough sounds (true positives) out of all cough sounds (true positives + false negatives). To prevent mistakenly reporting results, we decided to report the performance measure with the nomenclature used by the article.A standard nomenclature for PLF validation studies could be a solution since every researcher would use the same terms. Although using standard terms would be ideal, we believe that it is challenging to teach and require every researcher to use the same terms. The inclusion of the confusion matrix is helpful and likely the best option as, if presented, most performance measures could be calculated (Se, Sp, accuracy, precision, F1-score, PPV, NPV). The confusion matrix is also important to assess risk of bias in the results and could be used in future meta-studies on this topic.4.2. Reference TestIn this review, we decided to use “reference test” instead of “gold standard” because to our knowledge, there is no perfect test (gold standard) to identify respiratory diseases or to monitor the clinical signs. We assessed the reference test used by studies and stated what we consider to be reliable for validating PLF technologies for monitoring respiratory diseases.The reference test is essential for supporting the performance of a technology. A technology that achieved a high performance in a study based on an unreliable reference test is likely to perform badly when applied in a commercial setting. Poorly tested technology may provide incorrect insights, consequently leading users to wrong decision-making that can negatively impact animals [56].Vandermeulen et al. [13] and Carpentier et al. [48] used blood analysis as a reference test to confirm if the proposed technology was able to detect respiratory diseases by monitoring coughing in dairy calves. Both used the number of neutrophils as the reference test to determine if an animal was sick. Vandermeulen et al. [13] also used the Wisconsin health scoring criteria, a diagnostic tool that includes the clinical evaluation of six clinical signs caused by respiratory diseases. After comparing reference tests, blood analysis failed to correctly identify cases of bovine respiratory disease, and therefore assessing specific respiratory disease clinical signs is a better reference test to validate the technology for dairy calves [13]. It is observed in the results that the reference test affected the performance of the technology. The technology that just used blood analysis [48] as a reference test had lower Se when compared to the technology that used the Wisconsin health scoring chart [13]. If applied in field conditions, the technology with unreliable reference tests would be less effective to detect true positive cases of diseased animals.Schaefer et al. [47] combined different clinical signs to determine if an animal was affected by respiratory disease. If the animal presented at least three clinical signs, the illness was considered as a respiratory disease. The clinical signs were high temperature, a clinical score higher than 3 (moderate to severe nasal discharge, cough, and crepitant auscultation), and low levels of white blood cells and neutrophilia. The comparison of each of these clinical signs to detect respiratory disease was presented. Fever was the most effective reference test in this study (high Se and Sp). However, caution should be taken when interpreting these results. High temperature is a clinical sign that can be caused by many diseases, so just assessing this clinical sign for exclusively detecting respiratory diseases might not be ideal, especially if applied to younger animals as these animals are susceptible to many other diseases [57]. Furthermore, the technology was tested in conditions similar to auctions and could not perform so well when applied in different environments (e.g., commercial farm).For bovine production, there is a wider discussion on defining what could be considered a reliable reference test to detect respiratory diseases. The use of thoracic ultrasonography is considered accurate to detect respiratory diseases in bovine [58]. Studies comparing the efficiency of many clinical signs and clinical score systems to thoracic ultrasonography for diagnosing respiratory diseases were published [59,60]. Lowie et al. [60] found that spontaneous cough was the clinical sign that best indicated a calf to have respiratory diseases; however, the performance of this clinical sign is relatively poor (Se = 37.4%, Sp = 85.7%).The use of the Wisconsin scoring system was tested and achieved a Se of 62.4% and Sp of 74.1% [61]. This result shows that the scoring chart is also not a perfect measure to be used as a reference test for detecting respiratory disease. When applying scoring systems, the criteria for diagnosing respiratory diseases in different environments should also be taken into consideration [59].We categorized one type of reference test as audio labeled for those studies that did not assess clinical data other than cough. Overall, the methodology consists of labeling sounds from audio files that were previously recorded at animals’ sites. The labeling process is performed by a person that listens to the audio files and label cough sounds. Aerts et al. [29] compared the number of coughs that were labelled by an observer on the scene with an observer that used the audio labelled methodology previously described; this study found an underestimation of up to 94% on the number of coughs detected by the observer that labelled the audio files. Therefore, labeling sounds from recordings may not be an effective reference test. This may have artificially inflated the performance measures reported in the studies reviewed.4.3. Risk of BiasSeveral studies lacked key information that are considered important to assess studies’ risk of bias. Within these studies, the most common issues were related to not presenting information about the devices [32,33,43], and software [47] used. Hardware information is necessary since different sensors (i.e., different manufacturers and technology specifications) could perform differently. A full description of the software is needed to understand how the technology works and to present proof that the technology’s efficiency is supported by the software.A few studies did not show the data (i.e., raw data or confusion matrix) that could be used to cross-reference their results [37,41,44]. Other studies lacked descriptive information of the context where the technology was applied such as the number of animals used [32] or animal production phase [37].Descriptive information regarding animals is essential since the technology could not work as well when applied to different groups (different ages, weights, and different productions stage) or larger groups (applying the technology to a much larger group of animals). For example, for swine production most were developed for the fattening stage than the finishing stage. A technology developed and validated in a specific stage could not work well when applied in a different production stage. It is important to conduct tests in different stages to support the technology’s efficiency in different situations.One study [38] did not provide housing information and how the devices were installed. This is necessary because technologies applied in different environments could lead to different results. Providing data on how the sensor is installed is important to understand if the technology is useful, taking in consideration its positioning; for example, a technology could work well if placed close to animals but not work when placed more distantly.4.4. Can We Reliably Detect Livestock Respiratory Disease through Precision Farming?Based on our definition of reliability, which included high technology performance achieved through a comparison against reliable reference tests, conducted in field (or similar) conditions and having low risk of bias, we identified only one study that fit these criteria [15]. This study validated a technology able to monitor pigs’ cough sounds and specific disease detection (Mycoplasma hyopneumoniae, porcine reproductive and respiratory disease, and postweaning multisystemic wasting syndrome) in field conditions. The technology is described as a low-cost solution that is suitable for smaller farms. All information to assess risk of bias is described by the study. Two reference tests were used: labeling cough sounds from video and audio recordings (video was used to ensure that a sound was correctly labeled), and blood analysis to determine the type of respiratory disease affecting the pigs. The cough sound was linked to the disease affecting the pigs to validate the algorithm for specific disease detection. The technology achieved high performance for both cough sound detection and specific disease detection. The only limitation of the study is the low sample size.Other two studies complied with most of our criteria but had a high risk of bias [37,47]. One swine study [37] did not describe the study population and the raw data supporting their performance results. The other study monitored respiratory diseases with infrared thermography in bovine production [47]; however, it did not provide full software description.A limitation of our approach may have been that setting the performance threshold at >90% resulted in leaving several relevant articles out of further analysis. Another limitation is that we have not assessed all the PLF proceedings from conferences and companies’ internal validations available online.5. ConclusionsWe evaluated PLF technologies to monitor respiratory diseases in swine, bovine, and poultry productions. One study met all the criteria set in this review to be considered reliable. Some studies achieved high performance, but it is unclear if the technologies would perform well when applied in real-life situations. Many studies were only conducted on laboratory conditions or used an unreliable reference test. We identified issues with how studies report their validation tests, such as not reporting all relevant performance measures, low performance measures, different nomenclature for the same performance measure within studies, and not reporting all necessary information to assess risk of bias.We encourage future studies to improve how the methodology and results are reported so that all important information is provided for readers to have a full understanding on the effectiveness of a technology.The fact that some studies have had good performances indicate that, in the future, it might be possible to rely on PLF technologies for automatic monitoring of respiratory diseases.

animals : an open access journal from mdpi

10.3390/ani11082458

PMC8388640

The nematode species Anisakis simplex sensu stricto (s.s.) and Anisakis pegreffii are wormlike parasites found in commonly consumed fish and are the main cause of human anisakiasis. Outwardly, the two nematodes are extremely similar and difficult to distinguish, especially in their larval forms. Genetic analysis has discovered the existence of a hybrid between these two “sibling species”, but its identification is a controversial matter, as results differ according to the specific region of the DNA analysed. The aim of our work was to confirm the presence of this hybrid genotype in fish off the Spanish coast and to obtain fourth-stage larvae in the laboratory to study if different genotypes are associated with different growth behaviour. Our results confirm that hybrid genotypes can be overestimated if identification is based on a particular molecular marker. We also obtained fourth-stage larvae with a hybrid genotype, which has not been reported previously. These findings are valuable for the taxonomic classification of Anisakis species, and for further epidemiological and biomedical research.

The sibling species Anisakis simplex (s.s.) and Anisakis pegreffii are parasites of marine mammals and fish worldwide and the main causative agents of human anisakiasis. In sympatric areas, a hybrid genotype between the two species has been identified, mainly in third-stage larvae, but rarely in fourth-stage and adult forms. The aim of this study was to confirm the presence of hybrid genotypes in larvae parasitizing fish caught in sympatric and allopatric Spanish marine waters, the North-East Atlantic and West Mediterranean, respectively, and to study possible differences in the growth behaviour between genotypes. Of the 254 molecularly analysed larvae, 18 were identified as hybrids by PCR-RFLP analysis of the rDNA ITS region, 11 of which were subsequently confirmed by EF1 α-1 nDNA gene sequencing. These results therefore indicate an overestimation of hybrid genotypes when identification is based only on the ITS region. We also report the detection of a hybrid specimen in a host from the West Mediterranean, considered an allopatric zone. Additionally, fourth-stage larvae with a hybrid genotype were obtained in vitro for the first time, and no differences were observed in their growth behaviour compared to larvae with A. simplex (s.s.) and A. pegreffii genotypes.

1. IntroductionAscaridoid nematodes of the family Anisakidae include species with sanitary and/or commercial impact that are found worldwide in fish and marine mammals [1,2,3]. The life cycle of these parasites includes cetaceans and pinnipeds as definitive hosts, a wide range of fish and cephalopods as paratenic and/or intermediate hosts, and crustaceans as first intermediate hosts [2]. Species of the genus Anisakis are the main causative agents of human anisakiasis, an emerging disease with gastrointestinal and/or allergic symptoms acquired by eating raw or undercooked fish parasitized with larvae [4,5,6,7]. The disease is mainly caused by third-stage larvae (L3) of the sibling species Anisakis simplex sensu stricto (s.s.) and Anisakis pegrefii, which are included in the A. simplex sensu lato (s.l.) complex, together with A. berlandi, and are morphologically indistinguishable [8,9].A. simplex (s.s.) and A. pegreffii have a worldwide distribution, with allopatric areas where only one species is found, and sympatric areas where the two overlap [2]. In sympatric areas, such as the North-East Atlantic and the North-West Pacific, both species can also share hosts and commonly co-infect the same definitive or intermediate/paratenic host. In these areas, hybrid forms between the two sibling species have been detected [9,10,11]. However, despite the large number of L3 larvae described with a hybrid genotype, reports of hybrid fourth-stage larvae (L4) and adult specimens in definitive hosts are scarce [11].Until recent years, hybrid forms among A. simplex (s.l.) sibling species were identified mainly by PCR-RFLP of the rDNA ITS region, although it has been suggested that this method leads to an overestimation of hybrids [12,13]. Alternative multi-marker nuclear genotyping approaches have been proposed, such as the use of allozyme markers or sequencing of the elongation factor 1 alpha 1 (EF1 α−1) nDNA [12].The aim of the present study was to confirm the presence of hybrid genotypes between A. simplex (s.s.) and A. pegreffii in larvae parasitizing fish caught in sympatric and allopatric Spanish marine waters using a multi-marker genotyping approach. Moreover, to gain insight into the behaviour of the hybrid genotype, special emphasis was placed in obtaining further developmental stages of the two sibling species by in vitro culture.2. Materials and Methods2.1. Fish Samples, Larvae Collection, and Morphological IdentificationTwo teleostean fish species commonly consumed in Spain were analysed for the presence of A. simplex (s.l.) larvae: horse mackerel (Trachurus trachurus) (n = 52) and blue whiting (Micromesistius poutassou) (n = 98). All hosts were acquired dead from markets in Barcelona during 2015-17 and came from two fishing areas, the sympatric area of the North-East Atlantic Ocean, corresponding to zone 27.8 of the Food and Agriculture Organization of the United Nations (FAO), and the allopatric region of the Western Mediterranean Sea, corresponding to FAO zone 37.1.1. The fish were dissected and viscera were examined under the stereomicroscope for the detection and isolation of nematode larvae. Anisakis larvae were preserved in 70% ethanol and studied microscopically. Larval specimens were cut in three portions: the anterior and posterior parts were mounted in lactophenol for morphological identification, following the criteria of Berland [14], and the central portion was used for molecular analysis.2.2. In Vitro Culture of Anisakis LarvaeA selection of L3 larvae of A. simplex (s.l.) (n = 200), isolated from North-East Atlantic blue whiting, were cultured in vitro to obtain further developmental stages. Larvae were first placed in antibiotic and antimycotic solution for axenization, as reported by Iglesias et al. (1997), [15] and subsequently cultured in commercial medium RPMI-1640 + 20% (v/v) FBS (foetal bovine serum) + 0.1% w/v pepsin (1:10,000 NF), adjusting the pH to 4.0, at 37 °C with an atmosphere of 5% CO2 [16]. Culture plates were observed daily under the inverted microscope to monitor larval evolution. Once moulted into L4 larvae, a subsample of 90 specimens was fixed and preserved in 70% ethanol for morphological and molecular identification. The remaining larvae continued to be cultured in vitro to obtain adult forms.2.3. Molecular IdentificationGenomic DNA was isolated using the Pure PCR Template Extraction Kit (Roche), according to the manufacturer’s protocol. Specific identification of larvae was carried out by PCR-RFLP analysis of the complete rDNA ITS region using NC2 and NC5 primers with a hybridization temperature of 55 °C. DNA amplification products were digested with restriction endonucleases HinfI and HhaI at 37 °C for 90 min. The digested products were subjected to electrophoresis in 2% agarose gel and visualized by a UV light Illuminator [17]. To confirm the detected hybrid genotypes, the elongation factor 1 alpha 1 (EF1 α-1) of the nDNA region of these specimens, as well as of A. simplex (s.s.) and A. pegreffii larvae identified by PCR-RFLP, was amplified and sequenced from the isolated DNA using EF-F and EF-F primers with a hybridization temperature of 58 °C [12].2.4. Statistical AnalysisThe nonparametric chi-squared test (χ2) was used to assess differences in the molecular identification methodology, the PCR-RFLP and the EF1 α-1 sequencing. The statistical analysis was performed using SPSS v22 software and the level of significance was set to p < 0.05.3. Results3.1. Morphological Identification and In Vitro Culture of A. simplex (s.l.)All larvae obtained from the dissected fish were morphologically identified as L3 larvae of A. simplex (s.l.). After four days, 100% of the L3 larval specimens selected for in vitro culture moulted to L4 larvae, characterized by the loss of the cephalic tooth, the presence of developed lips, a marked striated cuticle, the absence of a mucron, and signs of moult in the culture media. The motility of L4 larvae declined progressively until their death, which occurred between days 7 and 55 of culture (on average, day 34), and adult specimens were not obtained.3.2. Molecular IdentificationA total of 254 A. simplex (s.l.) larvae, including L3 and L4 larvae, were molecularly identified as A. simplex (s.s.) (n = 140), A. pegreffii (n = 96) and hybrid forms (n = 18) according to genotype-specific patterns revealed by PCR-RFLP analysis of the rDNA ITS region (Figure 1A). The host and geographical origin of the identified specimens are shown in Table 1.Sequencing of the EF1 α-1 nDNA was performed in all hybrid genotypes and in a representative subsample of A. simplex (s.s.) (n = 5) and A. pegreffii (n = 5) L3 larvae previously identified by PCR-RFLP. EF1 α-1 sequencing confirmed the specific identification of A. simplex (s.s.) and A. pegreffii through the observation of the corresponding nucleotides at two diagnostic positions: base pairs 186 and 286. Hybrid genotypes were confirmed in 11 specimens by the observation of a heterozygote pattern at both diagnostic positions (Figure 1B), while 7 specimens presenting the heterozygote pattern by PCR-RFLP were finally identified as A. simplex (s.s.) (n = 3) and/or A. pegreffii (n = 4) (Table 1). The chi-squared test showed significant differences when comparing the hybrid genotype identification by PCR-RFLP with their identification by the multi-marker approach (p < 0.01). Among the confirmed hybrid specimens, four corresponded to L3 and seven to L4 larvae. Sequences of the studied larvae were deposited in GenBank under the accession numbers MZ517160-2.Using the multi-marker approach, A. simplex (s.s.) specimens were only identified in North-East Atlantic hosts, whereas A. pegreffii and hybrid genotypes were detected in fish from both Atlantic and Mediterranean areas. In the allopatric area of the West Mediterranean, one sample was identified with the hybrid genotype (1/44), while in the sympatric area of the North-East Atlantic, hybrid genotypes represented 4.8% of the total analysed larvae (10/210).4. DiscussionThe extensively used rDNA ITS region/gene is an efficient tool for the specific identification of nematodes [18,19]. However, it has been suggested that molecular techniques based only on this region have limited power regarding the identification of hybrid genotypes between A. simplex (s.s.) and A. pegreffii, and may lead to an overestimation of hybrid specimens [12,13]. This limitation is not exclusive to Anisakis hybrid genotypes, and the risk of using a single genetic marker has been highlighted for the specific identification of other parasites [20].For the species included in the A. simplex (s.l.) complex, the two polymorphisms of the ITS region, previously designated as species-specific, are not fixed diagnostic markers, so it is not clear whether the polymorphism shared by the two taxa is caused by incomplete lineage sorting, historical introgression, or current hybridization [12]. Therefore, to elucidate the hybridization phenomena between Anisakis sibling species, multi-marker nuclear genotyping approaches have recently been proposed, such as the use of allozyme markers and sequencing of the EF1 α-1 nDNA [12]. The results of the present study suggest that analysis based only on the rDNA ITS region can lead to an overestimation of hybrids, with fewer being detected by EF1 α-1 nDNA gene sequencing, supported by significant statistical differences. The identification of the hybrid genotype should therefore be confirmed by the analysis of another DNA region in addition to the rDNA ITS.The detection and/or origin of Anisakis hybrid genotypes remains a matter of controversy. Mladineo et al. [13], in a study using a microsatellite panel, suggested that the hybridization is caused by ancestral polymorphism, which could be a leftover of incomplete lineage sorting. However, this observation was questioned by Mattiucci et al. [8], who obtained divergent results using a different microsatellite panel. Far from disputing the utility of microsatellites as a tool for Anisakis sibling species identification, these authors argue that more studies are needed to clarify the levels of genetic differentiation between these species, which would open a wide field of study in Anisakis population genetics based on microsatellite analysis.More recently, several authors have identified hybrids between A. simplex (s.s.) and A. pegreffii using novel diagnostic markers from the nuclear genomes, such as the beta-tubulin gene and the nas10 nDNA region, thus expanding the range of markers available for the multi-marker genotyping approach to Anisakis species classification [21,22].The detection of hybrid genotypes between A. simplex (s.s.) and A. pegreffii is of great significance for the study of speciation in the A. simplex (s.l.) complex. Llorens et al. [23] observed strong parent-of-origin effects in the hybrid transcript repertoire, which would be important in the evolutionary biology of A. simplex (s.s.) and A. pegreffii through gene introgression. Other authors suggest that the presence of hybrid forms between the sibling species complicates the accurate definition of genomes, transcripts, and proteins, including the assignment of genes or gene products to a determinate species [24].Hybrid detection is also important in terms of the biomedical impact of the parasite on humans and its pathogenic potential for marine hosts. In this context, it has been suggested that A. simplex (s.s.), A. pegreffii and their hybrids have differential allergenic potential, and may induce overlapping disease responses, with a variable capacity to penetrate tissue [23,25,26]. Hybrid forms are also relevant from an epidemiological point of view, as hybridization can lead to the colonization of new hosts [27]. Moreover, the presence of hybrids reflects the dynamics of moving areas of sympatry and sheds lights on the microevolutionary processes of host–parasite ecology and definitive host migration [23].Hybrid specimens between A. simplex (s.s.) and A. pegreffii have previously been documented in fish from sympatric areas, such as the North-East Atlantic, in agreement with their detection in two Atlantic hosts in the current study, and the Sea of Japan [10,11,28]. This phenomenon has also been observed in Spanish Mediterranean waters, especially the sympatric area of the Alboran Sea, including in horse mackerel and blue whiting, the two hosts studied here [11,29,30]. The proportion of hybrid genotypes previously reported in North-East Atlantic waters is slightly higher (around 15%) than the value we obtained.In contrast with most reports, which indicate that hybrid genotypes are restricted to sympatric areas, in the present study a hybrid specimen was detected in the allopatric area of the West Mediterranean Sea. This could be explained by host mobility, as horse mackerel are highly migratory and could have travelled there from the Atlantic Ocean or the Alboran Sea, where A. simplex (s.s.) and hybrid specimens are present [31,32]. Therefore, the identification of a hybrid in the West Mediterranean highlights the possibility of finding hybrid genotypes in allopatric areas, particularly those near sympatric regions. Hybrid specimens have been previously reported in strictly allopatric areas, such as the Adriatic, the Tyrrhenian or the Aegean Sea [13,33,34], but all the studies, except Mladineo et al. [13], based the identification on the rDNA ITS region, which provides insufficient evidence for the presence of hybrids.In the present work, a hybrid genotype was also identified in seven L4 larval specimens obtained by in vitro culture, which showed no differences in in vitro culture behaviour or the moulting process in comparison with A. simplex (s.s.) and A. pegreffii. No L4 larvae with a hybrid genotype have been described in definitive hosts and few hybrid adults have been identified [10,21,35]. Consequently, it has been suggested that Anisakis hybrid specimens have low fertility and/or reduced fitness, which could affect their survival and therefore their detection and identification in definitive hosts [12,36]. Nevertheless, another factor could be the difficult recovery of these developmental stages in contrast with the readily obtainable L3 larvae. It has also been proposed that hybrid forms might only be able to survive in hosts when accompanied by their parental species [23], but in our case all in vitro-cultured L3 larvae, growing individually without competition, reached the fourth developmental stage, with no specific difference between genotypes.Although a hybrid genotype was identified in the cultured L4 larvae, reported here for the first time, no adult specimens were obtained on continuing the culture, regardless of the genotype. Despite this, in vitro culture of Anisakis species constitutes a useful tool in the study of developmental stages in the nematode life cycle. In several recent studies, in vitro culture was employed to analyse the transcriptome and gene expression in L4 larvae of A. simplex (s.s.), A. pegreffii and their hybrids, as well as for morphological and molecular characterization [9,37,38].5. ConclusionsOur results confirm that a multi-marker genetic approach is needed to identify hybrid genotypes between A. simplex (s.s.) and A. pegreffii, as analysis based only on the ITS rDNA region may lead to an over-detection. Notably, we identified a hybrid specimen in the allopatric area of the Western Mediterranean, which was likely due to the migration of horse mackerel. Additionally, fourth-stage larval hybrids, which were obtained in vitro for the first time in this study, presented no moulting or behavioral differences with regard to A. simplex (s.s.) or A. pegreffii larvae at the same stage of development.

animals : an open access journal from mdpi

10.3390/ani12010036

PMC8749609

As far as it is known, studies dealing with antimicrobial resistance in certain species of staphylococci, in particular, S. chromogenes and S. simulans, isolated from products made from unpasteurized milk are limited. In addition to that, little attention was paid to the resistance of staphylococcal isolates from regional sheep and goat cheeses. At this level, works are published that focus on the evaluation of resistance from only one sheep product, Bryndza. Other studies are only focused on the evaluation of resistance from raw sheep’s or goat’s milk. The study contributes to the knowledge of the possible spread of antimicrobial resistance from the farm to the final consumer in this area.

S. aureus and some species of coagulase-negative staphylococci, including S. chromogenes and S. simulans, commonly cause intramammary infections. However, little attention was paid to the antimicrobial resistance of these species with respect to their occurrence in dairy products, for example, popular sheep and goat cheeses made from unpasteurized milk. The aim of this study was to investigate such sheep and goat cheeses for the occurrence and antimicrobial resistance of the relevant staphylococci species. The staphylococcal isolates were identified by polymerase chain reaction (130 isolates) and matrix assisted laser desorption/ionization time-of-flight mass spectrometry. The most common species of S. aureus (56 isolates) were identified, as well as S. chromogenes (16 isolates) and S. simulans (10 isolates). Antimicrobial resistance to penicillin, oxacilin, ceftaroline, teicoplanin, gentamicin, erythromycin, tetracycline and ofloxacin was subsequently determined in these species using the agar dilution method. The highest resistance was confirmed in all species, especially to penicillin (91%) and erythromycin (67%). The highest sensitivity was confirmed to ofloxacin (83%). Due to the high incidence of penicillin and oxacilin-resistant staphylococci, the mecA gene was detected by polymerase chain reaction, which was confirmed only in S. aureus isolates (19%). Our study shows that the tested strains (77%) were resistant to more than one antibiotic at a time.

1. IntroductionOne of the oldest fermented foods is cheese [1,2]. Cheese has been part of the human diet for thousands of years, although there have been changes in dietary patterns associated with technical, social, and economic progress in individual geographical regions. The cheese fermentation process and use depend considerably on culture and tradition [3]. There are numerous variations in the characteristics in cheese, including texture, aroma, visual presentation, and flavour, that depend on the activity of microorganisms and the cheesemaking process [4].In addition to the beneficial microflora, undesirable microorganisms are also present in the cheeses. Among them, we also include bacteria of the genus Staphylococcus sp., which are the main causative agent of mastitis, with a higher prevalence in cases of clinical and subclinical manifestations. Recently, in addition to S. aureus, S. chromogenes and S. simulans species have also often been detected in mastitis [5,6,7,8].The presence of antimicrobial resistance has recently been confirmed in the above-mentioned staphylococcal species, which is an increasing problem, so the collection of information on pathogen resistance is very important from an epidemiological point of view. Especially, a serious risk is associated with multiresistant strains and their resistance to more than one antibiotic [9]. Recent studies also suggest the presence of methicillin-resistant staphylococci (MRS) that have been identified in unpasteurized milk and dairy products, including cheeses. The opportunistic ability of MRS strains to cause mastitis and serve as a source of zoonotic infections is a considerable threat to public health. Moreover, they present a reservoir of antimicrobial resistance genes on dairy farms. The most widely reported MRS species is Staphylococcus aureus (MRSA). Coagulase-negative staphylococci (CNS) were also identified as MRS isolates [10].At the same time, the degree of resistance to different antibiotics varies from species of staphylococci; therefore, their identification of them is important. Traditional methods used to identify and classify bacteria are based on the analysis of morphological, physiological and biochemical traits or genetic approaches (DNA-DNA or RNA-DNA hybridization, determination of G+C content in DNA). Those methods are currently supplemented by PCR methods to analyze the sequence of small rRNA subunits [11].Matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) currently serves as a method suitable for this purpose. Recently, considerable improvement has been made in MALDI-TOF MS tools intended for microbiological identification of potential pathogens and foodborne pathogens. These commercial, MALDI-TOF MS devices are convenient and utilize their own algorithms and databases Many authors described them as fast, cost-effective systems with an accurate and reliable performance [12,13].Based on the abovementioned, the aim of our study is to detect the presence of Staphylococcus sp. by PCR and MALDI-TOF mass spectroscopy of sheep and goat cheese produced on farms in Slovakia and, subsequently, determine their antimicrobial resistance. The cheeses tested by us belong to the group of so-called “ready to-it” made from raw milk with no special starter culture [14].2. Materials and MethodsStaphylococci were isolated from cheese samples (10 sheep cheese samples and 10 goat cheese samples) taken between May and August 2021. Cheeses were obtained from a farm located in the border area of Slovakia and Hungary in the region Slanské vrchy. This farm is producing milk and milk products in accordance with legal requirements in the EU and is an officially approved establishment. The farm raises Valachian sheep and white shorthaired goats, which were grazed in a wild pasture. The cheeses were made from non-pasteurized sheep and goat milk without adding a cheese starter culture and were ripened for 30 days.2.1. General Microbiological AnalysisPreparation of test samples of cheeses and decimal dilutions were prepared according to STN EN ISO 6887-5 (2011) [15] and, subsequently, for the tested samples, the total viable count (TVC) according to STN EN ISO 4833-1 (2013) using Plate Count Agar. After incubating the inoculated plates for 72 h at 30 °C, the inoculated plates where more than 10 and less than 300 colonies grew were selected for the TVC calculation.2.2. Isolation of StrainsSubsequently, basic suspension and decimal dilutions were prepared from all tested samples according to ISO 6887-5 [15].Isolates of staphylococci from the examined samples were obtained according to ISO 6888-1 [16] using Baird-Parker selective arbitration medium. The inoculated plates were incubated at 37 for 24 h. Subsequently, plates with more than 10 and less than 150 atypical and typical colonies were considered for staphylococcal counts. Based on their characteristic appearance, 2 typical colonies (1.0–1.5 mm colony, black or grey colonies with halo) and 2 atypical colonies (black or gray colonies without halo), each plate was inoculated with a sterile bacterial loop on the surface of Columbia blood agar (Oxoid Ltd., Hampshire, UK) and incubated at 37 °C for 24 to 48 h. After incubation, individual strains were used for identification by PCR and MALDI-TOF MS.2.3. Identification of Staphylococcal IsolatesTotal genomic DNA was isolated from staphylococcal strains as described by Hein et al. [17].The obtained supernatant was used as a DNA source in PCR reactions. The 16S ribosomal RNA gene specific of Staphylococcus sp. was amplified in a thermal cycler (Techne Touchgene FTGPO2TD, Techne, Cambridge, UK). Primers used to amplify a given 16S rRNA, 16S1 (CAGCTCGTGTCGTGAGATGT) and 16S2 (AATCATTTGTCCCACCTTCG), were synthesized (Amplia s.r.o., Bratislava, Slovakia) and used according to Strommenger et al. [18]. The reaction mixture in a volume of 20 µL contained 1 µL genomic DNA, 10 pmol.L−1 of each primer and HotFirepol® Master Mix (Amplia s.r.o., Bratislava, Slovakia). The amplification was terminated by cooling to 6 °C. The PCR protocol was as follows: initial denaturation at 95 °C for 12 min, 25 cycles consisting of denaturation at 95 °C for 20 s, annealing at 55 °C for 1 min and extension at 72 °C for 2 min. The final extension at 72 °C for 10 min followed the last cycle. PCR products (420 bp) were visualized using MiniBIS Pro®, (DNR Bio-Imaging System, Ltd., Jerusalem, Israel).The species identification of bacteria was subsequently provided with the help of MALDI-TOF MS according to the standard sample preparation protocol of manual Bruker Daltonics [19]. The analysis of the results was performed in a Ultraflex III device (Bruker, Billerica, Massachusetts, USA) using Flex Analysis software, version 3.0, and evaluated with by BioTyper software, version 1.1 (Bruker, Billerica, MA, USA).Retrospective PCR products were sequenced in species-identified S. aureus, S. chromogenes and S. simulans strains by MALDI-TOF MS. The sequencing of the PCR products was performed according to the Sanger method (GATC Biotech AG, Konstanz, Germany). The obtained strains were sent to the GenBank—EMBL database for comparison with the sequences available in the nucleotide database of the National Center for Biotechnology Information (NCBI) available at http://www.ncbi.nlm.nih.gov/BLAST (accessed on 20 November 2021).2.4. Detection of Antimicrobial ResistanceThe susceptibility of isolated bacterial strains to selected antibiotics was determined by the agar dilution method (ADM) according to the procedure described in the CLSI document [20]. ADM is performed on Petri dishes with Müeller–Hinton agar (Hi-Media, Mumbai, India) in duplicate. Test plates containing different concentrations of antibiotics were used to determine minimum inhibitory concentrations (MICs). Turbidity of 24-hour bacterial suspensions was adjusted to the 0.5 McFarland turbidity standard. Drops of such adjusted suspensions were applied in parallel onto the surface of the test plates which were then incubated at 37 °C for 24 h. After incubation, we visually determined the minimum concentration of respective antibiotics that inhibited the growth of the investigated staphylococcal strains. The results were evaluated according to document CLSI [20].In determining the minimal inhibitory concentrations (MIC) used in assay plates with final concentration of antibiotics on Staphylococcus spp.: Penicillin (PEN) 0,.12;0.25; 0.5 mg·L−1; Oxacillin (OX) 0.12; 0.25; 0.5; 1.0 mg·L−1; Ceftaroline (KF) 0.5;1.0; 2.0; 4.0; 8.0 mg·L−1; Teicoplanin (TEC) 4.0; 8.0; 16.0; 32.0; 64.0 mg·L−1; Gentamicin (GN) 2.0; 4.0; 8.0; 16.0; 32.0; 64.0 mg·L−1; Erythromycin (E) 0.25; 0.5; 1.0; 2.0; 4.0; 8.0; 16.0 mg·L−1;Tetracycline (TE) 2.0; 4.0; 8.0; 16.0; 32.0 mg·L−1; Ofloxacin (OFX) 0.5; 1.0; 2.0; 4.0; 8.0 mg·L−1.To confirm the MRS strains, the presence of the mecA gene was detected by PCR according to Poulsen et al. [21]. The primers used to confirm the presence of the mecA gene were MecA-1 (5-GGGATCATAGCGTCATTATTC) and MecA-2 (5-AACGATTGTGACACGATAGCC) (Amplia s.r.o., Bratislava, Slovakia). The size of the amplification product using these two primers was 527 bp.Reference strain of S. aureus CCM 4750 (Czech Collection of Microorganisms, Brno, Czech Republic) were used in this study as a positive control for the agar dilution method and S. aureus CCM 4223 for the polymerase chain reaction.2.5. Statistical AnalysisResults obtained in this study were analyzed statistically using GraphPad Prism 5.0 software (2007). Comparison of individual proportions was made using the Chi-square test (χ2 test). The dependence of the individual signs was considered significant at the level of α = 0.05, with critical values = 4.776 for Staphylococcus sp.3. ResultsThe total viable count and staphylococcal counts according to the relevant standards were determined by microbiological culture examination of individual cheese samples (Table 1). Subsequently, identification isolates by PCR method detected 130 isolates of Staphylococcus spp.In the study, subsequently, we carried out identification of the species S. aureus (56 isolates), S. chromogenes (16 isolates) and S. simulans (10 isolates) by the MALDI –TOF MS method. The score value of the identified strains ranged from 2.096 to 2.268 in S. aureus, between 2.076 and 2.105 in S. chromogenes and from 2.002 to 2.224 in S. simulans.A total of 82 Staphylococcus isolates, identified by MALDI-TOF MS, were subjected to partial 16S rRNA gene sequencing. All PCR products showed the expected size (420 bp). The 16S rRNA sequences were compared with the sequences deposited in GenBank. Sequence similarity values greater than 99% (≥99%) were considered reliable according to the CLSI (2018). In all cases, the sequence similarities with GenBank sequences ranged from 99 to 100%. The specific results of the partial sequencing of 16S rRNA and MALDI-TOF MS identification are shown in Table 2.Because goat and sheep milk products are a good substrate for the growth of resistant staphylococci, we also examined the identified isolates for their resistance to selected antibiotics. Bacteria of Staphylococcus spp. (82 isolates) showed the highest resistance to penicillin (98%; 80 isolates) (Table 3).In particular, resistance to penicillin (100%; 16 isolates), oxacilin (94%; 15 isolates), erythromycin (63%; 10 isolates) and tetracycline (75%; 12 isolates) was detected in S. chromogenes. A 100% sensitivity has only been reported for teicoplanin in these species. S. simulans strains, similar to S. chromogenes, showed 100% resistance to erythromycin, but no strain was resistant to tetracycline, teicoplanin and ofloxacin. S. aureus was also confirmed to have the highest resistance to penicillin (96%; 54 isolates) and oxacilin (98%; 53 isolates) than S. simulans and S. chromogenes. In this species, each strain was detected to be resistant to at least one antibiotic. Intermediate susceptibility in S. aureus was confirmed in 19% of isolates against ceftaroline, teicoplanin, erythromycin and 43% of isolates against gentamicin. The occurrence of resistant strains for each type of cheese is listed in Table 4.Based on the phenotypic manifestation of antimicrobial resistance, the presence of MRS strains was assumed, which we confirmed by PCR detection of the mecA gene. The presence of strains with the mecA gene was detected. All positive strains belonged to S. aureus (19%; 10 isolates). In S. chromogenes and S. simulans, mecA gene occurrence was not confirmed by PCR.Multiresistance was also confirmed in the examined strains (Figure 1) in S. aureus in 73%, S. simulans (80%) and S. chromogenes (87%). Most often, resistance to two antibiotics simultaneously (PEN-E) was detected. Multidrug resistance to six antibiotics (PEN-KF-E-GN-TE-OFX) was also confirmed in one strain of S. aureus.4. DiscussionThe basic requirement of the entire food chain is food safety. Each part of the chain is responsible for safe food. In the Dolezalova et al. [22] study, they describe the total number of microorganisms as a general indicator and the examined cheese samples ranged up to 108 CFU/g. Regulation (EC) No. 2073/2005 [23] and Government Regulation (EC) No. 312/2003 [24] lays down production limits for raw sheep’s and goat’s milk but not specifically for dairy products produced from raw sheep’s milk. However, the TVC values determined in the cheeses indicate a possible contamination during the production of such cheeses, which must be eliminated in the production process by the correct setting of the HACCP system in the operation of Carrascosa et al. [25]. Higher TVC in cheese samples was also observed in our study. This may be due to the ripening temperature of the cheeses at 30 °C, which is favorable for the growth of beneficial microflora but also for contaminating microflora [26]. Similarly, higher TVC values in cheese samples were recorded by Carrascosa et al. [25].Microbiological examination of cheese samples and subsequent identification of bacteria by PCR confirmed the presence of bacteria of the genus Staphylococcus spp. The 16S rRNA gene serves as an excellent target for most staphylococci. For some coagulase-negative staphylococci, the separation between species can be difficult due to the lack of sufficient heterogeneity within the 16S rRNA gene [27]. The obtained bacteria were subjected to species identification by MALDI-TOF MS, which confirmed the presence of S. aureus, S. simulans, S. chromogenes. Chen et al., Deng et al. and Cheng et al. [28,29,30] claimed that a score value of 2.000–2.300 in the MALDI-TOF MS indicated a highly probable identification of the species, a score value in the range of 1.7–1.999 was considered an identification of the genus and probable identification of the species, while a score value of 1.699–0.000 was considered a reliable identification. Based on these facts, we can confirm the exact species identification of isolates tested by us.Similar confirmation was reported by Prod’hom et al. [31]. These authors used blood culture to propagate staphylococci that were then identified by MALDI-TOF MS. In this way, they were able to identify 25 isolates of S. aureus. Other groups have also confirmed the high efficiency of staphylococci identification at the species level only using only MALDI-TOF MS [32,33,34]. Clerc et al. [35] also identified coagulase-positive staphylococci by MALDI-TOF MS with a high score over 2.000.Since the bacterium of the genus Staphylococcus has recently shown antimicrobial resistance, in our studies, to proceed to the detection of strains identified by us, it was detected in up to 64% of isolates. Similarly, Jamali et al. [36] reported that staphylococci isolated from raw milk and milk products was highly resistant to tetracyclin and penicillin but susceptible to oxacillin, lincomycin, klindamycin, erythromycin, streptomycin, cefoxitin, kanamycin, gentamycin and chloramphenicol. Kraemer et al. [37] confirmed a more frequent resistance to erythromycin compared to other antimicrobial agents. The agar dilution method showed that isolates of the Staphylococcus spp. resistance to macrolides (erythromycin) was higher (75%; 60 isolates).Resistance to β-lactam antibiotics in staphylococcal isolates was also confirmed by Sampimon [38] who found that resistance to penicillin was 18% in S. chromogenes. Similar proportions of β-lactamase production for S. chromogenes (18%) were also reported in a US study [39].Persson-Waller et al. [40] also confirmed resistance to β-lactam antibiotics in S. chromogenes (33%) in their study.Erythromycin resistance was also detected in S. chromogenes and S. simulans by Lüthje and Schwarz [41]. Erythromycin resistance in food isolates of S. simulans was also confirmed by Chajecka-Wierzchowska et al. [42]. At the same time, they confirmed tetracycline resistance in these isolates.The lowest resistance to teicoplanin was confirmed. Vasiľ et al. [43] confirmed a higher resistance to erythromycin (12%) in S. aureus. Intermediate susceptibility of erythromycin was confirmed in 14 isolates. For other antibiotics, the incidence of resistant strains of S. aureus was relatively low.Overall, common resistance to β-lactam antibiotics and the presence of the mecA gene were found in our study, similar to the study by Rajala-Schultz et al. [44] and Persson-Waller et al. [40] who tested the staphylococcal strains in raw milk and mastitis.The prevalence of the presence of the mecA gene varied markedly between CNS species and was significantly higher in S. epidermidis and S. haemolyticus (∼40%) than in S. simulans and S. chromogenes. Originally, methicillin-resistant S. aureus (MRSA) were detected primarily in humans, only later they were found also in animals [45]. Recently, the increase in resistance to methicillin, penicillin and oxacillin of staphylococci strains constitutes a serious clinical and epidemiological issue. The presence of methicillin-resistant S. aureus in milk is not considered a serious food safety issue as milk is commonly heat treated before consumption. However, there are also exceptions involving the consumption of raw milk by farmers and making milk products from unpasteurized milk [46]. Such practice may expose people to MRSA. Recent reports revealed that MRSA was also associated with cases of bovine and caprine mastitis [47,48]. Bogdanovičová et al. [49] also confirmed the presence of the mecA gene in 9.7% of the examined S. aureus strains examined.In our study, multidrug resistance was also detected in the isolates. Resistance to more than one antibiotic was also confirmed by Persson-Waller et al. [40], where multidrug resistance occurred in 9% of a total of 56 staphylococcal isolates. Nunes et al. [50] reported that out of 19 CNS strains, 14 showed multiresistance to antimicrobial agents. The authors detected resistance to β-lactams (oxacillin, penicillin and/or cefoxitin) and to vancomycin in 73% of the total CNS strains identified. Resistance to tetracycline and gentamicin was confirmed in nine strains (64%), neomycin, erythromycin and chloramphenicol in eight strains (57%), sulfamethoprim in seven strains (50%), linezolid in five strains (36%), rifampicin in three strains (21%), ciprofloxacin and cefepime in two strains (14%) and, finally, only one strain (7%) was resistant to clindamycin.Recently, attention has been focused mainly on resistance to penicillin-stable penicillins, which is referred to as “methicillin resistance” or “oxacillin resistance”. Most resistance to methicillin (oxacillin) is mediated by mecA encoding PBP2a. Isolates, in which the presence of the mecA gene is confirmed, should be classified as resistant to methicillin (oxacillin) according to CLSI [23].The multiresistance of CNS strains reported by the above authors is consistent with the observations obtained in the previous studies on coagulase-negative and coagulase-positive staphylococci that revealed the presence of several resistant and multiresistant staphylococcal strains in raw milk and products from unpasteurized milk [51].Our results are supported by the study by Seng et al. [52] who reported as high as 80% resistance to several types of antibiotics in staphylococci isolates. The study by Hleba et al. [53] also reported the presence of multiresistant bacteria in milk and milk products.It is common belief that the increased resistance to antimicrobials is associated with the extensive and sometimes even unreasonable use of these drugs in human and veterinary medicine [54].5. ConclusionsInvestigation of selected milk products brought new information on the prevalence of resistant, multi-resistant staphylococci and MRSA strains in sheep and goat cheeses produced from unpasteurized milk in Slovakia. Our study also indicated the need for rapid and effective identification of individual staphylococcal species because of differences in antimicrobial profiles of individual species and their relevance from the point of view of protecting the health of humans and animals.

animals : an open access journal from mdpi

10.3390/ani13050863

PMC10000030

Newly weaned pigs are typically fed combinations of cereals and proteins to maximise performance. In the absence of the use of certain antimicrobial compounds, combinations of cereals and protein sources can also be used strategically to reduce dysbiosis in the gastrointestinal tract. This experiment examined the impacts of offering either medium-grain or long-grain extruded rice or wheat, in combination with animal or vegetable protein sources, on postweaning performance, shedding of β–haemolytic Escherichia coli, and the coefficient of total tract apparent digestibility (CTTAD) in the 21 days after weaning. The experimental findings confirmed that extruded rice is an excellent cereal for young pigs but that vegetable protein sources decreased production in weeks two and three compared to the use of animal protein sources. Vegetable protein sources decreased the faecal E. coli score. The CTTAD of dietary components differed according to interactions between cereal and protein sources.

Different cereal types, in combination with different protein sources, are fed to pigs after weaning, but their interactions and possible implications are not well researched. In this study, 84 male weaned piglets were used in a 21-day feeding trial to investigate the effects of feeding either medium-grain or long-grain extruded rice or wheat, in a factorial combination with protein sources of either vegetable or animal origin, on postweaning performance, shedding of β–haemolytic Escherichia coli, and the coefficient of total tract apparent digestibility (CTTAD). Pigs fed either rice type performed the same (p > 0.05) as wheat-fed pigs after weaning. The use of vegetable protein sources reduced growth rate (p < 0.001) and feed intake (p = 0.007) and deteriorated the feed conversion ratio (p = 0.028) in weeks two and three compared to pigs fed animal protein sources. The number of antibiotic treatments given for clinical diarrhoea was similar (p > 0.05). However, the faecal E. coli score showed a trend for the main effect of protein source, with pigs fed animal proteins showing a higher E. coli score than pigs fed vegetable proteins (0.63 vs. 0.43, p = 0.057). There was also a tendency for an interaction (p = 0.069) between cereal type and protein source (p = 0.069), with this difference being associated with a greater faecal score in pigs fed diets with long-grain rice plus animal proteins and wheat plus animal proteins. Significant interactions occurred for the CTTAD when assessed in week three. In general, pigs fed diets with medium-grain rice or long-grain rice with animal proteins had a higher (p < 0.001) CTTAD for dietary components than pigs fed all other diets, and vegetable proteins depressed (p < 0.001) CTTAD compared to animal proteins (main effect of protein: p < 0.001). In summary, pigs tolerated the extruded rice-based diets well and performed equivalently to pigs fed wheat as the sole cereal, and the use of vegetable proteins decreased the E. coli score.

1. IntroductionPostweaning diarrhoea (PWD), typically caused by the activity of enterotoxigenic and (or) enteropathogenic strains of Escherichia coli in the small intestine, remains a problem in commercial pig production in some parts of the world. This is exacerbated by bans or restrictions on the use of antimicrobial compounds such as medicinal zinc oxide and prophylactic antibiotics [1,2]. In association with poorer performance after weaning, the postweaning growth check significantly impacts whole-of-life productivity. A plethora of nutritional strategies exists commercially to alleviate the negative impacts of weaning in the absence of these antimicrobial compounds [3,4,5]. Among the numerous dietary options available to the industry, the choice of cereal and (or) protein sources in diets can have a major impact on PWD and performance. In this regard, rice, either as the sole cereal or in combination with others, can be used successfully in piglet diets as an alternative to other cereals such as maize, sorghum and wheat [6,7,8,9,10,11,12,13] and can reduce the incidence of PWD and decrease postweaning mortality [14]. Rice is characterised by its high starch content and varietal differences in the ratio of amylose:amylopectin, low non-starch polysaccharide (NSP) and oligosaccharide contents, and lower protein content in comparison with other cereals [15,16,17].Newly weaned pigs are typically fed a mixture of plant (vegetable) and animal protein sources to provide amino acids and energy for body growth and development. Animal protein sources are generally used more efficiently by the weaned pig than vegetable protein sources [18] due mainly to the lack of anti-nutritive factors present and associated negative physiological effects in the gastrointestinal tract (GIT) [19,20]. Past studies have typically shown differences between individual cereal sources and protein sources on postweaning performance and aspects of GIT structure and function and PWD; however, we are unaware of studies specifically comparing different types of extruded white rice with wheat in conjunction with either animal or vegetable protein sources.The hypotheses tested in this study were: (a) rice-based diets using animal protein sources will cause less excretion of β–haemolytic E. coli than diets based on vegetable proteins; (b) diets based on extruded rice will cause faster growth and less faecal excretion of β–haemolytic E. coli than a diet based on wheat; and (c) diets based on a medium-grain rice, with a lower amylose:amylopectin ratio, will cause better performance, irrespective of the protein source, than diets based on a long-grain rice having a higher amylose:amylopectin ratio.2. Materials and Methods2.1. Ethics StatementThe Murdoch University Animal Ethics Committee and the Animal Ethics and Experimentation Committee of the WA Department of Agriculture approved this experiment (Number 05-02-22).2.2. Animals, Procedures, and HousingEighty-four entire male pigs (Large White x Landrace) aged ~24 days and weighing 6.7 ± 0.13 kg (mean ± SEM) were used in this experiment. On arrival at the Medina Research Centre, the pigs were ear-tagged, weighed, and stratified into pens of four pigs, each according to treatment and live weight. Pigs were offered their respective diets (Table 1) in groups of four for the first 7 days after weaning to accustom them to their new surroundings. For the final 2 weeks of the study, pigs were housed individually for the collection of faeces from the pen floor. Pens were of wire-mesh construction with slatted metal floors and measured 1.68 m2 in floor area (0.42 m2 per pig). Each pen was equipped with a nipple water drinker and a stainless-steel feed trough. The ambient temperature was maintained between 26 and 28 °C throughout the study using two reverse-cycle air conditioning units. The room containing the pens was cleaned daily.2.3. Experimental Design, Diets, Feeding, and Sample CollectionThe 21-day experiment was designed as a 3 × 2 factorial arrangement of treatments with the respective factors being (a) three cereal types, i.e., a medium-grain, lower-amylose rice (cultivar Amaroo), a long-grain, higher-amylose rice (cultivar Doongara), and wheat, and (b) two protein sources, namely vegetable and animal protein sources (Table 1). Diets are subsequently referred to as: MGAP: medium-grain rice plus animal proteins; MGVP: medium-grain rice plus vegetable proteins; LGAP: long-grain rice plus animal proteins; LGVP: long-grain rice plus vegetable proteins; WAP: wheat plus animal proteins; WVP: wheat plus vegetable proteins. None of the diets contained any antimicrobial compounds.Isonitrogenous and isoenergetic diets were formulated to contain adequate levels of energy and nutrients for pigs of this genotype and age. Extruded rice was sourced as described previously [16] and was passed through a hammer mill to reduce particle size before incorporation into meal-based diets. The diet was offered on an ad libitum basis to piglets in mash form (Table 1). Titanium dioxide (TiO2) was added as an inert marker for the estimation of the coefficient of total tract apparent digestibility (CTTAD). Faecal samples were collected from the wire-mesh floor of each pig at 0800, 1000, 1200, 1400, and 1600 h on days 18–21 of the experiment. Samples collected over the 3-day period were pooled, kept at −20 °C and later thawed, mixed, freeze-dried, and ground through a laboratory hammer mill (1 mm screen) prior to chemical analysis.2.4. Microbiological AssessmentsFaecal swabs were taken to record initial E. coli presence upon arrival and then again on days 2, 5, 6, and 8 after weaning. Faecal swabs were cultured, and plates were assessed for β-haemolytic colonies displaying morphology characteristic of E. coli following overnight incubation, according to standard procedures [21]. The presence of β-haemolytic E. coli was then scored from no growth (0) to heavy colonisation (5) [21]. Piglets were monitored daily for clinical signs of diarrhoea [7]. Affected pigs (as assessed by the stockperson, who was unaware of the treatment allocation of pigs) were treated for diarrhoea by intramuscular injection with Trisoprim-480 ((trimethoprim 80 mg/mL, sulfadiazine, 400 mg/mL), 1.5 mL/30 kg body weight; Troy Laboratories, Smithfield, NSW, Australia); treatment continued until the diarrhoea ceased. Records were kept of the duration of treatment required for each treated piglet.2.5. Chemical AnalysesThe dry matter (DM), nitrogen (N), gross energy (GE), total starch, resistant starch (RS), amylose and amylopectin contents of extruded rice were determined as described previously [11]. Crude protein (CP) content was calculated as N × 6.25. The DM, GE, and TiO2 content of diets and faecal samples were determined for the estimation of CTTAD [11]. The GE content of the rice, diet, and faecal samples was determined using a Ballistic Bomb Calorimeter (SANYO Gallenkamp, Loughborough, U.K.). The TiO2 contents of diet and faecal samples were determined using the method described previously [22].2.6. Statistical AnalysesTreatment effects were analysed by two-way ANOVA for a factorial design, with the main effects being cereal type (medium-grain rice, long-grain rice and wheat) and protein type (vegetable and animal). Average daily gain (ADG) in the first week after weaning was evaluated using the pen as the unit of replication. The ADG, average daily feed intake (ADFI) and feed conversion ratio (FCR) in weeks two and three used the individual pig as the experimental unit. For the CTTAD of DM, starch, GE, and CP, the individual pig was considered the unit of replication. All effects were considered fixed effects in the model. Fisher’s-protected least significant difference test was used (at a 5% significance level) for comparison between mean values of different variables. A p-value between 0.05 and 0.1 was considered a trend. All statistical analyses were conducted using the statistical package StatView 5.0 for Windows (AddSoft Pty. Ltd., Woodend, VIC, Australia).3. Results3.1. Antibiotic Treatments and Faecal Shedding of E. coliBoth rice-based diets fed with vegetable proteins had fewer antibiotic administrations given for clinical diarrhoea than pigs fed diets WAP or WVP, but statistically, the number of treatments was similar (p > 0.05) across treatments. Shedding of β–haemolytic E. coli, ascertained via faecal swabs, showed a trend for the main effect of protein source, with pigs fed animal proteins having a higher E. coli score than pigs fed vegetable proteins (0.63 vs. 0.43, p = 0.057). There was also a tendency for interaction between cereal type and protein source (p = 0.069), with this difference being associated with the greater score recorded in pigs fed diets LGAP and WAP (Table 2).3.2. Production PerformanceThere were no statistically significant differences between treatment groups for ADG in the first week after weaning (Table 3). In weeks two and three, no main effects of cereal type on any indices were recorded (p > 0.05). However, pigs fed animal proteins rather than vegetable proteins were heavier (p = 0.01) at the end of the experiment (11.8 vs. 10.4 kg) because they grew faster (317 vs. 242 g/day, p < 0.001). This was a consequence of a higher ADFI (580 vs. 500 g/day, p = 007) and improved FCR (1.87 vs. 2.31, p = 0.028). No interactions (p > 0.05) occurred for any of the production indices (Table 4).3.3. Coefficient of Total Tract Apparent DigestibilitySignificant main effects for cereal type (rice vs. wheat) were observed in the CTTAD for DM, energy, and CP. Nevertheless, significant interactions occurred between cereal type and protein source for DM, energy, and CP. The CTTAD for DM was higher in pigs fed diets MGAP and LGAP than in the wheat-based diet (WAP) (0.92 and 0.92 vs. 0.85, p < 0.001); however, DM digestibility was similar (p > 0.05) between all three diets when vegetable proteins were fed to pigs rather than animal proteins (0.83, 0.82 and 0.80 for diets MGVP, LGVP, and WVP, respectively). A similar interaction occurred between cereal type and protein source for the CTTAD of energy, with diets MGAP and LGAP having the highest coefficients compared to diet WAP (0.92 and 0.91 vs. 0.83, p < 0.001) (Table 5).In general, the CTTAD for starch was very high in all diets (range 0.989 to 0.999) and higher (main effect, p < 0.001) in both rice-based diets than in the wheat-based diets. The CTTAD for starch was higher in diet WVP than in diet WAP (0.993 vs. 0.989), which resulted in a significant interaction (p < 0.001). The CTTAD for CP was higher in pigs fed diet WVP compared to those fed diets MGVP and LGVP (0.76 vs. 0.67 and 0.66, respectively, p = 0.016) (Table 5).4. DiscussionThe hypotheses proposed in this study were, in general, supported by the experimental findings. Feeding vegetable proteins (as a main effect) showed a tendency to reduce faecal shedding of β-haemolytic E. coli in the first 8 days after weaning compared to pigs fed animal protein sources, as evidenced by the greater faecal swab score recorded in pigs fed diets LGAP and WAP. This suggests that in these two cereal sources, the presence of vegetable proteins reduced intestinal colonisation of pathogenic E. coli. Within a few days of piglets being weaned, enterotoxigenic and (or) enteropathogenic strains of E. coli may proliferate within the intestinal tract and induce diarrhoea [4]. Virtually all E. coli strains that cause PWD produce an alpha-haemolysin and show characteristic haemolysis on blood agar. These haemolytic strains also produce virulence factors that allow them to adhere to enterocytes (e.g., F4; F18; AIDA (Adhesin Involved in Diffuse Adherence)), and they generate one or more toxins (e.g., stable toxin (ST)a; STb; labile toxin (LT); Enteroaggregative E. coli heat-stable enterotoxin (EAST1)) that are responsible for inducing the diarrhoea [23]. In the current study, we did not investigate the individual attributes of the recovered E. coli strains but relied instead on the presence of characteristically strong β-haemolysis as a marker of strains capable of causing PWD.Feeding extruded rice failed to translate into a reduced number of therapeutic antibiotic treatments given for clinical diarrhoea as this was statistically the same for all treatments, although it was evident that pigs fed both rice-based diets with vegetable proteins received fewer antibiotic administrations than pigs fed diets WAP or WVP. The incidence of PWD was generally low in this study, which contributed to the lack of statistically significant differences in these indices between diets. These results concur in part with those reported previously, where no correlation was found between faecal shedding of β-haemolytic E. coli and the number of antibiotic treatments required for PWD in the first 14 days after weaning in pigs fed either rice- or wheat-based diets [24], suggesting that while the presence and activity of enterotoxigenic E. coli (ETEC) are of central importance in the aetiology of PWD, dietary, and (or) physiological contributions also may have an important impact on disease expression in post-weaned piglets.The vegetable proteins used in the current study contained considerable levels of soluble non-starch polysaccharides (NSP) and oligosaccharides, which have been shown to exacerbate the shedding of E. coli and cause diarrhoea [4,20,21]. Data from the present study showing an ameliorative effect of feeding vegetable proteins on faecal shedding of β-haemolytic E. coli suggesting that the types and quantities of dietary fibre (DF) fed in the postweaning period have a significant impact on the expression of diarrhoea. In this regard, data from the current study contrast with previous work using rice-based diets where the addition of different sources of DF increased the number of antibiotic injections required for the treatment of diarrhoea [25].In some previous studies, increased diarrhoea and faecal E. coli scores have been observed in pigs fed very highly digestible rice-based diets after weaning [10,11], but the use of a (mostly) insoluble NSP source such as oat hulls was found to ameliorate the condition [10,26]. Oat hulls included in an extruded rice-based diet with animal proteins decreased total biogenic amine concentrations commensurate with lower plasma urea concentrations [26], suggesting that oat hulls were able to decrease diarrhoea where a misbalance of carbohydrate to protein entering the hindgut may occur. Alternatively, adding oat hulls to a rice-based diet might not influence fermentation behaviours in the large intestine due to its highly insoluble and lignified nature, but rather, it may have modified motility and transit time of digesta that, in turn, reduced the availability of substrate for bacterial growth [10]. While DF was added to diets in the current study either in the forms of vegetable proteins or wheat rather than oat hulls, the observations are generally consistent with the current industry philosophy that the inclusion of fibre sources in postweaning diets aids in reducing the proliferation of pathogenic bacteria such as ETEC [1,3,27,28].Weaned piglets fed extruded medium-grain rice or long-grain rice performed similarly to pigs fed wheat in the first 3 weeks after weaning, indicating that extruded rice can replace wheat as the sole cereal in piglet diets after weaning. These data are consistent with previous studies [6,7,8,9,10,11,12,13,14] and reinforce the excellent nutritional value of rice for young pigs, especially with mild cooking that can enhance diet digestibility and ileal morphology [29]. Pigs fed rice-based diets also display lighter gastrointestinal organ weights and a greater carcase weight [11]. Vegetable protein sources in the diet depressed growth rate and feed intake and caused a deterioration in FCR in weeks two and three of the study compared to animal protein sources. These data are consistent with numerous previous observations. Weanling pigs fed a mixture of animal proteins (whey-protein concentrate and fish meal) performed better in the 2 weeks after weaning than pigs fed a mixture of plant proteins, including soybean meal, fermented soy protein and rice protein concentrate [18], commensurate with higher digestibilities of energy, DM, and CP. Moreover, post-weaned pigs fed an increased content of animal protein in an extruded rice-based diet displayed improved performance [30]. Animal proteins are more digestible than vegetable proteins [13,18,25], which are richer in anti-nutritive carbohydrate fractions, and hence more nutrients became available for body growth and development.Seemingly in contrast to some previous findings, Montagne et al. [21] reported no differences in ADG or FCR when pigs were fed diets based on cooked (autoclaved) white rice with either animal or vegetable proteins. The reason(s) for this difference is (are) hard to explain but could be attributable to the fact that these authors infected pigs experimentally with F4:ETEC that could have disturbed the intestinal milieu associated with digestion and absorption, and (or) there was an effect of cooking form on digestibility and subsequent growth rate. Extruded rice contains very low levels of RS, whereas cooked (autoclaved) white rice that is then cooled contains approximately 20 times the RS content of extruded rice [16]. The RS level of the extruded products was not considered in the derivation of the energy value of extruded rice used in the formulation of these diets; hence, there could have been a misbalance in energy contributions between the protein sources and the extruded rice that contributed to the inferior performance of the pigs fed vegetable proteins. In this regard, Montagne et al. [21] found no difference in faecal shedding of β–haemolytic E. coli in pigs fed a wheat and vegetable-proteins-based diet compared to pigs fed medium-grain rice-based diets with either animal or vegetable proteins. This may also reflect the difference in the form (i.e., extrusion vs. autoclaving) of rice used in the different experiments.Be this as it may, the presence of significant interactions between cereal and protein sources for CTTAD in the current study indicates that the dietary component responded differently to both dietary factors. For example, CTTAD for energy was significantly higher in both extruded rice-based diets irrespective of whether animal or vegetable proteins were added compared to diets WAP and WVP, whereas for CP digestion, the significant difference was caused by an apparently higher digestibility in pigs fed diet WVP compared to pigs fed diets MGVP and LGVP. It is difficult to explain the higher CP digestibility in pigs fed diet WVP compared to pigs fed diets MGVP and LGVP, given the higher DM and energy digestibilities observed in the extruded rice-based diets when vegetable protein sources were added. This might be attributable to a higher formation of microbial protein, causing an overall depressed total tract digestibility. Interpretation of total tract apparent digestibility coefficients for CP is fraught regardless because the formation of protein by the microbiota provides no real indication of the ileal digestibility of CP and absorption of amino acids.Rice-based diets fed to piglets after weaning provide a ready form of energy in the form of glucose through starch digestion. In a previous study [11], the effects of different types of cooked white rice on starch digestion, digesta and fermentation characteristics, shedding of β–haemolytic E. coli, and performance after weaning were examined. Pigs received one of three rice-based diets: (i) medium-grain, (ii) long-grain, and (iii) waxy, all with animal protein sources, and a fourth diet contained mainly wheat, barley, and Australian sweet lupins. The apparent digestibility of starch measured in the ileum 14 d after weaning was highest in the medium-grain and waxy rices containing the lower amylose contents and lowest, but the same, with the other two cereal sources, similar to findings in the current study, albeit that digestibility was measured in faeces. Starch digestibility in faeces was highest in all rice diets, and digesta viscosity was highest in pigs fed the wheat-based diet in both the ileum and caecum [11].5. ConclusionsWeaned piglets fed the extruded rice-based diets with either animal protein sources or vegetable protein sources performed equivalently to pigs fed either of the wheat-based diets in the first 3 weeks after weaning, confirming that extruded rice is an excellent cereal for young pigs. The use of vegetable proteins compared to animal proteins decreased production performance in weeks two and three after weaning. Vegetable protein sources displayed a trend to reduce faecal E. coli shedding, as evidenced by lowered faecal scores, implicating a role for implicated a role for DF sources in PWD. The CTTAD of dietary components differed according to interactions between cereal and protein sources, although the CTTAD of DM, starch and energy was generally improved by the use of extruded rice compared to wheat and was increased by the use of animal rather than vegetable proteins.

animals : an open access journal from mdpi

10.3390/ani12050589

PMC8909795

Acute phase proteins (APPs) are useful markers which can be evaluated in animals to assess health status and characterize inflammation, infection, and trauma. Among APPs, serum amyloid A (SAA) has been widely investigated in pets and food-producing species as a possible biomarker of inflammatory and infective conditions, especially in the field of animal reproduction. The aims of this paper are to review the literature available on the use of SAA for the diagnosis and monitoring of inflammatory reproductive disease in animals, critically appraising the usefulness of such marker and summarizing the current state of knowledge.

The investigation of acute phase proteins in veterinary medicine has opened the doors towards the identification and use of new markers for a timely assessment of health status in both companion and food-producing animals. The aim of this paper is to review the literature available on the use of serum amyloid A (SAA), an acute phase protein, for the diagnosis and monitoring of reproductive disorders in animals. This review critically appraises the usefulness of such marker in clinical practice and summarizes the current state of knowledge. Recent advances in the diagnosis and monitoring of reproductive diseases are presented, highlighting where SAA evaluation may enhance early diagnostic tools for dogs, cats, cattle, and equines.

1. IntroductionAn ongoing challenge in both human and veterinary medicine is discovering new biomarkers, for early identification of subclinical disease, while being able to identify healthy and ill animals [1]. Biomarkers are biological molecules (usually proteins) found in cells, tissues, and body fluids (blood, urine, feces, exhaled breath) that can be quantified as indicators of physiological and pathological conditions [2]. Ideally, a biomarker should be easily detectable and effective in both identifying the onset and monitoring the progress/outcome of the disease [2].Acute phase response (APR) is a systemic and dynamic process, including a wide range of pathophysiological responses, such as fever, leukocytosis, hormone alterations, and muscle protein depletion, which combine to minimize tissue damage while enhancing the repair process [3]. During APR, specific plasma proteins, known as acute-phase proteins (APPs), which are mainly secreted by the liver to provide an early and non-specific protection against insults [3]. In APR, cytokines act as messengers between the site of disorder/disease and the hepatocytes that synthetize the APPs [4].APPs represent a large and wide-ranging group of serum non-specific proteins, unrelated to immunoglobulins [5]. APPs may increase due to multiple causes unrelated to inflammation, such as a response to transportation, stress, malnourishment, and parturition, as well as other a stress-related causes. Both an increase in APPs and serum glucocorticoid levels occur with such conditions [6]. According to the definition of APPs, the concentration will increase by >25% in response to proinflammatory cytokines. It has been suggested that APPs are biomarkers of inflammation, infection, and trauma in human and veterinary medicine [3,6,7,8]. According to variations in serum when measuring APPs, they can be labelled as “positively reacting” or “negatively reacting”. Positive APPs, such as serum amyloid A (SAA), haptoglobin (Hp), C-reactive protein (CRP), and fibrinogen, show a “major”, “moderate”, and “minor” increase in their serum level upon initiation of APR. Major APPs, such as SAA, Hp, and CRP, are characterized by a concentration usually below 1 μg/mL in healthy animals which dramatically increases up to 1000 fold following an inflammatory stimulus, reaching a peak around 24–48 h and then decreases rapidly upon recovery. Moderate APPs as alpha1 acid glycoprotein (AGP) increase from 5- to 10-fold following an inflammatory stimulus, reaching a peak after 2 or 3 days and then decrease more slowly than other major APPs. Minor APPs, whose utility is still debated in veterinary medicine, gradually increases by 50 to 100% of its quiescent level [7]. Negative APPs, such as albumin and transferrin, decrease during the inflammatory response and, apart from albumin, have limited use in clinical pathology [7].Among APPs, SAA has been widely investigated in different species, including dogs, cats, horses, and ruminants, as a possible biomarker of inflammatory and infective processes, especially in the field of animal reproduction [9,10] (Table 1 and Table 2).Reproductive failure is one of the most significant factors that limit the productivity of animal production systems and negatively influence the welfare of pets and their owners [11]. For dogs and cats, subclinical endometritis, endometrial hyperplasia, and pyometra are the most frequently disorders faced in clinical practice [12]. While endometritis is a major cause of subfertility and infertility in horses and cows [13,14,15], subclinical mastitis represents the major reproductive disorder in sheep flock [16]. Since SAA has been evaluated in different aspects of domestic animals’ reproduction, the aim of this paper is to provide an updated review on the clinical relevance of using SAA in the detection and monitoring of inflammatory and infective diseases of reproductive importance in such species.animals-12-00589-t001_Table 1Table 1Studies about serum amyloid A evaluation in pets, ruminants, and mares.SpeciesAnimals/Disease/ConditionReferenceDOGHealthy bitches and bitches with pyometra underwent ovariohysterectomyDabrowski et al., 2006 [17]Bitches with pyometra, with normal and complicated post-ovariohysterectomy periodDabrowski et al., 2009 [18]Healthy bitches and bitches with pyometraHagman et al., 2009 [19]Healthy bitches and bitches suffering from systemic inflammation, including pyometraChristensen et al., 2014 [20]Septic and non-septic bitches with pyometraJitpean et al., 2014 [21]Healthy bitches and bitches with open- and closed cervix pyometraDabrowski et al., 2017 [22]MastitisKaszak et al., 2018 [23]Mammary tumorTecles et al., 2009 [24]CATHealthy cats and cats with pyometraYuki et al., 2020 [25]RUMINANTSHealthy cows and cows with postpartum endometritisChan et al., 2009 [26]Healthy cows and cows with endometritis Biswal et al., 2014 [27]Healthy cows and cows with subclinical endometritisBrodzki et al., 2015 [28]Healthy cows and cows with endometritisKaya et al., 2016 [29]Cows affected by endometritis, before and after treatmentAhmadi et al., 2018 [30]Healthy cows and cows with mastitisEckersall et al., 2001 [31]Mastitis experimentally induced with Streptococcus uberisPedersen et al., 2003 [32]Acute and chronic experimentally induced Staphylococcus aureus mastitisGrönlund et al., 2003 [33]Subclinical mastitis in dairy cowsGerardi et al., 2009 [34]Chronic subclinical mastitis in dairy cowsSubclinical experimentally induced mastitis with Staphylococcus epidermidis in sheepGrönlund et al., 2005 [35]Winter et al., 2003 [36]EQUINEBefore and after inseminationTuppits et al., 2014 [37]Subclinical endometritis Sikora et al., 2016 [38]Before and after artificial insemination in uterine lavage fluid Wojtysiak et al., 2020 [39]Experimentally induced ascending placentitisCoutinho da Silva et al., 2012 [40]Healthy periparturient mares and mares with ascending placentitisCoutinho da Silva et al., 2013 [41]animals-12-00589-t002_Table 2Table 2Comparison of concentration of serum amyloid A on serum samples in healthy and disease, considering the literature currently available.Species Disease/ConditionLiterature/References in Healthy AnimalsLiterature/References in DiseaseReferenceDOGPyometra27.1 ± 13.1 μg/mL184.2 ± 122.3 μg/mLDabrowski et al., 2006 [17]<5 μg/mL61.3 ± 31 (<5–>80) μg/mLHagman et al., 2009 [19]0.55 ± 0.54 μg/mL103.56 ± 35.2 μg/mL 1193.81 ± 20.7 μg/mL 2Dabrowski et al., 2017 [22]Mammary tumor1.69 (1.35–2.05) mg/LStage IV 75.87 (38.13–92.76) mg/LTecles et al., 2009 [24]CATPyometra0 µg/mL154.8 (0.1–182.4) μg/mLYuki et al., 2020 [25]RUMINANTSCattle endometritis48 ± 20 mg/mL 385 ± 23 mg/mL 3Chan et al., 2009 [26]16.80 ± 1.62 μg/mL33.97 ± 2.14 to 35.42 ± 0.58 μg/mLBiswal et al., 2014 [27]30 µg/mL §,450 µg/mL §,4Brodzki et al., 2015 [28]14.24 ± 0.52 µg/mLMild Endometritis = 20.25 ± 0.65,Endometritis = 28.17 ± 1.22,Severe Endometritis = 34.62 ± 1.28 µg/mLKaya et al., 2016 [29]Cattle mastitis5.1 (3.6–11) μg/mLMild mastitis = 13.8 (5.4–142) μg/mLModerate mastitis = 29.9 (5.9–141) μg/mLEckersall et al., 2001 [31]0.47–4.62 μg/mL 510-fold increase 5Pedersen et al., 2003 [32]2.82 ± 1.8 mg/LAcute mastitis = 376.9 ± 352 mg/LChronic mastitis = 11.3 ± 11.0 mg/L Grönlund et al., 2003 [33]50 mg/L §Subclinical mastitis = 104 mg/L §Clinical mastitis = 245 mg/L §Gerardi et al., 2009 [34]Sheep mastitis1.5 (0–29.4) μg/mL207 μg/mL §Winter et al., 2003 [36]EQUINEPost-inseminationBefore: 0.23 (0.05–10.16) mg/LAfter: 0.30 (0.02–1.81) mg/LTuppits et al., 2014 [37]0.001 g/L in uterine lavage fluid 40.0015 g/L in uterine lavage 4Wojtysiak et al., 2020 [39]§ Average on serum samples; 1 open-cervix pyometra; 2 closed-cervix pyometra; 3 a week after parturition; 4 no range was reported within the reference paper; 5 evaluated in milk, no data available on serum concentration.2. SAA in Reproductive Disease of PetsThe most common reproductive disorder affecting bitches is pyometra, a severe form of endometritis [42], often caused by Escherichia coli (E. coli), characterized by pus accumulation into the uterine lumen [43]. The disruption of the cell wall of E. coli causes different concatenated events, including the release of lipopolysaccharides (LPS), neutrophils activation, increased synthesis of proinflammatory cytokines [44], systemic inflammation, septic shock, and death [43]. Despite the availability of different medical conservative treatments, ovariohysterectomy remains the treatment of choice. Therefore, a proper monitoring of pre- and post-operative inflammation in bitches with pyometra becomes essential to promptly detect any sub-sequent pyometra complications, such as sepsis and septic shock, systemic inflammatory response syndrome (SIRS), peritonitis, disseminated bacterial infection, and multi-organ dysfunction [45]. The early identification and appropriate treatment of pyometra is crucial to obtain a favorable outcome which can increase the survival rate of this life-threatening condition; consequently, some authors have assessed the changes in serum APPs in bitches with pyometra. In one evaluation [17], Dabrowski et al. found increased serum levels of SAA in pyometra-affected bitches compared to healthy subjects, as a result of the chronic endometrial inflammation. After ovariohysterectomy, SAA levels slowly reduced during the postoperative period in association with the recovery of the bitches. According to their results, the authors concluded that SAA concentrations can be used as markers of inflammation in bitches with pyometra, as demonstrated also by Jiptean et al. [21], who found increased SAA serum concentrations in dogs with pyometra-induced sepsis. The monitoring of SAA levels after ovariohysterectomy can also provide valuable information about inflammatory response during postoperative course as confirmed by another study by Dabrowski et al. [18] who assessed the usefulness of APPs determinations for monitoring the severity of postoperative inflammatory responses in bitches with pyometra undergoing ovariohysterectomy. In that study, serum levels of different APPs, including SAA, remained over the upper limits or even increased in bitches with postoperative complications, compared to animals undergoing a normal post-operative recovery. Considering this fluctuation, authors concluded that SAA may be used as markers for surgical site infections and that determination of serum APPs is a good prognostic indicator which can enable the early detection of bacterial infections of postoperative wounds in bitches that underwent surgery for pyometra.In another evaluation [22], Dabrowski et al. found that SAA levels reflect the intensity of inflammatory processes in bitches with open- and closed-cervix pyometra regardless of the site of blood collection, suggesting once again the usefulness of evaluating peripheral blood APPs levels as indicators of uterine inflammation, with higher concentrations of APPs in the peripheral blood than in the uterine arterial blood of bitches with closed-cervix pyometra. Similar findings were already noted by Hagman et al. [19], who linked increased SAA concentrations with a strong upregulation of the SAA gene in the uterus, due to inflammatory stimuli.According to the studies by Dabrowski et al., serum levels of APPs can serve as valid indicators of uterine inflammation in bitches with pyometra providing useful information about the progression of recovery after surgical treatment [19].Similarly, Christensen et al. found that SAA is useful as a diagnostic marker of systemic inflammation in dogs, including pyometra-related inflammation [20]. Both markers showed comparable diagnostic capacity and achieved and excellent agreement in their performance. Interestingly, Christensen et al. found that SAA was characterized by a wider range of concentrations and a significantly superior overall diagnostic potential for systemic inflammation than other APPs.APPs has also been evaluated in mammary diseases which represent a frequent health problems of bitches [23]. APPs from serum and milk specimens may be used as accurate inflammatory biomarkers, since their concentrations are significantly higher in bitches suffering from mastitis than in healthy ones [23]. This is particularly important in the case of initial or subclinical mastitis, as standard diagnostic procedures, such as clinical examinations, blood tests, and cytological examinations of mammary gland secretion, may fail to diagnose a condition that, if misdiagnosed or left untreated, can become life-threatening.SAA fluctuation has also been tested in mammary gland tumors that represent up to 52% of neoplasm in female dogs, they and are diagnosed as malignant in up to 50% of cases [46]. In their work, Tecles et al. [24] found that, in female dogs with mammary gland tumors, APP secretion is dependent upon factors such as the presence of metastasis, the large size of the primary mass, ulceration, or the secondary inflammation of the neoplasm. The preliminary results obtained by Tecles et al. deserves to be further explored to better ascertain a possible application of APPs in the monitoring of therapy and long-term prognosis of mammary tumors in female dogs.Limited data are available on the evaluation of SAA levels in cats and the possible correlation with diseases [47,48]. Despite this, in a very recent paper, significant increases in SAA serum levels were observed in cats affected by pyometra [25].3. SAA in Ruminants ReproductionThe need for maintaining high-level performance in dairy herds has encouraged researchers to explore suitable indicators of herd health status, which may be represented by APPs [49,50]. In cows, stress-related factors along with damage to the reproductive tract have been shown to be responsible for increasing the concentrations of SAA, suggesting a possible use to monitor the onset of inflammatory and infective disease in this species [26].Humblet et al. [51] correlated APP concentrations with clinical health status for the diagnosis of disease in the post-parturition period in dairy cattle. The study tested 158 dairy cows from four different herds that underwent clinical and gynecological examination every two weeks over a six-month period; at each time point, serum SAA levels were assessed, and the results were classified as positive or negative based on ad-hoc cutoff points. Although interindividual variability was high, authors found a statistically significant increase in SAA concentrations in the first week after calving, confirming that parturition is associated with a physiologic acute phase response, as previously reported in this species [52,53]. This should be considered when SAA is evaluated as a marker of inflammation within the first week after calving, since it could be challenging to discriminate increases in physiological APPs following normal peripartum from pathological increases related to the onset of inflammatory disease. Authors concluded that, in the postpartum period, SAA was capable of identifying healthy animals; however, its ability to identify animals with ongoing pathologic processes was only fair, and further studies are needed [51].To better ascertain a possible use of SAA as indicators of uterine infection in dairy cows, Chan et al. [26] tested serum fluctuations over time in cows with metritis. Authors collected blood samples from 18 Holstein dairy cows which displayed acute puerperal metritis at stated intervals from one week prepartum to six months post-partum, using six clinically healthy cows as controls. In addition, authors included ten heifers used to highlight the threshold values of normality for SAA, which was 51.9 mg/mL. SAA concentrations increased significantly in cows after calving, confirming the serum increase in APPs in cows within 3 weeks after parturition, as a laboratory sign of postpartum metritis [54,55,56]. The increase in APPs is primarily due to subclinical endometritis which can seriously reflect its impact on reproductive performance [57]. Reproductive performance in cows with normal and increased APPs concentrations was evaluated by Gilbert et al. [49], taking into account the number of days open and the conception rate. Clinically healthy cows had a better conception rate than those affected by postpartum metritis. Furthermore, among the successfully pregnant cows, the number of days open was significantly higher in cows with Hp values above the threshold limit compared to the remaining cows with a normal range. Interestingly, a significant difference in reproductive performance was not observed between cows with SAA values within or exceeding threshold limit described above.Biswal et al. [27] evaluated serum levels for APPs at certain time points during the treatment of endometritis to test the possible usefulness of this marker to monitor the treatment progress when using different immunomodulators. Twenty-one cows diagnosed with endometritis underwent three different treatments for this disease, using seven healthy cows as controls, by collecting serum to evaluate SAA levels pre- and post-treatment. SAA decreased post-treatment, irrespective of the type of treatment and according to these results, authors concluded that SAA might serve as reliable biomarkers in the diagnosis and monitoring of endometritis in dairy cows.Brozki et al. [28] found that even cows affected by subclinical endometritis in the late post-partum period (sixty days after parturition) showed higher serum levels of inflammatory cytokines and APPs in comparison to healthy animals. SAA concentrations increased only in serum while it was not detected in uterine fluid; in fact, the local production of this protein in endometrial cells has only been postulated and not confirmed yet.A study by Kaya et al. [29] was carried out with the aim of providing a simple and effective tool to assess health herd, by correlating serum SAA levels and serum ceruloplasmin with endometritis which display various degrees of severity in cows. A total of 100 Brown Swiss cows, at 28–32 days postpartum, were divided into two groups—healthy (no endometritis) or endometritis (mild, moderate, and severe) based on ultrasonography, vaginoscopy, and cytology. According to study results, SAA and ceruloplasmin concentrations were significantly higher in cows with endometritis than in healthy controls, and positively correlated with the severity of the endometritis, by confirming the utility of APPs as markers of endometritis in cows.In 2018, Ahmadi et al. [30] examined the fluctuation of serum SAA in 81 lactating dairy cows affected by postpartum clinical endometritis 30 days after calving. Animals were treated with hyperimmune serum, and the serum concentration of SAA was evaluated before and two weeks after therapy, by showing significant lower levels of this APP following treatment. In addition, the increase in serum levels of SAA was directly correlated to the total milk production per standard lactation (305 days). Ahmadi and collaborators stated that hyperimmune serum administered to treat clinical endometritis could decrease SAA in dairy cows. Interestingly, factors such as milk production and pregnancy status could increase serum levels of SAA.APPs were found to be useful to assess both acute and subclinical mastitis in dairy cows. SAA milk concentrations were effective indicators of healthy udder quarters; they were within the normal limit in healthy mammary glands and increased in chronic subclinical mastitis [31,32,33,34,35].Mastitis is a common problem in cattle and sheep herds, representing a significant welfare and financial issue in farming [16]. This is particularly true in the case of subclinical mastitis which is still hard to promptly diagnose [58,59].In sheep, SAA has been preliminary appraised as a marker of udder health. Winter et al. [36] experimentally induced subclinical mastitis in ewes by inoculation of Staphylococcus epidermidis into the udder and then evaluated SAA concentrations in both serum and milk. Although not all the infected ewes showed an increase in serum SAA 24 h after bacteria inoculation, the SAA concentration peaked in milk between 24 and 48 h by returning to levels slightly above the control values within one week. Interestingly, Winter et al. found that basal serum level of SAA in healthy sheep was like those recorded in cattle by Gronlund and collaborators [35]. Starting from these results, Winter et al. further evaluated SAA as a diagnostic marker of subclinical mastitis in ewes, demonstrating that SAA levels in milk, but not in serum, can detect subclinical mastitis in individual ewes [60]. A study by Miglio et al. determined serum levels of SAA in lactating Lacaune sheep [61] of different ages, with the aim of providing a reference interval for SAA in healthy animals.4. SAA in Equids ReproductionThe use of APPs in equine practice has recently raised attention by different research groups with controversial opinions within the scientific community [62,63]. SAA has been one of the most studied APP in several disorders, including those related to reproductive sphere [64,65].A study [47] monitored the changes in serum SAA during infectious endometritis in mares, correlating these APPs to the local innate immune response. Experimentally induced bacterial endometritis with E. coli strains determined a significant increase in SAA, because of an up-regulated endometrial gene expression of SAA, as well as several pro- and anti-inflammatory cytokines [47].Conversely, other authors found that measurement of SAA was not suggestive of subclinical endometritis in mares, keeping the debate alive [37,38,66,67]. Tuppits et al. [37] investigated changes in serum SAA every 48 h during estrous and 5, 6, or 7 days after artificial insemination (AI) with frozen-thawed semen. Standardbred mares, each with a different reproductive status, were divided in groups accordingly (healthy in first postpartum estrus, healthy barren mares, and mares with endometritis). Post-AI endometritis was determined by bacteriological, cytological, and ultrasonographical examinations. Authors found no statistical difference in SAA levels during the study and after AI, concluding that frozen-thawed insemination did not affect serum APPs. Similar results were obtained by Sikora et al. [38], who found no significant increase in serum SAA of Icelandic mares with subclinical endometritis, thus questioning the usefulness of serum APPs for the diagnosis of uterine inflammation. Conversely, Wojtysiak et al. [39] evaluated the local uterine production of SAA before and after AI, observing a significant increase in SAA concentrations in uterine flushing after AI.APPs fluctuations have also been investigated in the periparturient period in healthy mares and in mares with placentitis, with the aim to possibly provide a reliable diagnostic tool to identify early onset of placental diseases that represent prevalent causes of abortion, premature delivery, and neonatal death in equine species.Coutinho da Silva [40] found that serum SAA concentrations significantly increased and remained elevated until abortion due to ascending placentitis experimentally induced by Streptococcus zooepidemicus. This study’s results suggest the use of SAA as possible diagnostic aid in spontaneously occurring placentitis during late-term gestation, which represents a hard-to-diagnose condition in clinical practice. In a further study, Coutinho da Silva et al. [41] confirmed these preliminary results. Fifteen healthy pregnant mares were evaluated weekly from 280 days of gestation to foaling and then at 12, 36 and 60 h post-partum. SAA levels remained within normal intervals during pregnancy, apart from three mares that showed higher SAA values the week before foaling and increased significantly at 12 and 36 h following parturition. In the second part of the same study, authors induced placentitis in 14 pregnant mares by intra-cervix inoculation of Streptococcus zooepidemicus between 280 and 295 days of gestation, dividing animals into treated (n = 9) and control (n = 5) groups. SAA concentrations were determined prior to inoculation and then weekly until abortion or parturition. In untreated mares, SAA increased within 96 ± 56 h from bacteria inoculation and remained elevated until abortion, whereas in the treated group, medical treatment maintained SAA levels within the limits in six out of nine mares [41]. Authors concluded that SAA may serve as both a prognostic indicator in cases of ascending placentitis in the mare and a marker to monitor the treatment efficacy. Similarly, Canisso et al. found higher levels of SAA in pregnant mares after the induction of placentitis and until abortion [68].Changes in APPs have also been evaluated in healthy mares during peripartum by Krakowski et al. [69] who found a certain degree of fluctuations in serum concentrations of SAA. However, the concentration remained within physiological ranges; thus, authors did not consider APPs suitable for indicating potential susceptibility to peripartum disorders in mares.Although horses are not the only equids among domestic animals, few studies investigated SAA in donkeys and mules [70,71] and only one has assessed SAA levels in healthy jennies and donkey foals in the post-partum period [72]. This study highlighted the effect of parturition on SAA levels, with both jennies and foals giving higher SAA values within 48 h from birth [72].5. DiscussionAPPs represent promising diagnostic aids for the early identification of disorders in different fields of veterinary medicine, allowing a continuous monitoring of disease progression and treatment response [4]. However, APPs should not be used as solely indicators of disease and they should be evaluated in adjunction to clinical workup [73]. In animal reproduction, SAA has been proven to be very useful in the detection of challenging diseases, such as subclinical endometritis, with the possibility to provide information regarding the development of the disease [26,74].The clinical application of SAA evaluation as a routine test has some limitations due to practical reasons. On the one hand, despite the availability of portable devices, the analysis of SAA is time-consuming and still relatively expensive, limiting the wide-scale use of APP evaluation in routine clinical practice [74,75]. Furthermore, the lack of recognized reference ranges for domestic species still limits the use of APPs in clinical practice [74,75]. Despite this, the development and optimization of rapid and economic devices for SAA measurement should be encouraged, considering the broad spectrum of possible applications of acute-phase protein-based diagnostics [75] in veterinary reproduction.SAA was found to provide diagnostic and prognostic support in the monitoring of the progression of sepsis and post-operative inflammation in subclinical endometritis, endometrial hyperplasia, and pyometra in dogs and cats [12].Endometritis is one of the first conditions that causes subfertility and infertility in cows, and the possibility to use a reliable indicator could facilitate clinician in both detection and monitoring of such conditions [75,76,77,78].In small ruminants, SAA can play a crucial role in the prevention of economic losses related to reproductive disorders, especially in rural household environments [79].In equine medicine, the understanding of serum and endometrial expression of APPs and other cytokines implicated in uterine defense mechanisms could lead to new therapeutic strategies for endometritis, and maybe identifying further diagnostic mediators/markers for the infertility of equids [13,14,80,81,82].6. ConclusionsDue to its rapid increase after the onset of APR, SAA is considered a sensitive and early indicator of inflammation in domestic species. These characteristics make the assessment of SAA an effective diagnostic aid in the field of animal reproduction, since the early detection and prompt treatment of reproductive diseases are essential to preserve fertility. Notwithstanding the usefulness of SAA assessment in animal diseases, further efforts are needed to include this APP in routine clinical pathology screening.

animals : an open access journal from mdpi

10.3390/ani12010079

PMC8749613

Metabolomics is a powerful approach that is based on the identification in biological samples of metabolites, which production and levels may vary due to factors intrinsic to the environment and the organism. For a correct data interpretation, it is, therefore, necessary to first evaluate the metabolome of the tissue/organ under investigation when it is exposed to no stressor. In this study, the complete set of metabolites of liver and gills of wild golden grey mullet (Chelon auratus) that were collected from a reference area was compared by using metabolomics, which was able to reveal metabolites that are commonly present in both organs but with different levels to be attributed to organ-specific functions. The same metabolomic approach was applied also to study the metabolite changes that were induced in mullet gills and liver after environmental exposure to mercury (Hg), and a variety of organ-specific metabolic disturbances were observed. The findings from this study validate the use of metabolomics in ecotoxicological studies to assess organ-specific functions and the cytotoxicity mechanisms of Hg in fish.

Metabolomics is a powerful approach in evaluating the health status of organisms in ecotoxicological studies. However, metabolomics data reflect metabolic variations that are attributable to factors intrinsic to the environment and organism, and it is thus crucial to accurately evaluate the metabolome of the tissue/organ examined when it is exposed to no stressor. The metabolomes of the liver and gills of wild golden grey mullet (Chelon auratus) from a reference area were analyzed and compared by proton nuclear magnetic resonance (1H NMR)-based metabolomics. Both organs were characterized by amino acids, carbohydrates, osmolytes, nucleosides and their derivatives, and miscellaneous metabolites. However, similarities and differences were revealed in their metabolite profile and related to organ-specific functions. Taurine was predominant in both organs due to its involvement in osmoregulation in gills, and detoxification and antioxidant protective processes in liver. Environmental exposure to mercury (Hg) triggered multiple and often differential metabolic alterations in fish organs. Disturbances in ion-osmoregulatory processes were highlighted in the gills, whereas differential impairments between fish organs were pointed out in energy-producing metabolic pathways, protein catabolism, membrane stabilization processes, and antioxidant defense system, reflecting the induction of organ-specific adaptive and defensive strategies. Overall, a strict correlation between metabolites and organ-specific functions of fish gills and liver were discerned in this study, as well as organ-specific cytotoxicity mechanisms of Hg in fish.

1. IntroductionMetabolomics is a powerful and sensitive tool to evaluate the health status of aquatic organisms in environmental ecotoxicological studies [1]. It provides complete and accurate information on the biochemical responses to contaminants exposure, complementing evidence that are provided by the use of conventional biomarkers [2,3]. Besides that, metabolomics has been contributing with valuable novel insights on the specific mechanisms of aquatic contaminants toxicity in fish. A paradigmatic example of such usefulness is demonstrated by the constant increase in the number of metabolomic studies dealing with ecotoxicology that have been produced in the last two decades, with about 900 papers on this topic published annually [4,5].Although mercury (Hg) is widely known for its toxicity in humans and wildlife, the number of studies using metabolomics to address its effects in fish is still scarce. In our previously published papers, we elucidated the mechanisms of environmental Hg-toxicity in wild golden grey mullet (previously named as Liza aurata, now Chelon aurata) by using an innovative triad approach that was based on Hg bioaccumulation, shifts in metabolite profiles, and conventional oxidative stress biomarkers [2,3,6]. This strategy was successfully applied to mullets inhabiting an Hg-contaminated system in the Aveiro lagoon, Portugal. As a consequence of the high accumulated levels of inorganic Hg (iHg) and methylmercury (MeHg), severe changes were documented in mullet gills in the metabolites that were related to the antioxidant protection, with depletion of reduced glutathione (GSH) and its constituent amino acids (i.e., glutamate and glycine) [3]. This was further corroborated through oxidative stress endpoints, by depletion of glutathione peroxidase (GPx) and superoxide dismutase activities, indicating massive GSH oxidation under Hg stress and an inability to carry out its regeneration or de novo synthesis [3]. Alterations in the structure of cell membranes were also suggested, together with alternative mechanisms for preventing lipid peroxidative damage [3,6]. A mechanistically-based assessment of Hg toxicity was conducted also in liver of the same wild mullets, providing new insights into the toxicological pathways underlying the oxidative stress at hepatic level [2]. Metabolomics revealed several levels of impact due to Hg exposure that triggered adaptive responses in the antioxidant system as observed by increased glutathione-S-transferase and catalase activities, and total glutathione content, which compensated for a decreased GPx activity. Despite these effects, the induction of lipid peroxidation was efficiently prevented in mullet liver [2]. Noteworthy, some of the adaptive and defensive strategies that were triggered by Hg contamination were similar in mullet gills and liver, while others reflected organ-specific responses. It is, therefore, of high relevance in ecotoxicological studies to deeply investigate and compare the metabolite profile of mullet gills and liver, and then observe any organ-specific deviations that are triggered by Hg exposure. Both these issues were addressed in the present work, conducted on the same individuals that were used in our previous studies.Metabolomics was also applied to unveiling the toxicity of Hg by Bridges et al. [7] that sequenced the gut microbiome of fathead minnows (Pimephales promelas) that were exposed to dietary MeHg to investigate its relation to neurotoxicity. Data suggested that environmentally relevant exposure scenarios might cause MeHg-mediated dysbiosis of the gut microbiome, contributing to neurotoxicity in fish. Recently, the effects of MeHg and iHg were investigated in zebrafish embryos [8]. The metabolism of galactose, starch, and sucrose was disturbed after exposure to both Hg forms, and the levels of the neurotransmitters tyrosine, dopamine, and tryptophan were reduced. In addition, oxidative stress was related to metabolite changes, such as alterations in the putrescine, niacinamide, and uric acid contents in zebrafish that were exposed to iHg, and squalene in the MeHg-exposed organisms [8].Fish gills and liver are organs that are widely used in studies addressing environmental biomonitoring of pollutants, including Hg [2,3,9,10]. While the gills are the main organ of waterborne Hg uptake, the liver accumulates high levels of Hg upon dietary exposures due to its main role in metabolism. Indeed, a toxicokinetic study of waterborne iHg in white seabream (Diplodus sargus) pointed out the gills as the most responsive organ by accumulating the metal faster than the other surveyed tissues/organs (eye wall, lens, blood, liver, brain, and bile), while accumulating the highest levels over the exposure time [11]. Moreover, a recent comparative study of dietary iHg and MeHg toxicity in the Korean rockfish (Sebastes schlegeli) comprised of the evaluation of Hg accumulation in several tissues (liver, kidney, dorsal muscle, and brain), reporting the highest levels in the liver upon exposure to MeHg [12]. This pattern was explained by the primary role of liver in detoxification that is related to metallothioneins or metallothionein-like proteins that bind specific metals (e.g., Hg, cadmium, copper) and sequester Hg ions [13]. Keeping in mind that, under realist exposure scenarios, fish are mainly exposed to waterborne iHg and dietary MeHg [14], the gills and liver should be combinedly considered in fish health assessment to better understand the extent of the effects that are triggered by these two Hg forms.Data that were obtained from metabolomic investigative studies reflect metabolic variations that are attributable to factors intrinsic to organisms (e.g., age and developmental stage, gender, size, feeding condition, genetic, specificities of tissues targeted, etc.) and to the environment (e.g., temperature, salinity, dissolved oxygen, etc.). Therefore, to avoid the interference of these confounding factors, it is crucial to evaluate first the metabolome of the tissue/organ under investigation when it is exposed to no stressor. Cappello, et al. [15] published the metabolome of farmed mussel (Mytilus galloprovincialis) through the investigation of three different organs, namely the digestive gland, gills, and posterior adductor muscle, reporting a total of 44 metabolites that were grouped in amino acids, carbohydrates, tricarboxylic acid (TCA) cycle intermediates, osmolytes, neurotransmitters, nucleotides, alkaloids, and miscellaneous metabolites. Multivariate statistics revealed that mussel organs clustered separately from each other, suggesting a clear differentiation in their metabolic profiles. Therefore, the study of Cappello, et al. [15] provided a better understanding of mussel organ-specific functions, while supporting future metabolomic investigations of marine mussel health and safety. A similar approach would be valuable in fish to better address the toxicity of aquatic contaminants [16].In the current study, the metabolomes of the gills and liver of the wild golden grey mullet (Chelon auratus) were analyzed by a proton nuclear magnetic resonance (1H NMR)-based metabolomic approach to achieve: (i) a comparison of both organs of fish that were captured at a putative reference area that was considered unpolluted for Hg to highlight the similarities and differences that are related to organ-specific functions; (ii) a comparison of both organs of mullets that were captured at an Hg-polluted area in respect to those from the reference area to detect any metabolite deviations that were triggered by environmental exposure to Hg.2. Materials and Methods2.1. Study AreaThe Aveiro lagoon (47 km2 of maximum surface area) is a coastal ecosystem on the northwest coast of Portugal (Figure 1). In its upper part, there is an inner and enclosed basin named Laranjo (LAR, with around 2 km2) that had received Hg effluents from a chloro-alkali plant for about five decades. Although the plant activities had ceased in 1994, high levels of Hg are still stored in sediments [11] and could be found in the biota [11,17,18]. Moreover, Hg biological effects have been reported in fish [2,3,10,11,17]. Laranjo basin is a “field laboratory” for Hg toxicity investigative studies since other relevant contaminant sources besides Hg are considered negligible [19], and, therefore, it offers a unique opportunity to evaluate Hg toxicity under realistic conditions [10,11,17].São Jacinto (SJ) is located near the lagoon entrance, distancing about 10 km from the Laranjo basin. In studies evaluating the effects of Hg in fish, SJ has been selected as a reference for comparison purposes with LAR, since it was considered unpolluted for Hg [17,18]. In fact, as already reported in our previously published papers [2,3,6], LAR presented higher levels of total dissolved Hg (1.5 ± 0.77 ng/L) and MeHg (0.040 ± 0.008 ng/L) in water than SJ (1.0 ± 0.02 ng/L and 0.016 ± 0.007 ng/L, respectively). The same spatial pattern was also recorded in the surface sediments for the total Hg (0.44 ± 0.25 µg/g at LAR and 0.025 ± 0.005 µg/g at SJ) and MeHg (0.008 ± 0.003 µg/g at LAR and 0.0001 ± 0.00002 µg/g at SJ). About the water parameters, similar temperatures of around 18 °C were recorded at both sites, while salinity was lower at LAR (average of 21) than SJ (33), thus reflecting the proximity of the sea (Figure 1). As previously reported, the levels for dissolved oxygen indicated undersaturation at LAR (average of 65%) whereas at SJ it was around 100% saturation [2,3,6].2.2. Fish SamplingSpecimens of the golden grey mullet (Chelon auratus) were captured in summer (June 2013) at two sites of the Aveiro lagoon (Figure 1), namely SJ and LAR. A total of eight wild fish were collected at each sampling site. Sampling was carried out during low-tide using a traditional beach-seine net named “chincha”. Juvenile specimens, being not sexually mature, were selected to minimize the interference of gender on metabolome data, and thus in results interpretation. Fish total length at LAR and SJ was 13.6 ± 2.1 and 16.5 ± 2.1 cm, respectively. Immediately after catching, the fish were anesthetized, sacrificed, and properly bled in compliance with the ethical guidelines of the European Union Council (Directive 2010/63/EU), and then the gills and liver were removed. The tissue/organ samples were immediately flash-frozen in liquid nitrogen in the field, while at the lab the samples were stored at −80 °C until further processing for metabolite determinations. As previously reported, the levels of tHg that were accumulated in the gills of mullets from SJ and LAR were 0.08 and 0.22 µg/g dry weight (d.w.), respectively, whereas the concentrations of tHg in the fish liver from SJ and LAR were 0.9 and 2.3 µg/g d.w., respectively [2,3]. Moreover, the level of bioaccumulation of MeHg in the fish gills from SJ and LAR was 0.02 and 0.1 µg/g d.w., respectively, whereas in liver from SJ and LAR it was 0.6 and 1.7 µg/g d.w., respectively [2,3].2.3. ChemicalsDeuterated water (D2O; Armar AG, Dottingen, Switzerland) and 2,2-dimethyl-2-silapentane-5-sulfonate (DSS; Sigma-Aldrich, Milan, Italy), as well as other chemicals (Sigma-Aldrich, Italy) were acquired to conduct the metabolomics analysis.2.4. 1H NMR-Based Metabolomics2.4.1. Gills and Liver Metabolite ExtractionTo extract polar metabolites, a “two-step” methanol/chloroform/water procedure [20,21,22] was applied to the frozen fish gill and liver tissue of 150 and 100 mg weight, respectively. The tissue samples were homogenized in cold methanol (4 mL/g) and cold water (0.85 mL/g) by a TissueLyser LT bead mill (Qiagen, Hilden, Germany) for 10 min at 50 vibrations/s. Then, chloroform (4 mL/g) and water (2 mL/g) were added, and the samples were vortexed for 1 min. The samples were kept on ice for 10 min, followed by centrifugation for 5 min at 2000× g at 4 °C to separate suspension into three phases: the water phase at the top, the denatured proteins in the middle, and the lipid phase at the bottom. The upper methanol layer with polar metabolites (600 µL) of each sample was transferred into clean microtubes and evaporated to dryness by a centrifugal vacuum concentrator (Eppendorf 5301) to be then stored at −80 °C. Before NMR analysis, the dried polar extracts were redissolved with 600 μL of sodium phosphate buffer (0.1 M, pH 7.0, 10% D2O) containing 1 mM of DSS, vortexed, and then transferred into a 5 mm diameter NMR tube for NMR spectroscopy. DSS was added as the internal standard providing a chemical shift reference (δ = 0.0 ppm) for the NMR spectra, whereas D2O provided the deuterium lock signal for the NMR spectrometer.2.4.2. 1H NMR Metabolomics and Spectral Pre-ProcessingThe 1H NMR spectra of all the fish samples were acquired on a Varian-500 NMR spectrometer that was operating at 499.74 MHz at 298 K. One-dimensional (1-D) 1H NMR spectra were obtained using a PRESAT pulse sequence to suppress the residual water resonance and 6983 Hz spectral width with a 2.0 s relaxation delay. A total of 128 transients were collected into 16,384 data points in a 10 min acquisition time. All data sets were zero filled to 32,768 data points and exponential line-broadenings of 0.5 Hz were applied prior to Fourier transformation. To reduce the complexity of the NMR data and facilitate the pattern recognition, 1H NMR spectra of all the fish samples were manually adjusted for phase, baseline, and calibration (DSS at 0.0 ppm) by Chenomx Processor, a module of Chenomx NMR Suite (version 5.1; Chenomx Inc., Edmonton, AB, Canada) software. The peaks within the 1H NMR spectra were identified with reference to known chemical shifts and peak multiplicities using public databases such as the Human Metabolome DataBase (HMDB) [23] and the Chenomx 500-MHz library database. Metabolite quantification was performed by Chenomx NMR Suite software that uses the concentration of the internal standard DSS to determine the concentrations of individual metabolites.All 1-D 1H NMR spectra were converted to a data matrix using Chenomx Profiler, which segmented each spectrum into 0.005 ppm bins between 0.5 and 9.5 ppm for the gill and liver NMR spectra, with regions from 0.60 to 0.64 and 2.87 to 2.93 ppm (DSS), and 4.66 to 5.19 ppm (water) removed from all spectra to prevent interference in subsequent multivariable analyses. To conduct comparisons between the spectra, the integrated spectral area of the remaining bins was normalized before pattern recognition analysis to the total integrated area of the spectra to eliminate the dilution or bulk mass differences among the samples due to the different weight of tissue. 2.5. Statistical AnalysisThe data were normalized to tissue weight and then expressed in mM as a mean ± standard deviation (S.D.). With the aim to clearly recognize the similarities and differences in the metabolome of the two examined fish organs, a Venn diagram was applied using the web application GeneVenn [24]. Moreover, to highlight any potential deviations from the metabolome profiles of the gills and liver of the fish that were triggered by environmental exposure to Hg, a principal component analysis (PCA) was conducted using MATLAB, version R2016a (The MathWorks Inc., Natick, MT, USA) to reduce the dimensionality of the metabolomic data and distinguish the two organs and the two groups of fish (from SJ and LAR). In fact, the PCA allows differences and similarities among the NMR metabolic fingerprints to be visualized in a score plot since samples that are metabolically similar are clustered together. The organ-specific metabolite changes were then calculated via the ratio between the means of LAR and SJ individual metabolites as measured within each organ. The metabolic dataset was tested for normality using the Shapiro-Wilk distribution test, and after confirmation of normal distribution, data homogeneity was evaluated through the Levene’s test. Univariate statistical analysis, namely the parametric Student’s t-test, was therefore carried out on metabolite data using the GraphPad software (Prism 5.0, San Diego, CA, USA). The threshold for significance was p < 0.05, recognized as the criterion of statistical significance.3. Results3.1. Metabolite Profiling of Fish Gills and Liver from the Reference AreaThe representative 1-D 1H NMR spectra of tissue extracts of the gills and liver from golden grey mullet (C. auratus) that were sampled from the reference area (SJ) are depicted in Figure 2. Numerous metabolites were identified within the metabolome of both organs, which were both characterized by a dominant presence of the organic osmolyte taurine (3.25 and 3.41 ppm), found at a concentration of approximately 45 times higher than other metabolites in both the gills and liver. Moreover, all spectra of fish gills were also dominated by lactate (1.33 and 4.12 ppm), which was about 20 times higher than other metabolites, whereas within the 1H NMR spectra of the fish liver the second dominant metabolite was glycerophosphocholine that was found to be approximately 15 times more concentrated than other metabolites. Other major classes of compounds that were found in the metabolome of both fish organs included amino acids (e.g., glutamate, alanine), energy storage compounds (e.g., glucose), glycolytic products (e.g., lactate), Kreb’s cycle intermediates (e.g., malonate, fumarate in gills; succinate in liver), and nucleotides and their derivatives (e.g., uracil).A list of the metabolites that were identified, with their relative chemical shifts and concentrations, is reported in Table 1.3.2. Metabolome Comparison between Fish Gills and Liver from the Reference AreaTo evaluate the similarities and differences in the metabolome content of the two organs under investigation, all the metabolites that were identified in the gills and livers of the fish from the reference area were compared by performing a Venn diagram, as shown in Figure 3. From the comparison of fish gills and liver metabolomes, a total of 40 polar metabolites that are involved in a variety of metabolic pathways were found in the two examined organs. In detail, 24 out of these 40 metabolites were commonly detected in both the fish gills and liver (Figure 3 and Table 2), even if the concentrations of most of these metabolites differed between the two tissues. Moreover, the Venn diagram revealed that ten metabolites were found solely in the fish gills (i.e., acetone, arginine, aspartate, betaine, choline, isobutyrate, lysine, N6-acetyllysine, serine, and UDP-glucose), while six metabolites were detected exclusively in the fish liver (i.e., acetate, glycogen, succinate, taurocholic acid, unknown resonance #1, and unknown resonance #3), as reported in Figure 3 and Table 2. It is worthy to note that the metabolites that were found exclusively in an organ may have been not detected in the other because of their low concentration with respect to the sensitivity of NMR spectroscopy.3.3. Metabolome Changes Induced by Hg in Fish Gills and LiverThe PCA scores plot of the 1H NMR metabolic profiles confirmed the metabolome differences between the fish gills (circles) and the liver (squares), as evidenced along the PC-1 axis and explained by 86.45% of variance (Figure 4). Moreover, a clear clustering of the fish groups from the two sampling sites was also noticed for both tissues along the PC-2 axis that, with the 7.32% of variance, separated mullets that were sampled from SJ (orange circles and light blue squares) from those that were collected at LAR (red circles and blue squares) (Figure 4). The metabolites that were responsible for the differences that were observed between the two groups of fish for each tissue under examination were identified as the result of statistical analysis. As reported in Table 3, the metabolic gill fingerprints of the wild fish that were sampled from LAR compared to those from SJ were characterized by significantly higher levels of alanine, creatine, lactate, uracil, and choline, together with significantly reduced levels of taurine, glycerophosphocholine, and glutathione. Conversely, the metabolic profiles of the livers of mullets that were collected at LAR in respect to those from SJ evidenced significantly elevated concentrations of alanine, phosphocholine, glucose, and glutathione, besides significantly decreased levels of tyrosine, phenylalanine, taurine, and hypoxanthine, as reported in Table 3.4. Discussion4.1. Metabolome of Fish Gills and Liver from the Reference AreaMetabolomics is a powerful approach in elucidating the interactions between organisms and the environment [4,5]. However, the necessity to correctly evaluate the metabolome of the biological system under investigation to accurately address the toxicity of aquatic contaminants avoiding interferences of confounding factors is undeniable. Therefore, a 1H NMR-based metabolomic approach was herein conducted on the gills and liver, the target organs that are strongly connected to each other in terms of metabolism, of the wild golden grey mullet (Chelon auratus) that were collected from a reference area explore and compare their metabolome to highlight any qualitative and/or quantitative differences or similarities in their metabolites, to be attributed to their major functions. Specifically, the gills of fish are mainly involved into the respiratory gas exchange activity, active ion transport [25], osmoregulation [26], acid-base balance, and excretion of nitrogenous wastes [27], besides being the main route for uptake of waterborne environmental pollutants. Contrarily, the liver of fish is a target organ for a number of metabolic activities and detoxification processes, besides playing a main role in accumulation, biotransformation, and cycling of environmental pollutants, including Hg [28,29,30,31].From a careful comparative evaluation of the 1H NMR spectra of the gills and liver of mullets from the reference area, a total of 40 metabolites were identified and, according to their metabolic roles, clustered into six groups including the metabolites that are involved in the metabolism of amino acids, comprising of 16 compounds, in the energy metabolism (8 compounds), in the osmoregulatory processes (3 compounds), in nucleotide and their derivative metabolism (4 compounds), in bile acid metabolism (1 compound), and miscellaneous metabolic pathways (5 compounds). Additionally, three unknown resonances were found in the fish liver, with only the unknown resonance #2 commonly present also in the gills but at a concentration that was notably higher than that which was recorded in the liver. This further supports the hypothesis of an organ-specific variation, both qualitative and quantitative, in metabolites between the two fish organs that were examined that is strictly correlated to physiological specificities. Interestingly, 60% of the identified metabolites were commonly present in fish gills and liver, although exhibiting different concentrations.It is worthy to note that the metabolome of both organs was found to be dominated by the organic osmolyte taurine, detected at a concentration of about 45 times higher than that of other metabolites. Taurine is a sulfur-containing amino acid that serves primarily in the maintenance of cellular homeostasis. For instance, in case of hypo-osmotic stress, the induced cell swelling is followed by a regulatory volume decrease due to cellular efflux of ions and organic osmolytes, including taurine [32]. It is, therefore, justified that the high level of taurine that was recorded in the metabolome of fish gills from the reference area that, being in direct contact with the surrounding water, constantly deal with fluctuations of water salinity. A high sensitiveness to environmental osmolarity is peculiar of euryhaline fish, such as golden grey mullets and European sea bass (Dicentrarchus labrax), as also demonstrated by in vitro investigations [33]. However, for the liver that is not directly challenged by the external medium, the dominant presence of taurine may be explained by its well-documented metal-chelating and antioxidant properties, and, therefore, its active involvement in detoxifying processes and hepatoprotective effects [34].Among organic osmolytes, besides betaine that was detected exclusively in the fish gills as further evidence of the osmoregulatory processes that actively occur within fish branchial epithelium, it is important to highlight the common presence in the two fish organs of glycerophosphocholine, found at a notably higher level in the liver than the gills metabolome. Indeed, besides its role into the mechanisms of osmoregulation, it must be noted that glycerophosphocholine, together with phosphocholine, are the main storage forms for choline within the cytosol [35] and, as a matter of fact, no choline was found in the liver of golden grey mullets.As stated above, the largest group of metabolites that was identified was that of the amino acids, even if it must be pointed out that, contrarily to gill metabolome, in the liver not all amino acids were detected. This discrepancy is likely to be associated to the different role of the examined organs. In detail, it is evident that there is a higher level of alanine in the liver compared to the gill metabolome, and this may be attributed to its role as main carrier of amino nitrogen to liver. However, the presence of alanine in the gills, although at a lesser extent than in the liver, may be associated with the nitrogenous waste excretion that is occurring at the branchial epithelium [27]. In regard to the branched-chain amino acids (BCAAs), namely leucine, isoleucine, and valine, they were all found at comparatively similar concentrations in both organs, although at a slightly higher extent in the liver than the gills. BCAAs are known essential amino acids for the immune system because, beyond their role in the regulation of protein turnover processes, they serve as proteinogenic amino acids, precursors for the biosynthesis of new molecules and cells, such as lymphocytes [36], but also as energy sources. Among the differences in amino acids content between the fish gills and liver, isobutyrate, lysine, arginine, aspartate, N6-acetyllysine, and serine were observed only in the gills. Among these, isobutyrate is recognized as a marker of anoxia, of which the concentration increases in the gills, is involved in the respiratory gas exchange activity when the organism is facing a hypoxic condition due to alteration in the oxygen intake [37,38]. The absence of isobutyrate in the metabolome of the fish liver is, therefore, not a surprise, due to the different physiological role in respect to gills.Metabolites that are involved in the energy metabolism were also found in both the fish gills and liver metabolome, though with evident differences between the organs in terms of the metabolites that were detected or in their concentrations. In detail, the fish liver was characterized by higher levels of glucose and glycogen, the latter of which was not recorded in the gills, and higher concentrations of Kreb’s cycle intermediates together with a lower level of lactate. This metabolite profiling of the liver finds explanation considering that, among the others, the liver plays a unique role in controlling the carbohydrate metabolism and lipid homeostasis [39].Another class of compounds that was found in the metabolite profiling of fish gills and liver was that of nucleotides and their derivatives, including uracil, uridine, hypoxanthine, and inosine, where the latter was found to be the only molecule where the concentration was higher in the gills than the liver. Overall, the different levels of these compounds as found between fish organs may reflect ongoing transcriptional activities. Thus, it is reasonable to hypothesize that the concentrations of nucleotides and their derivatives are highly variable, context-specific, and strictly associated with tissue-specific metabolism [40].Furthermore, a metabolite that was recorded in fish liver and not detected in the gills was the taurocholic acid, a bile acid representing the product of conjugation of cholic acid with taurine. Bile acids, known as essential organic molecules that are synthesized in the liver from cholesterol, are also recognized as valid indicators of hepatobiliary impairment because their synthesis and metabolism are strictly influenced by liver diseases [41].4.2. Tissue-Specific Metabolic Changes Induced by Hg in Fish Gills and LiverData on Hg accumulation in golden grey mullets that were collected at LAR and compared to SJ, as reported above and published previously [2,3], revealed a bioaccumulation of tHg and MeHg level of 10- and 17-fold higher, respectively, in fish liver compared to the gills, thus making it feasible to hypothesize the induction of metabolic changes in both fish organs. Therefore, a further purpose of this study was to elucidate similar or differential deviations from the metabolome profiles of the gills and liver of mullets that were triggered by environmental exposures to Hg. However, besides Hg, the influence on some key metabolites of the differential salinity and dissolved oxygen as recorded at the two sampling sites was also considered in this field study.A clear grouping was observed between the two fish organs that were examined, as well as between the two groups of fish that were collected at SJ and LAR, thus suggesting a clear differentiation in their metabolic profile. From the comparison of the major metabolites that were probably affected by the exposure to Hg, it was possible to notice a different extent of perturbation or an opposite or specific pattern in the changes of metabolites between the fish gills and liver. This reflects differential toxicological effects that were triggered by Hg exposure in each organ, as well as their ability to respond to the same stressor by activating organ-specific adaptive and defensive strategies.Of note, a significant reduction in taurine level that was induced by Hg exposure was observed in both the fish gills and liver, even if likely to be due to different Hg modes of action. In fact, in the gills it may be attributed to disturbances in osmoregulatory processes since taurine is known to be involved in the maintenance of cellular homeostasis [32,33]. Therefore, the decreased taurine level that was recorded in the gills from LAR with respect to those from SJ may reflect interferences in osmotic balance due to differences in salinity between the two sampling sites as well as to Hg bioaccumulated levels. Conversely, in the liver, that is an internal organ and thus not markedly challenged by the external medium, the influence of salinity in taurine level could be disregarded. Its depletion may be explained with the taurine metal-chelating property, and, therefore, its active involvement in detoxifying and hepatoprotective effects [34]. Moreover, taurine can act as a potent antioxidant and its reduced levels that were observed in both fish organs may thus be indicative of a protective mechanism against Hg-induced oxidative stress [42], which was demonstrated to occur in both the mullet gills and liver in our previous studies through oxidative stress endpoints [2,3,6]. It is also plausible to hypothesize that if taurine is affording antioxidant protection, this could compromise other taurine-dependent functions.A similar trend in the metabolic change that was induced by Hg in fish gills and liver was also found for alanine, which increased in both tissues after Hg exposure. Taking into consideration that alanine is the main nitrogen carrier to the liver, its augmentation in both organs may be explained as an adaptive response to deal with an excess of free ammonia, maybe resulting from protein catabolism to prevent intoxication in gills, which is actively involved also in nitrogenous waste elimination [27].Surprisingly, opposite changes, even if not all significant, were observed in mostly of the Hg-altered metabolites between the fish gills and liver. This is the case of glycerophosphocholine and phosphocholine, which depleted both in the gills, showing an opposite trend than choline, but both increased in the liver. The significant decrease of glycerophosphocholine that was measured in the gills is related to changes in osmoregulatory processes and, therefore, is affected, not only by Hg but also by changes in salinity as found between the two sampling sites. Moreover, it may be linked to ongoing membrane stabilization/repair processes due to the biosynthesis of phosphatidylcholine, the major structural phospholipid of cell membranes [6], as an adaptive cell-protecting antioxidant mechanism that is occurring in mullet gills from LAR to cope with decreased dissolved oxygen and environmental Hg contamination [3]. Contrarily, the opposite trend that was exhibited by the same metabolites in the fish liver suggests the breakdown of phosphatidylcholine and/or the occurrence of membrane turnover. The increase in phosphatidylcholine degradation products suggest the occurrence of oxidative insult at the hepatic level, more likely triggered by Hg exposure as specifically documented in our previous paper [2]. Interestingly, aberrant choline metabolism is a hallmark of oncogenesis and cancer progression that is characterized by increased phosphocholine, glycerophosphocholine, and total-choline-containing compounds that are also strictly related to the Kennedy pathway, which constitutes the biosynthesis pathway for membrane phosphatidylcholine [43].An opposite trend between the two organs was also noticed for glutathione, which significant dropped in the fish gills together with the reduced levels of its constituent amino acids (i.e., glycine and glutamate). This makes it feasible to hypothesize some vulnerability in mullet gills to Hg toxicity, as the reduction in glutathione enhances the risk of oxidative stress due to the accumulation of reactive oxidative species that is triggered by environmental Hg pollution as well as related to the decreased dissolved oxygen [6,44]. Interestingly, an accurate evaluation of the pro-oxidant status of mullet gills from LAR indicated the occurrence of massive GSH oxidation under Hg stress and an inability to carry out its regeneration or de novo synthesis [3]. On the contrary, the significant hepatic elevation in glutathione content suggests an adaptive response to Hg bioaccumulation [31], resulting in the activation of antioxidant defense mechanisms in mullets that are exposed to Hg, further supporting the organ-specific functions with liver being the major site of storage and detoxification of pollutants.Among metabolites exhibiting opposite patterns between the two fish organs, lactate must also be mentioned. Its significant increase in the gills of Hg-exposed mullets may be indicative of an enhanced anaerobic metabolism as an adaptive strategy to replenish insufficient energy supply due to environmental Hg contamination as well as to the lower oxygen availability that was recorded at LAR [6]. On the contrary, the observed hepatic dropping of lactate may be related to its use as a substrate for gluconeogenesis, as further supported by the elevated levels of glucose in the liver together with unaltered glycogen levels [31]. Therefore, this may be considered as a further organ-specific toxicological effect that is induced by Hg in fish.It is worthy to also note the significant rise in the level of creatine in the gills of fish that were exposed to Hg, which indicates changes in ion-osmoregulatory processes, and therefore, may also be influenced by the lower salinity that was recorded at LAR. In regard to the hepatic creatine level, no changes were detected in the fish from the Hg-contaminated area, LAR. However, considering that the liver is usually the site of creatine production, which has to be synthesized continuously [45], it is reasonable to hypothesize that as creatine is biosynthesized it is soon transported to the blood to be translocated to the gills [46] to cope with Hg-induced changes in ion-osmotic balance.5. ConclusionsProton NMR-based metabolomics allowed a successful comparison of the complete metabolome of the gills and liver of golden grey mullets (C. auratus), revealing similarities and differences in their metabolite profile to be related to their organ-specific roles. Among others, taurine was predominant in both organs and is mainly involved in osmoregulation in the gills and detoxification and antioxidant protective processes in the liver. Also, the high levels of glucose and glycogen that were observed in the liver confirmed its role in carbohydrate storage.Mercury environmental contamination triggered multiple and often differential metabolic alterations in the golden grey mullet gills and liver. In detail, severe disturbances in ion-osmoregulatory processes were highlighted in the gills, whereas differential impairments in energy-producing metabolic pathways and protein catabolism were found between the fish organs. Perturbation in membrane stabilization processes and alteration of the antioxidant defense system were also pointed out, thus reflecting differential cytotoxicological effects that were triggered by environmental Hg, as well as their ability to respond to the same stressor by activating organ-specific adaptive and defensive strategies.Overall, a strict correlation between metabolites and the organ-specific physiology of the gills and liver were discerned, which provides precious information for a better understanding of metabolite shifts when investigating the cytotoxicity of Hg in fish.

animals : an open access journal from mdpi

10.3390/ani11113250

PMC8614291

Melanomas are skin tumors of the pigment-producing melanocytes. Equine melanomas are among the most frequently diagnosed tumors affecting grey horses. The melanocytic tumors progress to malignancy in more than two-thirds of cases. Previous laboratory experiments and studies with horses utilizing the naturally occurring betulinic acid (BA) and its derivative NVX-207 showed promising results with respect to the topical (epicutaneous) treatment of equine melanoma. Therefore, the aim of this feasibility study was to gain first insights into the effect and safety of BA and NVX-207 in eighteen horses with early-stage melanocytic tumors after a 13-week-long topical application. The topical treatment was convenient and safe. Compared to a placebo, the data suggest a positive treatment effect from topical application of BA and NVX-207 on equine melanomas toward the end of the treatment period. However, the time period studied was too short to conclusively prove this. Further advancement of the investigational medicinal products studied herein could lead to an effective, topical and marketable novel drug which helps to relieve suffering and, consequently, improve the welfare of equine skin cancer patients.

The naturally occurring betulinic acid (BA) and its derivative NVX-207 induce apoptosis in equine melanoma cells in vitro. After topical application, high concentrations of the substances can be reached in healthy equine skin. With the aim to investigate the effect and safety of topically applied BA and NVX-207 in horses with melanocytic tumors, the longitudinal, prospective, randomized, double-blind, placebo-controlled study protocol included eighteen Lipizzaner mares with early-stage cutaneous melanoma assigned to three groups. Melanocytic lesions were topically treated either with a placebo, 1% BA or 1% NVX-207 twice a day for 91 days. Caliper measurements, clinical examinations and blood tests were performed to assess the effects and safety of the medication. The topical treatment was convenient and safe. The volumes of tumors treated with BA were significantly reduced over time as compared to tumors treated with the placebo from day 80 of the study. Although treatment with NVX-207 seemed to decrease tumor volume, these results did not reach statistical significance. The findings must be regarded as preliminary due to the limited group size and need to be replicated in a larger cohort with modified pharmaceutical test formulations. Accordingly, the treatment protocol cannot yet be recommended in its current form.

1. IntroductionThe susceptibility to melanoma development in grey horses is high due to genetic mutations [1,2]. Early stages of the melanomas located mainly in the dermis frequently occur as single, black-pigmented, firm nodules in glabrous skin under the tail root, around the anus, perineum, external genitalia, in the lips and eyelids, or in rare circumstances at other locations [3,4,5]. Economic and functional problems such as interference with harnessing and breeding impairment have been reported [6,7], but more severe and life-threatening visceral signs can occur with disease progression and metastasis [8,9,10,11,12,13]. The often slow-growing nature of the tumors, the proximity to important anatomical structures such as nerves, vessels or the anal sphincter, and the currently challenging or inefficient therapeutic options have led many practitioners to advocate benign neglect of small melanocytic tumor masses in horses [8,9]. However, every equine melanocytic neoplasm should be considered pre-cancerous or potentially malignant and, therefore, worthy of treatment [8,9,14]. The topical (epicutaneous) treatment of equine melanomas could be a feasible approach to treat early stages of the disease, aiming at preventing later malignant transformation. Topical therapies are characterized by their non-invasive nature and reduced systemic side effects [15,16]. Usually, they are affordable and can be performed with low logistical effort by the horse owners themselves, which reduces the stress factor on the horse significantly.Betulinic acid (BA) is a pentacyclic lupane-type triterpenoid of plant origin [17]. Considerable amounts of the substance can be extracted from the bark of certain tree species, for example, the plane or the white-barked birch tree [17,18]. A wide range of pharmacological properties have been described for BA [19], among which the antitumoral features have been particularly studied [17,20,21]. The main antitumoral effects of the substance are based on the ability to trigger the mitochondrial pathway of apoptosis in cancer cells [22,23] to inhibit the eukaryotic topoisomerase I and II [24,25,26] and suppress the angiogenesis within the tumor [27,28,29,30]. Among a variety of BA derivatives, the compound NVX-207 has been identified as one of the most biologically active and pharmacologically significant agents [31,32,33]. The efficacy and mechanisms of BA and NVX-207 as potential therapeutics against equine melanoma were evaluated by in vitro cell culture experiments [33,34,35]. Reported findings suggest that BA and NVX-207 may achieve anticancer activity in equine melanoma cells due to cytotoxic and antiproliferative effects, whereby cell death is induced by apoptosis [33,34,35]. Concentration profiles of BA and NVX-207, both of which have been determined in vitro and in vivo, further indicated that the compounds’ half-maximal inhibitory concentrations for equine melanoma cells can be achieved in healthy horse skin [34,35,36]. The in vitro and in vivo studies reported provide a promising basis for the use of BA and NVX-207 as topical drugs in clinical trials for equine melanoma treatment [33,34,35,36]. Consequently, the aim of this longitudinal, prospective, randomized, double-blind, placebo-controlled pilot study was to gain first insights into the effect and safety of BA and NVX-207 in horses with early-stage melanocytic tumors after a 13-week-long topical application.2. Materials and Methods2.1. Approval of the Animal ExperimentsThe longitudinal, prospective, randomized, double-blind, placebo-controlled study protocol was approved by the institutional ethics and animal welfare committee of the University of Veterinary Medicine Vienna, Vienna, Austria and the Austrian Federal Ministry of Education, Science and Research in accordance with the Austrian Animal Welfare Law (BMBWF-Reference number: 68.205/0197-V/3b/2019). Informed consent was obtained from the stud management.2.2. HorsesThe study was performed between January and April 2020 at a stud farm in Austria. Eighteen white, flea-bitten or dappled Lipizzaner mares with cutaneous melanomas were included in this study (Table 1). The number of animals was determined by a power analysis using G*Power 3 [37] with the following assumptions: effect size 0.25, type I error 0.05, type II error 0.2 and correlation between measurements 0.75. Considering the planned duration of the trial, the relatively low division rate of the melanoma cells, the pharmacokinetic properties of the drug formulations and the relatively short intervals between measurements, the expected effect size was in the low to medium range and the expected correlation between measurements high. These parameters resulted in a total number of 15 animals. The estimated number of 18 animals included one additional animal per group, as at this low group size the exclusion of one animal would dramatically affect the power of the study. To avoid seasonal changes in the coat and skin (e.g., length and texture of the coat, possible sweating, sun exposure, wetness, exercise, etc.), the additional animals participated in the study from the beginning.The median age of the horses was 14.5 years (range 6 to 28 years) and the median body condition score was 6 (range 4 to 8) according to the scoring scheme of Kienzle and Schramme [38]. None of the horses had ever been treated for melanoma in the past. Nine of the eighteen horses (horses 1, 5, 6, 9, 12, 13, 14, 15, 16) were in foal and the births of the foals were expected during or shortly after the study period. Horses were considered eligible for the study if they had cutaneous melanomas in localizations easy to treat (e.g., undersurface of the tail). Irrespective of the total number of melanomas identified on an individual horse, a maximum of two tumors per horse with a diameter of maximal 15 mm were treated. The tumors to be treated had to be easily distinguishable from each other and from other tumors. Clinical diagnosis was set at the beginning of the study on the basis of the localization and gross appearance of the lesions in conjunction with the horses’ signalment. Fine needle aspirations of tumor masses were performed in seven horses (horses 1, 5, 7, 12, 13, 17, 18) and revealed in 7/7 heavily pigmented cells, where the evaluation of nuclear criteria of malignancy was impossible. However, tumors were identified as pigmented lesions with unknown malignant potential. For the other horses, the procedure would only have been possible under sedation, which was not permitted by the stud management. Medical histories were obtained before the instigation of the topical treatment and a thorough physical examination was performed on each horse to ensure eligibility for the trial. The animals were kept in groups of 15 to 25 horses in stables overnight and on a paddock during the day. After birth, mothers and foals were separated in individual boxes for about seven days before they were kept together in groups with other mares and foals in large stables. All horses were fed a mix of muesli, oats and mineral feed daily, the quantity of which depended on body weight and performance. They had ad libitum access to hay and water.2.3. Topical TreatmentMelanomas were topically treated with pharmaceutical test formulations (creams) which had been previously tested for tolerability on eight healthy horses and in which a homogenous and stable distribution of BA and NVX-207 had been shown [36]. Treatment was performed twice daily and consisted of topical application of either 1% BA in “Basiscreme DAC” (amphiphilic cream as published in the German Drug Codex) with 20% medium-chained triglycerides, 1% NVX-207 in “Basiscreme DAC” or a placebo (“Basiscreme DAC” with 20% medium-chained triglycerides) for 13 consecutive weeks (91 days). Each tumor was completely covered with the cream and protected with an appropriately sized wound dressing (“Animal Soft” Snögg, Vennesla, Norway), which was fixed with “Fixomull stretch” (BSN medical GmbH, Hamburg, Germany) to prevent the cream from being rubbed off (Figure 1). As there are no safety studies on the topical application of BA and NVX-207 in humans, the person applying the cream wore gloves.The study protocol stipulated that if the melanoma(s) disappeared completely before the end of the 13 weeks (tumor no longer palpable), the tumor was treated with the preparation assigned to it for another 14 days after tumor regression. The treatment was discontinued before the end of the 13 weeks if the tumor(s) grew aggressively.The treatment was performed blinded by the first author (LAW). Horses 17 and 18 were treated by the stud’s staff from day 56 to 83 of treatment due to restrictions in the context of the SARS-CoV-2 crisis. The identical-appearing study medication was packed in identically number-coded jars. The patients were randomized into three groups (placebo, BA, NVX-207) of six horses. Randomization was performed by a person not involved in the study, assigning the horse numbers, which were written on paper sheets and blinded to him, to the blinded number-coded jars by random allocation. The numerical codes of the study medication were unblinded after all analyses had been completed.2.4. Clinical Safety Assessment of the TreatmentThe safety and tolerability of the topical treatment were evaluated by general clinical examinations, monitoring of the tumor and its surrounding clinically normal skin, assessment of the horses’ behavior during cream application according to [36], and hematologic and blood biochemistry profiles. The animals were examined clinically prior to each topical application (twice a day) in the first week of treatment. Thereafter, a general clinical examination was performed twice at 14-day intervals (day 21 and 35) and then twice at 30-day intervals (day 63 and 92). The melanoma to be treated and the surrounding tissue were assessed daily for local inflammation, ulceration and depigmentation. If observed, inflammatory skin reactions on treatment sites (redness, heat, swelling, pain) were classified into “mild”, “moderate”, and “severe” based on clinical experience. Blood was collected on days 0, 7, 21, 35, 63 and 92 for a complete blood count and serum chemistry profile, including electrolytes (sodium, potassium, chloride, calcium, magnesium), urea, creatinine, total protein, albumin, lactate, serum amyloid A and enzymatic activity of the alkaline phosphatase, glutamate dehydrogenase, g-glutamyl transferase and creatine kinase.The study protocol specified that treatment was discontinued if the horse showed moderate skin changes in the area treated for more than two days, if the horse showed mild abnormal physical examination parameters for five days or moderate abnormal physical examination parameters for three days, and in the case of significant illness or general deterioration in the condition of the horse.2.5. Tumor Response EvaluationTarget lesions were photographed and the length (mm, longest diameter) and width (mm, perpendicular to length) were measured with calipers (CONNEX GmbH, Oldenburg, Germany) prior to treatment (day 0) and on days 7, 21, 35, 63, 77 and 92. Measurements were performed in duplicate. Tumor volume (mm3) was calculated according to a formula described previously [39,40]: Tumor volume = length × width2 × 0.5. All tumors were measured by the first author (LAW). Follow-up examinations of the horses and tumors were performed four months after the last treatment.2.6. Statistical AnalysisA generalized additive model of the Gamma family with log link function was fitted using R 4.0.2 [41] and the ‘mgcv’ package [42] to estimate trends in tumor growth during the experiment. The model consisted of a smooth function of time interacting with group (treatment), which was also included as a main parametric effect. The individual melanomas were considered random effects nested within horse. The inclusion of a separate random effect for ‘horse’ was discarded based on the higher Akaike information criterion (AIC) of the corresponding model. The thin plate regression splines were parametrized by restricted maximum likelihood [43]. The fitted smooth functions were compared on the link scale for each pair of groups via the use of a prediction matrix, as described in [44]. The given confidence intervals correspond to approximate 95% pointwise confidence intervals. P-values for parametric effects were obtained from Wald tests using the Bayesian covariance matrix for the coefficients. Normality of the residuals was ensured visually using histograms and quantile–quantile plots.3. Results3.1. Tumor ResponseA total of 29 melanoma lesions (groups placebo n = 8; BA n = 12; NVX-207 n = 9) were treated twice a day for 13 weeks. Tumor diameters and volumes are described as medians and quartiles (QR; first quartile, third quartile). The median longest tumor diameter (length) measured on day 0 was 6.0 mm (QR; 5.0 mm, 7.5 mm) in the placebo group. Tumors in the BA group had a median length of 5.0 mm (QR; 5.0 mm, 8.0 mm). Tumor diameters in the NVX-207 group had QR of 5.0 and 6.0 mm with a median length of 5.0 mm. Supplemental data provide information on tumor length and width at the different measurement time points in the placebo group (Figure S1), BA group (Figure S2), and NVX-207 group (Figure S3).The median absolute tumor volumes in the placebo group reduced from 139.8 mm3 (QR; 68.8 mm3, 364.5 mm3) on day 0 to 117.0 mm3 (QR; 62.5 mm3, 191.0 mm3) on day 92. After treatment with BA, median tumor volumes decreased from 75 mm3 (QR; 62.5 mm3, 473.3 mm3) on day 0 to 51.3 mm3 (QR; 36.0 mm3, 322 mm3) on day 92. Median tumor volumes in the NVX-207 group decreased from 108.0 mm3 (QR; 62.5 mm3, 108 mm3) to 62.5 mm3 (QR; 40 mm3, 108 mm3). Figure 2 shows the relative tumor volumes over time for the individual horses.Disregarding time, no overall differences between groups were evident from the generalized additive model (Wald test: df = 2, F = 0.624, p = 0.537). In contrast, the group:time interaction distinguishes the time course of each group and may uncover differences at certain time points that are not sizeable enough to lead to an overall difference, but are still statistically significant and relevant to the interpretation. The group:time interaction indicated a significant difference in tumor volumes between groups BA and placebo from treatment day 80 (Figure 3). Although treatment with NVX-207 seemed to reduce tumor volumes in four horses (Figure 2), these effects did not reach statistical significance within the time:group interaction when compared to the placebo group (Figure 3).No new melanocytic lesions were detected in any of the horses during the treatment period.3.2. Clinical Safety Assessment of the TreatmentAll horses tolerated the topical drug application well and no active defense movements were observed during the treatments. The dressings covering the treatment areas reliably remained at the desired location. Based on the clinical examinations of the horses, the topical melanoma treatment was safe in all groups. Two horses developed a mild spasmodic colic on day 7 (horse 7; NVX-207 group) and on day 13 (horse 17; BA group). Both cases of colic were successfully treated with a single administration of mild spasmoanalgesics (50 mg/kg bodyweight metamizole sodium IV plus 0.2 mg/kg bodyweight butylscopolammonium bromide IV; “Novasul” Richter Pharma AG, Wels, Austria and “Buscopan compositum” Boehringer Ingelheim, Ingelheim, Germany). All pregnant mares gave birth to healthy foals. The blood results revealed no hematologic toxicity or clinically relevant abnormalities at any time point.Depigmentation of the melanomas or the melanoma overlaying skin was observed in four out of eight tumors treated with the placebo (tumors of horses 1, 15, 18). The same was noted in 7 out of 12 tumors treated with BA (tumors of horses 4, 6, 13, 17) and three out of nine tumors treated with NVX-207 (tumors of horses 7 and 14). An ulceration of melanoma II was observed in horse 6 from day 43 to day 70 after treatment with BA (Figure 4). In horse 2, the melanoma-surrounding skin revealed isolated, depigmented areas from day 24 to day 86 of treatment. In horse 18, an isolated to extensive depigmentation and mild redness of the skin around melanoma II was observed from day 16 to day 30 (Figure 5). Apart from this, none of the horses showed clinical signs of redness, heat or swelling in the area of the treated skin.Four months after the end of the last treatment, small, depigmented areas were apparent only in the two tumors of horse 17 (BA group). The skin of all the other horses was pigmented again.4. DiscussionIn the present pilot study, the topical application of 1% BA or 1% NVX-207 twice a day for 13 consecutive weeks in equine melanoma patients proved to be safe and was well tolerated. The topical therapy resulted in part in clinically visible and measurable changes in small melanocytic lesions, which were reflected in skin depigmentation and reduction in tumor diameters and volumes. A significant beneficial treatment effect could be shown after treatment with BA towards the end of the treatment period.Although most melanocytic tumors in horses show a slow growth pattern for many years, more than two-thirds are thought to progress to malignancy [8,45]. For this reason, and because smaller tumors are easier to treat than large ones, even early (pre-cancerous) stages of equine melanoma should be subjected to therapy. Previously reported in vitro cell culture experiments and in vitro and in vivo permeation studies on unaltered horse skin indicated that the naturally occurring BA and its derivative NVX-207 may exert anticancer effects against equine melanoma [33,34,35,36]. The findings of this preceding work prompted further evaluation of the safety and efficacy of the compounds in equine melanoma patients in the current study. Smaller tumors were deliberately treated to explore a potential therapy that can be used for early stages of the disease. In these tumors, a cytological assessment of the cell nuclei for the criterion of malignancy could not be performed with certainty, as a consequence of the strong pigmentation of the cells. However, due to the small size of the melanocytic tumors, it can be assumed that they were pre-cancerous rather than malignant. It has been demonstrated previously that the calculation of tumor volumes with caliper measurement and the formula used here correlates well with tumor volumes calculated using three-dimensional ultrasound measurements [39,40]. Thus, caliper measurements were considered sufficiently reliable for tumor volume assessment.A steady volume reduction was observed in most melanomas treated with BA. Despite the small number of cases, this resulted in a significant positive treatment effect in the time:group interaction compared to the placebo group towards the end of the treatment period. Although these first results after topical BA application on small equine melanocytic lesions are promising, the observations must be confirmed in larger studies with a longer treatment period and with a more diverse horse population in order to be able to draw sound conclusions regarding the effectiveness of the substance in melanoma-affected horses. Modifications in the test formulation, such as increasing the concentration of the active ingredient or incorporating permeation enhancers that transport large amounts of the compound through the fibrous tumor capsule of equine melanomas to the tumor cells, could also have a positive effect on tumor volume regression and may reduce the treatment period.A cautiously positive, albeit not significant trend in tumor volume reduction was shown after 91 days of topical treatment with a cream containing 1% NVX-207. The treatment period investigated was too short, however, to prove this conclusively. Regarding the existing in vitro and in vivo data of NVX-207, it seems surprising that this derivative appears to have fewer anticancer effects on the tumors than its parent BA [31,33,34,35,36]. The substance’s reported in vitro half maximal inhibitory concentrations which lead to antiproliferative and cytotoxic effects in equine melanoma cells are much lower than those determined for BA [33,34,35]. However, it should not be disregarded that tumor cells integrated in their native microenvironment can be much more robust against pharmacological influences than tumor cells cultivated under in vitro two-dimensional cell culture conditions and, therefore, a reliable transferability of in vitro to in vivo results is not always given [46,47]. In addition, permeation barriers, such as the firm tumor capsule often found around equine melanocytic tumors, which could hinder the active substance to diffuse into the tumor cell, are also missing in vitro [5,46,48]. While the half maximal inhibitory concentrations determined for equine melanoma cells were surpassed after topical application of 1% NVX-207 in the epidermis, superficial and deep dermis of healthy horse skin [35,36], a less potent permeation into melanoma-affected skin could, therefore, further explain the only moderate effects of the compound in this study. It is also likely that tumors were located in the deep dermis and the NVX-207 applied topically may not have reached the full depth of the tumor invasion. An analysis of the NVX-207 content in the study medication a few weeks after the study termination revealed that the NVX-207 concentration had decreased only negligibly and a correlation between the reduced active ingredient content and reduced effectiveness can, thus, be excluded (own laboratory controls; data not shown).The topical treatment of early stages of equine melanoma with 1% BA and 1% NVX-207 resulted in part in tumor volume and tumor diameter reductions and may represent—after modification of the pharmaceutical test formulations—an alternative to the frequently practiced approach of benign neglect of small solitary masses. Nevertheless, the results for lesions belonging to the BA and NVX-207 groups should be interpreted against the background that a few tumors in the placebo group also showed a decrease in size. It was stated previously that no reports about spontaneous melanoma regressions in horses exist [14,39]. When the growth behavior of 59 untreated melanomas was investigated in 17 Lipizzaner stallions from the same Lipizzaner stud as the Lipizzaner mares used in the present study, the tumor volume increased by 0.14% per day over an observation period of 162 days, but a slight reduction in tumor volume was sporadically observed in some lesions [49]. A trend in melanoma growth was observed in the placebo group of another study over only 64 days [50]. When the same pharmaceutical formulations as those used in this study were topically applied twice a day for seven days on eight healthy horses, an activation of the immune system by means of a perivascularly accentuated, lymphohistiocytic inflammation with a few neutrophils was observed in the superficial dermis of both the cervical and ventral tail skin [36]. As these alterations were noted in all treatment groups, an association with ingredients in the carrier cream “Basiscreme DAC” but no causative effect of the compounds BA or NVX-207 was suggested [36]. In the present study, the repeated topical application of the formulations for 13 consecutive weeks and the covering of the treatment areas could have led to an increased blood supply to the tumor area with increased immune cell infiltration. The presence of tumor-infiltrating lymphocytes has been associated with a favorable prognosis for human melanoma [51,52]. However, as no histopathological examinations of the melanomas treated with appropriate staining for vascularization markers or immune cell typing were performed in the present study, it remains unknown whether immunological adjuvant effects were involved in the tumor volume reduction. Since the tumor measurements were carried out by only one person, measurement variations can almost be excluded.As there are currently no topical treatment options for equine melanoma that rely on larger clinical evidence-based studies, the treatment regime in the current study could only be presumed. In equine sarcoids, treatment durations between three and 45 weeks are reported for the topical approach [53,54,55,56]. Previously determined in vitro data indicated that in vivo treatment regimens with short application intervals and long treatment durations could favorably influence the concentration and efficacy of BA and NVX-207 in the skin of equine melanoma patients [34,35]. In addition, the application interval of 13 weeks utilized in the recent study is similar to an 11-week topical application of frankincense oil to an Arabian mare with stage 3 equine melanoma, which resulted in a clear tumor volume reduction [57]. The data generated in this pilot study suggest that an even longer treatment interval with 1% BA or 1% NVX-207 may lead to more notable clinical effects. However, this has not been definitively proven and, consequently, the study protocol applied in the current study cannot yet be recommended for melanoma therapy in horses. A treatment twice a day over a period of 91 days was feasible within the frame of the current study. Generally, for reasons of practicability, faster treatment success and owner compliance shortening the treatment period by improving the pharmacologic formulations should be the aim of future studies.The topical melanoma treatment was safe and well tolerated in all groups as assessed by regular clinical examinations and serial blood sampling. The inconspicuous behavior may be related to the fact that the treatment did not cause painful skin inflammations. All horses showed an undisturbed general condition during the entire course of the study, apart from two horses with acute and medically resolved mild colic. Both horses that developed colic had a history of occasionally developing slight spasmodic colic at this time of the year. It seems very unlikely that the occurrence of colic was related to the topical melanoma treatment.Depigmentation was occasionally observed in the skin surrounding and overlaying the tumors. The decreased amounts of melanin in the epidermis might be caused by toxic effects on the melanocytes or disturbed melanization due to the treatment [58]. Observations from follow-up examinations four months after the last treatment revealed that the depigmentation was a temporary side effect. While toxicity data of BA or NVX-207 for normal equine melanocytes are missing, a cytotoxicity of BA for human melanocytes by induction of apoptosis was described to varying degrees [59,60,61]. As the cases of depigmentation were also observed in the placebo group, an association with the ingredients of the amphiphilic carrier vehicle “Basiscreme DAC” is likely. With regard to the evidence of a mild skin irritative potential reported here and previously [36] and with respect to the fact that few tumors in the placebo group reduced also in size, it is recommended to use another pharmaceutical formulation as a placebo and vehicle for BA and NVX-207 in future studies.The topical drugs used in this study are relatively affordable because the basic active ingredient BA is found in plants and is relatively inexpensive to produce. Although it is advantageous that topical medications are commonly low in cost and can be easily applied by horse owners, generally the benefits of topical therapies appear to be limited to only the lesions treated and no systemic antitumor effects can be achieved. Melanomas of the lip seem unsuitable for a topical therapy because of the risk that the animal will lick off the cream or ointment and absorb it orally. By contrast, the treatment of melanomas located on the ventral tail and in the perianal region has been proven to be very feasible in this study. Future trials should evaluate the effect, safety and feasibility of the topical medication investigated when applied to melanomas located in different anatomical regions.Particularly because there is no established gold-standard treatment for equine melanoma, comparative study protocols should be considered for prospective studies. The phytochemical therapy introduced here and other described treatment modalities including surgery [62,63], radiation [64,65], (electro)chemotherapy [66,67,68] or immunotherapy [39,50,69,70] could be investigated. Moreover, approaches combining the aforementioned therapies with BA or NVX-207 as an adjunctive topical treatment could be the subject of further research.The limitations of this pilot study include the usage of a single horse breed and sex, which limits predictability for a larger, more diverse population. In addition, due to the small number of animals, it was not further investigated whether the state of pregnancy in nine mares as well as the possibly different melanoma growth potential in horses with grey, flea-bitten or white coats could have an influence on the tumor response. Furthermore, no tissue samples of the melanomas treated and surrounding skin were taken, as this was not accepted by the stud management. Prospective studies should include tumor and skin biopsies in order to evaluate local treatment effects histopathologically. Another limitation of the study could be that the amount of cream applied per cm2 of tumor surface was not standardized. However, the application of only slightly different quantities of cream should not have had any effect on the results. On the one hand, there were always cream residues from the previous treatment, so more cream was applied per treatment than the skin could absorb. On the other hand, previous in vitro and in vivo studies showed that the concentration profiles of BA and NVX-207 hardly differ after 30 min and 24 h of incubation, at least in healthy equine skin, and that there is a reservoir effect due to the accumulation of the two substances in the stratum corneum [34,35,36]. Again, biopsies could be helpful in future studies to measure compound concentrations in tumor tissues.5. ConclusionsThe results presented in this pilot study indicate that topical treatment of early-stage equine melanoma with 1% BA and 1% NVX-207 twice a day over a period of 13 weeks is feasible and safe. Especially after BA application, positive effects were observed toward the end of the treatment interval. This suggests that this approach might be a potential therapy for early-stage equine melanoma and, thus, reduce the health risks associated with the possible malignant degeneration of the tumors. However, these findings must be regarded as preliminary due to the limited group size and need to be replicated in a larger cohort. Further, the long treatment duration could lead to poor owner compliance. Accordingly, the study protocol investigated in the current study cannot yet be recommended for melanoma treatment in horses. Modifications of the pharmaceutical formulations may further improve the clinical outcome and reduce treatment periods.

animals : an open access journal from mdpi

10.3390/ani11123377

PMC8698180

The Wolong National Nature Reserve in Sichuan province covers a unique mountainous ecosystem located on the eastern border of the Tibetan Plateau in China. We applied a popular non-invasive observational method, i.e., infrared-triggered camera trapping, to gain thousands of photographs of wildlife to monitor biodiversity over three years. Combined with data on the local abiotic factors, our integrative statistical analysis identified the key environmental drivers, i.e., temperature and vegetation, affecting the distribution and abundance of mammals and birds in the reserve. All species were classified into three main types by their tolerance of or fondness for different environmental conditions. The detectability of each species by camera trapping was quantified and ranked to provide insights on each species’ relative abundance in the area.

The high-altitude ecosystem of the Tibetan Plateau in China is a biodiversity hotspot that provides unique habitats for endemic and relict species along an altitudinal gradient at the eastern edge. Acquiring biodiversity information in this area, where the average altitude is over 4000 m, has been difficult but has been aided by recent developments in non-invasive technology, including infrared-triggered camera trapping. We used camera trapping to acquire a substantial number of photographic wildlife records in Wolong National Nature Reserve, Sichuan, China, from 2013 to 2016. We collected information of the habitat surrounding the observation sites, resulting in a dataset covering 37 species and 12 environmental factors. We performed a multivariate statistical analysis to discern the dominant environmental factors and cluster the mammals and birds of the ecosystem in order to examine environmental factors contributing to the species’ relative abundance. Species were generalized into three main types, i.e., cold-resistant, phyllophilic, and thermophilic, according to the identified key environmental drivers (i.e., temperature and vegetation) for their abundances. The mammal species with the highest relative abundance were bharal (Pseudois nayaur), Moupin pika (Ochotona thibetana), and Himalayan marmot (Marmota himalayana). The bird species with highest relative abundance were snow partridge (Lerwa lerwa), plain mountain finch (Leucosticte nemoricola), Chinese monal (Lophophorus lhuysii), and alpine accentor (Prunella collaris).

1. IntroductionThe main problems that constrain the efficacy of biodiversity conservation in China include habitat loss and fragmentation, overuse and environmental pollution. The underlying causes are unsustainable economic development and the pressures of land development that reduce biodiversity [1]. It is important to monitor wildlife and categorize the collected information on the multi-environmental gradient axes to perform better fitted conservation strategies. Due to the unique geographical and climatic conditions on the border of the Qinghai–Tibet Plateau, researchers have realized the importance of its biodiversity. The region provides quaternary glacial refugia for species, which has been investigated in many studies of ecological observation and biological conservation in the Wolong National Nature Reserve (WNNR; Sichuan Province, PR China). The high-altitude ecosystem of the eastern Qinghai–Tibet Plateau has a profound elevational gradient with high vertical uplift distance in a relatively limited space. Mammals and birds face severe challenges relative to habitat loss due to climate change and human interference, especially for mammals living at high altitudes, which have the highest proportion of threatened species [2,3]. Mammal diversity is an important part of biodiversity and is often used as an indicator of habitat quality and changes to provide an objective and measurable basis for biodiversity research [4].Under the pressure of population growth and urbanization, habitat loss has become more severe in recent decades as the landscape fragments [5]. The Chinese government made efforts to balance the land development and nature conservation by establishing nature reserves and creating an ecological ‘red-line’, i.e., a legally restricted area of land that is strictly prohibited to be farmed or developed [6] Biodiversity conservation requires the creation of national parks to protect wildlife habitats, which is a recent plan launched by the national government [7]. The designation and prioritization of such plans require the monitoring of mammals and their trends in populations, as well as those of other wildlife.Apart from habitat loss, another major challenge for the wildlife in China is climate change, especially in high-altitude mountainous areas where the limited vertical space reduces the flexibility of migration or emigration as a consequence [2,8]. More knowledge of the various responses of wildlife to climatic factors is needed, especially in the least developed and most depopulated areas, such as the mountainous areas of southwestern China. Under global warming, concerns over the biodiversity in alpine ecosystems are deeper than ever because of rapid climate change and acute consequences for local wildlife activities. Monitoring and generalizing the pattern of wildlife in alpine ecosystems is the first step to developing strategies to understand and manage threats with the aim to ensure the long-term persistence of species.Previous investigations have mainly been carried out in the forest ecosystem of relatively low-altitude areas of the WNNR. Research on alpine mammals and birds has been minimal. Infrared-triggered camera trapping, a non-invasive and highly efficient observational technique, has been popular compared to conventional tools such as wireless telemetry collars [9,10]. This technology has facilitated continuous observations of alpine wildlife activities in the southwestern mountains of the Qinghai–Tibet Plateau [11,12]. In this paper, we applied infrared-triggered camera trapping to monitor the biodiversity in the WNNR, mainly aiming to identify (1) the species that occur in the reserve and their activity patterns and (2) the relationship between environmental factors and species occurrence/abundance. The results have implications for habitat improvement measures and provide a baseline for future wildlife research.2. Materials and Methods2.1. Study Area and Sampling SitesBetween 2013 and 2016, a total of 27 Ltl-6210 MC infrared-triggered cameras was installed in the installation sites in the study area. Infrared-triggered camera trapping is a non-invasive method for monitoring the relative abundance of wildlife, with few interferences and relatively high efficiency. In practice, we first selected three valleys where wildlife was most frequently observed in the routine survey (Figure 1).The elevation of the study area ranged from ~3500 to ~4500 m. Then, we designated installation sites within the chosen valleys based on the following criteria: (1) the different habitats were diversified so we aimed to cover all the different main alpine ecosystem types in the nature reserve, i.e., shrub, meadow, and scree; (2) we picked locations where the local topographical and geological conditions allowed us to set up a hidden and stable camera observation and recyclable device. The specific site selection to distribute the cameras mainly depended on the reachability, safety, and micro-environmental conditions in the field.As a result, 27 sites, including 13 in Tizi Valley, 9 in Yinchang Valley, and 5 in Weijia Valley, were used for camera trapping (Figure 1). The elevation of these installation sites ranged from 3536 m to 4481 m. Cameras were set 40–50 cm above the ground in the tracks of wildlife (i.e., near traces of mammal scats and mammal or bird footprints), paths of beasts (i.e., the narrow passageways shaped by frequently passing beasts), and mountain ridges. The linear distances between cameras in the same valley ranged from ~100 m to ~500 m. An animal movement triggers a camera to take three photos over 20 s followed by a 10–20 s video. Every three months, we checked and collected the data captured by the cameras. During the monitoring period, four cameras were lost and three cameras were broken, which were replaced within three months.2.2. Environmental Factors MeasurementWe measured 12 environmental factors, encompassing geographical, vegetation, energy, and water environmental information, during camera installation in the trapping locations (Supplementary Table S2). We measured geographical and topographic metrics, including latitude, longitude, altitude, slope, and aspect, by a handheld GPS. The vegetational information, including species information, herbal coverage (0~100%), and shrub coverage (0~100%), were measured in an area of 5 m × 5 m within a radius of 20–30 m around the camera installation point (Figure 1). In the quantification of the habitat types, we integrated the herb coverage, shrub coverage, and community-dominant species information with the aim of transferring the categorical factor of the habitat types to numerical scores to reflect that the habitat selected was a continuous spectrum, naturally mixed with land cover of herbal grassland, shrub grassland, and nearly bare rocks—though most places at our study area were scree landscape. For evenness of distribution in the quantification, a formula was constructed to assess habitat types as follows: score = {10×shrub coverage×Wshub+herbal coverage,without the dominant species of alpine meadow 10×shrub coverage×Wshub+herbal coverage + 1,without the dominant species of alpine meadow where Wshub is the weight of the shrub species in the community. The dominant species of alpine meadow refers to the plants belonging to the genuses Pedicularis, Ligularia, and Care, as these are the major components (over approximately 50%) of the community around the site. Based on the score, the four main habitat types were generalized and labeled ‘alpine shrub’ (score > 12), ‘alpine meadow’ (12 ≥ score > 9), ‘alpine scree–meadow’ transition zone (9 ≥ score > 6), and ‘alpine scree’ (score ≤ 6). The cameras recorded the ambient temperature as wildlife activity temperature when triggered to take photographs of wildlife. For the vegetational environment, the score of habitat type reflected comprehensive living conditions as an integrated indicator, while shrub or herbal coverage reflected how plants grew in and covered the survey site from a single aspect. Based on these records, maximum activity temperature, minimum activity temperature, and average activity temperature were calculated for the subsequent analysis. The linear distance from the camera installation sites to the closest water source was measured in the projected map on the QGIS platform with the geographical coordinates of the camera installation sites.2.3. Calculation of Relative AbundanceThe photos captured by the cameras were analyzed and the captured videos were used to assist species recognition. Photos of the same species from one camera within 30 min or consecutive photos of different species were defined as individual effective detections [13]. The relative abundance (RA) is positively proportional to the wildlife population sizes [11,12] and was calculated with the following formula:RA = (A/N) × 1000(1) where A represents the number of efficient captures of certain species, and N represents the number of camera-working days. One camera-working day was defined when a camera worked continuously for 24 h. The RA was calculated for each species.2.4. Environmental Factors and Species AnalysisSpecies’ RA and environmental factors were used for detrended correspondence canonical analysis (DCCA) to detect principal environmental factors and to cluster species [14]. The X- and Y-axis were added to indicate key clustering factors and to classify species.3. Results3.1. The Overall BiodiversityThe cameras recorded for 7056 days and we retrieved ~90,000 photos and ~30,000 video clips, including 2251 effective detections (see the definition in Section 2) of wildlife, leading to the identification of 37 species (Figure 2). The invalid photographing (i.e., photos without wildlife detection) represented ~20% of all trigger actions. A total of four of the 37 photographed species are listed as category I species and seven as category II species in the Chinese National List of Protected Animals (Supplementary Table S1). Among them, the Sichuan takin (Budorcas tibetanus) and Chinese monal (Lophophorus lhuysii) are species endemic to China. Of the 37 photographed species, the snow leopard (Panthera uncia), Chinese monal, and beech marten (Martes foina) are endangered species in China according to the listing criteria of the Chinese National List of Protected Animals. The Sichuan takin and Chinese goral (Naemorhedus griseus) are vulnerable species, and we also identified 11 near-threatened species [15]. Other species, mostly small mammals and passerines, are listed as ‘Least Concern’ by the International Union for the Conservation of Nature (IUCN).3.2. Environmental Factors Analysis and Species ClustersThe results of the DCCA are shown in Figure 3. The 12 environmental factors orthogonal to each other form two principal component axes (X, Y), in which the environmental factor arrows are distributed in three main directions (X1, X2, and Y). The X-axis indicates the material environment information that changes with the elevation gradient, with X1 as a geographical factor and X2 as a vegetation factor. The Y-axis indicates the energy environment information that includes temperature-related factors and terrain gradients. The most dominant component of X1 is elevation, and other factors closely related to X1 include the distance between observation sites, water sources, and maximum activity temperature (Tmax on the graphs). X2 includes herbal vegetation coverage, shrub coverage, and habitat type. The opposite directions of X1 and X2 indicate that alpine vegetation coverage decreases with an increase in elevation. The angles among the factors on the Y-axis are less than 90 degrees, indicating positive relationships between these factors. Aspect, slope, and minimum activity temperature (Tmin on the graphs), which are close to each other, are relatively accordant with the principal component Y-axis (Figure 3a). In Figure 3b, aspect and minimum activity temperature are close and mostly accordant with the Y-axis, while slope could be partly projected to the reverse extension line of the Y-axis. Aspect and slope are related to energy balance by sunlight. Ambient temperature is directly related to the energy of the biological activity and metabolic processes of the individuals [16]. Therefore, the minimum activity temperature is one of the key environmental factors for the observed wildlife in alpine ecosystems.According to the results of the DCCA analysis, species are categorized into three groups. The first group is defined as cold-resistant species (purple oval in Figure 3), which are distributed on the positive side of the X1-axis and around the opposite extension line of the Y-axis. This indicates that they are distributed in high-altitude areas and have a lower minimum activity temperature. The second group is defined as phyllophilic species, which react positively to foliage (green oval). These species are distributed on the positive side of the X2-axis, indicating that they are highly dependent on habitats with high vegetation cover at low altitudes. The third group is defined as thermophilic species, which tend to select warmer places (yellow oval). These species are distributed in the positive side of the Y-axis and have a relatively shorter projection on the X-axis, indicating that they have a higher minimum activity temperature and less sensitivity to environmental materials (i.e., vegetation and water resources).3.3. Detection Probability of the Species GroupsThe three mammal species with the highest RA are bharal (Pseudois nayaur), Moupin pika (Ochotona thibetana), and Himalayan marmot (Marmota himalayana). The four bird species with the highest RA are the snow partridge (Lerwa lerwa), plain mountain finch (Leucosticte nemoricola), Chinese monal, and alpine accentor (Prunella collaris) (Figure 4). The mammals are partly grouped as phyllophilic species and partly grouped as thermophilic, while the birds have the highest relative abundance in both the cold-resistant and phyllophilic groups. For thermophilic species, the relative abundance of mammals is higher than that of birds.4. Discussion4.1. Ethology of the Wildlife in Alpine CommunityThe characteristics of alpine environments include low average temperatures, strong winds, and prolonged snow covers [17]. Wildlife living at high elevations must cope with these extreme environmental conditions. Their behavior, biology, and life-history traits may reflect species’ responses to alpine environments.Our cameras captured approximately eight wolves on the alpine scree of Weijiagou. These are the first images and videos of wolves in the WNNR. Beech martens were found to be nocturnal, and they showed collective behavior in July. This is consistent with a previous study [18] that found that beech martens were active at night and that their mating season is in July and August.Mountain weasels (Mustela altaica) were widely distributed; they were active throughout the year. Our cameras captured the moments at which a mountain weasel preyed on a Moupin pika and a snow partridge, which are essential for dietary analysis. The hog badger (Arctonyx collaris) and Himalayan marmot hibernate in winter; hog badgers spend approximately six months per year in hibernation, while Himalayan marmots spend approximately four months in hibernation. Swinhoe’s striped squirrel (Tamiops swinhoei) was observed in July at an elevation of ~4200 m. Liu et al. [19] also captured swinhoe’s striped squirrel at elevations ranging from 4200 m to 4500 m. Sichuan takins were observed on alpine scree at an elevation above 4200 m in summer, which partly confirms that Sichuan takin has a vertical distribution of habitats at elevations ranging from 1500 m to 4500 m as reported in previous observations [20]. Chinese gorals typically prefer a forest habitat [21]. We observed Chinese gorals on alpine scree at an altitude of ~4200 m. This is the first report of a Chinese goral in alpine scree.Snow partridge and Tibetan snowcock (Tetraogallus tibetanus) mainly stay in areas with an altitude of above 4200 m year round, which is not consistent with their seasonal vertical migration observed in previous studies [22,23]. Further research is needed to explore the reasons for this discrepancy. We found that Chinese monal, blood pheasant (Ithaginis cruentus) and passerine birds migrated vertically.4.2. Habitat Environmental Factors RelationshipsAccording to the results of the DCCA, bharal, wolf and snow leopard are cold-tolerant species; thus, their suitable habitats may overlap. We inferred that the bharal is the main prey of the wolf and the snow leopard, indicating that wolves may compete with snow leopards.The abundances of the other four carnivores, i.e., beech marten, red fox (Vulpes vulpes), hog badger, and mountain weasel, were similar. Both red foxes and hog badgers were frequently distributed in habitats with plenty of vegetation, and they may compete for resources. According to the intersectional part of ‘cold-resistant’ and ‘thermophilic’ analysis (Figure 3a), Swinhoe’s striped squirrel and beech marten are distributed over both cooler and warmer areas, indicating that these species may have two differentiated populations occupying different habitats or they might migrate seasonally. Additional information about food sources may help clarify ecological relationships in the alpine community. With that information, we could build alpine food-web models and predicate changes in local communities.According to the RA rankings of birds, cold-resistant species show the greatest abundance. For example, Chinese monal had the greatest abundance and snow partridge had the third greatest abundance followed by alpine accentor. The phyllophilic species had lower abundances than the cold-resistant species, although the plain mountain finch (Leucosticte nemoricola) had the second greatest abundance. The thermophilic species had the smallest abundance. Many bird species were distributed in more than one type of habitat, e.g., blue-fronted redstart (Phoenicurus frontalis), grandala (Grandala coelicolor), and plain-backed thrush (Zoothera mollissima).4.3. The Outlook and Implication of the Camera-Trapping ApproachEach capture of the camera trapping includes one video clip and three photos taken consecutively. These video clips are not merely replications or supplementary information but are quite useful observational materials for identifying species and studying the ethological features of wildlife. However, we have not yet performed further analysis to dig into animal ethology and their function in the communities and ecosystems, which could be investigated as a next step. To improve the efficiency and accuracy of the camera-trapping approach for wildlife monitoring, advanced digital image processing techniques such as machine learning could be applied to these photos and videos, which are typical ‘big data’ with high volume and dynamical streaming accumulation.Our monitoring and analysis of alpine biodiversity not only provide fundamental knowledge and evidence of how mammals and birds are distributed across various environmental gradients in the reserve but also present an example of collecting species information via a non-invasive observational technique in an alpine ecosystem and generalizing groups of wildlife species through multivariate statistical analyses. In addition, the quantified abundance information of wildlife can be referred to when setting out localized species conservation strategies, as well as policy and nature reserve planning. Admittedly, even at our best, in this stage of our pilot study, there is a limitation to the number and representativity of sampling sites, which restricts our analysis to the current depth. Further efforts could be made to develop models for mapping species distributions with data collection at a larger spatial range designed in a systematically gridded plan on our present basis.5. ConclusionsIn this study, we monitored and accessed the activities of mammals and birds in the mountainous forest in WNNR of Sichuan, China, by a camera-trapping approach. We classified the wildlife into three groups ((1) cold-resistant, (2) phyllophilic, and (3) thermophilic species) based on their responses to key environmental drivers, i.e., temperature and vegetation. We calculated and ranked the relative abundance of the observed species. The mammal species with the highest relative abundance were bharal, Moupin pika, and Himalayan marmot. The bird species with the highest relative abundance were snow partridge, plain mountain finch, Chinese monal, and alpine accentor. Our results offer fundamental knowledge on the distribution patterns of mammals and birds and their relationship with key environmental factors in the alpine ecosystems in southwestern China, which could assist further strategizing for local biodiversity conservation under environmental gradients.

animals : an open access journal from mdpi

10.3390/ani11030856

PMC8002700

Hemp (Cannabis sativa L.) is an annual herbaceous plant, globally distributed and cultivated in the past as an important source of fiber. Recently, the interest in hemp cultivation has significantly increased, considering its positive impact on the environment and the production of feed and food of high nutritional value. The authorized hemp varieties are registered in the EU’s Common Catalogue of Agricultural Plant Species and have content in psychotropic 9-tetrahydrocannabinol (THC) less than 0.2–0.3%. In this review, the chemical and nutritional values of hemp are presented and the effects of inclusion of the hempseed, and products obtained by the processing of seed (co-products), in diets for dairy ruminants (i.e., cows, ewes, goats, and buffaloes) are discussed. Hemp supplementation could be a good feeding strategy to improve the bioactive compounds in milk and derivatives: the content of n-3 fatty acids and isomers of conjugated linoleic acid, substances beneficial to human health, increased in milk and cheese obtained with hemp addition. However, up to now, few publications do not allow to suggest the optimal dosage of the co-products for the different species. In addition, no experiments are published on the use of whole plants as forage for dairy ruminants.

Recently, hemp (Cannabis Sativa L.) was rediscovery as a plant that offers a wide variety of applications (textile, pharmaceuticals, construction, etc.), including also the use in animal and human nutrition. The inclusion of whole seeds and co-products obtained by processing of seeds (cake, meal, and oil) in the diets of farm animals can allow the transfer of bioactive substances to human food. Few publications are available on the use of hemp in dairy ruminants but some authors reported a positive effect on the fatty acids profile of milk and cheese with an increase of n-3 fatty acids and c9,t11 conjugated linoleic acid. The protein content, amino acids profile, and rumen undegradable protein (RUP) of hempseed and co-products of hemp appear interesting and suitable for ruminant nutrition. Negative effects of anti-nutritional factors (i.e., phytate) are not observed. However, the researches on the effects of the use of hempseed and co-products in diets for dairy ruminants do not allow to suggest optimal levels of inclusion. In addition, no data are published on the use of whole or part of the hemp plant as forage, as another possibility to use the hemp in the perspective of the circular economy.

1. IntroductionThe consumption of animal products (meat, milk, and eggs) is growing globally mainly due to an increase in world population, greater incomes, and urbanization [1]. The growing demand for livestock products can have an undesirable impact on the environment, considering, in particular, low energy conversion ratio from feed to food and the high requirements of land and other input (i.e., water, nitrogen) to produce the feed for animals. Ruminants are animals with a lower efficiency to convert the energy of feed in food considering the losses due to rumen fermentation processes. On the contrary, ruminants play an important role in the bio-economy by converting food not edible by humans (i.e., forages, crop residues, and agricultural by-products) into high nutritional value food [2,3].On this basis, alternative plants have been recently rediscovered and reintroduced on the agricultural surfaces by exploiting (i) their higher resistance to the adverse conditions (i.e., drought, pathogens); (ii) their role as phytoremediation and soil revitalization [4]; and (iii) their lower nutritional requirements compared to traditional sources of energy and protein in ruminant feeding (mainly corn meal and corn silage, soybean meal, etc.). The hemp plant (Cannabis sativa L.) is undoubtedly one of the most cultivated plants throughout history in the world.The surface of hempseed in Europe, estimated by European Industrial Hemp Association (EIHA) [5], was about 50,081 hectares in 2018 with an increase of 3.3%, 70%, and 614% compared with 2017, 5-years average and 1993, respectively. The major producers in the world are Canada and USA with an estimated 315,000 and 1160 hectares respectively, as reported by Semwogerere et al. [6].In the past, hemp has been cultivated primarily to obtain fibers from the stem [7,8]. The seeds have traditionally been used for therapeutic purposes and in pharmaceutics and chemistry [9], and the cannabinoid-containing flowers have been utilized for medicinal, spiritual/religious, and recreational purposes [10].In Europe, the varieties allowed to be cultivated must be listed in the European Union (EU) Common Catalogue of Varieties of Agricultural Plant Species. The varieties must contain <0.2% delta-9-tetrahydrocannabinol (THC, in dry matter basis), which is the main psychoactive substance [11]. The interest to this plant is mainly oriented to produce seed for human and animal nutrition, shives for construction (green building) and animal bedding, and fibre for textile and paper industry (“industrial hemp”). In dairy ruminant nutrition, hempseed and derivatives (oil, cake and meal) can be used as a supplement in feed mainly as sources of essential fatty acids and essential amino acids [12].The aim of the present review paper was to report an update of data on the chemical and nutritional characteristics of hempseed and derivatives and a state of the art on the researches on the use in dairy ruminant feeding, considering their effects on the milk yield and quality.2. Chemical Composition and Nutritive Value of Hempseed and Derivatives2.1. Chemical Composition and Nutritive Value of Full-Fat HempseedThe whole (full-fat) hempseed (HS) can be used as fed in the animal feeding or after the treatments to removal lipid components using cold mechanical pressing in order to obtain hempseed cake (HC) or, less frequently, by chemical extraction using organic solvents to obtain hempseed meal (HM). Some authors use the term “hemp meal” or “hemp flour” to indicate the product obtained by the mechanical extraction because the cake is often subjected to grinding and then it is in the form of powder. In this paper, “hempseed cake (HC)” will be used for both of these products.Hempseed varieties, which are generally used for animal nutrition, are considered THC free; however, some studies have reported traces of THC present in the seed sample probably because it was contaminated with plant debris [13].In Table 1, data of the chemical composition of the full-fat hempseed reported in the literature are shown. The expected differences of the chemical composition in the published studies are due to the effect of variety/cultivar, preliminary treatments (i.e., decortication), different pedological and climatic situations, and agronomical practices. The ripened seed of hemp is an excellent protein source in animal feeding (on average 24.8 ± 2.0% on dry matter, DM). A similar value of crude protein (25% on DM) for hempseed was reported by European Food Safety Authority (EFSA) [11]. Considering other protein sources, largely diffused in animal feeding, the hempseed can be located as an intermediate crude protein (CP) source between soybean (39.2 ± 5.4% on DM) and sunflower seeds (19.2 ± 4.2% on DM) [14]. The average percentage of lipids in hempseed is very high and results in 30.9 ± 4.2% on DM. Lower values were found by Arango et al. [15], considering six different varieties, cultivated in the north of Italy (province of Rovigo) in 2019.The neutral detergent fiber (NDF) content (Table 1) showed a large variability among the authors, ranging from 29.7–37.2% on DM. Only four publications reported the energy value of hempseed, resulting on average 2422 ± 97 and 946 ± 117 kJ/100 of DM respectively for gross and net energy for lactation in sheep [16].Identification and characterization of hempseed proteins showed that edestin, rich in valuable amino acids (as glutamic acid and aspartic acid), is the main protein component in isolate hempseed protein fraction [24]. Another protein structure, rich in methionine and cystine, was found in hempseed and, subsequently, characterized as an albumin protein family member [25]. Callaway et al. [9] reported, for the first time, the amino acidic profile of hempseed (cultivar Finola) in comparison with the other protein sources. The composition of essential amino acids of hempseed, soybean, and rapeseed [9] compared with the reference pattern recommended by FAO/WHO/UNU [26] in human nutrition, is presented in Figure 1. The contents of the sulphur-containing amino acids and histidine of hempseed are very similar to those of the other two protein sources. Only levels of lysine, threonine, and tryptophan are lower in hempseed compare to soybean and rapeseed. Considering the reference pattern of FAO/WHO/UNU [26] for adults, the limiting amino acid of hempseed is lysine (chemical score: 0.23).Hempseed contains anti-nutritional compounds that reduce the absorption of protein and micronutrients. In particular, the phytate (inositol hexaphosphate) content in the seeds and cake of hemp can be over 5% [27]. The absorption of mineral elements and vitamins can be reduced by phytic acid, during the gastrointestinal passage, producing an insoluble final product [28]. Therefore, an additional amount of microelements is needed to maintain the efficiency of the metabolic processes that support growth, development, and a correct functioning of the organism [29]. Reggiani and Russo [30] observed that the replacement of 6.4% (on DM basis) of corn and soybean with hempseed or flax, maintaining the diets isonitrogenous, can increase the availability of iron in Alpine lactating goats. The authors speculate that some substances (i.e., inulin) contained in hemp or flax seeds can stimulate the absorption of iron.2.2. Chemical Composition and Nutritive Value of Hempseed MealAfter oil extraction, the hempseed cake (HC) can be used as optimal protein source for dairy ruminants. The chemical composition of HC has been reported by several authors (Table 2). As expected, crude protein content increases in HC in comparison with hempseed, and the average value is 34.3% on DM. As other oilseeds, cold mechanical extraction of seed produces a cake that is higher in oil compared with the corresponding chemically obtained meals. The method of extraction is very important not only to obtain an oil of good quality but also to have a high oil yield [31,32]. The percentage of residual oil in the cake is 11.7–12.5% on DM for all authors; only Silversides [20] found a higher concentration of lipids (17.9% on DM). The content of fiber fractions increases in the hemp cake in respect to hempseed (about +27% and +42% for NDF and ADF, respectively).2.3. Chemical Composition and Nutritive Value of Hempseed OilThe quality of the oil obtained by chemical extraction is lower than that obtained by mechanical extraction. For this reason, hempseed meal is used mainly in the industrial processes (lubricants, detergents, paints).In Table 3 is shown the fatty acid profile of whole hempseed (HS), hemp cake (HC) and hemp oil (HO) reported in the literature in order to compare the composition of fatty acids (FA) in the different products.The contents of saturated FA (SFA) were very variable within the different products (from 8.2 to 14.5; from 7.7 to 13.1, and from 7 to 11.6 % of total FA for whole hempseed, cake and oil, respectively). In hempseed products, palmitic (C16:0) and stearic (C18:0) acids represent the higher percentages of SFA (on average 65% and 24% respectively). As known, these long-chain SFA, if consumed in excess, have been associated with increased cardiovascular disease risk in the human population [36,37].The average values of the percentages of mono-unsaturated FA (MUFA) in the three products are very similar (13.4, 12.5, and 13.0% of total FA for hemp seed, cake, and oil, respectively). However, the variability within each group is very high, especially in oil (from a minimum of 9.0 to 20.7% of total FA). The oleic acid (C18:1) represents a very high percentage (from 93 to 98% of total MUFA).As shown in Table 3, the sum of polyunsaturated fatty acids (PUFA) of hemp products is around 75%, and this value is reported as a mean by other authors [9,38,39]. The differences of single PUFA among the three products are very small but, within group, the variability is high, especially for alpha linoleic acid (ALA) in whole hempseed (from 12.98 to 22.4% of total FA). Over 70% of the PUFA are linoleic acid (LA; 18:2 n-6) and ALA (18:3 n-3) [40]. Small amounts of gamma-linolenic acid (GLA; 18:3 n-6) and stearidonic acid (SDA; 18:4 n-3), the biological metabolites of LA and ALA, respectively, were found by some authors (on average 4 and 2% of total FA) [41]. In all publications, the n-6/n-3 ratio is lower than 5:1, which has been claimed as ideal for humans [41,42].3. Use of Hempseed and Derivatives in Dairy Ruminants3.1. Use of Hempseed and Derivatives in Dairy CowsThe interest in the development of different feeding strategies to improve the chemical-nutritional properties of dairy milk and milk products, assuming that nutrition can influence milk composition in ruminants, has grown in the last years [48,49,50,51,52,53]. Considering the high level of n-3 and n-6 fatty acids and the optimal n6/n3 ratio in hempseed, an increase of these PUFA could be expected in milk and derivatives. However, no papers are available to date on the effects of hempseed cake inclusion in the diet of dairy cows on fatty acid profile of milk and derivatives.There is only one published paper on the use of hempseed or its co-products in dairy cows. Karlsson et al. [33] evaluated the effects of increasing the proportion of hempseed cake (HC) in the diet of dairy lactating cows on milk production and composition. Four experimental diets (based on a ratio of 494:506 g/kg of DM between silage and concentrate mixture) were formulated to contain increasing concentrations of HC: 0 (HC0), 143 (HC14), 233 (HC23) or 318 (HC32) g/kg of DM. No effects in DM intake but significant linear increases in CP, fat, and NDF intakes were observed with the increase of the proportion of HC in the diets. Increasing HC dietary levels resulted in significant quadratic effects on the milk yields and energy-corrected milk, with the highest value for the HC14 group (Table 4). The milk protein and fat percentage decreased linearly (p < 0.05) with the increasing of HC in the diet. Furthermore, there was a significant (p < 0.001) linear increase in milk urea concentrations with the enhancement of HC inclusion due to the increase of CP intakes. A linear decrease in CP efficiency (milk protein yield/crude protein intake) was also observed. The best and maximum suggested level of HC inclusion in this experiment was 143 g/kg DM.Mustafa et al. [54] determined the DM and CP in situ degradability in two non-lactating rumen fistulated cows of four different protein sources (hemp, borage, canola, and heated canola meals). The results showed that hemp meal resulted in an excellent natural source of rumen undegradable protein (RUP) (774 g/kg of CP), equivalent to heat-treated canola meal but higher than borage and canola meals.In conclusion, further studies are required to determine the effects of including HC in dairy rations, suggesting to maintain the diets as isoenergetic and isonitrogenous, modifying the proportion of the other ingredients. In addition, the nutritional value of milk and derivatives (i.e., fatty acids profile, vitamins, bioactive substances) could be determined to know the possible nutraceutical effects of hempseed meal.3.2. Use of Hempseed and Derivatives in Dairy EwesEwes milk would naturally have a high content in substances beneficial to human health, such as n-3 fatty acids (FAs) and conjugated linoleic acid (CLA). The n-3 FAs, especially eicosapentaenoic acid (EPA, C20:5 n-3) and docosahexaenoic acid (DHA, C22:6 n-3), can reduce the risk of cardiovascular diseases and in experimental animals, c9,t11 CLA has been proved to possess anticancer and anti-atherosclerotic effects, as well as anti-obesity activities [55]. As above reported, to increase the concentration of PUFA in milk, different sources of unsaturated plant lipids (i.e., linseed, soybean, safflower, and sunflower) could be included successfully in the diet [56,57]. The disadvantage of milk enriched with PUFA is the possibility of oxidation owing to its high content of double-bonded molecules, which are prone to oxidation [58]. The delicate balance between anti- and pro-oxidative processes in milk is influenced by different factors such as ruminant feeding, degree of unsaturated fatty acids, contents of transition metal ions and antioxidants such as tocopherols and carotenoids [59].In this context, Mierlita et al. [18] carried out an experiment using 30 Turcana dairy sheep divided into three groups consisting of a control diet (C diet) based on hay and supplemented by mixed concentrates and two experimental diets designed to provide the same amount of fat using hempseed (180 g/d) (HS diet) or hempseed cake (480 g/d) (HC diet). The three diets were isoenergetic and isonitrogenous, and the two diets with hemp had the same amounts of PUFA. Hemp (HS and HC diets) increased milk yield and milk fat content but decreased milk lactose (Table 5). The hemp feeding increased the PUFA content (especially n-3 fatty acids) in ewes milk and improved the n-6/n-3 ratio. Total CLA content doubled in the milk of the ewes that received hempseed and increased by 2.4 times with the hemp cake inclusion (Table 6). The alpha-tocopherol and antioxidant activity increased using hempseed in the diets, reducing the risk of lipid oxidation in raw milk.Traditionally, ewes on farms are fed indoor or often on part-time grazing during much of the lactation period. During this period, the c9,t11 CLA and n-3 FA contents in milk are lower than that observed during grazing [60]. Mierlita et al. [19] studied the effects of a part-time grazing system or indoor feeding and the supplementation of hempseed in the diet on milk yield and quality, FA profile, and health lipid indices in the raw milk of dairy ewes. Forty ewes were used in this 10-week experiment and were divided into four groups: indoor feeding system with and without hempseed and part-time grazing with and without hempseed. Feeding with the addition of hempseeds significantly increased milk fat content and fat yield (Table 5). Milk protein content was not affected by dietary treatments. Hempseed supplementation increased the content of total PUFA, n-3 and n-6 fatty acids. In the indoor feeding system, the CLA content doubled with the hempseed addition (1.13 vs. 2.29% of total FA) but increased also in the milk of grazing sheep (+37%) (Table 6). As known, the availability of precursors (i.e., linoleic acid) for ruminal bio-hydrogenation and synthesis of CLA is high at pasture when the animals were fed fresh forage [61].Ianni et al. [50] evaluated the effects of a diet enriched with hempseed (5% on DM basis) on the chemical characteristics of milk and cheese from 32 half-bred dairy ewes. The enrichment of dairy ewes’ diet with HS increased the lactose concentration from 4.69% to 5.84% but no significant differences were observed in milk fat, protein, casein, and urea (Table 5). In addition, no changes were detected in total fat, protein, and ash in derived cheeses. During the experiment reported by Ianni et al. [50], the first RNA sequencing of the whole blood transcriptome on ewes of the two experimental groups (0 and 5% of hempseed on DM) was described by Iannaccone et al. [13]. Hempseed supplementation positively affects the pathways related to energy production in lactating ewes. This condition could also be potentially beneficial to increase the resistance to adverse climatic conditions such as low temperature.A digestibility experiment on sheep was conducted by Mustafa et al. [54] using hemp meal (5.2% of lipids on DM) at different levels of inclusion (0, 50, 100, 150, 200 g/kg of DM) in replacement of canola meal, maintaining isonitrogenous diets, based on barley. Voluntary DM intake was not affected by the hemp meal inclusion levels. Total tract DM and organic matter digestibility values were similar across treatments, suggesting that digestibility of hemp meal is equal to that of canola meal. The authors concluded that the hemp meal can be used up to 20% on DM with no detrimental effects on nutrient utilization by sheep.3.3. Use of Hempseed and Derivatives in Dairy GoatsGoat milk has high concentrations of caproic (C6:0), caprylic (C8:0), and capric (C10:0) acids, known to exhibit antiobesity and antidiabetic properties [62]. Also in dairy goats, the interest of modulating milk fat composition by diet leads to the supplementation with feed sources rich in PUFA as an efficient way to modify milk FA profile. The oils extracted by oleaginous seeds can directly affect the fatty acid composition of milk and derivatives but could also have negative effects in terms of animal health status and, in particular, on the efficiency of the rumen microorganisms.Cozma et al. [47] have evaluated the effect of a diet supplemented with hempseed oil in Carpathian goats during 31 days of experiment. No significant changes of milk yield were observed for ewes receiving the hempseed oil supplementation. Fat content increased significantly during the trial in milk produced by goats receiving hemp oil in comparison with the control group. The increase of milk protein content, due to the hemp oil addition, is significant just until day 15 of the experiment and then values remain the same (Table 7).Cremonesi et al. [63] carried out an experiment to evaluate the effects of the inclusion of 9.3% on DM of linseed or hempseed in diet for Alpine lactating goat. The milk yield was unaffected by dietary treatment but linseed and hempseed supplementation significantly increased the milk fat content. No differences were detected in milk protein, lactose and urea concentration (Table 7).Cozma et al. [47] found a significant increase of the PUFA concentrations (+45%) in milk produced by goats supplemented by hempseed oil, without an effect on n-3 fatty acids content. The content of cis-9, trans-11 CLA increased on average by over four times, reaching the peak during the second week of oil supplementation but then decreasing from the third week (Table 8). This transitory effect could be due to an adaptation of the rumen microorganisms to oil supplementation. Hemp oil inclusion had no effect on cholesterol concentration in milk (Table 8), even if plasma cholesterol concentration increased in the ewes fed with oil supplementation. The lack of a relationship between plasma and milk cholesterol concentration could be explained considering that a low proportion of the total milk cholesterol is derived from mammary de novo synthesis. In dairy cows, about 80% of the cholesterol in milk originates from the uptake of serum cholesterol obtained through hepatic synthesis [64]. The overall results of Cozma et al. [47] suggest, for the first time, that beneficial effects on human health can be obtained in goat milk with the inclusion of hempseed oil in the diets.3.4. Use of Hempseed and Derivatives in BuffaloesIn several countries, buffaloes are important species for the production of milk and derivatives for human consumption. There are not any studies related to hemp as feed for improving buffalo milk. Only one published study [65] reported, in the north of Pakistan, possible exposure to delta-9-tetrahydrocannabinol (THC) by the children consuming buffaloes milk. In this region, buffaloes graze in natural pastures, where Cannabis sativa L. with high levels of THC grows spontaneously and higher concentrations of THC metabolites were found in buffaloes milk. As above reported, in EU countries, the hemp varieties allowed for cultivation are registered in the EU’s Common Catalogue of Agricultural Plant Species and are characterized by THC value less than 0.2–0.3% [11].EFSA [11] recommended introducing a maximum THC content of 10 mg/kg to hempseed-derived feed materials to avoid risks for human health due to consumption of food of animal origin.4. ConclusionsThe chemical and nutritional characteristics of hempseed and hempseed derivatives (cake, meal and oil) are updated in the first section of this review. Protein content, aminoacids profile, and ruminal undegradable protein (RUP) make these products suitable for inclusion in ruminant diets. In addition, the fatty acid composition of hemp oil allows to transfer the PUFA and, in particular, n-3 fatty acid to the milk of dairy ruminants, as reported by several authors. Up to now, few publications are available on dairy ruminants to suggest the optimal dosage of hempseed or derivatives in the different species. No information about the use of the whole plant or the different botanical fractions (i.e., leaves) is published.

animals : an open access journal from mdpi

10.3390/ani13081326

PMC10134968

In dairy cows, twin pregnancies and twinning are highly undesirable as they compromise health, welfare, and productive lifespans. We propose that the negative effects of twinning can be avoided by inducing twin reduction. Among dairy cows in their third lactation or more, the incidence of twin pregnancies may be as high as 30%, and different-sex twins represent around 50% of all twin pregnancies. This study provides new unique information on twin pairs during the late embryonic period (28–34 days of pregnancy) that allows for sex selection at the time of embryo reduction, such as differential ultrasound size measurements in heterosexual twins, different intrauterine embryonic growth patterns, and the different vulnerability of female and male embryos following the induced reduction in their co-twins. After heterosexual twin reduction, when the survival embryos were male, pregnancy loss was null.

This study provides new unique information on bovine twin pairs during the late embryonic period (28–34 days of pregnancy) in relation to (1) a predictive ultrasound measurement that was differential for sexing heterosexual twins; (2) intrauterine embryonic growth patterns in twin pairs; and (3) a higher vulnerability of female embryos compared to males following an induced embryo reduction in heterosexual twins. The study population comprised 92 dairy cows carrying bilateral twins. A length difference between co-twins equal to or greater than 25% in around 50% of pregnancies served to determine the sex of embryos with 100% accuracy in heterosexual twins, which was assessed four weeks later on the remaining fetus after twin reduction. The apparent rates of growth of twin pairs and of individual male and female embryos from day 28 to 34 of gestation were similar to established growth pattern standards for singletons. Mean embryo sizes in relation to gestational age were smaller by some 5 days’ growth equivalent in twins compared to singletons. After the reduction in the female embryo in heterosexual twins, the risk of male embryo loss was null. This new information allowed for sex selection at the time of twin reduction.

1. IntroductionIn mammals, a male embryo develops faster than a female embryo before the differentiation of the gonads [1,2,3]. This phenomenon was already described in 1917 as the result of studying freemartinism in bovine twin pregnancies [4,5,6]. Subsequent work confirmed this finding [7,8,9]. Placental vascular anastomosis occurs in most twin pregnancies in cattle, and when heterosexual twins share the same circulation system, the female fetus becomes masculinized and, therefore, sterile [6,10,11]. Heterosexual twins account for some 50% of all twin pregnancies [12,13,14]. Twin pregnancies carry a greater risk of pregnancy loss than single pregnancies [15,16,17], and twin births lead to postpartum reproductive disorders [18,19,20,21], compromising the health, welfare, and productive lifespan of a cow. Consequently, twin pregnancies are detrimental to farming economics [22,23,24]. Multiple pregnancies, mostly twin pregnancies [25,26], may affect almost 30% (1827/6463) of cows in their third lactation or more [26,27]. Given that twin pregnancies are highly undesirable in dairy cattle, twin reduction by a manual rupture of the amniotic vesicle at the time of pregnancy diagnosis (28–34 days of gestation) is routine practice in commercial dairy herds under our surveillance [15,16]. Although this procedure is little used in dairy cattle, induced embryo reduction is a relatively common procedure in women [28,29,30] and mares [31,32,33]. Standards of late embryonic/early fetal growth and bovine conceptus viability have been extensively reported in ultrasound studies [34,35,36,37,38]. The embryonic period of gestation spans from conception to the completion of the stage of differentiation (approximately 45 days), while the fetal period runs from gestation day 45 until parturition [39]. Twin pregnancy is classed as bilateral (one fetus in each uterine horn, 44%) or unilateral (both fetuses in the same uterine horn, right or left, 56%) [26,27]. The likelihood of pregnancy loss may be up to five times higher for unilateral twins than for bilateral twins [40,41]. We designed the present study in bilateral twin pregnancies to examine the association between the co-twin-size differential at the time of induced embryo reduction (at 28–34 days of gestation) and fetal sex observed four weeks later. This new information could allow for sex selection at the time of twin reduction in identified heterosexual twin pregnancies. As the literature lacks embryo twin pair measurements during the late embryonic period in dairy cattle, a second objective was to explore the apparent rate of growth in individual twins. As a third objective, the possible effects of individual twin size, days of gestation, and embryo gender on heterosexual twin pregnancy loss rates were also examined.2. Materials and Methods2.1. Cows and Herd ManagementThis study was performed on three commercial Holstein-Friesian dairy herds kept on farms 1 km apart in north-eastern Spain (latitude 41.13 N, longitude −2.4 E). As individual twin measurements and the pregnancy loss rate was not significantly different among herds data, were grouped as derived from a single herd and “herd”, which were considered as a factor in binary logistic regression procedures. During the study period (October 2021 to January 2023), the mean number of lactating cows in the herds was 7330, and the mean annual milk production was 14,965 kg per cow. The mean annual culling rate was 30%. Cows were grouped according to age (primiparous plus secundiparous versus cows in their third lactation or more) and fed complete rations. The herds were maintained on a weekly reproductive health program in which pregnancy was diagnosed by transrectal ultrasonography from days 28 to 34 post-insemination using a portable B-mode ultrasound scanner equipped with a 5–10 MHz transducer (E.I. Medical IBEX; E.I. Medical Imaging, Loveland, CO, USA). Each ovary was scanned in several planes by moving the transducer along its surface to identify the ovarian structures, and the number and location of corpora lutea (CL) were recorded. The dorso/lateral surface of each uterine horn was scanned to detect the presence of twins. The pregnancy was checked again by ultrasound four weeks later. Pregnancy loss was recorded when this exam proved negative. The viability of an embryo/fetus was confirmed by the observation of a heartbeat in all exams. All gynecological exams and pregnancy diagnoses were performed by the same operator.2.2. Experimental DesignOnly cows in their third lactation or more carrying live bilateral twins (one embryo in each uterine horn) with two CL (one CL in each ovary) were included in this experiment. Primiparous and secundiparous cows were excluded because twins are much more common in cows in their third lactation or more [25,26,27], and the pregnancy loss rate is similar among cows with a parity of three or more [42,43]. The length of each embryo was measured on a frozen ultrasound frame using the electronic calipers of the ultrasound machine, and measurements were adjusted to the nearest one mm. As the embryo initially had the appearance of a straight line which developed a C shape on day 25 and an L shape on day 33 [35], the length of the long axis of the embryo visualized in a dorsoventral plane was recorded as the length of the embryo proper. For twins of different lengths, the length differential was defined as the percentage of difference with respect to the larger embryo. We assumed the larger embryo to be male and the smaller one to be female [1,2,3,37]. In these pregnancies, larger and smaller embryos were alternately ruptured on a weekly rotational basis, and fetal sex was determined from days 56 to 62 of gestation [44,45,46] in the pregnancy confirmation exam. Finding a size differential cutoff for the sexing of twins in the heterosexual group was the first objective of the present study. In the remaining pregnancies, in which a length that was differential between twins was not detected, the embryo located in the right uterine horn was ruptured. All cows received 1250 mg flunixin meglumine plus 100 µg GnRH immediately before or after reduction to promote the maintenance of gestation [15]. For induced twin reduction, the rupture of the amniotic vesicle of a twin embryo was guided by ultrasound and carried out by applying manual pressure. The amniotic vesicle was held and pressed with the thumb and half to cause a rupture (Figure 1). The time lapse between the vesicle subjection and the rupture was no longer than five seconds, and the cows did not show any sign of discomfort. Embryo death was assessed by the cessation of the heartbeat and was observed by ultrasound. All procedures of twin reduction were performed by the same operator.In single pregnancies, factors such as parity (heifers versus parous cows), milk production, body condition score, circulating hormones and metabolites, insemination number, the sire of the embryo, and the horn of pregnancy could not be associated with embryo size during the late embryonic period [37,38]. Therefore, our second objective was to compare growth patterns in twins with established growth standards for singleton pregnancies. As embryo measurements were taken only once between 28 and 34 days of pregnancy, only apparent rates of growth for twin pairs from day 28 to 34 of the pregnancy and embryo measurements for each day of pregnancy were compared to the growth standards for singletons. Examining the possible effects of individual embryo measurements and days of gestation on pregnancy loss rates was this study’s third objective. Two unplanned appraisals were initiated during the course of this study. Once a clear cutoff was observed, which allowed for the sexing of twins), the apparent growth rate for female and male embryos was included in the second objective, whereas the embryo gender of the heterosexual twins was examined as part of the third objective.In our geographical region of study, there are only two clearly differentiated weather periods: warm (May to September) and cool (October to April) [47,48]. The likelihood of pregnancy loss in twin pregnancies may be up to five times higher during the warm than cool period [49]. Temperatures for the study period were: 39 days of minimum temperatures < 0 °C and 7 days of maximum temperatures > 25 °C for October 2021 to April 2022 plus October 2022 to January 2023; and 0 days of minimum temperatures < 0 °C and 101 days of maximum temperatures > 25 °C for May to September 2022. To reduce the risk of pregnancy loss, cows were only included if they were healthy, as indicated by a body condition score of 2.5–3.5 on a scale from 1 to 5 [50], which produced more than 40 kg of milk per day, were free of clinical signs of disease during the study period (Days 28 to 62 of pregnancy), and became pregnant during the cool period of the year. Cows were included only once in the experiment. The final study population comprised 92 cows with twin pregnancies.2.3. Data Collection and Statistical AnalysesThe following data were recorded for each pregnancy: the day of pregnancy, length of individual twins, size differential between twins, semen-providing bull, pregnancy loss, and fetal sex in pregnancies with size differentials between twin pairs. Statistical analyses were performed using PASW Statistics 18 software (SPSS Inc., Chicago, IL, USA). Significance was set at p < 0.05. Variables are expressed as the mean ± standard deviation (S.D.).As a result of the evident size differential cutoff predicting fetal sex, descriptive statistics were only used on these values. Embryo measurements and size differentials for each time point were assessed by one-way ANOVA. A correlation was performed between mean embryo size values and the day of pregnancy. The relative contributions of each factor to the probability of pregnancy loss were determined by binary logistic regression. Pregnancy loss was considered to be the dependent variable, and the day of pregnancy (class variable: days 28, 29, 30, 31, 32, 33, and 34), the mean length of twin pairs, and the size differential between co-twins (continuous variables), and the sex of the embryo, coded as a class variable (non-sexed, female, and male), were considered factors in the analysis. The semen-providing bull (sire of the embryo in this study), a main factor influencing pregnancy maintenance independent of its fertilizing capacity [51,52,53], was not included in the model because a large number of sires was used (14 sires). Four sires provided semen for a single pregnancy each, with six sires for two pregnancies and four sires for three or more pregnancies. Any effect on the semen-providing bull was assumed to be randomly distributed and a possible component of error in the experiment. Regression analyses were conducted according to the method of Hosmer and Lemeshow [54]. Basically, this method consisted of five steps as follows: the preliminary screening of all variables for univariate associations; the construction of a full model using all the significant variables arising from the univariate analysis; a stepwise removal of non-significant variables from the full model and comparison of the reduced model with the previous model for model fit and confounding; an evaluation of the plausible interactions among variables; and assessment of model fit using Hosmer-Lemeshow statistics. Variables with univariate associations showing p values < 0.25 were included in the initial model. Model reduction continued until only significant terms according to the Wald statistic remained in the model at p < 0.05.3. ResultsA length differential was not observed in 34 (37%) of the 92 twin pairs. Table 1 shows length differentials on days 28–34 of pregnancy, which was used to determine embryo sex. Descriptions of embryo measurements (mean, S.D., minimum and maximum) and length differentials between male and female embryos are provided in Table 2. Sex was assessed in 45 co-twins (48.9%) which showed length differentials equal to or higher than 25% (Table 1 and Table 2). Length differentials from 8% to 15% detected in the remaining 13 twin pregnancies (14.1%) were not associated with subsequently determined fetal sex (Table 1). The cutoff for embryonic sex determination was set at a length differential equal to or greater than 25%. This meant that embryo sex could be determined in 48.9% (90 co-twins) of twin pregnancies during the late embryonic period (28–34 days of pregnancy) based on the embryo length differential. Figure 2 shows the embryo size increases produced over time. A significant positive correlation was observed both for the total sample of twins and for male and female embryos (p < 0.0001). The mean percentage of the difference between co-twins with respect to the larger embryo (length differentials) was similar for each day of pregnancy.Of the 92 cows enrolled in this study, 23 (25%) experienced pregnancy loss after twin reduction. Logistic regression analysis indicated no significant effects in terms of the day of pregnancy, the mean length of twin pairs, and the size differential between co-twins on pregnancy loss. Embryo sex was the only variable selected by the logistic procedure (Table 3). Taking non-sexed twin pairs as a reference, male embryos gave rise to an incidence of pregnancy loss that was significantly (p = 0.01) reduced by a factor of 0.08. As these results were unexpected, two further regression analyses were performed. First, when male embryos were used as the reference sex of the embryo variable, results were similar, and the incidence of pregnancy loss for female embryos was increased by a factor of 13. Finally, after removing the variable sex of the embryo, the length differential was the only variable remaining in the model, and one unit increase in the length differential significantly reduced the incidence of pregnancy loss by a factor of 0.92 (95% confidence interval: 0.86–0.97; p = 0.005).4. DiscussionThis study provides new and unique information on bovine twin pairs during the late embryonic period indicating that a predictive threshold value using ultrasound size differential measurements between co-twins can be used to determine the sex of heterosexual twins (50% of all twin pregnancies in cattle [12,13,14]); twin pairs showed some distinguishing intrauterine embryonic growth pattern characteristics compared to established standard singletons; and female embryos showed a greater vulnerability than their male partners following induced embryo reduction.In the present study, length differentials between co-twins equal to or greater than 25% proved valuable to predict the sex of embryos at pregnancy diagnosis (28–34 days of pregnancy) with an accuracy of 100% for fetal sex, determined four weeks later on the remaining fetus after twin reduction. In contrast, length differentials from 8% to 15% were not useful. In effect, based on differentials in this last range for twin pairs, only five out of a total of twelve fetuses were of the expected sex. In other words, these were homosexual twin pregnancies. In an extensive study, sex ratios for twin pairs were 24.9% (male/male), 48.7% (male/female), and 26.4% (female/female) [12]. Probably, homosexual twins showed small growth differences. The fact that no size differentials between 15% and 25% were observed reinforces this statement. Our study population consisted of cows with two CLs suggesting that most of the homosexual twins were dizygotic. This means independent double ovulation and independent embryogenesis. In fact, dizygotic twins have the same genetic relationship as ordinary brothers or sisters, which may be the reason for small growth differences.While we did not take measurements over time on the same co-twins, apparent rates of growth in twin pairs and of individual male and female embryos from day 28 to 34 of gestation resembled the average growth pattern observed in singletons [34,35,36,37,38]. When taking Curran’s work as a reference [35], however, the mean embryo size in relation to the age of gestation was smaller by about 5 days’ growth, which was equivalent in twins compared to singletons. For example, the mean lengths of a singleton on days 28 and 34 were reported at 9.5 and 16.5 mm, respectively [35], while means of 7.4 and 11.7 mm were observed here on the same days of pregnancy for twin pairs. On days 22 and 29 of pregnancy, singletons were 6 and 12 mm long [35]. This could suggest that, for any gestational age, twins are younger than singletons. The truth is that twin pregnancies reach parturition up to eight days earlier than single pregnancies [55,56], and although smaller calves are delivered, twin newborns show the same viability as singletons [56]. The smaller size of twins may be the result of competition for resources in the uterine environment and placental system.In our study, efforts were made to reduce the risk factors associated with pregnancy loss. Thus, we only included cows in their third lactation or more with a similar risk of loss [42,43] that became pregnant during the cool period of the year, when in our region, there was a very low risk of twin pregnancy loss [49]. In this way, we tried to examine the possible effect of the day of pregnancy and embryonic measurements on the risk of loss. A reduced embryo size could not be associated with pregnancy loss contrasting with the situation for single pregnancies, in which a small embryo was considered a predictor of pregnancy loss [57,58]. Only embryo sex was a factor included in the logistic regression. Unexpectedly, female embryos appeared to be adversely affected by the rupture of their male co-twin. Conversely, after a reduction in the female embryo, the risk of male embryo loss was greatly reduced by a factor of 0.08 (0% under our work conditions). In humans, pregnancy failure is more likely if the baby is a boy [59,60,61,62]. Boys grow faster than girls before implantation, and this makes them more vulnerable to compromised nutrition [63,64,65]. Probably, in all mammals, life in the maternal uterus is more dangerous for male embryos than for female embryos. The widely reported size difference in favor of the male embryo [1,2,3,4,5,6,7,8] may support this notion. The present results could also confirm the idea that male embryos are more resistant only in the case of heterosexual twins.The pregnancy loss rate was similar for unsexed twins (34%) and female embryos (30.4%), being slightly higher than the 25% described after induced embryo reduction in bilateral twin pregnancies [15,16]. Probably, losses for heterosexual twins have led to reduced reported incidences of pregnancy loss in previous studies. Total losses were also 25% (23/92) in the present study. The question that arises is why losses for male/male homosexual twins, accounting for 50% of all homosexual twin pregnancies [12], did not reduce the loss rate recorded in the set of unsexed twins. Could this be due to an XX/XY chimerism in heterosexual twins in addition to the more advanced development of the male compared to female co-twin, which would favor the maternal–embryo relationship and so the maintenance of gestation in the male? Bovine freemartinism likely remains the best-known rather than best-understood example of abnormal sexual differentiation in mammals [13,14].Although XX/XY chimeras have been described in both bull and heifer calves from single births [66,67,68], the single-born freemartin remains unknown [14,69]. Determining the level of XX/XY chimerism in males born after a reduction in their female co-twins may help clarify this point.Embryo sex selection at the time of twin reduction in cows in their third lactation or more has acquired particular relevance. The use of sex-sorted semen has increased enormously over the past two decades, particularly in heifers and primiparous cows [70,71,72]. As a result, producers generate their own replacement females from a smaller group of cows, leading to the massive use of conventional beef semen in dairy herds, particularly in multiparous cows [72,73]. In this situation, maintaining the male embryo partner in heterosexual twin pregnancies should maximize the market value of cross-bred calves after birth.5. ConclusionsThe length differentials between co-twins equal to or greater than 25% detected in around 50% of twin pregnancies proved valuable to determining the sex of embryos with 100% accuracy and assessed four weeks later on the remaining fetus after twin reduction. Apparent rates of growth for the twin pairs and of individual male and female embryos from day 28 to 34 of gestation were similar to growth standards for singletons. Mean embryo sizes in relation to gestational age were smaller by about 5 days’ growth equivalent in twins than established singleton standards. After heterosexual twin reduction, when survival embryos were male, pregnancy loss was null. This new information allows for sex selection at the time of heterosexual twin reduction in dairy cattle.

animals : an open access journal from mdpi

10.3390/ani11082183

PMC8388447

Fasting is usually used before metabolizable energy assessment in poultry. Recently, fasting-induced autophagy has been of concern because of the beneficial function of autophagy. In this study, we found that the intestinal autophagy gene Atg7 has a good quadratic fitting with fasting duration. We found that the serum metabolism pathways involved in glycerophospholipid, phenylalanine, GnRH signaling pathways, glycosylphosphatidylinositol anchor biosynthesis, autophagy, and ferroptosis changed with fasting. Furthermore, we found a correlation between intestinal autophagy and serum metabolite PE (18:3(9Z,12Z,15Z)/P-18:0).

Fasting-induced autophagy in the intestine is beneficial for body health. This study was designed to explore the relationship between the host metabolism and intestinal autophagy. Broilers were randomly assigned into 48 cages. At 0 (CT), 12 (FH12), 24 (FH24), 36 (FH36), 48(FH48), and 72 h (FH72) before 09:00 a.m. on day 25, eight cages of birds were randomly allotted to each fasting time point using completely random design, and their food was removed. At 09:00 a.m. on day 25, the blood and jejunum were sampled for serum metabolome and autophagy gene analyses, respectively. The results showed that the autophagy gene Atg7 has a good quadratic fit with fasting duration (R2 = 0.432, p < 0.001). Serum phosphatidylethanolamine (PE) and lyso-PE were decreased in the birds that were fasted for 24 h or longer. Conversely, the serum phosphatidylcholine (PC) and lyso-PC were increased in the birds that were fasted for 36 h or longer. Metabolism pathway analysis showed that the serum glycerophospholipid, phenylalanine, and GnRH signaling pathways were downregulated with the extended fasting duration. The serum metabolites involved in glycosylphosphatidylinositol anchor biosynthesis, autophagy, and ferroptosis were upregulated in all of the fasted groups. Correlation analysis showed that serum PE (18:3(9Z,12Z,15Z)/P-18:0) was a potential biomarker for intestinal autophagy. Our findings provide a potential biomarker related to intestinal autophagy.

1. IntroductionFasting is becoming more popular in the public because of its beneficial effects, such as anti-aging, anti-inflammatory, and body fat loss [1]. Additionally, fasting is also used to empty the gastrointestinal tract before analyzing the energy efficiency of certain feed ingredient in chickens [2,3,4,5]. During fasting, AMP-activated protein kinase is activated and inhibits the mechanistic target of rapamycin in order to suppress cell growth and activate autophagy [6]. Then, activated autophagy scavenges cytoplasmic materials for energy production [7] and regulates gut homeostasis [8,9]. It has been reported that anomalous autophagy is associated with inflammatory bowel diseases [10] and impaired absorptive enterocytes [8]. What is more, activated intestinal autophagy triggered by fasting protects against TNF-induced apoptosis during chronic colitis and improves life span [11,12]. In vitro, it has been reported that the autophagy level varies with fasting durations [13]. However, the autophagy level in the intestine in response to different fasting durations remain unknown.It has been suggested that 12 to 24 h of fasting causes a 20% or more decrease in the serum glucose, and depletion of hepatic glycogen in humans [14]. Consequently, fatty acids, ketone bodies, and amino acids are used as energy sources. Traditionally, blood sampling after overnight fasting (≥8 h) is used to monitor body metabolism in humans [15,16,17] and chickens alike [18]. However, broiler chickens are fed ad libitum, except for 1 to 4 h of darkness every day [19,20]. As at least 8 h are needed for emptying the gastrointestinal tract of chickens [21], the broiler’s systemic metabolism can be better characterized by a nonfasted serum. Nowadays, the recommendation of non-fasting blood samples for blood profile assessment in humans has been emerged [22]. Previous studies have been reported the influence of 58 to 96 h of fasting on the host metabolism in humans and mice [23,24]. Serum metabolites have been shown to be involved in digestive function [25,26], which means that serum metabolites may reflect some functions of the intestine. What is more, white blood cells have been recommended to monitor autophagic flux [24]. It is unclear whether there is a metabolite in the serum that is highly correlated with the level of intestinal autophagy.Liquid chromatography strip–quadrupole mass spectrometry (LC–MS) is regarded as the best choice for vast serum metabolite detection [27]. Therefore, in the present work, we aimed to study the impact of different fasting durations on intestinal autophagy and LC–MS-based untargeted serum metabolites. Furthermore, the potential relationship between serum metabolites and intestinal autophagy was explored.2. Materials and MethodsAll of the procedures were complied with the Beijing Regulations of Laboratory Animals, and the Laboratory Animal Ethical Committee of China Agricultural University approved this study (no. AW04110202-3).2.1. Bird ManagementA total of 240 one-day-old male broiler chickens (Arbor Acres Plus) were obtained from a commercial hatchery and were randomly assigned to cages with five birds in each cage. The broiler chickens were raised on wire net floors and received a lighting program 23L:1D (turn off at 23:00 p.m. and turn on at 24:00 p.m.) on arrival, which was transformed to a new lighting program 20L:4D (turn off at 22:00 p.m. and turn on at 02:00 a.m.) from day 8. The room temperature was initially set at 33 °C and then gradually decreased according to the age of the birds, until reaching 23 °C on day 21. All of the broilers had free access to water and feed. Broiler chickens were fed a standard corn–soybean meal-based diet, as shown in Table 1. Titanium dioxide (5 g per kg diet) added in the feed to collect digestibility data for other work.2.2. Sample CollectionOn day 21, several cages of birds were selected and weighed to get 48 birds with similar body weights. Then, these 48 birds were randomly divided into 48 cages with one bird per cage. At 0 (CT), 12 (FH12), 24 (FH24), 36 (FH36), 48(FH48), and 72 h (FH72) before 09:00 a.m. on day 25, eight cages of birds were randomly allotted to each fasting time point using completely random design, and the food was removed (Figure 1). At 09:00 a.m. on day 25, the blood samples were drawn from all of these birds’ wing veins. The birds were then euthanized by intravenous injection of sodium pentobarbitone (30 mg/kg). Middle jejunum (about 0.25 cm) was collected and washed with saline solution, then snap-frozen in liquid nitrogen immediately, and stored at −80 °C for mRNA analysis. After 4 h of routine temperature shelving, the serum was obtained by centrifugation at 3000 g for 10 min at 4 °C. Samples of jejunum and serum were stored at −80 °C for later analysis.2.3. Chemical AnalysisThe feed samples were finely ground and passed through a 40 μm strainer. The gross energy in the feed was determined using an oxygen bomb calorimeter (PARR 6400; Parr Instrument Company, Moline, IL, USA) with benzoic acid as the calibration standard. The nitrogen in the feed was determined using the Kjeldahl method (FOSS KT 200 Kjeltec, Hillerod, Denmark) [28], and CP was calculated using 6.25-fold of nitrogen. The crude fat in the feed was determined in double 1 g samples, packaged into filter-bags (Ankom XT4; ANKOM Technology, Macedon, NY, USA) and extracted by petroleum ether at least 50 times, using a Soxhlet extractor for the ether extract analysis. The starch in the feed was analyzed using a total starch assay kit according to the manufacturer’s instruction (KSTA; Megazyme, Bray, Wicklow, Ireland). Three technical replicate samples were used for the determination of the total calcium and total phosphorus content of the feed using the permanganate titration method and phosphovanoclonoly beate molecular absorption spectrometric method, respectively.2.4. Autophagy-Related Gene ExpressionThe total RNA was extracted from the chicken jejunum using RNAiso Plus (9109; TAKARA, Kyoto, Japan). The purity and concentration of the total RNA were measured with a nucleic acid analyzer (Nano-drop 2000, Thermo Fisher Scientific, Waltham, MA, USA) using the 260:280 nm absorbance ratio. The complementary DNAs of chicken Atg7 (Accession no. NM_001030592.1) were amplified with the primer as follows: Atg7 sense, CACTGCGGAACTTCCTGATCTTGG, and Atg7 antisense, CTTGCATGGTCCT GTCTCTGAAGC. According to the manufacturer’s instructions, cDNA synthesis was performed using a PrimeScript RT reagent kit with a gDNA eraser. Real-time PCR was conducted on a 7500-fluorescence detection system (Applied Biosystems, Carlsbad, USA) and set as described previously [29]. The mRNA level of Atg7 was normalized against the housekeeping gene β-actin (accession no. XM_027015741.1, amplified as sense CAACACAGTGCTGTCTGGTGGTAC and antisense CTCCTGCTTGCTGATC CACATCTG) using the 2−ΔΔCt method.2.5. Blood Serum PreparationThe blood serum samples stored at −80 °C were thawed at room temperature. Then, the samples were analyzed using LC–MS by OE Biotech. Co., Ltd., in Shanghai. Then, 100 μL of each sample was added to a 1.5 mL tube with 10 μL of 2-chloro-l-phenylalanine (0.3 mg/mL) dissolved in methanol as the internal standard, and then vor-texed for 10 s. Subsequently, 300 μL of the ice-cold mixture of methanol and acetonitrile (2/1, v/v) was added, and the mixtures were vortexed for 1 min, ultrasonicated at an ambient temperature (25 °C to 28 °C) for 10 min, and stored at −20 °C for 30 min. The extract was centrifuged at 13,000 rpm, 4 °C for 15 min. The supernatants (200 μL) from each tube were collected using crystal syringes, filtered through 0.22 μm micro-filters, and transferred to LC vials. The vials were stored at −80 °C until the LC–MS analysis. Quality control samples were prepared by mixing aliquots of all of the samples to be a pooled sample.2.6. Liquid Chromatography–Mass Spectrometry AnalysisA Shimadzu 30AD UHPLC system fitted with a Q-Exactive quadrupole-Orbitrap mass spectrometer equipped with a heated electrospray ionization source (Thermo Fisher Scientific, Waltham, MA, USA) was used to analyze the metabolic profiling in both the ESI positive and ESI negative ion modes. An ACQUITY UPLC BEH C18 column (1.7 μm, 2.1 × 100 mm) was employed in both the positive and negative modes. The binary gradient elution system consisted of (A) water (containing 0.1% formic acid, v/v) and (B) acetonitrile (containing 0.1% formic acid, v/v), and separation was achieved using the following gradients: 5% B over 0–2 min, 20% B over 2–24 min, 100% B over 24–28.1 min, and 5% B over 28.1–30 min. The flow rate was 0.2 mL/min and the column temperature was 40 °C. All of the samples were kept at 4 °C during the analysis. The injection volume was 5 μL.The mass range was from m/z 66.7 to 1000.5. The resolution was set at 70,000 for the full MS scans and 17,000 for the HCD MS/MS scans. The collision energy was set at 20, 40, and 60 eV. The mass spectrometer was operated as follows: spray voltage, 3000 V (+) and 2500 V (−); sheath gas flow rate, 35 arbitrary units; auxiliary gas flow rate, 10 arbitrary units; and capillary temperature, 320 °C. The QCs were injected at regular intervals (every 10 samples) throughout the analytical run in order to provide a set of data from which the repeatability could be assessed.2.7. Data Preprocessing and Statistical AnalysisThe acquired LC–MS raw data were analyzed by the progenesis QI software (Waters Corporation, Milford, NE, USA) using the following parameters. The precursor tolerance was set to 5 ppm, the fragment tolerance was set to 10 ppm, and the retention time (RT) tolerance was set to 0.02 min. The internal standard detection parameters were deselected for the peak RT alignment, isotopic peaks were excluded for analysis, the noise elimination level was set at 10.00, and the minimum intensity was set to 15% of base peak intensity. The resulting matrix was obtained with three-dimension data sets, including m/z, peak RT, and peak intensities, and the RT–m/z pairs were used as the identifier for each ion. The resulting matrix was further reduced by removing any peaks with a missing value (ion intensity = 0) in more than 50% of the samples. The internal standard was used for data quality control.Metabolites were identified by progenesis QI (Waters Corporation, Milford, NE, USA) Data Processing Software, based on databases such as http://www.hmdb.ca/; http://www.lipidmaps.org/ and self-built databases (accessed on 6 December 2018). The positive and negative data were combined to get combine data that was imported into the R ropes package. Principle component analysis (PCA) and orthogonal partial least-squares-discriminant analysis (OPLS-DA) were carried out to visualize the metabolic alterations among the experimental groups, after mean centering (Ctr) and Pareto variance (Par) scaling, respectively. The Hotelling’s T2 region, shown as an ellipse in the score plots of the models, defines the 95% confidence interval of the modeled variation. Variable importance in the projection (VIP) ranked the overall contribution of each variable to the OPLS-DA model, and those variables with VIP >1 were considered relevant for group discrimination.In this study, the default seven-round cross-validation was applied, with one/seventh of the samples being excluded from the mathematical model in each round, in order to guard against overfitting. The differential metabolites were selected based on the combination of a statistically significant threshold of variable influence on the projection values obtained from the OPLS-DA model and p values from a two-tailed Student’s t-test on the normalized peak areas, where metabolites with VIP values larger than 1.0, p values less than 0.05, and fold change over 2 were considered as differential metabolites. MBRole was used for the metabolite pathway enrichment analysis.The data of the Atg7 gene expression were analyzed by ANOVA and Tukey was used in the post hoc test. Pearson’s correlation was used to analyze the correlation between PE and autophagy gene Atg7 using an R statistical package named Corrplot. The values were means, with SEM represented by vertical bars. Treatments with different letters were significantly different (p ≤ 0.05).3. Results3.1. Intestinal Autophagy at Different Durations of FastingThis study showed that the mRNA level of Atg 7 was significantly (p < 0.05) upregulated at every fasting time point, except for 12 h of fasting (Figure 2). Linear regression did not provide a good fit between the Atg 7 and fasting durations (R2 = 0.092, p = 0.036), but quadratic fitting provided a good fit (R2 = 0.432, p < 0.001).3.2. PCA and OPLS-DA Analysis of Chicken Serum at Different Durations of FastingThe PCA score plots suggested a distinct separation among the different fasting durations (Figure 3A). All of the samples in the score plots of the samples were within the 95% Hotelling T2 ellipse. The OPLS-DA revealed a clear separation between the non-fasting and fasting groups (Figure 3B–F). The OPLS-DA models indicated that different fasting time points all significantly affected the serum metabolic patterns.3.3. Serum Metabolic Profiles Changes with Different Durations of FastingCompared with the birds in the CT, the number of changed metabolites decreased within 36 h of feed deprivation and increased with longer fasting durations (Figure 4B). The serum metabolomics showed the union set of the changed metabolites between different fasting groups and the non-fasting group reached 194 species, the commonly changed species were 23 species (Figure 4A and Table S1). The metabolites in the union set were fatty acyl groups (52 species, accounting for 26.8%), glycerophospholipids (62 species, accounting for 32.0%), carboxylic acids and derivatives (13 species, 6.7%), steroids and derivatives (7 species, 3.6%), organic oxygen compounds (7 species, 3.6%), organic sulfuric acid and derivatives (6 species, 3.1%), prenol lipids (6 species, 3.1%), indoles and derivatives (4 species, 2.1%), flavonoids (4 species, 2.1%), and other substances (37 species, 19.1%, Table S1).The volcano plots showed that fatty acyls were the most changed metabolites in the birds fasted for 12 h and glycerophospholipids were the most changed metabolites in the extended fasting (Figure 5). Table S2 shows that 90% of the fatty acyl substances were significantly changed in the birds fasted for 12 h, of which 91.5% were increased and 8.5% were significantly decreased. This effect was gradually lost as the fasting time extended.3.4. Changes in Serum Fatty Acyl Metabolism with Different Durations of FastingThe fatty acyl substances detected in this study mainly consisted of saturated fatty acids (SFA), monoun-saturated fatty acids (MUFA), polyunsaturated fatty acids (PUFA), eicosanoids, and fatty amides (Figure 6A). The level of SFA increased significantly (p < 0.05) in FH12 and FH24 and returned to the non-fasting levels in longer fasting durations (Figure 6B). The serum MUFA increased (p < 0.05) maximally in FH12, and then decreased to the non-fasting state with prolonged fasting duration (Figure 6C); the serum PUFA level gradually decreased (p < 0.05) with the fasting durations extending (Figure 6D). Serum eicosanoid was upregulated in FH12 (p < 0.05) and then returned to the non-fasting level in the prolonged fasting durations (Figure 6E and Table S2). It is worth noting that eicosanoids such as the analog of prostaglandin (PG) and leukotriene were significantly increased in the birds of FH12 when compared with the CT (p < 0.05, Table S2). The serum lipid amide levels significantly increased (p < 0.05) in FH12 and FH24, and then gradually decreased to the non-fasting level in FH72 (Figure 6F).3.5. Changes in Serum Glycerophospholipids Metabolism with Different Durations of FastingAs shown in Figure 7A, the changed serum glycerophospholipids detected in this study were mainly phosphatidylcholine (PC), phosphatidylethanolamine (PE), and their lyso-forms (Lyso-PC and Lyso-PE). The serum PC levels gradually increased with extended fasting durations (p < 0.05, Figure 7B). The serum Lyso-PC increased with prolonged fasting durations, except in FH72 (Figure 7C); both the serum PE and Lyso-PE levels gradually decreased with the fasting durations extending (all p < 0.05, Figure 7D,E).3.6. Changes in Serum Metabolic Pathway with Different Durations of FastingThrough the MBRole pathway enrichment analysis, the serum glycerophospholipid metabolism, glycosylphosphatidylinositol anchor biosynthesis (GPI-AB), autophagy, linoleic acid (LA), arachidonic acid (ADA), phenylalanine metabolism, GnRH signaling pathway, and ferroptosis were identified in the fasted birds when compared with the non-fasted (Figure 8).We further evaluated the effect of fasting on the host serum metabolic pathways. The abundance of one metabolic pathway was the sum of the abundance of the detected metabolites typically found in that pathway. The results showed that the abundance of serum autophagy, GPI-AB, and ferroptosis were increased significantly during food deprivation, regardless of fasting durations (all p < 0.05, Figure 9A,C,H). The serum glycerophospholipid metabolic pathway was decreased in the birds fasted longer than 24 h (p < 0.05, Figure 9B). Both the serum LA and ADA metabolism were increased (all p < 0.05) in FH12 and then decreased with prolonged fasting (Figure 9D,E). The abundance of the phenylalanine metabolic pathway was downregulated in all of the fasting groups compared with the non-fasting birds (p < 0.05, Figure 9F). The GnRH signaling pathway was gradually decreased with prolonged fasting durations (p < 0.05, Figure 9G).3.7. Pearson’s Correlations between Intestinal Autophagy and Serum PE during FastingThe serum autophagy pathway showed a quadratic fitting characteristic with fasting durations (R2 = 0.380, p < 0.001), which was similar to the intestinal autophagy in response to fasting. It implies that the serum autophagy can monitor the intestinal autophagy level. The phosphatidylethanolamine covalent binding to the Atg protein is essential for autophagy initiation, and its abundance positively regulates autophagy. Here, serum metabolome showed that PE (8:0/8:0) and PE (18:3(9Z,12Z,15Z)/P-18:0) were significantly improved during fasting, even fasted for 12 h. Reversely, PE (18:1(9Z)/0:0) and PE (18:2(9Z,12Z)/P-18:1(11Z)) were significantly decreased compared with the control group (Table S2). Correlation analysis showed that PE (18:3(9Z,12Z,15Z)/P-18:0) and PE (8:0/8:0) were positively correlated with Atg7, whereas PE (18:2(9Z,12Z)/P-18:1(11Z)) and PE (18:1(9Z)/0:0) were inversely correlated with Atg7 (Figure 10). According to the correlation coefficient, serum PE (18:3(9Z,12Z,15Z)/P-18:0) is the best biomarker for monitoring intestinal autophagy.4. DiscussionFasting is widely used in the adjuvant therapy of obesity, inflammation, cardiovascular disease, and so on in humans [1]. In this study, through LC–MS, we identified some altered metabolic pathways at different fasting time points and a potential biomarker in the serum that can predict the intestinal autophagy in chicken.Autophagy plays an important role in the maintenance of intestinal homeostasis [30]. Generally, autophagy is maintained at a fundamental level in different tissues [31]. Nutritional stress-induced elevated autophagy will harvest the nutrients or damaged organelles for cell survival [32]. Previous work has shown that adult laying hens subjected to 20 days of fasting exhibit large lysosomal autophagic vacuoles in the intestinal epithelial cells and atrophic intestinal villus [33]. It is widely known that Atg7 is a critical autophagy protein that plays an essential role in intestinal integrity [34]. In this study, the genetic expression of Atg7 showed a quadratic response to extended fasting duration, even though previous data indicated an increase in autophagy upon nutrient deprivation [35,36]. Our results showed that 12 h fasting is enough to activate autophagy, and 24 h of fasting had the highest autophagy process. The reverse shift from its peak could potentially point towards a new energy balance inside the body, because longer fasting reduces heat production in the chicken [37]. Intriguingly, a similar pattern in autophagy was also detected in the serum. Increased autophagy is an efficient metabolic process to alleviate inflammation and cell apoptosis [11,38]. It has been reported that fasting reduced inflammatory bowel disease by modulating the microbiota [39]. So, the increased autophagy may provide new evidence for the inflammation pathology. More so, blood cells are a suitable indicator for monitoring autophagic flux [24]. Given that PE covalently binds to the Atg protein during autophagy [40], correlation analysis was applied to the serum PE and intestinal Atg7. We found serum PE (18:3(9Z,12Z,15Z)/P-18:0) is a potential biomarker for intestinal autophagy assessment.Serum glycerophospholipid was the most detected chemical in the present study. However, the glycerophospholipid metabolism was decreased in the birds fasted for more than 24 h, which is consistent with the results from Maity et al., who reported that starvation decreased glycerophospholipid metabolism in Diporeia [41]. Glycerophospholipid consists of PE, PC, Lyso-PE, and Lyso-PC. It has been reported that Lyso-PC in the plasma is correlated inversely with body mass index [42]. What is more, people refed after food deprivation showed decreased PC and Lyso-PC [43]. Similarly, in the present study, PC and Lyso-PC were increased with fasting. These published data are in agreement with our study. Serum PC is the most abundant phospholipid component in lipoprotein, with the highest levels in the high-density lipoprotein fraction [44]. Increased PC here may imply changes in the lipoprotein metabolism. Serum Lyso-PC, derived from PC by the elimination of one fatty acid by phospholipase A2 and the transesterification by lecithin-cholesterol acyltransferase, accounts for 5–20% of the total phospholipid and participates in many cellular processes [45]. Recently, it has been reported that increased Lyso-PC is associated with suppressed homocysteine thiolactonase activity [45]. Homocysteine thiolactonase may protect against the vascular injury caused by elevated homocysteine levels by preventing its oxidation to the harmful compound, homocysteine thiolactone [46]. So, the increase in Lyso-PC in the present study may indicate that fasting may damage chicken vascular by suppressed homocysteine thiolactonase activity.Phosphatidylethanolamine is the second most abundant phospholipid in mammalian cells [44]. By catalyzing phospholipase A2, PE was broken into Lyso-PE [47]. A study has shown that cells cocultured with Lyso-PE can inhibit serum starvation-induced cell apoptosis [48]. The decreased Lyso-PE at prolonged fasting in this study may partly explain the atrophic intestinal villus during fasting [49]. Phosphatidylethanolamine participates in the synthesis of glycosylphosphatidylinositol-anchored protein [50] and mediates cell ferroptosis [51,52]. Glycosylphosphatidylinositol, a widely distributed glycolipid in eukaryotes, is the membrane protein’s anchor containing enzymes and receptors [53]. Pantetheinase is one of the glycosylphosphatidylinositol-anchored proteins that hydrolyzes pantetheine into cysteamine and pantothenic acid [54], and elevated pantetheinase has been reported in mice fasted for 24 h when compared with non-fasting mice [55]. Pantothenic acid is a substrate in coenzyme A synthesis that is a necessary cofactor in fatty acid oxidation [56]. So, energy supply during energy deficiency may partly be sourced from increased GPI-AB metabolism in the present study. In turn, serum PE negatively related to GPI-AB and ferroptosis may be attributed to more PE flow into organs and tissues during fasting.Lipids are the primary energy source for most tissues when glycogen is used up [57]. It was observed that previously birds fasted for 24 and 48 h showed a decreased concentration of plasma triglyceride and free fatty acid compared with the non-fasted [58]. It has been reported that food withdrawal caused a decrease in overnight fasting serum fatty acid metabolism, including SFA, MUFA, and PUFA in mice [59]. In this study, if we considered the birds fasted for 12 h as the control group, and the results aligned well with the previous work conducted in mice [60]. Furthermore, our results provide new evidence for recent guidelines that endorse non-fasting profiles for blood sampling [22,60,61].Fatty acid-related metabolism pathways such as ADA, LA, and GnRH signaling pathways were detected in this study. Linoleic acid, a precursor of ADA [62], was significantly increased in the birds fasted longer than 12 h and linearly decreased in the birds fasted longer than 12 h. Arachidonic acid can be converted to eicosanoids, such as PG and leukotrienes [62]. In the present study, 12 h and 24 h of fasting increased the serum prostaglandin F2, which is similar to previous work conducted in mice [63,64]. However, long-term fasting leads to decreased prostaglandin D2, E2, and F2. This can be explained by the long-term fasting-induced deficiency of insulin [65]. In summary, serum ADA and LA metabolism are fasting time-dependent.In addition, phenylalanine has been regarded as an energy source during food deprivation [66]. The stable level of phenylalanine in birds fasted for 24 h or longer suggests that the phenylalanine metabolism provides energy within 24 h of fasting.5. ConclusionsThe present study shows that extending the fasting duration from 12 to 72 h influences the serum metabolic pathways, such as the glycerophospholipid metabolism, GPI-AB, autophagy, ADA metabolism, LA metabolism, ferroptosis, phenylalanine metabolism, and GnRH signaling pathway. Our results also suggest that serum PE (18:3(9Z,12Z,15Z)/P-18:0) is a potential biomarker for intestinal autophagy monitoring, and 24 h of fasting activates intestinal autophagy well.

animals : an open access journal from mdpi

10.3390/ani13091560

PMC10177327

Selenium is an essential trace mineral for better performance, meat quality, and health benefits in farm animals. In broiler diets, selenium has been added in inorganic or organic form to maintain the selenium requirement of the chickens. Researchers have studied the effects and applications of both inorganic and organic selenium in broiler nutrition. However, there is a paucity of information pertaining to organic selenium application under coccidiosis conditions. In this study, we evaluated the effect of dietary organic selenium on growth performance, gut health, and tissue selenium concentrations in broiler chickens under coccidiosis conditions. Results indicated that the addition of organic selenized yeast improved the growth performance and enhanced the selenium concentrations in the tissue, regardless of the Eimeria challenge. Therefore, organic selenized yeast supplementation can be recommended to produce selenium-enriched organic broiler chickens.

A total of 252 one-day-old Ross broilers were randomly allocated to one of six treatments in a 2 × 3 factorial arrangement with respective Eimeria challenges (non-infection and infection) and three different selenium (Se) diets. Dietary treatments were as follows: (1) Se un-supplemented control (CON), (2) inorganic Se treatment (SS; 0.3 mg/kg as sodium selenite), and (3) organic Se treatment (SY; 0.3 mg/kg as selenized yeast). Six replicate cages were allocated per treatment. Chickens in the respective Eimeria infection groups were infected with an E. acervulina, E. tenella, and E. maxima oocyst mixture (15,000 oocysts/chicken) on day 16. Growth performance was measured on days 16, 22, and 24. On day 22, intestinal samples were collected from randomly selected chickens to evaluate gut lesion scores, antioxidant enzymes, and tight junction gene expression. Blood, breast, and liver samples were collected to analyze the Se concentrations on day 24. Dietary SY supplementation improved (p < 0.05) the growth performance of the chickens regardless of the Eimeria challenge. Moreover, independent of Eimeria infection, Se supplementation elevated (p < 0.05) the heme oxygenase 1 (HMOX-1) expression in jejunal mucosa at 6 days post-infection (dpi). Duodenal junctional adhesion molecule 2 (JAM-2) expression and jejunal occludin (OCLN) were elevated (p < 0.05) with dietary SY supplementation at 6 dpi. Among Se sources, broiler chickens fed with the SY diet showed higher (p < 0.05) Se concentrations in breast muscle and serum on 8 dpi. These results confirmed the beneficial effects of dietary Se and the efficiency of organic Se compared with inorganic Se for growth improvement and muscle Se enrichment in broiler chickens regardless of coccidiosis infection.

1. IntroductionTrace mineral selenium (Se) has been used in poultry diets for a few decades because of its well-known health benefits. Some of these benefits include maintaining the growth performance, redox potential, reproduction, and immune function of the chickens [1]. In broiler diets, Se is added in inorganic or organic form to maintain the Se requirement of the chickens. Conventionally, sodium selenite (Na2SeO3) and sodium selenate (Na2SeO4) are the most widely used Se sources in broiler diet formulations in the form of inorganic Se. However, organic sources of Se, such as seleno-methionine (C5H11NO2Se), seleno-cysteine (C3H7NO2Se), and Se-enriched yeast, have become popular and many broiler experiments have been conducted due to their higher bioavailability and tissue retention compared to inorganic Se [1]. In addition, a new form of mineral nanoelemental Se was also developed, tested, and showed improved performance in broiler chickens [2].The impact of dietary Se supplementation on broiler chickens varies to a great extent with different factors, such as the environment, Se source, and Se dosage levels. In recent years, broiler chickens were used to evaluate the multifactorial interaction of Se nutrition with the growth performance, immunity, gut health, and antioxidant potentials. The impacts of Se and Se sources on heat stress, high stocking density, and disease challenge conditions in broiler chickens were reported [3,4]. According to Sun et al. [3], organic Se facilitated better performance and immune status in broiler chickens under high stocking density and heat stress challenges. Similarly, the importance of organic Se in mitigating inflammation and oxidative stress under heat stress conditions was observed with commercial layer chickens [5]. Moreover, it was reported that broiler chickens fed organic Se (as Se yeast) had better oxidative stress resistance during Escherichia coli challenge and heat stress conditions [6]. In another study, dietary Se enhanced chicken immunity during the vaccine response to the low-pathogenicity avian influenza virus [7].As an enteric infection, coccidiosis is a major broiler disease condition caused by Eimeria protozoan parasites, resulting in poor nutrient absorption and hindering growth performance via gut epithelial damage [8,9,10]. In recent papers, the annual loss due to coccidiosis was reported to be over USD 14 billion for the poultry industry globally [10,11]. Before the regulatory antibiotic ban in the livestock industry, enteric infection conditions were successfully managed with in-feed antibiotics. Afterward, different dietary antibiotic alternative feed additives were tested, including probiotics, prebiotics [12], organic acids, enzymes [13,14], phytochemicals [15], and amino acids [16], to mitigate the coccidiosis in broiler chickens [10]. Moreover, dietary vitamins and/or trace minerals’ effects on coccidiosis control were documented [17]. El-Maddawy et al. [18] reported the beneficial effects of zinc oxide nanoparticles as anticoccidial agents in broiler chickens. Similarly, the influence of dietary zinc, copper, and manganese on the performance and intestinal health of coccidiosis-afflicted broiler chickens was reported [19]. Previously, we reported the beneficial effect of organic B-Traxim Se in protecting young broiler chickens against necrotic enteritis [20,21,22].In an early study [23], protective immunity against Eimeria tenella was reported with inorganic Se. Nevertheless, the response to dietary Se supplementation and organic Se sources during coccidiosis infection conditions in broiler chickens is scantly documented. With the mounting concern regarding organic poultry products in the market, it is important to understand the broiler responses to dietary organic Se and its tissue accumulation dynamics. The objective of the present study was to understand the effect of dietary organic Se supplementation on performance, gut health, and tissue Se concentrations in coccidiosis-infected broiler chickens.2. Materials and Methods2.1. Chickens and Animal CareDay-old male broiler chicks (Ross 708) were transferred from a local hatchery (Longnecker Hatchery, Elizabethtown, PA, USA) to the experimental farm. Upon arrival, a total of 252 healthy chicks were weighed individually and randomly allocated to six treatments in a 2 × 3 factorial treatment arrangement, maintaining the same body weight (44.03 ± 0.2 g) and weight distribution among treatments and replicates. Each treatment contained six replicate cages with seven chickens each. Up to 14 days of age, chickens were raised in electrically heated Petersime brooder units housed in a temperature-controlled closed-house environment. After 14 days, chickens were moved to experimental grower cages. All the management practices followed the Ross broiler management guide.2.2. Experimental Design, Diets, and TreatmentsThe experiment was performed using a completely randomized design in a factorial arrangement with respective factors being the three different Se diets and the two different Eimeria challenge conditions (non-infection and infection). Dietary treatments included (1) Se un-supplemented control (CON), (2) inorganic Se-supplemented treatment (SS; 0.3 mg/kg as sodium selenite), and (3) organic Se-supplemented treatment (SY; 0.3 mg/kg as selenized yeast, Sel-Plex®, Alltech, Nicholasville, KY, USA). All experimental diets were formulated and manufactured (NC State University feed mill, Raleigh, NC, USA) based on corn and soybean meal to meet or exceed the nutritional requirements and Se was supplemented accordingly (Table 1). Chickens were fed with the non-medicated respective treatment diets until the end of the experimental period. Ad libitum feed and fresh clean water were provided at all times. To obtain the two different Eimeria challenge conditions, chickens were infected with a freshly propagated and sporulated oocyst mixture (15,000 oocyst/mL/chicken; 5000/mL E. acervulina, 5000/mL E. tenella, and 5000/mL E. maxima) through oral gavage, except for the chickens in the uninfected treatments, on day 16 (Figure 1). In order to make the sporulated oocyst mixture, laboratory-maintained and genotypically verified pure Eimeria strains were mixed proportionately after enumeration. Uninfected chickens received phosphate-buffered saline (PBS) as a placebo in place of the sporulated oocyst mixture.2.3. Growth Performance EvaluationAll chickens were weighed on days 16, 22, and 24, and we recorded the pen basis body weights to calculate the average daily gain (ADG). Feed disappearance of the individual cages was recorded to measure the average daily feed intake (ADFI). The feed conversion ratio (FCR) was calculated and corrected for mortalities.2.4. Sample CollectionOn day 22 (6 days post-infection (dpi)), six chickens from each treatment group (1 bird from each replicate) were randomly selected for sample collection, considering this as the peak infection time. After sacrificing chickens by manual cervical dislocation, 15-cm-long mid-duodenal and jejunal samples together with whole ceca were obtained for gut lesion scoring. The remaining duodenum and jejunum tissue samples were scraped aseptically to collect the mucosa using a tissue scraper [9] for antioxidant enzyme and tight junction gene expression analysis. Collected mucosa samples were stored in RNA stabilization solution (RNAlater™ solution, Invitrogen Corporation, Carlsbad, CA, USA) at −20 °C.Blood samples were collected on day 24 (8 dpi) for serum separation (5 chickens/treatment) followed by cervical dislocation. Separated sera from collected blood samples were stored at −20 °C to analyze the Se content. At the same time, breast muscle samples (pectoralis major and minor) and liver samples were collected from the same chickens to analyze the Se concentration.2.5. Gut Lesion ScoringGut samples were evaluated by three independent scientists based on the gut lesion scoring technique, as described previously [24]. Briefly, intestinal samples and ceca were placed on a white background in good light and slit open. Both the unopened serosal area and the opened mucosal surface were examined for lesions. A four-point hedonic scale was used to enumerate the lesion severity.2.6. Se AnalysisCollected serum, breast muscle, and liver samples were frozen (−20 °C) and sent to the Michigan State University Veterinary Diagnostic Laboratory (Veterinary Diagnostic Laboratory, Lansing, MI, USA) for tissue Se analysis. After tissue preparation, the Se content of the samples was analyzed using an Agilent 7900 inductively coupled plasma–mass spectrometer (Agilent Technologies Inc., Santa Clara, CA, USA), as described previously [25]. In brief, an aliquot of each diluted tissue digest and calibration standard was diluted 25-fold with a solution containing 0.5% ethylenediaminetetraacetic acid and Triton X-100, 1% ammonium hydroxide, 2% butanol, and 5ppb of scandium and 7.5 ppb of germanium, rhodium, indium, and bismuth as internal standards. The ICP/MS was tuned to yield a minimum of 7500 cps sensitivity for 1 ppb yttrium (mass 89), less than a 1.0% oxide level as determined by the 156/140 mass ratio, and less than 2.0% double charged ions as determined by the 70/140 mass ratio. Elemental concentrations were calibrated using a 6-point linear curve of the analyte–internal standard response ratio. Standards were from Inorganic Ventures (Inorganic Ventures, Christainsburg, VA, USA). Bovine liver and mussel standards (National Institute of Standards and Technology, Gaithersburg, MD, USA) were used as controls. A second source calibration check standard from Alfa Aesar (Alfa Aesar, Tewksbury, MA, USA) was also used.2.7. Oocyst CountsFrom 6 dpi (day 22) to 8 dpi (day 24), fecal samples from every cage were collected separately to enumerate the oocyst shedding, considering this period as the peak oocyst production period. The collected fecal samples were processed according to the method described previously [26]. Briefly, total feces collected from individual cages were soaked and homogenized with 3 L of water. Two subsamples from each cage were separated into 50 mL tubes for oocyst counting. Subsamples were then subjected to serial dilutions before the enumeration of oocysts for each sample. Three individual scientists counted oocysts microscopically with a McMaster counting chamber (Challex LLC, Park City, UT, USA) using the sodium chloride flotation method [26]. The total number of oocysts shed per chicken was calculated using the following formula:Total oocysts = (oocyst count × dilution factor × fecal sample volume/counting chamber volume)/number of chickens per cage.2.8. Real-Time PCR AnalysisTotal RNA from each tissue sample was extracted and homogenized using TRIzol reagent (Invitrogen, Carlsbad, CA, USA), followed by DNase digestion, as described [12]. The RNA quantity and purity were assessed using a NanoDrop spectrophotometer (NanoDrop One; Thermo Scientific, Wilmington, DE, USA) at 260/280 nm. Extracted RNA was diluted to the same concentration and the synthesis of cDNA was performed using a QuantiTect® Reverse Transcription Kit (Qiagen, Hilden, Germany), as per the manufacturer’s instructions. The cDNA samples were diluted to 1:5, and 5 µL aliquots were used for qRT-PCR amplification. Each sample was analyzed using SYBR Green qPCR Master Mix (PowerTrack, Applied Biosystems, Vilnius, Lithuania) in triplicate, using the Applied Biosystems QuantStudio 3 Real-Time PCR Systems (Life Technologies, Carlsbad, CA, USA). The following PCR conditions were followed: denaturation at 95 °C for 2 min followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min s. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the housekeeping gene for gene expression. For the relative quantification of the gene expression levels, the logarithmic-scaled threshold cycle (Ct) values were used in the 2−∆∆Ct method before calculating the mean and standard error of the mean (SEM) for the references and individual targets.The encoded gene expression levels of tight junction proteins, including junctional adhesion molecule 2 (JAM-2), occludin (OCLN), and zonula occludens 1 (ZO-1), mucin (MUC-2) expression in the intestinal samples, and antioxidant gene expression, such as superoxide dismutase type 1 (SOD-1), catalase (CAT), and heme oxygenase 1 (HMOX-1), were investigated. All oligonucleotide sequences of the forward and reverse primers used in this experiment are listed in Table 2.2.9. Statistical AnalysisData were analyzed as a completely randomized design, using a general linear model procedure of two-way ANOVA in the SPSS software (Version 24; IBM SPSS 2016, Armonk, NY, USA). Eimeria infection and dietary Se were considered as the two main effects. The pen was used as the experimental unit for all growth performance measurements. The selected individual chicken was considered as the replicate unit for other measurements. Mean differences were considered significant at p < 0.05. When treatment effects were significant (p < 0.05), means were separated using Duncan multiple range test procedures.3. Results3.1. Growth PerformanceDietary Se or Eimeria infection did not show any interaction effect (p > 0.05) on the growth performance of broiler chickens from hatching to 24 days of age (Table 3). Regardless of Eimeria infection, dietary Se supplementation improved (p < 0.05) the ADG of broiler chickens from hatching to 22 days of age. Interestingly, chickens fed SY diets showed a higher ADG from the beginning of the experiment up to day 22. However, SS’ impact (p < 0.05) on ADG was shown after the Eimeria infection (Day 16). Se supplementation did not affect (p > 0.05) the ADFI of broiler chickens from day 1 to day 16. Afterward, from day 16 to 22, SS and SY diet-fed chickens showed a lower feed intake (p < 0.05) followed by better feed efficiency (p < 0.05) compared to CON chickens. Regardless of dietary treatment, Eimeria infection reduced (p < 0.05) the ADG and ADFI of broiler chickens from day 16 to day 24. Similarly, the FCR was increased (p < 0.05) in the Eimeria-infected broiler chickens from day 16 to day 22.3.2. Fecal Oocyst SheddingDietary Se and Eimeria infection did not show any interaction effect (p > 0.05) on fecal oocyst shedding. Moreover, dietary Se supplementation (or Se source) did not affect (p > 0.05) the fecal oocyst shedding in Eimeria-infected broiler chickens (Figure 2). Regardless of dietary Se, Eimeria infection significantly increased (p < 0.05) the fecal oocyst shedding from 6 to 8 dpi.3.3. Intestinal Lesion ScoresThere were no interaction effects (p > 0.05) between dietary Se and Eimeria infection for the intestinal lesion scores of broiler chickens at day 22 (6 dpi; Figure 3). For the main effect of Eimeria infection, broiler chickens infected with Eimeria showed increased (p < 0.05) gut lesion scores on day 24 (8 dpi), independent of Se treatments. However, dietary Se did not affect (p > 0.05) the intestinal lesion scores of broiler chickens, although Se supplementation lowered the jejunal and duodenal lesion scores numerically.3.4. Antioxidant Gene ExpressionThe effects of dietary Se and Eimeria infection on the mucosal antioxidant gene expression of broiler chickens are presented in Table 4. There were no interaction effects (p > 0.05) between dietary Se and Eimeria infection on the duodenum and jejunum antioxidant gene expression at 6 dpi. Moreover, as the main effect, dietary Se or Eimeria infection did not change (p > 0.05) the SOD-1 and CAT gene expression levels in the duodenum and jejunum mucosa. The main effect was that the Eimeria infection lowered (p < 0.05) the HMOX-1 gene expression in the duodenum mucosa compared to the uninfected chickens. Regardless of Eimeria infection, Se supplementation (either SS or SY) elevated (p < 0.05) the HMOX-1 expression in the jejunum mucosa.3.5. Tight Junction and Mucin Gene ExpressionThere were no interaction effects (p > 0.05) between dietary Se and Eimeria infection on the duodenum and jejunum tight junction gene expression of broiler chickens at day 22 (6 dpi; Table 5). Regardless of Eimeria infection, SY supplementation elevated (p < 0.05) the JAM-2 expression in the duodenum mucosa of broiler chickens compared to the SS counterpart. Similarly, SY supplementation elevated (p < 0.05) the OCLN expression in the jejunum mucosa of the broiler chickens. Independent of dietary treatments, Eimeria infection reduced (p < 0.05) the OCLN and MUC-2 gene expression in the jejunum mucosa.3.6. Tissue Se ConcentrationNo interaction effects (p > 0.05) were observed between dietary Se and Eimeria infection on Se concentrations in the tissue of broiler chickens at day 24 (8 dpi; Figure 4). Regardless of Eimeria infection, broiler chickens fed the SY diet showed elevated (p < 0.05) Se content in the serum and breast muscle compared to chickens fed other treatment diets on day 24. However, SS-fed chickens showed higher (p < 0.05) Se retention in liver tissue compared to CON- and SY-fed chickens.4. DiscussionThe chicken responses to dietary Se supplementation and different Se sources have been reported under various stress and challenge conditions [3,4,27]. However, there is a lack of knowledge and understanding of broiler responses to dietary organic Se supplementation following coccidiosis infection. In this regard, the present study was conducted to investigate the effect of dietary inorganic and organic Se in broiler diets on growth performance, oocyst shedding, gut health, and tissue Se concentrations in coccidiosis-infected broiler chickens.In this study, Se supplementation and Eimeria infection did not show any significant interaction effects on the growth performance of broiler chickens. Regardless of Se supplementation, mixed Eimeria infection reduced the growth performance of broiler chickens, similarly to previous studies [9,28]. As the main effect, the addition of Se into the diet (both organic and inorganic) improved the growth performance of broiler chickens in this study, which was identical to previous studies that reported the beneficial effects of dietary Se on chicken growth responses [22]. Dietary Se works as an activator and cofactor of key enzyme 5′ deiodinase for triiodo thryonine synthesis, where growth is controlled by the energy and protein assimilation of the animal [29]. Thus, it is postulated that Se supplementation improves the growth performance of chickens through improved protein and energy digestibility. Nevertheless, some other studies did not show significant growth differences in broiler chickens with dietary Se supplementation [30]. The disparity between these different observations may be explained by the initial Se content of the basal diets in the experiments. When the initial Se content in the diet is lower than the requirement for chickens, broiler chickens positively respond to dietary Se supplementation, and therefore studies conducted with sufficient Se in the control diet did not show any growth response in chickens [30]. In this study, we maintained 0.35 mg/kg Se in our Se-supplemented diets and 0.07 mg/kg in the Se-un-supplemented control diets. In our control diet, the Se level was lower than the National Research Council (1994) recommendation (0.1 mg/kg) for optimum growth in broilers. Looking into the dietary Se sources, previous researchers [30,31,32,33] reported no growth performance difference in chickens fed inorganic and organic Se. However, the results of this study showed an improved body weight and ADG in SY-fed broiler chickens compared to the SS-fed chickens up to day 16 of age. Similar to our findings, Upton et al. [34] and Sundu et al. [35] observed improved growth performance when broiler chickens were fed with an organic selenized yeast (Sel-plex)-supplemented diet. The inconsistency of these growth results with other studies may be caused by many factors, including the dietary Se content in basal diets, chicken breed, age, and experimental conditions. The mechanisms by which organic Se improves the growth performance of broiler chickens depend upon higher bioavailability, improved tight junction permeability, and reduced oxidative stress in young broiler chickens [22,36]. Moreover, it is worth noting that the chickens in the CON group showed higher feed intake compared to the chickens in other treatments from days 16 to 22. It is possible that the lower dietary Se may have increased the feed intake and thereby lowered the feed efficiency in broiler chickens. The observed feed intake results agree with a previous report [37] that showed that broiler chickens fed 0.1 mg/kg consumed more feed compared to those fed 0.25 mg/kg Se. The reason for the higher feed intake in low-Se-fed broiler chickens may be due to feather growth and the metabolism of thyroid hormones. As explained by Choct et al. [37], lower dietary Se concentrations reduce feather growth, leading to higher maintenance energy requirements in chickens. Consequently, chickens tend to eat more feed to maintain the energy requirement.To understand the effects of dietary Se supplementation and the source of Se on Eimeria oocyst shedding, we enumerated the fecal oocysts in this study. Eimeria-infected broiler chickens showed significantly higher oocyst shedding regardless of dietary Se, as we observed in our previous studies [9,28,38]. According to our observation, neither Se supplementation nor the Se source affected the fecal oocyst shedding of the broiler chickens. However, Mengistu et al. [39] reported lower oocyst shedding in E. tenella-infected chickens fed with sodium selenite-supplemented diets. Similarly, a previous study in mice [40] observed lower fecal oocyst shedding of E. papillate with a Se-supplemented diet. The possible mechanism of Eimeria inhibition by dietary Se could be explained as a direct inhibitory effect on Eimeria, or indirect suppression via local oxidative burst, altered microbiota, and the blocking of intracellular parasite development [40]. The disparity observed between our study and these previous studies may be explained by the different animal models used, the dose of Se, and/or different Eimeria species with different virulency.When Eimeria penetrates host intestinal epithelial cells to complete its life cycle, it causes visible gross lesions, and it can be used as a tool to assess the severity of coccidiosis infections [12]. Based on the site specificity of the Eimeria spp., we measured the intestinal lesions in the duodenum, jejunum, and caeca, as we used an Eimeria oocyst mixture that contained E. tenella, E. acervulina, and E. maxima. Broiler chickens infected with Eimeria showed higher gut lesions compared to the uninfected chickens in all three measured sites, regardless of dietary treatments. The invasion of Eimeria sporozoites and merozoites into intestinal epithelial cells and their intracellular replication led to local inflammation in the intestinal mucosa, accompanied by the infiltration of various types of leukocytes [41,42,43]. Therefore, our observation is in agreement with the previous literature that has described elevated gut lesions in Eimeria-infected chickens [38,44]. In this study, dietary Se supplementation did not significantly affect the intestinal lesion scores, regardless of Eimeria infection. However, although statistically not significant, jejunum and duodenum lesion scores were numerically lower in chickens fed with Se-supplemented diets compared to those of chickens without Se-supplemented diets. Similar to our results, Georgieva et al. [45] also failed to see a significant sodium selenite impact on the gut lesions of E. acervulina-infected chickens. Nooreh et al. [44] reported that a combination of vitamins E, C, and Se lowered the intestinal lesion scores of chickens infected with mixed Eimeria parasites to conclude that the beneficial effect may have been driven by the vitamins and not the Se. In another study, Mengistu et al. [39] observed lower gut lesion scores in E. tenella-infected chickens fed sodium selenite-supplemented diets compared to Se-un-supplemented control chickens. The reason that there was no significant difference in gut lesion scores could be due to the higher virulency of the Eimeria spp. that we used to challenge in this study.In modern poultry production, genetic selection for rapid growth, higher feed efficiency, and egg production in a confined cage system promotes high oxidative stress conditions, especially in newly hatched chickens [46,47]. In addition, in newly hatched chickens, until they develop adaptive immunity, the occurrence of enteric diseases such as coccidiosis and necrotic enteritis contributes to high levels of oxidative stress, leading to lower antioxidant gene expression in broiler chickens, especially when their maternal immunity starts to wane at around three weeks [9,48]. In this study, we observed lower HMOX-1 gene expression in the duodenal mucosa in Eimeria-infected chickens regardless of dietary treatments. HMOX-1 is a rate-limiting enzyme responsible for catalyzing the reaction that degrades heme to biliverdin, and we observed lower HMOX-1 gene expression in the mucosa and spleen in E. acervulina-infected broiler chickens in our previous study [9], similar to the present study. As an essential element for the antioxidant enzyme system, Se is vital for detoxifying lipid peroxide and reactive oxygen species, which are generated from the oxidative stress response [49]. Although antioxidant gene expression patterns in the mucosa were not reflective of the Se responses, as we expected, elevated HMOX-1 gene expression was observed in the jejunal mucosa regardless of the Eimeria challenge, showing the dietary Se effect on the mucosal antioxidant system. In a recent study [50], dietary supplementation of organic Se increased the serum T-AOC activity but not the levels of GSH-Px, SOD, or MDA activity. Similarly, we did not observe a significant difference in SOD-1 and CAT activity in the duodenum and jejunum mucosa of the Se-supplemented diet-fed broiler chickens in this study. However, previous studies [2,3] reported the positive effect of Se in terms of enhanced antioxidant defense systems in broiler chickens under different stress conditions. The disparity between these and our results may be due to fewer stress responses in our birds, as observed in the SOD-1 and CAT activity.To further understand the impact and interactions of dietary Se and Eimeria infection on broiler gut health, we assessed the tight junction and mucin gene expression in the duodenal and jejunal mucosa. The tight junction is an integral part of the intercellular junctional complex of intestinal epithelial cells that plays a vital role to maintain the paracellular permeability and thereby the gut health of the host animal [28]. In this regard, the comparative expression levels of the tight junctional genes such as JAM-2, OCLN, and ZO-1 were used as biomarkers to assess the gut health of the broiler chickens [42]. The results of the present study showed that chickens challenged with Eimeria demonstrated lower expression of OCLN and MUC-2 genes in the jejunum. Similar to the present results, previous studies from our laboratory [28,42] reported downregulated tight junction gene expression with Eimeria infection, indicating the detrimental effect of intracellular parasitism on the gut health of young broiler chickens. Moreover, our previous study also showed that the mucin (MUC-2) gene expression, which is a key encoded secretory protein for gut barrier protection, was downregulated when chickens were infected with coccidiosis [9]. As the main effect, chickens fed with an organic selenized yeast-supplemented diet showed upregulated JAM-2 and OCLN gene expression compared to their inorganic Se-fed counterparts in the duodenum and jejunum, respectively. These results support the notion that organic selenized yeast promotes chicken gut health by enhancing structural gut integrity and by decreasing intestinal epithelial permeability. According to Yang et al. [51], Se-enriched yeast inhibited the NF-κB signaling pathway and reduced the intestinal tight junction injury caused by ochratoxin A toxin in broilers. Likewise, the results of this study showing the upregulated tight junction gene expression support the important role of organic selenized yeast in improving gut health.According to our data, breast and liver tissue showed a higher Se concentration compared to the serum in all treatment groups, and it is not surprising that body tissue such as the liver accumulates more Se regardless of the Se source or the dietary level [52]. However, the bioavailability of each Se source is a key factor that governs the Se concentrations in host animal tissue. Organic selenized yeast has been reported to contain predominantly selenomethionine (50–70%) and more than 100 other unique Se species [27]. Therefore, selenized yeast shown to provide higher bioavailability compared to inorganic sodium selenite to promote higher Se concentrations in broiler chickens [3,29]. In agreement, organic selenized yeast showed a higher Se concentration in serum and breast meat [52,53]. Nevertheless, sodium selenite-fed chickens showed elevated Se retention in the liver compared to organic selenized yeast-fed chickens. Similar to our results, higher selenium retention in the liver was previously reported by Ibrahim et al. [54]. Presumably, the underlying reason for the higher level of Se in the liver may be caused by the absorption pathways, with lesser bioaccumulation in body muscles. In agreement with this idea, previous reports [29,36,55] confirmed that organic selenized yeast metabolizes and is then absorbed as selenomethionine and selenocysteine via an active mechanism, similar to methionine (neutral amino acid transport system), and deposition occurs in the muscle tissue as a source of methionine instead of Se and thereby lowers the Se concentrations in the liver. With regard to Eimeria infection, the tissue Se concentration was not affected by the infection in this study. Moreover, the reduced stress conditions affirmed by the antioxidant gene expression as discussed above may explain the lack of a significant Se concentration difference in the tissue between Eimeria-infected and uninfected chickens.5. ConclusionsIn conclusion, the present study on the role of dietary Se supplementation in coccidiosis-infected young broiler chickens showed that dietary Se supplementation did not interact with coccidiosis conditions in growth performance, intestinal health, oocyst shedding, and tissue Se concentrations. Regardless of the Eimeria challenge, Se supplementation in broiler diets improved the growth performance of the broiler chickens. Specifically, the addition of organic selenized yeast improved the growth performance and enhanced the Se concentrations in the tissue; therefore, organic Se supplementation is beneficial to produce Se-enriched organic broiler chickens.

animals : an open access journal from mdpi

10.3390/ani12010105

PMC8749769

Dog population estimates are necessary to design effective rabies and dog population control programs. Dog population sizes vary drastically from country to country and vary within a country based on human tolerance, pet ownership practices, culture, religion, and several other factors. Human density, level of urbanisation and human settlement types (urban, semi-urban and rural) also play a role in the size of the dog population. Humane dog management programs have shown that dog density per km street length is one measure to monitor the program’s impact. However, we argue here that efficient sterilisation and vaccination program planning also requires an estimate of the total dog population. In the Philippines, we have conducted owned dog population surveys (household surveys and dog demographic surveys), which have proven to be very effective in planning high-volume vaccination programs. Following the implementation of the dog surveys and the subsequent understanding by local officials that actual dog populations were far higher than originally assumed, a higher rabies vaccination coverage was achieved in two target cities due to a correction in the number of vaccines doses needed.

Understanding dog population dynamics plays a vital role in planning both rabies and dog management interventions. Establishing a human to dog ratio and an understanding how the urban/rural nature of the community might affect the overall dog population estimate provides an easy-to-use reference to estimate approximate dog populations in a range of communities. A total of 10,664 households were interviewed in 10 locations in the Philippines (2017 and 2018) to understand the dog population variations among the urban, semi-urban and rural areas. Epicollect5 and OSM tracker applications were used to conduct household interviews using a predesigned fixed set of questions. All answers were recorded directly using mobile phone applications. The survey results showed that for every 1000 humans, there are 256.3 dogs in rural areas, 213.8 dogs in semi-rural areas, 208.7 dogs in urban areas and 170.0 dogs on small islands of the Philippines. We estimate a total dog population in the Philippines of 23.29 million dogs (CI 95%, 22.51–24.07 million). Based on the survey findings from Quezon City and Cebu City, targets, resources allocations and vaccination approach were adjusted for the anti-rabies vaccination program at two locations in 2018, which lead to a 3- to 4-fold increase in the total number of dogs vaccinated in each city compared to previous years.

1. IntroductionDogs have evolved to be with humans or close to human settlements for survival [1]. Dogs are usually very dependent on human food provision and the relative dog population (e.g., dogs per 1000 people) fluctuates according to the level of human tolerance [2]. In many countries (mostly high income), the relationship between humans and dogs has reached the stage where dogs are part of the family and are mostly very well taken care of. However, there are still large populations of roaming dogs living in harsh conditions on the streets in Asia, Africa, and Latin America. It is usually difficult to distinguish the ownership status of free-roaming dogs; therefore, it is not unusual for owned but partly roaming dogs to be identified as unowned dogs. There is growing evidence that these free-roaming dogs are mostly owned not dependent on garbage, but rather dependent on the direct provision of food by humans [3,4,5].In the last ten to twenty years, the number of papers reporting dog populations (both owned and roaming on the streets) has increased significantly [6,7,8]. These dog census projects have often been conducted to provide estimates of the proportion of dogs that need to be vaccinated against rabies (it is usually assumed that one has to vaccinate 70% of a dog population to eliminate the transmission of rabies among dogs and hence to people).To plan an effective rabies management program, it is vital to have a realistic estimate of the number of dogs present in specific areas of interest. Unfortunately, estimating actual population size with high accuracy can be a resource-intensive and complicated process. For that reason, it is often more feasible to approximate dog population sizes using simpler kinds of metrics. One option is to conduct counts of roaming dogs along transects and use the resulting data to calculate an index of density describing the number of dogs seen per km of the transect. Although this index does not give true population size, it does provide a valuable indicator of the relative density of dogs in different areas or during different periods. Another approach is to use a proxy variable—usually human density, which is typically known—as an indicator of likely dog density. This is often expressed as the number of dogs present per 1000 people. Data to generate this metric can be obtained with transect surveys, household questionnaires or a combination of these methods. Each of these methods has its own strengths and weaknesses and, furthermore, they tend to sample overlapping, but not equivalent segments of the total dog population. If we know what knowledge exists among the people and what practices they follow, it will help planning a more effective program. It also provides an opportunity to program implementor to identify hurdles and find realistic solutions.Several methods to estimate dog population are available, often consisting of a combination of questionnaire surveys and street counts, depending on the dog demographics in a community [9]. Humane Society International (HSI) has conducted numerous dog surveys in Asia and Africa has reported the relative dog population in particular communities, i.e., the number of dogs per 100 or 1000 humans. HSI has also conducted numerous KAP (Knowledge, Attitude and Practice) surveys to understand dog ownership practices and to estimate sterilisation and vaccination program costs. One very simple method of estimating the impact of a sterilisation program on street dog population is the development of one or more index survey routes where the number of dogs observed along the route at set times and times of year are counted and plotted on a graph showing the changes in observed dog numbers and sterilisation status [10]. Taking only free-roaming dogs count and measuring dog density per km is not sufficient for owned roaming dog populations, as the roaming dog density is under the direct control of dog owners.Dogs on the street are at high risk of contracting disease from other dogs as well as creating a risk to the community. Zoonotic diseases, including rabies, are a serious concern for many governments [11]. Free-roaming dogs and rabies transmission are closely linked in many low-income countries and large unmanaged dog populations are a particularly daunting challenge for rabies control [12,13]. Understanding the demography of domestic dogs is essential when planning a dog population management and rabies control program [14,15].The lack of dog population estimates has led to ineffective rabies vaccination programs [16]. The US Centers for Disease Control and Prevention (CDC) developed a tool for vaccination campaign planning that requires dog demographic data to develop an appropriate vaccination program, including data on free-roaming owned, free-roaming unowned and owned confined dogs [17]. Most dogs in the developing world are short-lived and the high turnover rapidly reduces the level of vaccination coverage in a dog population [18]. Several anti-rabies vaccination programs in Asia and Africa have not achieved the required 70% vaccination coverage, the target level of vaccination that is projected to break rabies transmission in a dog population for twelve months [16].In the Philippines, vaccinating dogs against rabies and registering owned dogs with the local government authority is mandatory. Keeping a dog on a leash when in public places is also mandatory by law (Republic Act No. 9482 “Anti-rabies Act of 2007”—a system for the control, prevention of the spread and eventual eradication of human and animal Rabies shall be provided and the need for responsible pet ownership established) [19]. Nevertheless, free-roaming dogs on the street are common and dog-mediated rabies is prevalent across the Philippines [20]. The National Rabies Prevention and Control Manual of Procedures in the Philippines provides guidelines on the estimation of the dog population. The most common estimate when planning rabies control projects has been that there is 1 dog per 10 humans [20]. Rabies is endemic in the Philippines and remains a major public health concern. It has a fatality rate of almost 100% and at least one-third of these deaths occur in children aged 15 years old and below. Nationally, the number of animal bite cases in the country increased by 462%, from 2009 (206,253 bite cases) to 2018 (1,159,711 bite cases) [20]. The confirmed number of positive human rabies cases increased by 13.5% between 2009 and 2018, from the 243 cases reported in 2009 to 276 in 2018 [20]. 2. Materials and MethodsIn preparation for a new strategy of anti-rabies vaccination programs, the Department of Agriculture Regional Field Offices (DARFO) of the Bureau of Animal Industry (BAI) and Humane Society International conducted cross sectional surveys to generate dog population estimates and densities across urban and rural areas. Ten different urban, semi-urban and rural locations were surveyed either based on local governments request or as a part of an evidence-based rabies control program strategy that BAI and HSI agreed to develop. These surveys were conducted at different times and by different team combinations between 2017–2018. A standard method and established protocol were followed during each location’s survey. A short questionnaire to record dog demographics targeting randomly selected households in all locations was developed. However, in Quezon City, a longer questionnaire was added to explore Knowledge, Attitudes and Practices (KAP) in the selected barangays (A small territorial and administrative district forming the most local level of government).2.1. Phone Based Application and Survey QuestionnaireTwo different smartphone applications were used for household surveys: Open Street Map (OSM) Tracker with a layout specifically designed for a shorter version of the questionnaire and Epicollect5 (https://five.epicollect.net/, accessed on 27 December 2021) to conduct the longer version. The shorter version was designed to reach a large number of households and was used at all the survey locations. The longer questionnaire was used only in Quezon City in addition to the shorter version. The short version included an icon-based set of questions for dog-owning households (DOHH) and no dog-owning households (NDOHH). The surveyor selects an icon on the OSM Tracker display to record the answers (Figure 1). Each surveyed household was asked whether or not they owned one or more dogs and in the event of a dog-owning household, the surveyor further asked the sex of the dog, confinement practices followed by the owner, sterilisation and vaccination status and willingness to vaccinate and sterilize. The data in the form of GPX files were uploaded to a specifically designed Access database to produce results from uploaded GPX files. Screen A (Figure 1) was the main screen used to record if participant’s house is DOHH or NDOHH, whereas screen B (Figure 1) appeared upon selecting the ‘details’ icon on the main screen and was used to record each dog’s details in the case of DOHH.Each household approached for the survey was asked the questions below in the shorter version; the questions were in the form of icons on the OSM Tracker. The questions indicated by the icons on the screens are listed below.Figure 1Screenshots of OSM Tracker application with modified layout showing icons-based screen used in shorter version to conduct household surveys at all the locations.Screen A: DOHH—Dog owning household, NDOH—Non dog owning household; Screen B: In the case of DOHH, further, we asked below detail for each owned dog. Gender of dog—Female, MaleConfinement practices followed by owner—Confined all the time (confine yes), not confined at all the time (confine no) and Female dog confined during the heat (confine heat)If the dog was vaccinated against rabies in the last one year and, in case of un-vaccinated dog, surveyor to further asked owner’s willingness for vaccination if it were to be provided free—Vaccinated in the last one year (vacc), willing to vaccinate (vacc OK) and not willing to vaccinate (vacc NO)If the dog was sterilized or not and in case of an unsterilized dog, owner’s willingness to sterilize if it were to be provided free—Sterilized (steril), willing to sterilize (steril OK) and not willing to sterilize (steril NO)It was considered culturally insensitive to ask if the female dog was tethered during the heat (confine heat) and so it was not asked at any surveyed locations.The longer version of the questionnaire (conducted only in Quezon City) collected more detailed information on household demographics, knowledge, attitudes regarding dogs, rabies prevention and dog bite wound care knowledge should a household member be bitten and knowledge concerning the rabies program and human attitudes to rabies and dog management practices in Quezon City. The set of questions was created in the Epicollect5, which allows questions not applicable to the participant to be skipped. For example, if a household does not own a dog, upon selecting NDOHH, the questionnaire skipped the questions related to dog details. To maintain consistency in data collection at each location, surveyors were given two days of indoor (theoretical) and outdoor in situ (practical) survey protocol training before starting household survey. Each survey team consisted of two individuals, either two government employees or one government employee and one Humane Society International staff. For each location, several teams were trained and employed for household surveys.2.2. Survey Design and Household SelectionThe local government provided digital map files for each survey location to identify and locate each barangay via Google Maps. Using the most recent census data, barangays were stratified into human density categories (rural and urban) and an Excel sheet was used to draw a random sample of the barangays for survey by generating unique random numbers in excel using the RAND function. Each randomly selected barangay was marked on Google Maps and shared with the survey team. Table 1 provides a listing location and the sample size for each survey. For example, Cebu City areas were divided into rural and urban areas based on human population density. Rural areas were defined as those with a human population below 5000 humans per square kilometer. Urban areas were defined as those with more than 9000 humans per square kilometer. Thirty rural barangays and fifty urban barangays were classified. The barangays were distributed across the North and South regions, with 46 barangays located in the north and 34 barangays in the south of Cebu City. This gave us four identified categories: North Rural (16 barangays), North Urban (30 barangays), South Rural (14 barangays) and South Urban (20 barangays). Using the free online sample size calculator, Raosoft®, it was determined that 2020 households would need to be surveyed to achieve a confidence level of 95% for the owned dog population survey. The final sample included 4 North Rural barangays, 5 North Urban barangays, 3 South Rural barangays and 6 South Urban barangays that were randomly selected using Microsoft Excel. The household sample size to be surveyed per barangay varied from 40 to 240. This was dependent on the barangay’s population density and the number and spatial distribution of households.Survey teams received a map/area boundary in Google Maps with marker pins indicating where they would conduct the surveys. Households were selected in a stratified random sampling method around the pins. A target number of households were interviewed, walking zigzag in small rectangle or square area around the marker pin, making sure not to reach closer to another marker pin and interviewing every 5th or 10th household either on the right or left side only. Additionally, barangays in Quezon City, where we conducted the KAP survey were selected based on the available bite and rabies statistics for both humans and animals (provided by the City Veterinary Offices). Priority was given to barangays with the highest incidence of confirmed animal rabies cases, followed by dog bite cases admitted to the hospital from the barangays. The sample size for each KAP barangay was 200 households.Across the Philippines, several locations were surveyed (Figure 2) using the same methodology. The sample size was based on the urbanisation of the location and resources available for the survey. 2.3. EthicsThe data collection was launched to generate baseline data for the anti-rabies vaccination programs in the Philippines. Verbal consent was collected from each survey participant. For consistency and accuracy in delivering the consent text, the surveyor read a pre-written text explaining the reason they were asked to participate, the scope of the questionnaire and confirming that no personal information identifying any individual would be collected during the interview. Participants were informed that they can withdraw their consent at any point during and after the interview and that they can skip questions they do not wish to answer. If participants did not agree to participate, the surveyor would record a “no” and the questionnaire automatically ended.3. Results3.1. Dog Demographics and Dog OwnershipHousehold dog ownership rates across all study areas were generally high, varying between a low of 29.2% of households owning dogs on Malapascua Island to a high of 63.16% in Lingayen town. The 10,664 households surveyed across all locations owned 9916 dogs. The average number of dogs per dog owning household (DOHH) was over one in all categories of locations (Figure 3). There are an average 0.89 dogs per household (HH) across the Philippines. There was a slight preference for male dogs, but this preference was not significant (Table 1).In Table 1, estimated dog population for each surveyed location was calculated by extrapolating dogs per household number to total households of surveyed location. A Microsoft Access database (https://www.microsoft.com/en-ww/microsoft-365/access, accessed on 27 December 2021) generated average dogs per household for the location from number of dogs counted during the survey with reference to total households surveyed for the location. 3.2. Sterilisation and Vaccination Status Vaccination levels varied between locations (Figure 4). There were no clear reasons identified why the percentage of dogs vaccinated differed.The percentage of dogs that were sterilized as a proportion of their sex is reported in Figure 5. Higher proportions of male dogs were sterilized compared to female dogs.Across all locations and urbanization levels, dog owners were overwhelmingly willing to have their dogs vaccinated (Table 2). However, around a quarter to a third of owners in urban Quezon City (75.6%) and the islands Cabilao (67.9%) and Pitogo (79.4%) were unwilling to have their dogs vaccinated. The level of willingness to have dogs sterilized varied from around 10% to over 90%. It was not clear why there were such large differences in owner attitudes towards dog sterilisation from one location to another.3.3. Dog Population by Human Density and Type of Human SettlementThere was some decrease in relative dog populations as one moves from rural to urban locations, but the island locations had the lowest relative dog populations, even though human density on the islands was low to medium (Figure 6). Locations combined by the settlement type provide a range of dog densities per 1000 humans (Figure 6). We found 256.2 dogs (n = 1386, CI 95%, 249.6–262.4) per 1000 humans in rural areas, 213.8 dogs (n = 2354, CI 95%, 208.7–219.4) per 1000 humans in semi-urban areas, 208.6 dogs (n = 6224, CI 95%, 203.8–214.2) per 1000 humans in urban areas and 170 dogs (n = 700, CI 95%, 163.7–176.3) per 1000 humans in small islands of the Philippines. This indicates that more privately owned dogs are kept per 1000 people in rural areas compared to urban areas.3.4. Dog Population and Human DensityDog density (dogs per 1000 people) may vary with human density (humans per km2). Table 3 examines this relationship. Of all ten locations we surveyed, human population density data up to barangay level were only available for the three locations we presented in Table 3. We observed fewer dogs per 1000 people as human density increased in each of these three locations: Quezon City, Lingayen Town and Zamboanga rural area. Across our sample of 10,664 households (2086 rural and 8578 urban) we counted 2076 (95% CI: 1987–2165) and 7840 (95% CI: 7666–8014) dogs in rural and urban samples respectively. Extrapolated for the total human population and number of households present in the sample regions, as reported by the World Bank [21], we calculate an estimated average dog density of 221.1 (rural) and 209.2 (urban) dogs per 1000 people. Using the World Bank’s total human population figure (divided into rural and urban totals) across the Philippines as a whole [21], we calculate a rough estimate of 23,295,301 (95% CI, 22,515,832–24,074,770) dogs across the country, following from our rural and urban dog density estimations.3.5. Level of Confinement in Different Types of Human SettlementsThe level of confinement differed significantly between survey locations (X-squared = 1999, df = 9, p-Value < 0.001) and between different settlement types (urban, semi-urban, rural and island: X-squared = 1256.5, df = 3, p-Value < 0.001), appearing to increase with the degree of urbanisation (Figure 7). A chi-squared pairwise comparison (with Bonferroni p-Value correction) of the four grouped settlement categories reveals significant differences in confinement proportions between all settlement categories (p < 0.001 for all pairwise comparisons) with the exception of the semi-urban and rural category comparison (p = 0.0533). This supports the hypothesis of increasing confinement practices with increasing degree of urbanisation.3.6. Knowledge Attitude and Practice (KAP) Survey Finding from Quezon City3.6.1. Human DemographicsIn Quezon City, 1741 households were interviewed using a detailed KAP questionnaire prepared in Epicollect5. The sample consisted of 1026 (63.6%) female and 587 (36.4%) male interviewees. Of these, 938 (58.2%) came from dog owning households and 675 (41.8%) from non-dog owning households. Of the 936 dog owning households, 651 (69.6%) owned one dog, 187 (20%) owned two dogs, 61 (6.5%) owned three dogs, 18 (1.9%) owned four dogs, 11 (1.2%) owned five dogs and 14 (1.5%) owned six or more dogs. Over 90% of interviewees (865) were the main caretaker for the dog(s). The main reasons for owning a dog were Pet/Companionship (49.1%, 645) and protection of the property/crops (49.0%, 645). Another 1.8% (23) said that they owned dogs to breed, and one household (0.1%) had a dog for food/to be eaten. Most of the homes had a fenced-in yard (63.4%).3.6.2. Rabies and Dog BitesRegarding symptoms of rabies in dogs, 96.4% (1552) reported that they have heard about rabies. When asked “Do you think it is possible for you or your family to get rabies?” 82.7% (1332) said yes and 8.8% (142) and 8.5% (136) either said no or did not know, respectively. When asked if the interviewee knew how rabies in dogs could be prevented, 60% (1259) knew that dogs should be vaccinated annually against rabies and another 10% (201) knew that dogs should get an injection but did not know what the injection would be. A small proportion (8%, 179) thought that impounding dogs would be the best method to prevent rabies in dogs.3.6.3. Dog DemographicsDogs were acquired in several ways (Table 4). The most common way of adding a dog to a household was by receiving it from someone (65.9%, 880) followed by being born in the household (11.8%, 158). Therefore, most owners acquired their dogs by “accident”. Less than twenty percent either adopted or bought their dog intentionally. Puppies born at the owner’s home were given mainly as gift to family and friends (43.6%) and only 15.6% were sold (Table 5). About one third of the 578 female dogs (33.2%, 192) in the sample had at least one litter in their life (Table 6). It is unclear, however, how many puppies survived from each litter and grew into adult dogs. Puppies that were born to bought female dogs appear to be more likely to be also sold compared to puppies born to other female dogs (Figure 8).3.7. Application to the National Rabies Vaccination ProgramAfter the survey results were shared internally with the local government veterinary offices, a joint vaccination program was initiated in Cebu City and District-2 of Quezon City in 2018 in coordination with the Bureau of Animal Industry and DARFO-7. In earlier rabies vaccination programs (2013–2017), an average of 25,585 and 20,198 dogs were vaccinated annually in Cebu City and Quezon City, respectively. In 2018, the vaccination program used the survey data to plan and resource (with ample vaccination doses) a new rabies vaccination drive (Figure 9). In Cebu City, the new vaccination drive lasted from February 2018 to end of August 2018; whereas in Quezon City the vaccination drive started later towards the beginning of August 2018 and continued until the first week of December 2018. As a result, numbers of rabies cases reported during 2018 were high since there was no vaccination effort during the first half of the year, yet fewer numbers of cases observed in both places during 2019 indicate the efficacy of the program.In both cities, the rabies incidence in animals and humans has declined (Figure 10 and Figure 11). For 2019, the cities reported zero cases of rabies in humans. If rabies vaccination programs continue to reach the 2018 numbers, rabies cases should continue to decrease in animals and human rabies cases should remain at zero.4. DiscussionOur household surveys across all the locations found that 50% of households own one or more dogs. This was independent of the location’s urbanisation status (rural, semi-urban, or urban). Only on small, isolated islands did we find a lower ownership rate of approximately 40% (Table 1). This suggests that the size of the owned dog population in an area of the Philippines is dependent on the number of dogs owned per dog owning household. This statistic differed by urbanization status: highest in rural areas (2.21 dogs/DOHH) and lowest in urban areas (1.73 dogs/DOHH). Dog owners were found to possess a marginal bias to owning male dogs in preference to female. Confinement practices varied dramatically with level of urbanisation, with the percentage of dog owners willing to allow their dogs to roam the street increasing with decreasing urbanization: 29.3% in urban areas, 54.7% in semi-urban, 60% in rural and 84.5% on the small islands. Often these free roaming owned dogs are mistaken for un-owned street dogs during the planning of a large-scale mass vaccination programs and targeted for catching. However, door to door vaccination programs may simultaneously provide a simpler and more cost-effective means to vaccinate such a population and help educate owners concerning the necessity of annual vaccination. Dog owning data also have the potential to inform projected vaccination goals in planning large scale mass vaccination programs, using dog to household ratios (Table 1) combined with the total household numbers available from human census data.We have also presented data in the form of dogs per 1000 humans for all types of settlements, which may be extrapolated to any area of the Philippines. Further, based on the assumption of sample representativeness to the country as a whole, we used urban and rural dogs per 1000 humans to estimate the owned dog population of the Philippines country. Such total country estimates are necessary for central government agency plans for large scale vaccination programs, especially with respect to securing sufficient vaccine numbers. While more data from other parts of the Philippines are needed to validate our estimate, in the meantime, it serves as a guide for other Asian and African countries to establish similar ratios for each region and, thus, plan effective large scale vaccination programs.The sterilization rate was found to be very low in all surveyed locations; a comparatively higher sterilisation rate was found in male dogs. Owners’ willingness for male dog sterilisation was also found to be higher suggesting that dog owners prefer intact female dogs with the potential to breed. In Quezon City, when we used the longer version of our questionnaire, we found that the gifting of dogs is a common cultural practice: 65.9% of dogs were acquired as gifts (Table 4) and upon asking about the fate of newborn puppies, 43.6% of dog owners gave them away as gifts (Table 5). On the other hand, we found a high willingness for vaccinating owned dogs against rabies in all the surveyed locations. This is encouraging for the prospect of potential large scale rabies vaccination programs.In Quezon City, we found that people are keeping dogs either as a companion (49%) or as a guard (49%), with the majority (69.6%) keeping only a single dog. Almost all the interviewed households (96.4%) knew about rabies disease, but only 60% of them knew about the necessity of annual re-vaccination. This shows that there is broad scope for and benefit in an awareness and education program alongside the annual vaccination program.There has been a shift in rabies and humane dog population management in the past decade. Successful programs now demand more evidence of impact beyond just the number of vaccine doses administered and the number of reported rabies cases in animals [22]. There are long-standing and successful anti-rabies programs in other parts of the world that did not start by determining dog populations. However, in the Philippines, South Asia and Africa, the persistence of rabies begs for a re-evaluation of the approaches used to eliminate the virus.Despite the lingering rabies caseload in the Philippines, the government is committed to eradicating the virus with mass-dog-vaccination programs. The locations in this study were chosen specifically because they were places where local veterinary offices decided their programs needed to be more evidence-based and sought support from Humane Society International. The results of the surveys in this report indicate that there are better approaches to estimating the dog population than the 1 dog per 10 people rule commonly used by municipalities in the Philippines.This study shows that standard estimation of the dog population (using 1 dog per 10 people) undercounts local dog populations by 2–3-fold. As a result, dog vaccination programs carried out in the past by Philippines municipalities never achieved a 70% coverage of dogs in the community. A revised human-dog ratio to estimate local dog populations would help secure higher (and more effective) coverage of rabies vaccination. A single parenteral dog rabies-mass vaccination campaign achieving coverage of at least 70% appears to be sufficient to interrupt rabies transmission to humans for up to 6 years [23].In addition to evidence-based planning, Quezon City and Cebu City used a specifically designed mobile phone application for household surveys and to track their vaccination program. The increased data accuracy and simplified data compilation provide additional information that can be used for monitoring and evaluation purposes.5. ConclusionsDogs are closely associated with humans across the world. Based on this study, we conclude that using a dog per 1000 humans metric would be a helpful approach when determining the needs of community rabies vaccination programs. The metric is easy to understand and easy to use. Furthermore, a breakdown into dog—human ratio for rural, semi-urban, urban and small islands has potential to aid the planning of mass vaccination and dog sterilisation programs across the Philippines. The existence of human rabies cases in Quezon City despite people having knowledge of rabies disease and knowing the risk of rabies transmission demand a better rabies vaccination approach. Implementing the baseline dog surveys in Cebu City and Quezon City ultimately led to plans for a broader vaccination campaign in both urban centers and, due to a correction in the number of vaccines doses needed, a 3- to 4-fold increase in dog vaccination coverage. Nationally, such a survey-driven vaccination program could improve rabies elimination efforts and result in a greater likelihood that the Philippines could achieve rabies-free status by 2030.

animals : an open access journal from mdpi

10.3390/ani11051403

PMC8156665

Inflammatory diseases of the reproductive tract of bitches are a common problem in veterinary practice. Vaginitis, which is an inflammation of the vagina, may have various causes and degrees of severity. The aim of this study was to analyze whether the inflammation of the vagina in dogs causes changes in the parameters of white blood cells (leukogram, which is the percentage of individual types of white blood cells) and red blood cells (erythrogram, changes in the shape of erythrocytes). The obtained results suggest that leukogram and erythrogram analyses may be useful diagnostic tools in veterinary practice.

Vaginitis in female dogs is a problem most veterinarians face in their practice. It manifests as localized inflammation, and its variable etiology and different severities often make diagnosis problematic. The study consisted of comparing blood smears taken from 16 animals: 8 healthy bitches and 8 bitches with confirmed vaginitis. We analyzed the percentage of different types of white blood cells (leukogram) and changes in the shape of red blood cells (erythrogram) in both groups. We observed changes in red blood cell morphology, i.e., a higher percentage of lacrimocytes and schistocytes in female dogs with vaginitis compared to their healthy counterparts. The observed hematological changes may illustrate the severity of inflammation. The analysis of erythrograms showed a significantly higher percentage of lacrimocytes and schistocytes in diseased bitches (1.58 ± 1.19% and 0.13 ± 0.12%) compared to healthy animals (0.58 ± 0.38 and 0.00 ± 0.00, respectively). The obtained results may indicate that the analysis of erythrograms throughout the course of vaginitis in bitches may constitute a diagnostic tool, as opposed to the analysis of leukograms, which is more sensitive when it comes to the systemic inflammatory response of the organism. It seems that simultaneous analysis of erythrograms and leukograms may facilitate the diagnostic process in clinical practice.

1. IntroductionVaginitis is a disease that does not occur frequently, although any practicing veterinarian is likely to encounter it. Vaginitis can occur in any age or breed, and in both intact and spayed bitches. The causes of vaginitis can be very diverse, and include bacterial infection, viral infection (e.g., herpes virus type I–HVC I), fungal infection (however, it is very rare), hyper- or hypoestrogenism, urinary tract infection or urinary incontinence, genital infections such as pyometra, metritis or uterine stump abscess, vaginal trauma, chemical irritation due to urovagina, mechanical irritation caused by foreign bodies or tumors, anatomic abnormalities of the genitourinary system and vaginal atrophy after being spayed [1,2]. We distinguish juvenile vaginitis from adult bitch vaginitis depending on age and sexual maturity. Juvenile vaginitis comprises 40–52% of reported cases of vaginitis. Affected animals usually do not demonstrate systemic involvement, and immaturity of the reproductive and immune system are considered to be its causes [3]. Bacterial infection, which is the most common case of adult bitch vaginitis, is usually caused by the overgrowth of the normal microbiota of the vagina [2]. In our previous study, we reported that Escherichia coli, Staphylococcus pseudintermedius, Streptococcus canis, Enterococcus spp., and Mycoplasma spp. are the most commonly isolated bacteria from both healthy and infected (vaginitis) bitches [2]. Other authors have also shown the presence of Proteus mirabilis and P. vulgaris, Pseudomonas aeruginosa, Haemophilus, and Pasteurella haemolytica [1].Vaginitis may be self-limiting, and treatment, when necessary, includes therapy with an antibiotic (or antibiotics), vaginal cleaning, and/or surgical correction of predisposing abnormalities. If left untreated or if improperly treated, it may lead to subfertility or infertility [1].The analysis of hematological parameters is one of the basic elements in assessing the state of health in both human [4] and veterinary medicine [5]. Changes in hematological parameters can be physiological, may be a result of environmental stress factors, or may accompany various diseases [6,7,8,9,10,11,12,13]. Uyarlar et al. [14] showed that dairy cows with ketosis exhibited a higher number of total leukocytes, neutrophils, and monocytes in comparison to control individuals. Lambert et al. [12] revealed a connection between inflammatory response syndrome in horses and the presence of a toxic neutrophil, both of which were more common in diseased animals than in healthy individuals. In addition, leukocytosis, neutrophilia, and lymphopenia were reported in dogs with pyometra [15]. Stacy et al. [16] revealed that the inflammatory process may result in morphological changes in white blood cells in Asian and African elephants (Elephas maximus and Loxodonta africana). Silva et al. [17] showed the occurrence of erythrocyte deformations (cell membrane scrambling, cell shrinkage, and membrane blebbing) in Swiss mice infected with Salmonella enterica serovar Typhimurium. Christopher et al. [11] found higher percentages of schistocytes, microcytes, keratocytes, and spherocytes in rabbits with different organ and systemic diseases in comparison to healthy individuals. It is known that metabolic syndrome induces changes in the shape of red blood cells in humans [18].To our knowledge, the data regarding the relationship between inflammation in dogs and hematological changes are insufficient. The effect of vaginitis on hematological indices, especially on the morphology of erythrocytes, is still unknown. Thus, the aim of the current study was to investigate if vaginitis results in changes in leukogram and/or erythrogram (erythrocyte morphology) outcomes in bitches.2. Materials and Methods2.1. AnimalsClient-owned dogs were transferred to the Veterinary Clinic of the University Center of Veterinary Medicine JU-AU, University of Agriculture in Cracow, Poland in connection with reproductive problems, estrous monitoring, determination of mating date, or routine gynecological examination of breeding bitches. Eight healthy bitches (4 in proestrus/oestrus, 1 in diestrus and 3 in anestrus) and eight bitches with vaginitis (2 in proestrus and 6 in anestrus) were qualified for this study (mixed breeds, aged from 1 to 7 years). The qualification standards were set according to a protocol described by Golińska et al. [2]. Briefly, qualification was based on a clinical and gynecological examination, which included vaginal cytology. Subsequently, the phase of the estrous cycle was determined according to the guidelines described by Concannon [19]. These examinations were carried out by an experienced veterinarian. The bitches with mucusy, milky-white, yellowish, or greenish vaginal discharge and the presence of numerous neutrophils in a cytological smear were classified as animals with vaginitis. In bitches with vaginitis, owners often observed increased licking of the vulva by the animal and polydipsia/polyuria. Moreover, strands of mucus in the cytological smear were visible. The bitches without vaginitis had physiological vaginal discharge (from blood to straw colored) or no discharge at all. The presence of a few neutrophils in the cytological smear was acceptable in two cases: at the beginning of the proestrus phase (during bleeding) and during the diestrus phase, when the bleeding had ceased. To the best of our knowledge, no additional detailed classifications of canine vaginitis have been published to date. Only bitches in good condition, showing no systemic or organ diseases (excluding symptoms of vaginitis), were selected for this study, and bitches undergoing any treatment were excluded. The animals were not given antibiotics for at least 2 weeks prior to collection. The owners were informed about the purpose of the study and gave their written consent for their dogs to participate in the study.2.2. Sample CollectionThe blood for hematological tests was collected from 8 healthy (control) and 8 diseased bitches from the cephalic vein into EDTA plastic tubes. Analyses were performed using the LaserCyte, Vetlab station (IDEXX, Westbrook, ME, USA) with settings for canine blood after testing for trueness and precision. Next, blood smears were prepared. The smears were stained with a modified Wright Giemsa stain (Hemacolor® staining kit, Merck, Darmstadt, Germany) according to the manufacturer’s instructions. After drying, the smears were analyzed under optical microscopes, Leica DM2500 (Leica Microsystems Inc., Buffalo Grove, IL, USA) and Nikon Eclipse Ci (Nikon Instruments Inc., Melville, NY, USA), for leukogram and erythrogram determination, based on the hematological atlas [20]. The leukogram included differentiation of individual types of leukocytes into: segmented neutrophils, band neutrophils, hypersegmented neutrophils, toxic neutrophils, non-activated lymphocytes, activated lymphocytes, monocytes, eosinophils, and basophils; 100 white blood cells were analyzed each time. The differentiation of erythrocytes into normal (unchanged) and altered (with a changed shape) cells was included in the erythrogram. The following cell types were classified as altered erythrocytes: acanthocytes, echinocytes, elliptocytes, keratocytes, lacrimocytes, schistocytes, and spherocytes, as well as poikilocytes (erythrocytes of irregular shapes that could not be classified into any of the above categories). Each time, 600 red blood cells were analyzed. The altered erythrocytes are shown in Figure 1.2.3. Statistical AnalysisDue to the lack of compliance of the analyzed data with the normal distribution, the data were analyzed statistically using a non-parametric Mann–Whitney U test. The level of significance was set at α = 0.05. The analysis was conducted with the Statistica software. The results of the analyses are presented as percentage values (mean ± SD).3. ResultsThe analysis of leukograms did not reveal statistically significant differences. We failed to detect basophils in any of the studied animals (only in one diseased individual was 1 basophil observed in 100 analyzed cells) (Table 1).The analysis of erythrograms showed a significantly higher percentage of lacrimocytes and schistocytes in diseased bitches (1.58 ± 1.19% and 0.13 ± 0.12%) compared to healthy animals (0.58 ± 0.38 and 0.00 ± 0.00, respectively). No acanthocytes or spherocytes were observed in either of the animals (Table 2).4. DiscussionOur study did not reveal any significant effects of vaginitis on the leukogram in the case of female dogs. It is known that the phase of the estrus cycle has no influence on hematological parameters [21]. Bauer et al. [22] noticed that carrageenan-induced local aseptic moderate inflammation did not lead to a significant systemic inflammatory response in bitches. Despite the fact that our study involved infectious inflammation as opposed to the aseptic inflammation studied by Bauer et al. [22], and the authors themselves strongly emphasized that their empirical study involved only five dogs, the results were comparable. The increase in toxic neutrophil counts in horses with systemic inflammatory response syndrome (SIRS) observed by Lambert et al. [12] may be a result of the release of band cells from the bone marrow, which can occur in acute systemic inflammatory processes [23]. According to Aroch et al. [24], analysis of neutrophil cytoplasmic toxicity can provide useful clinical information and may serve as a good prognostic predictor in dogs with different pathological states. Saleh and Allam [25] observed that Barki ewes diagnosed with pneumonia exhibited higher counts of neutrophils, eosinophils, basophils, and monocytes in comparison with healthy individuals. It was observed that leukocytosis with left shift was more marked in dogs with closed cervix pyometra complex than in dogs with open cervix pyometra complex (with discharge of pus) [15,26].Our study showed that localized inflammation contributed to a significantly increased percentage of lacrimocytes and schistocytes in the erythrogram. Although many authors have indicated changes in the basic hematological parameters, such as RBC count, Hb concentration, and Ht values in the course of different inflammatory diseases in animals [15,17,25], studies concerning changes in the erythrogram are very scarce. Acanthocytosis is often seen in cats with liver disorders [27]. Acanthocytes were also detected in dogs with hemangiosarcoma, glomerulonephritis, as well as disseminated intravascular coagulation [28]. Leptocytes and codocytes were noted in hepatic insufficiency, especially in dogs and cats with portocaval shunts [27,29]. Sepsis leads to decreased red blood cell deformability in rats and mice [30,31] and altered erythrocyte morphology, with the formation of echinocytes, sphero-echinocytes, and spherostomatocytes [32]. These changes are thought to either be a combined result of oxygen free radicals and toxin action, or depletion of ATP in cells [32]. It is also known that the erythrocyte membrane interacts with some inflammatory molecules, which may result in eryptosis phenomena such as cell shrinkage, membrane deformities, and other structural disorders [17,33].5. ConclusionsChanges in both the leukogram (percentages of individual types of white cells) and the erythrogram (changes in the shape of erythrocytes) seem to constitute significant markers of homeostasis disruptions in mammals. The analysis of our results and the results obtained by other authors may suggest that leukograms are sensitive markers of systemic inflammatory response, while erythrograms seem to change both in the localized and systemic inflammatory process. Neither changes in the leukogram nor changes in the morphology of erythrocytes are specific. Thus, further studies are necessary to assess the clinical utility of these markers. The presented studies are preliminary, and we are aware that they are only an introduction to the development of a diagnostic method. To the best of our knowledge, no studies on the relationship between vaginitis and the erythrogram and leukogram have been published so far.

animals : an open access journal from mdpi

10.3390/ani11051329

PMC8148579

The transport and general handling of slaughter animals are associated with a series of events that cause stressful and unfavorable conditions that can compromise animal welfare. All these stressful events start at the farm and end with the death of the animal. In this experiment, we evaluated the effect of two finishing strategies and two contrasting lairage times through the combination of several indicators regarding productivity, physiology, behavior and postmortem variables. Individual temperaments and their impact on welfare and carcass quality were also considered. Animal welfare was not compromised in any diet during the finishing period. Individual temperament had a positive impact on the productivity and on all physiological indicators at different preslaughter stages. For that reason, we consider that it should be given paramount importance when talking about animal welfare. According to our results, with pasture-based animals, without fasting on the farm and after a short time of transportation (3.5 h), a longer preslaughter resting time (15 vs. 3 h) is desirable from the animal welfare perspective. Furthermore, our results suggest that this longer resting period, would also be more convenient from the carcass quality perspective.

The objective of this experiment was to evaluate the effect of two different pasture-based finishing strategies and lairage time on steers welfare in Uruguayan conditions. Sixty Hereford (H) and Braford (B) steers were assigned to two different diets for finishing purposes: (D1) native pasture plus corn grain (1% of live weight) (H n = 15, B n = 15) and (D2) high-quality pasture (H n = 15, B n = 15). The average daily gain was registered every 14 days, and temperaments were individually assessed one week before slaughter by three individual tests: crush score, flight time and exit speed, building a multicriterial temperament index (TIndex). Animals were slaughtered the same day in two groups (50% from D1 and 50% from D2 in each group) after traveling for 3.5 h and staying 15 (long lairage) and 3 h (short lairage) in the lairage pens, respectively. The behaviors were observed during lairage, and physiological indicators were used to assess stress at the farm after transport, after lairage and at slaughter. Bruises incidence and final pH were registered at the abattoir as a means of assessing the overall animal welfare. Calmer animals had higher average daily gains with no differences either between diets or between breeds. Calmer animals also had a lower stress response during all preslaughter stages, regardless of the time in lairage. Transport did not imply psychological stress (cortisol) for any slaughter group, but physical stress was evident after transport in both groups through NEFA and CPK increases. Bruise incidences did not differ between lairage groups. The short lairage group did not have enough time to cope with the environment before slaughter, with the consequent deleterious effects on the carcass pH. Animals from the long lairage group had a higher metabolic response shown through NEFA values, but they had enough time to rest and recover overnight, reaching final pH values lower than 5.8, considered the upper limit of the normal range. According to this experiment, with pasture-based animals without fasting on the farm and after 3.5 h of transportation, a resting period of 15 h in lairage should be better than a 3-h one.

1. IntroductionThe transport and handling of slaughter animals are associated with a series of events that cause stressful and unfavorable conditions that can compromise animal welfare, increase the chance of spreading disease [1,2] and reduce the meat quality [3,4]. All these potentially stressful events start at the farm and end with the death of the animal. They involve physical stress like food deprivation; fatigue due to transport to the abattoir; collision with equipment and psychological stress because of gathering and mixing, lairage and repeated handling, unfamiliarity to the environment and social disturbance because of the disruption of the rearing group [5]. Stress, whether physical or psychological in origin, induces behavioral and physiological changes [5] that can have a significant impact also on the quality of meat via their effects on muscle energy metabolism [6].The significant relationship between preslaughter stress and meat quality has been widely documented [5,7,8,9,10,11,12,13]. Regarding the effects of lairage time on animal welfare and meat quality, controversial experimental results have been reported, depending on the production systems and the general context of the meat production chain. Several authors sustain that the time in lairage brings about several positive benefits and potentially allows cattle to replenish muscle glycogen concentrations, reduce the dehydration of body tissues and carcass weight loss and to rest and recover from the effects of transport [14,15,16,17,18,19]. Other authors have reported that the lairage environment itself may inhibit the ability of cattle to rest or recover from the effects of feed and water restriction [20,21,22,23]. These varying results should be expected, given the multifactorial characters of these traits, leading the different study designs (preloading fasting at the farm, transport distance and transport time and lairage conditions) to produce different outcomes [14].In this context, strict regulations and directives have been issued to promote animal welfare during preslaughter stages and some international bodies, probably, not considering either the mentioned different realities or differences among species (mainly ruminant vs. nonruminants), which recommend that all livestock animals should be slaughtered immediately after their arrival at the abattoir [24,25]. Therefore, in many European countries, it is common to slaughter animals on the day of arrival, whereas, in South America, due to the climatic, geographic and sociocultural conditions, among other factors, is common that animals are slaughtered the day after arrival. In turn, within South America, there are different country sizes, vast differences between countries in livestock transport durations and in the average distances between farms and abattoirs [26,27] and in preslaughter lairage time regulations. In Uruguay, where meat production is mainly based on grasslands, because of national meat safety regulations, animals are more typically slaughtered the day after arrival, reaching 12 h in lairage as the mean [28] and after having traveled relatively short distances from the farms to the packing plants (250 km in average) [29].In this context and looking for a proper lairage duration in Uruguayan conditions, the objective of this experiment was to evaluate the effects of different pasture-based finishing strategies and two contrasting lairage times on steers welfare. The relationships between temperament and variables related to animal welfare were also assessed in the present study.2. Materials and MethodsThis study was run by the National Institute of Agricultural Research at INIA Tacuarembó Research Station, Tacuarembó, Uruguay (Latitude South 32°02′12.4″; Longitude West 57°09′15.2″) over a period of 5 months (through the end of summer, autumn and the beginning of winter). Sixty Hereford and Braford steers 2.5 years old were assigned to the following diets with finishing purposes according to live weight and breed: (D1) rangeland plus corn grain with the grain supplied at 1% of live weight (LW) (Hereford n = 15, Braford n = 15) and (D2) high quality pasture composed mainly of lotus (Lotus corniculatus) with a small proportion of white clover (Trifolium repens) (Hereford n = 15, Braford n = 15). In D1, Paspalum notatum, Botriochloa laguroides, Stipa setigera and Paspalum dilatatum made up 2/3 of the paddock total forage production. The area for each finishing strategy (35 hectares, 1.16 hectares/animal) was divided into two plots by electric fencing and animals alternated plots every 14 days. The system was planned in order to avoid overgrazing.2.1. Field Determinations2.1.1. ProductivityAnimals were weighed early in the morning without previous fasting every 14 days. For D1, amounts of corn grain were adjusted at this time according to LW. The supplement was provided once a day early in the morning (6 a.m.). Animals from both finishing strategies had ad libitum access to water.2.1.2. TemperamentHair whorl position (HWP) was recorded on the first day of the experiment, looking for a correlation with temperament. If the center of the hair whorl was above the top of the eyes, the animal was categorized as “excitable”, “medium” if the center was located at eye level and “calm” if the center was located below the bottom of the eyes [30].Individual temperament was assessed one week before slaughter by 3 individual tests: Crush score (CS), Flight time (FT) and Exit speed (ES): (a) (CS)—the animal behavior is scored while it is in a chute, using a 1 (calm)–5 (combative) scale, adapted from Hearnshaw and Morris [31]. The categories took into account the general state of the animal, including movements of limbs, head and tail, as well as behavioral signs of stress, attributing one of the following scores: (1) animal does not offer resistance, remaining with tail, head and relaxed ears; (2) animal has little limb movement, keeps head up and ears erect; (3) animal has frequent but not vigorous movements of limbs, head, ears and tail; (4) animal offers great resistance, with sudden movements of head and tail, can jump and fall, with audible breathing; (5) paralyzed animal, with muscle tremor (freezing). The measurement was performed after the animal entered the chute. Only the rear (entrance) and front (exit) gates remained closed for the test, without the use of any of the containment structures (side walls, fisheries and coasters). The records were taken 4 s after closing the gates; (b) (FT)—the amount of time (in seconds) it takes an animal to cover a known distance (5 m) immediately after leaving a confinement situation was recorded. A manual stopwatch was used, and registration started when the chute gate was opened, and the animal had the chance to exit. Animals with shorter flight times were considered more excitable; and (c) (ES)—data were obtained through a nominal scale scoring cattle exit gait: 1 (walk), 2 (trot) and 3 (canter). Animals that canter were considered more excitable. A multicriterial temperament index (TIndex) was built from (FT), (CS) and (ES), following Saaty [32]. For that, a matrix was established with the relative importance of the (FT), (ES) and (CS) characteristics to each other, according to our criteria. This matrix was normalized. A standardized ranking of the animals was generated for each of the variables, on a scale from 1 to 100. Then, the index was constructed according to the following equation:TIndex = ∑1jWjdi where “W” is the weight of each of the variables according to the researcher’s criteria applying Analytic Hierarchy Process-AHP [32], and “d” is the value of each normalized record. Considering that (FT) is an objective test, it was assigned a relatively higher ranking in the index, meaning that the higher the TIndex, the calmer the animal.2.1.3. Health StatusPathological event or trauma and the corresponding medical treatments were daily observed and registered throughout the entire experimental period.2.2. Transport and Slaughter PlantAll animals were slaughtered the same day in a commercial abattoir licensed to export meat, following standard animal welfare procedures. Each slaughter group was composed of 50% of animals from D1 and 50% from D2, remaining in pens for 3 (short lairage) and 15 h (long lairage) preslaughter, respectively. Animals from both slaughter groups remained grazing until loading (without fasting on the farm) and transported for 3.5 h in a commercial truck with two compartments, allowing 420 kg/m2 (1–1.2 m2/head) according to the abattoir protocol (based on international recommendations). Steers from different diets within each slaughter group were not mixed either in the truck or at the abattoir. The same truck and driver were used for both journeys. Distance from the farm to the slaughterhouse was 140 km, and the average driving time was 3 and a half h, with 3 stops of 3 to 4 min for animal monitoring. No problems were registered during loading and unloading, being fluid in both groups. After arriving at the abattoir, animals from each diet (n = 15) within each slaughter group were taken to a 37.5-m2 pen with 2 divisions (8 and 7 animals per division). The space allowance in lairage pens was 420 kg/2.5 m2, according to the protocol mentioned above. Animals from the long lairage group waited from 3 p.m. of day 1 to 6 a.m. of day 2 (slaughter day), and those animals from the short lairage group waited in lairage during the morning of day 2 (from 10 a.m. to 1 p.m.), being the first and the last group sacrificed that same day in the abattoir, respectively.2.2.1. Physiological IndicatorsThree blood samples were taken four times from all animals, looking for basal values in welfare indicators and their respective changes, according to the following periods:Time A: before leaving the farm (basal values)Time B: immediately after arriving at the slaughterhouse (transport effect)Time C: after lairage (lairage effect)Time D: during bleeding immediately post slaughter (effect of the last handling procedures)For bleeding, animals were conducted to a portable chute strategically located near the pens of both slaughter groups. One of the three samples was collected into anticoagulant (Becton, Dickinson and company, Franklin Lakes, New Jersey, USA) cooled and immediately sent for hematocrit determination. The other 2 samples were kept cool until they arrived at the laboratory. Serum was extracted following centrifugation at 3000 rpm for 15 min. The serum fractions were frozen and immediately sent for analysis:Sample 1. Hematocrit. It was determined by the micro hematocrit technique at the University Veterinary Faculty in Uruguay. Results are expressed in percentages.Sample 2. Cortisol and Creatine kinase (CPK). Serum samples were assayed in the Nuclear Techniques Laboratory at the University Veterinary Faculty in Uruguay.Cortisol. Method: it was determined by a direct solid-phase radioimmunoassay (RIA) using DPC kits (Diagnostic Product Co., Los Angeles, CA, USA). All samples were determined in the same assay. The RIA had a sensitivity of 8.2 nmol/L (0.91 log nmol/L). The intra-assay coefficients of variation for low (36 nmol/L–1.56 log nmol/L), medium (224 nmol/L–2.35 log nmol/L) and high (427 nmol/L–2.63 log nmol/L) controls were 10%, 6.8% and 4.6%, respectively. Results are expressed in log nmol/L.CPK. Method: CK NAC liquid UV. Liquid test for creatine kinase determination (EC 2.7.3.2.) activated by NAC and measured by spectrophotometry at 340 nm. Results are expressed in U/L.Sample 3. Non-esterified fatty acids (NEFA) and ß-hidroxibutirate (βHB). Serum samples were assayed at the Rubino Laboratory (Ministry of Agricultural affairs) in Uruguay.NEFA. Method: ACS-ACOD (acil-CoA sintetasa-acil-CoA oxidasa). WAKO laboratory kits (WAKO Chemicals, Richmond, VA, USA) were used (references 999-34691, 995-34791, 991-34891 and 993-35191)—lots TK 365, TK 366, TK 367 and TK 368. This method was adapted for use in a VITALAB Selectra 2 Autoanalyzer (Wiener Lab Group, Buenos Aires, Argentina). Results are expressed in nmol/L.βHB. Method: D-3-hidroxybutyrate oxidation into acetoacetate through the 3-hidroxibutirate dehydrogenase enzyme. As a consequence, NAD+ from the reactive is reduced to NADH, and the absorbance changed to 340 nm. RANDOX laboratory kits were used (reference RB 1008)—094293 in a VITALAB Selectra 2 autoanalyzer. Results are expressed in nmol/L.2.2.2. Behavior in Lairage PenCattle behavior was evaluated by 8 trained observers working in pairs, who rotated between divisions each hour to minimize the observer effect. Direct observation was performed within each pen division (experimental unit) combining the instantaneous scan sampling and the behavior sampling techniques [33]. Due to operative restrictions, animals were observed for 1.5 h in the short lairage and 7.5 h in the long lairage group. At each scan, the following behaviors (body postures and activities) were recorded: walking (without rumination-wr), lying (wr), standing (wr), ruminating, drinking water, conflicts (bumps with the head and mounting), positive social behavior and self-grooming. Results from the scan are shown as a percentage of the total time spent on each behavior. Conflicts are considered very important from the welfare perspective, being relevant to record each occurrence. Due to their possible short duration, these events would tend to be missed by scan sampling [33]. Therefore, in this experiment, conflicts (bumps with the head and mounting) were also registered with the behavior sampling technique at each pen division, between 2 scan periods. Each consecutive sample interval took 7.5 min. Animals were individually identified with a number painted on both sides of the body.2.2.3. Carcass Traits Indicators at the Abattoir as a Means of Assessing Overall Animal WelfareBruisingBefore carcasses were dressed, they were visually inspected, recording the number and severity of bruises at the individual level. Severity was scored as major or minor, depending on whether they involved tissue remotion (minor: subcutaneous or no tissue remotion; major: affecting muscle).pHCarcass pH was measured at 24 h post-mortem (pm) at the Longissimus dorsi (LD) between the 12 and 13th ribs, using a pHmeter (Orion 210A; Cole-Parmer, Vernon Hills, IL, USA) with a gel device.2.3. Statistical AnalysisExploratory analyses were performed for all variables using Statgraphics (Statgraphics Technology Inc., The Planes, VA, USA) and SAS packages (SAS Institute Inc., Cary, NC, USA).Productivity. A general linear model (PROC GLM) [34] was used to determine the effects of diet, breed and TIndex on ADG. Initial and final liveweights were included in the model as covariates. Interactions were considered and, if not significant, were removed from the model. ADG means were compared by the least-squares method (PROC LSMEANS) [34].Physiological data. Due to absence of normality, the cortisol and CPK values were normalized by taking a natural logarithm. The effects of diet, breed and TIndex on physiological indicators through time (4 consecutive times) were evaluated through analysis of variance using a mixed model with repeated measures considering the animal as a random effect inside each diet (PROC MIXED) [34]. Initial and final liveweight were included in the model as covariates. As explained above, bleed samples were obtained four consecutive times. To model the correlation between repeated measures for each animal, a general linear mixed model was used (PROC MIXED) [34]. For each physiological indicator, means were compared by the least-squares method (PROC LSMEANS) [34].For ADG and physiological indicators, several covariance structures were tested (variance components (VC), first-order autoregressive structure (AR (1)) and compound symmetry (CS)), in order to fit the best model. Goodness of fit was defined by the lower Akaike’s Information Criteria (AIC) value [35]. After each model was adjusted, robustness was tested, excluding from data standardized, residual values higher than 2 and lower than −2. The model was considered robust when explanatory variables stayed in the model after data filtering and model rerunning.A regression analyses was performed to evaluate TIndex, lairage duration and final liveweight effects on cortisol concentration during slaughter.Behavioral data. Scan sampling technique: Binomial data from each activity was modeled assuming a binary distribution and a logit link function, using the pen division as the subject/experimental unit. A general linear mixed model was used to study the effect of lairage time, diet, breed and temperament, on the frequency of each behavior (PROC GLIMMIX) [34]. Conflicts data from the Behavior technique was modeled, and a log-link function was set, assuming a Gamma distribution. A general linear mixed model was used to study the effect of lairage time, diet, breed and TIndex on conflicts in the first hour in lairage (PROC GLIMMIX) [34]. Hypothesis tests (binomial proportion) were performed to analyze the differences in conflict frequency (number of events per hour) between consecutive and nonconsecutive hours in lairage.Carcass quality. Bruising frequency was compared by the χ2 test (PROC FREQ) [34], and the regression analysis was performed to study the effect of independent variables on the bruising frequency (PROC LOGISTIC) [34] and pH values (PROC REG) [34]. pH means were compared by the least-squares method (PROC LSMEANS) [34].3. Results and DiscussionNeither breed nor HWP had an effect on the evaluated variables. In turn, none of them were associated with temperament. Therefore, breed and HWP discussion is omitted in this paper.3.1. Field Determinations3.1.1. ProductivityADG did not differ between diets (0.63 ± 0.02 in D1 and 0.64 ± 0.02 in D2). The crude protein content of Uruguayan rangeland pastures seems not to be restrictive for animal production [36] covering cattle and sheep maintenance requirements [37], but low ADG, especially in autumn, are usually due to the unbalanced chemical composition of native pasture, with low energy availability for the digestive process [38]. In our experiment, grazing was not restricted in any finishing strategy, crude protein contents were above critical values (9.22% and 22% in D1 and D2, respectively) and energy restrictions in D1 were compensated by the energetic supplementation, providing the animals with adequate daily gains.3.1.2. Temperament and ADGAverage TIndex was not different between diets (61.2 ± 5.6 in D1 and 54.1 ± 5.6 in D2) (p > 0.05). Calmer animals had higher ADG regardless of diet and breed (p < 0.05). These results are consistent with Voisinet et al. [39], who reported that calmer Bos indicus-cross and Bos taurus cattle had higher ADG than steers with excitable temperaments. Barnett et al. [40] and Hemsworth et al. [41] sustained that a fall in the rate of growth is the consequence of a series of acute or chronic responses to human presence and is probably more accentuated in temperamental animals. Regardless of temperament, gentler animals are known to be less susceptible to stress generated by management practices in which the human presence is involved [42], and their productivity is therefore less affected. In this study, all animals had been subjected to good animal husbandry practices prior to and during the test, probably contributed to the satisfactory ADG levels obtained.3.1.3. Health StatusHealth was compromised in two animals from D2 during the experiment, not being related to feed problems in none of the cases. Immediate and effective control measures were applied, with no incidence on ADG of the involved animals. Good physical health is undoubtedly a necessary condition for animal welfare. However, health is more than the absence of disease, and understanding the relationship between health and welfare depends on drawing inferences about subjective feelings such as pain, discomfort and distress [43]. Based on productive and behavioral observations, it was considered that these events did not have a strong negative impact on welfare, and both animals remained in the experiment. No deaths were registered during the experimental period.3.2. Transport and Slaughter Plant3.2.1. Physiological IndicatorsCORTISOLTime A—cortisol at the farm.Cortisol did not differ between diets at the farm (Figure 1).Time B—Transport effect on cortisol.Cortisol did not increase after transportation in any slaughter group. According to several authors, the major factors determining the welfare of cattle during road transport are: vehicle design, stocking density, trailer design and ventilation, driving and handling quality, transport duration, road and environmental conditions [12,44,45,46]. In our experiment, all these factors were standardized, and based on these data, it is assumed that the proper animal handling during all the preslaughter transportation process (including procedures at the farm, during transport and unloading) and the use of suitable equipment and facilities, contributed to our results. Similar results were reported by Ishiwata et al. [47], who did not find differences in plasma cortisol concentration before and after travelling, suggesting that transport had no severe effects on cattle. Fazio et al. [48] suggested that the effects of short-distance road transport on the increase in cortisol levels in cattle, probably depend on preliminary contact with staff during handling. Trunkfield and Broom [49] reported a sharp response in cortisol levels in calves during the first 2 h of transport, mainly due to the loading procedure. These authors suggested that cattle are stressed during the initial period of transportation (on short journeys) and that the degree of stress is greater after long-distance road transport. Villaroel et al. The authors of reference [50] also reported that cortisol was higher after 1 to 2 h of transportation compared to journeys that were less than 1 h or more than 2 h long. After this initial period on short journeys (less than 4 h), animals are thought to become accustomed to the new situation. In our experiment, animals from both slaughter groups showed a good habituation to transport (Figure 1, Time B). In short, on the basis of the comparative response of circulating levels of cortisol before and after transportation, our data do not agree with results that consider transport to be one of the most potent stressors for cattle [51]. That confirms that using best management practices for transportation will contribute to animal welfare, decreasing, and like in this case, avoiding the expected psychological stress response, even with animal coming from extensive conditions.Cortisol concentrations did not differ between slaughter groups at Time B (Figure 1).Time C—Lairage effect on cortisol.Serum cortisol concentrations significantly increased respect to basal values after lairage in both lairage groups (p < 0.05; Figure 1, Time C). All animals were stressed, probably due to the inherent noises and movement of animals and people in the yards during routine handling at the abattoir. It is known that after a stressful event, hematological variables can return to basal levels within 30 min if animals are in their familiar environment [52]. In this study, probably due to the new environment, higher values of cortisol were registered even after 15 h in lairage. Cortisol concentrations did not differ between slaughter groups at Time C (Figure 1).Time D—Preslaughter effect on cortisol.Both slaughter groups also had a considerable preslaughter stress response, increasing to 1.95 and 2.11 log nmol/L in the short and the long lairage group, respectively (p < 0.05; Figure 1, Time D). This is consistent with results from Boissy and Le Neindre [53] and Lay et al. [54], who reported that cortisol levels in response to a stressor could increase up to 1.78–2.30 log nmol/L in cattle. Some authors believe that the increase in cortisol concentrations at Time D, during bleeding, are mainly a response to handling in the race when driving the steers to the stunning box [55] and that this stress depend on length and design of the chute and the quality of the human–animal relationship [11]. It is worth noting that they could also represent the cumulative effects of all stages of the preslaughter handling. Moreover, due to food safety requirements, cattle in Uruguay are washed on their way to the stunning box to remove hide or fleece contaminants, such as excreta and dirt. The process of handling and washing the animals would have elicited a stress response, which could partially explain the cortisol rise in both slaughter groups. Although the distance between washing and stunning is short, it could have been enough to raise HPA axis activity. In addition, the effect of the process of stunning itself cannot be disregarded. Cortisol concentrations did not differ between slaughter groups at Time D (Figure 1).Temperament and CortisolCalmer animals showed lower cortisol concentrations in serum throughout the whole experiment (p < 0.05; Times A, B, C and D), regardless of diet or slaughter group. Figure 2 shows the effect of TIndex on log cortisol values at slaughter, where a relevant rise in cortisol was registered in both slaughter groups (p > 0.05). As it was previously mentioned, Time D could represent the cumulative effects of all stages of the preslaughter handling, implying that temperament and a proper handling are very important at this stage. Our results are consistent with those reported by Curley et al. [56], Café et al. [57] and Burdick et al. [58], who indicated that the functional characteristics of the HPA axis vary with animal temperament and sympathoadrenal-medullary responses can be more intense in excitable animals [57,59,60,61]. Stress may activate the pituitary-adrenocortical system [62], and these hormonal changes may affect cellular metabolic processes [63].Our results provide support to the recognized influence of temperament in modulating the adrenal response of cattle to different stressful situations.Cortisol concentrations did not differ between slaughter groups at Time D.CPKTime A—CPK did not differ between diets at the farm.Time B—Transport effect on CPK.Transportation had a significant effect on CPK (p < 0.05; Table 1; Time B). Absolute values in U/L increased 2 times at Time D, with respect to basal values in both slaughter groups, being consistent to several other studies [18,50,64,65,66].CPK is a muscle-specific enzyme whose activity in the blood is useful for indicating leakage from the muscle as a result of trauma, physical exercise and stress and/or other muscle damage in animal production [67,68,69]. Transportation is a physical demanding factor, because animals have to maintain balance and the contact between animals produces fatigue and bruising, affecting the permeability of the muscle membranes and the liberation of the enzymes into the blood [11,68]. Even if driving is smooth, animals need to make a considerable physical effort during transportation to keep their balance (stability) and posture. Vibration and motion might also have caused stress. The increased activity of the enzyme in this experiment could represent possible trauma during loading, transport and unloading, or it could have increased as a result of behavioral interactions between steers [67,70].Therefore, in our experiment, traveling could have been an accumulation of the nonspecific stress response and the physical effort.CPK concentrations did not differ between slaughter groups at Time B.Time C—Lairage effect on CPK.After lairage, CPK values did not increase in any slaughter group (Table 1, Time C). Similar results were found by Tadich et al. [55] who found higher CPK activity after transport (with 0, 3 and 16 h) but did not find an additional increase during lairage (in different combinations of transport: 0, 3, 16 h and lairage duration: 0, 3, 12, 16 and 24 h). Even considering that an important frequency of conflicts (bumps with the head and mounting) was registered during the first hour in lairage in both slaughter groups (see behavioral analysis), it was apparently not enough to increase serum CPK concentrations.CPK concentrations did not differ between slaughter groups at Time C.Time D—Preslaughter effect on CPK.Preslaughter handling procedures had a significant effect on CPK (p < 0.05; Table 1; Time D). Absolute values in U/L increased four times at Time D, with respect to basal values in both slaughter groups. The higher presence of CPK implies constant muscle movement, both voluntary and those that are controlled by the autonomic nervous system (heart and lungs). Elevated plasma CPK activity is also associated with strenuous or unaccustomed muscular exercise [71]. For this reason, we considered that the stunning process itself could have had a considerable effect on our results (tonic and clonic phases).CPK concentrations did not differ between slaughter groups at Time D.Temperament and CPKCalmer animals had lower CPK values in serum throughout the whole experiment (p < 0.05; Times A, B, C and D). Animals with the most excitable temperament are most susceptible to stress generated by routine handling practices, such as loading and unloading, transport or the new environment in the abattoir [72] with the consequent effects on CPK.NEFATime A—NEFA did not differ between diets at the farm.Time B—Transport effect on NEFA.NEFA concentrations significantly increased respect to basal values after transportation in both slaughter groups (p < 0.05; Table 2; Time B). Similar results were found by Warris et al. [4], who reported that transport of cattle for between 5 and 15 h was associated with increases in blood concentrations of free fatty acids. Changes in these blood metabolites are indicative of body energy reserves mobilization, a mechanism necessary to maintain homeostasis [73]. Fasting and stressful events are typically associated with increased energy demands, and this leads to depletion of energy stores—in particular, liver glycogens and body fat [74]. Free fatty acids may also increase in response to catecholamine release following acute stress [75]. Although, in this experiment, cortisol concentrations did not increase after transport, it is possible that sudden moments of extremely acute stress (like sudden truck movements or vibrations) provoked the activation of the autonomic nervous system with the consequent increase in NEFA, although it was not enough to activate the HPA axis. According to Mellor and Stafford [76], the relatively slow response time of the HPA axis may make it insensitive as a means of discriminating different level of stress within the first few minutes after a noxious stimulus. The physiological changes elicited by the sympathetic adrenomedullary system may be more accurate in assessing the early stages of distress response [77]. In our experiment, physical stress was evident after transport, according to CPK and NEFA concentrations, but the results showed that the situation did not involve the HPA axis activity. The activation of the HPA axis is mainly dependent on the emotional involvement of the animal; stressors do not necessarily activate the HPA system when the animal does not perceive the situation as stressful [78]. Therefore, is not possible to conclude from the results obtained, that animals were suffering during transport. The physiological changes registered in this stage indicate that the adaptive mechanisms were functioning.NEFA concentrations did not differ between slaughter groups at Time B.Time C—Lairage effect on NEFAAfter lairage, NEFA concentrations increased in the long lairage group (p < 0.05; Table 2; Time C), and NEFA values were higher than those of the short lairage group (p < 0.05), suggesting that food deprivation was not long enough to cause a lasting rise in NEFA in the short lairage. These results indicate a greater energy demand to restore homeostasis because of the longer food deprivation [79]. These differences could therefore be explained as a result of fat reserves being mobilized to supply energy requirements, probably with the additional effect of psychological stress due to the new environment, as shown in Figure 1. However, as has been mentioned, HPA axis activity increased but did not differ between groups during lairage.Time D—Preslaughter effect on NEFA.At slaughter, NEFA concentrations did not increase in any slaughter group (Table 2, Time D), but animals from the long lairage had greater NEFA concentration values at slaughter than the short lairage group (p < 0.05). Undoubtedly, the long lairage group presented higher energy demands. However, as has been mentioned, the HPA axis activity did not differ between slaughter groups at Time D, not being possible to infer more suffering in this group. Results from this experiment are consistent with those from Jarvis et al. (1996), who reported higher concentrations of NEFA during bleeding in animals that spent more than 16 h in the abattoir (overnight) when compared to animals that spent 5 h in lairage pens previous to slaughter (0.28 and 0.33 mmol/L, respectively). Cockram and Corley [80] also found that cattle held overnight in lairage had significantly greater plasma-free fatty acid concentrations than those slaughtered on the day of arrival.Temperament and NEFACalmer animals had lower NEFA values throughout the whole experiment (p < 0.05; Time A, B, C and D). Results from all physiological indicators show that the magnitude and quality of the stress response will be greatly affected by individual differences [81] and that the stress response mechanisms are much more active in excitable animals than in their calmer counterparts.Considering that temperament has been validated as a consistent trait that can be easily assessed on a farm [82] by direct observation [83] and due to its positive effect on all physiological indicators at different preslaughter stages, it should be given paramount importance when talking about animal welfare. β-HIDROXIBUTIRATE (βHB)βHB did not differ between slaughter groups at any Time (A, B, C and D). Ketonic bodies, like βHB, are excellent fuel for tissue respiration—in particular, when glucose levels are limited (fasting). However, under these circumstances, these tissues can easily use NEFA energy sources. In the present study, probably fasting was not long enough to cause a strong and clear βHB stress response and to determine differences between slaughter groups.3.2.2. Behavior in Lairage PenAccording to the Scan sampling technique, animals did not drink water during lairage at the abattoir unless they had the opportunity to rehydrate after arrival [84]. It is possible that this behavior was suppressed as a result of unfamiliarity with the new environment, being consistent to several authors who reported that not all animals will drink water [8,85,86], as the priority is to settle down and explore the pen rather than drinking [22,55,87]. However, in the present experiment, hematocrit values at slaughter showed that animals were not dehydrated. If cattle are fully hydrated and fed before transport, it is likely that food deprivation rather than water will be the greater stressor over the initial 24 h, since this is more likely to disrupt rumen function [88]. In this study, animals did not perform any positive social behavior, self-grooming or lie down during the scan. Steers from both slaughter groups spent around 80% of total time in lairage, standing (wr) and ruminating (Figure 3), and no differences were registered in the percentage of time for walking (wr) and conflicts between slaughter groups. The short lairage group had a higher frequency of rumination (Figure 3, p < 0.05). Animals are known to ruminate while resting [89], and time spent ruminating is a direct indicator of animal welfare [90]. However, our results could be better explained by the experimental schedule defined to reach the stipulated preslaughter waiting hours, as mentioned in Section 2.2. Although grazing behavior is affected by various environmental conditions [91], most grazing behavior studies show similarity in daily grazing patterns, with the major grazing period occurring early in the morning and another later in the afternoon, with intermittent grazing occurring throughout other periods of the day and night (baseline ethogram) [92]. In the present experiment, the short lairage group took advantage of the grazing peak of the afternoon on the preslaughter day and kept grazing until dawn (it was loaded at 6 a.m. of the slaughter day). The 15 h group was loaded at 11 a.m. of the preslaughter day, not being able to perform the afternoon grazing peak on that day. Despite the aforementioned differences in the rumination frequency between laughter groups, high rumination frequencies were recorded in the long lairage group, up to the last hour of observation (7.5 h). A similar experiment developed in Uruguay with steers fed on the pasture and comparing 3 vs. 12 h in lairage, registered high frequencies of rumination until the tenth hour in lairage [93]. Results from both experiments suggest that animals did not experience the hunger sensation during the evaluated corresponding periods.In the present experiment, results from the GLIMMIX procedure showed that animals from D2 spent more time ruminating than supplemented steers (D1) (p < 0.05) in both slaughter groups. These results could mainly be explained by the fact that animals from D2 were strictly fed on pasture (without supplementation), probably implying a higher fiber consumption, a slower digesta passage and, therefore, a larger rumination period [14,94].According to the behavior sampling technique, conflicts frequency (bumps with the head plus mounting/hour) did not differ between slaughter groups, but results from the GLIMMIX procedure showed that regardless of the slaughter group, supplemented steers (D1) were more aggressive than those from D2 (p < 0.05). These results are consistent with the higher rumination time registered in animals from D2, in both slaughter groups.When analyzing conflicts frequency during the first hour in lairage, no differences were found between slaughter groups (Figure 4). The frequency of this activity in consecutive hours (in the long lairage group) was therefore compared to conflicts frequency during the first hour. Results from each binomial proportion comparison showed that conflicts frequency in the first hour in pens was significantly higher than the second, third, fourth, fifth, sixth and seventh hours, respectively (Figure 4; p < 0.05). The first hour in pens was a critical adaptation stage for both groups, but animals that remained in pens became calmer afterwards. According to these results, we could have expected the same evolution in conflicts frequency in the short lairage group. The lowest conflicts frequency in the 15 h group (with respect to the first hour) was registered during the 4th and 7th hour (Figure 4, p < 0.05).Both groups were situated in quiet environments far from the unloading facilities, but the long lairage group waited overnight, with greater opportunities to rest. Noise generated by the normal abattoir activity was noticeably higher during the morning and mid-day because of the slaughter procedures. This could have contributed to a higher excitability in the 3-h group, not having the opportunity to rest or to get used to the pens. However, considering that there were no differences between groups in conflicts frequency during the first hour of observation, we consider that lack of resting time was probably the most important reason for these results. Conflict may be beneficial in the long run but will still be unpleasant while it lasts [95], especially considering those animals that did not have enough time to cope with the new situation (3-h group).Temperament and BehaviorTIndex did not have an effect neither on time budget nor on conflicts frequency during the first hour in lairage, suggesting that this first hour was a critical adaptation stage for all animals.3.2.3. Carcass Traits Indicators at the Abattoir as a Means of Assessing Overall Animal WelfareBruisingIncidence of bruising was not significantly affected by lairage time, with 14 bruises registered in the long lairage and 15 in the short lairage group, respectively. These results are not consistent with those of Mc. Nally and Warris [96], who reported higher bruise incidence in carcasses from cattle that remained for longer lairage periods. Results from the third Uruguayan Beef Quality Audit (2013–2015) show that, at the commercial level, 71% of the carcasses in Uruguay had at least one bruise [97], costing the Uruguayan cattle industry 13 million of dollars in lost carcass value, annually, being 37% of the total losses of the meat chain [97]. Bruises are a very good indicator of animal welfare, and when a bruise affects muscle tissue, the affected area is trimmed during postmortem processing, leading to economic losses due to decreased carcass value from reduced carcass yield and, depending on bruise location, potential devaluing of cuts [29,98]. In the present experiment, major bruises affecting the carcass and meat quality were only registered in the long lairage group (1 bruise in two animals) and both steers jumped through the chute while being bled. In spite of not having differences between slaughter groups, 50% of bruises incidence is very relevant, implying that Uruguay must identify causes and stages where bruises are provoked and strengthen corrective capacitation strategies for diminishing its incidence.Temperament and Bruising.TIndex was not related to bruise incidence. The good management practices followed during the whole experiment including the abattoir, could have contributed to these results. These results are not consistent to Barnett et al. [99] who reported that the vigorous avoidance response of cattle with poor temperament in confined areas during handling, transport and preslaughter increases the likelihood of falling and of collision with yard or stock crate structures and, also, with other cattle, increasing the chance of bruising.pHCarcasses from the short lairage group had higher values of final pH (5.83 ± 0.04 vs. 5.68 ± 0.04 in the long lairage group; p < 0.05). It seemed that their excitability without having the opportunity to recover implied a significant depletion of muscle glycogen reserves with a profound effect on pH at 24 h post-mortem. Stressors appear to be additive [100], so that multiple stressors without the opportunity to recover during lairage resulted in a greater elevation of muscle pH. At the commercial level, the last pH measurement taken is one of the most important reference values to measure meat quality and is related to the depletion of glycogen reserves and the release of lactate caused by stressful handling [63]. It is also the most used instrumental indicator in studies that evaluate preslaughter handling, because it takes into account metabolic routes and muscle energy stores [11]. The Uruguayan National Beef Quality Audit estimated that pH higher than 5.8 costs the Uruguayan cattle industry 16.5 million of dollars annually, with 48% total losses of the meat chain [97].The digestive process has a longer lag phase when animals are pasture-based fed [14,94]. In the present experiment, animals from the long lairage, ruminated during the night, thus, glycogen levels after 15 h could probably have been an important component of glucose availability. They probably had the opportunity to rest overnight when the environment of the slaughterhouse was quieter and could also have achieved some control over possible stress-induced energy intake caused by the new environment. In addition, these animals could have restored their muscle reserves from mobilized liver glucose and would have had greater time to restore muscle glycogen from gluconeogenesis during the resting period [14]. A similar experiment developed in Uruguay, with steers fed on pasture and comparing 3 vs. 12 h in lairage after 1.5 h of transport, did not find differences in pH values between lairage groups, but the glycogen content was lower in the short lairage group. This lower glycogen content was not enough to affect quality, but being consistent to the present study, it suggested a higher level of stress in the short lairage group, becoming a warning flag regarding animal welfare and higher risk regarding meat quality [14].In addition of being used as “iceberg” or “key” welfare indicators during meat inspection, as a means of assessing and ensuring overall animal welfare from the farm of origin to the abattoir [11,101,102], is important to emphasize that bruises and the carcass pH values above 5.8, imply 85% of total economic losses in the Uruguayan meat chain [97]. Therefore, being an exporting country, it is mandatory for Uruguay both from an ethical and economic point of view to strengthen corrective actions and educational strategies throughout the entire chain, as well as to develop research initiatives to minimize the incidence of both problems. In this context, and from the present experiment results, the preslaughter lairage time of 15 h seems to be better than the shorter period of 3 h.Temperament and pH.In the present study, TIndex was not related to final pH values, not being consistent with Lensink et al. [72], who indicated that excitable animals may be most susceptible to stress generated by routine handling practices, such as loading and unloading, transport, and the new environment in the abattoir, reducing the muscle glycogen level in vivo [103] because of energy expenditure due to physical exercise or psychological stress, which may, in turn, increase the ultimate pH of muscles [104].4. ConclusionsConsidering the average daily gains, environmental conditions, animal health performance and mortality rate, it is possible to make the preliminary inference that animal welfare was not compromised in any diet during the finishing period. Due to the positive effect of temperament on productivity and on all physiological indicators at different preslaughter stages, it should be given paramount importance when talking about animal welfare. The psychological stress response of transportation may be minimized in 3.5-h travels by using best management practices, even with animals coming from extensive conditions. Increases in energy demands are unavoidable in fasting animals, especially with longer lairage, but adequate conditions and a calm environment may allow cattle to rest and recover while waiting in lairage pens until 15 h, with positive effects on the animal welfare and carcass quality. The emotional involvement or the psychological stress response did not differ between the contrasting lairage times evaluated, but the insufficient resting period from the short lairage in this experiment contributed to glycogen depletion and higher pH values. According to the experiment results, with pasture-based animals not fasting on the farm and after a short time of transportation (3.5 h), a longer preslaughter resting time (15 vs. 3 h) is desirable from the animal welfare perspective. Furthermore, the results suggest that this longer resting period would be also more convenient from the carcass quality perspective, with its consequent positive effects on the meat quality. Based on our results, international organizations should consider different realities and, therefore, contextualized scientific information when writing worldwide regulations or recommendations, as suggested by Costa [14].

animals : an open access journal from mdpi

10.3390/ani13050878

PMC10000046

In this study we addressed the analysis of human breast cancer and canine and feline mammary tumors with regard to the expression, at either gene or protein level, of some molecules that are related to the capacity of an epithelial cell to become mesenchymal (epithelial-to-mesenchymal transition), acquiring higher ability to metastasize. In our samples, some typical markers of this transition were not higher at mRNA levels in tumors than in healthy tissues, indicating that some other markers should be investigated. Instead, at protein levels, some molecules such as vimentin and E-cadherin were indeed associated with higher aggressiveness, being potential useful markers. As already described in the literature, we also demonstrated that feline mammary tumors are close to an aggressive subtype of human breast cancer called triple negative, whereas canine mammary tumors are more similar to the less aggressive subtype of human breast cancer that expresses hormonal receptors.

Epithelial-to-mesenchymal transition (EMT) is a process by which epithelial cells acquire mesenchymal properties. EMT has been closely associated with cancer cell aggressiveness. The aim of this study was to evaluate the mRNA and protein expression of EMT-associated markers in mammary tumors of humans (HBC), dogs (CMT), and cats (FMT). Real-time qPCR for SNAIL, TWIST, and ZEB, and immunohistochemistry for E-cadherin, vimentin, CD44, estrogen receptor (ER), progesterone receptor (PR), ERBB2, Ki-67, cytokeratin (CK) 8/18, CK5/6, and CK14 were performed. Overall, SNAIL, TWIST, and ZEB mRNA was lower in tumors than in healthy tissues. Vimentin was higher in triple-negative HBC (TNBC) and FMTs than in ER+ HBC and CMTs (p < 0.001). Membranous E-cadherin was higher in ER+ than in TNBCs (p < 0.001), whereas cytoplasmic E-cadherin was higher in TNBCs when compared with ER+ HBC (p < 0.001). A negative correlation between membranous and cytoplasmic E-cadherin was found in all three species. Ki-67 was higher in FMTs than in CMTs (p < 0.001), whereas CD44 was higher in CMTs than in FMTs (p < 0.001). These results confirmed a potential role of some markers as indicators of EMT, and suggested similarities between ER+ HBC and CMTs, and between TNBC and FMTs.

1. IntroductionMammary gland cancer is the most common tumor in women [1] and in female dogs [2], and the third most common neoplasia in cats [3]. Human breast cancer (HBC) is classified into four main subtypes according to the expression of estrogen receptor (ER), progesterone receptor (PR), and epidermal growth factor receptor ERBB2, as follows: (i) Luminal A tumors (ER+ and/or PR+, ERBB2-); (ii) Luminal B tumors (ER+ and/or PR+, ERBB2+); (iii) ERBB2-overexpressing tumors (ER-, PR-, ERBB2+); and (iv) triple-negative (ER-, PR-, ERBB2-) breast cancer (TNBC) [4]. TNBCs are typically high-grade carcinomas characterized by an aggressive behavior and a poor prognosis, with high risk of distant metastasis and death [5]. Canine mammary tumors (CMTs) are classified based on morphologic features [6]. Fifty per cent of CMTs are malignant with a 20% risk of metastasis [7]. The majority (80–90%) of feline mammary tumors (FMTs) are characterized by a highly aggressive behavior that leads to rapid progression and distant metastasis development [8,9]. Typically, FMTs lack the expression of ER, PR, and ERBB2, and have been considered a remarkable spontaneous model for TNBC [10,11,12,13,14,15,16]. In all three species, mammary tumors exhibit both inter- and intra-tumor heterogeneity as a consequence of genetic and non-genetic aberrations [17].Over the past 20 years, the investigation of cell differentiation/phenotypic markers has been used in both human and veterinary medicine, primarily to improve our knowledge of the histogenesis of mammary tumors [18]. In the normal human, canine, and feline mammary gland, two cell subpopulations are present: luminal epithelial cells, positive for cytokeratin (CK) 7, CK8, CK18, and CK19; and basal/myoepithelial cells, variably positive for CK5, CK6, CK14, CK17, SMA, calponin, vimentin, and p63 [19]. In HBC, the evaluation of cell differentiation proteins is frequently performed in association with routine diagnostic markers (ER, PR, ERBB2, and Ki-67) to better classify this tumor. The identification of HBC subtypes has a diagnostic, prognostic, and therapeutic value, and is associated with the cell differentiation and epithelial-to-mesenchymal transition (EMT) status of the neoplastic population according to a hierarchical model [20]. EMT is a key event that neoplastic epithelial cells use to acquire a mesenchymal phenotype [21]. As a result, tumor cells obtain the ability to detach from the primary tumor mass, invade the surrounding tissue, migrate throughout the body, and eventually give rise to metastases in distant organs [22]. The classical EMT is characterized by a decreased expression of epithelial markers and a complementary upregulation of mesenchymal markers. Classical EMT transcription factors, namely snail family transcription repressor 1/2 (SNAIL), TWIST, and zinc-finger-enhancer binding protein 1/2 (ZEB) are known to orchestrate EMT by regulating cell adhesion, migration, and invasion, also interacting with different signaling pathways and microRNAs [22,23]. Although this is a well-de-scribed process that promotes metastasis formation, accumulating evidence suggests the existence of an intermediate state called partial EMT or hybrid E/M, whereby both epithelial and mesenchymal markers are co-expressed in cancer cells [23,24,25]. The aim of this study was to investigate the mRNA expression of classical EMT-related transcription factors SNAIL, TWIST, and ZEB in human, canine, and feline mammary tumors. Additionally, we studied the expression of key proteins involved in the EMT process, including E-cadherin and vimentin, and of proteins related to the tumor phenotype, such as ER, PR, ERBB2, Ki-67, cytokeratin (CK) 8/18, CK5/6, CK14, and CD44.2. Materials and Methods2.1. Tissue CollectionHuman samples were collected from the Istituto Oncologico Veneto (IOV, Padua, Italy), whereas canine and feline samples were collected from local veterinary clinics. The human sample collection was approved by the IOV Ethics Committee. All patients or patients’ owners provided informed, written consent to use their samples for this study. Specifically, samples from 5 healthy human mammary gland tissues (MGTs), 5 ER+ HBCs, 5 TNBCs, 4 healthy canine MGTs, 10 canine mammary tumors (CMTs) (5 grade I and 5 grade II), 6 healthy feline MGTs, and 6 grade III FMTs were collected. In this study, to avoid contaminations with other tumor cell subpopulations, we selected only simple tubular carcinomas (STC), which are composed of only one tumor cell subpopulation (luminal epithelial cells) [6]. Healthy MGTs were collected from tumor-bearing patients during the therapeutic/diagnostic surgical procedures, with no additional sampling performed only for the study. Sampling was performed by surgeons. At the time of sampling, most of the tissue was fixed in 4% formaldehyde for histopathology and immunohistochemistry, whereas a peripheral small portion of tumor and normal tissues (approx. 0.5 cm2 each) was collected and preserved in RNALater (Ambion, Austin, TX, USA), according to manufacturer’s instructions. In the lab, before RNA extraction, a small portion of each RNALater-preserved sample was fixed in 4% formaldehyde and embedded in paraffin to check the content of the samples themselves. Four-μm tissue sections were stained with hematoxylin and eosin, and slides were visualized under the microscope to further confirm the presence of healthy tissue in the samples labelled as “healthy” and of tumor tissue in the samples labelled as “tumor”.2.2. RNA Extraction and Real-Time Polymerase Chain ReactionFor gene expression analysis, a small portion of each tissue sample preserved in RNALater was used for RNA extraction using Trizol Reagent (Invitrogen, Carlsbad, CA, USA), following the manufacturer’s protocol. The extracted RNA was treated with RNAse-free DNAse I (New England Biolabs, Ipswich, MA, USA). Five-hundred ng of total RNA from each sample was reverse transcribed using the RevertAid First Strand cDNA Synthesis Kit (Invitrogen). The cDNA was then used as a template for quantitative real-time PCR using the ABI 7500 Real-Time PCR System (Applied Biosystem) to evaluate the mRNA expression of the following EMT-related genes: SNAIL1, SNAIL2, TWIST1, TWIST2, ZEB1, ZEB2. All the samples were tested in triplicate. ACTB was used as a house-keeping gene. The primer sequences are reported in Table 1. The primers were designed using NCBI Primer-BLAST. To examine primer specificity, the dissociation curves of qPCR products were assessed to confirm a single amplification peak. The qPCR reactions were then purified using the ExoSAP-IT PCR product cleanup (Applied Biosystems) and sequenced at the BMR Genomics (Padua, Italy). The sequences were then verified using the NCBI BLAST database. For data analysis for each sample, the ΔΔCt value was calculated and expressed as a relative fold change (2−ΔΔCt), as described in [16]. Real-time PCR efficiency was calculated by performing a dilution series experiment and applying the following formula to the standard curve: efficiency = 10(−1/slope) − 1 [26,27]. Real-time PCR efficiency was between 90 and 100% for all the samples.2.3. ImmunohistochemistryImmunohistochemistry (IHC) was performed on the above-mentioned samples as well as on additional human breast tissue samples from the Division of Anatomic Pathology archive of the University of Padua Hospital, and on additional canine and feline mammary tissue samples from the anatomic pathology archive of the Department of Comparative Biomedicine and Food Science of the University of Padua. Specifically, IHC was per-formed on the following tissue samples: 10 ER+ HBC, 11 TNBCs, 11 CMTs grade I, 11 CMTs grade II, 12 FMTs grade III. Sections (4 μm) were processed with an automatic immunostainer (BenchMark XT, Ventana Medical Systems), as previously described [11]. Briefly, the automated protocol included the following steps: a high-temperature antigen unmasking (CC1 reagent, 60 min), primary antibody incubation (1 h at RT, see below for dilutions), an ultrablock (antibody diluent, 4 min), hematoxylin counterstain (8 min), dehydration, and mounting. Negative controls omitted the primary antibody, whereas adnexa, epidermis, and non-tumor mammary gland, when present, were used as positive controls for CK8/18, CK5/6, CK14, E-cadherin, vimentin, and Ki-67. For ERBB2, an additional technical external positive control was used (ERBB2 3+ HBC), whereas the species-specific cross-reactivity was previously tested in dogs and cats [10,28]. For ER and PR, feline and canine uterus as well as ovary were also stained as positive controls. For CD44, the lymph node was used as positive control. Positive control tissues, typically collected from necropsies, were derived from the same archive as the canine and feline mammary tumor samples. The following antibodies were tested: anti-ER alpha (anti-ERα) (NCL-ER-6F11 1:40, Novocastra in human and feline species—NCL-ER-LH2 1:25, Novocastra in canine species); anti-PR (NCL-PGR-312 1:80, Novocastra in human and feline species); an-ti-ERBB2 (A0485 1:250, Dako in canine and feline species); anti-CK8/18 (NCL-L-5D3 1:30, Novocastra); anti-CK5/6 (D5/16 B4 1:50, Dako); anti-CK14 (NCL-LL 002 1:20, Novocastra); anti-E-cadherin (610182 1:120, BD Biosciences); anti-CD44 (550538 1:100, BD Biosciences); anti-vimentin (M0725 1:150, Dako); and anti-Ki-67 (M7240 1:50, Dako). In the human species, ERBB2 immunolabeling was performed with Bond Oracle HER2 IHC System for BOND-MAX (Leica Biosystems), containing the anti-ERBB2 antibody (clone CB11, ready-to-use). IHC positivity was semi-quantitatively and separately evaluated by ECVP-boarded (V.Z.) and experienced (L.C.) pathologists. Specifically, cytoplasmic and nuclear positivity were measured as a percentage of positive cells for all markers (100 cells per field in 10 high-power fields were counted). ERBB2 was scored as 0, 1+, 2+, and 3+ according to the American Society of Clinical Oncology (ASCO) 2018 recommendations [29] (10% cut-off), with 2+ and 3+ cases considered weakly and strongly positive for complete membrane immunolabeling, respectively. The protein expression of the studied markers was evaluated in the epithelial/luminal component. Additionally, immunolabeling was observed in healthy/hyperplastic adjacent mammary tissue, and in this case normal basal/myoepithelial cells were also evaluated.2.4. Statistical AnalysisStatistical analyses were performed using Prism version 9.3.1 (GraphPad Software, San Diego, CA, USA). To verify mean differences among groups, either the Student’s t-test or the one-way ANOVA with Tukey’s multiple comparison test was used, when values were normally distributed. A Mann–Whitney test or Kruskal–Wallis test were used when values were not normally distributed. Normality was tested using the Shapiro–Wilk test. The Spearman’s rank correlation analysis was used to analyze associations between variables. The level of significance was set at p < 0.05.3. Results3.1. Gene ExpressionWe sought to investigate the mRNA expression of the EMT transcription factors SNAIL, TWIST, and ZEB in mammary tumors compared with healthy tissue. In HBC (Figure 1), SNAIL1 showed a higher mRNA expression in TNBCs when compared with ER+ (p < 0.05). Conversely, the mRNA expression of TWIST1, TWIST2, and ZEB1 in ER+ and TNBCs was significantly lower than in healthy MGTs (p < 0.05). Additionally, TNBCs had a significantly lower mRNA expression of SNAIL2 and ZEB2 when compared with healthy MGTs (p < 0.05).In CMTs (Figure 2), SNAIL1 showed a higher mRNA expression in STC II when compared with healthy MGTs (p < 0.01) and STC I (p < 0.001). The mRNA expression of SNAIL2, ZEB1, and ZEB2 was lower in tumors than healthy MGTs, although not statistically significant.In FMTs (Figure 3), tumors showed a lower mRNA expression of SNAIL1, SNAIL2, TWIST1, TWIST2, ZEB1, and ZEB2 when compared with healthy MGTs, which was significant only for ZEB1 (p < 0.05). 3.2. ImmunohistochemistryNext, we aimed to study the expression of key proteins involved in the EMT process. The expression of the studied markers was evaluated in the tumor epithelial luminal cell population.CD44 and ERBB2 staining was membranous, whereas CK8/18, CK5/6, CK14, and vimentin staining was cytoplasmic. E-cadherin staining was present in either or both membrane and cytoplasm and it was separately evaluated. Ki-67, ER, and PR staining was nuclear. As expected, epithelial luminal cells of healthy MGT in all three species were diffusely positive for CK8/18, membranous E-cadherin, ER, PR, and occasionally positive for CK5/6, CK14, and CD44. The basal/myoepithelial cells of healthy MGT in all three species were diffusely positive for CK5/6, CK14, CD44, and vimentin, and occasionally also positive for ER and PR. Results for the human, canine, and feline mammary tumors are summarized in Table 2, Table S1 and are graphically represented in Figure 4.In HBC (Figure 4A), ER+ tumors had a high protein expression (roughly 100%) of CK8/18, whereas they were negative for basal cytokeratins CK5/6 and CK14. In TNBCs, the protein expression of CK8/18, although fairly heterogeneous, was lower than in ER+ (p < 0.001) and the protein expression of CK5/6 was higher than in ER+ (p < 0.05). In ER+ tumors the protein expression of E-cadherin was predominantly membranous (Figure 5A), whereas in TNBCs E-cadherin protein expression was often lost from the membrane and pre-dominantly cytoplasmic (Figure 5B). Membranous E-cadherin protein expression was higher in ER+ than in TNBCs (p < 0.001), whereas cytoplasmic E-cadherin protein ex-pression was higher in TNBCs when compared with ER+ (p < 0.001) (Figure 4A). Overall, the expression of this protein was quite heterogeneous across the samples. Interestingly, a strong negative correlation between membranous and cytoplasmic E-cadherin protein expression was found in ER+ (r = −1, p < 0.001) (Figure 4B) and in TNBCs (r = −0.9, p < 0.001) (Figure 4C). CD44 protein expression was lower in ER+ (Figure 5C) than in TNBCs (Figure 5D), although not statistically significant. Notably, in TNBCs, a strong positive correlation between CK5/6 and CK14 expression (r = 0.8, p < 0.01), and a moderate positive correlation between CD44 and vimentin (r = 0.6, p = 0.05), were found. All CMTs (Figure 4D) were positive (>1%) for ER and, therefore, classified as ER+. ER protein expression was lower in STC II than in STC I (p < 0.01). The protein expression of E-cadherin was quite heterogeneous across the samples. As in HBC, a strong negative correlation between membranous and cytoplasmic E-cadherin protein expression was found in the CMTs (r = −0.974, p < 0.001) (Figure 4E). In addition, in STC II, a strong positive correlation between CK8/18 and membranous E-cadherin (r = 0.8, p < 0.01) and a strong negative correlation between CK8/18 and cytoplasmic E-cadherin (r = −0.8, p < 0.01) were found. Interestingly, in STC II, Ki-67 expression was positively correlated with CK8/18 (r = 0.7, p < 0.05) and membranous E-cadherin (r = 0.8, p < 0.01) expression, and negatively correlated with cytoplasmic E-cadherin expression (r = −0.7, p < 0.05). All FMTs (Figure 4D) were negative for ER (<1%), PR (<1%), and ERBB2 (either 0 or 1+), and were therefore classified as triple negative. E-cadherin protein expression was quite heterogeneous. As in the HBCs and CMTs, a strong negative correlation between membranous and cytoplasmic E-cadherin protein expression was found (r = −0.984, p < 0.001) (Figure 4F). In addition, a strong negative correlation between CK5/6 and vimentin expression was found (r = 0.8, p < 0.01). CD44 protein expression was higher in the CMTs (Figure 5E) than in the FMTs (p < 0.001) (Figure 5F). Vimentin and Ki-67 protein expression was lower in the CMTs than in the FMTs (p < 0.001) (Figure 6). The expression of the studied markers was not associated with other histopathological features, such as vascular invasion or regional lymph node metastases (data not shown). Moreover, no significant correlations were found between gene and protein expression of the analyzed markers.4. DiscussionIn this study, we investigated the expression of genes and proteins involved in one of the processes thought to play a major role in cancer progression: epithelial-to-mesenchymal transition [22]. EMT is an evolutionally conserved morphogenetic program during which epithelial cells undergo a series of changes allowing them to acquire a mesenchymal phenotype [21]. During classical EMT, epithelial cells lose the expression of tight junction molecules such as membranous E-cadherin and acquire mesenchymal properties such as migration, invasiveness, and elevated resistance to apoptosis. Transcription factors like SNAIL, TWIST, and ZEB regulate this process and are activated by a variety of signaling pathways, including TGF-α, Notch, and Wnt/β-catenin [30,31,32,33]. SNAIL is a classical regulator of EMT that represses E-cadherin transcription in both mouse and human cell lines [34]. In HBC, it has been associated with tumor recurrence and metastasis [35], and with poor patient prognosis [36]. In contrast to the findings of other authors [37], we found that the mRNA expression of SNAIL2 was significantly lower in TNBCs than in healthy MGTs. In CMTs, SNAIL1 expression was higher in STC II when compared with healthy MGTs and STC I, indicating a possible association of EMT with a higher aggressiveness of these tumors. SNAIL2 in CMTs did not show any difference between healthy MGT and tumor tissue, confirming what other authors have also found [38,39,40]. Conversely, in FMTs, there was a trend such that STC III had a lower mRNA expression of SNAIL1 and SNAIL2 when compared with healthy MGTs. To the best of our knowledge, SNAIL has never been investigated in feline tumors. It is believed that TWIST plays an essential role in cancer metastasis [33]. In HBCs and FMTs, the mRNA expression of TWIST1 and TWIST2 was lower in tumors than in healthy MGTs, which differs from what some authors have found in HBC [41], but is similar to what other authors have found in HBC [42] and in FMTs [43]. ZEB1 has been implicated in carcinogenesis in breast tissue [44] because it enhances tumor cell migration and invasion [45]. In our samples, ZEB1 mRNA expression was lower in tumor than in healthy MGTs, as previously reported by other authors in HBC [42]. Although one study examined the expression of ZEB1 and ZEB2 in five canine mammary carcinoma cell lines [46], to the best of our knowledge, ZEB mRNA expression has never been studied in CMT and FMT tissues. Overall, our data suggest that these transcriptional factors are often downregulated in tumors compared with healthy MGTs, except for SNAIL1 in TNBCs and in CMTs STC II. The RNA isolated from healthy tissues came from the whole mammary gland, which is composed of different cell populations, namely epithelial cells, connective tissue, and fat. Although these transcription factors are barely detectable in normal mesenchymal cells of adult tissues [47], adipose tissue expresses these genes variably [48]. As a result, the mRNA levels of these genes in healthy samples can be dramatically influenced by the presence of non-mammary gland tissues, such as fat.Moreover, it is possible that the number of cells undergoing classical EMT is low when compared with the tumor bulk, which is known to be characterized by a remarkable intra-tumor heterogeneity [22]. Furthermore, some authors believe that these genes are regulated post-transcriptionally [35,49,50,51]. Furthermore, accumulating evidence suggests the existence of cell populations with a hybrid E/M state, which exhibit increased plasticity and metastatic potential, characterized by the co-expression of epithelial and mesenchymal markers [23,24,25,52]. However, the expression of some of these markers may be associated with a complete EMT status, whereas others may be associated with a partial EMT status. For example, it is believed that SNAIL1 is a stronger inducer of complete EMT than SNAIL2, which is rather associated with a hybrid E/M state [53,54]. This suggests that the choice of the markers to be analyzed is fundamental and may help in identifying intermediate EMT states more precisely. In addition, in order to study the EMT process, it would be interesting in the future to investigate the expression of these markers at a single cell level, using single-cell omics approaches such as Laser Capture Microdissection or single-cell RNA sequencing.In the present study, we also assessed the protein expression of several phenotypic as well as EMT-related markers, such as ER, PR, ERBB2, CK8/18, CK5/6, CK14, E-cadherin, CD44, vimentin, and Ki-67, in a subset of HBCs, CMTs, and FMTs.The HBC ER+ samples showed a high expression of luminal CK8/18, and a negative expression of basal CK5/6 and CK14, confirming the strong association between ER+ tumors and highly differentiated glandular cells (CK8/18+), as well as null expression of basal CKs (CK5/6, CK14). In the TNBCs, the protein expression of CK8/18 was highly heterogeneous, whereas the expression of CK5/6 and CK14 was low in most of the samples. This result, in concordance with another study [55], supports the idea that the terms “basal-like cancer” and “triple-negative breast cancer” are not interchangeable. Indeed, only a small percentage of TNBCs are basal-like [56]. The CMTs were positive for ER, whereas the FMTs were negative for ER, PR, and ERBB2. Despite only a few samples being analyzed, these data suggest, as already proposed by other authors [11,57], a similarity between CMTs and HBC ER+ and between FMTs and TNBCs. In CMTs and FMTs, the protein expression of CK8/18, CK5/6, and CK14 was highly heterogeneous, confirming the high inter- and intra-tumor heterogeneity [16,57]. Basal CK14 protein expression was higher in FMTs than in CMTs, confirming that FMTs are more “basal-like” when compared with CMTs [11,12].E-cadherin is a cellular adhesion molecule, and its disruption may contribute to the enhanced migration and proliferation of tumor cells, leading to invasion and metastasis [58,59,60,61,62]. In our samples, E-cadherin protein expression was evaluated in the membrane and in the cytoplasm of tumor cells, separately. Overall, the expression of E-cadherin was highly heterogeneous across the samples of the three species, confirming once more the high inter-tumor heterogeneity of mammary cancer in the three species. In human ER+ tumors, E-cadherin protein expression was predominantly membranous, whereas in TNBCs it was predominantly cytoplasmic, confirming that the delocalization of the protein is associated with increased tumor aggressiveness [56,63]. These results confirm that it is not only the loss of E-cadherin that correlates with increased tumor aggressiveness, but also the protein translocation from the membrane to the cytoplasm, as already described [64,65,66,67].Together with E-cadherin, CD44 has been extensively studied in tumor cell differentiation, invasion, and metastasis, and is thought to be involved in the EMT process in HBC [68,69]. Although a few studies on HBC have shown that protein overexpression of CD44 is associated with poor prognosis and metastasis [70], others have shown that downreg-ulation of its expression is correlated with an adverse outcome [68,71]. For this reason, the role of CD44 in the behavior and prognosis of HBC is controversial [71,72]. In our study, CD44 expression was heterogeneous and lower overall in ER+ tumors compared with TNBCs. This trend agrees with study findings by Klingbeil and collaborators, who found high levels of CD44 expression in tumors with a basal-like or triple-negative phenotype, suggesting an association of this protein with an aggressive phenotype in HBC [73]. CD44 was highly expressed (roughly 85%) in our CMT samples, regardless of the tumor grading, as well as in the healthy mammary gland tissues. Moreover, other authors found no differences between benign CMTs, malignant CMTs, and normal mammary gland tissues, suggesting that CD44 is not associated with aggressiveness in canine mammary tumors [74,75,76,77,78]. In FMTs, the expression of CD44 was low overall (approximately 5%). Sarli and collaborators evaluated the intramammary/intratumoral and extramammary/extratumoral expression of CD44 in feline normal mammary tissues, benign tumors, and malignant tumors in relationship to lymphangiogenesis [79]. They found that CD44 had a significantly higher expression in intramammary/intratumor areas compared with extramammary/extratumor areas in both benign and malignant tumors. Additionally, no statistically significant differences in CD44 expression between normal mammary gland, benign tumors, and malignant tumors were found. To the best of our knowledge, no other studies on CD44 expression in FMT tissues are present within the literature. These data, together with our findings, suggest that CD44 is not a useful marker of malignancy in cats.Another protein that is well-studied and plays a central role in the EMT process, and therefore in tumor invasion and metastasis, is vimentin [51]. Vimentin is one of the major intermediate filament proteins and is ubiquitously expressed in normal mesenchymal cells [80]. Recent studies have reported that vimentin knockdown causes a decrease in genes linked to HBC metastasis, such as the receptor tyrosine kinase Axl [81]. In our study, we also evaluated the expression of vimentin in HBCs, CMTs, and FMTs. We found a higher expression of vimentin in TNBCs compared with ER+, although not statistically significant. This result suggests that vimentin expression is associated with the triple-negative subtype, aggressive behavior, and a poor prognosis of HBC, as previously reported by many authors [82,83,84,85]. In CMTs, vimentin expression is low (approximately 15%), con-firming the low aggressiveness of mammary tumors in dogs, which is in concordance with the findings of other authors [86]. Conversely, in FMTs, the expression of vimentin, although heterogeneous, was quite high (approximately 70%), suggesting the high aggressiveness of mammary tumors in this species [9], as well as their similarities with TNBCs [11].Unfortunately, as a limitation of this study, only grade I and II CMTs were included. No RNALater-sampled canine tumors were diagnosed as grade III. For possible IHC analyses in our archive of paraffin-embedded tissues, a very limited number of grade III simple CMTs were found (14 cases over five years) that were often already vascular/lymph node invasive (10/14). This study would not benefit much from adding only IHC analysis of grade III CMTs that already have invaded the vascular system or with metastases. We still believe that the study allowed the collection of some new data on the most frequent FMTs and CMTs in comparison with HBC samples assessing both gene and protein expression. 5. ConclusionsIn summary, this study showed that most of the classical EMT-related transcription factors SNAIL, TWIST, and ZEB are downregulated in tumor tissues compared with healthy tissues, although additional analyses should be performed to better investigate them in neoplastic clones and in a larger set of samples. IHC analyses indicated a potential role of some markers, namely vimentin and E-cadherin, but not of others (i.e., CD44) as indicators of EMT (including loss of cell differentiation and increased malignancies). Moreover, all the IHC data seem to support the already proposed similarities between FMTs (grade III) and TNBCs, as well as between CMTs (grade I and II) and ER+ HBCs. The two species are widely discussed as potential spontaneous models of specific HBC subtypes [11,12,15,16,57,87,88,89,90].

animals : an open access journal from mdpi

10.3390/ani11102920

PMC8532946

Anesthesia is often required for the medical management of large tortoise species, but little has been published regarding effective anesthetic regimens for these species. The purpose of this study was to review anesthetic regimens that have been used safely and effectively in Galapagos (Chelonoidis nigra), Aldabra (Aldabrachelys gigantea), and African spurred (Centrochelys sulcata) tortoises, with the aim of improving medical management.

Tortoises belong to the taxonomic family Testudinidae, which is considered one of the most imperiled families of the order Testudines. Anesthesia is often required for the medical and surgical management of large tortoises. The objectives of this retrospective study were to review drug regimens used to successfully anesthetize Galapagos (Chelonoidis nigra), Aldabra (Aldabrachelys gigantea) and African spurred (Centrochelys sulcata) tortoises, and to compare the times to effect and to extubation in tortoises administered different premedication protocols. Anesthetic records of giant tortoises admitted to the University of Florida College of Veterinary Medicine between January 2009 and December 2019 were reviewed. A total of 34 tortoises (six Aldabra, 23 Galapagos, and five African spurred) were included, resulting in 64 anesthetic events. Frequently used premedication protocols included an α2-adrenergic agonist and ketamine combined with either midazolam (group α2−adrenergic agonist, midazolam, ketamine, AMK; n = 34), a μ-opioid receptor agonist (group α2−adrenergic agonist, μ-opioid receptor agonist, ketamine, AOK; n = 13), or a μ−opioid receptor agonist and midazolam (group α2−adrenergic agonist, midazolam, μ-opioid receptor agonist, ketamine, AMOK; n = 10). Inhalant anesthetics (isoflurane, n = 21; sevoflurane, n = 23) were frequently used for maintenance of anesthesia following premedication. Out of the 34 total tortoises, 22 had only one anesthetic event, five had two anesthetic events, three had three anesthetic events, and four had four or more anesthetic events. Few adverse effects were observed and there was no mortality reported during the peri-anesthetic period. Sedation and general anesthesia of giant tortoises can be successfully performed with a combination of an α2-adrenergic agonist and ketamine in combination with midazolam and/or a μ−opioid receptor agonist.

1. IntroductionPhysical examination and diagnostic evaluation can be difficult in conscious giant tortoises because of their unique anatomy, large size, and behavior. Consequently, sedation and anesthesia are often essential to perform medical and surgical procedures. Few published studies describe successful anesthesia of the largest tortoise species. Techniques for anesthesia and analgesia have been described for Galapagos tortoises (Chelonoidis nigra) undergoing surgical sterilization. These regimens included ketamine and medetomidine in females, and intrathecal lidocaine in males [1,2]. Medetomidine and ketamine have also been used to anesthetize Aldabra tortoises (Aldabrachelys gigantea) [3]. Anesthesia of African spurred tortoises (Centrochelys sulcata) has been reported using combinations of midazolam, medetomidine, ketamine and morphine [4,5]. All of these reports provide limited information on time to sedation and recovery and do not describe any potential adverse effects.The main objectives of this study were to describe anesthetic protocols used in three giant tortoise species (Aldabra, Galapagos, and African spurred), and to compare the dosages used between these species. 2. Materials and MethodsAs a retrospective clinical study, approval from the Institutional Animal Care and Use Committee of the University of Florida was not required. Data retrieval was limited to complete records of anesthesia of Galapagos, Aldabra, and African spurred tortoises that underwent general anesthesia at the University of Florida College of Veterinary Medicine from 2009 to 2019. Records were considered complete if all drug dosages were recorded. No records were excluded from our analysis. Digital records were identified by searching for all tortoise anesthetic events from this time period, and then sorting them by species. The data was retrieved and organized by one author using Microsoft Excel software. Information retrieved from medical and anesthesia records included species, sex, age (as reported by the owner or caretaker), weight, presenting complaint, preanesthetic hematologic analysis, comorbidities, medications, and any anesthetic event within the past seven days. Anesthetic drugs used for premedication, induction, and maintenance, antibiotics, fluids, anesthetic reversal agents, and route of administration of all medications were also recorded. Premedication protocols were categorized based on drug class: α2-adrenergic agonists, benzodiazepines, dissociatives, and μ−opioid receptor agonists. Time to intubation and type and size of the endotracheal tube were recorded for animals that underwent endotracheal intubation. Time to intubation was defined as the time from administration of the first premedication combination to placement of the endotracheal tube. In addition, the mode of ventilation (spontaneous, manual or both) was recorded. Monitoring devices, recorded physiologic variables, and any peri-anesthetic complications were also noted. Time to effect (minutes) was defined as that from initial drug administration to either the start of maintenance anesthesia or when sedation was adequate to facilitate the intended procedure. Time to extubation (minutes) was defined as either time from discontinuation of inhalant anesthetic gas to removal of the endotracheal tube, or time from the end of the procedure to removal of the endotracheal tube for patients that were not administered inhalants. To accomplish the main study objective, data from all 64 anesthesia episodes were included in the analysis. Descriptive statistics were calculated from the signalment and historical data, which included species, age, sex, and body weight. Drug dosages are reported as mean with a dosing range if appropriate, and all time variables are reported as median [interquartile range (IQR)]. Ages and weights are reported as mean ± standard deviation (range). The non-parametric Kruskal–Wallis test was used to test the null hypothesis that dosages of drugs administered were not significantly different between species. In addition, a power analysis was performed to evaluate the strength of these comparisons. Analyses were performed using Sigmaplot Version 14.0 (Systat Software Inc., San Jose, CA, USA).3. ResultsThe 64 anesthetic events included six Aldabra tortoises (aged 32 ± 15 (16–65) years, weighing 139 ± 32 (52–210) kg), 23 Galapagos tortoises (aged 42 ± 28 (6–85) years weighing 119 ± 56 (13.5–250) kg), and five African spurred tortoises (aged 11 ± 7 (1–20) years, weighing 22 ± 23 (0.5–57.7) kg). There were 18 females, 14 males and two that were undetermined sex. Indications for anesthesia included computed tomography (n = 29), esophageal feeding tube placement (n = 10), radiography (n = 10), wound care (n = 10), biopsy (n = 8), venipuncture (n = 6), echocardiography (n = 5), endoscopy (n = 5), cloacal prolapse repair and reduction (n= 3), fine needle aspiration (n = 3), tracheal or bronchiolar lavage (n = 3), ultrasonography (n = 3), carapace fracture repair (n = 2), cystolithotomy (n = 2), and one each for bronchoscopy, cloacolith removal, surgical exploration of an inguinal mass, and tail amputation. Many of the tortoises had multiple procedures performed during a single anesthetic event. Out of the 34 total tortoises, 22 had only one anesthetic event, five had two anesthetic events, three had three anesthetic events, and four had four or more anesthetic events. Drugs and dosages, route of administration, sedation scores, and complications associated with the anesthetic events using common drug protocols are summarized in Table 1 and Table 2. Premedication drugs were administered either intravenously (9%, IV), intramuscularly (78%, IM), or the route of administration was not reported (13%). The most commonly used drug combination for all anesthetic events was an α2-adrenergic agonist, midazolam, and ketamine (Group α2−adrenergic agonist, midazolam, ketamine, AMK; n = 34, 53%). In this group, medetomidine was the most commonly used α2-adrenergic agonist (76%), followed by dexmedetomidine (18%) and detomidine (6%). Administration of an α2-adrenergic agonist, a μ-opioid receptor agonist, and ketamine (Group α2−adrenergic agonist, μ-opioid receptor agonist, ketamine, AOK) occurred 13 times (20%). Medetomidine (46%) and dexmedetomidine (46%) were the most common α2-adrenergic agonists used with this drug combination, followed by detomidine (8%). Morphine was the most common opioid used with this drug combination (62%) followed by methadone (23%) and hydromorphone (15%). Finally, a combination of an α2-adrenergic agonist, midazolam, a μ-opioid receptor agonist, and ketamine (Group α2−adrenergic agonist, midazolam, μ-opioid receptor agonist, ketamine, AMOK) was used 10 times (16%). For this drug combination, dexmedetomidine and hydromorphone were most commonly used (50%), while dexmedetomidine was administered with morphine one time (10%). Medetomidine and morphine were used twice (20%) while hydromorphone (10%) and methadone (10%) were combined with medetomidine one time each. The remaining 11% of anesthetic events involved an alternative premedication drug protocol. For all tortoises, regardless of the premedication protocol utilized, the median (IQR) time to effect was 62 min (47–90) Dosages of various α2-adrenergic agonists and ketamine were compared between Aldabra, Galapagos, and African spurred tortoises, as these two classes of drugs were included in the anesthetic protocols of most tortoises. This comparison supported the null hypothesis that dosages of drugs administered were not significantly different between species. Medetomidine was the only α2-adrenergic agonist administered to Aldabra tortoises, and dosages administered to this species (0.05, 0.02–0.1 mg/kg) were not significantly different compared with dosages used for Galapagos tortoises (0.05, 0.02–0.2 mg/kg; p = 0.69). Medetomidine was administered one time to an African spurred tortoise at a dosage of 0.05 mg/kg. Dosages of dexmedetomidine administered to Galapagos tortoises (0.03, 0.02–0.06 mg/kg) were similar to dosages administered to African spurred tortoises (0.03, 0.02–0.05 mg/kg). Ketamine dosages between different tortoise species were not significantly different (Aldabra 4.6, 2–10.3 mg/kg, Galapagos 4.2, 1.2–10 mg/kg, African spurred 4.2, 2–9.6 mg/kg; p = 0.99). Power analysis calculated from the study population, assuming a clinically significant difference in dosages between the species (α 0.05) of 5 mg/kg or greater for ketamine, 0.03 mg/kg or greater for medetomidine, and 0.02 mg/kg for dexmedetomidine indicated a power of 0.79 for comparing ketamine dosages between species, 0.87 for comparing medetomidine dosages, and 0.94 for comparing dexmedetomidine dosages.Sedation scores assigned by the anesthetist following administration of anesthetic agents included no effect, mild, moderate, and profound. These scores were subjectively determined by the individual anesthetist rather than a structured sedation scoring system. Distribution of anesthetist-assigned sedation scores for different anesthetic protocols are described in Table 2. The score of profound was only assigned to protocols in the AMK group, and eight of those nine protocols were used for Galapagos tortoises, with the other for an Aldabra tortoise.Additional premedication or anesthetic agents were required in 25 of 64 (39%) cases to improve the quality of sedation or facilitate endotracheal intubation. Of those 25 cases, seven tortoises required sevoflurane via face mask (2–5% sevoflurane based on the vaporizer setting) to facilitate endotracheal intubation. Additional injectable drugs were required in 18 cases. Of those 18 cases, 10 required a second premedication dose, five were administered a third, and three required four or more injections. The drugs used for supplementary anesthesia by IM injection included ketamine (2.5, 0.1–5.0 mg/kg), medetomidine (0.04, 0.02–0.07 mg/kg), midazolam (0.2, 0.1–0.2 mg/kg), dexmedetomidine (0.03, 0.02–0.03 mg/kg) and tiletamine–zolazepam (2.1, 1.5–2.5 mg/kg). Anesthetic agents administered IV following an IM premedication included propofol (1.1, 0.7–1.6 mg/kg), propofol (1.0, 0.9–1.4 mg/kg) with ketamine (0.3, 0.09–1.0 mg/kg), and ketamine alone (5 mg/kg). Only one tortoise (5.5%) in group AOK needed additional injectable anesthetic agents, three (17%) tortoises in group AMOK, 10 (56%) in group AMK, and four (22%) after administration of another premedication protocol. There was no significant difference between the dosages of either α2-adrenergic agonist, opioid, or ketamine administered as premedication agents between these groups. All Aldabra, 11 Galapagos, and one African spurred tortoise required additional premedication agents beyond the initial premedication drugs. Table 3 illustrates which protocols required additional injections or inhalant anesthesia for each species. A total of 48 animals were intubated and the median (IQR) time to intubation for all tortoises was 80 (42–129) min.Inhalant anesthetics (isoflurane 48%, sevoflurane 52%) were administered to 44 of 48 (92%) intubated tortoises. Nitrous oxide was administered to three animals (7%) in addition to oxygen as the carrier gas, and the remainder were administered only oxygen. Supplemental drugs administered IV to maintain an appropriate anesthetic depth were required for eight tortoises. These included either a single injection (0.05–1.14 mg/kg; n = 6), or a constant rate infusion (CRI) of propofol (0.05–0.1 mg/kg/hour; n = 1), a ketamine bolus (0.8–3.18 mg/kg; n = 3), or midazolam bolus (0.8 mg/kg; n = 1). Additional analgesics were administered to seven tortoises, including meloxicam (0.1–0.3 mg/kg; n = 3; subcutaneously [SQ] or IM), hydromorphone (0.02 mg/kg; n = 1, IM), or morphine (0.2 mg/kg; n = 2, IV). Infiltration with local anesthetics was performed twice; once with lidocaine (0.3 mg/kg) and once with a combination of lidocaine (0.95 mg/kg) and bupivacaine (0.1 mg/kg). Coccygeal intrathecal administration of medication was performed in two tortoises: once with lidocaine (0.26 mg/kg) and once with morphine alone (0.01 mg/kg). Pharmacologic intervention for bradycardia, hypoventilation, or apnea that occurred during general anesthesia was administered to six animals, including antagonist drugs (atipamezole 0.1 mg/kg, n = 2, flumazenil 0.002 mg/kg, n = 1), atropine (0.02–0.03 mg/kg, n = 4), glycopyrrolate (0.008 mg/kg, n = 1), and epinephrine (0.08 mg/kg, n = 1). Monitoring was performed using a variety of different multivariable monitors and included capnography, doppler assessment of heart rate (HR, Parks Medical Electronics, Beaverton, OR, USA), electrocardiography, pulse oximetry, visual assessment of respiratory rate (RR), and temperature monitoring with either an infrared thermometer gun (Raytek, Santa Cruz, CA, USA) or an esophageal temperature probe (Smiths Medical, Minneapolis, MN, USA; Philips, Cambridge, MA, USA). Thermal support was provided by the use of a heat lamp or warming blanket (Bair HuggerTM, 3M, St. Paul, MN, USA). Monitoring results at the start of anesthesia were evaluated. Across all species and protocols HR was 16 ± 8 bpm, spontaneous RR was 6 ± 4 bpm, manual ventilation was 6 ± 4 bpm, end-tidal carbon dioxide tension was 27 ± 9 mmHg, and body temperature was 25 ± 3 °C. Manual ventilation was performed for 24 (38%) tortoises, five ventilated spontaneously (8%), and 21 had a combination of both (33%). Ventilation was not reported in six (9%) cases, and eight tortoises were not intubated (12%).Reversal agents, including atipamezole, flumazenil, and naloxone were administered at the beginning of the recovery period in 50 cases, not administered in three cases, and not reported in 11. Atipamezole was administered 48 times (0.17, 0.004–0.5 mg/kg). Flumazenil was administered 33 times (0.01, 0.001–0.01 mg/kg), and naloxone three times (0.01, 0.003–0.02 mg/kg). Reversals were administered IM except in two tortoises where atipamezole was administered IV. Epinephrine (0.03, 0.0008–0.1 mg/kg) was administered during recovery 10 times across nine cases and was administered IV once, intraosseous (IO) once and IM eight times. For all anesthetic events, the duration of anesthesia ranged from 45 to 405 min (125 [80–185] min) and time to extubation ranged from 7 to 225 min (58 [30–73.5] min). There were four outliers (3 Aldabra, 1 Galapagos tortoise) in this data set that had significantly longer times to extubation than the other tortoises (range 101–225 min). There were no consistent differences in clinical condition, anesthetic protocols, or anesthetic duration between these outliers versus animals with more rapid recoveries. All of the tortoises in this outlier group were administered reversal agents in the recovery period and none of these tortoises were administered epinephrine. In the recovery period, ventilation was noted to be supported manually in 22/64 (34%) anesthetic events while temperature support was noted in 3/64 (5%) events.Reported complications were noted in 18 out of 64 total anesthetic events (28%), and included prolonged recovery (seven tortoises, 11%), apnea (seven tortoises, 11%), bradycardia (three tortoises, 5%), hypothermia (two tortoises, 3%), endotracheal tube obstruction (one tortoise, 2%), expectoration of blood and mucus (one tortoise, 2%), hypocapnia (one tortoise, 2%), and hypoventilation (one tortoise, 2%). The distribution of complications in the most commonly used protocols are outlined in Table 2. These complications were subjectively determined by the anesthesia team involved in each individual anesthetic event. There were no mortality events during the anesthetic period in this population of tortoises.4. DiscussionThe majority of premedication protocols used in the giant tortoises under study involved combinations of ketamine and an α2-adrenergic agonist. These premedication drugs were then co-administered with either a benzodiazepine alone or with a μ-opioid receptor agonist. The dosages of medetomidine and ketamine used in the tortoises were lower than dosages described in previous studies [1,3,6]. This is likely due to the clinical opinion that there would be synergistic effects of co-administered anesthetic drugs, allowing lower dosages of ketamine and α2-adrenergic agonists. Previous reports of midazolam administration with an α2-adrenergic agonist and ketamine in chelonians resulted in a reduction in all drug dosages, less cardiac depression, and shorter recoveries compared with use of α2-adrenergic agonists and ketamine alone [7]. The opioid dosages reported in this study were also lower than those that have been published in previous studies [4,8]. This may be due to the clinical perception of differences in species, patient size, the existence of comorbidities in this population of tortoises, or the co-administration of multiple different premedication agents permitting the use of lower dosages to achieve an adequate anesthetic plane. It is also possible that as tortoise body mass increases, metabolism decreases, therefore, necessitating lower dosages to achieve therapeutic results [9]. Also, these previous studies described dosages in relatively healthy populations of tortoises, while all the tortoises under study in this review were anesthetized to address underlying medical concerns, which may have motivated anesthetists to use lower dosages of premedication agents.The sedation score following initial premedication was subjectively assigned by the individual anesthetists, and inter-individual variation in scoring is expected. Additionally, a standardized scoring system was not used and several of the sedation categories likely overlapped. The score of profound was only reported for AMK protocols, which was the most frequently used protocol. This may have contributed to more variation in sedation scores associated with this protocol. Also, AMK was used most frequently for Galapagos tortoises, and eight of the nine profound scores were assigned to this species. It is possible that Galapagos tortoises are more sensitive to the effects of this anesthetic protocol, contributing to a higher frequency of profound sedation. Both isoflurane and sevoflurane were used to successfully maintain anesthesia in the majority of giant tortoises. A lack of difference in recovery variables using inhalants of varying blood–gas solubility has been previously reported in other reptiles, including Dumeril’s monitors (Varanus dumerilii) and green iguanas (Iguana iguana) [10,11]. There are several factors beyond inhalant solubility that can influence the speed of anesthetic recovery from inhalant anesthetics in chelonians. This includes anesthetic duration and maintenance dose, co-administration of other premedication or anesthetic agents, the use of reversal agents, body temperature, minute ventilation rate in the recovery period, and the degree of right-to-left cardiac shunting that delays the development of an appropriate inhalant partial pressure within the functional lung unit to facilitate anesthetic wash-out [12,13,14]. Pulmonary shunting can be reduced or eliminated by parenteral administration of atropine or epinephrine, which can, respectively, significantly reduce the delivered concentration of isoflurane required to maintain general anesthesia, and hasten the return of spontaneous ventilation, spontaneous movement, and extubation times during recovery from inhalant agents in chelonians [12,13,15]. No mortalities were reported in this review and all complications were easily addressed with either reversal of anesthetic drugs, intubation and ventilation, or thermal support. The majority of tortoises required ventilatory support during the anesthetic period and into recovery, and apnea was one of the most commonly reported complications. Respiratory depression is a known sequela of anesthesia with α2-adrenergic agonists and μ-opioid receptor agonists in mammals and in reptiles [8,16,17,18]. Hypoventilation has been reported in desert tortoises (Gopherus agassizii) following medetomidine sedation, and IV administration of medetomidine and ketamine induced moderate hypoventilation in gopher tortoises (Gopherus polyphemus) [6,16]. These anesthetic agents were used in this population of tortoises, and likely contributed to hypoventilation and apnea. However, there are other factors influencing ventilation requirements in anesthetized tortoises, including impairment of muscular motion of the limbs that normally facilitates movement of gases through the respiratory system and the use of high concentrations of oxygen as a carrier gas during inhalant anesthesia, which may suppress the ventilatory drive of chelonians [5,7,19,20,21]. Body temperature and positioning also have a profound effect on ventilation in reptiles [22,23]. In freshwater turtles (Chrysemas picta bellii), pulmonary ventilation and oxygen uptake both increased along with a rise in body temperature [19]. Variations in ambient and patient temperature may have also contributed to hypoventilation and apnea observed in the present report. Additional premedication drugs were necessary to achieve an appropriate plane of sedation or anesthesia in several tortoises. Fewer animals required supplemental drugs in the AOK group compared with either AMK or AMOK groups, but similar dosages were used for the individual drugs across the different premedication protocol groups. A myriad of other factors exist that contribute to the ability of injectable anesthetics to achieve a desired plane of anesthesia in chelonians, including route and location of drug administration, ambient and body temperature, the degree of activity and excitation prior to drug administration, the body condition of the patient, body position during anesthesia, and the overall clinical status of the patient [7,23]. Due to the retrospective nature of this review, these factors could not be controlled. This study had several limitations. Complications and sedation scores reported here were assigned by individual anesthetists and were not always clearly defined or assigned based on the same criteria. Data collected during anesthesia could not be standardized across anesthetic events, due to the retrospective nature of this study; consequently, information such as body temperature was often omitted. Due to these omissions, more in-depth statistical analysis of the data, such as factors affecting time to recovery, were not performed. Analysis was also impacted by the small sample size for sulcata tortoises. This study was slightly underpowered, especially to detect subtle differences in ketamine dosing between the species. However, there are many other factors influencing the dosage of ketamine beyond species differences, including other medications administered, health status of the animal, and body temperature. In addition, this review relied on anesthetic records from a single referral veterinary hospital, where the majority of the animals included in the study were clinically ill or injured. Therefore, information gained from this study may not translate to a healthy population. Pharmacokinetic and pharmacodynamic studies on anesthetic drugs are warranted to better elucidate their clinical effects in giant tortoises.5. ConclusionsAnesthesia of Galapagos, Aldabra, or African spurred tortoises was safe and effective with any of the drug combinations reported here. A combination of an α2-adrenergic agonist, midazolam, and ketamine was the most common induction protocol. No mortalities were reported in this review and all complications were resolved using appropriate interventions.

animals : an open access journal from mdpi

10.3390/ani11113307

PMC8614385

Cryptosporidium is a parasite that infects humans and a broad range of animals. There are few diagnostic features that can be used to identify and differentiate between species and therefore DNA-based detection and genetic typing methods are required. This is important as some species are transmitted from animals to humans (a process called zoonotic transmission) and understanding this is central to control. These DNA-based tools have greatly facilitated our current understanding of which of the many Cryptosporidium species and genotypes that have been identified are capable of infecting humans. More studies need to be conducted in areas where the potential for zoonotic transmission is greatest and recently developed genetic tools should be applied more extensively to further our understanding of the zoonotic transmission of this important parasite.

The enteric parasite, Cryptosporidium is a major cause of diarrhoeal illness in humans and animals worldwide. No effective therapeutics or vaccines are available and therefore control is dependent on understanding transmission dynamics. The development of molecular detection and typing tools has resulted in the identification of a large number of cryptic species and genotypes and facilitated our understanding of their potential for zoonotic transmission. Of the 44 recognised Cryptosporidium species and >120 genotypes, 19 species, and four genotypes have been reported in humans with C. hominis, C. parvum, C. meleagridis, C. canis and C. felis being the most prevalent. The development of typing tools that are still lacking some zoonotic species and genotypes and more extensive molecular epidemiological studies in countries where the potential for transmission is highest are required to further our understanding of this important zoonotic pathogen. Similarly, whole-genome sequencing (WGS) and amplicon next-generation sequencing (NGS) are important for more accurately tracking transmission and understanding the mechanisms behind host specificity.

1. IntroductionCryptosporidium species are enteric parasites with a global distribution and a wide range of hosts. Transmission is by the faecal–oral route via contaminated water, food or direct contact with humans and animals [1]. Although first described in 1907 by Tyzzer [2], Cryptosporidium did not come to prominence until the early 1980s, when it was identified as a cause of severe protracted diarrhoea and death in HIV+/AIDS patients [3]. It is now recognised as a major pathogen in children and immunocompromised adults [4,5] and after rotavirus is the most important diarrheal pathogen in young children [6,7]. In 1993, Cryptosporidium was responsible for a large waterborne outbreak affecting over 400,000 residents of Milwaukee, Wisconsin [8] and although still frequently under-reported, is a well-known and major cause of both waterborne and foodborne outbreaks of gastroenteritis globally [9,10,11]. This is due in part to the resistance of the environmental stage, the oocyst, to disinfectants including chlorine treatment of both drinking and recreational water [12,13].The parasite has a complex life cycle that initiates upon ingestion and excystation of oocysts and involves both asexual and sexual phases, which culminate in the shedding of infectious thick-walled oocysts in faeces [14]. In humans, in addition to watery diarrhoea, cryptosporidiosis can cause abdominal pain, vomiting, headaches, joint pain, malnutrition, failure to thrive and cognitive deficits and has been linked with colon cancer [15,16,17,18,19,20]. Although usually self-limiting in immunocompetent hosts, cryptosporidiosis can become chronic, persisting for 2 years or more [5,17]. In neonatal livestock, cryptosporidiosis can cause profuse diarrhoea, weight loss and death [11,21,22,23], resulting in significant production losses [22,24,25,26]. The only US Food and Drug Administration approved drug, nitazoxanide, is ineffective in the most affected populations (children and immunocompromised individuals) and there is no FDA-approved vaccine [27,28,29,30]. In animals, nitazoxanide is also largely ineffective and while Halocur® (halofuginone lactate) has been licensed in some countries as a prophylactic, its effectiveness is variable and it cannot be given to animals that already have diarrhoea [11,22].Molecular typing of Cryptosporidium has advanced significantly over the last few decades and is particularly important for this parasite as many species are morphologically identical [31]. Identification to species level is most commonly conducted using the 18S ribosomal RNA locus due to the existence of both hypervariable and conserved sequences [31,32]. Other loci examined include actin, heat shock protein 70 (hsp70) and the Cryptosporidium oocyst wall protein gene (COWP) [31]. Currently, at least 44 species of Cryptosporidium and >120 genotypes have been identified, with C. hominis and C. parvum being the most important species infecting humans [33].Contact with animals has been identified as a risk factor in many studies [34,35,36,37]; therefore, zoonotic transmission plays a major role in the epidemiology of cryptosporidiosis [1,23,38,39,40,41]. In order to track transmission, multi-locus sequence typing (MLST) tools have been developed, including analysis of the highly polymorphic glycoprotein 60 (gp60) gene, the most commonly used locus [31]. The gp60-based typing system is based on a combination of tandem serine-coding trinucleotide repeats (TCA, TCG and TCT) and extensive sequence divergence in non-repeat regions [31,42,43]. The nomenclature system [43], starts with a Roman numeral and lower-case letter for each Cryptosporidium and genotype (e.g., C. hominis, Ia, Ib, etc., C. parvum IIa, IIb, etc.), followed by uppercase letters denoting numbers of repeats. For example, C. hominis subtype IfA19G1R5 has 19 TCA (A) repeats, 1 TCG repeat (G) repeat and 5 AAGAAGGCAAAGAAG repeats (R). For some divergent species and genotypes, including C. ubiquitum, C. felis, C. canis, C. ryanae, C. xiaoi, C. bovis and Cryptosporidium chipmunk genotype I and skunk genotype, whole-genome sequencing (WGS) has been required to identify the gp60 loci in order to develop typing systems [44,45,46,47,48,49,50]. The trinucleotide repeats (TCA, TCG, TCT) are absent from gp60 genes in C. ubiquitum, C. felis, C. canis, C. ryanae, C. bovis and C. xiaoi [44,46,47,48,49,50].This review aims to summarise the currently available data on the zoonotic transmission of Cryptosporidium species and genotypes as well as outlining future studies that are required to better understand the transmission dynamics of this ubiquitous parasite. This review could provide a valuable reference for One Health scientists and help guide future research.2. Zoonotic Cryptosporidium Species and GenotypesIn total, 19 species (including C. hominis and C. parvum) and 4 genotypes have been reported in humans including C. meleagridis, C. canis, C. felis, C. ubiquitum, C. cuniculus, C. viatorum, C. muris, C. andersoni, C. erinacei, C. tyzzeri, C. bovis, C. suis, C. scrofarum, C. occultus, C. xiaoi, C. fayeri, C. ditrichi, Cryptosporidium chipmunk genotype I, mink genotype, skunk genotype and horse genotype [1,33] (Table 1). Where possible, subtyping analysis at the gp60 locus has been used to document gp60 subtypes common to humans and animals, to support zoonotic transmission (Table 2). Cryptosporidium hominis and C. parvum are responsible for ~95% of human infections, followed by C. meleagridis, C. felis, C. canis and C. ubiquitum [1,38] (Figure 1). Some species such as C. meleagridis, C. canis, C. felis, C. viatorum and C. muris are more commonly identified in humans in developing countries, whereas others such as C. ubiquitum, C. cuniculus and chipmunk genotype I are predominantly seen in developed countries [1].2.1. Cryptosporidium hominisCryptosporidium hominis is the dominant species in humans in many industrialised countries and in developing countries, whereas, in the Middle East, European countries and New Zealand, C. parvum occurs at similar rates to C. hominis [1], indicating that zoonotic transmission is more prevalent in countries with intensive farming [122]. Although largely anthropologically transmitted, there have been numerous reports of C. hominis in animals (including non-human primates, cattle, sheep, goats, horses, donkeys, Bactrian camels, birds, marsupials, a dugong, badger, dingo, foxes, flying fox, rodents, and fish) and experimental infections has been established in calves, lambs, piglets, gerbils, and mice [123,124,125]. In human infectivity trials, the 50% infectious dose (ID50) was as low as 10 C. hominis oocysts [126], however, in animal models, much larger numbers of oocysts were required to achieve infections, suggesting that higher doses are required to cause infection in animals [125].Cryptosporidium hominis has been detected at a relatively low frequency in livestock [22,23,125], although in one study, 80% of Cryptosporidium positives in calves in New Zealand were identified as C. hominis [127] and C. hominis was the only Cryptosporidium sp. detected in goats in South Korea [128]. These studies demonstrate that cattle, sheep and goats can potentially serve as animal reservoirs for infections with C. hominis [22,23,125]. In Australia, in addition to cattle and sheep, C. hominis has been detected in deer, dingoes and marsupials inhabiting drinking water catchments [129,130,131]. This likely resulted from human spill-back either via direct contact or via sewage contamination of pastures or drinking water.The C. hominis subtypes identified at the gp60 locus in non-human hosts has been extensively reviewed [125]. Briefly, the C. hominis IbA10G2 subtype dominates in livestock that are positive for C. hominis [22,125] and has also been detected in hedgehogs and mice in Europe [132,133,134]. As IbA10G2 is one of the most prevalent subtypes in humans in many countries and is also a major cause of waterborne outbreaks [1,11,38], this further supports a human origin for C. hominis in animals. In equines and particularly donkeys and horses, the Ik C. hominis subtype family (which is relatively rare in humans) is endemic and a reservoir of infection for humans [135,136,137,138]. Studies in Australia also suggest that C. hominis may become endemic in marsupials [131,139,140]. Continued human encroachment into wildlife areas may result in C. hominis becoming increasingly endemic in livestock and wildlife populations escalating the risk of zoonotic C. hominis transmission.2.2. Cryptosporidium parvum and Other Livestock-Associated SpeciesCryptosporidium parvum has a very wide host range infecting ungulates, wildlife (including carnivores, rodents, non-human primates, marine mammals and fish) and is the most important zoonotic species in humans, particularly in rural areas with frequent contact with livestock [1,11,22,23,124,141,142]. Calves, sheep and goats are major contributors to zoonotic C. parvum transmission with significant differences in prevalence [11,22,23,143]. In cattle, the dominant species are C. parvum, C. bovis, C. ryanae and C. andersoni, although other species (C. hominis, C. felis, C. suis, C. scrofarum, C. meleagridis, C. tyzzeri, C. serpent is, a novel genotype and C. occultus) have been reported [11,22]. In sheep and goats, C. parvum, C. ubiquitum and C. xiaoi are the predominant species, but a range of other species including C. bovis, C. hominis, C. andersoni, C. suis, C. muris, C. fayeri, C. baileyi, C. ryanae, C. scrofarum, C. canis, C. occultus, Cryptosporidium sheep genotype I and rat genotype II have occasionally been reported [11,22,23]. In pre-weaned calves, C. parvum is the dominant species in most studies [22], with the exception of China where C. bovis mostly dominates [144]. There does not appear to be an age-related prevalence of C. parvum in sheep and goats and while C. parvum dominates in these hosts in Europe and Australia, in other regions including the Americas, Asia, C. xiaoi and C. ubiquitum dominate in sheep and goats [23].Over 20 C. parvum gp60 subtype families have been identified with geographic variation in subtype distribution as well as host adaptation demonstrated [23,31]. For example, the three dominant C. parvum gp60 subtype families in humans are IIa, IId and IIc, and of these, IIc appears to be almost exclusively anthropologically transmitted whereas IIa and IId are zoonotic [1,33,145]. WGS also supports this and it has been suggested that the C. parvum IIc subtype should be considered a separate subspecies (Cryptosporidium parvum anthroponosum), while the zoonotically transmitted IIa and IId subtypes referred to as Cryptosporidium parvum parvum [146]. In cattle, IIa is the dominant C. parvum subtype family in most countries and particularly in Europe and contact with calves has been identified as a risk factor for infection with the IIa subtypes [34]. In other countries, however, including China, Egypt, Malaysia, Romania and Sweden, the IId subtype dominates in cattle [31,144]. Similarly in sheep and goats, IIa dominates in European countries but in Africa, Asia and Australia, IId is the dominant C. parvum subtype family [23,144]. Subtyping studies have confirmed the identification of identical IIa and IId subtypes in humans and animals indicating zoonotic transmission [11,22,23,31,117,139]. For example, the relatively rare IIaA19G1R1 was identified in children, goat kids, and lambs during an outbreak of cryptosporidiosis at a holiday farm in Norway [147]. Similarly, subtypes IIaA15G2R1, IIaA18G2R1 and IIaA19G1R1 were responsible for several outbreaks in the UK, with the same subtypes identified from livestock on the same farms where the outbreaks occurred [38]. Of these, IIaA15G2R1 is the dominant C. parvum subtype in humans and cattle in industrialised countries and MLST analysis indicates that genetic recombination is the driving force behind its emergence as a dominant and hyper-transmissible subtype [148].Cryptosporidium bovis is prevalent in cattle and other bovid species, while C. xiaoi is a related species in sheep and goats worldwide [11,22,23]. A gp60 typing system has been established for C. bovis based on WGS sequences. Sixty-eight subtypes in six subtype families (XXVIa to XXVIf) have been identified, with apparent genetic recombination among subtype families [50]. There have been several reports of C. bovis in humans (Table 1). It was first reported in a dairy farm worker (and dairy calves) in Bengal, India in 2010 [108] and again in 2012 in a study in Australia, which collected and screened faecal samples from diarrheic calves in rural NSW and farmers from these farms [109]. In that study, C. bovis was identified in a 3-year-old child and a 23-year-old adult, from separate farms and both were asymptomatic [109]. Both individuals drank raw milk and had frequent contact with calves [109]. In another study, a mixed C. parvum and C. bovis infection was identified in a diarrhoeic child (<6 years) (along with mixed infections in cattle) from Ismailia province in Egypt [110]. A gp60 typing system has recently been established for C. xiaoi based on WGS, with 12 subtype families, (XXIIIa to XXIIIl) and high subtype diversity identified [48]. It has previously been reported in two HIV/AIDS patients from Ethiopia based on 18S sequences only [34]. Both C. bovis and C. xiaoi appear to have narrow host ranges but further research is required to understand the zoonotic potential of these common livestock species.Like C. parvum, C. ubiquitum (previously cervine genotype) also has a wide host range and is commonly detected in both domestic and wild ruminants, rodents, carnivores, primates, and humans, particularly in industrialised countries [44,51,52,139,149,150]. Both gp60 and MLST tools have been developed for C. ubiquitum [44,150]. At the gp60 locus, C. ubiquitum lacks the TCA, TCG and TCT repeats and to date, eight subtype families (XIIa–XIIh) have been identified [44], yet the transmission routes between animals and humans are not well understood. In the UK, XIIa is the dominant C. ubiquitum subtype in humans, which is also the main subtype in small ruminants [23,44]. In contrast, human C. ubiquitum infections in the US are caused by XIIb, XIIc and XIId subtype families which are common in rodents [44]. Subtype families XIIe and XIIf are also found in rodents but at present have only been identified in the Slovak Republic [44]. In France and Sweden, XIIb and XIId subtype families have been reported in humans [51,52]. MLST analysis has also re-affirmed that XIIa is ruminant-adapted and XIIb, XIIc and XIId subtype families are rodent-adapted [150].Cryptosporidium andersoni, a gastric species, was originally thought to be C. muris due to its morphological similarity, until it was established as a separate species [151]. It is commonly reported in ruminants, particularly adult cattle and other bovids and is also common in sheep and goats, mainly in China [23] and has been reported in deer [152] and rodents, particularly hamsters [153,154,155,156]. As gastric Cryptosporidium species do not appear to have the gp60 gene, a gp60 typing tool has not been established for C. andersoni. However, an MLST tool has been developed and a mostly epidemic population structure has been identified [157,158,159]. There have been numerous reports in humans, but the extent of zoonotic transmission remains to be determined (Table 1).Cryptosporidium suis and C. scrofarum are the dominant species infecting pigs (and wild boars) and mostly cause subclinical infections (although C. parvum, C. muris, C. tyzzeri, C. felis, C. hominis, C. andersoni and C. meleagridis can also infect pigs) [160,161,162]. Cryptosporidium suis is mainly found in pre-weaned animals and C. scrofarum in post-weaned pigs. Neither C. suis nor C. scrofarum are commonly identified in humans. There have been a few reports of C. suis in HIV-positive individuals in China, Peru and Thailand [79,92,101,102,103], as well as children in Cambodia [99] and symptomatic individuals in the UK and Madagascar [96,100]. There has only been one report of C. scrofarum in a symptomatic immunocompetent 29-year-old male co-infected with Giardia [121]. A gp60 typing tool has not yet been developed for C. suis or C. scrofarum and need to be developed to explore the zoonotic potential of C. suis and C. scrofarum in pigs and transmission routes via water and sewage.The horse genotype is commonly reported in donkeys and horses [135,137,138]. Typing of the horse genotype at the gp60 locus has identified subtype families VIa, VIb and VIc [31,51]. Subtype VIaA15G4 is one of the most common subtypes identified in horses and donkeys, with subtype VIbA13 and VIaA14G4 identified in a hedgehog and calf respectively [135]. The horse genotype (subtype VIcA16) has been identified in a traveller to Sweden [51], an 18-year-old symptomatic patient from the US (subtype VIbA13) [104], a symptomatic 30-year-old immunocompetent woman from the rural UK (subtype VIbA13) [105] and in two adult females from the UK [82].2.3. C. meleagridisAvian cryptosporidiosis was first described in 1929 [163], but the first avian species, Cryptosporidium meleagridis was not named until 1955 in turkeys [164]. The parasite has a broad host range and is commonly reported in wild birds but less so in poultry [165,166]. It has also been reported in foxes, minks, cattle, wallabies, gorillas, and dogs [167,168,169,170,171,172]. Experimental transmission studies have also confirmed its infectivity for calves, pigs, rabbits, rats and mice and humans [173,174]. It is the third most common Cryptosporidium species in humans [1,175] and has also been identified in colon adenocarcinoma tissue in an immunocompetent Polish patient [176]. Phylogenetic analysis suggests that C. meleagridis may originally have been a mammalian parasite that secondarily became established in birds [177]. In children and HIV+ individuals in developing countries, C. meleagridis is frequently the dominant Cryptosporidium species [1,5,178], whereas, in developed countries, C. meleagridis is usually responsible for ~1–4% of human infections [38,66,174,179]. Despite being so commonly identified in humans, the extent of zoonotic transmission is not well understood. One study in an organic farm in Sweden identified identical C. meleagridis 70 kDa Heat Shock Protein (hsp70) gene sequences in samples from one human, three chickens and one hen, suggesting zoonotic transmission [180].Typing at the gp60 locus indicates that humans are susceptible to most C. meleagridis gp60 subtypes, as of the 10 known subtype families (IIIa to IIIi), eight (IIIa to IIIh) (and 30 subtypes) have been identified in humans with IIIb among the most common [31,181,182,183,184]. Relatively few subtyping studies have been conducted in birds but both MLST and gp60 studies have provided some support for zoonotic transmission. For example, subtype IIIgA31G3R1 has been reported from four poultry and one human from a Swedish farm [181]. MLST analysis of C. meleagridis from humans and birds from Peru did not find evidence of host segregation [182] and the same gp60 subtypes (IIIbA26G1R1b and IIIbA22G1R1c) have been found in children with diarrhoea and in farmed chickens in Hubei province, China, supporting zoonotic transmission [183,184]. Similarly, subtypes IIIeA17G2R1, IIIeA19G2R1, IIIeA21G2R1 and IIIeA22G1R1 has been reported in Swedish patients (and IIIeA21G2R1 in Canadian patients) and in rodents and chickens in Asia [51,52,79,183] (Table 2).2.4. Companion Animal-Associated Species (C. canis and C. felis)Companion animals, particularly dogs and cats, have close relationships with their owners contributing to improved mental health and social support but can also contribute to zoonotic disease transmission [185]. There are currently ~470 million pet dogs and 370 million pet cats in the world (www.statista.com/statistics/1044386/dog-and-cat-pet-population-worldwide/accessed on 20 October 2021). Cryptosporidium canis and C. felis are the main species infecting dogs and cats respectively and are among the top five Cryptosporidium species infecting humans [142,186]. There have also been numerous reports of C. parvum and a few reports of C. meleagridis, C. muris, C. andersoni, C. scrofarum, C. ryanae, C. hominis and rat genotype III in cats and dogs, some of which may be due to coprophagy [40,187]. Prevalence rates of C. canis and C. felis vary widely but are commonly below 10% [40,187,188]. Few studies have examined the species of Cryptosporidium in humans and pets living in the same household, but one study identified C. canis in a 32-month-old girl, her 6.5-year-old brother and a dog from the same house in Lima, Peru [189]. Another study reported identical 18S, HSP70 and COWP C. felis sequences from a cat and her 37-year-old immunocompetent owner in Sweden [190]. A case-control study of HIV-infected individuals with and without cryptosporidiosis reported only a weak association between dog ownership and cryptosporidiosis [191].Cryptosporidium canis and C. felis gp60 loci have recently been characterised and have been shown to lack the serine-coding trinucleotide repeats normally used for typing (similar to C. ubiquitum) [40,46,49,186]. Within C. canis, five gp60 subtype families have been identified (XXa, XXb, XXc, XXd and XXe), with subtypes XXa1 and XXa4 detected in both humans and dogs [49]. The previously identified household transmission of C. canis between two children and a dog in Lima, Peru [190] was confirmed by gp60 subtyping [49]. Five subtype families have also been identified in C. felis (XIXa, XIXb, XIXc, XIXd and XIXe) [46] and of these, two subtypes (XIXa and XIXb) have been reported in both humans and cats supporting zoonotic transmission, with the remaining three subtypes (XIXc, XIXd and XIXe), possibly transmitted anthropologically [40,46,186,192]. 2.5. Wildlife-Associated Species and GenotypesDue to their wide geographical distribution and abundance and close contact with humans, rodents are considered an important zoonotic reservoir for Cryptosporidium and the global prevalence of Cryptosporidium in rodents has been estimated at ~17% [193]. To date, cryptosporidial infections from five rodent-derived species (C. viatorum, C. muris, C. tyzzeri, C. occultus and C. ditrichi) and one genotype (Cryptosporidium chipmunk genotype I) have been reported in humans (Table 1). Of these, there have only been a few reports of C. tyzzeri, C. occultus and C. ditrichi in humans (Table 1).Cryptosporidium muris, a gastric species, has a very wide host range including rodents, ruminants, cats and dogs, horses, Bactrian camels, pigs, birds, and non-human primates [22,23,40,162,194,195,196,197,198]. There have been numerous reports in humans, particularly in developing countries (in children and HIV patients) [1,5], (Table 1). As with C. andersoni, an MLST tool has been developed for C. muris [157,158], which revealed that the genetic diversity of C. muris was greater than C. andersoni, possibly reflecting the much smaller host range of C. andersoni and that host clustering was evident, suggesting that some C. muris isolates have co-evolved with their hosts over a long period of time [158]. A C. muris human infectivity trial has been conducted in which six healthy adults were challenged with 105 C. muris oocysts each (isolate RN66) and all became infected (two were symptomatic), confirming the zoonotic potential of C. muris [199].Typing of C. tyzzeri is at the gp60 locus has identified three subtype families; IXa, IXb and IXc [31,200], including subtype IXbA22R9 from a horse [195]. Cryptosporidium tyzzeri (subtype IXaA6R2) was identified from a child in Kuwait [43] and subtype family IXb has been detected in three human patients in New Zealand [54]. A mixed C. parvum (subtype IIaA13G1R1) and C. tyzzeri (subtypes IXaA8 and IXbA6) infection was identified in a symptomatic 25-year-old female conducting fieldwork trapping wild rodents in the Czech Republic [113]. The same subtypes were identified in the trapped mice supporting zoonotic transmission [113]. Cryptosporidium tyzzeri is genetically very closely related to C. parvum [201], which further supports its potential for zoonotic transmission.A gp60 typing system has not been established for C. ditrichi or C. occultus. Cryptosporidium ditrichi has been reported in three patients from Sweden, two of which were symptomatic and one had possible contact with mice [51,116]. Although predominantly a rodent species, C. occultus appears to be common in bovids including cattle, yaks (Bos gunniens) and water buffalo (Bubalus bubalis) [202] and was the dominant species detected in Alpacas in China [203]. It has been reported in one human from the UK (HQ822146), two humans from British Columbia [115] and a young child in China [114]. Cryptosporidium viatorum was originally reported in the UK in travellers returning from India and formally described as a human species in 2012 [73]. Since then, C. viatorum have been described in humans from Australia [66], China [114], Columbia [68], Ethiopia [34,64,74], India [63,67], Mozambique [62], Myanmar [65], Nigeria [69,70] and Swedish patients (who had returned from Kenya or Guatemala) [51,71,72]. It has also been detected in urban wastewater in China [204]. Due to the relatively high prevalence of C. viatorum in rats in Australia and China [205,206,207], it is now thought that C. viatorum was originally a rodent species and therefore rodents are likely to be an important reservoir host. A gp60 typing tool has been established for C. viatorum and to date four subtype families have been identified; XVa, XVb, XVc and XVd [205,206,207,208]. Subtyping supports the potential zoonotic transmission between rodents and humans. For example, human-derived C. viatorum isolates have been subtyped to date as XVa3a to XVa3h [66,114,209] and XVcA2G1c [65], and subtype XVaA6 was identified in both sewer overflow and wastewater in Shanghai, China [204]. In rodents, XVa (XVaA6, XVaA3g, XVaA3h), XVb (XVbA2G1), XVc (XVcA2G1a, XVcA2G1b) and XVd (XVdA3) subtype families have been identified [205,206,207]. Thus, three subtypes (XVaA3g, XVaA3h and XVcA2G1) are common to humans and rodents and in addition, the XVaA3g subtype identified from wild rats [206] was 100% homologous to an XVaA3g subtype identified in a human patient in Australia [66]. Screening of rodents and rats across wider geographic areas is essential to better understand their role as reservoirs for C. viatorum.To date, five chipmunk genotypes have been identified in rodents but only one of these, chipmunk genotype 1 has been identified in humans [33]. Chipmunk genotype I was first identified in New York storm water as genotype W17 [209] but was renamed as chipmunk genotype I, when it was detected in rodent faecal samples in 2007 [210]. It infects a range of rodents particularly squirrels, chipmunks, and deer mice and is considered an emerging zoonotic pathogen in humans [45,75], being the third most commonly identified Cryptosporidium sp. in Sweden [75]. A gp60 typing system has been developed for chipmunk genotype I, with one subtype family identified (XIVa) [45]. Chipmunk genotype I was first identified in two humans in 2004 in Wisconsin (subtype XIVaA16G2T1) [77], and in a 41-year-old HIV-positive male in France [60]. In 2013, it was reported in a two-year-old female and 56-year-old male from Sweden [71,72] and typed as XIVaA20G2T1 by Guo et al. [45], who also identified 19 US human isolates and typed 17 as XIVaA14G2T1 (n = 1), XIVaA16G2T1 (n = 1), XIVaA14G2T2 (n = 1), XIVaA16G2T2 (n = 4), XIVaA18G2T1b (n = 2), XIVaA17G2T2 (n = 1), XIVaA19G2T2a (n = 1), XIVaA20G2T2 (n = 2), XIVaA19G2T2b (n = 1), XIVaA15G2T3 (n = 2) and XIVaA17G2T3 (n = 1) [45]. In that study, in the three rodent samples analysed, subtypes XIVaA18G2T1a and XIVaA18G2T2 were detected. It was identified in a human in Nebraska [37] and between 2014 and 2015, chipmunk genotype 1 was identified in five adults (four women and one man) in Sweden and all were typed as XIVaA20G2T1 [51]. Subsequently, it was identified in sixteen sporadic cases, three outbreak-related cases, and one zoonotic case, as well as in two squirrel samples in Sweden between 2018 and 2019, and subtyping of nineteen humans and two squirrels identified subtype XIVaA20G2T1 in all samples, supporting zoonotic transmission [75]. Of the other wildlife species and genotypes identified in humans, C. cuniculus (previously known as rabbit genotype) is a common species in rabbits worldwide [211,212,213] and has also been reported in kangaroos and alpacas in Australia [130,131,214]. It is the only species (other than C. parvum and C. hominis) known to have caused a waterborne outbreak of disease [56,61]. It is more commonly reported in humans in industrialised countries and particularly in the UK (Table 1). Two subtype families (Va and Vb) have been identified [31]. The waterborne outbreak caused by C. cuniculus in the UK was typed as VaA18 [56], however, most C. cuniculus cases typed in humans and rabbits have been caused by subtype family Vb [51,52,53,54,55,58]. In humans, subtypes VaA11, VaA18, VaA19 to VaA22, VaA23, VbA13, VbA15 VbA17, VbA20, VbA22, VbA23, VbA24 to VbA34, VbA36, VbA37 and VbA38 have been reported [51,52,54,58,211,212]. Of these, subtypes VbA19, VbA22 to Vb26, Vb28, VbA29 and VbA31 to VbA33 have been reported in rabbits supporting zoonotic transmission [130,131,140,212,215,216,217,218]. More recently an MLST typing tool has been developed and a clonal population structure identified [219]. Cryptosporidium erinacei (previously hedgehog genotype 1) was first described in European hedgehogs in 2014 [220]. It has also been reported in horses in Algeria and rats in China [206,221] and was detected in cattle and kangaroo faecal samples by next-generation amplicon sequencing in Australia [131]. Cryptosporidium erinacei has been reported in an immunocompetent 26-year-old man from the Czech Republic [107], an immunocompromised patient in France [106], in two cases from Sweden (one locally acquired and one in a traveller from Greece) [51] and in thirteen cases from New Zealand [53,54]. The higher number of cases in New Zealand may be due to the greater abundance of European hedgehogs (Erinaceus europaeus) in New Zealand compared to Europe, but as C. erinacei can infect other hosts, the range and prevalence of this parasite in reservoir hosts remain to be determined.At the gp60 locus, one subtype family (XIIIa) has been identified [31,220]. Zoonotic transmission has not been demonstrated as the gp60 subtypes identified in humans (XIIIaA20R10, XIIIaA23R12, XIIIaA24R9, XIIIaA24R10, XIIIaA25R10 XIIIaA25R11, XIIIaA26R9 and XIIIaA26R10) [51,54] have not been identified in hedgehogs and horses (XIIIaA19R12, XIIIaA19R13, XIIIaA21R10, XIIIaA21R11, XIIIaA22R9, XIIIaA22R11 [221,222,223]). However as only a small number of isolates have been subtyped, further research is required to better understand its zoonotic potential.The skunk genotype was first identified in striped skunks (Mephitis mephitis) in the US [177], but has a broad host range and has been described in numerous species of wild mammals [111,223], and is common in surface water in the US and Canada [111]. There have been five reports in humans; an 18-month-old child attending a UK daycare centre [112], a 25-year-old woman from a rural area in the UK [105], an adult female from the UK [82] and a human in Nebraska [37,111], and all except the child attending daycare were symptomatic.Typing at the gp60 locus has identified four subtype families (XVIa, XVIb, XVIc and XVId) with a high subtype diversity (n = 14) [111,224]. In humans, subtypes XVIbA16G2b and XVIcA22 have been identified, which have not been identified in wildlife. However, a closely related subtype (XVIbA16G2a subtype) has been found in raccoons and storm runoff samples in the US [111]. There have only been two reports of the marsupial derived C. fayeri in humans; a symptomatic 29-year-old immunocompetent female in New South Wales (NSW), Australia in 2010 and a decade later, in 2020, it was identified in a symptomatic 37-year-old immunosuppressed female treated for acute myeloid leukaemia in Western Australia (WA) [119,120]. Eight subtype families have been identified within C. fayeri (Iva–IVh) [31,130,225,226]. In the initial study by Waldron et al. [119], the same C. fayeri subtype (IVaA9G4T1R1) was identified in both the woman and eastern grey kangaroos (Macropus giganteus) inhabiting the main drinking water catchment for Sydney [225] and the woman had regular contact with marsupials supporting zoonotic transmission [119]. In the latter study in WA, C. fayeri subtype IVgA10G1T1R1 was identified in the female [120] and the same subtype had previously been identified in western grey kangaroos (Macropus fuliginosus) in WA and rabbits (Oryctolagus cuniculus) in NSW drinking water catchments [131], again supporting the potential for zoonotic transmission. Marsupials are the dominant animals inhabiting drinking water catchments in Australia and further research is necessary to better understand the zoonotic implications, particularly in addition to C. fayeri and C. macropodum, marsupials can also be infected with C. hominis as discussed above [131].The Cryptosporidium mink genotype was first identified in minks (Mustela vison) inhabiting a New York City drinking water supply watershed [210] and is a common parasite in minks [227]. Typing at the gp60 locus has identified four subtypes, including XaA5G1, XbA5G1R1, XcA5G1R1 and XdA4G1 [172,227,228]. There have been two reports of the mink genotype in humans [117,118]. However, no gp60 sequences are available and therefore the possibility of zoonotic transmission is unknown.animals-11-03307-t002_Table 2Table 2Zoonotic gp60 subtypes common to humans and animals (excluding non-human primates).Species Namegp60 SubtypesReferences C. hominis IbA9G3, IbA13G3, IbA14G2, IbA10G2, IdA15G1, IbA10G2R2 and Ik subtype family[22,125,127,131,132,133,134,139,221] C. parvum Many subtypes but mainly the IIa (particularly IIaA15G2R1) and IId subtype families[11,22,23,31,117,140] C. meleagridis IIIbA21G1R1b, IIIbA22G1R1cIIIbA23G1R1b, IIIbA23G1R1c, IIIbA24G1R1, IIIbA26G1R1b, IIIeA17G2R1, IIIeA19G2R1, IIIeA21G2R1, IIIeA21G2R1 and IIIgA31G3R1[51,52,79,181,182,183,184] C. felis XIXa and XIXb[40,46,186,192] C. canis XXa1 and XXa4[40,49] C. ubiquitum XIIa, XIIb, XIIc and XIId[23,44,51,52] C. cuniculus VaA18, VbA19, VbA22 to Vb26, Vb28, VbA29 and VbA31 to VbA33[51,52,53,54,55,56,58,210,211] C. viatorum XVaA3g, XVaA3h and XVcA2G1[65,66,114,207,209]Chipmunk genotype IXIVaA18G2T2[37,45,51,60,71,72,75,76,77] C. muris Gastric Cryptosporidium species do not appear to have the gp60 gene C. andersoni C. suis A gp60 typing tool has not yet been developed Horse genotypeVIbA13 and VIcA16[51,58,82,104,105] C. erinacei XIIIaA20R10, XIIIaA23R12, XIIIaA24R9, XIIIaA24R10, XIIIaA25R10 XIIIaA25R11, XIIIaA26R9 and XIIIaA26R10 (humans only)[53,54] C. bovis No gp60 sequences available from humans Skunk genotypeXVIbA16G2b and XVIcA22 (humans only but XVIbA16G2a in wildlife)[111,224] C. tyzzeri IXaA8 and IXbA6[43,113] C. occultus A gp60 typing tool has not yet been developed C. ditrichi A gp60 typing tool has not yet been developed Mink genotypeNo gp60 sequences available from humans C. fayeri IVaA9G4T1R1, IVgA10G1T1R1[119,120] C. xiaoi No gp60 sequences available from humans C. scrofarum A gp60 typing tool has not yet been developed 3. Knowledge Gaps and Future StudiesThe prevalence and clinical impact of Cryptosporidium and the potential for zoonotic transmission is highest in developing countries, yet relatively few molecular epidemiological studies have been conducted in these areas [1,33]. Large-scale molecular studies are therefore urgently required to explore the extent of zoonotic transmission [1,33]. Relatively little information is also available on the pathogenicity of some of the less common zoonotic species and genotypes, as most reports do not provide detailed clinical information, and this is another important knowledge gap that needs to be addressed. The majority of gp60 subtyping studies conducted to date have relied on Sanger sequencing, which cannot reliably identify the extent of mixed subtypes. For example, a recent study applied the bioinformatic program TIDE to deconvolute gp60 chromatograms generated using Sanger sequencing and identified previously unrecognised mixed subtype infections and has the advantage of also being able to be applied to retrospective data [229]. Similarly, another bioinformatics tool called “CryptoGenotyper” has been developed to read both 18S and gp60 Sanger sequence data, which can also resolve double peaks in mixed infections and increase the accuracy of sequence identification [230]. Amplicon next-generation sequencing (NGS), which can identify low-abundance sequences in mixed infections, has shown that it can identify additional Cryptosporidium gp60 subtypes not identified by Sanger sequencing in various hosts [141,231]. This has important implications for tracing zoonotic transmission as Sanger sequencing may not detect zoonotic species and subtypes that are present at low abundance and therefore incorrect conclusions regarding zoonotic transmission may be made. Future studies need to examine the extent of intra-sample diversity more closely to better understand zoonotic transmission dynamics.The application of CRISPR-Cas (Clustered Regularly Interspaced Short Palindromic Repeats-CRISPR associated protein) genome editing technology to Cryptosporidium has been a major advance in the ability to genetically modify the parasite [232]. More recently CRISPR technologies have been applied to the detection of infectious diseases by identifying specific pathogen sequences and then cleaving them in order to produce a readable signal [233]. For example, in the HOLMES (one-HOur Low-cost Multipurpose highly Efficient System) detection platform, if a target DNA is present, a Cas protein called Cas12a forms a complex with the guide CRISPR RNA, which then binds the target DNA and cleaves a non-target ssDNA reporter in the system, resulting in a fluorescent signal [233]. A lateral flow biosensor has been developed which incorporated isothermal amplification of Cryptosporidium DNA with this CRISPR detection method to identify C. parvum sequences belonging to the gp60 IId subtype family (from both humans and cattle), with high sensitivity and specificity [234]. This type of system has the potential to be employed in field situations to rapidly identify zoonotic subtypes and transmission hot-spots for more targeted analysis.A more complete understanding of the molecular epidemiology of Cryptosporidium and zoonotic transmission dynamics can be gained from WGS studies, which are also increasingly being used to identify gp60 loci in divergent species and genotypes [235]. Currently, WGS data has been generated from the following zoonotic species and genotypes; C. parvum, C. hominis, C. muris, C. meleagridis, C. ubiquitum, C. andersoni, C. tyzzeri, C. cuniculus, C. viatorum, C. felis, C. canis, C. bovis, C. xiaoi, Cryptosporidium skunk genotype and chipmunk genotype I [44,45,46,47,48,49,50,236]. Comparative WGS has identified a variety of invasion-associated proteins including mucin glycoproteins, insulinase-like proteases and MEDLE secretory proteins, which differ between species with narrow and broad host ranges and hence may play a role in host specificity [76,237,238,239]. However, many challenges remain including the development of a universal MLST system that uses polymorphic markers distributed across all chromosomes [236]. Future studies need to conduct WGS on a wider range of species and genotypes and to examine the biological functions of diverse genes in the Cryptosporidium genome to gain a better understanding of the genetic determinants of host specificity.WGS is expensive and another alternative may be to mine existing raw untargeted (shotgun) sequencing metagenomic data deposited in public databases generated for other purposes (such as the microbiome of humans and animals) [240,241]. For example, one study successfully mined 12 large data sets on the human gut microbiome to examine the population genomics of Blastocystis [242]. As a proof of principle, C. parvum (chromosome 6) was identified and mapped from data sets derived from water concentrates and calf microbiomes, although further bioinformatic pipeline optimisation is required [243]. With the rapid expansion of publicly available metagenomes, this approach offers a cost-effective and unprecedented opportunity to unravel the extent of zoonotic transmission and environmental contamination of Cryptosporidium across wide geographic areas.4. ConclusionsOur understanding of the extent of genetic diversity and zoonotic potential of Cryptosporidium species and genotypes has been greatly enhanced by the development of molecular detection and typing tools, particularly the gp60 typing tool. The six most common species identified in humans are C. hominis, C. parvum, C. meleagridis, C. canis, C. felis and C. ubiquitum, yet the extent of zoonotic transmission in many of these species is still not clearly understood. Although primarily anthropologically transmitted, C. hominis is increasingly detected in animals, with evidence that it may be endemic in marsupials (IbA10G2 subtype) and equines (Ik subtype family). Cryptosporidium parvum is the most important zoonotic species, with IIa and IId the major zoonotic subtypes. Evidence to date supports the zoonotic transmission of C. meleagridis (IIIbA26G1R1b and IIIbA22G1R1c subtypes), C. canis (XXa1 and XXa4 subtypes), C. felis (XIXa and XIXb subtypes) and C. ubiquitum (XIIa, XIIb, XIIc and XIId subtype families). Cryptosporidium viatorum was originally described as a human species but rodents are now thought to be the major host species with subtypes (XVaA3g, XVaA3h and XVcA2G1) common to both humans and rodents. The rodent-derived chipmunk genotype I is considered an emerging human pathogen in the US and Sweden with evidence supporting the zoonotic transmission (subtype XIVaA20G2T1) between humans and squirrels. Cryptosporidium cuniculus in rabbits is another important zoonotic species, as it was responsible for a waterborne outbreak in the UK, with evidence supporting zoonotic transmission of multiple subtypes. Cryptosporidium erinacei in hedgehogs is a minor zoonotic species in most areas except New Zealand and the limited gp60 subtyping conducted to date has yet to support zoonotic transmission. Similarly, identical skunk genotype gp60 subtypes have not been identified in skunks and humans. Two studies have identified identical C. fayeri subtypes (IVaA9G4T1R1 and IVgA10G1T1R1) in humans and kangaroos in Australia. Gastric species (C. muris and C. andersoni) seem to lack the gp60 gene but gp60 typing tools that are still lacking for C. suis, C. scrofarum, C. ditrichi, C. occultus and the mink genotype are essential to better understand their transmission dynamics. The application of WGS and amplicon NGS in future studies are important for more accurately tracking transmission and understanding the mechanisms behind host specificity.

animals : an open access journal from mdpi

10.3390/ani12040508

PMC8868372

A severe trauma of unknown aetiology was suspected as the cause of death in an adult female Sowerby’s beaked whale found floating dead in the Canary Islands in December 2016. Many bruises in the skin and muscles (contusions) were observed in the chest wall and bone fractures, mainly located in the mandible and ribs. The broken rib bones also affected thoracic muscles, which escaped into the blood circulation once ruptured, reaching several organic locations, including the lungs, where they became trapped within the small lumen of pulmonary blood vessels, leading to a systemic and pulmonary skeletal muscle embolism. An embolism occurs when a piece of intravascular internal or foreign material obstructs the lumen of a blood vessel, starving tissues of blood and oxygen. An embolism necessarily needs cardiac function, indicating a survival time after trauma. This case report aimed to include the diagnosis of skeletal muscle embolism as a routine tool to determine if the traumatic event occurred before or after death. This is especially valuable when working with dead animals because no other evidence of traumatic injury may be recorded if carcasses are in advanced decay.

An adult female Sowerby’s beaked whale was found floating dead in Hermigua (La Gomera, Canary Islands, Spain) on 7 December 2016. Severe traumas of unknown aetiology were attributed, and the gross and microscopic findings are consistent with catastrophic trauma as a cause of death. Rib fractures affected the intercostals, transverse thoracis skeletal muscles, and thoracic rete mirabile. Degenerated muscle fibres were extruded to flow into vascular and lymphatic vessels travelling to several anatomic locations into the thoracic cavity, including the lungs, where they occluded the small lumen of pulmonary microvasculature. A pulmonary and systemic skeletal muscle embolism was diagnosed, constituting the first description of this kind of embolism in an animal. The only previous description has been reported in a woman after peritoneal dialysis. Skeletal pulmonary embolism should be considered a valuable diagnostic for different types of trauma in vivo in wild animals. This is especially valuable when working with decomposed carcasses, as in those cases, it is not always feasible to assess other traumatic evidence.

1. IntroductionSeveral types of pulmonary embolism have been described in human and animal medicine, but generally, they can be grouped as thrombotic and non-thrombotic [1,2,3,4]. The latter is defined as the embolisation to the pulmonary vasculature of a wide range of non-thrombotic agents, including endogenous and exogenous material. Skin, fat, bone, bone marrow, amniotic fluid, trophoblast, decidua, brain, liver, fat, and bile are among the biological endogenous extravascular tissues previously reported to embolise the pulmonary vessels [5]. In addition, a single case of skeletal muscle pulmonary embolisation has been described in the literature, which occurred after peritoneal dialysis in a 75-year-old woman [6]. As in that case, most of these embolic tissue fragments usually happen due to a severe musculoskeletal and/or soft tissue trauma [7]. After trauma, cells from the damaged tissue enter the bloodstream through ruptured vessels (veins and venules) in the injury or fracture site, and they may reach the lung circulation, where they become trapped within the pulmonary microvasculature [8]. However, several mechanisms have been suggested for this process: emboli can be forced to enter the vessels or travel from an extravascular place inside the venous system due to a pressure gradient and/or a perforated vessel wall. The exact mechanisms of transvascular migration remain unclear [5,9].Imaging diagnosis, such as computed tomography, can become difficult in cases of microemboli, in which the diagnosis is then based on parenchymal findings. A definitive diagnosis is reached by microscopic emboli visualisation, although indirect signs of obstruction include pulmonary oedema, infarction, empyema, or pulmonary hypertension [5,10]. The detection of endogenous biological origin embolism substances in a corpse indicates survival after trauma [7].2. Materials and MethodsAn adult female Sowerby’s beaked whale (Mesoplodon bidens) (total body length of 470 cm) was found floating dead in Hermigua (La Gomera, Canary Islands, Spain) on 7 December 2016 (CET 827). Presumably, this same animal had been previously sighted alive, with a calf, close to a rocky coast. After floating in the sea for a few days, the animal was finally towed to shore, where a necropsy was performed according to standard procedures [11,12,13] (Figure 1).Biological parameters (stranding epidemiology (type, location, and date), life history data (species, age category, and sex)), body condition, and decomposition code were recorded. Age categories (neonate, calf, juvenile, and adult) were based on total body length and gonadal development [11,12,13]. Body condition (very poor, poor, fair, or good) was estimated based on anatomical landmarks [14]. Five codes of conservation condition were established: Code 1 (extremely fresh carcass, as an animal that has recently died or been euthanised), Code 2 (fresh carcass), Code 3 (moderate decomposition), Code 4 (advanced decomposition), and Code 5 (mummified or skeletal remains) [11,12,13]. During the necropsy, lesions were described and photographed.2.1. Evidence of Ethical ApprovalPermission for the management of stranded cetaceans was issued by the environmental department of the Canary Islands Government and the Spanish Ministry of Environment.2.2. Gross, Histological and Histochemical AnalysesDuring necropsy, formalin-fixed samples for histopathologic processing were collected from selected tissues. The fixed tissue samples were then trimmed, routinely processed, embedded in paraffin, sectioned at 5 μm, and stained with haematoxylin and eosin (HE) for examination by light microscopy. Additionally, several tissue sections were cut at 4 μm for Phosphotungstic Acid Hematoxylin (PTAH) staining, as previously described [15]. PTAH is ideal for demonstrating striated muscle fibres, stained in blue, in contrast to collagen, stained in red.2.3. Immunohistochemical AnalysesTo confirm the presence of skeletal muscle fibres inside blood vessels, immunohistochemical stain with α-actin (HHF35; mouse anti-human monoclonal; Enzo Biochemical, New York, NY, USA) was employed. Briefly, 3 µm-thick sections were deparaffinised and rehydrated through a series of graded alcohols. Endogenous peroxidase activity was blocked by incubation in 3.0% hydrogen peroxide (30%) in methanol for 30 min. Protease digestion at 0.01% for 10 min was used for antigen retrieval. The sections were blocked with 10% normal rabbit serum in phosphate-buffered saline (PBS) for 30 min. The primary antibody (1/200 dilution) was incubated overnight at 4 °C. The sections were washed in PBS and incubated for 30 min with a 1:20 biotinylated rabbit anti-mouse secondary antibody. The label used was an avidin–biotin–peroxidase complex (Elite ABC kit, Vector Laboratories, Burlingame, CA, USA), following the manufacturer’s instructions. 3-Amino-9-ethyl-carbazole (AEC, Sigma Chemical Co., St. Louis, MO, USA) was used as the chromogen. Slides were pre-incubated in acetate buffer for 10 min and incubated for 1 min with a filtered solution of AEC (0.025 g AEC was dissolved in 5 mL N, N-dimethylformamide, 70 mL acetate buffer, and 75 µL hydrogen peroxide at 30%). Sections were rinsed in tap water, counterstained with Mayer’s haematoxylin, and mounted with an aqueous medium. Nonimmune homologous serum was used as the negative control. Anti-muscle-specific actin recognises alpha and gamma isotypes of all muscle groups. Cross-reactivity in cetacean skeletal muscle has been previously demonstrated [16]. Anti-von Willebrand factor (RRID: AB_2811207, rabbit polyclonal; Thermo Fisher Scientific) antibody was used as previously described [17] and visualised using a Chemmate Dako EnVision detection kit (Glostrup—Denmark).2.4. Molecular Virological AnalysesMolecular analysis for screening of herpesvirus and cetacean morbillivirus (CeMV) were performed in selected frozen tissues (i.e., skin, skeletal muscle, lung, liver, mesenteric lymph node, kidney, brain, and spleen). Herpesvirus DNA was detected by conventional nested PCR using degenerate primers designed to amplify a region of the DNA polymerase gene [18]. Molecular detection of cetacean morbillivirus was performed by a one-step reverse transcription-quantitative PAN RT-qPCR method based on SYBRN® Green dye that successfully detects different CeMV strains and amplifies a 205 bp partial region of the P gene [19].3. Results3.1. Gross FindingsThe postmortem examination revealed several haematomas in the dorsal and ventral aspects of the head and thorax (Figure 2A), affecting the hypodermis, the fascia, and the muscular tissues. The skeletal muscles (epaxial and hypaxial skeletal musculature) were extremely friable, particularly close to their attachment to the thoracic vertebrae (Figure 2B). In addition, locally extensive haemorrhages in epaxial muscles near costochondral joints and within the thoracic wall (affecting the intercostal muscles) were observed. Multiples fractures were found in the right ribs (Figure 2B), which were simple in nature with a slight degree of external bone displacement into adjoining soft tissues. Simple fractures with bone displacement were also observed in both mandibles. A moderate haemoabdomen and systemic vascular changes (congestion, haemorrhages, and clots) were also present (Figure 2C,D). The lungs were collapsed with subpleural emphysema, most remarkably in the right upper lobe.3.2. Histopathological and Histochemical FindingsUpon histopathologic and histochemical examinations, the main findings were moderate to severe skeletal muscle segmentary hypercontraction and myonecrosis with the presence of intravascular myofibers in different locations, specifically in the rete mirabile (Figure 3), subcapsular sinus in the prescapular lymph node, lymphoid capsule in pulmonary lymph node (Figure 4), thymus capsule (Figure 5) and lung (Figure 6). Severe haemorrhages were found in the prescapular lymph node and rete mirabile. In the latter, these haemorrhages were associated with disrupted tissue intermixed with free adipose tissue and necrotic muscle fibres, as well as vessels wall damage. Myofibers showed blue staining by PTAH, turning to red-purple when disrupted (degenerated), confirming myofibers both free in the rete mirabile and inside blood vessels.3.3. Immunohistochemical FindingsMyofibers were also identified by α-actin antibody displaying different levels of immunostaining (weak/moderate/strong) (Figure 4, Figure 5 and Figure 6). The reduction in immunostaining intensity observed in some of the embolised skeletal muscle fibres could be due to the loss of skeletal muscle integrity (degeneration). As expected, α-actin immunolabeling was also observed in the smooth muscle cells of vessel walls. In addition to the anti-von Willebrand factor antibody (Figure 5C), this immunolabeling substantiates that muscle tissue was placed in a vessel.3.4. Molecular ResultsAll the tested samples were negative to HV [20] and CeMV (under review).4. DiscussionSevere pulmonary and systemic gas and fat embolism, compatible with decompression sickness (DCS), has been described in beaked whales related to mid-frequency active sonar during military manoeuvres [21]. In those cases, gas bubble-associated lesions and fat embolisms were observed in the vessels and parenchyma of vital organs. A modified diving behaviour and/or physiological stress reaction overriding the diving response in response to sonar exposure has been postulated as an explanation to trigger gas bubble formation in those animals. Nitrogen can dissolve easily into the bloodstream and tissues under high pressure at great depths. Still, it can come out of solution when the sum of partial pressures exceeds the environmental pressure, forming gas bubbles in the blood (gas emboli) and body tissues, causing local and systemic damage [22]. In contrast, pulmonary and systemic fat emboli are usually the consequence of direct injury (trauma) to fat depots [23,24,25], as has also been proved for sperm whales that have died due to vessel collisions [26].Beaked whales are extreme deep divers, a condition that renders them prone to a higher susceptibility to suffering from DCS-like syndromes [22]. Here, we present a pulmonary and systemic muscle embolism in a Sowerby’s beaked whale. Gross and microscopic lesions were consistent with massive trauma, which would have been sufficiently severe to account for the death of this animal. The nature of the trauma in the Sowerby’s beaked whale of our study remains undetermined. Still, it should have been strong enough to cause several right ribs fractures and affect their associated thoracic muscles (intercostals and transverse thoracis origin and insert on the ribs) [27] and the rete mirabile. The rete mirabile is a complex structure in which an artery branches into multiple smaller vessels that eventually reconstitute one (or a few) larger vessels, thus creating a direct continuation of the artery that generated the rete [8,28]. The cervicothoracic rete mirabile is supplied by the brachiocephalic trunk and the intercostal and internal thoracic arteries [29]. It continues with the cervical rete mirabile, which enters the occipital foramen with the spinal meningeal arteries for cerebral blood circulation supply. The cervico-thoracic rete in marine mammals has afferent and efferent limbs (ambiocentric type) [28].The force of the traumatic impact in the right thoracic wall of the Sowerby’s beaked whale of our study presumably forced muscular and fat fragments into the vessel channels. Skeletal muscles could arise from the thoracic musculature, while fat (not histochemically demonstrated in our study) could be released at fracture sites or may be derived from adipocyte-rich soft tissue injuries [8,30], as it is present in the histologically normal stroma of the rete mirabile [28,29]. These fragments are then likely to have been transported by the blood and lymphatic circulation, during the agonal period of survival, towards several anatomic locations, including the lungs, where they occluded the small lumen of pulmonary blood vessels [31]. The lung blood vessels receive the venous return from all the peripheral organs; thus, emboli arising in any part of the systemic circulation necessarily pass through the lungs [32]. Embolisation occurs most commonly in the lungs, and for that reason, the lung is considered the target organ for emboli detection. The thoracic scan is the method of choice in cases of suspected embolism [33].The thorax has been described as the most affected body region with fractures involving the ribs in a retrospective study of traumatic intra-interspecific interactions in stranded cetaceans in the Canary Islands [34]. However, in cases of advanced autolysis carcasses, it is challenging to distinguish if the trauma occurred ante- or post-mortem. The presence of muscle and fat emboli within the lungs in our case constitutes evidence of antemortem injury, as cardiac function is needed, even for a short period, to disseminate these droplets to reach the lung circulation [35].5. ConclusionsSkeletal pulmonary embolism should be considered a useful diagnostic criterion for different types of trauma in vivo in wild animals. This is especially valuable when working with decomposed carcasses, as assessing other traumatic evidence is not always feasible. This case is, to our knowledge, the first description of skeletal muscle embolism in an animal. The only previous description was reported in a woman.

animals : an open access journal from mdpi

10.3390/ani11092579

PMC8466328

Leishmaniosis is an important zoonotic protozoan disease. Leishmania infantum is a protozoan species that accounts for the majority of cases in the Mediterranean. In this study, we analyzed the prevalence of infection in different dog breeds from Ibizan Island. Our results showed that the Doberman Pinscher and Boxer breeds present a higher prevalence of infection, and the relationship between antibodies’ serum titer and staging of disease was confirmed. Differences between age and sex were not found.

Leishmaniosis is an important zoonotic protozoan disease primarily spread to the Mediterranean region by Leishmania infantum, the predominant protozoan species, which accounts for the majority of cases. Development of disease depends on the immune response of the definitive host and, predictably, their genetic background. Recent studies have revealed breed-typical haplotypes that are susceptible to the spread of the protozoan parasite. The objective of this study was to analyze the prevalence of leishmaniosis on a Mediterranean island and determine the relationship between disease prevalence and breed. In addition, information on seropositive animals was recorded to characterize animals affected by the disease. To study the prevalence, a total of 3141 dogs were analyzed. Of these, the 149 infected animals were examined for age, sex, antibody titer, and disease stage. We observed a prevalence of 4.74%, which varied between breeds (p < 0.05). The Doberman Pinscher and Boxer breeds had the highest prevalence of leishmaniosis. Significant differences were observed between breeds with common ancestors, emphasizing the important genetic component. Finally, regarding the characterization of seropositive animals, the distribution is similar to other studies. We discovered a relationship (p < 0.05) between the number of antibody titers and the clinical disease stage, which was also present in Leishmania infantum, suggesting that the development of the disease depends on the humoral or Th2 immune response with ineffective antibodies.

1. IntroductionLeishmaniosis is a parasitic disease caused by different genera of Leishmania, including Leishmania infantum (L. infantum), the most prevalent causal agent of leishmaniosis in the Mediterranean area stretching as far as the Iberian Peninsula. This zoonotic disease is endemic to 88 countries and is considered the most relevant vector-borne disease in the Mediterranean, affecting between 63% and 80% of the domestic dog population [1,2,3,4,5]. Akohundi et al., (2016) identified 53 different species of the Leishmania genera and more than 800 vectors capable of transmitting the infection stadium of the parasite (promastigote) to the host [6]. Although domestic dogs are the definitive hosts, different studies have documented L. infantum in other mammalian species, including horses, cats, pigs, wild rodents, hares, wild canids, and humans [7,8,9,10,11,12,13,14]. The host immunity response determines the infection severity of L. infantum, which can be mild, moderate, severe, or very severe [15]. Solano–Gallego et al. (2009) reported that the most common clinical signs of infection by L. infantum are skin lesions, generalized lymphadenomegaly, progressive weight loss, muscular atrophy, exercise intolerance, decreased appetite, lethargy, splenomegaly, polyuria and polydipsia, ocular lesions, epistaxis, onychogryphosis, lameness, vomiting, and diarrhea [15]. In terms of spatial distribution, the seroprevalence of L. infantum varies in different Spanish areas, ranging from 3.7% to 34.6% in the north and south of the country, respectively [16,17]. There are some studies on the prevalence of L. infantum on Mediterranean islands [18,19]. Burnham et al., (2020) studied the relationship between anti-Phlebotomus pernicious saliva antibodies in Ibizan hounds and susceptibility to canine leishmaniosis without analyzing seroprevalence data in different dog breeds from the Mediterranean coast of Spain [20]. The global seroprevalence in domestic dogs of the Spanish Mediterranean islands was estimated at approximately 57.1% [21]. Although numerous studies (in both visceral and cutaneous forms) explored a possible genetic influence or resistance to disease in murine and human models, similar studies in dogs are scarce [22,23,24,25,26,27,28]. However, some studies found resistance and/or susceptibility to disease depending on the dog breed, thereby influencing the pathological process [29,30]. Recently, Vasconcelos et al., (2019) published a review on dogs’ genetic resistance to infection by Leishmania spp., particularly the Ibizan Hound breed, but did not obtain clear evidence [31]. These data complement Solano-Gallego et al.’s study (2000), which demonstrated that the Ibizan Hound presents a greater immune response than dogs of other breeds [30].The clinical pathogeny of L. infantum infection is a multifactorial process, where the activation of Th1 or Th2 pathways depends on the host–pathogen interaction. The cellular immune response or Th1 pathway eliminates amastigotes within the macrophage more efficiently since stimulation by cytokines, mainly interferon-gamma (IFN-γ) and interleukin-2 (IL-2), cause a de-inhibition of nitric oxide production and deactivate the action of the arginase enzyme. The relationship between nitric oxide and arginase is essential for the progression of L. infantum inside macrophages [32,33,34]. However, recent studies have linked clinical symptoms to an increase in the expression of transforming necrosis factor-alpha (TNF-α), IL-1β, IL-4, IL-10, and IL-12, suggesting that the balance between different cytokines plays an important role in regulating inflammation and clinical presentations of the disease [35,36]. Hosts who activate the humoral immune response or Th2 pathway develop the disease with greater prevalence and severity since the antibodies are ineffective and stimulation by cytokines is based on the production of IL-10. This situation favors the development of the parasite along with the arginase enzyme and synthesis of polyamides [37]. Studies in vitro have demonstrated that blocking the activity of receptor IL-10 may be a new effective treatment [38,39].Some authors have associated this differential immune response to the genetic backgrounds of these canine breeds. Additionally, different canine breeds present different prevalence rates [40,41]. Genes related to the activation of Th1 pathways include IL12RB1, JAK3, CCRL2, CCR2, CCR3, and CXCR6, whereas COMMD5 and SHARPIN are associated with the activation of Th2 pathways [37]. Other genes, such as Slc11c11 or Major Histocompatibility Complex I gene (MHCI), also affect the susceptibility or resistance to L. infantum infection [31]. Batista et al., (2020) associated genetic markers with antibody response in infected dogs [42]. Furthermore, different dog breeds show differences in gene expression. Sánchez-Roberts et al., (2008) postulated that some breeds, such as the Boxer, are most susceptible to infection [43] due to gene expression. These authors conducted a study with 19 breeds (including the Ibizan Hound and Boxer), where they analyzed the presence of polymorphism in the Slc11a1 gene, which is related to autoimmune diseases in humans [44]. They concluded that two of the 24 polymorphisms found in this gene showed greater susceptibility to infection. Six single nucleotide polymorphisms (SNPs) in three different genes were detected in the Beagle breed, correlating with susceptible infection phenotypes [29].Despite several studies regarding the genetic resistance and susceptibility of canine breeds, few have determined the prevalence of this infection in different breeds. Therefore, we aimed to investigate the effect of breed on L. infantum prevalence and characterize the profile of seropositive animals on a Mediterranean island as well as to identify and characterize the profile of seropositive animals.2. Materials and Methods2.1. Ethics ApprovalThe experiments involving animals were conducted according to the guidelines of the Declaration of Helsinki and approved by the Animal Experimentation Ethics Committee of the Universidad Cardenal Herrera CEU, with code 2020/VSC/PEA/0216.2.2. Data and Collection of SamplesThe study was conducted on Ibiza (38°54′31.79″ N, 1°25′58.66″ E; an area of 572 km2), a Mediterranean island located east of Spain. Data and blood samples from 3141 dogs with symptoms of L. infantum infection were collected between September 2020 and May 2021 (Table 1) [2]. Blood samples were obtained in 5 mL EDTA tubes from all symptomatic animals by cephalic venipuncture for serological analysis. Serological analysis was performed for all animals, including those with a positive serological test. The data recovered included prevalence (2 levels), breed (25 levels), sex (2 levels), age (divided into 3 categories: young (<4 years), adult (4 ≤ years < 10), and older dogs (≥10 years)), presentation of disease (3 levels: cutaneous, visceral, or both), and clinical stage. Four levels (Table 1) were assigned to each animal in a physical exam evaluating clinical signs according to Solano-Gallego et al. [15].2.3. Serologic TestsBlood samples were collected from the cephalic vein in 5 mL EDTA tubes. Plasma was obtained and preserved at −20 °C before analysis. Serological testing for L. infantum detected specific antibodies using the indirect immunofluorescence antibody test (IFAT), which was conducted by an external laboratory, and the IFAT for anti-Leishmania-specific immunoglobulin G (IgG) antibodies (MegaFLUO LEISH®, Megacor Diagnostik GmbH, Hörbranz, Austria). Seroprevalence was calculated as the percentage of dogs testing positive for L. infantum antibodies. Dog serum was considered seropositive with IFAT titer ≥ 1:80, following the manufacturer’s instructions [45,46]. The titer of detected antibodies was classified into 3 levels: low or questionable (<1:100), medium (between 1:100 and 1:400), and elevated (>1:400).2.4. Statistical AnalysisSeroprevalence status was analyzed using the GENMOD procedure of the statistical program SAS (North Carolina State University, USA). Within positive animals, antibody titers were analyzed using a binary (in the case of sex) or multivariable (in the case of breed, age, presentation of disease, and clinical stage) logistic regression model using each factor as a fixed effect in their respective statistical analysis. The statistical significance was set at p-value < 0.05.3. ResultsTable 2 shows the seroprevalence (%) observed in different breeds using the cut-off of an L. infantum antibody titer ≥1:100 to denote seropositivity. The total seroprevalence was 4.74% (149 to 3141 animals studied) and varied between the canine breeds studied (p-value < 0.05). The results showed three groups evaluated for seroprevalence: The first group (Table 2, superscript a) with the lowest prevalence (on average 1.73%) comprised the Pug (0.72%), French Bulldog (1.41%), Maltese Bichon (1.47%), Chihuahua (2.33%), and crossbreeds (2.74%). The second group (Table 2, superscript b), with medium prevalence, contained only the German Shepherd (7.38%). The third group (Table 2, superscript c) with the highest prevalence (34.26% on average) comprised the Pointer (24.49%), Dogue de Bordeaux (33.33%), Great Dane (33.33%), Majorcan Shepherd (33.33%), Weimaraner (33.33%), Boxer (39.13%), and Doberman Pinscher (42.86%) breeds. The animals that do not present statistically significant differences are indicated by more than one superscript. Dogs belonging to the Beagle breed (Table 2, with superscripts a and b) only show statistically significant differences compared with the Boxer breed (Table 2, superscript a) and German Shepherd (Table 2, superscript b). Table 3 describes the seropositive animals based on sex, age, presentation of the disease, and clinical status. The number of seropositive males was higher than females. Younger dogs and adults had higher seropositivity than older dogs, with a cutaneous presentation of the disease in clinical stage II (Table 3).No observable effects were detected between antibody titers, and canine breed, sex, or age. However, antibody titers were recorded in seropositive animals ranging from 1:100 to 1:1280, which is associated with presentation of disease (p-value < 0.05) but not canine breed, sex, or age (p-values of 0.97, 0.46, and 0.43, respectively). Only one dog with a high level of antibodies (>1:400) presented as clinical stage I (p-value < 0.05). Animals with a low antibody titer (<1:100) had clinical stages III or IV (Figure 1).4. DiscussionIn this work, we observed a seroprevalence of L. infantum infection of 4.74%. This result agrees with data observed in other reported studies on Mediterranean areas, where the prevalence was estimated between 5% and 57.1% [16]. In the Balearic Islands, the prevalence was approximately 1% of 813 dogs in 1999 [47] and 40% of 40 dogs ranging from 2011 to 2016 [21]. These relevant differences may be due to the sample size used and asymptomatic animals. Our study and others, which demonstrated high prevalence, only analyzed animals with symptoms, suggesting that the actual prevalence may be higher. Our work is the first to analyze and document more than 3000 dogs from the island of Ibiza. In other Spanish regions such as Northern Spain, the prevalence rate is lower than in Mediterranean regions at approximately 4% [17,45], resembling our data. These differences are likely due to environmental conditions and the vector distribution. Gálvez et al., (2020) analyzed the seroprevalence of L. infantum in dogs and sand fly distribution in Spain. They confirmed that the Balearic Islands are hyperendemic with P. perniciosus as the most abundant sand fly vector [21]. On the one hand, the climate and vegetation of Northern Spain are typically oceanic (warm summers and cool winters) and distribute rain throughout the year. On the other hand, the Mediterranean climate is characterized by mild and rainy winters, and dry and hot summers, with variable autumn and spring seasons (both in temperature and rain). This situation promotes the dispersion of L. infantum transmitted vectors [48]. Although the sample sizes were different between the breeds, the robustness of our statistical analysis indicates that differences exist. However, other authors conducted similar studies with similar sample sizes, and the results obtained are comparable to those of our study [20,49,50]. Regarding canine breeds, the high seroprevalence found in Boxers and Doberman Pinschers, 39.13% and 42.86%, respectively, agree with other study results [2,41,49,51], where authors attributed the genetic susceptibility and high prevalence of the Boxer breed to the TAG-8-141 haplotype. Abranches et al., (1991) conducted a study in Portugal showing high prevalence in the Doberman Pinscher and German Shepherd breeds [51]. However, the latter breed had a moderate seroprevalence in our study. Other authors suggested that the elevated susceptibility of the Boxer and Doberman Pinscher is due to three single nucleotide polymorphisms (SNPs) in the Slc11a1 gene, specifically in the promoter regions T151C, A180G, and G318A, associated with these breeds and others, such as the Cocker Spaniel [40,43]. In our study, Cocker Spaniels had a high prevalence of infection (26.83%). The authors of [40,43] explained that these three SNPs of the Slc11a1 gene are associated with a visceral presentation of the disease, which was similar in the different breeds analyzed. Although there are fewer studies on the presence of these haplotypes and SNPs in all dogs, the breeds with the lowest prevalence (Bichon, French Bulldog, Pug, and Chihuahua) and the highest prevalence (Doberman Pinscher, Boxer, Great Dane, Dogue de Bordeaux, and Boxer) appear to have common ancestors. A cladogram showed that their bootstrap supported more than 70% with over 150,000 SNPs [52,53]. Only the French Bulldog, genetically similar to the Boxer and Dogue de Bordeaux, presents a low seroprevalence of L. infantum infection in our study. Currently, there are no published scientific data to explain (in a genetic manner) the low prevalence of infection by L. infantum in this breed. To describe the infected population, although the cutaneous presentation is the most common in the dogs studied in this work regardless of breed and other factors, other authors found no relationship between sex and prevalence of L. infantum infection in Spain and Portugal [45,51]. Our results show similar seroprevalence in males and females, whereas some authors found a higher prevalence in males than females [54,55,56,57]. Viol et al., (2012) claimed these differences result from breeding males’ preference for surveillance or hunting activities, which exposes them to insect bites and infection by L. infantum [55]. Our data likely show similar seroprevalence because the dogs included in our study were not engaged in surveillance or hunting activities. The same authors [55] also found lower prevalence in adult dogs than younger and older dogs (>7 years), whereas Dantas-Torres et al., (2006) demonstrated a correlation between early age and prevalence [55,57]. Some studies indicated that the L. infantum infection presented two peaks in younger dogs (1–2 years) and a second peak at 7–8 years [58]. However, our results show no statistical difference between ages. According to studies performed in Brazil and Tunisia, the highest prevalence was found in dogs with clinical stage II, i.e., moderate disease [59,60].We observed that antibody titers are associated with disease presentation; however, some authors ascribed the clinical picture to other factors such as eosinophil and alpha-globulin values [61]. We showed that animals with low antibody titers contracted the cutaneous form. These results are concordant with those obtained in a recent meta-analysis indicating that the cutaneous form could lead to more advanced stages of visceral disease [54]. The antibodies’ titer is related to clinical stage according to Martinez-Orellana et al., (2017) [49]. The authors argued that the clinical stage depends on the host’s immunologic response; therefore, the Th2 response produces ineffective antibodies and increases disease. Solano-Gallego et al., (2011) associated the clinical stage with antibody titers. Therefore, the clinical stage system is based on clinical signs, clinical–pathological abnormalities, and serologic status, directly associated with antibody titers of L. infantum [2].5. ConclusionsOur study concludes that the total seroprevalence of L. infantum infection is 4.74% on a Mediterranean island, Ibiza Island. Our results confirm that the seroprevalence of L. infantum is linked to the canine breed but not with antibody titers. The differences in prevalence and breeds may be explained by three polymorphisms present in the Slc11a1 gene. More studies should endeavor to document these interactions. Regarding the profile of seropositive animals, the distribution was similar to other studies. However, a relationship was observed between the number of antibody titers and the presentation of the disease. These data may indicate that the cutaneous presentation is anterior to the visceral one; however, future trials are needed.

animals : an open access journal from mdpi

10.3390/ani13061118

PMC10044088

Ammonia emission in poultry farms leads to environmental or health problems. Additionally, the sustainability and health of poultry depend on the improvement of the surrounding microclimate conditions of birds (ammonia, ambient temperature, heat index, and relative humidity). Therefore, the objective of this study was to investigate the influence of the addition of natural zeolite to the Japanese quail litter on microclimate parameters, growth performance, blood gases, and other blood biochemical indicators in two Japanese quail lines. Both the zeolite-treated group and the selected line performed better than the untreated group and control line in growth performance. The microclimate characteristics were improved by adding natural zeolite in the quail litter, revealing better growth performance and health of birds. Based on this study, it can be concluded that the use of natural zeolite in the Japanese quail litter is effective in reducing the atmospheric ammonia level.

Microclimate parameters (ammonia, ambient temperature, heat index, and relative humidity) surrounding birds affect the production and health status of poultry. Therefore, this study aimed to evaluate the impact of adding natural zeolite to the litter of Japanese quail on improving microclimate parameters and its reflection in growth performance, blood gases, and blood biochemical parameters. A total of 1152 chicks were obtained from the same hatch at the 20th selection generation. Chicks were allocated into two groups based on the litter composition: Group 1: wheat straw as litter (untreated group); Group 2: 80% wheat straw + 20% zeolite (treated group). Each group consisted of 576 chicks: 410 selected line chicks and 166 control line chicks. Significant and favorable effects of the treatment on microclimate parameters during tested periods were found to favor the treated group. Either the treated group or the selected line had significantly better growth performance than the untreated group and control line. Zeolite-treated quails had significantly desirable blood gases and lower blood acidity and serum total iron binding capacity compared to the untreated group. Thus, adding natural zeolite to the litter enhanced the microclimate parameters that improved growth performance, blood gases, and blood biochemical parameters and reduced ammonia emission.

1. IntroductionAnimal welfare scientists have emphasized the importance of controlling microclimate and air quality as relevant aspects for most categories and species of farmed animals for good housing [1]. The surrounding microclimate conditions of poultry, along with standardization, scale, and intensive husbandry processes, are essential elements influencing their growth and limiting their genetic potential [2]. Ammonia (NH3), ambient temperature, and relative humidity are considered among the most important microclimate parameters. Atmospheric NH3 is known as the major aerial pollutant in poultry barns [3], as high NH3 levels can severely affect broiler performance [4], increase disease risk, and jeopardize animal welfare [5]. Abnormal serum biochemical indices, hepatic damage, and reduced performance linked to oxidative stress have been found in broilers exposed to high NH3 concentrations [5,6,7]. Furthermore, the pressure of genetic selection for growth traits may lower the anti-stress capabilities of birds [8]. Therefore, it is crucial to improve the environmental conditions, such as temperature, humidity, harmful gases, and NH3 levels, in poultry houses, especially in the flocks, which are under genetic selection to obtain their best genetic potential. To decrease the adverse influence on the microclimate in the poultry house, it is required to make conditions in which the content of ammonia in the air of the poultry house is low [9]. A new approach to improve microclimate conditions is to use natural zeolite in the litter of poultry [10,11].Natural zeolite has been used in aspects related to poultry production because of its chemical and physical properties. Zeolite is a crystalline aluminosilicate compound classified according to the framework structures’ common features. Using natural zeolite is developed by utilizing the features of gas and water absorption and ion exchange [12]. In this regard, zeolite was found to be among the most useful litter amendments due to the NH3 absorption, ammonium adsorption, and water retention properties of this natural mineral. In poultry, zeolite displayed affirmative results. For broiler chicks, the inclusion of zeolite as a feed supplement exhibited a favorable influence on the performance of broilers and improved their litter quality [13]. Moreover, using zeolite in the litter reduced its moisture and NH3 volatilization [14]. Zeolite can augment the productive performance of birds under different conditions [15,16].It is hypothesized that the inclusion of zeolite in the litter will have a positive impact on growing quail under the genetic selection program. Thus, the purpose of this study was to evaluate the impact of adding natural zeolite to the litter of two Japanese quail lines on improving environmental conditions and its reflection in growth performance, carcass yield, blood gases, and blood biochemical parameters.2. Materials and MethodsThis study was conducted at the Poultry Research Center, Faculty of Agriculture, Fayoum University, Egypt, and the study protocol was approved by the Fayoum University Institutional Animal Care and Use Committee (FU-IACUC), Egypt (approval no. AEC2207).2.1. Birds and Experimental DesignA total of 1152 Japanese quail chicks (average weight ± standard deviation = 8.02 ± 0.77 g) were obtained from the same hatch at the 20th selection generation. The newly hatched chicks were wing-banded using small size plastic bands and brooded on the floor until day 42 of age. As shown in Figure 1, the birds were divided into two groups based on litter composition (5 cm deep): Group 1: wheat straw as a litter (20 kg) (the untreated group); Group 2: 80% wheat straw (16 kg) + 20% zeolite (4 kg) were properly mixed (the treated group). The wheat straw length was about 3.5–12 mm, and the moisture content was 6.0%. Zeolite granules ranged in size from 2.22 mm to 9.8 mm and were generally light-colored and hard. Each group consisted of 576 chicks: 410 selected line (SL) chicks and 166 control line (CL) chicks. Birds in the two groups were housed at a stocking density of 50 chicks/m2. In the same quail house, the chicks were grown in two adjacent rooms (the 1st room was treated with zeolite and the 2nd room was untreated (without zeolite)) that had the same dimensions (4 × 3 × 3 m) and orientation. Each room had two opposite windows perpendicular to the wind’s direction, providing natural ventilation. Each group consisted of selected and control chicks reared together in the same room. The birds had ad libitum access to water and feed. Quails were fed on a diet containing 24% crude protein and 2900 kcal metabolizable energy [17]. Background of the two Japanese quail lines was as follows: The selection criterion is high growth rate during 1–21 days of age according to the estimated aggregated breeding values through generations. The selected line was based on the estimated aggregated breeding values for 20 consecutive generations, while a control line was maintained via random mating without any selection.2.2. Characteristics of the Product (Zeolite)Natural zeolite (Ca, K2, Mg)4 (Al8 Si40) O96·24H2O has a cation exchange capacity of 1.50 mol/kg and it has high affinity and selectivity for NH4+ ions. Zeolite was purchased from A & O trading, Giza, Egypt.2.3. Measurements of MicroclimateMicroclimate parameters were measured once daily (during 1–2 pm) (5 measurements: 4 corners and the center of each room by one sensor for each microclimate parameter/room) when the chicks were aged from 14 to 42 days. All measurements were taken at the bird’s height (approximately 30 cm from the floor) according to Sohsuebngarm et al. [18]. Using Arduino microcontrol board ((Elecrow Company, Shenzhen, China), which has an embedded DHT11 sensor to measure ambient temperature (°C), relative humidity (%), and an MQ135 gas detector sensor for monitoring harmful gases and to detect the level of NOx, alcohol, Benzene, smoke, and CO2 in the poultry house (according to Orakwue et al. [19]). The heat index (HI) was mathematically calculated according to the formula cited by Wicaksono et al. [20]. The NH3 level was measured by an NH3 gas detector (Smart Sensor GM8806, BENETECH, Nanjing, China).2.4. Growth PerformanceBody weights at 14, 21, 28, 35, and 42 days of age (BW14, BW21, BW28, BW35, and BW42, respectively) were individually recorded to the nearest 0.01 g. The growth rate (GR) during different periods was calculated according to the following:(1)Growth rate=[W2−W1]×1000.5[W2+W1] where W2 is the weight at the end of the period, and W1 is the weight at the beginning.2.5. Carcass TraitsOn day 42, forty-eight quails (6♂ SL + 6♀ SL, 6♂ CL + 6♀ CL/each treatment) were randomly taken from the two treatments for the slaughter test. The relative weight of organs (liver, gizzard, heart, and lung) was recorded as a proportion of the slaughter weight. Dressing% was calculated as the following:(2)Dressing %=[Carcass weight+Giblets weight]×100[Bodyweight] where Giblets (g) = the weight of three organs (gizzard, heart, and liver).2.6. Blood Sampling and Laboratory AnalysesForty-eight blood samples (6♂ SL + 6♀ SL, 6♂ CL + 6♀ CL/ each treatment) were collected from the wing vein using sterilized syringes. The blood sample/quail was collected into 2 tubes. The 1st tube contained the whole blood with an anticoagulant to determine blood gases and hematological indices (red blood cells (RBCs), white blood cells (WBCs), packed cell volume (PCV), and hemoglobin (Hb) concentration). Saturated O2, total CO2, and pH were measured in the collected whole blood using a blood gas and electrolytes analyzer (GEM PREMIER 3000, Mumbai, India). The 2nd tube (without an anticoagulant) was centrifuged (3000 rpm/10 min) to separate the serum that was saved at −20 °C until the time of laboratory analyses.Serum total protein, albumin, urea, creatinine, NH3, alanine aminotransferase (ALT), aspartate aminotransferase (AST), iron, and total iron-binding capacity (TIBC) were determined using colorimetric methods and Biosystems A25 (BioSystems A.S, Barcelona, Spain). All blood biochemical indices were measured using commercial diagnosing kits (produced by Spectrum Diagnostics Company, Giza, Egypt).2.7. Statistical Analyses2.7.1. Measurements of MicroclimateData were analyzed using GLM (SAS [21]), and measurements of microclimate were subjected to a one-way analysis of variance with treatment as a fixed main effect as follows:Yij = μ + Ti + eij(3) where Yij = the observations for a trait; μ = the overall mean; Ti = the fixed effect of ith treatment; and eij = the random error term.Group size: each group consisted of 576 chicks: 410 selected line (SL) chicks and 166 control line (CL) chicks.2.7.2. Growth, Blood Constituents, and Carcass TraitsThe recorded data of growth, blood constituents, and carcass traits were analyzed by PROC MIXED (SAS [19]) to calculate the treatment-, line-, and sex-specific means using the following model:Yijklm = µ + ai + Tj + Lk + Sl + Tj × Lk + Ti × Sl + Lk × Sl + Tj × Lk × Sl + eijklm(4) where Yijkim: the observation for a trait, µ: the overall mean, a: the random additive genetic effect of the ith animal, T: the effect of jth treatment, L: the effect of kth line, S: the effect of lth sex, Tj × Lk: the effect of the interaction of the jth treatment with the kth line, Tj × Sl: the effect of the interaction of the jth treatment with the lth sex, Lk × Sl: the effect of kth line with the lth sex, Tj × Lk × Sl: the effect of the interaction of the jth treatment with the kth line with e lth sex, and eijklm: is the random error term; the random variable was the birds within each line.3. Results3.1. Microclimate MeasurementsThe effects of treatment on NH3, harmful gases, HI, temperature, and relative humidity% during different tested periods are shown in Table 1. It was clear that there are significant and preferable effects of the treatment on NH3 and harmful gases during different tested periods favoring the treated group compared to the untreated group. Furthermore, HI and relative humidity% during different tested periods were lower for the treated group than the untreated group, except for the period from 21–28 days of age. Despite the increases in NH3 and harmful gases with increasing age, the air quality remained in favor of the treated group compared to the untreated group (Table 1).3.2. Growth PerformanceThe effects of treatment, line, sex, and their interactions on body weight (BW) at different ages are shown in Table 2. Quails of the SL had significantly heavier BWs than those of the CL at all ages. Except for BW14, the treated group had significantly heavier BWs compared to the untreated group, also sex significantly affected BWs in favor of females compared to males except at 14 days of age. Additionally, Figure 2 displayed that the SL and CL treated with zeolite had heavier BW than the SL and CL untreated (without zeolite). As a result, any significant differences observed between the treated and untreated groups’ performance can be attributed to the treatment itself (or addition of zeolite).As shown in Table 3, the SL had a significantly faster growth rate (GR) than the CL during different tested periods of growth. Quails of the treated group had significantly higher GR14–21 and GR14–42 than the untreated group. Females had significantly faster GR28–35, GR35–42, and GR14–42 than males. Additionally, there were significant effects of treatment by line interaction on BW42 and GR21–28.3.3. Blood GasesTable 4 showed the effects of treatment, line, sex, and their interactions on blood gases. Zeolite-treated quails had significantly and preferably higher saturated O2 and lower total CO2 and NH3 compared to their counterparts in the untreated group. Regarding sex effect, males had significantly lower NH3 than females, whereas the line had no significant effects on blood gases. Both the saturated O2 and total CO2 showed significant impacts of the treatment by line interaction.3.4. Blood Biochemical IndicesThis study investigated the effect of microclimate parameters on the hematological indices, serum profile of proteins, and lipids of quails through the use of zeolite. The results in Table 5 and Table 6 show the effects of the treatment, line, sex, and their interactions on the performance of some blood biochemical parameters. Regarding the treatment effect, blood pH was significantly declined in the treated group compared with the untreated group. The SL had significantly lower blood creatinine and albumin values than the CL as affected by the line effect. Sex significantly affected both ALT and AST where males had lower estimates than females. Moreover, significant effects of treatment by line interaction appeared on ALT, creatinine, urea, and albumin.3.5. Iron Profile and Haematological ParametersSignificantly lower serum iron levels and a rise in TIBC levels were seen in the untreated group than in the treated group. However, according to the statistical analysis, there were no significant differences in Hb, PCV, RBCs, and WBCs between the two groups. The line significantly affected PCV with a higher value for the CL than the SL. Males had significantly higher Hb and PCV than females as shown in Table 6.3.6. Carcass TraitsThe performance of some carcass traits as affected by treatment, line, sex, and their interactions is presented in Table 7. Treatment had a significant effect on gizzard % favoring treated quails compared to those from the untreated group. Additionally, the previous results refer to numerically and preferably higher dressing % for the treated than the untreated groups. However, the line significantly affected the dressing % and gizzard %, favoring the SL compared to the CL. Females had significantly higher liver % than males. Effects of significant treatment by line interaction were found for liver %.4. DiscussionPoultry house air quality contributes to sustainability due to its influence on the health of birds and humans working there, and as a factor affecting the environment directly [22]. Thus, controlling the microclimate conditions is one of the innovative solutions in public health and environmental protection necessary to encourage the development of modern poultry farming. Measurements of the microclimate during different tested periods were significantly influenced by the addition of zeolite to the litter. This may be due to the fact that zeolite has large porosity and surface area, which have a beneficial effect on the absorption of liquids, such as water, NH3, organic liquids, and gases, such as volatile organic compounds and hydrocarbons [23]. Additionally, Schneider et al. [14] stated that zeolite can easily absorb toxic gases from the air, such as ammonia. Finally, the current study developed a new litter amendment composed of 80% wheat straw plus 20% zeolite which may be used as an effective solution for reducing NH3 emissions from the quail litter and improving microclimate conditions. Moreover, zeolites have unique chemical and physical characteristics that can be used as a part of the litter in poultry production to mitigate the pollution. Thus, improved microclimate conditions can boost bird comfort and reduce the hazard of ammonia toxicity in the birds’ house [24], leading to the improvement of productive performance. This is of great importance for the biosafety and hygiene of production.In the current study, the growth performance of the treated group was significantly enhanced over the untreated group by the addition of zeolite to the litter. Zhu et al. [25] and Wei et al. [26] reported similar trends in ducks and chickens. Basha et al. [27] mentioned that the natural zeolite addition in the litter in the broiler house increased the productive performance of chicks. This may be attributed to higher hematic NH3 levels, which can accelerate the detoxification of NH3 in the muscle, brain, and liver, which is a very energy-intensive process [26], which translates into a reduction in the energy required for growth and production, which has an adverse impact on growth performance. Zeolite treatment of the litter resulted in significant differences between the treatment averages in final body weight and average daily weight [11]. Similarly, Eleroğlu and Yalçın [28] revealed statistically significant differences in broiler live weight, which were consistent with the findings of the current study. This may be due to the structure of zeolite allowing for ion exchange, molecular “sieving,” absorption, dehydration, diffusion, catalysis, and reversible dehydration, which improve microclimate conditions surrounding the quails. The litter treatment using zeolite as a management practice to boost broiler performance and reduce ammonia emission was able to improve the performance of birds by having a negligible impact on the main odor-producing culprits and microbial activities in the litter [11]. In contrast, Altan et al. [29] used zeolite to treat litter and reported insignificant differences in the live BW gain of birds among the treatments.The effects of sex and selection on the growth performance parameters were examined in the current study, the literature revealed differences between lines of various genetic backgrounds. The selection is a successful approach for quail performance and genetic enhancement, which may result in some alterations of physiological and metabolic processes that affect characteristics related to growth or egg production [30,31,32,33]. Since the growth trends of the sexes varied, females had significantly better BW and GR than males. According to the results of Narinc et al. [34] and Elkomy et al. [35], females were consistently heavier than males for BW at various ages. Furthermore, there was a noticeable sexual dimorphism in Japanese quail BW, favoring females over male counterparts as a result of male sexual activity caused by hormonal alterations. Furthermore, Taskin et al. [36] found that across selection generations, sex was a substantial source of variance for BW at all ages. On the contrary, Mahmoud et al. [37] found an insignificant sex effect on BW at all ages. In addition, treated quails showed significantly faster GR14–21 and GR14–42 than the untreated group; Zhu et al. [25] and Wei et al. [26] reported similar trends in ducks and chickens.Improved microclimate conditions can boost the blood physiology of birds. Determining numerous blood physiology parameters such as blood gas concentration and blood biochemical indices may indicate the suitability of housing for birds. The zeolite-treated group had significantly better blood saturation levels of O2, CO2, and NH3 when compared to the untreated group. These results are in harmony with the findings obtained by Wei et al. [38,39] and Zhu et al. [25], who indicated that exposure to high NH3 levels had a significant impact on plasma NH3 levels in chickens and ducks. Zeolite can actively adsorb carbon dioxide, ammonia, mercaptans, and hydrogen sulfide, remove toxins, and improve immunological responses [15], thus the inclusion of zeolite in the litter can decrease blood gases (saturated oxygen, total carbon dioxide, and ammonia).The inclusion of zeolite in the litter can enhance the chemical, microbiological, and physical integrity of the litter, and consequently, can boost the performance, hygiene, and ambience of poultry [40], improving the physiological status of birds. The recent findings may validate that zeolite litter treatment decreased the health risks for quails, which was reflected in some blood biochemical traits. The blood pH of the treated group significantly declined compared to the untreated group. A similar trend that blood pH was higher for the untreated group than the treated group was reported by Borges et al. [41] and Wasti et al. [42], indicating the occurrence of slight alkalosis which may have resulted from the higher panting rate in birds of the untreated group during challenging higher heat stress. The current study indicated that the serum concentrations of AST, ALT, blood urea, creatinine, and protein profile did not differ significantly between the two treatment groups; this may be due to the fact that atmospheric NH3 did not exceed the harmful level in the two treatment groups which contradicted those results of the previous studies reported by Zhu et al. [25], Chen et al. [43], and Lu et al. [44]. Furthermore, Zhang et al. [7] reported that birds exposed for an extended period to high concentrations of atmospheric NH3 may have chronic liver and renal damage. Blood gases and serum biochemical examinations indicate the affirmative influence of the addition of zeolite in the litter of Japanese quail. Currently, adding zeolite to the litter for the genetically selected quail based on the growth rate could be proposed. However, it is equally significant to continue this kind of research due to its scarcity in science. There was evidence of significant sex-related differences in ALT and AST estimates, demonstrating that females had higher values of AST and ALT than males, which may be explained by the physiological changes in metabolism in female birds due to maturation and attaining egg laying [45]. Similar findings were reported by Udoh et al. [46]; however, Scholtz et al. [47] reported significant differences in ALT and AST with males exhibiting lower estimates than females.Iron is a necessary constituent of Hb, which is the O2-carrying protein in RBCs [48]. After improving microclimate conditions, serum iron concentration was significantly increased in birds treated with zeolite compared with the untreated birds. The untreated group has shown a significant decrease in serum iron associated with the increase in TIBC level, indicating that birds of this group were experiencing some stressful conditions such as psychological stress which has been shown to activate the hypothalamic–pituitary–adrenal axis system, which resulted in elevated levels of adrenocorticotropic hormone in the blood and decreased levels of serum iron, hepatic iron enrichment, and the enrichment of iron overload [49,50]. This study found increases in Hb, PCV, and RBCs together with decreased blood oxygen and increased carbon dioxide saturation in the untreated group, which may be associated with the increased metabolic activity required to fulfil the energy requirements for growth, especially when birds are kept under challenging inhalation conditions, including higher atmospheric NH3 and harmful gases levels along with higher HI measurements. Similar findings were reported by Olanrewaju et al. [51] and Asif et al. [52]. The enhancement in the hematological parameters in the zeolite-treated birds can be attributed to the role of zeolite in improving air quality in the birds’ housing. There is no available literature on using zeolite in the litter for quails and its influence on serum iron, TIBC, and blood hematology; consequently, further investigations on this point are needed.The current experiment indicated that carcass traits were decreased in the untreated group, which may be associated with suffering from some stressful conditions. This is in agreement with some previous studies [5,53]. The current results indicated that improving microclimate conditions during the growth period non-significantly boosted the dressing % at the marketing age of Japanese quail. The present study indicated that the inclusion of zeolite in the litter resulted in a favorable impact on gizzard, possibly due to the effect of the birds in this group eating the zeolite in the litter. Similarly, the addition of zeolite to the diet significantly increased the relative weight of edible organs including the gizzard [54]. In broiler chickens, Banaszak et al. [55] indicated a significantly higher body weight and carcass weight in the treated group with aluminosilicates in feed or litter compared to the control group. Furthermore, the aluminosilicates-treated group had significantly high weights of wings and neck with the skin but did not affect the dressing percentage. Therefore, future research should deeply address how the zeolite group showed better growth performance but not a significant increase in the dressing percentage.The present results showed that the genotype of a bird and the microclimate surrounding the quails had a substantial impact on a bird’s physiological and productive performance. The results also showed a significant effect of line-by-treatment interaction for some studied traits (BW42, GR21–28, saturated O2, total CO2, ALT, creatinine, urea albumin, and liver %). Similarly, Erdem et al. [56] indicated that young layer chickens’ genotype and dietary environment influenced their disease resistance. Interaction between the genetics and environment might decrease the effectiveness of breeding programs [57]. Because of this, genotype by environment interaction is important for breeding programs’ effectiveness and sustainability.Finally, microclimatic measurements are necessary constituents for boosting the healthy environment for poultry. The current results suggest that the birds of the selected line were likely trying to take advantage of the improved microclimate conditions. The presented findings provide a tentative signal that zeolite influences ammonia release in quail houses, decreasing the adverse health significances of ammonia emissions. In our opinion, this solution, adding natural zeolite to the litter (ratio: 20% zeolite: 80% wheat straw), to improve microclimate conditions and reduce the ammonia emissions from the litter, can be considered promising in poultry farms. The present study suggested that improvement of microclimate conditions by adding zeolite to the litter of genetically selected quails for fast growth rate could be a promising area for future research.5. ConclusionsUsing natural zeolite as 20% of the quail wheat straw litter is considered a promising solution for the surrounding problems related to the microclimate conditions of poultry, since it led to an enhancement of the quails’ health, physiological status, growth performance, and microclimate parameters, which was accompanied by a decrease in the ammonia emission.

animals : an open access journal from mdpi

10.3390/ani11051310

PMC8147431

Methane produced by ruminants contributes to increased greenhouse gas effect. There are various nutritional strategies to reduce methane emission, such as supplementing fat or changing starch levels in the diet. Understanding the interactions of these strategies on methane emission, as well as performance, digestibility, and rumen fermentation is important. The present study aimed to assess the effects of starch level with or without a mixture of sunflower and fish oils on nutrient intake and digestibility, milk yield and composition, rumen fermentation, ruminal CH4 emissions and microbial ecology in dairy cows. Oil mixture rich in polyunsaturated fatty acids supplemented to low- or high-starch diets reduced dry matter intake and increased energy digestibility of lactating cows. High starch level improved nutrient digestibility and tended to reduce ruminal acetate:propionate ratio but did not affect rumen pH, molar propionate ratio, or ruminal CH4 emissions. Oil decreased absolute ruminal CH4 emission or tended to decrease CH4 per energy corrected milk.

Four multiparous dairy cows were used in a 4 × 4 Latin square to examine how starch level and oil mixture impact dry matter (DM) intake and digestibility, milk yield and composition, rumen fermentation, ruminal methane (CH4) emissions, and microbial diversity. Experimental treatments comprised high (HS) or low (LS) levels of starch containing 0 or 30 g of a mixture of sunflower and fish oils (2:1 w/w) per kg diet DM (LSO and HSO, respectively). Intake of DM did not differ between cows fed LS and HS diets while oil supplementation reduced DM intake. Dietary treatments did not affect milk and energy corrected milk yields. There was a tendency to have a lower milk fat concentration due to HSO compared with other treatments. Both high starch level and oil supplementation increased digestibility of gross energy. Cows receiving HS diets had higher levels of total rumen VFA while acetate was lower than LS without any differences in rumen pH, or ruminal CH4 emissions. Although dietary oil supplementation had no impact on rumen fermentation, decreased CH4 emissions (g/day and g/kg milk) were observed with a concomitant increase in Anoplodinium-Diplodinium sp. and Epidinium sp. but a decrease in Christensenellaceae, Ruminococcus sp., Methanobrevibacter ruminantium and Mbb. gottschalkii clades.

1. IntroductionThe growing human population is boosting the demand for milk and meat as sources of animal protein, resulting in several challenges for ruminant production systems, including the need to reduce their contribution to greenhouse gas emissions. This calls for special attention to solutions for reducing methane (CH4) emissions from ruminants without negative effects on productivity. Additionally, it has been shown that 5–7% of gross energy (GE) intake is lost through CH4 production from dairy cows [1]. A number of strategies, including management, dietary approaches, and genetics have been proposed for CH4 mitigation [2,3,4]. In fact, chemical composition of the feed [5] and changing the starch content of the concentrate has been proposed as a CH4 reducing strategy [2,4]. This effect was attributed to an increase in the number of amylolytic bacteria and a drop in the number of methanogens and fibrolytic bacteria, changing ruminal volatile fatty acids (VFA) in favor of propionate production [3,4] and creating an alternative hydrogen sink to methanogenesis [6].On the other hand, unsaturated fat supplementation is another feeding strategy which not only reduces enteric CH4 production [1,2] but can improve milk monounsaturated (MUFA) and polyunsaturated fatty acid (PUFA) composition [7] with potential benefits to human health [8]. Unlike starch’s mode of action, lipid sources are not fermented in the rumen; rather, they lower fermented organic matter (OM), leading to a drop in CH4 production. Furthermore, it has been shown that medium-chain fatty acids (C14–C17) also affect the number of methanogens while unsaturated fatty acids (linoleic and α-linolenic acids) shift rumen fermentation towards production of propionate and, therefore, reduce CH4 production through toxic effects on cellulolytic bacteria and protozoa [4].Thus, our hypothesis was that higher starch level and oil supplementation would have additive effects on reducing ruminal CH4 production in dairy cows. Therefore, the present study aimed to assess the effects of starch level with or without a mixture of unsaturated fatty acids (sunflower and fish oils) on nutrient intake and digestibility, milk yield and composition, rumen fermentation, ruminal CH4 emissions, and microbial ecology in dairy cows.2. Materials and MethodsAll experimental procedures were approved by the National Ethics Committee (ESAVI/4342/04.10.03/2011, Turku, Finland) in accordance with the guidelines established by the European Community Council Directive 86/609/EEC [9].2.1. Animals, Experimental Design and DietsA 4 × 4 Latin square with 2 × 2 factorial arrangement of treatments was applied to four multiparous Nordic Red cows in mid-lactation (76 ± 10.4 days in milk; mean ± SD) producing 35.2 ± 2.10 kg milk/d. The cows were fitted with rumen cannula (#1C, i.d. 100 mm; Bar Diamond Inc., Parma, ID, USA) and each experimental period consisted of 14 days diet adaptation, five days as sampling period, and a 16-day washout to avoid carry-over effects to the next period. The cows were randomly allocated to the diets. Diets, formulated to be isonitrogenous, were used based on grass silage (forage to concentrate ratio 55:45 on a dry matter (DM) basis) consisting of low starch (LS) or high starch (HS) levels (16.1 and 202 g/kg DM) with 0 or 30 g of unsaturated fatty acid mixture (sunflower oil-fish oil mixture; 2:1 w/w)/kg diet DM (LSO and HSO, respectively) (Table 1). Sugar beet pulp and barley feed in LS diets were replaced with rolled barley and ground wheat to provide different levels of starch, and urea was added to make the diets as iso-nitrogenous as possible.The oils were stored in +4 °C until incorporated into the low-or high-starch TMR to avoid oxidation of unsaturated fatty acids and the oil replaced concentrate ingredients. The oil mixture and starch levels were selected based on our previous experiences [10] where satisfactory induction of milk fat depression was realized. One of the main objectives of this work was to study milk fat depression phenomenon (not reported in this paper). The grass silage was prepared from timothy and meadow fescue (54:46) grown at Jokioinen (60°49′ N, 23°28′ E), and ensiled with a formic acid-based ensiling additive (AIV2 plus, 5 L/t; AIV Valio Ltd., Helsinki, Finland) to allow for a restricted fermentation. In order to avoid selection of dietary components and to maintain the target forage to concentrate ratio, the diets were prepared as TMR. Experimental diets were formulated to meet or exceed metabolizable energy and protein requirements of lactating cows producing 30 kg milk/d [11] offered ad libitum to result in 5–10% refusals, and fed in four equal amounts at 06:00 h, 09:00 h, 16:30 h, and 19:30 h. Cows were kept in individual tie stalls, had free access to water and salt blocks, and were milked at 07:00 h and 16:45 h.2.2. Feed Intake, Milk Yield and Chemical AnalysisDaily feed intake was measured by subtracting the refusals from the offered feed throughout the study but intakes during d 14–17 of each experimental period were used for statistical analysis. Representative samples of silage and concentrate ingredients during the sample collection period were used for chemical analysis. The samples were pooled within each period before chemical analysis using the standard methods described by Shingfield et al. [12]. In addition, the method proposed by Huida et al. [13] was used to correct silage DM content for the loss of volatiles. Indigestible neutral detergent fiber (iNDF) of silage and concentrates was determined by 12 d of ruminal incubation using nylon bags (60 × 120 mm, pore size 0.017 mm; Swiss Silk Bolting Cloth Mfg. Co. Ltd., Zurich, Switzerland) followed by neutral detergent fiber (NDF) analysis excluding ash. Chemical analysis of silage, concentrates, and oils plus their proportion in each diet were used to calculate chemical composition of each experimental TMR. Bomb calorimetry (1108 Oxygen bomb, Parr Instrument Co., Moline, IL, USA) was conducted to determine the GE of silage, concentrates, oils, and excreta. Milk samples were collected over 10 consecutive milking during d 15–19 of each experimental period, treated with preservative (Bronopol, Valio Ltd., Helsinki, Finland). Milk fat, crude protein (CP), and lactose were predicted using infrared analysis (MilkoScan 133B, Foss Electric, Hillerød, Denmark). Milk composition was calculated based on the weighted average of morning and afternoon milk yields.2.3. Rumen FermentationOn d 18 of each period and at 1.5 h intervals from 06:00 until 16:30 h, a suction pump with a Büchner flask was used to collect samples of ruminal fluid (150 mL; n = 8) through the rumen cannula. Then, pH was measured using a portable pH meter (pH 110, VWR International). Rumen liquid was filtered through two layers of cheesecloth and 5.0 mL ruminal fluid was preserved with 0.5 mL of saturated HgCl2 and 2.0 mL of 1 M NaOH to determine VFA. Furthermore, ammonia-N concentration was analyzed by collecting additional ruminal fluid (15.0 mL) preserved with 0.3 mL of 50% (vol/vol) sulfuric acid. The ruminal samples were stored at −20 °C until the time of analysis. Analysis of VFA and ammonia-N were performed as described by Shingfield et al. [12].2.4. Apparent Total-Tract DigestibilityFeces were collected over a 72-h interval starting at 18:00 h on d 14 of each experimental period and then used to determine total tract apparent digestibility coefficients. A light harness and flexible tubing attached to the vulva was used to separate urine and feces. Representative fecal samples were collected daily and composited, dried in an oven (55 °C, 48 h). The chemical composition of fecal samples was determined using the same methods for the feed samples as described earlier.2.5. Ruminal Gas ProductionRuminal CH4 and carbon dioxide (CO2) emissions were recorded over 6 days period (d 11–17 of each period) using sulfur hexafluoride (SF6) as a tracer marker as described by Bayat et al. [14] and validated against respiration chambers by Bayat et al. [15]. Briefly, gases in the rumen headspace were drawn continuously (1.7 mL/min) over every 24-h period into evacuated 5.5 L air-tight canisters using a capillary tubing (PEEK 1.6 mm × 0.13 mm i.d., VICI Valcro Instruments Co, Houston, TX, USA). Tubes used to collect the ruminal gas were anchored securely to the neck of the rumen cannula allowing gas collection at approximately 5 cm above the rumen mat. No correction was made for background SF6, CH4 and CO2 concentrations because cows were housed in a well-ventilated facility (72 m3/min) and fitted with custom-made sponges placed between the outer edge of the cannula flange and the abdominal wall to minimize the exchange of surrounding air with ruminal contents. Gas chromatography (Agilent 6890N, Agilent Technologies, Santa Clara, CA, USA) proposed by Regina and Alakukku [16] was applied to sub-samples of ruminal gases to analyze them in triplicates for CH4, CO2, and SF6 concentrations. Actual release rate of SF6 (1.16 ± 0.19 mg/d) in the rumen during the experiment as well as the concentrations of CH4, CO2, and SF6 measured by GC were used to calculate daily ruminal CH4 and CO2 emissions as following:CH4 (L/d) = SF6 (L/d) × [CH4]/[SF6] CO2 (L/d) = SF6 (L/d) × [CO2]/[SF6]2.6. Microbial AnalysisSamples of ruminal digesta (2 kg) were collected from four sites (anterior dorsal, anterior ventral, posterior dorsal, and posterior ventral rumen sacs) during each period on d 15 at 15:00 h and d 17 at 09:00 h. Samples were mixed and 50 g subsample was placed into a plastic bag and snap-frozen in liquid nitrogen. Samples were stored at −80 °C until DNA extraction. Total DNA was extracted from 250 mg of ruminal digesta following Yu and Morrison [17] protocol. Rumen bacterial, archaeal and ciliate protozoan community composition was determined using 16S and 18S rRNA gene sequencing. Primers used for amplicon library preparation, sequencing conditions and sequencing data quality control were performed as described in Tapio et al. [18]. Sequencing data was further processed using Qiime v 1.9.1 [19]. Briefly, quality filtered sequences were clustered into operational taxonomic units (OTU) at 97% similarity using UCLUST [20]. Chimeric reads were filtered out using ChimeraSlayer [21]. Bacterial and archaeal OTUs taxonomy was assigned using the Greengenes 13_8 [22] and RIM-DB [23] databases, respectively. Ciliate protozoa OTUs were assigned using ciliate protozoa database [24]. Singleton OTUs were removed and the data from each sample were rarefied to the similar sequencing depth prior to further analyses.2.7. CalculationsThe difference between nutrient intake and fecal outputs was used to calculate total tract digestibility coefficients. Energy losses as CH4 were calculated using the factor 55.24 kJ/g [25]. Potentially digestible NDF (pdNDF) was calculated as the difference between NDF and iNDF. Energy corrected milk (ECM) yield was calculated according to Sjaunja et al. [26]. Methane (or CO2) emissions as proportions of organic matter intake (OMI), organic matter digestibility (OMD), and milk and ECM yields, were calculated by dividing daily CH4 emissions (g/d) by OMI, OMD, and milk and ECM yields, respectively. Methane emissions as percentage of GE intake (GEI) was calculated as CH4 energy (MJ/d) by GEI (MJ/d).2.8. Statistical AnalysisExperimental data was analyzed by ANOVA for a 4 × 4 Latin square with a 2 × 2 factorial arrangement of treatments through the mixed procedure of SAS (version 9.2, SAS Inst. Inc., Cary, NC, USA) with a model that included fixed effects of period, starch level, oil level, and starch by oil interaction, and the random effect of cows assuming an autoregressive covariance structure fitted on the basis of Akaike information and Schwarz Bayesian model-fit criteria. The averages of data for cow within period were calculated before statistical analysis. The values reported are least square means ± SEM. The significance level p ≤ 0.05 was used to determine significant effects of starch, oil, and their interaction. In addition, probabilities at 0.05 < p < 0.10 were considered as a trend.For microbial community analysis, taxa with less than 0.01% relative abundance across all samples were filtered out before further analyses. Data was normalized by cumulative-sum scaling and log2 transformation to account for the non-normal distribution of taxonomic count data, as implemented in Calypso [27]. Microbial community alpha diversity was estimated using Shannon, Simpson diversity indices, richness and evenness estimates. Redundancy analysis (RDA) and analysis of similarity (ANOSIM), calculated based on Bray-Curtis dissimilarities, were used to identify if diet, oil or starch can be explanatory factors for the rumen microbial community composition. Permutation multivariate analysis of dispersion (Permdisp2) was used to tests whether the dispersion between the groups is significant. Analysis of variance (ANOVA) followed by subsequent pairwise comparisons with Tukey test was performed to look at diet, oil or starch effect on individual taxa. Spearman correlation was used to explore associations between rumen fermentation, methane production phenotype data and individual microbial taxa. Comparisons were counted as significant with p < 0.05. p values were further adjusted for the false discovery rate (FDR).3. Results3.1. Dry Matter and Nutrient Intake and Milk YieldIntakes of DM and GE did not differ (p ≥ 0.18) between cows fed LS and HS diets while oil supplementation reduced (p < 0.01) DM and GE intakes (Table 2). In comparison to LS diet, HS diet tended to increase (p = 0.087) OM intake while oil supplementation reduced (p < 0.01) OM intake. Starch intake was much greater for HS compared with LS diets as planned by design, but dietary oil inclusion reduced starch intake more in HS diet (p < 0.01 for the interaction of starch and oil). Intakes of NDF and water-soluble carbohydrates (WSC) were greater (p < 0.01) in cows fed LS than those fed HS diets. Intake of CP in cows fed HS diet was slightly greater (p < 0.05) and again oil supplementation led to a lower (p < 0.01) CP intake. As expected, intakes of saturated fatty acids (SFA), MUFA, and PUFA were greater (p < 0.01) with oil supplementation of both low- and high-starch diets. Milk and ECM yields were not influenced (p ≥ 0.11) by dietary treatments (Table 3). However, ECM yield was noticeably (2.7 kg/d) yet numerically lower for HSO diet than other treatments. There was a tendency (p = 0.07 for interaction of starch level and oil supplementation) to cause lower milk fat concentration due to HSO compared with other treatments and oil supplementation reduced (p < 0.01) milk protein concentration. Inclusion of oil in the diet tended to reduce (p = 0.087) milk fat yield and reduced (p < 0.05) milk protein yield. Milk production efficiency expressed as ECM/DMI was not affected (p ≥ 0.28) by dietary treatments.3.2. Apparent Total-Tract DigestibilityThe high starch level, but not oil supplementation, increased (p < 0.01) apparent total tract digestibility of DM, OM and starch while decreased (p < 0.01) NDF and pdNDF digestibility (Table 4). Both high starch level and oil supplementation increased (p ≤ 0.02) digestibility of GE and tended to increase (p = 0.058) CP digestibility. There was no interaction between starch level and oil supplementation for any of digestibility measurements.3.3. Rumen FermentationThe experimental treatments had no impact (p > 0.05) on rumen pH while HS diet tended to increase (p = 0.056) total VFA concentration compared with LS diets (Table 5). However, no significant change (p = 0.21) was observed in total VFA as a result of oil supplementation. Compared with the cows receiving LS diet, cows fed HS diets had lower (p < 0.01) molar proportion of acetate and greater (p < 0.05) molar proportions of butyrate, isobutyrate, valerate, isovalerate, and caproate. The experimental treatments did not influence (p ≥ 0.18) molar proportion for propionate. Acetate to propionate ratio tended to be lower (p = 0.07) for HS compared with LS diets. Ruminal ammonia-N was greater for HS compared with LS diets (p < 0.01), and dietary oil inclusion increased ammonia more in HS diet (p < 0.05 for the interaction of starch and oil).3.4. Ruminal CH4 and CO2 EmissionInclusion of the oil mixture in LS and HS diets reduced (p = 0.05) daily ruminal CH4 emissions (Table 6). Cows receiving oil supplements had lower CH4 emission intensity calculated as g/kg milk (p < 0.05) and g/kg ECM (p = 0.067) than their control counterparts. No difference (p ≥ 0.15) was found between the treatments in terms of CH4 emissions calculated as proportion of GE intake or g/kg OM digested. The experimental treatments were not different (p ≥ 0.16) in terms of daily ruminal CO2 emission and g/kg OM digested.Ruminal CO2 emissions expressed as g/kg milk or ECM was greater for LS compared with other diets (p = 0.086 and 0.036 for the interaction, respectively).3.5. Rumen Microbial EcologySequencing yielded 5222–7362 good quality sequences per sample for bacteria, 1730–18,249 for archaea and 3339–12,123 for ciliate protozoa. Rumen bacterial community was represented by 18 phyla. Bacteroidetes (51–60%), Firmicutes (20–25%), Proteobacteria (1–12%), and Spirochaetes (0.9–5%) were the dominating phyla. Remaining phyla were detected at the abundance below 1%. Among bacterial genera, Prevotella was predominant in all dietary groups (40–50%) with other more abundant genera being unclassified Succinivibrionaceae (0–11%), unclassified Clostridiales (5–6%), Treponema (1–5%), unclassified Ruminococcaceae (3%), unclassified Lachnospiraceae (2–4%), Succiniclasticum (2%), Ruminococcus, Fibrobacter, CF231 and Butyrivibrio (altogether 1–2%). The remaining genera were detected at an abundance below 1%.The Archaea community was dominated by Methanobrevibacter gottschalkii (50–63%) and Methanobrevibacter ruminantium (12–24%) in all the groups. Other more dominant archaea were Methanosphaera sp. ISO3F5 (6–24%), Methanimicrococcus blatticola (1–15%) and Methanobacterium alkaliphilum (1–3%). Archaea groups belonging to the Methanomassiliicoccaceae (Mmc) family were observed at the abundance below 1%.Ciliate protozoa community was dominated by Entodinium (35–56%) in all dietary groups. Other predominant ciliate genera detected at the abundance above 5% in at least one of the diets were: Polyplastron, Ostracodinium, Metadinium, Isotricha, Eudiplodinium-Eremoplastron, Epidinium, Charonina, and Anoplodinium-Diplodinium.Dietary treatments had little effect on microbial alpha diversity estimates. Only bacterial richness was significantly (p = 0.01) reduced in high starch diets and archaeal richness tended to be numerically lower (p = 0.06) in dietary treatments with oil additive. No significant diet, oil or starch effect was observed on richness of ciliate protozoa (Table 7).Beta diversity analysis was performed using RDA and ANOSIM, and identified significant clustering of bacteria with respect to diet (p = 0.006) and starch (p = 0.006). Amount of starch in the diet was also a significant explanatory factor (p = 0.05) for the clustering of ciliate protozoa. No distinct clustering of archaea was identified due to diet, oil or starch. Beta dispersion was significantly different between high and low starch bacterial communities (p = 0.009).3.6. Taxa Affected by Diet, Oil and StarchDiets with oil additive significantly (p < 0.05) increased abundance of Anaerovibrio sp., (Spirochaetes) PL-11B10 and ciliate protozoa Eudiplodinium-Eremoplastron (AB536716). Oil caused a significant reduction in (Bacteroidetes) RF16, (Proteobacteria) GMD14H09, Anaeroplasma sp., Prevotellaceae, (TM7) F16, Bacteroidales, and archaea belonging to Mmc. Group 8 sp. WGK1. However, only increase in Anaerovibrio sp. remained significant after correction for multiple testing using false discovery rate (FRD = 0.035) (Figure 1, Table S1).Diets with high starch content had significantly higher relative abundance of bacteria (Bacteroidetes) S24-7, Succinivibrionaceae spp., Ruminobacter sp., Selenomonas sp., Moryella sp., Ruminococcus flavefaciens, and ciliate protozoa Isotricha sp. and Entodinium sp. Contrary, diets with low starch content were significantly enriched in Treponema sp., (Bacteroidetes) F16, SR1, Lachnospira sp., Clostridium sp., Acholeplasmatales, Desulfovibrio sp., (Tenericutes) RF39, (Actinobacteria) OPB41, Lactobacillus sp. and ciliates affiliated with Eudiplodinium-Eremoplastron (OTU12), Charonina ventriculi, Ostracodinium sp. and Dasytricha sp. After multiple testing correction, only (Bacteroidetes) S24-7, Succinivibrionaceae sp., Isotricha sp. and Eudiplodinium-Eremoplastron (OTU12) retained significant differences (Figure 1, Table S1).Diet composition had significant (p < 0.05) effect on 27 bacterial, one archaeal and seven ciliate protozoan taxa, but after FDR correction, only seven bacterial and two ciliate protozoan taxa remained significant (Figure 1). In pairwise comparisons, Anaerovibrio sp. was significantly more abundant in diets containing oil, in particular HSO diet when compared with HS, LS or LSO diets. Eudiplodinium-Eremoplastron (OTU12) was significantly more abundant in LS compared with HS or HSO diets, while Isotricha sp. was significantly more abundant in HS compared with LS or LSO diets. Bacteria (Bacteroidetes) F16, (Proteobacteria) GMD14H09, (Bacteroidetes) RF16 and SR1 were detected at the lowest abundance in HSO diet and showed significant differences between HSO and all other diets. Bacteria (Bacteroidetes) S24-7 had highest abundance in HSO diet and showed significant differences in pairwise comparisons with LS and LSO diets. Among archaea Mmc. Group 8 sp. WGK1 was more abundant (p = 0.03, FDR = 0.4) in HS diet compared with HSO or LSO diets (Figure 1, Table S1).3.7. Microbiota Association with Rumen Fermentation and Methane Production TraitsTo look at the associations between rumen microbiome and rumen fermentation as well as methane production traits, we applied Spearman correlations on the microbial taxa present in all samples (n = 16 observations). From 50 bacterial, two archaeal and 25 ciliate protozoan significant associations (p < 0.05), eight bacterial and 10 ciliate protozoan correlations remained significant after correction for multiple testing using false discovery rate (FDR < 0.1).An increase in acetate was negatively correlated with (Bacteroidetes) S24-7 (FDR = 0.005), but positively with Clostridium sp. (FDR = 0.035), ciliates Dasytricha sp. (FDR = 0.053), and Charonina ventriculi (FDR = 0.057) (Figure 2). Increase in ammonia-N was positively associated with (Bacteroidetes) S24-7 (FDR = 0.011) and Moryella sp. (FDR = 0.049) but negatively associated with Treponema sp. (FDR = 0.044), (Bacteroidetes) RF16 (FDR = 0.049), Clostridium sp. (FDR = 0.049), and Charonina ventriculi (FDR = 0.07). Charonina ventriculi was negatively correlated with butyrate (FDR = 0.043) while Epidinium sp. was negatively correlated with isobutyrate (FDR = 0.062) and isovalerate (FDR = 0.016). On the contrary, Entodinium sp and Entodinium caudatum were positively correlated with isovalerate (FDR = 0.089 and 0.016, respectively). Anoplodinium-Diplodinium sp. and Epidinium sp. were negatively correlated with daily CH4 emissions (FDR = 0.064), while Christensenellaceae and Ruminococcus sp. correlation was positive. Ruminococcus flavefaciens was positively correlated with valerate (FDR = 0.063). Methanobrevibacter ruminantium and Mbb. gottschalkii clades were both positively correlated with CH4 intensity (g/kg milk) (p < 0.03, FDR = 0.12) (Figure 2).4. Discussion4.1. Dry Matter Intake, Milk Yield, and Nutrient DigestibilityThe results of this experiment showed that diets with oils rich in PUFA in a moderate amount reduced DM intake, which in turn resulted in lower intakes of other nutrients (OM, CP, NDF, WSC, and starch). Similar observations have been reported previously in experiments with PUFA-rich plant oils [28,29], fish oil [30,31], or a mixture of fish oil and plant seed oils [32,33,34]. In this experiment, dietary oil supplementation reduced DM intake by 9.7% on average. In line with these results, Shingfield et al. [33] reported approximately 20.5% reduction in DM intake as a result of adding a mixture of fish and sunflower oils to corn silage-based diet (45 g oil/kg DM and 63 g ether extract (EE)/kg DM) compared with a non-supplemented diet (EE content of 33.5 g/kg DM). However, another study reported no difference in terms of DM intake of cows between the control diet based on alfalfa and corn silage (27.8 g EE/kg DM) and the same diet supplemented with mixture of fish or canola oils (20 g oil/Kg DM and 46.7 g EE/kg DM, respectively; [31]). It has been well shown that the effect of oil supplementation of a diet on DM intake can be a function of combination of including oil content and diet composition, source of oil, and type of basal diet [29,35,36].The lack of responses on milk yield due to dietary starch level and unsaturated oil mixture were expected as the diets were designed based on our previous experiences [10] to cause no effect on milk yield but lower fat and ECM yields due to the combination of high starch level and oil (i.e., HSO diet) known as milk fat depression effect. However, ECM yield was not influenced by dietary treatments despite being noticeably yet numerically lower for HSO diet (2.7 kg/d) compared with the average of other treatments. The reason for numerically lower ECM was the lower milk fat concentration (11% lower compared with LS diet).The lack of oil supplementation effect on nutrient digestibility, with exception of CP and GE digestibility is consistent with some previous findings [14,37,38].While in our experiment high-starch diet did not affect DM intake, an increase was observed for CP, OM, and starch intakes with drops in NDF and WSC intakes, which are due to differences in nutrient contents of the diets. The results of our study are consistent with the findings of Pirondini et al. [39] and Philippeau et al. [40]. In contrast, a 4.3% drop in DM intake for cows fed high- compared with low-starch diets based on grass silage (212 vs. 116 g starch/kg DM) has been reported [41]. Starch effect on feed intake can be mediated by a number of factors including starch fermentation rate, forage content of diet, amount of metabolic fuel absorbed from the rumen (for instance VFA), rumen pH, and rumen fermentation parameters [42,43]. It should be noted that the source of starch in diet and processing method can also contribute to the mixed results of different experiments. Hatew et al. [41] attributed the lower DM intake in high starch diets to increased propionate concentration in the rumen since hepatic oxidation of propionate influences DM intake [44]. Therefore, as in the aforementioned studies, no significant change was found in ruminal propionate concentration which is consistent with unaffected DM intakes.In our experiment, digestibility of DM, OM, starch, and GE was higher and CP digestibility tended to be higher in cows fed HS than those receiving LS diets. Beckman and Weiss [45] observed linear increase in DM and OM digestibility as a result of reduction in NDF:starch ratio and linked it to replacement of highly digestible carbohydrates (i.e., starch) with low-digestible carbohydrates (i.e., NDF). In the current experiment, lower NDF and pdNDF digestibility along with improved starch digestibility beside the lack of effect on rumen pH as a factor influencing fiber digestion, might reflect the competition between rumen microbial population to utilize more easily nonstructural carbohydrates when available. As shown by the results of the microbiological analysis, the abundance of Ruminobacter sp. or Selenomonas sp. increased in high starch diets while the abundance of amylolytic Prevotella or cellulolytic bacteria Ruminococcus and Fibrobacter in LS-fed cows was not significantly different from that in HS-fed cows. Another explanation might be that observed differences in fiber digestibility reflect changes in the composition of diets; in HS diet rolled barley and ground wheat were used instead of sugar beet pulp and barley feed which might have different fiber characteristics.It should be noted, however, that in some experiments where starchy feeds were replaced by fibrous by-products, there has been a confounding effect of both starch:NDF ratio and forage proportion of the diet [45]. However, in our experiment, the forage proportion in both LS and HS diets were fixed to remove such a source of difference between diets.4.2. Methane Production and Rumen FermentationThe oil supplementation to both low- and high-starch diets resulted in lower CH4 production compared with non-supplemented diets (475 vs. 552 g/d, on average). This is in line with the findings of the experiments that used oilseeds [29,46,47], or fish oil [30] to supplement dairy cow diets. In the current experiment, enteric CH4 emission (g/d) reduced by on average 4.7% for each 10 g/kg DM unsaturated oil in the diet. Similarly, Martin et al. [4] indicated that CH4 production reduces by 3.8% as a result of every 10 g/kg increase in dietary lipid supplementation. The most of this response in our experiment was caused by lower feed intake as CH4 yield (g/kg OMI) was not affected by the oil supplementation (average numerical reduction of 1.7% in CH4 yield). Ramin and Huhtanen [48] in a meta-analysis showed that 1 g/kg of DM increase in dietary EE concentration decreased CH4 yield by 0.043 L/kg of DM. The equivalent value in our experiment was on average 0.045 g/kg of OM intake which is very close to the reported value. Apparently both biohydrogenation of unsaturated fatty acids and amount of unfermentable organic matter introduced by oil to the diet should have caused the lower CH4 yield. Oil potential in reducing CH4 production seems to be a function of factors such as source of oil, fatty acid composition and level of supplementation in the diet [2,4]. Several mechanisms are known to influence the impact of fats and oils in reducing CH4 production including reduced OM fermentation in the rumen, unfavorable effects of C12:0 and C14:0 on protozoa community, and inhibition of methanogens by 18-carbon unsaturated fatty acids [4,49,50] and the competition for using hydrogen for biohydrogenation of unsaturated fatty acids. Patra [51] reported that C12:0 and C18:3 fatty acids are stronger inhibitors of methanogenesis compared with other fatty acids. Furthermore, as far as reduction in CH4 production is concerned, MUFA and PUFA are more effective than SFA in the diet. In this experiment, oil supplementation to both low- and high-starch diets resulted in higher intakes of both MUFA and PUFA compared with non-supplemented diets.Unaffected CH4 yield for cows fed oil-supplemented diets in our experiment is consistent with the lack of differences in rumen pH, acetate, propionate, and VFA concentration and acetate:propionate ratio, although acetate and propionate concentrations and acetate:propionate ratio decreased numerically. However, oil supplementation is expected to not only reduce CH4 production but also lower acetate:propionate ratio. The non-significant change in acetate:propionate ratio observed in our experiment can be attributed to the lower DM and concentrate intakes, and the forage portion of our experimental grass silage-based diet. Changes in rumen fermentation pattern due to oil supplementation with diets based on restrictively fermented grass silage (using the silage additive based on formic acid) may be resistant to lipid supplementation [7].Even though the intake of GE increased as a result of including oil in LS and HS diets, there were no significant differences between experimental treatments in terms of CH4 production as a percentage of GE intake while CH4 intensity calculated as g/kg milk or ECM decreased and tended to decrease by oil supplementation, respectively. In fact, the reduction in CH4 intensity indicates that net energy is partitioned more towards milk production, leading to lower CH4 intensity [3]. The values of CH4 production as a percentage of GE intake for un-supplemented diets (LS and HS; 7.25 and 7.04%) having DM intake of 23.0 and 23.5 kg/d are higher than 6.4% measured from dairy cows receiving rather similar diets with similar DM intake in respiration chambers (Bayat et al., unpublished data).Our findings showed that starch level did not influence ruminal CH4 and CO2 production (g/d) and emission intensity (g/kg milk yield). Previous studies [39,40,41] have reported lower or tendencies towards lower daily CH4 production and CH4 emission intensity with the exception of Hatew et al. [41] reporting a non-significant CH4 emission intensity due to increasing dietary starch level. Increased starch content influenced rumen fermentation parameters, with a significant decrease in molar proportion of acetate and a tendency to reduce acetate:propionate ratio. Although high starch diets are expected to reduce acetate and increase propionate molar proportions and to lower CH4 production [4], this mechanism may not invariably apply to all experimental conditions. We did not observe any differences in rumen pH between cows fed low- and high-starch diets while it has been shown that high starch content lowers rumen pH, thereby limiting growth or activity of methanogens and cellulolytic bacteria [4]. The higher rumen ammonia N concentration due to higher starch level can be attributed to slightly higher dietary CP level whereas the higher rumen ammonia N concentration with oil-containing diets might have arisen from adding urea, which is highly rumen-degradable, to the diets in an attempt to make them isonitrogenous.4.3. Rumen Microbial EcologyStarch level had stronger effect on bacterial richness compared with those caused by inclusion of oil in the diet. Despite reports of toxic effects of oils on microbial community and, therefore, greater expected changes in the rumen microbial composition [52], this study did not observe significant oil effect on the reduction of alpha diversity in bacterial, archaeal or ciliate protozoan communities. Effects of oil on the microorganisms may depend on type and amount of oil, type of fatty acids in diet, and type of dietary forage fed. Bayat et al. [7] showed that supplementation of dairy cow diets with plant oils like rapeseed, safflower, linseed or myristic acid reduced CH4 production, with each type of oil having a different impact on bacterial community. Furthermore, Martin et al. [37] reported that adding extruded linseed to hay-based diets or corn silage-based diets for dairy cows reduced CH4 production without a notable change in abundance of rumen methanogens and cellulolytic bacteria. Looking at individual microbial taxa, addition of oil to both low- and high starch diets in this study, provided a suitable ecological niche for lipid hydrolyzing Anaerovibrio sp. [53] and Spirochaetes order PL-11B10. PL-11B10 was detected at significantly higher abundance in the diet with myristic acid supplement [7] and outside ruminants has been found positively correlated with methane production in methanogenic oil wells [54]. Nevertheless, our understanding of PL-11B10 ecology in rumen is limited. Oil supplementation also increased relative abundance of ciliate protozoa Eudiplodinium-Eremoplastron (AB536716). Similarly, a significant increase in Eremoplastron dilobum abundance was detected in an in vitro experiment with linseed oil but not rapeseed oil additive [55], suggesting that a positive or negative oil effect is depends on protozoa species and oil type.High starch diets increased abundances of known starch utilizers like Ruminobacter sp., Selenomonas sp. or ciliate protozoa Isotricha sp. and Entodinium sp. that are involved in the utilization of non-structural polysaccharides and soluble sugars [53]. The Bacteroidetes family S24-7 was also enriched in HS diets. Current research suggests that members of S24-7 family are versatile with respect to complex carbohydrate degradation, but starch utilization trait is common to all family members and increased abundance of S24-7 is correlated with increased propionate production [56]. Inclusion of oil in LS and HS diets did not change total VFA, acetate, propionate, and butyrate but reduced CH4 production, suggesting lower H2 availability in these ruminal ecosystems. S24-7 was significantly more abundant in HSO diet, was strongly negatively correlated with acetate and tended (p = 0.054) to be positively (R = 0.49) correlated with propionate concentration in the rumen (data not shown). In addition to S24-7, also Prevotella sp., Moryella sp. and members from Paraprevotellaceae family were positively correlated with propionate concentration in the rumen (Figure 2). Given that propionate concentration was numerically the highest in HSO diet, these bacteria may have contributed to the sink in reduction of H2 availability for methanogenesis.Low starch diets were enriched with bacteria directly or indirectly involved in fiber degradation. Clostridium, Treponema, (TM7) F16, (Tenericutes) RF39, and Desulfovibrio were found to be tightly attached to switchgrass [57] or wheat straw [58] during degradation process. In co-cultures with Fibrobacter succinogenes, Treponema bryantii has been shown to utilize soluble sugars released from cellulose degradation [53]. In addition to bacteria, ciliate protozoa Ostracodinium sp. and Dasytricha sp., are known to contain cellulolytic and hemicellulolytic activities, respectively, and were significantly enriched in LS diets. Charonina ventriculi is a holotrich protozoa not frequently observed in the rumen and with limited information about its metabolism. In our experiment, Charonina ventriculi was significantly enriched in LS diets and was negatively correlated with ammonia-N and butyrate, but positively correlated with acetate concentration in the rumen. Correlation profile of Charonina ventriculi was similar to Treponema sp. and Clostridium sp. (Figure 2) suggesting that these microorganisms require similar rumen conditions for thriving or are involved in similar metabolic processes.Methanobrevibacter gottschalkii and Mbb. ruminantium were the predominant archaea without being significantly affected by starch level or oil supplement, although oil numerically reduced their abundance. Methanobrevibacter are hydrogenotrophic methanogens that convert H2 and CO2 produced by protozoa, bacteria, and fungi to CH4. In our study, numerical decreases in both Mbb. gottschalkii and Mbb. ruminantium correlated with decrease in CH4 intensity (g/kg milk) in oil supplemented diets. With higher abundances of bacteria Moryella sp., Anaerovibrio sp., (Bacteroidetes) S24-7, (Spirochaetes) PL11B10, Selenomonas sp., Ruminobacter sp., (Paraprevotellaceae) YRC22 and ciliate protozoa Anoplodinium-Diplodinium which had a tendency to be positively correlated with propionate concentration in the rumen, we can hypothesize having an ecosystem with less hydrogen available for methanogenesis.Reduced daily CH4 emissions (g/d) were positively associated with reduction in Entodinium caudatum. Although Entodinium was the most abundant ciliate protozoa in all diets and smaller Entodinium spp. have been suggested to contribute more to CH4 production compared with larger protozoa in in vitro studies [59], a deeper subdivision of Entodinium into OTUs suggests functional versatility and differences in host dependency inside this genus Contrary to our results, Belanche et al. [60] investigated holotrich protozoa role in CH4 production compared to the natural flora and concluded that holotrichs were responsible for increased methanogenesis more than the entodiniomorphids. In our study, OTU affiliated with Isotricha prostoma was negatively correlated with daily CH4 emissions (OTU detected in 10 samples and therefore not included in Figure 2), while other OTUs affiliated with holotrichs did not produce significant associations with CH4 production. Discrepancies in observations could relate to the differences in the basal diet and dietary treatment as well as host impact on the general microbial community composition.Based on the results of this experiment, it can be argued that inclusion of the mixture of fish oil and sunflower oil at 30 g/kg of diet DM does not have profound toxic effects on bacteria, archaea, or ciliate protozoa, and it is possible that the minor reduction in CH4 yield caused by inclusion of oil in diet may be linked more to the functional networks of microbiome possibly leading to a lower availability of hydrogen in the rumen which is required for CH4 production.5. ConclusionsOverall, the results of this experiment show that starch level modified rumen fermentation and nutrient digestibility without influencing DM intake or methane emissions. Inclusion of unsaturated oil mixture (sunflower and fish oils, 2:1 w/w) reduced DM intake and some ruminal methane emission indices without influencing rumen fermentation characteristics or nutrient digestibility. The findings of this experiment show that feeding more starch originating from concentrate portion instead of fiber at a moderate level in dairy cow diets does not favor lower methane production, and oil supplementation is similarly effective on reducing methane in low- and high-starch diets. Therefore, our hypothesis that starch level and oil supplementation would have synergistic effects on CH4 emission could not be proved as increasing dietary starch level did not influence CH4 emission whereas oil supplementation did. Inclusion of moderate amount of the unsaturated oil mixture in the diet did not have profound toxic effects on bacteria, archaea, or ciliate protozoa, which is in line with the minor effect on methane yield.

animals : an open access journal from mdpi

10.3390/ani11113100

PMC8614346

Parathyroid hormone (PTH) is involved in many metabolic diseases, such as chronic kidney disease (CKD) and calcium disorders, and its measurement could be of clinical utility. However, available methods for the measurement of feline PTH are limited and not widely accessible. The aim of this study was to perform the analytical validation of a new method for PTH measurement in cats. Thirty-eight cats affected with CKD were included. The analytical protocol provided an evaluation of the precision, accuracy, and storage stability at different temperatures. The method investigated showed good precision and accuracy and good stability for 1 week of storage at freezing temperatures. The method was validated in cats, allowing its future use in diagnostic procedures.

The determination of parathyroid hormone (PTH) in cats could be of clinical utility in many metabolic disorders, such as renal diseases, hypercalcemia, or nutritional imbalances. However, the available methods for the measurement of feline PTH are limited, not widely available, and need radioimmunoassays. The aim of this study was to perform the analytical validation of a new immunoenzymatic method for the measurement of feline PTH. Thirty-eight cats affected with chronic kidney disease (CKD) were included. PTH was measured using a two-site immunoenzymatic method validated in humans and dogs (ST AIA-PACK® Intact PTH, Tosoh Bioscience, Tessenderlo, Belgium). The analytical validation provided the evaluation of precision (intra-assay and inter-assay), accuracy (linearity under dilution (LUD) and spike recovery test (SRT)), and the storage stability of serum samples at 20 °C, 4 °C, and −20 °C. The method showed good precision (intra-assay CVs (coefficient of variations) 3.19–9.61%; inter-assay CVs 9.26–15.28%). In both the intra- and inter-assays, the highest imprecision was found with the low concentration pool (9.61% and 15.28%) and accuracy (LUD and SRT r2 = 0.99, p < 0.001), while the stability was optimal up until 7 days at −20 °C (−7.7%). The method was successfully validated in cats, allowing its future use in diagnostic procedures.

1. IntroductionParathyroid hormone (PTH) is a single-chain, 84-amino-acid polypeptide produced by chief cells in the parathyroid glands and highly conserved among mammalian species; feline PTH is more than 83% identical to the canine and human molecules [1].PTH is the main regulator of ionized calcium concentration through its direct effects on renal tubular reabsorption and bone resorption of calcium and indirect intestinal absorption of calcium mediated by calcitriol or active vitamin D. PTH also has a role in phosphorus metabolism, increasing bone resorption and decreasing reabsorption in renal tubules [2].PTH is involved in many pathological conditions in which calcium and phosphorus metabolism is imbalanced; among these, primary hyperparathyroidism, renal secondary hyperparathyroidism (RHPT), nutritional secondary hyperparathyroidism, and hypercalcemia of malignancy. Specifically, RHPT is secondary to chronic kidney disease (CKD), a common disease in feline medicine affecting especially old cats (30–40% over 10 years) due to different underlying aetiologies; CKD induces many metabolic disorders involving calcium and phosphorus metabolism, resulting in hypersecretion of PTH [3].The determination of PTH concentrations is therefore a fundamental tool in the diagnosis of calcium metabolism disturbances, in monitoring the treatment response, and in the evaluation of prognosis. In this context, routine measurement of PTH improves the ability to diagnose parathyroid disorders [1,4]. However, the currently available validated methods for the measurement of feline PTH are limited and not widely available.To date, different generations of assays for the measurement of PTH in people and veterinary species have been reported. Briefly, first-generation PTH immunoassays used a single polyclonal antibody directed against the PTH C-terminal part or midterminal part. The finding of many biologically inactive C-terminal fragments detected by these assays led to the development of more specific second-generation assays using two antibodies directed at two different regions of the molecule. However, also these assays were found to measure some C-terminal fragments and a third-generation radioimmunoassay, able to detect specifically the full-length PTH of all 84 amino acids, has been recently developed [5].In human medicine, an immunochromatographic test strip method able to detect tissue PTH for intraoperative parathyroid identification is also currently available, without current application in veterinary medicine [6].Specifically, for PTH measurement in cats, a second-generation validated method (Allegro Intact PTH, Nichols Institute Diagnostics, San Juan Capistrano, CA, USA) is no longer available and only a third-generation assay (Duo PTH kit, Scantibodies Laboratory, Inc., Santee, CA, USA; total intact PTH IRMA-coated bead version, Part number 3KG600, Scantibodies Laboratory, Inc., Santee, CA, USA) is currently available, with the validation data reported by different authors [7,8,9]. However, also this method has several limitations, and it also has now been discontinued. First, radioimmunoassays are not available in all veterinary laboratories due to the high costs of the equipment and reagents needed. Moreover, the assays evaluated by Pineda et al. and Williams et al. showed a lower limit of detection of 3 pg/mL and 5.2 pg/mL, respectively, higher than the lower reference limit for cats [8,9]. Considering this scenario in which radioimmunoassays are discontinued, the possibility to expand the range of validated methods for the routine measurement of feline PTH would be useful.The aim of this study was the analytical validation of a new immunoenzymatic method, already validated in humans and dogs, for the measurement of the feline PTH.2. Materials and Methods2.1. Animals and SamplesThis study was carried out including 57 serum samples obtained from 38 cats affected with chronic kidney disease referred to the University of Milan Veterinary Teaching Hospital during routine clinical activity between February 2020 and April 2021. Fifteen cats were sampled more than once during the monitoring of the disease, each cat being sampled from 1 to 4 times.As the severity of renal hyperparathyroidism is assumed to increase with the degree of azotaemia, cats at different CKD stages were included in the study, to increase the probability of obtaining samples with different PTH concentrations [10]. For all cats, the diagnosis and staging of CKD were performed according to the International Renal Interest Society (IRIS) guidelines: serum creatinine (sCr) persistently above 1.6 mg/dL and/or presence of renal proteinuria (urinary protein:creatinine ratio higher than 0.4) and/or presence of ultrasonographic abnormalities compatible with CKD. All cats included in the study were staged according to the IRIS guidelines [11].All cats underwent a physical examination according to standard veterinary procedures and blood withdrawal was performed for diagnostics purposes after the owner’s consent. According to the Ethics Committee of the University of Milan (EC decision 29 October 2012, renewed with the protocol n° 02-2016), biological samples collected in this setting could also be used for research purposes and additional approval was not required. In total, 2 to 3 mL of blood were collected from the cephalic or jugular vein and placed into methacrylate tubes without anticoagulants, pre-filled with a gel separator and clot activator (FL Medical, Torreglia, Padua, Italy), followed by centrifugation (10 min, 2500× g) within 30 min from collection, and execution of routine analyses (including sCr) within 2 h. The leftover serum was utilized for the measurement of PTH.2.2. PTH MeasurementThe measurement of PTH was performed after routine analyses using an automated analyser (AIA 360®) and a two-site immunoenzymometric assay (ST AIA-PACK® Intact PTH, Tosoh Bioscience, Tessenderlo, Belgium) validated in humans and dogs [12]. Briefly, the analysis is performed entirely in the test cups according to the procedure described. Intact PTH present in the sample is bound with a polyclonal antibody immobilized on the magnetic solid phase and an enzyme-labelled polyclonal antibody. The magnetic beads were washed to remove the unbound enzyme-labelled polyclonal antibody and were then incubated with the fluorogenic substrate 4-methylumbelliferyl phosphate (4MUP). The amount of enzyme-labelled polyclonal antibody that binds to the beads is directly proportional to the intact PTH concentration in the sample. The standard curve was constructed using six calibrators, provided by the manufacturer, with increasing concentrations of PTH: 0 pg/mL, 15.3 pg/mL, 48.3 pg/mL, 198 pg/mL, 770 pg/mL, and 2280 pg/mL. Three levels of control sera provided by the manufacturer, with concentrations of 10.70, 33.60, and 213.50 pg/mL, were run before each work session.2.3. Analytical ValidationThe analytical validation was performed using the calibrators, control materials, and reagents provided by the manufacturer, which are all human-based. All the pooled sera were obtained by merging specimens from cats included in the study and grouped based on their PTH concentration, following preliminary assessment through measurement on single samples.The intra-assay imprecision was determined by measuring PTH in feline pooled sera with low (3.65 pg/mL), medium (24.12 pg/mL), and high (102.90 pg/mL) PTH levels; 5 replicates of each measurement within a single run of analysis were done on each pool. The inter-assay variability was assessed by analysing the same samples, in duplicate, on 5 consecutive working days. The mean value, standard deviation (SD), and coefficient of variation (CV = SD/mean × 100) were calculated.The accuracy was determined by evaluating the linearity under dilution (LUD) and the spike-recovery test (SRT): LUD was performed by measuring PTH on a pool of feline sera with a high PTH concentration (86.5 pg/mL) after serial dilutions with a 0.9% NaCl saline solution to obtain solutions containing 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20%, 10%, 5%, 2.50%, and 1.25% of the serum sample, respectively. In the absence of an accepted gold standard to measure canine PTH, nor any available purified feline PTH, SRT was performed by mixing a pool with a low PTH level (1.1 pg/mL) with increasing percentages (10% to 100%) of a pool with a medium PTH level (14.1 pg/mL). The association between the obtained and the expected values in the LUD and SRT tests was assessed using a least-squares regression analysis performed by the statistical software JMP 16 (SAS Inc, Cary, NC, USA).The lower limit of detection (LLOD) of the method was obtained using the following formula: LLOD = blank + 1.645 (SDlow concentration sample), where blank is the assay sensitivity provided by the manufacturer (1 pg/mL) and the low concentration sample is the mean of the lowest values obtained by LUD [13]. The LLOD was calculated with this method because in the measurement of the zero calibrator or blank the instrument does not provide numerical results.2.4. Storage StabilityStorage stability was evaluated using a pool of fresh feline sera with a mean of 11.05 pg/mL, analysed in duplicate immediately after sampling and then split in different aliquots and stored under different temperatures: room temperature (about 20 °C), refrigeration temperature (4 °C), and freezing temperature (−20 °C). The analysis was repeated in duplicate after the following times: 6 h (within the same work shift) and 24 h (between a work shift and the following) at room temperature; 6 h, 24 h, 48 h, and 72 h at refrigeration temperature; and 1 week and 1 month at freezing temperature. A deviation from the baseline of ±10% was considered tolerable, as reported by previous validation studies [14,15].3. ResultsThe population was composed of 34 Domestic Shorthairs, 1 American Curl, 1 Birman, 1 Bengal, and 1 Maine Coon; the age ranged from 1 to 19 years with a median of 10 years; according to sex, 18 neutered males, 16 neutered females, 3 entire males, and 1 entire female were present.The CKD staging according to the IRIS staging system at first examination was as follows: 1 cat was classified in Stage 1, 26 in Stage 2, 4 in Stage 3, and 7 in Stage 4.Serum creatinine ranged from 1.5 to 16.4 mg/dL, with a median of 2.4 mg/dL.Serum PTH ranged from 1.3 (excluding values below the LLOD) to 377.6 pg/mL, with a median of 11.9 pg/mL. According to the IRIS stage, the only cat in Stage 1 had a concentration of PTH of 11 pg/mL; the median PTH values were 9.8 (1.3–362) pg/mL in Stage 2; 18.9 (1.4–130.1) pg/mL in Stage 3; and 94.8 (3.9–377.6) pg/mL in Stage 4. In nine cats, the PTH value was lower than the limit of detection of the instrument; these samples were used in the low pools.The complete data are reported in Supplementary Materials Table S1.3.1. Precision and AccuracyThe results of the intra- and inter-assay precision assessment evaluated on the pooled sera at low, medium, and high PTH concentrations are reported in Table 1. The intra-assay CVs were <10% at the three levels of PTH concentration; the inter-assay CVs were between 9.26% and 15.28%. In both the intra- and inter-assays, the highest imprecision was found with the low concentration pool (9.61% and 15.28%, respectively).The results obtained after LUD and SRT are reported in Figure 1. Both tests fitted the linear model (r2 = 0.99, p < 0.001 for the LUD test; r2 = 0.99, p < 0.001 for the SRT), showing a satisfying agreement between the observed and expected results. This part of the study allowed the determination of the LLOD of the method, which was about 1.1 pg/mL.3.2. Storage StabilityStorage stability was evaluated using a pool of sera with a medium PTH concentration (11.05 pg/mL) tested at different storage conditions (Table 2). The deviation from baseline exceeded 10% after 6 h of storage at both 20 °C and 4 °C, ranging between 14.5% (6 h at 4 °C) and 51.1% (24 h at 20 °C). At freezing temperature (−20 °C), the deviation from the baseline was minimal until 7 days storage, whereas it exceeded 10% after 1-month storage.4. DiscussionFeline hyperparathyroidism is a documented disease, either in its primary or secondary form, i.e., nutritional, and renal secondary hyperparathyroidism [16]. Specifically, the prevalence of RHPT is reported to be between 47% and 100% of cats affected with CKD (mean 84%), with a possible increased PTH concentration also in the early stages of the disease, before evidence of calcium and phosphorus imbalances [17,18]. It is also reported that the evaluation of PTH concentrations could help predict the development of azotaemia in geriatric normoazotemic cats [19]. According to these data, to achieve our aim, we included cats with azotemic CKD (except one in Stage 1) in which the PTH levels are often higher compared to healthy cats.However, PTH measurement is not routinely performed due to the limited availability of methods specific for feline PTH, and the known instability in serum samples, as previously reported [7]. The current reference method for the determination of feline PTH is the expensive third-generation whole-PTH radioimmunoassay [16]. For these reasons, the validation of a new immunoenzymatic method able to detect feline PTH, avoiding the employment of a radioimmunoassay, would be useful.In the present study, the ST-AIA PACK® Intact PTH assay showed satisfying results in the detection of feline PTH in serum samples and could be used also on plasma samples, even if specific data about possible discrepancies are not available. Intra- and inter-assay CVs were within the limit of 15%, which is considered acceptable for an immunoassay [20], especially for hormones that are usually affected by elevated variability. Only the inter-assay CV with the low concentration pool was very slightly above the limit (15.3% vs. 15%); however, even if the diagnostic accuracy of this method has not been evaluated in other studies and was beyond the scope of the present study, such low concentrations are generally found only in healthy cats and thus the imprecision may be considered acceptable from a clinical point of view. Overall, these results are quite similar, and in some cases lower, compared with those reported in other validation studies, in which the CVs were approximately constantly near 10% [7,8,9,21], underlining that the investigated method shows acceptable analytical repeatability and reproducibility.In absence of a current gold standard for the measurement of feline PTH, the indirect evaluation of accuracy using SRT and LUD showed an optimal concordance between the obtained and expected results, suggesting that this method can be considered reliable for PTH measurement also in cats.Interestingly, the lower limit of detection identified, about 1.1 pg/mL, was lower than those obtained with other methods, in which LLOD was 2 pg/mL, 3 pg/mL, and 5.2 pg/mL [8,9,22]. This analytical sensitivity will allow to correctly measure the PTH concentration in healthy cats and in cats with early stages of CKD, in which RHPT may be absent.The evaluation of storage stability was done to define the best storage condition that allows accurate measurements, considering that only limited and not recent data about feline PTH stability are available. Barber et al. reported good stability at −20 °C up until 30 days and suggested the freezing of the sample within 2 h of collection [7]. In our study, at room and refrigeration temperatures the deviation from the baseline exceeded 10%, which was established as a cut-off of tolerance, already after 6 h, leading to lower PTH values. This decrease was even more evident after 24 h, discouraging the use of long-time stored samples for PTH measurement in cats. This observation has two possible explanations: firstly, the proteolytic degradation of the molecule could decrease the serum PTH concentration, as reported for other hormones such as adrenocorticotropic hormone [23]; secondly, the high deviation from the baseline could be attributable to the relatively low concentration of PTH (11.05 pg/mL) in the sera used for the evaluation of storage stability. However, the decrease of 14.5% after 6 h at 4 °C and about 20% after 6 h at 20 °C and after 24 h at 4 °C could be considered acceptable for clinical purposes, especially in cats affected with CKD in which the PTH concentrations are significantly higher [19]. The good stability observed for 7 days at −20 °C suggests freezing the samples as soon as possible after collection and storing at freezing temperature before PTH measurement, as already suggested and required by some laboratories for other hormones [7].5. ConclusionsThis study demonstrates that the ST-AIA Pack Intact PTH is a precise and accurate immunoenzymometric assay for the detection of PTH in feline serum and, considering the very limited availability of assays for the measurement of feline PTH, this could be relevant in clinical practice for the diagnosis of feline primary and secondary hyperparathyroidism.The storage stability is limited but compatible with working turnaround times as determinations should be performed on the same day of sampling when stored at room temperature or refrigerated but can be performed within one week when stored at freezing temperature.

animals : an open access journal from mdpi

10.3390/ani11123409

PMC8697932

The potential of near infrared reflectance spectroscopy (NIRS) to predict the nutritive value of chickpea straw was identified. Spectral data of 480 samples of chickpea straw (40 genotypes) were scanned with a spectral range of 1108 to 2492 nm. The samples were reduced to 190 representative samples based on the spectral data then divided into a calibration set (160 samples) and a cross-validation set (30 samples). All 190 samples were analysed for dry matter, ash, crude protein, neutral detergent fibre, acid detergent fibre, acid detergent lignin, Zn, Mn, Ca, Mg, Fe, P, and in vitro gas production metabolizable energy using conventional methods. The prediction equations were generated by multiple regression analysis. The NIRS prediction equations in the study accurately predicted the nutritive value of chickpea straw (R2 of cross validation > 0.68; standard error of prediction < 1%). Chickpea straw nutritive value could be predicted using NIRS.

Multidimensional improvement programs of chickpea require screening of a large number of genotypes for straw nutritive value. The ability of near infrared reflectance spectroscopy (NIRS) to determine the nutritive value of chickpea straw was identified in the current study. A total of 480 samples of chickpea straw representing a nation-wide range of environments and genotypic diversity (40 genotypes) were scanned at a spectral range of 1108 to 2492 nm. The samples were reduced to 190 representative samples based on the spectral data then divided into a calibration set (160 samples) and a cross-validation set (30 samples). All 190 samples were analysed for dry matter, ash, crude protein, neutral detergent fibre, acid detergent fibre, acid detergent lignin, Zn, Mn, Ca, Mg, Fe, P, and in vitro gas production metabolizable energy using conventional methods. Multiple regression analysis was used to build the prediction equations. The prediction equation generated by the study accurately predicted the nutritive value of chickpea straw (R2 of cross validation > 0.68; standard error of prediction < 1%). Breeding programs targeting improving food-feed traits of chickpea could use NIRS as a fast, cheap, and reliable tool to screen genotypes for straw nutritional quality.

1. IntroductionChickpea is one of the key pulses in the world [1]. Cultivated chickpeas are categorized into two main groups: Desi and Kabuli [1]. Desi grains are small, dark in colour, and smooth or wrinkled and are preferred for use as flour [1]; however, they are used for direct cooking. Kabuli grains are large and cream in colour and contain less fibre than Desi grains. Thus, Kabuli grains are used for whole grain cooking [1]. The world yield of chickpea grains was 14,246,000 t in 2019, which accounts for 12% of the world pulse grain. Chickpea grains have high levels of protein, minerals, and vitamins for human consumption [1]. Moreover, growing chickpea improves soil fertility, land use intensity, and provides households with cash supply [2]. It has been reported that the production of 1 kg of chickpea grains is associated with production of 1.55 kg of straw [3]. Accordingly, chickpea cultivation generates large amounts of straw (~221,001,000 t), which is used for ruminant feeding. Chickpea straw contains 65 g/kg of crude protein (CP), 694 g/kg of neutral detergent fibre (NDF), 516 g/kg of acid detergent fibre, 111 g/kg of acid detergent lignin, and 7.7 MJ/kg of metabolizable energy [1]. Chickpea straw has a high content of antinutritional factors like tannins and oxalates [1].Yet, ruminant animals were reported to have a high tolerance to tannins and oxalates if the diet was adequately supplemented with protein and energy [1]. Chickpea straw is also used as soil mulch in mixed farming systems [4]. Accordingly, leveraging grain yield alongside straw yield and nutritive value will improve the biomass supply for human and livestock consumption and will trigger greater use of cereal straw for soil mulching [4]. Varietal selection to increase the yield and nutritive value of chickpea straw is promising [3,5,6]. The improvement programme of chickpea leaded by the International Centre of Agricultural Research in the Dry Areas (ICARDA) and national research centres in developing countries started recently to recognize the importance of chickpea straw’s nutritive value and started to include it as a selection criterion in their breeding programs. These programs are expected to produce large sets of chickpea straw samples to be analysed for nutritional value [7]. The conventional methods of feed evaluation cannot cope with this huge number of samples since they are expensive and time consuming. The near infrared spectroscopy (NIRS) technique has been applied for the nutritional characteristics of animal feeds [8,9]. Unlike most conventional analytical methods, the NIRS technique is fast, low-cost, and nondestructive to the sample. Additionally, NIRS requires very little sample preparation and no chemicals [10]. It is consistent, accurate, and fast [10]. Furthermore, NIRS can be used to analyse multiple feed nutrient properties at one time [8]. Generally used as a quantitative and qualitative analysis method, NIRS technology requires the development of prediction models, which involves multivariate analysis and analytical chemistry to extract the most relevant information [10]. The ability of NIRS technology to determine the nutritive value of chickpea straw was evaluated by Dereje et al. [11]. Yet, the application of the results of this study is limited because the number of chickpea samples was marginal and the genotypic and locational characters of the samples were unknown.Developing NIRS equations for chickpea straw nutritional analysis would facilitate the process of chickpea improvement for food and feed production, which would promote sustainable food production in mixed farming systems. Although the reliability of NIRS has been well investigated for temperate feeds, no studies have reported on the feasibility of the use of NIRS to determine the nutritive value of chickpea straw. Thus, the goal of the current study is to determine the accuracy and robustness of NIRS to determine the nutritive value of chickpea straw for screening the genotype in multi-dimensional improvement programs.2. Materials and Methods2.1. Sample Description and Experimental LayoutA total of 480 samples of chickpea straw from preliminary and national variety trials were used in the current study. The description of the sampling areas is presented in Table 1. The straw samples were collected after harvest, naturally dried for 7 days, and then stored in paper bags until they were analysed.2.2. Spectral Analysis of Chickpea Straw SamplesAll samples were ground (1 mm sieve size), dried at 60 °C overnight in an oven to standardize the moisture content, and then scanned using Foss NIRS 5000 with software package WinISI II in a spectral range of 1108 to 2492 nm (Win Scan version 1.5, 2000, intrasoft international, L.L.C, Luxembourg). Optical values were recorded as log 1/Reflectance.2.3. Chemical Analysis Using Conventional MethodsIn total, 190 samples were subsampled out of the 480 samples using the CENTRE algorithm based on NIRS spectra data [12]. The samples were analysed for dry matter (DM), ash, and crude protein (CP) according to the methodology of AOAC [13]. The sample was ashed in a muffle furnace at 500 °C overnight (method 942.05) to determine the ash content. The Kjeldahl method using Kjeldahl (protein/nitrogen) Model 1026 (Foss Technology Corp., Hilleroed, Denmark) was used to identify the level of nitrogen of the sample (method 954.01). The nitrogen content was converted to crude protein by a conversion factor of 6.25. Van Soest et al. [14] was used to determine the NDF and ADF content. Neutral detergent fibre and ADF were expressed exclusive of residual ash. Neutral detergent fibre analysis did not include a heat-stable amylase. The lignin content of the sample was analysed by solubilisation of cellulose with sulphuric acid. Metabolizable energy was measured in rumen microbial inoculum using the in vitro gas production method. The in vitro gas production method ([15]) was used to prepare the buffer solution. Rumen fluid was obtained before morning feeding using a vacuum pump from three ruminally cannulated cows. The cows were fed grass hay (790 g/kg), wheat bran (203 g/kg), salt (3.2 g/kg), and a mineral and vitamin mixture (4.6 g/kg). Handling of the cows and rumen fluid sampling was approved by the Environmental and Occupational Health and Safety unit of International Livestock Research Institute. Fluids were composited (1:1, v/v), filtered through four layers of cheesecloth, and added to the buffer solution (1:2, v/v), which was kept in a water bath at 39 °C under continuous CO2 flushing. The buffered rumen fluid (30 mL) was transferred into 100 mL syringes containing 0.2 g of straw sample and immediately placed into a water bath at 39 °C. The 24 h gas production was recorded and used to calculate ME according to Menke and Steingass [16]. Phosphorous, calcium, magnesium, manganese, iron, and zinc contents were determined by an atomic absorption spectrophotometer (A. Analyst 300, Perkin Elmer, Shelton, CT, USA).2.4. Statistical AnalysisThe spectral data were not subjected to any mathematical treatment. The CENTRE algorithm was used to calculate the GH value (Mahalanobis global distance to the centre of the population) of each sample. Samples with GH ≥ 3 were considered as outliers. No outliers were found in the data; thus, all 190 samples were divided into two groups: a calibration group (160 samples) and a cross-validation group (30 samples). Multiple regression analysis was used to build the equations using the calibration group. The chemical composition of the validation set was predicted using the prediction equations and then the standard error of prediction was calculated (SEP). Calibration equations were evaluated using the coefficient of determination (R2), standard error of calibration (SEC), and SEP. Statistical analysis of the data was performed using WinISI II software.3. ResultsThe means and standard deviation of the chemical composition, mineral analyses, and ME of chickpea and straw for both predicted and reference samples are presented in Table 2. There was wide variation in the chemical composition and ME of chickpea samples, which cover most of the variability reported in the literature.3.1. CalibrationResults of the NIRS calibration are presented in Table 3. The R2c of the chickpea straw’s chemical composition ranged from 0.84 in DM to 0.99 in ADL. The R2c of Me was high (0.99). R2c Zn, Ca, Fe, Mn Mg, and P ranged from 0.71 to 0.93. The SEC of the proximate analysis ranged from 0.19% in DM to 0.85% in NDF. The SEC of ME was relatively small (0.06 MJ/kg DM) and the SEC of the mineral composition of chickpea straw ranged between 0.33% and 1.92%.3.2. ValidationThe results of the NIRS equations validation are shown in Table 3. R2v of the CP, NDF, ADF, ADL, and ME was higher than 0.96. Although, DM had a lower R2v of 0.78. The mineral composition of chickpea straw had a high R2v ranging between 0.68 and 0.92. The SEP of the proximate analysis and ME of chickpea straw ranged from 0.036% to 1.3%. The SEP of the mineral composition of chickpea straw Zn, Ca, Mg, Mn, Fe, and P was less than 3%.4. DiscussionImproved varieties of chickpea produce significantly higher yields of grains compared to the local genotypes. They have high tolerance to drought and disease [17]. Thus, adopting these varieties would enhance the food security of developing countries and improve the livelihoods of farmers relying on chickpea production. National and international research centres have recognised that many improved varieties of crops are rejected by farmers in mixed farming systems due to the low palatability and nutritive value of straw [18,19]. This is a huge threat to the agricultural production and food security of a large number of households in mixed farming systems. Accordingly, straw’s nutritive value as a livestock feed became a priority in crop breeding programs to raise the rate of adoption of the improved varieties on the one hand and to enhance the nutritive value of the straw for livestock feeding on the other hand. Chickpea breeding programs generate huge sets of samples requiring screening for nutritive value in a short time [7]. Under such a scenario, conventional lab methods are not feasible as they are expensive, time consuming, and environmentally destructive. Here, NIRS offers a cheap, fast, and reliable method to accurately determine the nutritive value of a range of animal feed [20]. Furthermore, NIRS technology for feed analysis does not include dealing with any chemicals and does not have any animal welfare issue related to ruminal cannulation.R2c and R2v of the chemical composition of chickpea straw were 0.78 for DM and higher than 0.96 for ash, CP, NDF, ADF, and ADL. This indicates a high match between the NIRS spectral data and conventional lab data of DM, ash, CP, NDF, ADF, and ADL of chickpea samples. In addition to this, SEC and SEP of NIRS equations for proximate analysis and cell wall constituents of chickpea straw were lower than 1%. Thus, these NIRS equations could predict the chemical composition of chickpea straw using spectral data. This is in agreement with previous studies, which reported that NIRS is an accurate method to predict the nutritive value of a range of animal feed [21,22,23,24,25,26].R2c and R2v of ME of chickpea straw were high (0.99), indicating a strong correlation between the spectral data and in vitro ME data. The NIRS prediction equation of ME had low SEC and SEP (<0.1%). Therefore, the NIRS model produced from the current study is accurate in determining ME of chickpea straw samples. This is in agreement with results in the literature, which reported on the possibility of predicting some biological parameters including digestibility and ruminal gas production of feed using their spectral data ([22,26], respectively). Theoretically, the mineral composition of feed is not detectable by NIRS because their structure does not have organic bonds. However, minerals can be predicted if they are included in organic complexes [27] or due to the change that minerals cause in the water region of the spectrum [27,28,29]. The mineral composition of chickpea straw included in the current study had high R2c and R2v. Accordingly, the association between the spectral data and mineral composition of chickpea straw is high. Furthermore, both SEC and SEP were low (<1%). Therefore, the mineral composition of chickpea straw could be accurately predicted via the NIRS models generated by the current study. This is in disagreement with Goi et al.’s [30] study on the mineral composition of dog food, where Ca, P, Na, and Mg were poorly predicted using the NIRS equations. On the contrary, K and Na were accurately predicted using NIRS data [30]. Furthermore, minerals were predicted from NIRS data with low error in cheese (Ca, P, S, Mg, Zn, and Cu [31]; Na and K [32]), meat (Ca and Zn [33]), and processed meat products (Na [34,35,36]). This could be because most of the minerals in chickpea straw might be linked to oxalate in a form of oxalate salts or they are bound to other organic compounds. Near infrared reflectance spectroscopy could be used to facilitate chickpea breeding programs for food-feed improvement by offering an accurate, cheap, and fast tool for nutritional value determination.5. ConclusionsStationary NIRS is a feasible replacement technology to the conventional nutritional analyses to determine the chemical composition, ME, and mineral content (Zn, Ca, Mg, Mn, Fe, and P) of chickpea straw. This would decrease the cost (as it requires only technician time and cost of sample processing) and time of the analysis and environmental hazard associated with the conventional nutritional analysis methods. Chickpea breeders can screen chickpea genotypes for straw nutritive value using NIRS when chickpea programs target improving food and feed traits. These programs would release chickpea genotypes that are superior in grain yield and straw nutritive value, leading to more sustainable food production and food security in mixed farming systems.

animals : an open access journal from mdpi

10.3390/ani11051223

PMC8146882

The price of feed ingredients is one of the most important and most costly constraints facing the aquaculture industry, and fishmeal is one of the most important and most expensive of these ingredients. As a result, recent research has focused on the use of alternative low-cost feed ingredients that are less reliant on fishmeal. Therefore, there has been considerable interest in the use of various types of plant protein (PP) meals in aquafeeds (soybean and sunflower meal). This study reveals that soybean and sunflower meal with methionine and lysine supplementation can be used on a larger scale in the aquafeed industry as substitutes of up to 50% of fishmeal.

African catfish (Clarias gariepinus) were subjected to a 30-day feeding trial to determine the appropriateness of using plant protein (PP) (soybean and sunflower meal) as a fishmeal (FM) replacement in the diet and its effects on immune status, antioxidant activity, pituitary adenylate cyclase-activating polypeptide (PACAP) gene expression, and disease resistance. A total of 150 C. gariepinus (51.01 ± 0.34 g) were randomly distributed among five groups in triplicate. Five experimental diets were formulated to replace 0 (control), 33.5, 50, 66.5, and 100% FM with soybean and sunflower meal to form the experimental diets (R0, R33.5, R50, R66.5, and R100, respectively). After 30 days, the diet containing PP for FM had no significant impact on total, and differential leukocyte counts determined at the end of the feeding period. The total globulin concentration showed significantly greater differences in the following order R0 > R33.5 > R50 > R66.5 > R100. The R0 group had the highest concentration of serum γ-globulin, while R100 had the lowest concentration. The antioxidant status complements 3 (C3), lysozyme activity (LYZ), and antiprotease activity were not significantly different between R0, R33.5, and R50 groups, while they were significantly lower in R100. The serum nitric oxide activity (NO) exhibited significantly greater differences in the following order R0 > R33.5 > R50 > R66.5 > R100. PACAP was significantly higher in the R33.5 group. The highest cumulative mortality caused by Aeromonas sobria was recorded in the R100 group (60%) and the lowest in the R0 group (30%). In conclusion, the results indicate that the immunological responses and antioxidant status of C. gariepinus were not affected when they consumed a diet with FM replaced by up to 50% with PP (SBM and SFM) with methionine and lysine supplementation, but total globulin, NO, and cumulative mortality were impaired with a diet containing a 100% FM replacement.

1. IntroductionThe development of the fish industry and the expansion of aquaculture have become necessary internationally, particularly in developing countries, because of the nutritional, social, and economic value of fish. A settled source of fish is crucial for the nutrition and health of a wide segment of the globe [1]. Moreover, this sector provides employment and income in many developed and developing countries [2].African catfish, Clarias gariepinus is a popular commercial fish species found in Africa, Europe, China, and Brazil, and is one of the most promising species for aquaculture production due to its distinct genetic and productive characteristics. C. gariepinus is mainly characterized by its high growth rate, which is estimated to be at least three times that of other fish species [3]. It can be raised in a high-intensity culture without daily water exchange because its airbreathing capacity allows it to withstand poor water quality and low dissolved oxygen [4]. Furthermore, C. gariepinus is distinguished by a wide range of tolerance to high salinity levels [5], the ability to withstand various stresses [6], high fecundity [7], and high palatability.Achieving the highest profit percentage with the lowest cost percentage is critical to the growth of any industry. In the aquaculture industry, the prices of certain feed ingredients are among the major challenges affecting the industry’s success or failure. Among them, fishmeal (FM) is the main ingredient in aquafeed and the costliest [8,9]. FM, a conventional, highly palatable dietary protein source in aquafeed, is prepared from pelagic fish and contains balanced amino acids, essential n-3 fatty acids, vitamins, and minerals [2,10,11]. Over the last few years, researchers and fish feed manufacturers have increased their efforts to find the best alternative sources of FM, especially plant protein (PP) sources for various reasons, including a decline in pelagic and wild fish stocks leading to a shortage in their supply for FM production, and increased prices [12]. In addition, a greater proportion of pelagic species are increasingly being marketed as human food [2]. In addition, aquafeed consisting of more than 20% FM is hazardous to the ecosystem because the phosphorus content exceeds the requirements for fish and is released in urinary and fecal excretion, but PP contains less phosphorus than FM [13,14,15].Therefore, there has been considerable interest in the use of various types of PP meals in aquafeed for the partial or total replacement of FM, such as soybean, cottonseed, pea seed, cassava leaf, sunflower, rapeseed meal, and rice protein meal [16,17,18,19,20,21]. Soybean meal (SBM) is the most widely used substitute protein source for fishmeal in fish and shrimp diets because it has many features and characteristics. SBM is readily available, inexpensive, rich in highly digestible protein with a balanced amino acid profile, easy to digest, and low in phosphorus compared to FM [22,23,24]. Although soybeans have some antinutritional factors, they can be reduced or eliminated through a variety of methods, including heat treatment during their manufacturing [25], via fermentation [26,27], the application of solvents [28], and alcohol extraction [29]. The cost of aquafeed can be further reduced by using other plant protein sources such as sunflower meal (SFM), which is less costly than FM and SBM [19]. SFM contains highly digestible protein with many amino acids (especially glutamic and aspartic acids) except lysine, and sulfur amino acids [30,31]. The essential amino acids that cannot be synthesized by the body, methionine and lysine, are obtained from the diet. Supplementing a fish diet with the recommended amounts of synthetic methionine and lysine will help minimize the oxidation of other amino acids and increase their utilization [32]. Dietary methionine and lysine are needed for protein, lipid, and energy metabolism, as well as carnitine biosynthesis [33]. Furthermore, the digestibility of feed ingredients is a vital factor that will indicate the nutritional balance of an aquatic animal’s diet [34]. Supplementation with amino acids, especially methionine and lysine, in diets high in plant protein ingredients has been shown to improve digestibility and, as a result, improve the growth and health of various fish species [35,36]. Previous research has also shown that any diet with unbalanced amino acid concentrations causes increased protein degradation [37].Aeromonadaceae are an opportunistic Gram-negative bacterium that cause mass mortality and great economic losses in different fish species [38,39,40]. A balanced diet helps to protect fish from pathogens. Any imbalances resulting from the substitution of FM are therefore considered potent stressors that adversely affect fish welfare, immune status, and resistance to pathogens. Furthermore, several studies have shown that the duration of experimental diet administration had an impact on the immunological parameters, with immunological reactions often appearing in the first days of feeding (from 7 to 30 days) [41,42,43].While FM replacement studies have mainly focused on growth, digestibility, and body composition, few have investigated the impact of this replacement on fish immunity, health, and disease resistance [19,44,45]. Accordingly, the current study was aimed at estimating the possibility of integrating PP (SBM and SFM) into practical diets C. gariepinus by replacing FM for 30 days and assessing its effects on immune status, antioxidant activity, pituitary adenylate cyclase-activating polypeptide (PACAP) gene expression, and disease resistance.2. Materials and Methods2.1. Ethical StatementThe experimental procedure was approved by the Ethics Committee of Institutional Animal Care and Use, Zagazig University, Egypt (ZU-IACUC/2/F/139/2020).2.2. FishA total of 150 African catfish (51.01 ± 0.34 g) were procured from a private fish farm in Sharkia province, Egypt. The fish were transported in polyethylene bags filled with one-third dechlorinated water and two-thirds oxygen before sunrise to avoid heat and sunshine. When the fish arrived at the laboratory of the Fish Diseases and Management Department, Faculty of Veterinary Medicine, Zagazig University, Egypt, the water in the laboratory aquaria was slowly mixed over a period of time with the water the fish were transported in so as avoid a sudden change in water quality. To allow acclimation to the laboratory conditions, fish were randomly distributed in glass aquaria (80 × 40 × 30 cm, water capacity 60 L) for 14 days and fed twice daily with a basal diet (at 3% of biomass).2.3. Experimental Diet and DesignFive isocaloric and isonitrogenous (32% crude protein) diets were formulated to replace 0 (control), 33.5, 50, 66.5, and 100% FM with plant protein sources (SBM and SFM) to form the experimental diets (R0, R33.5, R50, R66.5, and R100, respectively) [46,47] as shown in Table 1. The experimental diets were supplemented with methionine and lysine (as synthetic amino acids) and dicalcium phosphate to maintain the dietary requirements. The ingredients were mixed well and pelleted by passing them through a meat mincer (3 mm). This was then dried at room temperature and stored in the refrigerator (4 °C) during the feeding period. The proximate amino acid and chemical analyses were performed according to the Association of Official Agricultural Chemists [48] and Llames and Fontaine [49].The fish were randomly distributed into 5 groups, and each group was in triplicate (30 fish/group with 10 fish/replicate). The first group (R0) was fed a control basal diet consisting of FM as the sole source of protein. The other 4 groups (R33.5, R50, R66.5, and R100) were fed experimental diets, in which FM was replaced by PP sources (SBM and SFM) for 30 days as mentioned in Table 1. The fish were fed manually at a rate of 3% from total fish biomass twice daily (08:00 and 15:00), and the quantity of food was adjusted every 2 weeks according to changes in fish body weight. Water parameters were kept at standard values throughout the experimental period (water temperature 27.5 ± 0.5 °C, pH 6.7 ± 0.2, ammonia 0.020 ± 0.001 mg/L, and nitrite 0.013 ± 0.003 mg/L). One-fourth of the volume of aquarium water was exchanged daily.2.4. Sampling and Analytical Methods2.4.1. Blood Sample CollectionBlood samples were collected for serum separation after 15 and 30 days of experiment. At each point in time, 9 fish/group were collected and anesthetized with 100 mg/L of benzocaine solution (Al-Nasr pharmaceutical Chemicals Co., Oubour, Qalyubia, Egypt) [50]. Blood samples were collected from caudal vessels without anticoagulant for the separation of serum (centrifugation at 3000 rpm/15 min at 4 °C). The serum samples were held at −20 °C until analysis. For the analysis of total and differential leukocyte counts, the blood samples (9 fish/group) were collected after 30 days using EDTA-rinsed 1-mL syringes. After completing the blood sampling collection, the 9 fish were dissected to obtain samples for gene expression analysis, as explained below.2.4.2. Leukogram, Serum Total Protein, and Electrophoretic Fraction AnalysisTotal and differential leukocyte counts were determined by using a Sysmex XT-2000iV Automated Hematology Analyzer (Sysmex Corporation, Hyogo, Japan) at the Animal Health Research Institute, Zagazig Branch, Egypt. Serum total protein (g/dL) [51] and albumin (g/dL) [52] were estimated, and total globulin was calculated by subtracting albumin from total protein. Serum protein was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), according to Laemmli [53].2.4.3. Serum Oxidant/Antioxidant StatusSerum levels of total antioxidant capacity (TAC) (mM/L), superoxide dismutase (SOD) (U/mL), catalase activity (CAT) (U/L), and reduced glutathione (GPx) (mmol/L) were estimated using commercial ELISA test kits (Cusabio Biotech Co., Ltd., Wuhan, China) following the manufacturer’s instructions.2.4.4. Nonspecific Immune AnalysisNonspecific immune parameters such as serum alternative complement pathway C3 activity were examined using a C3 kit (Zhejiang Elikan Biological Technology Co., Ltd., Wenzhou, Zhejiang, China). Reagent 1 was slowly blended with distilled water (2 mL), standard liquid, and serum. The absorbance was estimated at 340 nm after a 5-min incubation at 37 °C. The tubes were then filled with Reagent 2 and the absorbance was measured. A C3 standard curve was created using the same process. C3 levels were determined according to [54]. For serum nitric oxide (NO) (µmol/L), Griess reagent was applied to 100 mL of each serum sample in a microtiter plate and incubated at 27 °C for 10 min, then NO level was measured using a spectrophotometric method [55]. Using a turbidimetric assay, the activity of serum lysozyme (LYZ) (µg/mL) was determined by dissolving 0.75 mg/mL Micrococcus lysodeikticus (Sigma-Aldrich Chemie GmbH, Darmstadt, Germany ) in 0.1 M sodium phosphate buffer (pH = 5.9) and placing 175 mL of the suspension on a microtiter plate and incubating at 30 °C for 5 min. After that, 25 L of serum was applied to the microtiter plate. A lyophilized hen egg white lysozyme (Sigma) standard curve was used to measure the serum lysozyme levels (μg/mL) [56]. The antiprotease activity (mg/dL) was measured by incubating 10 mL of serum with 20 mL of trypsin solution (0.25% bovine pancreatic trypsin in 0.02% EDTA; Beyotime, Jiangsu, China) for 10 min at 25 °C. Then, 500 mL of 2 mM BAPNA (sodium-benzoyl-DL arginine-p-nitroanilide HCl; Himedia, Giza, Egypt) was added. TriseHCl (0.1 M, pH 8.2; Beyotime, China) was added to a final volume of 1 mL. The mixture was incubated at 22 °C for 25 min. The reaction was halted with 150 mL of 30% acetic acid, and the optical density (OD) was calculated at 415 nm against a blank using a BioRad microplate reader (Hercules, CA, USA). The inhibitory activity of the antiprotease was expressed as the percentage of trypsin inhibition, calculated as follows: ((optical density "OD" of trypsin blank_OD sample)/OD of trypsin blank) × 100. according to Bowden et al. [57].2.4.5. Expression of Pituitary Adenylate Cyclase-Activating Polypeptide GeneThe same 9 fish used for the collection of blood samples at the end of the experiment (30 days) were euthanized using a benzocaine solution overdose (250 mg/L) and used for gene expression analysis to collect spleen samples. The total RNA was extracted using the QIAamp RNeasy Mini Kit (Qiagen, Hilden, Germany, GmbH) according to the manufacturer’s protocol. Complementary DNA was synthesized following the manufacturer’s instructions (Quantitect® Reverse Transcription kit, Qiagen, Germany). Quantitative real-time polymerase chain reaction (PCR) analysis was performed with the SYBR green PCR master mix (Step One Plus, Applied Biosystem, Waltham, MA, USA). Full-length elongation factor 1 alpha (EF-1α) (GenBank accession number AB075952.1) Available online: https://www.ncbi.nlm.nih.gov/nuccore/AB075952.1, (forward primer: 5′-CCTTCAACGCTCAGGTCATC-3′; reverse primer: 5′-TGTGGGCAGTGTGGCAATC-3′) was chosen as internal standard as suggested by Gröner et al. [58]. The target gene was PACAP (GenBank accession number = EF524513) Available online: https://www.ncbi.nlm.nih.gov/nuccore/EF524513, (Forward primer: 5′-CACTCGGACGGCATTTTCACGG-3′; Reverse primer: 5′–TTTGTTTCTAAACCTCTGTCTGTAC-3′) [59]. The amplification conditions were as follows: 40 cycles of 94 °C for 15 s, 65 °C for 30 s, and 72 °C for 30 s. The amplification efficiency of the used primer was determined by standard curve assay. Amplification efficiency was above 98% for each group.2.5. Aeromonas sobria Challenge TestSeven fish/replicate (21 fish/group) were challenged by intraperitoneal injection with 1.5 × 107 cells/mL (adjusted with McFarland standard tubes) of Aeromonas sobria (A. sobria) previously isolated from moribund fish at the Department of Fish Diseases and Management, Faculty of Veterinary Medicine, Zagazig University (project no. 5589) and confirmed to be pathogenic. The fish were observed for 15 days, and their clinical signs, postmortem lesions, and daily mortality were recorded. For confirmation, the pathogenic bacterial strain was re-isolated from the liver, kidney, and intestine of dead fish.2.6. Statistical AnalysisThe data were first examined for homogeneity and normality. One-way ANOVA (SPSS version 16.0, IBM, Chicago, IL, USA) was performed to detect the difference among experimental groups within the same period. The group means were analyzed with Duncan’s multiple range tests and presented as the mean ± SE. Two-way ANOVA (Statistica software, version 8, StatSoft, Inc., 2008, Tulsa, OK, USA) was used to determine the interaction between the levels of FM substitution with PP and the feeding duration. For every test, the level of significance chosen was p ≤ 0.05.3. Results3.1. Leukogram, Serum Total Protein, and Electrophoretic FractionThe data in Table 2 indicate that there were no significant differences in total and differential leukocyte counts of fish fed the experimental diets compared to the control group. Table 2 displays the concentrations of total serum proteins and their fractions. After 30 days of feeding, the highest total serum protein was recorded in the R0 and R50 groups, while there was no significant difference between R33.5, R66.5, and R100 in total serum protein. The total globulin concentration was significantly different in the order of R0 > R33.5 > R50 > R66.5 > R100. The concentrations of α1 and α2 globulin decreased significantly in the R100 and R66.5 groups compared to other groups, while the levels were not significantly different in R0, R33.5, and R50. The serum ß- globulin concentration was significantly higher in the serum of R0 and R33.5 groups, followed by the R50 group, while the lowest concentration was recorded in the R66.5 and R100 groups. The highest concentration of serum γ- globulin was recorded in the R0 group, and the lowest in the R100 group.3.2. Serum Oxidant/Antioxidant ActivityAfter 15 days of feeding, the serum TAC and GPx levels were significantly higher in the R0 group and significantly lower in the R50 group, while there was no significant difference in TAC in the groups that were fed substituted diets. SOD and CAT in serum were significantly higher in the R0 group and lowest in the R66.5 group. After 30 days of feeding, there was no significant difference in the TAC of the R0, R33.5, and R50 groups, which was significantly higher than that in R66.5 and R100 groups. After 30 days, the serum SOD and CAT levels were significantly higher in the R0 and R33.5 groups and significantly lower in the R100 group. The serum GPx level was significantly higher in the R0 group and significantly lower in the R66.5 group (Figure 1).The effects of the interaction between feeding duration and levels of FM substituted with PP were assessed, and the results reveal that, in all groups, the TAC level was dependent on time and level of FM substitution except in the R66.5 and R100 groups, where the TAC demonstrated better results after 15 days than 30 days of feeding (Figure 2A). This interaction between feeding duration and level of substituted PP had a different effect on the SOD level, which differed between groups, although it was not significantly different between R33.5 and R50 after 15 and 30 days. Over time, the SOD level decreased significantly in R0 and R100 and increased significantly in the R66.5 group (Figure 2B). There was a significant increase in serum CAT in all groups after 30 days compared to 15 days (Figure 2C). In contrast, the GPx level decreased significantly at the end of 30 days compared to 15 days of feeding in all groups except R50 (Figure 2D).3.3. Nonspecific Immune ParametersThe nonspecific immune parameters after feeding C. gariepinus diets in which FM was substituted with PP are shown in Figure 3. After 15 days of feeding, serum C3 was significantly higher in the control group (R0) than the other experimental groups, and there was no difference between other groups. Serum NO and LYZ activity was significantly higher in R0 and significantly lower in R100, and there was no difference between other groups fed substituted diets, which were lower than R0. There was no significant difference in serum antiprotease activity among the groups. Similarly, after 30 days of feeding, serum C3 was significantly highest in R0, and there was no difference among other groups, which were lower than R0. Serum NO activity exhibited a significantly greater difference in the following order R0 > R33.5 > R50 > R66.5 > R100. Serum LYZ activity showed no significant differences between groups. Serum antiprotease activity decreased significantly in R100, and there was no significant difference between the other experimental groups.Serum C3 and NO activity in fish was significantly higher after 15 days than after 30 days of feeding on the substituted diet. Serum LYZ activity increased significantly with time and feeding on the substituted diet only in R33.5 and R50. Similarly, serum NO activity increased significantly only in R50 and R66.5 with time when fish were fed the substituted diets (Figure 4).3.4. PACAP Gene ExpressionPACAP cDNA was significantly higher in the R33.5 group and was not significantly different in other groups in comparison to R0 (Figure 5).3.5. Resistance to Aeromonas sobriaAt the end of the feeding trial (30 days), the C. gariepinus were challenged with A. sobria. Different external clinical signs were recorded, such as sluggish movement and destruction of fin rays, especially dorsal and caudal fins, with hemorrhagic spots at the base. Internally, congestion and enlargement of internal organs, especially liver and kidney, were the main findings. The highest cumulative mortality was observed in the R100 group (60%), followed by R66.5 (50%), R33.5 and R50 (both 40%), and R0 with the lowest percentage (30%) (Figure 6).4. DiscussionAquafeeds incur high production costs; thus, the interest in using low-cost food processing by-products has increased. This new source of aquafeed production has reduced the need for expensive waste management systems and the reliance on expensive protein sources such as FM [60]. Not only does the wrong choice of protein source result in reduced weight gain and a longer time to reach marketable size, it also raises stress levels, leading to a declined physiological and immunological status of fish, with increased susceptibility to disease [24,61]. The current study, therefore, evaluated the effects of substituting FM with economic sources of PP on immune status, antioxidant activity, PACAP gene expression, and disease resistance.Generally, the analysis of white blood cells and protein fractions serves as a diagnostic tool to measure the physiological status, nutritional condition, and health status of fish. Any changes in these parameters indicate an adverse condition contributing to stress on fish health [62,63,64]. The present study indicates that consuming a diet in which FM was substituted with PP did not have a significant effect on total and differential leukocyte counts. In an earlier study, a diet in which FM was replaced with PP sources (wheat gluten, corn gluten, and soybean meal) by 40, 70, and 100% did not significantly affect the blood parameters hematocrit, hemoglobin, white blood cell, heterophil, and lymphocyte counts of Oncorhynchus mykiss after 60 days of feeding [65]. Moreover, a diet in which SBM was replaced with SFM and sesame seed meal by 15, 30, and 45% had no significant effect on the leukocyte, eosinophil, monocyte, and lymphocyte counts of C. gariepinus [66]. Yue and Zhou [67] reported a substantial improvement in the white blood cell and hematocrit values of hybrid tilapia, Oreochromis niloticus × O. aureus when the fish consumed feed with cottonseed meal (CSM) replacing SBM from 0 to 45% but these values decreased dramatically when the amount of CSM was increased from 60 to 100%.Antioxidant defense in fish involves enzymatic and non-enzymatic activities. The radical scavenging enzymes have different modes of action; some, such as SOD, act on superoxide (O2–), while others, such as CAT, act on hydrogen peroxide (H2O2), and glutathione peroxidase (GSH-PX) scavenges H2O2 and lipid hydroperoxides [68], which protects cells and tissues from oxidative damage. Antioxidant enzymes are important biomarkers of fish health and immune status [69]. Nutritional factors influence antioxidant defense in fish but, until now, there have been conflicting reports on the types and levels of dietary ingredients that exert such influence [70].In the present study, no statistical difference in antioxidant activity was observed in the experimental groups, except for the dietary group receiving the lowest level of FM. Compared to the control group (R0), the antioxidant activity of the R66.5 and R100 groups diminished, and this corroborates with earlier research. O. mykiss spleen SOD and kidney CAT activities decreased significantly after feeding for 8 weeks on a diet with 20 and 30% of FM substituted with soy protein and meat bone meal [71]. Hepatic antioxidant activities (TAC, SOD, CAT, and GSH-Px) of Pseudobagrus ussuriensis decreased with increase in FM substituted with cotton meal, especially at 60% replacement after 8 weeks of feeding [72]. TAC in the serum of Scophthalmus maximus decreased significantly after feeding for 66 days on a diet with 45% of FM replaced by SBM [73]. CAT and TAC in the hepatopancreas of Litopenaeus vannamei decreased significantly in the group fed a low-FM diet (FM substitution of 15 to 25%) after 8 weeks of feeding [74]. Similarly, SOD activity decreased significantly in the hepatopancreas of L. vannamei fed a diet with 25% of FM substituted with SBM for 12 weeks [75].The reduction in antioxidant index with a decrease in FM levels in the R66.5 and R100 groups indicates a disturbance in the antioxidant activity. This breaks the dynamics of free radical production and removal in fish, which could be linked to undegraded anti-nutritional factors (ANFs) in PP sources [72,73,74]. On the contrary, some studies have shown increased antioxidant activity in fish fed diet with FM substituted with a PP source [76,77,78,79,80], which they attributed to the presence of phenolic [76] and flavonoid [81] compounds in the plant constituents. These compounds could be stimulators increasing the antioxidant activity. A variety of factors such as feed ingredients and their levels, time, fish species, size, feeding behavior, and environmental factors, could reflected in the variations of antioxidant defense activity in fish [82].The nonspecific immune system in fish is more important than the specific immune system in terms of tolerating diseases because the latter takes relatively more time to produce antibodies and induce specific cellular activations [83]. In our study, we focused on C3, NO, LYZ, and antiprotease activity. The complement system consists of 35 plasma proteins and plays a crucial role in innate and adaptive immunity by alerting and the host to the presence and the clearance of possible pathogens [84]. NO is normally derived from fish macrophages, which, due to its powerful killing effects and its function as a deactivator of some particular enzymes involved in macrophage cytotoxic reactions, is integral to fish antimicrobial immunity [85,86]. LYZ is a mucolytic enzyme of leucocytic origin that primarily inhibits the invasion of Gram-positive bacteria by breaking the cell wall linkages between N-acetylmuramic acid and N-acetylglucosamine [87,88]. Meanwhile, the blood protein antiproteases protect fish tissues from lysis by antagonizing the proteolytic activity of microorganisms [89]. No doubt, there is a close relationship between the dietary components, especially proteins, ingested by the organism and its immunity. Proteins consist of nitrogen, carbon, hydrogen, and oxygen and are present in the cell-foundation, enzymes, certain hormones, and the defense mechanism of any organism [90,91]. Besides the substantial role of protein in the defense mechanism of the immune system by producing natural and acquired immunity against pathogens, the immune systems relies on it for the development of active protein compounds and cell replication [91].Some studies have investigated the impact of PP on fish immunity, and there are a few inconsistencies in their findings. The two main issues that lead to these discrepancies are the imbalance of important amino acids, especially methionine and lysine, whose levels are always below what is needed by the fish [92,93]. The second issue is the presence of ANFs in PP [93,94]. We believe that the different approaches to overcoming these two key issues caused the discrepancies in the findings on the impact of using PP on fish immunity, in addition to several other factors. Among the other factors, those that involve nutritional requirements include fish species, body size, feed ingredients other than PP, differences between fish species in handling and digesting various nutrients, and differences in environmental factors [42,80,95,96].In the current study, after 15 days of feeding, the nonspecific immune parameters were significantly higher in R0 than in the other groups fed a diet with FM substituted by PP. By the end of the experiment (after 30 days), there was no significant difference between R0, R33.5, and R50 groups in C3, LYZ, and antiprotease activity, while R100 had significantly lower values for these parameters. Moreover, Jalili et al. [65] reported an insignificant effect on lysozyme activity and total antibodies in O. mykiss fed a diet with 40% FM substituted with PP sources (wheat gluten, corn gluten, and SBM) for 60 days. Alternative complement activity and total serum antibody were significantly lower in the groups fed diets with 70 and 100% FM substitution compared to the control [65]. Likewise, nonspecific immune responses were suppressed in O. mykiss when they consumed a diet containing more than 60% of FM substituted with soybean proteins for 53 days [97]. Serum LYZ of S. maximus decreased significantly in the group fed with 60% SBM for 66 days [73]. Moreover, the hemolymph phagocytic activity of Macrobrachium nipponense decreased when the proportion of FM decreased after 8 weeks of feeding [79]. Storebakken et al. [98] reported that the fish immune system was weakened when fed with high levels of PP, especially SBM, and attributed this to the decreased availability of vitamins and minerals, which are cofactors for metabolic enzymes, antioxidation, and the immune response.PACAP is a regulatory neuropeptide that belongs to the glucagon superfamily [99]. A bidirectional relationship exists between the neuroendocrine system and its influence on immune functions. Neuropeptides are produced from the lymphoid tissue microenvironment, and corresponding immune cell neuropeptide receptors mediate neuroimmune interactions [100]. Some studies investigated the role of PACAP in the stimulation of innate and acquired immunity of fish [59,100,101]. However, to the best of our knowledge, this is the first study to investigate the effect of substituting FM with PP on PACAP gene expression. In the present study, the PACAP level was significantly higher in the R33.5 group, but there was no substantial difference between the other groups relative to R0. Mokrani et al. [80] indicated that dietary PP did not cause inflammation in blunt snout bream hepatopancreas and recorded a significant decrease in expression levels of interleukin 8 (IL-8), tumor necrosis factor alpha (TNF-a), and nuclear factor-kappa B (NF-kB) in FM-reduced groups. On the other hand, several intestinal transcriptomes of immune-related genes were upregulated in Salmo salar when they consumed a diet containing FM substituted with 20% SBM [102,103]. Moreover, Hedrera et al. [104] recorded an increase in the pro-inflammatory cytokine mRNA levels in Danio rerio larvae when they consumed SBM-containing diets at 50%, which may be related to SBM ANFs [103].The bacterial challenge test is mostly used after the feeding trial as a final measure of fish health status [105]. A. sobria is a hemorrhagic septicemic Gram-negative bacterium that has been used as an indicator of fish immunonutrition in some studies [106,107]. In the present study, the highest and lowest percentage of cumulative mortality caused by A. sobria were recorded, respectively, for the R100 group at 60% and the R0 group at 30%. The results of cumulative mortality in this study could be confirmed by the immune parameter studies (Figure 3). Jalili et al. [65] recorded no major change in the mortality rate of O. mykiss consuming FM diet meal for 60 days, when 0, 40, 70, and 100% FM was substituted with PP (wheat gluten, corn gluten, and SBM). The mortality rates of S. salar infected with Vibrio anguillarum were not affected after 56 days of feeding on a diet with FM replaced with dehulled lupin meal [44]. Wedemeyer and Ross [108] reported that consuming a diet with FM substituted with PP (maize gluten or cottonseed meal) did not affect corynebacterial kidney disease infection in O. mykiss. On the other hand, Ding et al. [79] reported that cumulative mortality in M. nipponense increased significantly when they consumed a diet with decreased FM percentage and increased fermented SBM.From the aforementioned results, it can be said that up to 50% fishmeal in the diets for C. gariepinus could be replaced by plant protein (soybean and sunflower meal) along with supplementation of methionine and lysine without producing any significant differences in the immunological responses and antioxidant status (Figure 7). We used nonspecific immune system parameters in this study because they are considered the first line of protection for fish and are more critical for disease tolerance than a particular immune system that takes longer to activate [109]. The activity of the nonspecific immune response appeared in the resistance to A. sobria, which was considered pivotal in reducing cumulative mortality to 40% in R33.5 and R50. Furthermore, a recent study reported that replacing FM with soybean and sunflower meal with the addition of methionine and lysine in catfish resulted in comparable growth and body composition as well as greater economic performance when compared with the control group [17]. Protein digestibility was found to be better or equivalent to the reference diet with lysine and methionine + cystine supplementation, and up to 50% replacement of fishmeal protein with fermented linseed meal protein for rohu fingerlings [35].5. ConclusionsFishmeal is the costliest component of aquafeed. Reducing the amount of fishmeal in the diet without reducing the performance of fish is beneficial for fish production. The current study concludes that the replacement of fishmeal with a plant protein source (soybean and sunflower meal) by up to 50%, alongside methionine and lysine supplementation, showed a similar immune response with good disease resistance, as well as a decrease in the mortality rate. As such, this type of supplementation is recommended for use in the breeding and production of African catfish.

animals : an open access journal from mdpi

10.3390/ani13081332

PMC10135370

Bacteria in the Anaplasma genus are intracellular parasites of mammal blood cells transmitted by ticks of genera Amblyomma, Dermacentor, Hyalomma, Ixodes, and Rhipicephalus. In this study, the presence of Anaplasma marginale and A. phagocytophilum in ticks was molecularly confirmed in Rhipicephalus sanguineus s.l. and Rhipicephalus bursa ticks, suggesting that these tick species are of importance in the transmission of potential zoonotic infections. Due to the nature of the Anaplasma species detected here, our results, together with data obtained to date in Sardinia, suggest that from a public health point of view, the potential zoonotic Anaplasma species should be further investigated in the island. Additional studies are needed to clarify whether these tick species can transmit these zoonotic bacteria both to human and animal hosts.

Ticks are hematophagous ectoparasites that are recognized for their ability to vector a wide variety of pathogens of viral, bacterial, protozoal, and helminthic nature to vertebrate hosts. Among the different diseases transmitted by ticks, also called “Tick-Borne Diseases” (TBD), many are zoonotic. Pathogens of the genus Anaplasma refer to obligate intracellular bacteria within the Rickettsiales order transmitted mainly through tick bites and considered as well-established threats to domestic animals, livestock, and humans, worldwide. In this retrospective study, 156 ticks collected from twenty goats, one marten, and one cattle from several Sardinian sites, were examined by molecular analyses to detect the presence of Anaplasma species. A total of 10 (10/156; 6.4%) ticks were shown to be Anaplasma-positive by PCR screening. After sequence analyses, A. phagocytophilum was detected in four Rhipicephalus sanguineus s.l. (3.3%) and four Rh. bursa (11%) ticks from goats, while one Rh. sanguineus s.l. (0.8%) and one Rh. bursa (2.8%) collected from the marten and cattle, respectively, exhibited 100% of identity with A. marginale strains. In this study, we provide the first description and molecular detection of A. marginale and A. phagocytophilum in ticks of the Rhiphicephalus genus in Sardinia. Considering the growing impact of tick-borne Anaplasma pathogens on human health, further studies are necessary to monitor the prevalence of these pathogens in Sardinia.

1. IntroductionTicks are hematophagous ectoparasites that transmit protozoan, bacterial, and viral pathogens of medical and veterinary importance [1]. Among bacterial pathogens, members belonging to the genus Anaplasma (Rickettsiales: Anaplasmataceae) are obligate intracellular organisms that replicate within parasitophorous vacuoles in the cytoplasm of both vertebrate and invertebrate host cells [2].In the vertebrate hosts, these organisms infect blood cells including erythrocytes (RBCs), monocytes, platelets, and neutrophils and constitute a major public health threat in humans and animals [3]. The different Anaplasma spp. exist in a zoonotic cycle that involve both Ixodidae ticks and vertebrate hosts, which can be reservoirs of infection [4]. Trans-stadial transmission from nymph to adult exists in nature while transovarial infection has been suggested only for few Ixodidae species. Transmission of these pathogens occurs due to the action of ticks during their blood meal on infected animals and they can transmit the agent to other mammals at the next stage. Although Anaplasma spp. are mainly transmitted by tick bites, other modes of transmission such as hematophagous insect bites and exposure to blood-contaminated fomites have been proven [5].The genus Anaplasma has been especially studied for its pathogenicity in farm animals since various species of Anaplasma such as A. marginale, A. ovis, and A. bovis, along with A. phagocytophilum, are regarded as one of the biggest threats to livestock [6]. In fact, these agents significantly affect animal productivity and cause considerable economic losses to farmers due to the reduction in reproductive performance, decreased milk and meat production, abortion, and death of the infected animal [7]. However, the clinical identification of infected animals is challenging because immunocompetent hosts do not exhibit symptoms while clinical signs ranging from subclinical infections with mild to high fever, anorexia, and respiratory symptoms have been described in immunocompromised hosts [8]. Although anaplasmosis is rarely fatal, leukopenia and impaired neutrophil and lymphocyte function of the bacteria can make animals more susceptible to life-threatening secondary infections [8].Currently, there is a high diversity of Anaplasma organisms, which includes six validated species (A. phagocytophilum, A. marginale, A. centrale, A. ovis, A. bovis, and A. platys).Moreover, the genus contains two new species, namely A. odocoilei and A. capra, that have not been cultured yet, as well as species of ‘Candidatus’ status and many other new unclassified Anaplasma genovariants that have been recently detected [9].Among them, several species that were previously considered non-pathogenic were recently found to be zoonotic and associated with human diseases, suggesting that the number of Anaplasma species that are infecting humans is increasing. Specifically, A. phagocytophilum, the causative agent of tick-borne fever in sheep and granulocytic anaplasmosis in dogs (CGA) and horses (EGA), is also responsible for human infection (HGA) [9]. A. bovis, previously found to infect bovine monocytes, has been recently detected in humans in China [10]. Anaplasma platys, which infects platelets and is the etiological agent of infectious cyclic thrombocytopenia in dogs, has been documented in two women from Venezuela who were exposed to Rhipicephalus sanguineus [11]. A variant of the erythrocytic anaplasmal A. ovis was identified in a Cypriot patient with clinical signs including fever, hepatosplenomegaly, and lymphadenopathy [12]. Finally, Anaplasma capra, a novel, tick-borne pathogen which was detected in China in 2010–2012, causes zoonotic infections and infects many different animal species, including humans [13].Anaplasma distribution is correlated with the presence of tick vectors, hosts, and competent reservoirs. Therefore, determining the density of ticks and the incidence of the infectious agents they transmit is important to prevent and avoid the transmission of possible diseases to animals and humans. Sardinia is the second biggest island in the Mediterranean Sea covering a surface of 24.090 km2 and with different habitat types. It has an annual mean temperature of 22 °C and a typical Mediterranean climate that allows the survival of ticks during the whole year. Furthermore, the island, which is located approximately halfway between Spain, Italy, and North Africa, is an important stopover area for migratory birds which pose a risk for the introduction and dispersal of ticks and TBD. Specific studies of the prevalence of Anaplasma spp. in Sardinian ticks are limited. It was previously observed that ticks belonging to the Rhipicephalus and Hyalomma genera are the most frequent hosts for A. ovis, A. phagocytophilum, A. platys, and A. platys-like, suggesting that these species could serve as potential vectors for these pathogens [14].The aim of this study was to verify the distribution of Anaplasma species in ticks collected from mammals from Sardinia, Italy, and provide epidemiological data to develop strategies and control programs for anaplasmosis prevention and monitoring in the island.2. Materials and Methods2.1. Tick CollectionIn this retrospective study, 156 tick specimens opportunistically removed between June 2011 and October 2013 from 20 goats, 1 cattle, and 1 marten in Sardinia (Italy) were analyzed for the detection of Anaplasma species. The sites belonging to Ogliastra and Sassari provinces were randomly chosen and ticks collected from this geographic area were included in this study. Ticks were removed from each host with tweezers and placed in vials containing 70% ethanol at room temperature. Ticks from one marten that was found dead were provided by hunters who removed specimens from the animal. Morphological identification of the ticks was conducted down to the species level using identification keys [15] with a binocular microscope at a magnification of 50×. Ticks were also sorted by stage and animal host and then stored at −80 °C until further analyses. Details of collection sites, species, and the sex of each tick were collected. The origins and hosts from which each tick was sampled are summarized in Table 1.2.2. DNA Extraction and PCRTo remove environmental contaminants, the ticks were rinsed with 70% ethanol and then immersed in deionized water to remove the ethanol. The ticks were then longitudinally cut in two equal parts using sterile instruments for each individual dissection, and one half was used for DNA extraction. The half tick was homogenized with a Tissue Lyser (TissueLyser II) in 200 μL of PBS. One hundred microliters of genomic DNA was extracted using QIAgen columns (QIAamp tissue kit; Qiagen, Hilden, Germany, cod.69504), according to the manufacturer’s instructions. PCR amplification of the 16S ribosomal RNA gene was carried out on all genomic DNA samples by using oligonucleotide primer pairs (Eurogentec, Seraing, Belgium), which amplified a 293-bp fragment of Anaplasma species [16]. All reactions and amplification conditions used in this study were confirmed from studies previously published [14]. Negative and positive controls were included in each amplification assay. Eight microliters of each 293-bp PCR product was electrophoresed in 1.5% agarose gel with SYBR™ Safe DNA Gel Stain (Invitrogen, Carlsbad, CA, USA) in one × TAE buffer against a DNA ladder. The gel was then visualized and photographed using Alliance LD2 gel documentation system (UVITEC, Cambridge, UK).An additional PCR was performed on positive DNA tick samples using 16SANA-F (5′-CAG AGTTTG ATC CTG GCT CAG AAC G-3′) and 16SANA-R (5′-GAGTTT GCC GGG ACT TCT TCT GTA-3′) primers that amplify 16S rRNA gene of Anaplasma spp., as reported in De la Fuente et al., 2005 [17]. The reaction was made up to 25 μL containing 12.5 μL of 2× PCR Master Mix (Quantitect Probe PCR Master Mix; Qiagen, Hilden, Germany), 1 μL of 25 μM of each primer, and 1 uL of template DNA. DNA extracted from uninfected ticks and DNA previously extracted from A. phagocytophilum IFA slides were included in each PCR test as negative and positive controls, respectively. Thermocycler conditions were performed in automated DNA thermal cyclers (GeneAmp PCR Systems 2400 and 9700; Applied Biosystems, Courtaboeuf, France) with the cycling conditions as follows: 95 °C for 15 min, 40 cycles of 94 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min, with the final elongation step at 72 °C for 5 min. The amplicons were then subjected to electrophoresis in a 1.5% of agarose gel at 110 V for 30 min and visualized using a Syber safe nucleic acid staining solution, under UV light.2.3. Purification, Sequencing, and Phylogenetic AnalysesThe Anaplasma positive samples were selected and purified using the QIAquick Spin PCR purification kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. The purified PCR products were then sequenced using the 16S primer pairs in both directions on an automated DNA sequencer (ABI-PRISM 3500 Genetic Analyzer; Applied Biosystems, Seevetal, Germany). The DNA sequencing kit (dRhodamine Terminator Cycle Sequencing Ready Reaction; Applied Biosystems) was used according to the manufacturer’s instructions. Chromatograms of nucleotide sequences generated in this study were assembled and edited with ChromasPro software (version 1.34; Technelysium Pty Ltd., Tewantin, Queensland, Australia). Sequences were then aligned with CLUSTALX [18] due to assign them to unique sequence types and checked against the GenBank database by using BLASTn analysis tool [19]. Pairwise/multiple sequence alignments and sequence similarities were calculated using the CLUSTALW [20] and the identity matrix options of Bioedit [21], respectively. The phylogenetic tree was constructed using the neighbor-joining method in MEGA software version 6.0. The distance matrix was calculated by use of Kimura-2 parameters. The statistical analysis was performed using the Bootstrap method with 1000 replicates. The method used to calculate a confidence interval for a proportion is the Wilson score method without continuity correction [22].3. Results3.1. Tick CollectionTicks were morphologically identified at the species level as Rh. sanguineus s.l. (120 specimens) and Rh. bursa (36 specimens). Although Rh. bursa is not included in the Rh. sanguineus group, the species has been differentiated by the shapes of adanal plates for males and by the genital aperture, porose areas in the dorsal surface of basis capituli, spiracle plates, and the presence of dense setae in the spiracle areas for females. Rhipicephalus bursa shows adanal plates guttiform with maximum width at the posterior margin and with quite convex and divergent lateral margins and obtuse and broadly rounded posterior inner angles. Rhipicephalus sanguineus presents adanal plates with rectilinear or weakly inclined posterior margins and posterior inner angles almost right. The genital aperture posterior lips of Rh. bursa have a narrow “V” shape with divaricate and slightly rounded lateral margins, while Rh. sanguineus shows a genital aperture like a broad “U” with divergent lateral margins. Porose areas are nearly circular with a broad distance separating them in Rh. sanguineus and are oval with a narrow separation in Rh. bursa. Rhipicephalus sanguineus shows spiracles plates with narrow tails and the presence of sparse setae in spiracles areas while Rh. bursa have spiracles plates with broad tails and dense setae in spiracles areas. All ticks were adult specimens and no larvae and nymphs were removed from collected animals. Tick species and number, host source, collection sites, stage, and sex of ticks collected from this study are shown in Table 1.3.2. Molecular Detection of Anaplasma spp.A total of 10/156 tick samples (6.4%; 95% CI 3.5–11.4) tested positive for Anaplasma DNA using 16S rRNA PCR (Figure 1).To molecularly determine the identity of Anaplasma spp. detected in ticks from this study, the PCR products of the 10 PCR-positive samples were directly sequenced.A total of four sequences from Rh. sanguineus s.l. (3.3%; 95% CI 1.3–8.3) and four from Rh. bursa ticks (11.1%; 95% CI 4.4–25.3%), all collected from goats, were readable and chromatograms generated a clear sequencing signal containing an Anaplasma that showed 100% similarity with the 16S ribosomal RNA fragment of A. phagocytophilum strains after BLAST search analyses.One Rh. sanguineus s.l. from marten (0.83%; 95% CI 0.15–4.57) and one Rh. bursa tick from cattle (2.78%; 95% CI 0.49–14.17) contained Anaplasma DNA that exhibited 100% similarity with A. marginale strains deposited in GenBank. All 16S rRNA PCR product sequences resulted in the establishment of two 16S different genotypes, as shown in Table 2.Specifically, eight sequences from ticks sampled from goats were identical to each other and to A. phagocitophilum strains from GenBank, while two sequences from two ticks collected from one marten and one cattle were identical to A. marginale strains isolated worldwide (Table 2). The two different sequence types generated in this study named AP-SAR2011 (Anaplasma phagocitophilum sequence type) and AM-SAR2011 (Anaplasma marginale sequence type) were deposited into GenBank under the accession numbers KP877313 and KP877314, respectively (Table 2). The phylogenetic analysis based on the partial 16S rRNA (Figure 2) showed that the A. phagocitophilum strain AP-SAR2011 found in this study was in the same clade as strains isolated in South Korea, Russia, and China from humans, ticks, rats, and goats. The A. marginale strain AM-SAR2011 was in the same clade as the Italian strain BS16, isolated from bovines, and close to the strains isolated in Philippines and South Africa.4. DiscussionThe rapid identification of different tick species and the bacteria they carry contributes substantially to the clinical diagnosis, treatment, and surveillance of tick-borne diseases. Current knowledge of the incidence of Anaplasma species in Sardinian ticks is limited, and obtained results, based on specific molecular typing, highlighted that Rhipicephalus and Hyalomma ticks harbored several species of Anaplasma (A. ovis, A. platys-like, A. platys and A. phagocytophilum), raising concerns regarding their potential to transmit these pathogens to humans, domestic hosts, and wildlife. The presented data show that within the five monitored sites, two different tick species were identified. Specifically, Rhipicephalus sanguineus s.l. ticks were the predominant identified species documented here, and this result was consistent with data from the previous literature [23]. Rhipicephalus species are widely distributed across the Mediterranean region, as well as in Sardinia where the Dermancentor, Haemaphysalis, and Hyalomma genera are also well represented and adapted to the ecosystem [24]. In this study, the presence of Anaplasma species was recorded in 6.4% of tested ticks. It does not mean that these ticks are competent vectors for the bacteria, as the ticks may have been infected by feeding on bacteremic animals or by cofeeding with tick vectors.In particular, the obtained results indicated that, like previous studies in which A. marginale was detected in Rh. bursa and Rh. sanguineus s.l. ticks from Italy, Portugal, and Spain [25,26,27], Sardinian tick species are hosts for A. marginale, suggesting that both tick species could serve as potential biological vectors for A. marginale infection. To the best of our knowledge, this is the first molecular evidence of A. marginale in Rh. bursa and Rh. sanguineus s.l. ticks collected from goats and a marten in Sardinia, a region in which species of the Rhipicephalus genus are widely distributed.Anaplasma marginale, the aetiologic agent of bovine anaplasmosis, represents one of the most important tick-borne diseases in ruminants worldwide, mainly in the tropical and subtropical regions. In the southern regions of Italy (Sicily, Puglia, Campania, and Basilicata), where bovine anaplasmosis is endemic, the presence of A. marginale in Rh. turanicus and Haemaphysalis punctata collected from cattle has been previously reported [27]. This agent infects circulating erythrocytes of domestic and wild ruminants [28]. Infected cattle serve as a reservoir of A. marginale providing a tick blood source for the efficient biological transmission of the pathogen [28]. Although approximately 20 tick species are reported as biological vectors of A. marginale worldwide [9], most of them are able to transmit A. marginale only under experimental conditions, which does not necessarily imply transmission in the field [29]. However, it has been demonstrated that the pathogen can be mechanically transmitted by blood-contaminated mouthparts of biting diptera of the Tabanus and Stomoxys genera, or via fomites [30]. In this study, ticks that tested positive for A. marginale were removed from asymptomatic goats. However, we cannot know the health status of the marten whose ticks tested positive for A. marginale since it was found dead for unknown causes. Results from a previous study conducted in northeastern Hungary highlighted the presence of Anaplasma sp. in spleen and liver samples of the European pine marten in which the zoonotic ecotype I of A. phagocytophilum has been identified [31]. More investigation on A. marginale in Rhipicephalus ticks and in domestic and wild vertebrate hosts could help to highlight the possible role of these ticks as vectors of Anaplasma species. Although phylogenetic analysis based on Anaplasma sequences obtained with 16S rRNA gene amplification revealed that this target gene can be widely used for the identification of Anaplasma species and can be considered a valuable phylogenetic tool, more discriminative genes will be used for the confirmation of these results.This study also reports the first molecular detection of A. phagocytophilum in Rh. sanguineus and Rh. bursa ticks in Sardinia, indicating a potential role for these tick species in the epidemiology of the disease. Anaplasma phagocytophilum is the cause of granulocytic anaplasmosis in humans (HGA) [32], which severity ranges from asymptomatic infection to mild or severe febrile illness and involvement of multiple organ failure or even death [33]. Therefore, several genetic variants of this pathogen have been determined and all of these differ from each other for the different host specificity, vectors, pathogenicity, and geographical distribution [34]. In fact, all variants can infect different species as demonstrated by experimental studies in which it has been proven that A. phagocytophilum strains isolated from different matrices are not capable of infecting different hosts [4]. The analysis of this genetic variability has been made through molecular methods by using different loci such as 16s rRNA, groESL msp2, msp4, and ankA genes. Thanks to the use of one or more of these molecular markers, A. phagocytphilum has been divided into different genetic variants that can be involved in different epidemiological cycles, distribution, and host spectra [4]. Although 16s gene rRNA is the most used target gene, the phylogenetic study based only on the use of 16S rRNA could be deficient. It is related to the genetic recombination of this gene that undergoes several variations [4]. Moreover, if used alone, it may not be discriminative enough to correctly differentiate the different types of Anaplasma species, although different variants have been identified by using it as a reference point. The analysis of nucleotide sequences using the groESL gene has allowed for the identification of four different ecotypes of A. phagocytophilum in Europe, and different genetic variants adapted to the different hosts and vectors present in a specific geographical area [35]. All these ecotypes can infect both vertebrate and invertebrate hosts. In particular, the ecotype I has a wide host spectrum, being associated to multiple animal species including humans. This ecotype with zoonotic potential has the largest range in wildlife reservoirs but can also infect domestic animals. Specifically, hosts of this pathogen include cattle, sheep, goats, horses, dogs, hares, yaks, and rodents [36] and evidence of the pathogen in several mammalian and invertebrate hosts have been reported in Italy as well [37,38,39]. However, one limitation of this study was the lack of identification of A. phagocytophilum ecotypes, which is essential information for defining the zoonotic relevance. Further studies are needed to better characterize strains by analyzing more discriminative genes and to identify the main vectors implicated in the transmission of Anaplasma species in Sardinia.Moreover, the A. phagocytophilum strain detected here was close to A. phagocytophilum strains isolated in China, Korea, and Russia. It could be related to anthropogenic activities that contribute directly or indirectly to the emergence and re-emergence of tick-borne pathogens (e.g., animal production, animal–human interfacing, and globalization). Moreover, the role of migratory birds in the spread of ticks and their role in the circulation and dissemination of pathogens in Europe cannot be ruled out. During seasonal migrations, birds that cover short, medium, or long distances within one or more distant geographical regions can carry ticks and related pathogens, introducing ticks and pathogenic species to new areas [40].Although I. ricinus is the main vector of A. phagocytophylum in Europe [41], the zoonotic pathogen has been also associated with Rhipicephalus and Dermacentor spp. ticks from other parts of the world [42]. However, since the Mediterranean climate could be a limiting factor for Ixodes distribution and it would explain the low population of I. ricinus in Sardinia, the abundance of Rhipicephalus ticks in the Mediterranean basin prompted us to suggest that the Rhipicephalus species can serve as vectors of A. phagocytophilum and may transmit the pathogen to animal hosts. Therefore, the number of ticks was very low and they were collected from the hosts. Therefore, no conclusion can be drawn whatsoever about the circulation of the pathogens within the tick population, as every detection could be the result of infected ingested blood. These results confirmed the presence of A. phagocytophylum in Rhipicephalus ticks feeding in goats. In Sardinia, where ruminant breeding represents a zootechnical reality of primary importance, the increase in the incidence of anaplasmosis transmitted by vectors could represent a serious threat to company profitability. According to the National Italian Database 2020 (BDN) (established by the Ministry of Health at the National Surveillance Centre of the IZS in the Abruzzo and Molise Region), Sardinia has an estimated population of more than 3 million sheep and 0.3 million goats, and losses due to abortion of ruminants are estimated to be around EUR 10 million per year. Veterinarians should not overlook the presence of A. phogocytophilum in Sardinian goats and evaluation of the potential role of Anaplasma species as abortifacient agents should be also taken into account. Improving the entomological surveillance program is necessary to establish and maintain a dialogue with farmers, including listening to and addressing their concerns and sharing an adequate diagnostic and therapeutic path for the good health of farm management through innovative solutions that will reduce the economic losses in this area and ensure the efficiency of vector control interventions. Furthermore, since the risk of transmission of vector-related diseases is extended to the entire population, particular attention must be paid to professional categories who carry out their activities outdoors.5. ConclusionsThe knowledge and characterization of the diversity of Anaplasma strains circulating in the island are fundamental to design epidemiological studies and control strategies for both HGA and bovine anaplasmosis. The results of this study showed that Rh. sanguineus s.l. and Rh. bursa harbor two Anaplasma spp., of which A. marginale was not yet reported in the territory, and support the hypothesis that these tick species could act as vectors for A. marginale and A. phagocytophilum in Sardinia. Further investigation to fully understand a possible role of Rhipicephalus ticks in the A. phagocytophilum and A. marginale strains circulation are warranted. Furthermore, future studies may yield more insight into the seroprevalence of Anaplasma species in ruminants and in the dog population of Sardinia considering their potential role in the transmission of the disease to humans.

animals : an open access journal from mdpi

10.3390/ani12010048

PMC8749748

The lipid deposition and health status of egg-laying hens is crucial to the development of the poultry industry. This study aimed to evaluate the effects of genetic variations in the flavin-containing monooxygenase 3 (FMO3) on the lipid metabolic diseases of laying hens during the late laying period. The results showed that the T329S mutation in FMO3 moderated the lipid parameters and decreased the atherosclerotic lesions and hepatic steatosis in laying hens with homozygous T329S mutation. In conclusion, the T329S mutation in FMO3 is closely associated with the improvement of lipid metabolic diseases in laying hens during the late laying period. The results of this study may contribute to overcoming the challenge of lipid metabolic diseases in laying hens during the late laying period.

The T329S mutation in flavin-containing monooxygenase 3 (FMO3) impairs the trimethylamine (TMA) metabolism in laying hens. The TMA metabolic pathway is closely linked to lipid metabolic diseases, such as atherosclerosis and fatty liver disease. We aimed to evaluate the effects of the T329S mutation in FMO3 on lipid metabolism in chickens during the late laying period. We selected 18 FMO3 genotyped individuals (consisting of six AA, six AT, and six TT hens) with similar body weight and production performance. The lipid metabolism and deposition characteristics of the laying hens with different genotypes were compared. The T329S mutation moderated the serum-lipid parameters in TT hens compared to those in AA and AT hens from 49 to 62 weeks. Furthermore, it reduced the serum trimethylamine N-oxide concentrations and increased the serum total bile acid (p < 0.05) and related lipid transporter levels in TT hens. Moreover, it significantly (p < 0.01) decreased atherosclerotic lesions and hepatic steatosis in TT hens compared to those in the AA and AT hens. Our findings may help improve the health status in laying hens during the late laying period.

1. IntroductionFlavin-containing monooxygenases (FMOs, EC 1.14.13.8) are an important class of oxidases that are responsible for the oxygenation of soft nucleophilic heteroatom-containing organic substances, such as nitrogen, sulfur, and phosphorous. Accordingly, FMOs play a significant role in the metabolism and detoxification of drugs, endogenous substances, and dietary-derived compounds [1]. FMO3, which plays an important role in the metabolism of endogenous substances, is the most important member of the FMO family and is the predominant isoform that is involved in the trimethylamine (TMA) metabolic pathway [2]. This pathway involves the oxidation of TMA to form trimethylamine N-oxide (TMAO) by hepatic FMO3, which is closely linked to many metabolic characteristics [3]. It has been reported that some mutations in the FMO3 gene in humans can either abolish or diminish the catalytic activity of the enzyme and inhibit the oxidization of TMA, a fishy odor substance, which results in the fish-odor syndrome [4]. Similarly, a threonine to serine substitution at the 329th position (T329S, FMO3 c.984 A > T) of FMO3 was detected in chickens, whose function is similar to those of humans and associated with the fishy-odor eggs traits [5,6]. Thus, we speculated that the T329S could decrease the circulating TMAO levels in homozygous T329S (TT) hens compared to those of homozygous wild-type (AA) hens. This mutation has a low mutation frequency in several Chinese local chicken varieties, such as Huainan Mahuang hens (3.8%), Hebei Chai hens (6.8%), and Wenchang hens (9.6%) [7,8]. In parallel, it was also detected in CAU-3, a strain of brown-egg dwarf chickens, that reaches to 5% production at 21 weeks of age and lays approximately 180 eggs from hatch to 52 weeks.Recently, a novel function of the TMA metabolic pathway on lipid metabolic diseases in mammals was identified [9,10,11,12]. TMAO was first suggested as an independent marker for atherosclerosis (AS), and dietary TMAO supplementation confirmed its role in promoting the development of AS and thrombosis in mice [9,10]. Furthermore, the circulating TMAO levels have been associated with adverse effects on fatty liver disease (FLD), and increased liver inflammation and damage have also been reported in human studies [11,12]. A possible mechanism underlying the involvement of TMAO in AS and FLD pathogenesis has been proposed, where TMAO affects lipid absorption and cholesterol homeostasis by decreasing the total bile acid (TBA) pool size [13,14]. Hepatic FMO3 plays a direct role in the cholesterol and triglycerides (TGs) metabolism in hyperlipidemic mice [15]. Further, Hepatic FMO3 has been demonstrated to be a central regulator of hepatic cholesterol balance by regulating reverse cholesterol transport (RCT), in mouse models [16]. In mice, the natural variability of FMO3 not only reduces TMAO production but also stimulates the RCT process to promote cholesterol efflux [15]. There is increasing evidence that supports the role TMA metabolic pathway in hyperlipidemia, AS, and FLD, and that the FMO3 genetic variants may have a positive effect on improving lipid metabolism in mammals [15,16,17].It has been suggested that the regulation of cholesterol disposal in poultry is highly similar to that in mammals [18]; however, the effect of the TMA metabolic pathway on lipid metabolic disease in poultry remains unclear. Lipid metabolic diseases, such as AS and FLD, negatively impact the poultry industry, as they reduce the production performance of laying hens in the late laying period [19,20]. They can even induce sudden death, resulting in major economic losses [21]. Previous studies have demonstrated that FLD can cause up to 5% mortality in commercial layers during the laying cycle and 74% of the total mortality in caged laying hens in Queensland, Australia. Meanwhile, the mortality of laying hens with AS has not been reported [22,23]. The occurrence and development of these diseases are associated with adiposity in the laying process [19,20]. To meet the needs of egg yolk formation during the egg-laying period, hepatic lipid synthesis and metabolism are strongly activated, resulting in elevated levels of TG and phospholipids. Under the action of multiple lipid transfer proteins, these lipids are assembled into very low density lipoprotein (VLDL) particles that are transported to the follicle [18,24]. For this reason, and coupled with the extension of the laying cycle from the first 72–80 weeks, some breeding companies have even prolonged the laying age to 100 weeks [25,26]. Continuous egg production with a high dietary consumption of carbohydrate can markedly induce adiposity, increasing the risk of FLD, as well AS [25,27]. Several studies have demonstrated that an increasing level of bile acid production or an exogenous bile acid intake can play a dramatic role in promoting lipid absorption and cholesterol excretion for improving this situation [28,29,30,31]. However, until now, the impacts of SNPs in FMO3 have been largely unstudied regarding lipid metabolism in chickens, other than fishy taint in eggs.Given the alleviating effects of FMO3 mutations in the TMA metabolic pathway on lipid metabolic diseases, we hypothesized that the TMA-metabolic-pathway-involved FMO3 genetic variants may have a positive effect on improving the health status in chickens. In chickens, only the T329S mutation in FMO3 is highly associated with the TMA metabolic pathway [5]. In this study, we compared the lipid metabolism and deposition characteristics among laying hens with different genotypes in the late laying period. We aimed to investigate whether the T329S mutation in FMO3 can improve the health status by modulating the lipid metabolism in laying hens.2. Materials and Methods2.1. Birds and DietA pure line of CAU-3, maintained mainly for egg production for over 10 years in the Poultry Genetic Resource and Breeding Experimental Unit of China Agricultural University [32,33], was used in this study. This population was produced from pedigree mating, using 50 sires and 7 to 8 dams per sire. After the strain of CAU-3 was established, each generation was “bred” by a random process.At 47 weeks of age, a total of 208 CAU-3 healthy hens were selected for blood sampling via wing vein puncture. Blood samples were collected for genotype analysis (A/T polymorphism at position 1034 of the chicken FMO3 exon 7, chromosome 8, accession number: AJ431390). Each hen was fitted with a leg ring marked with a unique identification number. A polymerase chain reaction-restriction fragment length polymorphism assay, as described in Reference [7], was used to determine the individual FMO3 genotypes (AA, AT, and TT) at this position. During the two-week acclimation period, 18 FMO3 genotyped individuals (consisting of six AA, six AT, and six TT hens) with similar body weight and egg production were selected and raised in individual cages (cage size: 45 cm × 45 cm × 45 cm). The formal experimental period ranged from 49 to 62 weeks of age. During the experimental period, hens with these genotypes were fed a basal diet, which was formulated to meet the National Research Council requirements (NRC, 1994) [34], and were offered in mash form ad libitum, with each hen ingesting approximately 100 g/day. The composition and nutrient levels of the basal diet are shown in Table S1. Water was supplied by nipple drinkers. Egg production was recorded daily for each hen. The room temperature was maintained between 22 and 26 °C, and light exposure was controlled with a light/dark cycle of 16:8 h. Illumination was provided by incandescent lamps with an intensity of 10 lx (at bird-head level). All procedures, as well as the care, housing, and handling of the animals, were conducted according to accepted commercial management practices. All birds remained healthy during the feeding period. No birds were culled, and none received any medical intervention. The animal experiment was conducted in the Experimental Unit for Poultry Genetic Resource and Breeding of China Agricultural University.2.2. Sample CollectionAt 49 weeks of age, after an 8-h fast, blood samples were collected from all 18 genotyped individuals in the morning. At 62 weeks of age, after an 8 h fast, blood samples were collected from the same individuals in the morning. Blood samples were stored in a vacuum blood-collection tube without anticoagulant. The serum was separated by centrifugation at 3000× g for 15 min and stored at −20 °C until analysis. Then, the six AA, six AT, and six TT hens were humanely euthanized. The whole aorta and liver tissues were isolated and fixed in formalin for 48–72 h, before being processed for AS lesion analysis and liver histopathological observation.In addition, three healthy AA pullets (18 weeks old) were sacrificed, and the whole aorta and liver tissues were isolated in the same manner. These samples were used as a negative control for AS lesion analysis and liver histopathological observation, since our pre-experiment found that lipid droplets were scarcely observed in the aortic wall and liver of pullets at 18 weeks.2.3. Serum Index Measurement2.3.1. Serum-Lipid Parameters and Bile AcidSerum TG, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and TBA levels were determined by using commercial kits (Shanghai Jingkang Bioengineering Co., Ltd., Shanghai, China). The automatic biochemical analyzer used was a KHB ZY-1280 manufactured by the Shanghai Kehua Bio-engineering Corporation (Shanghai, China).2.3.2. Nontraditional Serum-Lipid ParametersNontraditional serum-lipid parameters were calculated as previously described in Reference [35]: non-HDL-C = (total cholesterol minus high-density lipoprotein cholesterol); TC/HDL-C = (total cholesterol/high-density lipoprotein cholesterol); atherogenic index (AIS) = [(total cholesterol minus high-density lipoprotein cholesterol)/high-density lipoprotein cholesterol]; lipoprotein combined index (LCI) = (total cholesterol × total TG × low-density lipoprotein cholesterol/high-density lipoprotein cholesterol).2.3.3. Serum TMA and TMAOPretreatment serum was prepared as previously described in Reference [36], with minor modifications. In brief, 500 μL of the serum was immediately acidified upon collection with 500 μL of hydrochloric acid (0.01 M), vortex mixed for 30 s, and stored at −20 °C until analysis of TMA and TMAO. The serum TMA and TMAO levels were measured by using headspace gas chromatography–mass spectrometry (GC–MS), as previously described in Reference [6]. A Shimadzu GC-2010 Plus gas chromatograph (Shimadzu, Kyoto, Japan) fitted with a 60 m × 0.20 mm ID fused silica capillary column coated with a 1.12 μm film of HP-VOC (Agilent Technologies, CA, USA) and a Shimadzu TQ8040 mass spectrometer were used. The detection was performed on a Shimadzu GCMS-QP2010 Plus system (Shimadzu Technologies, Kyoto, Japan) equipped with an HS-20 headspace sampler.2.3.4. Lipid-Related Transporters and VLDLSerum protein concentrations of phospholipid transfer protein (PLTP), lecithin–cholesterol acyltransferase (LCAT), cholesterol ester transfer protein (CETP), and serum VLDL concentrations were measured by using chicken PLTP, LCAT, CETP, and VLDL enzyme-linked immunosorbent assay kits from Shanghai Jingkang Bioengineering Co., Ltd. (PLTP ELISA kit JLC23302, LCAT ELISA kit JLC10740, CETP ELISA kit JLC23020, and VLDL ELISA kit JLC10779), following the manufacturer’s instructions. The antigen, which the VLDL ELISA kit measured, is apolipoprotein B (ApoB), a fraction of VLDL particles. VLDL levels were calculated from a standard curve, which was plotted according to the OD of the concentration of standards (Standard concentration was followed by 16, 8, 4, 2, 1, and 0.5 mmol/L).2.4. Lipid-Deposition Characteristics2.4.1. AS Lesions AnalysisAS lesions were quantified by using en face analysis of the aorta (including the aortic arch, thoracic, and abdominal regions) and cross-sectional analysis of the aortic arch, as previously described in References [37,38], with minor modifications. For en face analysis, the aorta was longitudinally opened and stained with Oil red O (Wuhan Service Biotechnology Co., Ltd., Wuhan, China) to detect lipids and to determine the lesion area. AS lesions of the aorta were expressed as percentages of the total surface area. For cross-sectional analysis, a small segment of the aortic arch (in the same area) was embedded in OCT compound (Sigma-Aldrich, St. Louis, MO, USA) and frozen at −20 °C. Sections (8 µm thick) were collected. Lesions from ten alternating sections were stained with Oil Red O and hematoxylin. For each section of the aortic arch, ten randomly selected areas were assessed, using light microscopy at ×40 magnification.2.4.2. Liver Histopathological ObservationLiver sections were examined for steatosis, using Oil Red O staining, as previously described in Reference [39], with minor modifications. For cryo-section cutting, fixed samples were embedded in frozen OCT (Sigma-Aldrich, St. Louis, MO, USA) and then sectioned at 10 µm; all operations were carried out under frozen conditions. Samples were then stained with Oil Red O (Wuhan Service Biotechnology Co., Ltd., Wuhan, China), differentiated with isopropanol, washed with distilled water, and stained with hematoxylin. For each section of the liver, 10 randomly selected areas were assessed by using light microscopy at ×80 magnification.Images of aortic en face, aortic arch cross-sections, and liver sections were taken by using a Canon EOS 7D digital camera (Canon, Tokyo, Japan), and the aorta, aortic arch, and liver lesions were quantified by using computer-assisted image analysis (ImageJ, NIH Image, National Institutes of Health, version 1.8.0), according to procedures described in Reference [40].2.5. Statistical AnalysisThe genotype distributions in the flock were tested for Hardy–Weinberg equilibrium. Statistical analyses were conducted by using the R software (version 4.0.3), and the figures were plotted by using GraphPad Prism (version 7.04; GraphPad Software, San Diego, CA, USA). Due to the small sample size, nonparametric procedures were used. Basic descriptive statistics: An independent-sample Kruskal–Wallis test was used to analyze the average number of laid eggs, the serum indices, and the lipid-deposition characteristics, including aortic lesion (AL), aortic arch lesion (AAL), and hepatic lipid deposition (HLD), in chickens with different genotypes. Rate of change (RC) of serum indices of chickens in the same genotype between 49 and 62 weeks of age was obtained by applying the following formula: RC (%) = (X62 − X49)/X49 × 100%, where X62 includes the levels of the serum TG, TC, LDL-C, HDL-C, and TBA of chickens at 62 weeks; and X49 includes the levels of the serum TG, TC, LDL-C, HDL-C, and TBA of chickens at 49 weeks, corresponding to those of 62 weeks.In addition, correlations between nontraditional serum-lipid parameters (AIS, LCI, etc.) and lipid-deposition characteristics, as well as correlations between TMAO, TMA, TBA, and lipid-deposition characteristics, were analyzed by using Pearson’s correlation coefficient. The p-Values < 0.05 and < 0.01 were considered statistically significant and extremely statistically significant, respectively. Data are expressed as the mean ± standard deviation.3. Results3.1. Hardy–Weinberg Equilibrium Test and Egg ProductionA total of 151 AA, 51 AT, and six TT hens were obtained from the whole flock. The genotype frequencies of FMO3 were 72.59% (AA), 24.52% (AT), and 2.88% (TT), which complied with the Hardy–Weinberg equilibrium (p = 0.59). The egg production of the 18 sampling chickens from 49 to 62 weeks showed that TT hens produced 52 eggs each, which was higher than (p > 0.05) the number of eggs produced by AA (49.8 eggs per hen) and AT (47.2 eggs per hen) hens.3.2. Serum-Lipid Parameter and TBA LevelsThe changes in serum-lipid parameters and TBA levels in laying hens at the ages of 49 and 62 weeks are shown in Table 1. The serum-lipid parameter levels, including TG, TC, LDL-C, and HDL-C, and the serum TBA levels were all increased at 62 weeks, compared to those at 49 weeks, regardless of the genotype of the chickens. However, the T329S mutation in FMO3 prominently (p < 0.05) decreased the RC of TG, TC, and LDL-C levels in TT hens. In addition, at 62 weeks, as an effect of the T329S mutation, the serum TG and LDL-C levels were decreased, especially the serum TG levels (p < 0.05) in AT and TT hens compared to those of AA hens. Additionally, the T329S mutation of FMO3 decreased the serum TC levels (p < 0.05) but increased the serum TBA levels (p < 0.05) in TT hens compared to those of AA and AT hens.3.3. Metabolic CharacteristicsThe metabolic characteristics of chickens for the AA, AT, and TT genotypes at the age of 62 weeks are shown in Table 2. As a result of the T329S mutation, the nontraditional lipid parameters, including non-HDL-C, TC/HDL-C, AIS, and the LCI ratios, in TT hens were decreased compared to those of AA and AT hens. In particular, the AIS decreased by 30% and 40% (p < 0.05), and the LCI decreased by 70% and 35% (p < 0.05), respectively, compared to those of AA and AT hens. The serum TMAO concentration was also reduced by approximately 12%, with increasing serum TMA concentrations in TT hens compared to those of AA and AT hens, although the difference was not statistically significant (p > 0.05). Moreover, the T329S mutation in TT hens increased the serum PLTP, LCAT, CETP, and VLDL concentrations by approximately two-fold (p < 0.01) compared to those of AA and AT hens.3.4. Lipid-Deposition CharacteristicsThe pathological observations of lipid deposition in the aortic wall and liver of chickens for the AA, AT, and TT genotypes are shown in Figure 1. Compared to the controls (AA pullets, 18 weeks), all hens had varying degrees of lipid deposition in the aorta, aortic arch, and hepatocytes at 62 weeks (Figure 1AB). However, it was noticeable that the aorta and aortic arch cross-section lesion proportion in TT hens was less (p < 0.01) than half of that in AA and AT hens as an effect of the T329S mutation in FMO3 (Figure 1A,B,D,E). Oil Red O staining of the liver showed that TT hens had the least (p < 0.01) hepatic lipid droplet accumulation compared to that of AA and AT hens (Figure 1C,F). In addition, the trends of AL, AAL, and HLD of the 18 chickens were all positively correlated with the nontraditional lipid parameters (AIS, LCI, etc.), as shown in Figure 2.3.5. Correlations among the Serum TMAO, TMA, and Lipid-Deposition CharacteristicsThe correlation matrix of the relationships among serum TMAO, TMA, TBA, and lipid-deposition characteristics is shown in Figure 3. Of the 18 samples, the serum TMAO levels were positively (p > 0.05) correlated with AL (r = 0.26), AAL (r = 0.32), and HLD (r = 0.45), but negatively (p > 0.05) correlated with the serum TBA levels (r = −0.34). Serum TBA levels were negatively (p < 0.01) correlated with AL (r = −0.64), AAL (r = −0.76), and HLD (r = −0.64), whereas they were positively (p < 0.01) correlated with serum VLDL (r = 0.95) levels.4. DiscussionThe present study showed that the T329S mutation in FMO3 was associated with the improvement of lipid metabolic diseases in TT hens during the late laying period. This result could be attributed to the T329S mutation moderating the serum-lipid parameters, reducing TMAO production, and alleviating lipid deposition in TT hens. The present study proposed a new role of the T329S mutation in improving the health status in egg-laying hens. Previous studies have focused on the fact that feeding TT hens a high-level TMA precursor diet increases the risk of fish-odor eggs [5,6]. The effect of the T329S mutation on lipid metabolism and health status has not yet been considered. To the best of our knowledge, our study is the first to investigate the effect of the T329S mutation in FMO3 on health status in laying hens. Our results may contribute to overcoming the challenges of lipid metabolic diseases in laying hens during the late laying period.In chickens, the lipid metabolic diseases, such as AS and FLD, primarily occur in the late laying period and are mainly caused by the imbalance between deposition and removal of lipids [21,41]. Serum-lipid parameters, such as TG, TC, and LDL-C, are the direct indicators of lipid metabolism balance during the laying period [42]. In this study, the increases in serum TG and TC levels in TT hens were less than those in AA and AT hens, and the increases in serum LDL-C in TT hens were less than those in AA hens from 49–62 weeks. TG, TC, and LDL-C are all major risk factors for AS and FLD [43,44]. Persistently high levels of serum TG, TC, and LDL-C can increase the risk of AS and FLD, resulting in the decreased production performance of laying hens and even death [45]. In contrast, modulating the TG and cholesterol levels can reduce the incidence of these diseases in aged laying hens [21]. Thus, more stable lipid parameters could be one of the origins of health status improvement in TT hens. In parallel, the stability of lipid parameters could be attributed to the action of the T329S mutation. In our study, the decreases of the TG and LDL-C levels by T329S in both TT and AT (FMO3, c. 984 A > T) hens at 62 weeks confirmed this assumption. This is similar to the findings of Reference [41], who showed that T329S upregulated the mRNA levels of genes involved in cholesterol and TG transport (e.g., apovitellenin 1, and ATP binding cassette transporters G5 and G8) in TT hens. Therefore, we assumed that the T329S mutation in FMO3 could moderate the changes in lipid parameter levels.Compared with the conventional lipid parameters, the non-HDL-C, TC/HDL-C, AIS, and LCI ratios are better indicators of AS, a disease that is positively associated with FLD [46]. These ratios have been suggested to be accurate predictors of AS [47]. For example, a lower AIS ratio in healthy controls (3.19) than that in the disease onset group (3.51) indicated a low risk [35]. In our study, the decreases of non-HDL-C, TC/HDL-C, AIS, and LCI ratios demonstrated that the T329S mutation in FMO3 had an anti-AS effect on TT hens to some extent, as these ratios were positively correlated with AL, AAL, and HLD in chickens (Figure 2). The least aortic and aortic arch lesions in TT hens (Figure 1A,B; p < 0.01) provided evidence for this conclusion. At the same time, our results also showed that TT hens had only mild degrees of hepatic steatosis (Figure 1C; p < 0.01), thus indicating a lower risk of FLD in TT hens. Therefore, these results demonstrated that the T329S mutation could reduce the risk of AS lesions and FLD in laying hens.The decreased AS lesion and hepatic steatosis could be attributed to the decrease of circulating TMAO concentrations in TT hens. Mutation T329S diminishes the ability of FMO3 to oxidize TMA to TMAO [5,48]. Thus, circulating TMAO, the oxide of TMA, was reduced in TT hens. Previous studies have suggested that circulating TMAO levels are adversely associated with AS and fatty liver events, the increase of which could promote the formation of AS and FLD [10,11,12]. In our study, there was no significant (p > 0.05) correlation between TMAO concentrations and AL, AAL, or FLD in the chicken model, meaning that there was no direct association between TMAO concentrations and these diseases. However, the observations related to the bile acids could provide an alternative mechanism for the differences observed in AL, AAL, and FLD among different genotyped chickens. It is generally accepted that TMAO inhibits bile acid synthesis by decreasing cholesterol 7α-hydroxylase expression [13]. Bile acid synthesis and excretion are the major pathways of cholesterol and lipid catabolism (see Figure 4 for details). The obstruction of bile acid synthesis limits the efflux of cholesterol and increases lipid deposition, resulting in the occurrence and development of AS and fatty liver [13,14]. In contrast, the reduction of TMAO increases the synthesis of bile acids, and this increase can alleviate lipid deposition [38]. In our study, reduced TMAO concentrations were associated with increased serum TBA concentrations in TT hens. The negative correlation between TMAO and TBA concentrations supports the interpretation that T329S can promote bile acid synthesis due to reductions in TMAO concentrations in TT hens. Moreover, the preferential precursor for bile acid biosynthesis is HDL-C [49]. HDL-C needs to be delivered to hepatocytes by high-density lipoprotein (HDL) particles, a process mediated by scavenger receptor class B type I. The formation of mature HDL requires the action of PLTP and LCAT [13,50]. In TT hens, the T329S mutation significantly (p < 0.01) increased the serum PLTP and LCAT protein concentrations. However, the serum level of HDL-C in TT hens did not decrease. Therefore, we suggest that T329S increased bile acid synthesis by reducing circulating TMAO concentrations and subsequently improving RCT in laying hens.Additionally, our study also detected that the serum CETP protein and VLDL concentrations were increased in TT hens. CETP can transfer HDL cholesterol esters (CEs) into VLDL or LDL [50]. The increased CETP may be a response to meet the demand for more CE uptake for the increased VLDL, which has already been secreted from the liver in TT hens. The major VLDL classes present in laying hens are the yolk-targeted small-diameter triacylglycerol-rich particles (VLDLy). VLDLy delivers all its triacylglycerol intact to the oocyte, which finally develops into the mature egg yolk [24]. The elevation of serum VLDL in TT hens may attribute to an increase in demand for VLDLy assembly, since it was detected that TT hens had an increasing trend in production performance in our study. A previous study reported that T329S increased the expression of apolipoprotein VLDL-II, a specific apolipoprotein of VLDLy [41], may support this statement. The association between T329S mutation and the process of yolk formation should be further explored. At the same time, the elevation of hepatic VLDL secretion may protect the liver from TG accumulation [41,51]. Consistent with it, a negative correlation (r = −0.55, p < 0.05) between serum VLDL and TG concentrations was detected in our study. However, this negative correlation does not mean that peripheral lipolysis of VLDL is increased [24]. The lack of correlation (r = 0.11, p > 0.05) between serum VLDL (ApoB) and LDL-C concentrations could just as reasonably be hypothesized to indicate that peripheral lipolysis is reduced. At the same time, it could be implied that most of the VLDL (ApoB) is being deposited in eggs not metabolized by the hen. The TT hens drive the negative correlation between VLDL and TG, and the reduction of TG may rely on the improved egg production observed in those hens. These hypotheses warrant further research.In addition, the T329S mutation diminishes the ability of FMO3 to oxidize TMA to TMAO, and this may increase the risk of fishy eggs in TT hens when they are fed a high-level TMA precursor diet [5]. Hence, we should weigh the pros and cons and apply T329S to the poultry industry appropriately or take advantage of T329S by taking a reasonable diet in the late laying period.5. ConclusionsIn conclusion, we proposed an association between the T329S mutation in FMO3 and lipid metabolic diseases during the late laying period. Our analysis of the lipid metabolism and deposition characteristics showed that the T329S mutation could alleviate AL, AAL, and FLD in laying hens, and this could, in part, through downregulation, reduce the circulating TMAO level concentrations in association with increased circulating bile acids, revealing new insights into the role of T329S in association with health-state improvements in laying hens.

animals : an open access journal from mdpi

10.3390/ani12010021

PMC8749688

To achieve conservation goals with ex situ programs, zoos have two alternatives: government funding and private donations. By using published data from 2018 on the amount of money received by zoos through adoption programs (if any), we investigated the factors influencing donations in all Slovak zoos. Generalized linear mixed models were applied throughout statistical analyses. Although the majority of animal species in the zoos included in this study had low a conservation status, a few others, like amphibians, were rarely listed as threatened species. In general, vertebrates received more funding than invertebrates, and mammals were the preferred taxa by private contributors. Mammals were sponsored more frequently than non-mammal species, except for reptiles. We submit that zoo managers could concentrate their efforts on the breeding of threatened animals to support their reintroduction to the wild and to enhance people’s awareness of these animal species.

Anthropogenic disturbance causes biodiversity loss, and consequently the captive conservation (ex situ) of threatened animals may be an effective strategy in protecting species. We used estimated body mass, phylogenetic closeness with humans, International Union for Conservation of Nature (IUCN) conservation status, and species attractiveness scores, to examine the factors influencing the adoption likelihood of a species in all Slovak zoos. In general, vertebrates received more funding than invertebrates, and mammals were the preferred taxa by private contributors. In terms of funding, we propose that the perception of mammals as phylogenetically close to humans, and attractiveness factor, contribute to an advantage over less attractive and phylogenetically distant species. Conservation status also contributed to the amount of donations; however, the magnitude of these relationships was weak when compared to the effect of animal taxa. These results suggest that Slovak zoos might be more successful in raising donations by breeding threatened species, and raising public awareness about these animal species. Displaying popular, flagship species of non-mammal taxa may increase interest among the public as well, and may translate into a significant growth in the amount of donations.

1. IntroductionThe human exploitation of Earth’s biodiversity has resulted in continuous habitat loss, environmental pollution and an overall decline in the population sizes of wild species [1]. In response to this issue, efforts to mitigate these losses can be implemented in two ways. First, through in situ conservation, which is defined as the conservation/protection of a species in the wild; second, by pursuing ex situ conservation of a threatened species in captivity. Of the two options, the latter is more commonly implemented in botanical gardens, aquariums and zoos [2].The main functions of zoos include entertainment, research, education and conservation to achieve the sustainability of populations that rely on our care [3,4]. Nevertheless, reaching these goals constitutes a financial challenge for zoos. Given that zoos and aquariums are visited by about 700 million people annually [5], they serve as suitable places for generating donations to support animal conservation [6]. Although, in general, in situ conservation is more beneficial and less costly [7], ex situ conservation may be very effective in protecting species [8,9,10] to be reintroduced to the wild [11,12]. Indeed, ex situ conservation breeding programs have successfully enhanced the survival of certain species of mammals [13], birds [14], reptiles [15], amphibians [16], and fish [17], which have resulted in high conservation status of these taxa. However, to achieve the sustainability goals of ex situ conservation programs, additional funding and private donations are necessary to support the maintenance of captive species and in situ conversation initiatives [18,19]. In addition, in some cases, in situ programs are not feasible, particularly when habitats have been devastated as a result of armed conflicts [20]. Thus, understanding and taking advantage of zoo visitor preferences for captive animals may significantly increase donation proceeds to support particular species.The propensity of private individuals to support the conservation of particular species, which could be better utilized in managing ex situ programs, is a scarce line of research in the literature. Moreover, although a good volume of research on sponsorship programs using hypothetical scenarios exist, there is a paucity in studies dealing with actual fundraising programs [21,22,23,24,25,26].In order to increase the amount of donations, visitor’s passive interest in nature [27] needs to be addressed, particularly by building on the human natural affinity for certain species and/or its status [28,29]. It has been documented that willingness to protect animals is a factor associated with the species’ appearance, and its similarity to humans [30,31,32,33]. For instance, animals that look dangerous [34], disgusting [35], single-colored [25], or unattractive [21,36], receive less conservation support than other animals. Phylogenetic proximity relates to our ability to empathize with animals [37], which seems to trigger our interest in conserving animal species similar to us [30,31,38].It is also known that human preferences and conservation initiatives are biased toward large-bodied mammals [33,39,40,41,42,43,44], often considered charismatic species. Interest in these species contributes to a bias in the breeding of large-bodied animals in zoos [45,46]. In fact, large-bodied animals are included in zoo collections more frequently than small-bodied animals, irrespective of whether they are threatened or not [12,47,48,49,50,51]. Therefore, one would expect large-bodied, charismatic animals to receive far more donations than smaller, less charismatic species.The purpose of this study was to determine which animal-related factors influence actual donations that may support fundraising programs in zoos, and ultimately strengthen the effectiveness of conservation marketing [52] of ex situ conservation programs. Willingness levels were measured using actual donations. In this study, we seek to advance this line of research in three different ways: first, our data are not restricted to one zoo, as has been typical in similar research studies. Instead, we used representative, publicly available data on animal sponsorship from all Slovak zoos. Second, we used actual donation figures, rather than hypothetical donations. Third, we did not focus exclusively on sponsored animals, but rather we compared sponsored against non-sponsored animals to better understand which features made the sponsored animals attractive to donors. Fourth, these animal species were compared using a range of key characteristics (phylogenetic distance, estimated body mass, appeal factor and conservation status).2. Materials and Methods2.1. Defining SponsoringWe analyzed publicly available data published in 2018 in the yearbooks of three Slovak zoos (Bratislava, Bojnice and Spišská Nová Ves) [53,54,55]. The fourth Slovak zoo, in Košice, does not publish yearbooks. In this case, we requested data directly from the zoo headquarters. All yearbooks, including the data from the Košice zoo, contained lists of animal species kept in the zoo, their actual numbers, the conservation status, and information about animal sponsorship. In the context of this study, sponsoring could be defined as the act of willingly transferring personal (institution or company) funds to a conservation organization on an annual basis to sustain a particular species.Any contribution, large or small, is accepted by the zoos. Funding received from sponsors is predominantly used for species nutrition, and less is used for the reconstruction of the breeding facility, or for the development of the zoo in general. In terms of securing monetary donations, this mostly occurs by people contacting the zoo via e-mail or telephone. When a donation is made, the individual, family, or institution name is added to a placard, which is placed in front of the animal’s enclosure. Each sponsoring period lasts one year, and one species can have multiple sponsors. If an individual donates more than what it is required for a particular species per year, his/her donation is utilized in subsequent years, and the sponsoring is extended. By making the sponsoring status visible to visitors, the zoo underlines a privileged relationship between the sponsoring individuals or institutions and a particular animal species, compared with other non-sponsoring zoo visitors.Animal adoption programs are common sponsoring strategies; however, other approaches exist. Additional sponsoring methods include donations made by large companies or corporations. In this kind of sponsorship, the contribution does not support a particular species, but the entire zoo. Therefore, hereafter we will use the term “sponsoring” (or “sponsorship”) to refer to animal adoption programs, predominantly made by individuals or small companies. In all these cases, the funds are to be used in the support of specific species. Sponsorship programs are advertised by zoos on their websites. According to the sponsoring individual, family or institution preferences, the zoos allow the sponsoring of different species. Thus, although the actual patterns of sponsorship are partly predetermined by zoos, they also reflect individual preferences. Published yearbooks do not indicate which species were advertised by zoo; thus, we obtained this information directly from the zoos.The four zoos are located in different Slovak regions (Figure 1). The Bojnice Zoo from the western region is a 42-hectare facility, and is the oldest and most frequently visited zoo (~350,000 people annually); compared with the other zoos, it has the highest species diversity (~400 species, Table 1). The zoo is located in Bojnice, a small town of ~5000 inhabitants, and does not specialize in breeding any specific species. The zoo in Košice is the third largest zoo in Europe (its area is 288 hectares), and is located in East Slovakia in the second largest Slovakian city (~240,000 inhabitants). The zoo is home to about 200 species of animals, and its visitation rate is about 200,000 people per year. The Bratislava Zoo, in the country’s capital city (~450,000 inhabitants), has an area of 96 hectares. It has more than 150 animal species. It also has a high visitation rate, approximately 300,000 people annually. The zoo in Spišská Nová Ves is the smallest of the four zoos (8 hectares) included in this study, and is located in East Slovakia, with about 100 species of animals. The city has about 40,000 inhabitants. It has the lowest visitation rate: about 100,000 people per year come to visit this zoo.2.2. Measured Variables2.2.1. Sponsoring (Dependent Variable)Sponsoring was examined using the following variables: total and mean contribution amounts for all species per zoo in 2018 (in Euros) and sponsorship frequency in 2018. Whether the sponsoring occurred or not was defined as a binary variable.2.2.2. Conservation Status (Independent Variable)The conservation status of all listed animals in each zoo was derived from IUCN categories [56]. For statistical purposes, we transformed these categories into a threatened/non-threatened binary variable, where threatened corresponds to VU, EN and CR as per IUCN terminology.2.2.3. Body Mass and Phylogenetic Closeness with Humans (Independent Variable)The body mass of each species was calculated using data from Jones et al. [57]. Data were log-transformed, following the recommendations from Smith et al. [42]. The phylogenetic divergence time from humans (in millions of years) was obtained for each species from timetree.org [58]. A few specimens were apparently hybrids, or their scientific names were incorrect. In such cases, body mass or phylogenetic closeness data were omitted.2.2.4. Appealing Species (Independent Variable)The appealing species factor was defined according to the appeal scores available for 4320 species of mammals [59]. These scores express participants’ preference for each species in the context of conservation. Higher scores mean greater appeal (range = −0.77–5.01). Further details about the score can be found in MacDonald et al.’s (2017) study [59].2.3. Statistical AnalysisThe data were analyzed using Generalized Linear Mixed Models (GLMMs) where zoo and species identity were defined as random factors. Dependent variables were binomial (occurrence of sponsoring or not) or continuous (amount of donated money) with Poisson distribution. Data were checked for normality using the Kolmogorov–Smirnov test. As at least some of them were not normally distributed, additional comparisons of body mass and species appeal scores between threatened and non-threatened species were performed with the Mann–Whitney U-test. Correlations between species appeal score, body mass, phylogenetic closeness and financial support were performed with the Spearman correlation coefficient. It is important to note that in this study, sample sizes varied due to the lack of data on conservation status, body mass or appeal scores for certain species. We mainly investigated sponsored species in comparison to all non-sponsored species residing in zoos. In order to address whether possible differences could be confounded by species advertised on zoo web pages, we ran additional analyses with a subsample of advertised species. Statistical tests were performed in SPSS version 23.3. Results3.1. Sponsoring of Vertebrates versus InvertebratesThe total number of species across all zoos was 813. Of this number, nearly 25% of the species (N = 237) were reported in more than one zoo. A combined total of 1055 animal species were listed in all four zoos (Table 1, Figure 1).Considering reports from all the four zoos, vertebrates (148/984, 15%) were sponsored more frequently than invertebrates (1/71, 1.4%) (GLM with binomial distribution of data, Walds χ2 = 8.1, df = 1, p = 0.004). Of the 813 animal species in the four zoos, the Emperor scorpion (Pandinus imperator) was the only sponsored invertebrate species. Vertebrates were sponsored exclusively by individuals (57%), corporations (33%) or both (10%). As there were no significant differences between vertebrate classes being sponsored by individuals, corporations or both groups (Chi-square test, χ2 = 3.8, df = 6, p = 0.7), we are not referring to sources of sponsorship in subsequent analyses.Considering reports on all 682 animals advertised on zoo web pages, once again, vertebrates (148/662, 22%) were sponsored more frequently than invertebrates (1/26, 4%). This difference certainly approached significant levels (F2,571 = 2.85, p = 0.059).Sponsoring occurrences were more frequent in Bratislava (50%), followed by Spišská Nová Ves (24%), in Košice (18%) and Bojnice (5%). Therefore, we defined the effect of zoo location (city) along with species ID as random factors in the subsequent statistical analyses. Taking into account that the sponsoring rate of invertebrates was exceptionally rare, we made the decision to continue with analyses exclusively on vertebrates.3.1.1. Vertebrate SponsorshipVertebrate classes and conservation status were predictors in the GLMM model; species ID and zoo were random factors, and the occurrence of sponsoring was the dependent variable. The model was significant (F5,769 = 3.56, p = 0.003). Vertebrate classes (F4,769 = 4.22, p = 0.02), but not conservation status (F1,769 = 3.0, p = 0.08), played a significant role in sponsoring occurrence. The interaction term was not statistically significant (F4,765 = 0.26, p = 0.90). The effect of the vertebrate class factor was very clear. Mammals were sponsored more frequently than non-mammal species (all p < 0.01), except for reptiles (p = 0.07) (Figure 2). When the total sponsoring number (range: 0–7, mean = 0.24, SE = 0.02, N = 984) was set as a dependent variable, the model was significant (F9,765 = 2.74, p = 0.004), but neither vertebrate class (F4,765 = 2.02, p = 0.09), nor the conservation status (F1,765 = 0.38, p = 0.54), or the interaction term (F4,765 = 0.46, p = 0.76) showed any significant influence on the dependent variable.When we considered only a subset of species that were advertised on zoo websites, the results were almost identical; the model was significant (F6,549 = 2.24, p = 0.038). Vertebrate classes (F5,549 = 2.65, p = 0.02), but not conservation status (F1,549 = 0.03, p = 0.87), played a significant role in sponsoring occurrence. The interaction term was not statistically significant (F4,545 = 0.25, p = 0.91).3.1.2. Amount of Financial Support by Animal ClassBy looking at the large amounts of money contributed to the mammal class, our previous analysis suggests that in terms of sponsoring occurrences, this was the preferred class among vertebrate animals. In subsequent analyses, we calculated whether there were any differences in the mean amount of financial support among vertebrate classes in both sponsored and non-sponsored animals. The GLMM with Poisson distribution was set with the same predictors as in the previous analysis, and with the mean amount of money contributed per species as the dependent variable. The model was significant (F9,644 = 698.4, p < 0.001). The vertebrate class was a strong predictor of the mean amount of proceeds (F4,644 = 1367.4, p < 0.001). Contrast analysis between vertebrate classes showed that mammals received the greatest amount of donations compared to all the other vertebrate classes (all p < 0.001, Table S1). Species with a high conservation status received greater support than those with a lower conservation status (F1,644 = 94.1, p < 0.001). The interaction term was significant (F4,644 = 148.7, p < 0.001), suggesting that threatened mammals received greater support than non-threatened mammals. In contrast, no differences between animals with low and high conservation status among remaining vertebrate taxa were observed.When considering advertised species, the results were almost identical (whole GLMM model, F7,96 = 771.8, vertebrate class, F4,96 = 768.9, conservation status, F1,96 = 1135.5, interaction term, F2,96 = 232.4, all p < 0.001).When the total amount of money was defined as the dependent variable, the model remained significant (F8,572 = 934.6, p < 0.001). Again, vertebrate class (F4,572 = 1835.6, p < 0.001) and conservation status (F1,572 = 444.4, p < 0.001), along with the interaction term (F3,572 = 21.3, p < 0.001) were significant predictors of total donation amount. Again, mammals received more donations than other vertebrate classes (analysis of contrasts, all p < 0.001). The top 10 mammal and bird species with the highest mean amount of donated money are shown in Table 2.Considering the advertised species, results were once again almost identical (whole GLMM model, F7,103 = 1545.8, vertebrate class, F4,103 = 1865.4, conservation status, F1,103 = 579.9, interaction term, F2,103 = 70.8, all p < 0.001).3.1.3. Factors Influencing the Sponsoring of MammalsWe conducted an analysis restricted to mammals, with occurrence of sponsoring as the dependent variable, and conservation status, body mass, phylogenetic closeness to humans and species appeal scores as predictors, and the ID of species and zoos as the random factors. The GLMM model was not significant (Table 3). Although all main effects were non-significant, the interaction term Conservation Status × Species Appeal certainly approached significance (Table 3). This finding suggests that the appeal scores of sponsored threatened species were higher than those of sponsored, but non-threatened, species. In contrast, there was a small difference in appeal scores in non-sponsored threatened or non-threatened mammals.The GLMM with a subset of advertised species was not statistically significant (F8,142 = 0.99, p = 0.44, all p > 0.14 for fixed effects including the interaction terms).3.1.4. Additional Relationships between Sponsoring and Species CharacteristicsIt was also noted that body mass had a positive and moderate correlation with the appeal score (Spearman r = 0.42, p < 0.001, N = 177) and that threatened species showed higher appeal scores (median = 2.23, 95% CI [2.02, 2.55], N = 76) than non-threatened species (median = 0.6, 95% CI [0.43, 0.71], N = 161) (Mann–Whitney U-test, U = 9039, p < 0.001). Although threatened species tended to be heavier (median = 154.3 kg, 95% CI [14, 324.1], N = 58) than non-threatened species (median = 10 kg, 95% CI [32.6, 68.3], N = 125), this difference was not significant (Mann–Whitney U-test, U = 4045.5, p = 0.23).Correlations and comparison with a subset of advertised species showed almost identical results as in the cases above mentioned.Spearman correlations did not show significant associations between the mean or total amount of financial support and species body mass, phylogenetic closeness, or appeal score (all p > 0.2). The same results were obtained with a subsample of the advertised species.4. Discussion4.1. Phylogenetic Closeness to HumansWe investigated the relationship between donations and species phylogenetic closeness to humans at various levels. First invertebrates, which are phylogenetically distant from humans, were sponsored at lower rates compared with those species phylogenetically closer to vertebrates. This result corroborates what prior research studies have found, that people have negative attitudes and perceptions towards invertebrates [33,60,61,62,63,64]. Low conservation status of invertebrates in Slovak zoos, and to certain extent, their lack of promotion campaigns among the general public contribute to being ignored in sponsoring programs. It is possible that zoos do not engage in fundraising in favor of invertebrate species through sponsorship programs. We submit that flagship species, such as butterflies, dragonflies or corals, which are highly regarded by zoo visitors [44,63,65], should be used to increase the awareness of these species among the public and the willingness to support their conservation.Second, across animal orders, mammals that are phylogenetically closer to humans had higher sponsoring rates than phylogenetically distant vertebrates. Thus, the likeability or appeal of mammals at the expense of other taxa plays a role in species preferences [66]. A global analysis focused on crowdfunding platforms also showed that mammals, followed by birds, received more funding than other animal taxa [67]. When the relationship between phylogenetic distance and the likelihood of sponsoring/amount of donations were analyzed (exclusively within the mammal class), the correlation was not significant as was expected, according to the phylogenetic closeness hypothesis. In particular, we noted that the likelihood of a species being sponsored was weakly associated with phylogenetic closeness to humans. Although the amount of monetary contributions was greatly influenced by phylogenetic closeness [26], the magnitude of this relationship was much weaker when compared with species appeal. These results suggest that the willingness of an individual or organization to commit to an annual sponsoring contribution in Slovak zoos can be only partly explained by phylogenetic closeness. However, certain animal physical attributes seem to be relevant when securing annual sponsorships. It should be noted that “similarity” does not necessarily mean the same as “phylogenetic closeness”. For instance, a chimpanzee, the phylogenetically closest mammal to humans is much less preferred by children than parrots or dolphins [64], despite both parrots and dolphins being more phylogenetically distant to humans than chimpanzees. In fact, neither chimpanzees nor orangutans, as the phylogenetically closest species to us (note that gorillas, another phylogenetically close relative, are not breed in Slovak zoos) were sponsored in Slovak zoos. Perhaps this can be explained on the basis of the perceived beauty of the parrots [48] and the playful and altruistic nature of parrots and dolphins (unlike chimpanzees) that make them popular and attractive species in zoo collections [68]. This factor should be taken into consideration in future assessments of human–animal relationships, to better understand the willingness of individuals and organizations to become involved in species conservation programs.4.2. The Effect of Species Appeal and Body SizeAppealing species were no more frequently sponsored, and did not receive greater amounts of donations than less-appealing species. This finding was also reported in a recent study conducted in Finland [69]. Moreover, body mass of mammals had a positive correlation with their appeal score, and threatened species showed higher appeal scores than non-threatened species. This finding is in correspondence with the overall human affinity towards large charismatic animals [39,40,70]; however, the willingness to support conservation efforts focused on these species [23,26,44,71] was not confirmed. Some researchers suggested [72,73] that these charismatic species would be beneficial by maximizing shared benefits with other species. This feature can also reflect the neglect of non-charismatic/Cinderella species in zoos [29,42,74].4.3. The Effect of Conservation StatusSpecies conservation status played a significant, but not prominent role in sponsoring, and in the amount of donations. Our analyses suggest that endangerment of vertebrates did not automatically result in sponsoring occurrences. Moreover, threatened mammals were not likely to be sponsored, compared with non-threatened mammals. Regarding mean donation amounts, the conservation status of vertebrates was a significant factor, but this was much weaker than vertebrate class, with mammals receiving most of the donations. Similar results were reported in another study [59] in which appealing species did not necessarily have a high conservation status. Sponsored threatened mammals tended to be more attractive when compared with sponsored non-threatened or non-sponsored (threatened or not) mammals, which once again suggests that species attractiveness and conservation status play a non-trivial role in sponsoring decisions. In terms of sponsoring occurrence, these findings do not lend significant support to the assumption that people seem to prefer large-bodied over small-bodied vertebrates [39,40,41,70].The majority of animals in the four Slovak zoos had low conservation status (20% of vertebrates with an IUCN status were threatened). This was true for invertebrates, fish and, in particular, amphibians. These taxa are globally neglected, not only by zoo visitors [66], but also in fundraising programs taking place around the globe [67]. It is apparent that zoos did not value the conservation priority of amphibians, which are subject to captive breeding [75]. This trend reflects the low numbers of amphibian species in European zoos [76]. We propose that increasing the proportion of endangered species and the use of aesthetically appealing amphibians in advertisement [77] can contribute to an increase in sponsorship rates, and ultimately strengthen the investment capacity of zoos. A similar approach can be applied to other neglected taxa.In a survey with Australian participants, the authors found that people preferred in situ management of wild populations of native bird species, over captive breeding [78]. In fact, it is recommended that zoos consider an in situ–ex situ combined approach in their conservation activities [12]. Employing in situ conservation can be complicated, as determined by the strong preferences for exotic animals over native ones [79], except for vertebrates including the brown bear, the wild boar and the golden eagle. The remaining 17 species listed with top mean contributions (Table 1) were exotic. Out of these 20 species, only the golden eagle has an endangered status, and we suggest it could be used for in situ programs directly in Slovakia. Therefore, zoos can benefit from designing and implementing management strategies that promote public interest in threatened native species, and develop purposeful campaigns to generate monetary donations.4.4. Gaps and Recommendations for Future ResearchAnimal collections in zoos are predominantly focused on popular or well-known and non-threatened animals. We propose that zoo managers could concentrate their efforts on the breeding of threatened animals to support their reintroduction to the wild and to enhance people’s awareness of these animal species. Future research could involve sponsoring individuals and organizations to gather information regarding their decision to participate in sponsoring programs.5. ConclusionsIn conclusion, increasing the number and amounts of donations in Slovak zoos through sponsoring programs is an initiative largely influenced by animal taxa. Moreover, motives are partially influenced by animal conservation status and by species attractiveness, rather than by animal size. In all probability, individuals and organizations’ willingness to sponsor animals in zoos is multidimensional [80]. Displaying popular, flagship species of non-mammal taxa may also raise public awareness and interest, and in turn increase the amount of donations. Interviewing individuals and organizations participating in sponsoring campaigns (and those who did not) may provide further insights into individuals’ motivation to sponsor animals in zoos.

animals : an open access journal from mdpi

10.3390/ani11113027

PMC8614390

The chemical Nano-Selenium (Che-SeNPs) is a good example of applied nanotechnology used in the area of nutritional supplements due to its advantages and properties. From our results, dietary supplementation with Che-SeNPs could improve the performance of growing quails; the best level was 0.4 g Che-SeNPs/kg feed. Thus, this study supports the application of Che-SeNPs in quail diets in an effort to improve the productive and physiological performance. The results revealed that Che-SeNPs boosts the growth, blood biochemistry, antioxidant indices, immunity, and bacterial environment of the intestine of quail.

Nano-minerals are used to enhance mineral bioavailability, which helps improve animal growth and health. The use of chemical nano-selenium (Che-SeNPs) has lately attracted great scientific interest, mainly due to its potential benefits for poultry. The current study was conducted to investigate the impact of the dietary supplementation of Che-SeNPs on the growth performance, carcass traits, blood constituents, antioxidant status, immunity, and gut microbiota of Japanese quails. A total of one week-old 180 Japanese quails were randomly distributed into four equal groups, and each group consisted of 45 unsexed birds with five replications (nine birds each). The first group was fed a basal diet without supplementation (0 g/kg Che-SeNPs), and the second, third, and fourth groups were fed diets containing 0.2, 0.4, and 0.6 g/kg Che-SeNPs, respectively. The results showed that the dietary supplementation of Che-SeNPs significantly (p < 0.0001) increased body weight, body weight gain, and feed conversion ratio, but decreased feed intake (p < 0.0001) compared to the control group. The highest values of growth performance were recorded in the group fed 0.4 g Che-SeNPs g/kg feed. Che-SeNPs levels did not affect the carcass traits, relative organs (except liver), or blood hematology (except platelet count and hemoglobin level) of quails. Plasma total protein, albumin, aspartate amino transferase (AST), and urea values were not affected by dietary Che-SeNPs, but alanine aminotransferase and lactate dehydrogenase values declined. Globulin and creatinine values were linearly increased with the inclusion of Che-SeNPs (0.4 and 0.6 g/kg) in quail diets compared to the control. The supplementation of Che-SeNPs in quail diets significantly improved (p < 0.05) the plasma lipid profile and activities of antioxidant enzymes compared to the control group. Immunoglobulin G values of Che-SeNPs (0.4 and 0.6 g/kg) were higher (p < 0.05) than those in the control group. The groups fed diets supplemented with Che-SeNPs showed lower (p < 0.0001) total bacterial count, total yeast and molds count, Coliform, Escherichia coli, Enterococcus spp., and Salmonella spp. colonization, and higher (p = 0.0003 and 0.0048) lactic acid bacteria counts than those in the control group. In conclusion, Che-SeNPs supplemented up to 0.4 g/kg can improve the performance, lipid profile, antioxidant indices, and immunity, as well as decrease intestinal pathogens in quails during the fattening period (1–5 weeks of age).

1. IntroductionSelenium (Se) is one of the elements that can be used in diets as the chemical nano-selenium (Che-SeNPs). Se is required for the maintenance of physiological functions, growth, and health of birds. It also plays a crucial role in nutritional value and feed metabolism, leading to considerable growth [1]. Che-SeNPs has attracted more attention because of its strong adsorbing ability, high catalytic efficiency, high surface activity, and low toxicity compared to that of other chemical Se forms [2]. The high absorption of Che-SeNPs from the intestinal lumen into the body was observed. Shirsat et al. [3] highlighted that Che-SeNPs has antioxidant, anticancer, antibacterial, and antiprotozoal properties. El-Deep et al. [4] stated that dietary Che-SeNPs supplementation enhanced growth performance by improving immune or antioxidative properties in broiler chicks. Additionally, Ahmadi et al. [5] revealed that the dietary supplementation of Che-SeNPs improved growth performance and immune function without the deleterious effects on the internal organs of broiler chickens.Previous investigations exhibited that Che-SeNPs augmented body weight gain and improved antioxidant functions of Arbor Acres broilers [3,6]. Se nanoparticles have also been utilized in food preservation methods such as packing food items and antiseptic coating over food materials. Studies have been conducted to highlight the disinfectant properties of Se nanoparticles against Pseudomonas aeruginosa, and Proteus mirabilis [7]. On the other hand, natural agents and trace elements including nanoparticles as feed additives may affect the diversity of gut microbiota and health [8]. Se is one of the important elements that can help microbiota to complete its action within the gut [9]. In this concern, the caecal counts of Salmonella and E. coli of quails were decreased in birds fed diets containing nano-curcumin when compared to the control diets [8].Selenium can be considered an essential trace element and micronutrient for living creatures at low concentrations, but it becomes toxic and harmful at higher dose [2]. The extensive use of nano Se in nanotechnologies and medicine has increased the risk of their contamination in the environment, which could harm living species; however, it is useful to understand the assessment of Se-NPs toxicity to the biological ecosystem [10]. Nano-Se has lower toxicity than selenomethionine and is now the least toxic of all supplments of Se. Nano-Se has a threefold lower toxicity than organic Se and a sevenfold lower toxicity than inorganic Se [10].The positive impacts of nanotechnology involving Se are well-known in many pathological conditions [11]. However, the inclusion of Che-SeNPs in quail diet during the growth period remains limited. It is hypothesized that the dietary addition of Che-SeNPs is expected to exert beneficial effects on growing quails. Therefore, the purpose of this study was to evaluate the antibacterial and antifungal activities of Che-SeNPs, and its beneficial effects on the growth, feed utilization, carcass traits, hematology, blood constituents, and cecal microbiota of growing quails.2. Materials and Methods2.1. Source of Selenium NanoparticlesThe study was carried out at Zagazig University, Zagazig, Egypt in conjunction with King Abdulaziz University, Jeddah, Saudi Arabia under protocol no: (FP-73-43). In this study, Che-SeNPs were prepared using wet chemicals. Sodium selenite (Na2SeO3) was used for producing Se nanoparticles with ascorbic acid (C6H8O6) as a reducing agent. A stock of aqueous solution of 100 mM Na2SeO3 and 50 mM C6H8O6 was prepared in a 1: 4 ratio. The solution was kept under a magnetic stirring condition at different rpm and ambient temperature for 30 min. The mixtures were allowed to react with each other in the concentrated form until the mixture changed from colorless to red. Next, the solution was centrifuged at 3000 rpm, pellets were collected, and Che-SeNPs was obtained [12,13]. Chemically synthesized nano-selenium was determined via UV–Vis spectroscopy using an automated spectrometer (Spectro UV–Vis double beam UVD 3500). The morphology and element percentage of selenium nanoparticles were measured using transmission electron microscopy and an energy dispersive X-ray analytical instrument. Fourier transform infrared spectroscopy (JASCO) was used to determine the properties of produced selenium nanoparticles including size, shape, charge, and stability. Characterization of Che-SeNPs; maximum UV absorbance at 300 nm, spherical shape by TEM, size (75.68 nm) and charge (−23.26 mV) by zeta seizer, and zeta potential, respectively.2.2. Antibacterial Activity of Che-SeNPsListeria monocytogenes ATCC 15313, Staphylococcus aureus MTCC 1809, Bacillus cereus ATCC 11778, E. coli ATCC 25922, P. aeruginosa ATCC 27853, and Salmonella enterica MTCC 1253 were used in this study. The antibacterial activity of Che-SeNPs against animal and human pathogenic Gram-negative bacteria, P. aeruginosa ATCC 27853, E. coli ATCC 25922, and S. enterica MTCC 1253, and Gram-positive bacteria, L. monocytogenes ATCC 15313, B. cereus ATCC 11778, and S. aureus MTTC 1809, were estimated using the disc diffusion assay method. Mueller–Hinton agar medium consisting of peptone, beef extract, yeast extract, NaCl, and agar with 5, 3, 5, 5, and 20 g, respectively in 1 L of distilled water was prepared in slant to preserve all bacterial strains. Mueller–Hinton broth was used to activate bacterial cells; one hundred microliters of each bacterium (1 × 109 CFU/mL) were spread with sterile swabs in Mueller–Hinton agar plates. Freshly prepared selenium nanoparticles with different concentrations (50, 100, 200, 400, and 800 μg/mL) were loaded on paper discs (disc diameter was about 6 mm) and then were placed on the Muller–Hinton agar plates. Sodium selenite (50 μg/mL) and sterilized deionized water were loaded on paper discs and used as a positive and negative control, respectively. Mueller-Hinton agar plates were incubated for 24 h at 37 °C. After incubation, the obtained zones of inhibition surrounded the Che-SeNPs discs were measured and recorded as the mean ± standard deviation if they were greater than 6 mm. The minimum inhibitory concentration (MIC) of the Che-SeNPs was calculated based on a broth micro dilution method. Briefly, six pathogenic bacteria were cultured overnight at 37 °C in Mueller–Hinton broth and were adjusted to a final density of 109 CFU/mL by 0.5 McFarland standards. The Che-SeNPs (1 mg/mL) were homogenized with sterilized deionized water and dilutions of 50, 100, 200, 400, and 800 μg/mL were made. Next, 10 μL of different concentrations of Che-SeNPs was mixed in sterile test tubes contain 10 μL of bacterial inoculum and 90 μL of Mueller–Hinton broth. The test tubes were incubated for a day at 37 °C. The lower concentration of Che-SeNPs which inhibited bacterial strains growth or turbidity was considered the MIC. The lower concentration of Che-SeNPs which totally killed bacterial strains was defined as the minimum bactericidal concentration (MBC). The experiments were carried out in triplicate [8].2.3. Antifungal Activity of Che-SeNPsThe antifungal activity of the Che-SeNPs was tested against animals and human pathogenic Candida strains. Candida albicans ATCC 4862, C. glabrataATCC64677, C. parapsilosis ATCC 22019, and C. guilliermondii ATCC 6260 were used in this study. The antifungal activity of Che-SeNPs against these four strains was evaluated via the disc diffusion method [14] using sterile cotton swab lawn cultures of selected fungi that were prepared on Sabouraud Dextrose agar (SDA) plates. Che-SeNPs was loaded on paper discs (disc diameter was about 6 mm) and then was placed in SDA surface. Selenium selenite and sterilized deionized water were used as the positive and negative controls, respectively. The plates were then incubated for 36h at 30 °C. The Che-SeNPs were tested for MIC using the broth dilution method [14]. Sabouraud broth was used as diluents for fungal species. About 106 CFU/mL cells could be inoculated. The Che-SeNPs levels (50 to 800 μg/mL) were prepared in sterilized deionized water and homogenized. Next, 10 μL of different concentrations of Che-SeNPs were mixed in sterile test tubes containing 10 μL of Candida inoculum and 90 μL of Sabouraud broth. The test tubes were incubated at 30 °C for 36 h. The obtained turbidity was estimated at 600 nm to determine the MIC values. The minimum concentration of Che-SeNPs that reduced fungi growth by 90% was considered the minimum inhibitory concentration (MIC). The concentration of Che-SeNPs at which complete fungal growth was not observed was defined as the minimum fungicidal concentration (MFC). The experiments were replicated in triplicate.2.4. Animals, Design, and Diets A total of 180 one-week-old Japanese quails with an average weight of 27.17 ± 0.075 g were used. Quail chicks were randomly allocated into four groups, and each group consisted of 45 unsexed birds with five replications (nine birds each). Quails were kept in conventional cages (90 × 40× 40 cm), and feed and water were open during the study (4 weeks). The treatments were as follows: the first group was fed the basal ration which containing 150 mg of Se as Se selenite, whereas the second, third, and fourth groups were fed diets supplemented with 0.2, 0.4, and 0.6 g/kg of Che-SeNP, respectively. The Che-SeNP was added at the top of the basal diet at the highest level and then diluted with the unsupplemented basal diet to achieve the desire concentration. The basal diet (Table 1) was formulated to meet the birds’ requirements according to NRC [15]. The Ethics statement for Animal care and maintenance were in accordance with the guidelines of the Egyptian Research Ethics Committee and the guidelines for the Care and Use of Laboratory Animals by Zagazig University (ZU-IACUC/2/F/56/2021).2.5. Growth Performance and Carcass MeasurementsAll growth parameters and feed utilization were measured at 1, 3, and 5 weeks of age. For carcass examinations, at 5weeks old, 20 birds (5 per treatment) were randomly selected, weighed, and euthanized. All edible parts were weighed and expressed as a percent of the live body weight before slaughter.2.6. Microbiological AnalysisTen grams of quail cecum samples (five samples per each treatment) were homogenized and transferred to a 250 mL conical flask containing 90 mL of sterile physiological saline solution consisting of 0.1% peptone and 0.85%NaCl; the mixture was well-mixed to obtain a 10−1 dilution. Serial dilutions from the previous dilution (10−1) were prepared to obtain up to 10−6 dilution. The total bacterial count was counted using plate count agar medium at 30 °C for 24 h, and the total count of Enterococcus spp., was counted using Chromocult enterococci agar medium [16,17]. Total coliforms were enumerated by using MacConkey agar medium. Biochemical methods such as indole test, citrate reactions, methyl red, and Voges–Proskauer were used to identify Escherichia coli. DeManRogosa Sharpe agar was used to enumerate the lactic acid bacteria. Salmonella Shigella agar (SSA) media (Oxide CM 99) was used to count the Salmonella spp. The appearance of black colonies on SSA indicated the presence of Salmonella spp. SSA plates were incubated at 37 °C for 1 day. Sabouraud Dextrose agar (SDA) was used to count the molds and yeasts. SDA plates were incubated at 25 °C for 3–7 days. All the obtained microbiological results were then converted to logarithmic colony-forming units per gram (CFU/g) [8,18,19].2.7. Blood Chemistry After euthanization, blood samples were randomly collected from five quails per treatment into heparinized tubes. Hematological parameters were measured. Regarding biochemical parameters, we used a centrifuge (Janetzki, T32c, 5000 rpm, Germany) at 2146.56× g for 15 min to separate the plasma. The biochemical blood parameters were determined using commercial kits from Biodiagnostic Company (Giza, Egypt).2.8. StatisticsThe statistical analyses were carried out using SAS. The data of growth rate, feed efficiency, carcass parameters, hematology, blood chemistry, and microbiology were analyzed with a one-way analysis of variance using a normal distribution and the replicate as the experimental unit. Orthogonal polynomial contrasts were used to test the significance (linear and quadratic) of the gradual levels of dietary Che-SeNPs using the post-hoc Tukey’s test (p < 0.05).3. Results3.1. Antibacterial Activity of Che-SeNPsThree animal and human pathogenic Gram-negative bacteria (E. coli, P. aeruginosa, and S. enterica) and three Gram-positive bacteria (L. monocytogenes, B. cereus, and S. aureus) were selected to test Che-SeNPs antibacterial activity (Table 2). The maximum zones of inhibitions were observed in the three Gram-positive bacterial strains L. monocytogenes ATCC 15313, B. cereus ATCC 11778, and S. aureus MTTC 1809. The antibacterial activity of Che-SeNPs increased with increasing concentrations of Che-SeNPs. The effect of Che-SeNPs was superior to that of sodium selenite as an antimicrobial agent against tested pathogenic microorganisms; moreover, the deionized water did not show any antimicrobial activity. The highest MIC of Che-SeNPs against E. coli ATCC 25922, P. aeruginosa ATCC 27853, and S. enterica MTCC 1253 was 45, 40, and 50μg/mL, respectively, whereas, the lowest MIC was 30, 35, and 25 μg/mL against L. monocytogenes ATCC 15313, S. aureus MTTC 1809, and B. cereus ATCC 11778, respectively (Table 2).3.2. Antifungal Activity of Che-SeNPsChe-SeNPs showed acceptable antifungal activity, ranging from 50 μg/mL to 800 μg/mL, against all the tested fungal strains. C. albicans ATCC 4862 was the most sensitive strain to Che-SeNP when compared to other tested strains. The MICs for C. albicans ATCC 4862, C. glabrata ATCC64677, C. parapsilosis ATCC 22019, and C. guilliermondii ATCC 6260 were 70, 90, 80, and 100 μg/mL, respectively (Table 3). 3.3. Growth Performance The effects of dietary Che-SeNPs supplementation on the growth performance of Japanese quails are presented in Table 4. From the results, quails fed diets containing Che-SeNPs had significantly higher body weight (BW) (linear, p < 0.0001 and quadratic, p = 0.0004) and body weight gain (BWG) (linear, p < 0.0001 and quadratic, p = 0.0005) during the whole experimental period. The group fed diets containing Che-SeNPs (0.4 g/kg diet) had the highest BW and BWG. Feed intake was decreased (linear and quadratic, p < 0.001) in the Che-SeNPs groups compared with that in the control group. The feed conversion ratio was linearly and quadratically improved with the addition of Che-SeNPs in quail diets during all periods. Generally, the best growth performance parameters were recorded in the group fed 0.4 g Che-SeNPs/kg feed.3.4. Carcass Traits As indicated in Table 5, dietary Che-SeNPs levels did not affect carcass traits and relative organs (except liver) of Japanese quails. Supplementation of Che-SeNPs significantly increased the relative weight of the liver (linear and quadratic, p < 0.05) compared to the control group.3.5. Blood HematologyThe effects of the addition of Che-SeNPs on the blood hematology of growing quails are presented in Table 6. Dietary Che-SeNPs levels did not affect (p > 0.05) white blood cells (WBCs), lymphocytes, mid-range, granulocytes, red blood cells (RBCs), hematocrit, and mean corpuscular volume of the growing quails. Platelet count and hemoglobin (Hb) levels were increased (linear, p < 0.05) by the addition of Che-SeNPs at 0.4 and 0.6 g/kg. The mean corpuscular volume value was augmented (linear, p < 0.05) by the addition of Che-SeNPs level compared to the control group (without Che-SeNPs). The supplementation of dietary Che-SeNPs at levels of 0.6 g/kg decreased values of red blood cell distribution width linearly (p = 0.0091). In contrast, the dietary levels of 0.4 and 0.6 g/kg declined the values of red blood cell distribution volume linearly (p = 0.0019) compared to the control group.3.6. Blood ConstituentsLiver and kidney function data are presented in Table 7. The total protein and albumin were not affected (p > 0.05) by dietary Che-SeNPs. The globulin value was lowest (linear, p < 0.05) in the Che-SeNPs levels of 0.4 and 0.6 g/kg compared with the Che-SeNPs levels of 0.2 g/kg and the control group. The quails fed diets containing Che-SeNPs had higher alanine aminotransferase (ALT) and lactate dehydrogenase (p < 0.05) than those in the control group. Dietary Che-SeNPs had no significant effect on AST and urea values. The inclusion of Che-SeNPs (0.4 and 0.6 g/kg) in quail diets increased the creatinine value (linear, p < 0.05) compared with that in the control and 0.2 g/kg Che-SeNPs groups.The effects of Che-SeNPs inclusion in diets on the lipid profile of quails are presented in Table 8. Total cholesterol, triglyceride, and very-low-density lipoprotein were significantly decreased (p < 0.05) in Che-SeNPs-treated groups compared to those in control. The dietary supplementation of Che-SeNPs (0.2 and 0.4 g/kg) quadratically increased high-density lipoprotein (HDL) (p = 0.0019).3.7. Antioxidant IndicesThe results of the antioxidant indices in the serum are given in Table 8. The activities of superoxide dismutase (SOD) and glutathione peroxidase, and the levels of reduced glutathione (GSH) were significantly increased (linear and quadratic, p < 0.05) by the dietary supplementation of Che-SeNPs compared with those in control. Dietary Che-SeNPs levels decreased malondialdehyde (MDA) levels linearly (p < 0.0001) compared to the control group. The values of immunoglobulin G (IgG) of Che-SeNPs-treated groups were higher (linear and quadratic, p < 0.05) than those in the control group. IgM and IgA values of Che-SeNPs (0.4 and 0.6 g/kg) were higher (p < 0.05) than those in the control group. The quails fed a diet supplemented with Che-SeNPs showed higher plasma selenium concentrations when compared to those fed the control diet (linear, p = 0.0001).3.8. Microbiological AnalysisThe different Che-SeNPs levels significantly affected the cecal microbiota of growing Japanese quails (Table 9). The groups fed a diet supplemented with Che-SeNPs showed lower total bacterial count, total yeast and molds count, Coliform, E. coli, Enterococcus ssp., and Salmonella spp. colonization than those in the control group (linear and quadratic, p < 0.0001). However, the dietary supplementation of Che-SeNPs levels increased the lactic acid bacteria count (linear and quadratic, p < 0.05) compared to the control group.4. Discussion Antimicrobial agents are critical in the pharmaceutical and textile industries, water purification, and food packaging. One notable disadvantage of organically synthesized compounds is toxicity in the body; therefore, the trend is to use inorganic nanoparticles such as Che-SeNPs with antimicrobial activity [20]. These Che-SeNPs have an inhibitory effect on many microorganisms. Currently, antimicrobial drugs are becoming less effective for many diseases globally because of the drug resistance capability of microbes. Microorganisms use their biofilm to resist antimicrobial drugs, and the membranes are the primary source of food contamination. Che-SeNPs have been used to control the growth and formation of biofilms of food spoilage bacteria, including B. cereus, Enterococcus faecalis, S. aureus, E. coli O157:H7, S. typhimurium, and S. enterica [21]. The development more effective antibacterial agents is vital for a wide range of applications in various diseases for better public health. However, the emergence of multiple antibiotic-resistant bacteria presents a public health threat. Many developed antimicrobial drugs have limited effective applications due to chemical imbalances, low biocompatibility, and poor long-term antibacterial efficiency. Che-SeNPs conjugated with quercetin and acetylcholine have shown a tremendous antimicrobial effect on the pathogen [22]. Probiotics are microorganisms that can improve intestinal microbial balance and benefit poultry health after consumption in adequate amounts. Lactobacillus plantarum and L. johnsonii cells are resistant against selenium dioxide, and their cell-free extracts were tested against C. albicans ATCC 14053 [7]. Selenium particles extracted from cultures of S. carnosus stabilized by their natural protein coating, for instance, show considerable activity against the nematode Steinernema feltiae, Saccharomyces cerevisiae, and E. coli. Natural SeNPs were found to be more active than mechanically generated selenium particles and can be applied as antimicrobial materials in medicine and agriculture [23]. Antimicrobial tests show SeNPs activity against S. epidermidis, but not against E. coli in a low Se concentration of 2 ppm. S. aureus is an important bacterium commonly found in numerous infections. S. aureus infections were difficult to treat due to their biofilm formation and defined antibiotic resistance. SeNPs were used effectively in the prevention and treatment of disease caused by S. aureus [8].The antifungal activity of SeNPs was evaluated against C. albicans ATCC 4862, C. glabrata ATCC64677, C. parapsilosis ATCC 22019, and C. guilliermondii ATCC 6260 using the disk diffusion method [14] (Table 2). The common antifungal agents are enormously irritant and lethal, and it is necessary to formulate newer types of safe and cost-effective fungicidals. Accordingly, the present study illustrates that selenium nanoparticles have good antifungal activity against all pathogenic animals and human Candida species. Selenium nanoparticles showed better activity against C. albicans ATCC 4862 compared to other Candida species used in this study. In addition, it was proved that SeNPs ranging in size from 100 to 550 nm, with an average size of 245 nm, have low toxicity and high biological activities [24]. A similar observation was reported by Shakibaie et al. [7], who studied the antifungal activity of selenium nanoparticles against Aspergillus fumigatus and C. albicans, and found that the MICs for A. fumigatus and C. albicans were 100 and 70 μg/mL, respectively. However, the high surface-to-volume ratios and their nanoscale sizes provide better activity against biological materials. In addition, Che-SeNPs have significantly lower toxicity than other inorganic and organic forms of supplemental selenium [7].The current data demonstrated that dietary supplementation with Che-SeNPs substantially affected BW, BWG, feed intake, and feed conversion ratio (FCR). A similar observation was stated by Zhou and Wang [25], who clarified a significant improvement in the FCR and growth performance by supplementation with Che-SeNP up to a 0.5-mg/kg basal diet. Khazraie and Ghazanfarpoor [26] illustrated that weight gain was significantly increased in quail chicks fed the Che-SeNPs-supplemented diet compared to the control. Selim et al. [27], using the Che-SeNPs form (0.15 to 0.30 ppm), showed a marked improvement in BW, BWG, and FCR of broiler chicks. Ibrahim et al. [28] indicated that dietary Che-SeNPs supplementation significantly improved BW, BWG, and FCR of broiler chicks compared to the control group. The improved performance may be due to (1) higher utilization of Che-SeNPs associated with the unique properties of the nano form, such as excellent bioavailability, higher solubility, high cellular uptake, and greater surface activity [2]; (2) the involvement of Se in regulating several enzymatic systems, which interfere in energy metabolism and metabolism of the essential fatty acid apurinic and apyrimidinic base; and (3) Che-SeNPs having high biological activity, immune regulation, and oxidation resistance [22]. In addition, the improved FCR can be elucidated by the Che-SeNPs role in enhancing the activity of intestinal microbiota to digest and absorb the nutrients via the intestinal barriers [9].The results of the present study in carcass traits and relative organ weight of growing Japanese quails were in line with the study of Khazraie and Ghazanfarpoor [26], who stated that the supplementation of Che-SeNPs to the diet did not affect carcass traits of chicks. Additionally, Cai et al. [6] reported no significant effect of Che-SeNPs on the weights of carcass parts in broilers. Selim et al. [27] indicated that giblets were not affected due to the inclusion of Che-SeNPs in the diet. Recently, Bakhshalinejad et al. [29] reported that neither carcass yield nor carcass yield parts such as thigh and breast muscles and liver, gizzard, and heart of broilers were affected by different Che-SeNPs levels at 42 d of age. In the present study, the relative liver weight was significantly increased with Che-SeNPs; this increase (21–28% relative to control) may be due to the increase in live body weight in Che-SeNPs groups. However, the widespread use of Nano Se in medication and nanoelectronics has increased the risk of their environmental contaminations, which might affect animal species and humans, although it is useful to understand the assessment of the toxicity of Se-NPs to the biological ecosystem. It should be mentioned that the increase in WBCs was insignificant in the Che-SeNPs supplemented-groups; these change along with the change in liver percentage, even if not significant, warrant further investigation to confirm the safety of Che-SeNPs in animal and human nutrition.Boostani et al. [30] exhibited that packed cell volume, RBCs and WBCs were not different between the birds supplemented with Che-SeNPs and the control birds, which is in line with the current results. Likewise, Chen et al. [31] revealed no significant difference in WBCs, RBCs, and packed cell volume of broilers fed different Se sources. Additionally, Mohamed et al. [32] illustrated that using Che-SeNPs in the diet of Sinai chicks did not significantly affect WBCs, eosinophils, and monocytes. However, our study indicated that Hb level was increased by adding Che-SeNPs, in agreement with Khazraie and Ghazanfarpoor [26], who reported a significant increase in Hb concentration in quails fed a diet containing Che-SeNPs. These findings may be caused by Se enhancing the activity of hemopoietic organs [33]. Se protects the neutrophils, RBCs, WBCs, and other blood components against peroxidative damage [34]. Deficiency of Se can increase ROS in body tissues, the significant adverse impacts on the consistency of immunity cells’ performance and biological membranes [35].The results of the current study on the blood biochemistry of quails were in agreement with previous studies. Serum total protein and albumin were not significantly affected due to Che-SeNPs supplementation to the broiler diet [27]. However, serum globulin levels were increased with the addition of Che-SeNPs in the diet [36]. Additionally, no significant difference in serum AST activity was observed of chicks fed a diet supplemented with Che-SeNPs [27]. However, our results are similar to the study of Elsaid [37], who reported increased serum ALT activity in birds fed a diet supplemented with Che-SeNPs. Selim et al. [27] found that increasing the Che-SeNPs level in broiler diets increased plasma creatinine levels compared to the control group. However, some studies showed that blood creatinine levels declined in birds fed a diet containing Che-SeNPs Elsaid [37]. The potential reason for these differences is possibly related to the dose and time of animal exposure. We conclude from the current study that the higher Che-SeNPs level is the cause of increased ALT and creatinine as indicators of liver and kidney oxidative damage, whereas lower levels showed less damage.Selenium has a hypocholesterolemic activity. A significant reduction in plasma TC and an increase in HDL were detected in the Che-SeNPs-treated birds. The dietary addition of nano forms of selenium for hens caused substantial declines in serum levels of cholesterol as compared to that of the control [38]. Rizk [39] stated that Che-SeNPs addition in the chicken diet decreased cholesterol, triglycerides, and low-density lipoproteins and increased HDL compared with the control group. These results might be attributed to the lipolysis that increased with Se intake. Additionally, the reduction of cholesterol may be due to the role of Se in the activation of peroxisome proliferator-activated receptor-γ that can decrease sterol regulatory element-binding protein-2 level, resulting in decreased cholesterol synthesis [40].The nutritional status of an animal greatly influences the antioxidant system. Se nanoparticles have vital roles in protecting the body cells from reactive oxygen species abundance by decreasing the production of free radicals and lipid peroxidation [41]. Se is well-known for its ability to boost the antioxidant capacity as it forms selenocysteine, a portion of the active center of GSH-peroxidase (Px) [42]. Therefore, a dietary supplementation of Se is essential to improve Se-dependent antioxidant enzymes. These enzymes can help in decreasing the concentration of lipid peroxides and hydrogen peroxide. Dietary Che-SeNPs enhanced oxidative stability and antioxidant ability in broilers [6]. Mohamed et al. [32] reported a positive effect on birds’ plasma total antioxidant capacity when fed a diet containing Che-SeNPs. Aparna and Karunakaran [43] detected an increase in glutathione peroxidase and SOD cellular activity in birds fed Che-SeNPs compared to the control group. El-Deep et al. [4] displayed that Che-SeNPs enhanced the activities of SOD and GSH-Px and reduced MDA content in the liver of broilers. The improvement of antioxidant status in quails fed Che-SeNPs in the current study may be attributed to the fact that (1) Che-SeNPs had high antioxidant activity, because it has an augmented ability to trap free radicals with better antioxidant influence, (2) Che-SeNPs can act as a chemopreventive agent when administered at a smaller particle size, (3) Se plays a vital role as an antioxidant that could protect intestinal mucosa against pathogens and oxidative damage, and (4) Se has immunomodulation properties [44].Nanominerals such as Che-SeNPs can increase immune parameters and disease resistance [4]. In the current study, we presented a potential approach to the application of Che-SeNPs to improve the immunity of quails. These findings can be due to the higher absorption of selenium nanoparticles. The present data are in harmony with the study of Cai et al. [6], who stated that dietary Che-SeNPs supplementation improved humoral immunity by increasing the levels of IgG and IgM of broiler chicks. Dietary Che-SeNPs supplementation showed immunostimulatory impacts in broiler chicks [45]. The improvement in serum immunoglobulins levels may be attributed to the essential biological role of Che-SeNPs in increasing T helper cells and enhancing the secretion of cytokines [46].Additionally, Se plays a crucial role in the production of GSH-Px. Selenium inhibits arachidonic acid peroxidation and protects cells and tissues of the immune system from damage caused by free radicals. Therefore, it can be stated that Che-SeNPs boosts birds’ immunity and antioxidant metabolites [45]. Studies have shown that the use of nanominerals in poultry production and its effect on performance and immunity, and reproduction is promising [47,48]. It has been suggested that the application of Se can help to strengthen immunity and decrease inflammation [49,50]. Se, according to Rooke et al. [51], may be involved in a variety of immune functions at the cellular and molecular levels, including lowering immunosuppressive markers such as glucocorticoids; reducing the duration and rate of intramammary infections; and regulating the function of lymphocytes, neutrophils, and natural killer cells. Our results suggest that feeding a diet enriched with Che-SeNPs might have immunostimulatory impacts on quails.The regulation of microbiota in the gut can be achieved through dietary supplements that can encourage the growth of beneficial bacteria or selectively suppress pathogenic bacteria. Trace elements and natural agents as feed additives may affect the diversity of gut microbiota [8]. The present study found that supplementation of Che-SeNPs in quail diets declined harmful bacteria and increased beneficial bacteria. Se is one of the critical elements that can help microbiota complete its action within the gut [9]. Furthermore, Se supplementation augmented the population of caecum such as Bifidobacterium spp. and Lactobacillus spp. compared to the basal diet [9]. Therefore, using Che-SeNPs is one of the recommendations for reducing the population of harmful gut bacteria due to its inhibitory effect against many pathogenic bacteria.Nanotechnology has been found to have advantageous uses in the food chain of humans, mainly through enhancing the bioavailability and delivering enough levels of vital nutrients, vitamins, and minerals in animal products used by humans [10,52,53,54,55,56]. Moreover, the consumers’ demand for foods and their knowledge has been enhanced as consumers want safe and high-quality foods with high sensory quality, favorable health qualities, and prolonged shelf life [57]. Several studies proved the possibility of supplementing nanomaterials to improve mineral contents in animal products; nevertheless, most of these studies were carried out on chicken, meat, and eggs [58,59]. Therefore, more research is needed to analyze the ability of nanomaterials to affect the quality and nutritional content of meat and egg. In addition, the influence of nanomaterials on the environment and health needs further examination [60,61]. Thus, the application of nanoparticles in the poultry industry must be further investigated before they can be applied.5. ConclusionsThe current study’s findings demonstrated that dietary supplementation with Che-SeNPs could improve the performance of growing quails. The highest values of growth performance were recorded in the group fed 0.4 g Che-SeNPs g/kg feed during the fattening periods (1–5 wk of age). Moreover, the dietary addition of Che-SeNPs improved the lipid profile, antioxidant indices, and immunity and decreased the intestinal pathogens of growing quails. The groups fed diets supplemented with Che-SeNPs showed lower total yeast and mold count, Coliform, Escherichia coli, Enterococcus spp., and Salmonella spp. colonization, and higher lactic acid bacteria counts than those in the control group. However, further studies are warranted to understand the effect of nanominerals and their mechanisms of action, sites of absorption, and transcript expression analysis of distribution.

animals : an open access journal from mdpi

10.3390/ani13061093

PMC10044250

Dietary supplementation with fatty acids benefits the high productivity of sows and plays an essential role in piglet growth. Considering that the mechanisms of the fatty acid types on animal physiology differ, combined supplementation may have additive effects. Therefore, we investigated the effects of different combinations of fatty acids with different chain lengths on the reproductive performances of sows and on the antioxidant capacity, immune function, and intestinal health of their offspring. Taken together, the dietary supplementation of sows with different combinations of SB, MCFAs, and omega-3 PUFAs to the sows during late gestation and lactation can efficiently improve the growth performance, immune function, antioxidant capability, and intestinal microbiota and decrease the incidence of diarrhea in the suckling piglets. Additionally, dietary SMP supplementation has better effects on piglet intestinal health and is likely through gut microorganism alterations.

The aim of the study was to investigate the comparative effects of different combinations of sodium butyrate (SB), medium-chain fatty acids (MCFAs), and omega-3 polyunsaturated fatty acids (n-3 PUFAs) on the reproductive performances of sows, as well as on the biochemical parameters, oxidative statuses, and intestinal health of the sucking piglets. A total of 30 sows were randomly allocated to five treatments: (1) control diet (CON); (2) CON with 1 g/kg of coated SB and 7.75 g/kg of coated MCFAs (SM); (3) CON with 1 g/kg of coated SB and 68.2 g/kg of coated n-3 PUFAs (SP); (4) CON with 7.75 g/kg of coated MCFAs and 68.2 g/kg of coated n-3 PUFAs (MP); (5) CON with 1 g/kg of coated SB, 7.75 g/kg of coated MCFAs and 68.2 g/kg of coated n-3 PUFA (SMP). The results showed that sows fed the SP, MP, and SMP diets had shorter weaning-to-estrus intervals than those fed the CON diet (p < 0.01). The piglets in the SM, SP, and MP groups showed higher increases in the plasma catalase and glutathione peroxidase activities than those of the CON group (p < 0.01). The diarrhea incidence of piglets in the SM, SP and SMP groups was lower than that of piglets in the CON group (p < 0.01). Additionally, the addition of SM, SP, MP, and SMP to the sow diets increased the contents of immunoglobulin A, immunoglobulin G, fat, and proteins in the colostrum (p < 0.01), as well as the plasma total superoxide dismutase activities (p < 0.01) in the suckling piglets, whereas it decreased the mRNA expressions of tumor necrosis factor-α, interleukin-1β, and toll-like receptor 4 in the jejunum mucosa of the piglets. The relative abundances of Prevotella, Coprococcus, and Blautia in the colonic digesta of the piglets were increased in the SM group (p < 0.05), and the relative abundances of Faecalibacterium increased in the SMP group (p < 0.05), compared with the CON group. The relative abundances of Collinsella, Blautia, and Bulleidia in the MP group were higher than those in the CON group (p < 0.05). Collectively, dietary combinations of fatty acids with different chain lengths have positive effects on the growth performances and intestinal health of suckling piglets.

1. IntroductionThe reproductive performances of high-yielding sows and the growth performances of piglets are the two most important aspects that influence the economic efficiency of the modern pig breeding industry. With the development of breeding techniques, the reproductive performances of sows have been improved. However, highly prolific sows often suffer from a range of issues, including insufficient nutrient intake, excessive weight loss, longer weaning-to-estrus intervals (WEI), shortened service lives, and metabolic disorders, which result in the retarded growth of their piglets [1,2,3,4]. Additionally, in late pregnancy, the fetal growth rate dramatically accelerates [5]. Diets that contain supplemental fatty acids have been particularly effective at improving the body conditions of sows and the birth weights and growth of suckling piglets [6,7,8]. Therefore, meeting the nutrient requirements of prolific sows is a vital consideration for the development of pig farming.Fatty acids have biological functions, such as the regulation of metabolic disorders, the intestinal barrier, and the immune function of animals [7,9,10]. Fatty acids are categorized according to the length of the carbon chain and the degree of saturation. Short-chain fatty acids (SCFAs), and especially butyrate (SB), which is a major source of energy for colonic epithelial cells [9], can improve the immune function of piglets through the nuclear factor-kappa B (NF-κB) signal pathway [11,12]. In addition, appropriate doses of butyrate can alleviate diarrhea symptoms and reduce the intestinal permeability to maintain intestinal health [13]. The majority of MCFAs are carried straight to the liver via the portal vein, granting a rapid energy supply, which is vital for piglets [7]. In addition, MCFAs decrease intestinal colonization by opportunistic pathogens and modulate the colonic microbiota of piglets [14,15]. n-3 PUFAs are essential for embryonic and fetal development [16]. In addition, n-3 PUFAs also play an important role in shortening the weaning-to-estrus interval (WEI) and enhancing the immune function in sows [16,17,18]. Previous studies have shown that dietary single fatty acid supplementation could shorten the WEIs of sows and improve the intestinal health and growth performances of suckling piglets [19,20]. However, studies on the effects of dietary combinations of SB, MCFAs, and n-3 PUFA on prolific sows during late gestation and lactation are still lacking. Therefore, the purpose of this study was to investigate the effects of different combinations of SB, MCFAs, and n-3 PUFAs on the reproductive performances of sows and on the biochemical parameters, oxidative status, and intestinal health of their offspring during late gestation and lactation. We speculated that the combination of fatty acids would have additive effects on the reproductive performances of the sows and the growth performances of their offspring.2. Materials and MethodsAll animal operations were carried out in compliance with protocols approved by the Animal Ethics Committee of Southwestern University (Chongqing, China). The behavior and health of the experimental animals were continuously monitored during the trial period, and no negative impacts were observed. More precisely, there were no clinical problems that could have necessitated pharmaceutical treatment for pathologies after the investigation began, and all the animals were deemed suitable for the study. The experimental animals were disposed of safely following the Experimental Animal Handling Procedure of Southwest University (Ethics Approval Code: IACUC-20210120-03).2.1. Animals, Materials, and Feeding ManagementA total of 30 third-parity sows (Landrace × Large White hybrid; 200 ± 15 kg) were used in this study. From mating to d 109 of gestation, the sows were kept in individual stainless-steel cages (0.60 × 2.15 m) in the gestation house, and on approximately d 110 of gestation, they were transferred to the farrowing stalls (1.20 × 2.15 m) in a thoroughly sterilized farrowing house with iron fencing and plastic flooring. The SB, MCFAs, and omega-3 PUFAs in this study were supplied by Xingao Agribusiness Development Co., Ltd. (Xiamen, Fujian, China), with purities of 98%, 70%, and 20%, respectively. The primary active constituents of the n-3 PUFAs were docosahexaenoic acid (DHA), α-linolenic acid (ALA), and eicosapentaenoic acid (EPA). Additionally, the sows were fed twice a day at 08:00 and 16:00, with 2.5–3.0 kg of feed per day, which was restricted in late gestation based on the body condition, while 2 kg was fed on d 1–2 of lactation, with an increase of 0.5 kg per day from d 3 to d 7 of lactation, and an increase of 0.8 kg per day from d 8 to d 14 of lactation, with no further increase thereafter. All sows were allowed to consume water at any time during the study. Heaters and exhaust fans kept the room at a comfortable temperature (from 20 to 25 °C).2.2. Diets and Experimental DesignThe sows were randomly allocated to five treatments (six replicate pens per treatment and one sow per replicate) in a completely randomized experimental design. The sows were fed a basal diet (control, CON), a basal diet supplemented with 1 g/kg of coated SB and 7.75 g/kg of coated MCFAs (SM), a basal diet supplemented with 1 g/kg of coated SB and 68.2 g/kg of coated n-3 PUFAs (SP), a basal diet supplemented with 7.75 g/kg of coated MCFAs and 68.2 g/kg of coated n-3 PUFAs (MP), and a basal diet supplemented with 1 g/kg of coated SB, 7.75 g/kg of coated MCFAs, and 68.2 g/kg of coated n-3 PUFAs (SMP). The dosages were chosen based on the company’s recommended dosages. The piglets were housed in farrowing stalls, with one litter per pen. After the piglets were born, they were manually attached to the nipples to guarantee that they received breast milk for growth and development. This study shared the control group with Chen et al. [20], and the sows in this trial and Chen’s sows were kept in the same barn. The trial started on d 85 of gestation and ended with the weaning of the piglets (d 21 of lactation). From d 85 to d 110 of gestation, the test sows were fed given the gestation diet, followed by the lactation diet from d 110 of gestation and throughout weaning. The nutritional content of the baseline diet met or surpassed the nutritional recommendations of the National Research Council (2012) [21]. The dietary ingredients and nutritional levels for the sows throughout gestation and lactation are shown in Table 1. The gestation diet (approximately 100 g) and lactation diet (approximately 100 g) were collected. Then, the feed samples were analyzed for the crude protein (CP), crude ash (Ash), dry matter (DM), ether extract (EE), calcium (Ca), crude fiber (CF), available phosphorus (AP), and total phosphorus (Total P), according to the procedures followed by the standard of the AOAC (2000) [22].2.3. Recording and Sample Collection2.3.1. Reproductive Performances of SowsDuring the animal experiment, the feed wastage was recorded every day after the meal to calculate the average daily feed intake (ADFI). The individual neonatal weight of the born alive was weighed within 12 h of delivery. The stillborn and mummified fetuses were not weighed, and the mummified fetuses were counted as stillbirths (piglets that died before birth). We measured the numbers of born alive, stillborn, mummified fetuses, and total born, the litter weights, birth weights, and the weaning weights of the piglets, and the WEIs of the sows. The total litter size included mummified fetuses, born alive, and stillborn. The born-alive rate was calculated as the number of born alive/total born × 100%.2.3.2. Growth Performances of Piglets and Diarrhea IncidencePiglets were cross-fostered after altering the litter sizes of the sows within the same treatment within 24 h of farrowing. Piglets were weaned on d 21 of lactation, the number of weaned piglets was counted, and the piglets were weighed to determine the ADGs and weaning survival rates. The diarrhea severity was determined by daily observation of the piglet feces, as previously described [23,24]. The formula for calculating the diarrhea incidence was as follows: the diarrhea rate (%) = Σ[(diarrhea days in piglets × number of diarrhea piglets)]/(total number of piglets × 21) × 100%.2.3.3. Sample CollectionThe colostrum (about 40 mL) was manually collected after the alcohol sterilization of the sow teats within 2 h of the first piglet’s birth. For each repetition, piglets that met the average weight were randomly selected for blood and slaughter sampling. On d 22 of lactation, one piglet per pen was randomly selected for the collection of blood samples from the anterior vein. Then, the blood samples were centrifuged (4 °C, 3000× g, 15 min) and the plasma was stored at −80 °C for subsequent analysis. Next, the piglets were humanely killed after anesthesia by intravenous injection with sodium pentobarbital (50 mg/kg BW). Tissue samples from the liver and middle jejunum were taken, flushed with 0.9% saline, and deposited in a 4% formaldehyde solution for morphological examination. The colonic digesta (approximately 10 g) was collected in sterile tubes for the microbiota analysis. After that, jejunal mucosa samples were carefully scraped off using a sterile glass slide, flash-frozen in liquid nitrogen and were maintained at −80 °C for subsequent analysis.2.4. Analytical Methods2.4.1. Colostrum Composition AnalysisOne colostrum sample (approximately 20 mL) diluted 3 times with purified water was determined in triplicate for milk fat, milk protein, lactose and solids-not-fat (SNF) using a FOSS Multifunctional Dairy Analyzer (MilkoScan TM FT120, Foss Electric A/S, Hillerød, Denmark). The other colostrum sample (approximately 20 mL) was centrifuged (4 °C, 3000× g, 20 min), and the supernatant was aspirated and kept at −80 °C. The colostrum supernatant was thawed at room temperature and utilized to determine the concentrations of immunoglobulin A (IgA), immunoglobulin G (IgG), and immunoglobulin M (IgM) using swine reagent ELISA kits provided by the Nanjing Jiancheng Bioengineering Institute (Nanjing, Jiangsu, China). The intra- and inter-assay CVs for these ELISA kits were both less than 9.0%.2.4.2. Blood Biochemical ParametersThe total antioxidant capacity (T-AOC); (Code: A015-2-1), glutathione peroxidase (GSH-Px); (Code: A005-1-1), malondialdehyde (MDA); (Code: A003-1-2), total superoxide dismutase (T-SOD); (Code: A001-1-1) and catalase (CAT); (Code: A007-1-1) in the plasma of the pigs were analyzed by using commercial assay kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, Jiangsu, China). The concentrations of total protein (TP); (Code: A045-1-1) and albumin (ALB); (Code: A028-2-1) in the plasma of the pigs were determined using a por-cine-specific commercial kit with microplate test methods and an enzyme-labeled instrument (Thermo Electron Corporation; Rochester, NY, USA). The plasma concentrations of high-density lipoprotein cholesterol (HDL-C); (Code: A112-1-1), glucose (GLU); (Code: F006-1-1), total cholesterol (TC); (Code: A111-1-1), urinary nitrogen, and triglycerides (TG); (Code: A110-1-1) were determined using colorimetric method diagnostic kits. In accordance with the manufacturer guidelines, all protocols were strictly carried out, ensuring the highest level of safety and accuracy.2.4.3. Intestinal MorphologyJejunum tissues were collected from weaning piglets and fixed in 4% formalin for the analysis of the intestinal morphology. In short, the villus height (VH) and crypt depth (CD) of the jejunum were measured using an Axio Scope A1 microscope (Zeiss, Oberkochen, Germany) with 40× combined magnification. The VH was calculated by measuring the distance between the top of the villus and the villus-crypt junction, and the CD was calculated by measuring the distance between the villus-crypt junction and the bottom of the crypt. The averages of the measurements (at least 10 normative measurements) were used for the statistical analysis, and all intestinal mucosal morphometric analyses were executed by the same operator.2.4.4. Quantitative Real-Time PCRThe relative expressions of the genes related to the factors involved in the regulation of inflammation in the jejunal mucosa of the piglets were determined. Total RNA was isolated from frozen jejunal mucosa samples by using SteadyPure Uni-versal RNA Extraction Kits II (Code: AG21022; Accurate Biotechnology (Hunan) Co., Ltd., Changsha, China). The specific RNA extraction procedure was performed using the manufacturer’s recommendations. The concentration of total RNA was measured with a NanoDrop-ND2000 spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA), and the reverse transcription was performed with the qualified RNA samples using AMV First Strand cDNA Synthesis Kits provided by Sangon (Shanghai, China), according to the manufacturer’s instructions.Real-time PCR analysis was performed to quantify the claudin-1 (CLDN 1), interleukin-6 (IL-6), tumor necrosis factor-α (TNF-α), occludin (OCLN), interleukin-1β (IL-1β), zonula occludens-1 (ZO-1), interleukin-10 (IL-10), toll-like receptor 4 (TLR-4), NF-κB, myeloid differentiation factor 88 (MγD88), and glyceraldehyde-3-phosphate (GAPDH) mRNA levels in the jejunal mucosa. The primer sequences for all the target genes and predicted product sizes are shown in Table 2. The real-time PCR analysis was conducted using the SYBR Green approach combined with an ABI 7900 Sequence Detection System. The following thermal cycling parameters were used: initial denaturation at 94 °C for 30 s, followed by 40 cycles at 94 °C for 5 s, annealing temperature for 20 s, and extension at 72 °C for 20 s. Moreover, the melting curve analysis was used to ensure that the PCR products remained specific and pure. A standard curve was generated using LightCycler software and the amplification of serially diluted cDNA, and the quantification of the target gene expression was calculated using the 2−ΔΔCT method based on the standard curve with the GAPDH gene as the reference gene [25].2.4.5. 16S rRNA Gene Sequencing and Microbiota AnalysisTotal DNA from the digesta of the colon was extracted using Power Fecal DNA Isolation Kits (Mobio, Carlsbad, CA, USA) following the manufacturer’s instructions. Briefly, the 16S rDNA gene was presented in the genome of all bacteria and was highly conserved. Microbial profiling was performed on an Illumina HiSeq2500 platform (Novogene, Beijing, China) by the PCR amplification of a segment of a highly variable region sequence (V3 region) following 600 amplification cycles. Then, the raw data sequences for the 16S rRNA gene were collected and filtered with the software tools FLASH and QIIME. UPARSE was used to assess the valid sequences and establish the practical classification units (OTUs). Moreover, singletons and OTUs below 0.005% were eliminated. Subsequently, the species-level classification was determined by the taxonomic alignment of high-quality sequences with the National Center for Biotechnology Information (NCBI) nucleotide database (ver. 2.20) at a 90% confidence threshold. The alpha diversity index (Shannon, ACE, Chao1, and Simpson) and β-diversity (Bray Curtis) analyses were calculated with the QIIME software tool.2.5. Statistical AnalysisAccording to the post hoc power analysis for the ADFI of the sows, ADGs of the piglets, and IgA in the colostrum and plasma in this study, the calculated statistical power was > 0.90; thus, 6 pigs per treatment were enough to provide sufficient statistical power (α < 0.05; β = 0.80). All data analyses were performed using ANOVA analysis, with the dietary treatments used as the fixed factor. Diarrhea rate data assessments were translated using the arcsine square root transformation for subsequent statistics. Data were subjected to Duncan’s test method using SPSS 19.0 software (SPSS Inc., Chicago, IL, USA), and they are presented as means and standard errors of means (SEMs) unless otherwise noted. The histograms were created using GraphPad Prism 8. (GraphPad Prism Inc., San Diego, CA, USA). Statistical significance was identified when p < 0.05, and trends were considered when 0.05 < p ≤ 0.10.3. Results3.1. Reproductive Performances of SowsAs shown in Table 3, there were no significant changes in the number of born alive, stillborn and total born among the dietary treatments (p > 0.05). Compared with the CON group, the ADFI of the sows was significantly increased in the SM and SMP groups (p < 0.01). Moreover, the sows fed the SP, MP, and SMP diets showed significantly shorter WEIs compared with those fed the CON diet (p < 0.01).3.2. Growth Performances of PigletsAs shown in Table 3, there were no differences in the survival rates of the suckling piglets among the dietary treatments (p > 0.05). The final BWs of the piglets in the SMP group were significantly higher than those of the piglets in the CON, SM, and MP groups (p < 0.01). The ADGs of the piglets in the SMP group were higher than those of the suckling piglets in the CON and MP groups (p < 0.01). Of note, the suckling piglets in the SMP group showed higher final BWs and ADGs than the other groups. Moreover, the diarrhea incidence of the suckling piglets in the CON group was higher than those of the piglets in the SM, SP, and SMP groups (p < 0.01).3.3. Colostrum Composition of SowsAs shown in Table 4, compared with the sows fed the control diet, the dietary addition of SM, SP, MP, and SMP increased the concentrations of fat and protein in the colostrum (p < 0.01). The SNF concentrations in the colostrum of the sows in the SM and MP groups were higher than those of the sows in the CON, SP, and SMP groups (p < 0.01). In addition, the concentrations of IgA, IgG, and IgM in the colostrum of the SM, SP, and SMP groups were higher than those of the sows in the CON group (p < 0.01).3.4. Plasma Biochemical Index of Suckling PigletsAs shown in Table 5, the plasma TP, FFA, and HDL contents of the suckling piglets were significantly increased in the SM, SP, MP, and SMP groups compared with those of the piglets in the CON group (p < 0.01). In addition, the TG and TC contents were decreased in the plasma of the suckling piglets in the SM, MP, and SMP groups (p < 0.01) compared with those in the plasma of the piglets in the CON group. For the immunoglobulin levels, the piglets in the SM group showed higher IgA concentrations in their plasma than the other groups (p < 0.001). Compared with the CON group, the SM, SP, MP, and SMP groups showed significantly increased IgG concentrations in the plasma of the piglets (p < 0.01).3.5. Plasma Antioxidant Capacity of Suckling PigletsAs shown in Table 6, the piglets in the SM, SP, MP, and SMP groups showed significantly increased plasma T-SOD and T-AOC activities in comparison with those in the CON group (p < 0.01). The piglets in the SM, SP, and MP groups had higher plasma CAT and GSH-Px activities than those in the CON group (p < 0.01), and the piglets in the MP group had the highest plasma GSH-Px activity among the five groups. However, the content of plasma MDA was higher in the MP group than in the CON group (p < 0.01), with no significant difference in the SM, SP, and SMP groups (p > 0.05).3.6. Intestinal Morphology of Sucking PigletsAs shown in Figure 1B, there was no significant difference in the VH of the jejunum among the five groups (p > 0.05). The CD of the jejunum in the SM, SP, MP, and SMP groups was significantly lower than that of the jejunum in the CON group (p < 0.01, Figure 1C). The VH/CD ratio of the jejunum in the SM, SP, MP, and SMP groups was significantly increased compared with that of the CON group (p < 0.01, Figure 1D). Moreover, the piglets in the SMP group showed a higher VH/CD ratio in the jejunum mucosa than those in the other groups (p < 0.01).3.7. mRNA Expressions of Intestinal Tight Junction Protein and Inflammatory Cytokines of Suckling PigletsCompared with the CON group, the piglets in the SM, SP, MP, and SMP groups showed significantly upregulated mRNA expressions of CLDN-1 and ZO-1 in the jejunal mucosa (p < 0.01, Figure 2A), and the piglets in the SMP group had significantly upregulated mRNA expressions of OCLN (p < 0.01). There was no significant difference in the mRNA expressions of IL-6 among the five groups (p > 0.05, Figure 2B). However, the mRNA expressions of TNF-α, IL-1β, and TLR4 of the jejunum were significantly downregulated in the SM, SP, MP, and SMP groups compared with those in the CON group (p < 0.01, Figure 2B,C), and the mRNA expressions of NF-κB in the jejunal mucosa were downregulated in the SM, SP, and SMP groups compared with those in the CON group (p < 0.01, Figure 2C).3.8. Intestinal Microbial Flora in Colonic DigestaThe microbial flora in the colonic digesta was analyzed. As shown in Figure 3, a total of 222 OTUs were shared among the five treatment groups (Figure 3A). The piglets in the CON, SM, SP, MP, and SMP groups had 167, 317, 189, 164, and 205 unique OTUs, respectively (Figure 3A). However, there was no significant difference in the α-diversity (Shannon, Chao1, ACE, and Simpson indexes) of the colonic digesta in the suckling piglets among the five groups (p > 0.05, Figure 3B). The SM, SP, and MP groups resulted in significant changes in the beta diversity of the colonic microbiota, as shown by the NMDS based on the UniFrac distances (Figure 3C). Dietary supplementation with SMP was associated with increased relative abundances of Faecalibacterium at the genus level (Figure 3F, LDA score >2). Additionally, the piglets in the MP group were associated with increased relative abundances of Collinsella, while the piglets in the SM group were associated with increased relative abundances of Catenbacterium, Coprococcus, and Bulleidia at the genus level (Figure 3F, LDA score >4).At the phylum level, the relative abundances of Bacteroidetes were increased in the SMP group (p < 0.05, Figure 4A), while the relative abundances of Firmicute in the SM, MP, and SMP groups were significantly lower than those of the CON group (p < 0.05, Figure 3B). At the genus level, the relative abundances of Prevotella, Coprococcus, and Blautia were increased in the SM group (p < 0.05, Figure 4C,F,H), and the relative abundances of Blautia and Bulleidia in the MP group were higher than those in the CON group (p < 0.05, Figure 4F,H). Compared with the piglets in the CON group, the piglets in the SMP group had increased relative abundances of Faecalibacterium (p < 0.05, Figure 4E).4. DiscussionThe ADFI of the sows and the ADGs of the piglets were considered the principal limitation factors for the growth performances of the neonatal pigs. In the present study, the dietary addition of SMP led to the highest ADG (236 g in the SMP group vs. 186 g in the CON group) and final BWs of the piglets (6.48 kg in the SMP group vs. 5.36 kg in the CON group), which agree with Jin et al. [26], who reported that the addition of fish oil could produce positive effects in the ADGs of piglets. Gebhardt et al. [27] found similar results: MCFAs improved the growth performance of nursery piglets by increasing the ADG, ADFI, and feed conversion ratio in a linear dose-dependent manner. Conversely, in some studies, researchers found no effect on the ADGs of piglets using 0.2% or even lower fatty acid products [27,28], which indicates that the ADGs were affected by the dietary fatty acid contents. It has been documented that SB has a distinct cheese flavor, which regulates appetite and food intake [10,29]. Our results demonstrated that the dietary addition of SM and SMP significantly increased the ADFI of the sows, whereas the dietary addition of SP and MP obtained the reverse results. Hanczakowska et al. [30] found similar results: a mixture of SCFAs and MCFAs produced better results on the ADFI. Additionally, Smit et al. [31] indicate that n-3 PUFAs have a higher energy density and boost overall energy intake, which allows sows to eat less. Thus, a mixture of SB and MCFAs might have potential additive effects to mitigate the negative effect of n-3 PUFAs on the ADFI.In addition, diarrhea is the critical factor that causes the retardation of growth performances and increased mortality of piglets [32]. Feng et al. [13] and Lerner et al. [33] indicated that SB and MCFAs have the potential to replace antibiotics to control pathogenic bacteria, while n-3 PUFAs have a positive immunomodulatory effect in the gut [34,35]. In this study, the diarrhea incidence of the suckling piglets in the SM, SP, and SMP groups was lower than that in the CON group. Similar results were reported by Chen et al. [20]. The reason might be that SB can provide conditions for the growth of beneficial intestinal bacteria to reduce diarrhea incidence [19]. Intriguingly, Li et al. [36] indicate that organic acid combined with MCFAs showed a better reduction in the diarrhea incidence and growth-promoting effects that were comparable to those of antibiotics. It is noteworthy that supplementing the diets of the sows with SMP had the obvious and unexpected effect of decreasing the diarrhea incidence of the nursing piglets, which indicates that a blend of SB, MCFAS, and n-3 PUFAs may have a synergistic effect, although the mechanism still needs further study.The composition and intake of the colostrum are crucial factors that affect the early weight gain, immune function, and survival of neonatal pigs with limited energy reserves [37]. Previous studies have shown a positive association between the content of fat in the milk and the piglet BW, and maternal fat supplementation could improve the piglet weaning weight [38,39]. Consistently, dietary supplementation with fish oil has increased the concentrations of fat in the colostrum of sows [40]. In this study, compared with the CON diet, sows fed the addition of SM, SP, MP, and SMP diets showed increased concentrations of fat and protein in the colostrum, which is in agreement with previous studies. In addition, the immunoglobulins in the colostrum are the only source of passive immunity for neonatal piglets. The colostral IgA and IgG concentrations are major factors that influence the passive immune protection. Jin et al. [26] indicated that dietary fish oil supplementation increased the IgG and IgM concentrations in the colostrum of suckling pigs, which improved their immune function. Similarly, a diet supplemented with SB increased the concentration of IgA in the colostrum [12]. Our results indicated that the concentrations of IgA and IgG in the colostrum of the SM, SP, and SMP groups were higher than those of the sows in the CON group. Furthermore, similar results were seen in the piglet plasma, which indicated that fatty acids played a positive role in regulating the immune status and providing health benefits. Additionally, the addition of SMP to the sows’ diets increased the IgA and IgM concentrations in the colostrum compared with those of the sows fed the MP diets, with no significant changes observed for the SM, SP, or SMP additions. SB and omega-3 PUFAs exert multiple beneficial effects, including immunomodulatory effects. He et al. [12] also observed that the IgA concentrations in the colostrum increased in SB-treated gilts. A possible explanation is that the blends of SB, MCFAs, and omega-3 PUFAs fed to the sows had additive effects. Collectively, dietary supplementation with fatty acids could enhance the growth performances of piglets by improving the colostrum composition.The plasma biochemical parameters can reflect the nutritional status of the organism, and they can be influenced by changes in internal and external factors. The contents of TP and BUN are used as an index for protein utilization and metabolism. In this study, the piglets in the fatty acid-supplemented groups showed increased contents of plasma TP, which was partially due to the increment in the plasma globulin content. Meanwhile, the addition of fatty acids with different chain lengths to the sow diets increased the contents of BUN in the piglets, which indicated increased nitrogen metabolism. Moreover, previous studies on rats and pigs have shown the beneficial effects of fatty acids on lipid metabolism [41,42]. It is well documented that the TC and TG contents reflect the synthesis and metabolism of lipids in the organism and are associated with diseases linked to dyslipidemia [43]. Allyson et al. [44] reported that HDL cholesterol has a strong transport function that delivers cholesterol from peripheral cells to the liver cells. Similarly, Yu et al. [41] showed that the addition of SCFAs to the diet promotes lipid catabolism. In this study, the piglets in the SM, MP, and SMP groups showed increased HDL-cholesterol levels and lower TG and TC contents, which suggest that fatty acids might reduce cholesterol deposition in the blood.GSH-Px, SOD, and CAT are considered to be important endogenous antioxidant enzymes that scavenge endogenous free radicals produced by the body, maintain the body’s oxidative balance, and play an important role in the oxidative and antioxidative status [45]. In this study, the piglets in the SM, SP, and MP groups exhibited increases in the activities of T-AOC CAT, GSH-Px, and T-SOD in the plasma. Famurewa et al. [46] also found that the dietary addition of coconut oil relieved oxidative stress in a dose-dependent manner by significantly increasing the antioxidant enzyme activities (SOD, CAT, and GSH-Px). Similarly, Nguyen et al. [47] reported the benefits of a diet rich in n-3 PUFAs in stimulating antioxidant enzyme activities to reduce excess ROS production. MDA is considered to be the main product of lipid oxidation, and it is a commonly used indicator of lipid peroxidation [47]. Li et al. [48] showed that the addition of MCFAs to the diet linearly reduced the plasma MDA concentration. Unexpectedly, in this study, the dietary addition of a combination of mixed MCFAs and n-3 PUFAs increased the plasma MDA levels. The discrepancies between studies might be due to the different doses of n-3 PUFAs and the lengths of time that the n-3 PUFAs were supplied. Diets rich in n-3 PUFAs may undergo peroxidation, which leads to free radical-dependent cellular damage, as evidenced by elevated plasma MDA levels [49,50]. These results indicate that the fatty acid alleviation in the oxidative stress statuses of piglets might be related to the improvement in cholesterol metabolism.Tight junction proteins play an important role in intestinal barrier function. It has been reported that OCLN and ZO-1, which are tight junction proteins, are vital in the regulation of intestinal permeability [51]. Feng et al. [13] indicated that SB significantly increased the intestinal ZO-1 and OCLN expressions in vivo and in vitro. A prior study has demonstrated that different fatty acid treatments were beneficial to the intestinal epithelial barrier integrity and intestinal barrier function [20]. In the current study, we observed that the piglets in the SM, SP, MP, and SMP groups had significantly upregulated mRNA expressions of CLDN-1 and ZO-1 in the jejunal mucosa, which contributed to the alleviation of diarrhea in the suckling piglets. The piglets in the SMP group showed higher mRNA expressions of CLDN 1, ZO-1, and OCLN than those in the other groups, which implies that these three fatty acids may have a synergistic effect in strengthening the intestinal barrier function. In addition, the intestinal mucosa morphology is an evaluation of the nutrient digestion and absorption ability, which has a direct impact on the nutrient usage efficiency [23]. It is well-known that SB can reduce some of the negative effects of the intestinal mucosal morphology by providing the preferred energy [10]. Keyser et al. [52] indicated that MCFA supplementation restored the villus height in postweaning piglets with LPS challenges. In addition, n-3 PUFAs may repair the gut damage induced by oxidative stress and enhance the intestinal morphology in piglets [34]. These combined findings indicate that fatty acid supplementation could slightly improve intestinal development by enhancing the barrier integrity and intestinal morphology.Fatty acids play a major role in the inflammatory response of intestinal mucosa [25,53]. Many studies indicate that n-3 PUFAs can alleviate the inflammatory status in animals [17,35]. Carlson et al. indicated that medium-chain triglycerides reduced the mRNA expressions of IL-6 and TNF-α in mice and alleviated the inflammatory response [54]. Similarly, Kuang et al. [55] demonstrated that the addition of mixing MCFAs with SCFAs to the base diet significantly reduced the mRNA expressions of TNF-α and IL-1β. In the present study, the mRNA expressions of the inflammatory factors (TLR4, IL-1β, MγD88, TNF-α, and NF-κB) in the jejunal mucosa were reduced, while the mRNA expression of the anti-inflammatory factor (IL-10) was increased. Researchers have hypothesized that fatty acids of different chain lengths likely attenuate the inflammatory response through the NF-κB and TLR4 signaling pathways [56,57]. Fatty acids inhibit inflammatory factors by regulating the MγD88-dependent route. In addition, the receptor for TLR4 upregulated the MγD88 expression, which can lead to the production and release of inflammatory factors, inducing an immune response in the intestinal mucosa [58,59,60]. Of note, the mRNA expressions of IL-1β, TLR4, and TNF-α in the SM, SP, MP, and SMP groups were lower than those of the CON group, while the mRNA expressions of IL-10 had an opposite result in the SM, SP, and SMP groups. Butyric acid may be related to an inhibitor of a histone deacetylase and result in better anti-inflammatory effects [9].Intestinal microbes play important roles in the host health and performance, and they can profoundly impact the host nutrient metabolism, intestinal development, and immunological functions [61]. Researchers have extensively demonstrated that fatty acids can modulate the abundance and composition of intestinal microbes. Firmicutes, Bacteroidota, Proteobacteria, and Actinobacteriota predominated in the colonic contents of the suckling piglets, which is consistent with previous studies [19,62]. Researchers have reported that Bacteroidetes was able to significantly reduce the diarrhea incidence [63]. In the present study, supplementing the sow diets with SMP significantly increased the relative abundance of Bacteroidetes in the colonic digesta of the suckling piglets, which may partly explain the reduction in the diarrhea incidence. At the genus level, the high abundance of bacteria provides an opportunity to understand how microbiota metabolites affect the host physiology. Prevotella belongs to Bacteroidetes, while Faecalibacterium, Blautia, Bulleidia, and Coprococcus belong to Firmicutes, in which Prevotella and Coprococcus are mostly involved in complex polysaccharide metabolism [64]. Prevotella, Faecalibacterium, Blautia, and Coprococcus can produce high levels of SCFAs, mainly including propionate, butyrate, and acetic acid [64,65,66]. Butyric acid has been identified as a major energy source for colonic epithelial cells [67]. In this study, we found that the piglets in the SM group had higher relative abundances of Coprococcus than those in the CON group, which indicates the potential for the increased intestinal availability of butyrate. Similarly, a reduced diarrhea incidence has been proven to be one of the strategies by which Prevotella improves the intestinal immunity and promotes animal growth [68,69]. Our results indicated that the relative abundances of Prevotella in the SM group were higher than those in the other groups, which partially agrees with Li et al., who observed that the combination of SCFAs and MCFAs increased the relative abundances of Prevotella [36]. We speculated that SB, along with MCFA supplementation, might modulate the gut microbiota composition and benefit the host’s health. Moreover, it has been reported that Faecalibacterium is an anti-inflammatory intestinal commensal microbe that can suppress the TLR4/NF-κB signaling pathway in intestinal epithelial cells. Importantly, we found that the relative abundances of Faecalibacterium were particularly increased in the colonic digesta of the piglets when the sows were fed the SMP diet, and it might play an anti-inflammatory role and could promote intestinal development.5. ConclusionsIn conclusion, our results indicate that diets supplemented with different combinations of SB, MCFAs, and omega-3 PUFAs during late gestation and lactation can efficiently improve the growth performance, immune function, antioxidant capability, and intestinal microbiota, as well as decrease the incidence of diarrhea, in suckling piglets. Additionally, dietary SMP supplementation had better effects on piglet intestinal health and probably through gut microorganism alterations. In the future, attention should be focused on the dosage of fatty acid additives to sow diets during lactation, the synergistic effect of multiple fatty acids, and the mechanism of the interactions between fatty acids and intestinal microorganisms.

animals : an open access journal from mdpi

10.3390/ani12030314

PMC8833449

Territorial behavior is closely correlated with population fluctuations in territorial species, which is influenced by the density of conspecifics. Relevant research in aquacultural species, such as swimming crab (Portunus trituberculatus), is still lacking. In this study, we quantified the territorial behavior of the crabs according to a different number of intruders. This study provides a behavioral perspective for understanding and predicting the population dynamics of marine benthonic animals.

Territorial behavior of animals is affected by numerous factors, one being the number of intruders. The swimming crab (Portunus trituberculatus), an important commercial and ecological species on the continental shelf of Asia, usually needs to defend its territory from intrusion by other crabs, especially in habitats with high densities of conspecifics. To clarify the underlying patterns of how P. trituberculatus protects its territory, we assessed the territorial behavior of occupant crabs (territory holders) when presented with different numbers of intruders using an indoor observation system. We calculated the territory size of the occupants and quantified their behavioral responses to intruders. With an increased number of intruders, the territory size the occupants owned significantly decreased, and their behavior adjusted accordingly. Besides, the territorial behavior score, reflecting the territoriality of crab, decreased significantly. Furthermore, in a high density group that had seven intruders, the occupants showed a higher dominance hierarchy than the intruders, indicating the ascendancy of occupants in territorial competition with intruders. These results revealed that as the number of intruders increased, the territory size of P. trituberculatus shrunk because the fight for territory became more intense.

1. IntroductionTerritory is an essential resource that many animals require to survive, grow, and reproduce [1]. Territorial behavior, including patrolling and exhibiting aggression towards intruders, is an important feature of territorial species [2,3,4]. Individuals with territories have a greater chance of survival because they can more effectively monopolize limited resources including food, shelter, and mates within their territory [5]. For example, crayfish (Orconectes propinquus) occupying a territory may acquire increased access to shelter and food resources [6]. Nevertheless, as the territory expands, there will be more intrusions from others, especially from the conspecifics due to the high resource use overlapping within a similar size or age class [7]. Frequent intrusions usually cause the occupants to incur higher costs, in terms of time, energy, injury risk, and even death, in expelling intruders and maintaining a dominant position [8,9,10]. Accordingly, the number of intruders is a vital regulator affecting the territory quality and the territory size that an animal can have [11,12].The contender pressure hypothesis suggests that there is a negative correlation between the territory size and conspecific density [12]. The intrusion and expulsion on territory will be more intensify as the density of conspecifics increases [12]. Additionally, the net benefits of occupants rapidly decline as maintenance costs increase [13,14]. This hypothesis has been tested in some aquatic species such as the crayfish (Orconenectes propinquus) and Mediterranean trout (Salmo trutta) [6,15]. Within territories, dominance hierarchies may develop after individuals repeatedly interact with one another.A dominance hierarchy is developed after individuals repeatedly interact [16]. During competition, the winner becomes the dominant individual and the loser becomes the subordinate individual [17]. These relationships are retained and lay a foundation for the development of the dominance hierarchy in the population, which closely reduces the frequency of fighting for resource and contributes to the explanation of population fluctuations [16,17,18]. Research into dominance hierarchies in crustaceans has generally focused on the influences of resource types including females and food [17,19], and rarely addresses the relationship with territory, although the dominance hierarchy between occupants and intruders may regulate the territorial behavior by affecting the prior-residence effect [20,21]. This effect demonstrates that the occupants usually have a strong internal motivation to expel intruders and are more likely to gain advantages in such contests [6]. However, as previously mentioned, the increasing number of intruders may reduce the benefits of protecting the territory of the occupants [22], forcing them to decide whether to continue to maintain or abandon their territory once the benefits fall below a certain threshold [16,23]. Despite the related discussions that are arising, the connections between dominance hierarchy and territory in aquatic crustaceans are still poorly understood.The swimming crab (Portunus trituberculatus) is a territorial species in the shelf of West Pacific Ocean. Additionally, it is also an important aquacultural and ecological species [24,25]. The density of swimming crabs is usually high in most food-rich habitats [24]. It is vulnerable to intense fighting and cannibalism due to its aggressive nature [26,27], resulting in high mortality and fluctuations in the population [3]. Previous studies have found that cannibalism can attributed to lack of food, crowded space, and stresses generated by an unhealthy environment [28,29]. Territory may have potential in regulating cannibalism of territorial crustaceans since it has been indicated to affect competition [25]. However, the systematic underlying patterns of how these behaviors associated with aggression interact with territory are rarely examined [30]. Therefore, in order to better understand and predict the population fluctuations of P. trituberculatus, we quantitatively analyzed how the territorial behaviors of occupants are influenced by the different number of intruders in an indoor observation system. We predicted that with the increasing intruders, the occupant will shrink its territory size and reduce the activity associated with territory protection. The results of this study will improve our understanding of the territorial behavior of crustaceans and the predictive power regarding their population fluctuations.2. Materials and Methods2.1. Animal Collection and MaintenanceThe experiment was carried out from August to September of 2020 at the Key Laboratory of Mariculture, Ministry of Education, Ocean University of China, Qingdao, China. A total of 130 male swimming crabs (carapace widths (CW) = 9.33 ± 1.42 cm; wet weights = 112.63 ± 26.67 g) were collected from an aquaculture facility in Jiaozhou, Shandong Province. Crabs were acclimatized in separate aquaria (40.50 L, 45 × 30 × 30 cm) for one week. The temperature of the filtered seawater (80 µm) was 22 ± 1 °C, and the salinity was 30. The photoperiod was 12:12 light/dark. Aquaria were aerated continuously, and 1/2 of the water was exchanged daily. Crabs were fed with adequate Manila clams (Ruditapes philippinarum) each day. Excrement and leftover clams were removed from the aquaria 4 h after feeding.2.2. Experimental Design and ProcedureAn indoor behavioral observation system was set up in a quiet room (Figure 1) [25]. The system consisted of an infrared camera (DS-2CD864, Hikvision, Hangzhou, China), a monitor (233i, Philips, Zhongshan, China), a cylindrical experimental aquarium (white PVC, diameter = 120 cm; height = 100 cm), light source (5W, Philips, Zhongshan, China), shading cloth (white polyester, 2.5 m × 2.5 m), and shelter (black PVC board, 15 cm × 10 cm). The camera and light source were installed 1.0 m above the experimental aquarium. The aquarium was filled with 340 L of filtered seawater (depth = 30 cm). The shelter was fixed horizontally on the wall of the aquarium 15 cm from the bottom. To ensure that the system was not affected by external interference, the whole system was covered with a shading cloth. During recording, the temperature and salinity were the same as during acclimatization, but the aquarium was not aerated.After 1 week of acclimatization, crabs with intact appendages in the intermolt stage were selected. Three treatments were established, and the groups had 2, 4, and 8 crabs, denoted as group low (L), medium (M), and high (H), respectively. The crabs whose carapace widths differed by less than 5% were assigned to the same group to reduce the impact of size differences (group L: CW = 9.09 ± 1.17 cm; group M: CW = 8.87 ± 1.76 cm; group H: CW = 9.82 ± 1.90 cm). One crab was randomly selected from each group and designated as the occupant, and the remaining individuals were designated as intruders. The numbers of intruders in the three treatments were 1, 3, and 7, respectively. Crabs were not fed for 24 h before the experiment. A white number was painted on the carapace of each crab for identification (Liquitex HB, Liquitex Artist Materials, Athens, OH, USA). Each treatment had eight replicates, and each crab was only used once in the experiment.At the beginning of the experiment, ten Manila clams were randomly added into the aquarium as prey to stimulate the crab to establish a territory [25]. An occupant crab was then added to the aquarium, allowing it to move freely. The intruders were placed gently into the aquarium after 24 h of introducing an occupant. After that, the experiment continued running for 1 h [20]. Each experiment lasted 25 h.2.3. Data Acquisition and QuantificationOnly the videos taken after the addition of the intruders (1 h) were analyzed to quantify the territorial behavior of occupants.2.3.1. Territory SizeTo assess territory size, one picture was captured from the video every minute, resulting in 60 total frames being obtained from each video. The coordinates of occupants during that hour were determined using the GetData Graph Digitizer 2.26, and these data were used for the calculation of the territory size [25]. Using Matlab 2019, the size of each point was linked up, the territory sizes were calculated using the 95% fixed kernel method, and the least-squares cross-validation value was used for smoothing [31,32]. The bivariate density for each normal distribution at any given point was calculated (Formula 1) [31,32,33] as (1)Dn=12πσxσy1−ρ2×exp{−12(1−ρ2)[(x−μxσx)2+(y−μyσy)2−2ρ(x−μxσx)(y−μyσy)]}In this formula, σx and σy is the stand deviation of the coordinates. μx and μy are the coordinates of the center of the experimental aquarium. ρ is the X-Y covariance.2.3.2. Defense Time and Occupying TimeDefense time, which is equal to the time spent on patrolling (move in its territory) and expelling intruders (dash to the intruder, direct its cheliped, and fight with it), was calculated [14,34]. Besides, time occupying shelter (more than 2/3 of the carapace under the shelter) of each occupant was determined [25].2.3.3. Quantification of Territorial Behaviors, Bouts, and Prior Residence EffectThe total number of territorial behaviors in occupants, which includes aggressive dash, aggressive wave, and fight, was calculated using the EthoVision XT 10 software, and each behavioral component was identified (Table 1) [34]. Furthermore, we counted the number of bouts initiated by the occupant and intruder in each replicate separately, as well as the winning percentage for each type of crab. One bout is defined as the process of crabs moving to, contacting, and retreating. The crab that caused its opponent to retreat repeatedly in a bout was considered the winner [29]. The crab gained a “dominance sign” when it caused its opponent to retreat in the last bout [20]. If the occupant accumulated six consecutive dominance signs and did not lose in any fights during the next 10 min, it was categorized as exhibiting the prior residence effect. If the occupant was defeated while accumulating dominance signs, the sign returned to zero. If the occupant did not gain 6 consecutive dominance signs in the selected period, it was categorized as not showing the prior residence effect. Besides, if it was difficult to tell winner and loser apart in several bouts, the occupant was categorized as “not clear” [20]. Finally, the number of bouts between the occupant and intruders at different distances from the shelter was calculated. For the convenience of analysis, the distances were measured in units of 25 cm.2.3.4. Dominance HierarchyDominance hierarchy was analyzed using videos in high density group (1 occupant + 7 intruders, 1 h × 8 replicates). The number of winning bouts of each individual was counted, and a matrix of wining bouts was established. Nij is the number of bouts in which individual i defeated individual j. Mij is the number of bouts in which individual i was defeated by individual j. Pij is the winning percentage of individual i when it fought against individual j. David’s Score of individual i in this group was calculated (Formula (2)) using the following (2)DS=w+w2−l−l2 (3)w=∑Pij (4)w2=∑(Pij×Nij) (5)l=∑(1−Pij) (6)l2=∑(1−Pij)×∑Mij where w (3) is the sum of the winning percentages of individual i in the fights with each individual; w2 (4) is the weighted sum of the individual’s w; l (5) is the sum of failure rate of individual i in the fights with each individual; and l2 (6) is the weighted sum of the individual’s l value [16].2.3.5. Territorial Behavior ScoreThe occupant’s territory size, total number of territorial behaviors, defense time, number of bouts, and winning percentage were input into the principal component model. The first component (PC1) was designated as the “territorial behavior score” because it reflected the territoriality of occupants [25,35].2.4. Statistical AnalysisAll data were analyzed using SPSS Statistics 24.0. A chi-square test was used to analyze the proportion of prior residence effect. A Spearman test was used to identify correlations in the number of bouts between occupant versus intruder and distance from the shelter. The generalized linear mixed model (GLMM) was used to analyze the territory size, defense time, occupying time, number of territorial behaviors, the total number of bouts in the selected period and bouts initiated by occupant, the winning percentage of occupant, the difference of David’s Score between occupants and intruders, and the territorial behavior score of occupant. When analyzed in GLMM, a binomial distribution was used for the number of bouts initiated by the occupant and the winning percentage of occupant; a skewed distribution was used for territory size, defense time, occupying time, David’s Score between occupants and intruders, and territorial behavior score; and a negative binomial distribution was used for number of territorial behaviors and bouts. For models with skewed distributions, the residuals were checked to for normal distribution to assess model fit. Models with binomial error structures were checked for overdispersion. For number of territorial behaviors and bouts, negative binomial distributions were chosen to deal with overdispersion. When analyzing the difference of David’s Score between occupants and intruders, the ID of crabs and number of parallel groups were included as random factors, and the classifications of occupying crabs and intruders were taken as fixed factors. In the rest of the analyses, the ID of crabs was included as a random factor, and the number of intruders was taken as a fixed factor. Post hoc comparisons between different treatments were conducted with the calculation module in SPSS. For all tests, p < 0.05 was considered to be statistically significant.3. Results3.1. Territory SizeThe territory size of occupants was significantly affected by the number of intruders (F = 3.851, p = 0.034, Figure 2). The territory size of occupants in group H was significantly smaller than those of group L (t = 4.173, p < 0.001) and group M (t = 6.146, p = 0.047).3.2. Defense Time and Occupying TimeWith an increased number of intruders, the defense time of occupants increased significantly (F = 5.217, p = 0.044, Figure 3A), but the occupying time did not (F = 2.364, p = 0.729, Figure 3B). The defense time of occupants in group M (t = 3.129, p = 0.011) and group H (t = 4.467, p = 0.016) was significantly longer than in group L.3.3. Quantification of Territorial Behaviors, Bouts, and Prior Residence EffectThe number of intruders significantly affected the number of territorial behaviors (F = 4.868, p = 0.039, Figure 4A), bouts (F = 8.322, p < 0.001, Figure 4B), and the winning percentages of occupants (F = 5.297, p = 0.047, Figure 4C). With an increased number of intruders, the number of territorial behaviors increased significantly, and the number in group H was significantly more than group L (t = 4.739, p = 0.002) and group M (t = 5.551, p = 0.029). The number of bouts between occupants and intruders also increased significantly, and the number was significantly higher in group H than group L (t = 6.869, p = 0.003) and group M (t = 7.748, p = 0.044). In group L, the number of fights initiated by occupants was significantly higher than those initiated by intruders (F = 8.517, p = 0.024).Among the three treatments, the proportion of the occupants in group L and group M exhibiting the prior residence effect was higher (Table 2). The prior residence effect in group H was not obvious (Table 2).3.4. Dominance HierarchyIn all three treatments, there were significantly negative correlations between the number of bouts between the occupant versus intruder and distance from the shelter (Figure 5). With increased distance from the shelter, the number of bouts in all groups significantly decreased, but the decreases in group M and group H were much more pronounced compared to group L (group M: R = −0.789, p < 0.001, n = 8; group H: R = −0.908, p < 0.001, n = 8; group L: R = −0.505, p = 0.02, n = 8). Among the eight repeats in group L, five occupants attained the highest David’s Score, two occupants ranked second, and one occupant ranked third (Figure 6). In group L, the David’s Scores of occupants were significantly higher than those of intruders (t = 24.236, p < 0.001, n = 8, Figure 6).3.5. Territorial Behavior ScoreThe territorial behavior score (i.e., PC1) included dimensionality of the territory size, total number of territorial behaviors (aggressive dash, aggressive wave, and fighting), defense time, number of bouts, and winning percentage of occupants, and explained 54.8% of the total variance (Table 3). The territorial behavior scores of occupants were significantly affected by intruder number (F = 19.462, p = 0.038, Figure 7). With an increasing number of intruders, the territorial behavior score of occupants decreased, and the score of group H was significantly lower than group L (t = 5.432, p < 0.001).4. DiscussionWhen the number of intruders increased, the number of territorial behaviors and bouts increased (Figure 4A). Meanwhile, the proportion of fights initiated by occupants decreased as the number of intruders increased (Figure 4B). These results demonstrate the increasing aggression by the occupant and costs of expelling. With increased intruders, the territory size of occupants decreased (Figure 2), which likely reflects the costs to benefits balance of occupying a territory, that is, in highly competitive scenarios it is no longer beneficial to occupy a large territory [36]. Moreover, the proportion of individuals that exhibited the prior residence effect decreased as the intruder number increased (Table 2), suggesting that the costs of occupying correspondingly enhanced. In high density group that has 7 intruders, the occupant may not have the expected advantage in competition with intruders due to the high costs. These results were aforementioned with the territorial behavior score-when the number of intruders increased, the scores significantly decreased (Figure 7), indicating the territoriality of occupants became weaker. The results are in agreement with the fact that P. trituberculatus is an energy maximizing species–meaning it acquires energy and reduces costs of energy acquisition whenever possible [8]. Defense time increased when the number of intruders increased (Figure 2 and Figure 3A), which is also consistent with the response to the change of external conditions of energy maximizing [14]. However, in our study, the winning percentage of occupants did not significantly decrease with the increase in the number of intruders (Figure 4C), consistent with studies on behaviors (Allopetrolisthes spinifrons) [37]. We concluded that the reduced territory size may not be a forced choice caused by pressure from intruders, but a voluntary abandonment of part of the territory by the occupant.Shelter is the core of the territory and the main target of competition for resources in the wild [38]. For example, when an intruder cannot be evicted from a territory, fiddler crabs (Leptuca leptodactyla) will preferentially protect their cave shelter instead of territory [39]. In our study, when the number of intruders increased, the number of bouts farther away from the shelter decreased (Figure 5), which reflected a prioritization of the shelter. Despite this, the occupying time of the occupant did not increase as the intruder count increased (Figure 3B), which can be explained by the fighting that frequently occurred in the vicinity of the shelter causing the occupant to lose and regain shelter intermittently. Similar to reports studying king crab (Paralithodes Camtschaticus) and spider crab (Maja squinado) [40,41], adjusting the population density and total number of shelters may be an effective way to reduce the territorial disputes among P. trituberculatus, in other words, to diminish their territoriality.In a population, the dominance hierarchy is often associated with preferential access to resources [6,42,43]. In studies of territorial behavior, occupants usually have higher levels on dominance hierarchies than intruders [16]. The top five David’s Scores were gained by occupants in high density group, and those occupants also exhibited the prior residence effect. That is to say, those five occupants had absolute dominance of their territory (Figure 6, Table 2), which is similar to that found in territorial competition in gobies (Elacatinus prochilos) and crayfish (Austropotamobius pallipes), respectively [18,42]. Dominators occupy larger and better territories, and expel subordinates from it, which negatively affects the growth of the subordinate and even the survival rate in the population. All the dominator–subordinate dualities can also be combined to form the dominance hierarchy of the population [44]. In research of population of marine biology, it is significant to understand the influence of the dominance hierarchy on territorial behavior. For example, it was found that the density and yield predictions of Norwegian lobsters (Nephrops norvegicus) could be assessed using the dominance hierarchy [45]. In addition, an individual’s rank on the dominance hierarchy is usually related closely to its size [46]. Classification in crustacean cultures can therefore be based on individual sizes, this can be utilized with the principal goal to reduce differences in the dominance hierarchy among individuals, which is also an effective means to improve the survival rate and yield by adjusting territorial behavior.Territorial behavior of intruders and occupants also plays an important role in regulating population density [15,47,48]. In the wild, a region with abundant resources and few competitors will encourage animals to aggregate. When the conspecific density becomes high, territories the individual owns will shrink and overlap, leading to an intensified territorial dispute [49], which causes frequent emigration, fight and cannibalism and reduces density and competitive pressure until the density stabilizes [47]. In our study, the swimming crabs were unable to emigrate to reduce the density due to limited observation areas, therefore, territorial competition was the only means for intruders to acquire territory and for the occupant to maintain their territory. This also explained why territory size decreased and the number of territorial behaviors and bouts increased when the number of intruders was increased (Figure 4A, 4B). The high density promotes animals to invade nearby territories and engage in fierce fighting, accompanied by increased mortality and reduced growth rates [50,51]. In extreme cases, fish give up their territory when the density is too high, and they cannot emigrate [52]. However, relevant results have not been reported in crustaceans, and the trends of decreasing numbers of territorial behaviors were not observed in the multiple intruder treatments in our study (Figure 3A). A possible reason is that swimming crabs live mainly on the seabed, a two-dimensional surface, which enhances the importance of territory and the intensity of territorial dispute, and male crabs reluctant to abandon their territory [53]. From the perspective of territorial behavior, a high-density crab population may be unstable and therefore unlikely to last for long in the community.5. ConclusionsIn our study, when the number of intruders increased, the number of a swimming crab’s territorial behaviors increased, but their territory size decreased. When the density of crabs was high, there was a clear dominance hierarchy. According to the results, due to the intensification of territorial competition, an increase in density may lead to a decrease in survival rate. However, the territorial behavior of crabs is greatly influenced by vision [54], thus increasing the complexity in the seabed, such as establishing artificial shelters or planting algae, which may be an effective means by which to limit territorial behavior [46,55]. Therefore, optimizing the habitats in marine ranching using environmental enrichment will ensure the densities of crabs rationally increases by reducing conflict over territory [56]. The specific approaches to optimizing environments need to be examined further to assist in the exploration of population dynamics of swimming crabs.

animals : an open access journal from mdpi

10.3390/ani11102951

PMC8533022

West Nile fever is an arthropod-borne viral disease that is transmitted from birds to humans and animals by mosquitoes. Humans may develop a severe disease, which sometimes can be fatal. At the end of the 20th century, the first outbreaks of West Nile fever among humans in urban environments in Eastern Europe and the United States were reported. The epidemics were characterized by the neurological form of the disease with a fatal outcome. Since the first outbreak of the disease in Serbia, the highest number of cases occurred in 2018. West Nile fever spread is driven by location and time, which means nearby locations and periods have similar features. Recognition of patterns of spread of the disease has the potential to facilitate the mosquito control program and disease prevention. This study aimed to examine the geographical and temporal similarities of registered cases during the epidemics in the period 2017–2019 in South Banat District, Serbia. We identified the following factors as crucial for the prediction of possible outbreaks: the presence of virus in natural reservoirs, mosquito abundance; precipitation, high water level of rivers followed by a consequent sudden decrease of precipitation and withdrawal of rivers into the main bed, and favorable temperatures.

West Nile virus (WNV) is an arthropod-born pathogen, which is transmitted from wild birds through mosquitoes to humans and animals. At the end of the 20th century, the first West Nile fever (WNF) outbreaks among humans in urban environments in Eastern Europe and the United States were reported. The disease continued to spread to other parts of the continents. In Serbia, the largest number of WNV-infected people was recorded in 2018. This research used spatial statistics to identify clusters of WNV infection in humans and animals in South Banat County, Serbia. The occurrence of WNV infection and risk factors were analyzed using a negative binomial regression model. Our research indicated that climatic factors were the main determinant of WNV distribution and were predictors of endemicity. Precipitation and water levels of rivers had an important influence on mosquito abundance and affected the habitats of wild birds, which are important for maintaining the virus in nature. We found that the maximum temperature of the warmest part of the year and the annual temperature range; and hydrographic variables, e.g., the presence of rivers and water streams were the best environmental predictors of WNF outbreaks in South Banat County.

1. IntroductionWest Nile virus (WNV) is a single-stranded RNA arbovirus from the Flaviviridae family, genus Flavivirus. WNV is primarily a zoonotic agent transmitted between birds as the main reservoir hosts and mosquitoes as vectors. Humans and horses can be infected as spillovers host, but the infection in these organisms ends without further transmission of the virus, i.e., these organisms are dead-end hosts [1,2]. WNV is today considered the most important cause of viral encephalitis in humans worldwide [3].The WNV has been isolated or identified serologically in many vertebrate species [4]. Since 1998, cases of viral encephalitis caused by WNV in horses have been reported in Italy, France, and North America. Research conducted in parts of Europe and the Middle East has established that as much as one-third of the tested horses were exposed to the WNV, with or without symptoms of the disease [5,6].Since the discovery of the first infection in humans in 1937 [4], there has been a significant spread of the disease globally. In most cases, the infection was asymptomatic and only a few cases of severe neurological forms of the disease were reported [2]. At the end of the 20th century, the first West Nile fever WNF outbreaks among humans in urban environments in Eastern Europe and the United States were reported [2]. Since 2012, human WNF cases have been recorded every year in Serbia. In terms of the number of patients, the years 2013 and 2018 stand out, when the largest numbers of clinical cases of infection were registered [7]. The epidemics were characterized by a serious neurological form of the disease with a fatal outcome. During these epidemics, several common characteristics were observed at the time of the disease outbreaks: the Culex pipiens mosquito was identified as a vector of the disease, significantly less precipitation was registered than the usual multi-year average, significantly higher than normal summer temperatures were recorded, and all areas were located near large rivers which provide adequate living conditions for residential and migratory species of wild birds [2]. Besides, the epidemic in the northeastern United States was accompanied by the epizootic in birds, especially crows [8]. The sudden appearance of WNF in the United States can also be related to changes in the feeding habits of Cx. pipiens, the dominant enzootic species. Shifts in feeding habits from competent avian hosts early in the early stage of the epidemic to incompetent humans after mosquito infection, resulted in synergistic effects that greatly amplified the number of human infections [9].Given that WNF is an arthropod-borne disease, the spread of the disease is conditioned by climatic variations and landscape changes to the natural habitats of the mosquitoes and available bird habitats [1,4], geospatial data can be used for risk prediction and risk mapping [10]. With the development of geographic information system (GIS) software and other related tools for spatial data analysis, substantial progress has been made in risk analysis [11]. Numerous studies have used geospatial data and GIS software for detecting, analyzing, and predicting spatial patterns of disease occurrence [10,12,13,14]. The advantage of the GIS analytical tool is the ability to integrate and analyze risk factors and create WNV risk maps for humans and other susceptible hosts [15]. In addition to visualizing epidemiological data and presenting the geographical distribution of the disease in a much more predictive manner, the spatiotemporal analysis identifies spatial and temporal clusters and identifies risk geographic areas. Spatial inquiry can also inform vector control policy, including where to prioritize limited control resources [16,17,18]. By identifying locations of the WNF hot spots, the mosquito control treatments can be rationally and more efficiently applied primarily in high-risk areas and, therefore, will more efficiently decrease virus transmission [17,19].The analyses conducted in this paper aimed to discover spatial, temporal, and spatiotemporal clusters of WNV infection in humans and animals in South Banat District, Serbia, also researching for causal relationships, i.e., identifying such clusters of WNV cases and researching possible causal or associative relationships with meteorological factors. Our focus was on understanding spatiotemporal patterns and risk factors associated with WNV infection in humans and domestic animals. An additional aim was to investigate the epidemiological characteristics of WNV outbreaks and provide a scientific basis for the effective control of this disease. Spatial research into the geography of the West Nile disease in Serbia includes WNV surveillance in mosquitos (collected at trap locations), surveillance of WNV infection in horses, wild bird surveillance for WNV presence, and collection of human case residential addresses [20,21]. In this study, the data on disease occurrence in humans and animals and locations where WNV-positive mosquito pools are found were combined with predictor variables from geospatial data sets to develop risk maps and identify hotspots, i.e., high-risk areas of WNV.2. Materials and MethodsAs shown in Figure 1, the analytical workflow involved collecting, processing, and descriptive analysis of cases of WNV infection in humans and animals; retrieving, manipulating, and processing environmental data; processing geospatial data, and; developing negative binomial regression models to measure the association between WNV infection and climatological factors. It comprises five major parts, i.e., collection and processing of epidemiological data, collection and processing of environmental data, exploratory analysis of WNV case data, statistical modeling, and mapping the distribution of WNV cases.Cases of WNV infection and climatological factors in South Banat District, were analysed using ESRI’s ArcGIS 10.5 software (ESRI Geographic information system company, West Redlands, CA, USA), then SaTScan v.9.6 (Martin Kulldorff and Information Management Services Inc., Boston, MA, USA), IBM SPSS Statistics v.26 (IBM, SPSS Inc., New York, NY, USA), and Pavanaarekh v.5. (Envitrans Infosolutions Private Limited, Ghaziabad, India). Data were manipulated temporally for the entire period and through spatial programming. Local patterns for grouping of animals, humans, and mosquitoes that were positive over time for the presence of WNV were investigated, using three types of models, from heat maps to complex space–time models.2.1. Data and Sources of InformationFor the spatiotemporal analysis, we collected data for the time of the onset of the disease, incidences, and geographic locations of registered cases. Those data were aggregated based on geographic locations. The final input data had to be compliant with the software in use, i.e., ArcGIS 10.5 and SaTScan v 9.6. Data on WNV cases, animals seropositive for WNV, and WNV-positive mosquito pools were georeferenced and transformed into shapefiles that were used for spatial, temporal, and spatiotemporal analysis. A variety of data sets were compiled that contained land cover information and environmental data relevant to disease transmission and the survival of WNV in nature. As base maps, polygon shapefiles and rasters of the geography of South Banat District were used [22,23].Land cover data for South Banat District were acquired from Copernicus Land Monitoring Service (CLMS). Their CORINE Land Cover (CLC) inventory consists of an inventory of land cover in 44 classes, uses a minimum mapping unit of 25 hectares for areal phenomena, and a minimum width of 100 m for linear phenomena. Land cover data include land use, vegetation, topography, soils, and wetlands [24].Climatic data, i.e., daily temperature, precipitation, and the water level were acquired from the Meteorological Service of the Republic of Serbia. We calculated two-weekly mean temperature (°C), total monthly precipitation (mm), and monthly river water levels (cm) for the entire study period [25]. Then, we imported these data into regression models, and tested them against registered WNV human cases and registered WNV-positive mosquito pools. The data on registered WNV infection in animals and WNV-positive mosquito pools obtained during our research were used in the analysis, as were data taken from the Institute of Public Health Pančevo and the European Center for Disease Prevention and Control (ECDC), i.e., official data on registered WNV human cases [9,20]. In negative binomial regression models, WNV human cases and data on WNV-positive mosquito pools were aggregated monthly. Negative binomial regression analysis was used to test for associations between predictors (explanatory variables) and dependent count variables, i.e., climate factors and WNV human cases and positive mosquito pools, respectively. Given that it is the variance of the dependent variables was higher than the mean (variance = 36.68, mean = 3.24, CV = 187%), the data related to WNV human cases were first tested to determine whether they followed Poisson’s theoretical distribution. The Kolmogorov–Smirnov (KS) test was used to determine whether the data on WNV human infections originates from a population with a specific distribution, i.e., the Poisson distribution. The obtained result of the KS test (asymptotic p-value of 0.000) confirmed that the data do not follow the Poisson distribution. Moreover, the significant deviation of variance from the mean value of registered WNV human cases (the deviance over the degree of freedom equal to 1.008) indicated overdispersion of the data. Taking into account these observations and the results of statistical tests, the negative binomial regression model was found appropriate for the analysis. The goodness of fit determined how well the selected model fits a set of observations compared to the other model (Poisson regression model). In assessing whether a given distribution is suited to a data set, the following tests were used: Likelihood test, Bayesian information criterion and Akaike’s information criterion. The omnibus test, i.e., likelihood ratio chi-square test was used to test whether the negative binomial regression model represents a significant improvement in fit compared to the null model. The results of the omnibus test showed the significant improvement of the model based on the negative binomial distribution, i.e., likelihood ratio chi-square statistic equal to 12.48, p-value of 0.002. The same procedure was applied when it came to data on WNV-positive mosquito pools. The regression model was validated in IBM SPSS software by performing a cross-validation method with the same data divided into two parts, i.e., the training sample which covered 55% data and the test sample which covered 45% of the data.2.2. Study AreaThe study area was the South Banat Administrative District with its settlements. The area is rich in running and standing waters, has a network of irrigation channels, and incorporates the Danube and Tamis rivers, with the Danube river forming a delta right next to the city of Pančevo, as well as in several locations downstream. The administrative center of the district is the city of Pančevo, which is located about 2.5 km upstream from the mouth of the river Tamis into the Danube. The district lies in the region of Banat and has a population of 293,730 inhabitants [26,27]. The area is also a significant habitat for wild birds [28]. The South Banat Administrative District extends to Serbia’s north, in the southeastern part of the Autonomous Province of Vojvodina. It is bordered by the Danube and Tamis rivers, and on the east by the state border with Romania. South Banat covers an area of 4245 km2 [26].Livestock production in South Banat is heterogeneous. Different types of domestic animals are bred and held on rural holdings and industrial farms. Domestic animals in rural areas are usually reared extensively on backyard holdings, but intensive production on industrial farms is also present [29].2.3. Mosquito Pooled SamplesTo collect pooled mosquito samples, modified Centers for Disease Control and Prevention (CDC) mosquito light traps were set in the afternoons at selected location sites. Criteria for site selection were accessibility, an unkempt natural environment, with lots of trees and vegetation, and the vicinity of farms and backyard holdings, wetlands, temporary standing water, water-filled containers, and shady places. The mosquitoes were brought into the Laboratory for Medical and Veterinary Entomology, Faculty of Agriculture, University of Novi Sad. The mosquitoes were kept alive on dry ice to prevent virus degradation. In the laboratory, mosquitoes were sorted, identified, and grouped on a cold table. Only the Cx. pipiens mosquitoes (biotypes pipiens and modestus) were separated and transported on dry ice to the Department for Virology of the Scientific Veterinary Institute Novi Sad to be tested for WNV presence. Mosquitoes were tested in pools of up to 50 individuals. When traps contained more than 300 Cx. pipiens mosquitoes, two pools of up to 50 individuals were tested per trap.Sampling of mosquitoes from selected localities was performed in 2017 and 2018 in four sampling periods, one in June, two in July, and one in August. A total of 80 samples were collected and tested for WNV presence.2.4. Surveillance Targeting Dead Wild BirdsDead wild birds found in the natural environment, shot targeted wild birds, and tracheal/pharyngeal swabs of wild birds, live-trapped during the ringing and other activities of bird protection societies, particularly the resident species most susceptible to infection, were collected in the period from June to December 2017 and 2018 and tested for the presence of WNV genomic RNA.2.5. Blood SamplesBlood serum samples from cattle, pigs, and chickens were collected at several locations in South Banat District. The blood of domestic animals was sampled on industrial farms and backyard holdings. Sampling was performed at equal monthly intervals from March to October in 2018 and March to October in 2019. To determine seroconversion, sera were tested for the presence of specific IgG antibodies against WNV using the ELISA test (ELISA test INGEZIM West Nile Compac, INGENASA, Madrid, Spain).2.6. WNV Genome Detection by Molecular MethodsMosquito pools (up to 50 individuals in one pool) and wild bird samples (tissues and tracheal swabs) were tested for WNV RNA presence by TaqMan-based one-step reverse transcription real-time PCR (RT-qPCR) that amplified both lineage 1 and 2 strains of the virus, as described elsewhere [21].2.7. Spatiotemporal AnalysisWNV infections in humans, animals, and mosquitoes were analyzed for both space and time patterns using the following analytical approaches: kernel density analysis (Parzen windows-cores), incremental spatial autocorrelation (Moran’s I test), hot spot analysis (Getis-ord GI* statistics), and space–time aggregation (temporal, spatial, and spatio-temporal analysis of clusters using Kulldorff spatial scan statistics).2.7.1. Kernel Density EstimationKernel density estimation (KDE) is a method used to create “heat maps”, commonly applied in epidemiological research of WNF and other diseases [16,30,31,32,33]. KDE was performed using the spatial analysis/kernel density tool of the ArcGIS desktop 10.5 software package. The KDE tool was used to calculate the density of cases of WNV infection in the immediate vicinity (neighbourhood) around each entity (point object on the map that represents a registered case of WNV infection). It takes known quantities, i.e., the numbers of WNV infections, and extrapolates them across the landscape based on the quantity that is measured at each location and the spatial relationship of the locations of the measured quantities. The KDE technique creates a smoothed map of the density of cases of WNV infections at each location where positive cases were registered. The output raster surface is transformed into contour maps to allow for overlay onto other geographic layers. Although different sizes of kernels were analyzed, results for 660 m are reported, which was estimated as a standard distance by ArcGIS software itself. It is the default search radius (bandwidth) based on Silverman’s rule-of-thumb bandwidth estimation [33,34,35].2.7.2. Spatial Autocorrelation AnalysisSpatial autocorrelation analysis is a precondition for the in-depth spatial analysis of WNV infections. This method included global autocorrelation and local autocorrelation. The global autocorrelation was used to analyze whether the attributes specified in the study area were relevant at the level of the entire South Banat District, while the latter accurately determines where such attributes are gathered and reveals the spatial distribution pattern and the approximate spatial aggregation range. Moran’s index I, which ranges from −1 to +1, is an indicator of global autocorrelation analysis. A value close to 1 or to –1 indicates, respectively, a strong positive or negative spatial autocorrelation. Moran’s I can be tested based on Z-score and p-value, to determine whether or not the null hypothesis, that the incidence of WNV infection was randomly scattered in space, should be rejected [36,37].2.7.3. Hot Spot AnalysisHot spot areas are concentrations of incidents within a limited geographical area that appear over time. Hot spot analysis is also statistically known as cluster analysis. The most intuitive type of cluster is when only the location of incidents is considered. Thus, the location with the highest number of incidents is considered to be a hot spot [38]. In our investigation, one cluster represented a group, i.e., a series of cases of WNV infection in one geographical area in different animal species and humans.The hot spot analysis was based on a statistical calculation known as Getis-ord Gi * statistics. Getis-ord Gi * statistics are based on the calculation of z-statistics and p-values, where z-statistics determine whether the examined values of a particular attribute of a cluster entity are above or below the average values at the level of the whole geographical area (necessary to determine “hot” or “cold” points), while p-values determine whether the grouping of units into clusters is the result of a random event or a consequence of some external influence, a phenomenon that conditions the grouping. Hot spot analysis was performed using the cluster mapping/spatial statistics tools of the software package ArcGIS desktop 10.5 [39,40,41,42].2.7.4. Space–Time AggregationThe most sophisticated space–time analyses were conducted using SaTScan v9.6., a method that has been previously used in many WNF studies to analyze spatial, temporal, and spatio-temporal data using spatial, temporal, or spatio-temporal scanning. It detects geographical areas where WNV cases are grouped more densely than their usual distribution. Within the cluster, all cases of the disease are associated with the same epidemic [15,43,44,45,46,47]. An important feature of spatial scanning is the ability to detect cluster locations and make inferences about clusters, that is, locating the geographic areas of the most likely clusters and the secondary clusters [48].3. Results3.1. Epidemiological Characteristics of WNV Outbreaks in 2017, 2018, and 20193.1.1. Descriptive StatisticsA total of 68 human clinical, laboratory-confirmed cases of WNV infection were included in this study. The case number of WNV in humans maintained a seasonal variation in the study period, ranging from 0.24 to 2.48 cases per 100,000 person-months at risk. We estimated that the average incidence rate of WNV in 2017 was 0.24 per 100,000 person-months at risk, whereas, in 2018, the average incidence rate was 2.48 per 100,000 person-months at risk, and in 2019 it was 0.58 per 100,000 person-months at risk. The period prevalence in 2017 was 0.0017%, while in 2018 it was 0.017%, and in 2019 it was 0.0041%. The period prevalence in 2018 in populated areas ranged from 0.0028% to 0.3636% (Supplementary Figure S10). The largest number of registered WNV cases was recorded in Pančevo, a total of 18 cases. The highest prevalence of WNV infection was recorded in the villages of Mali Žarm and Dupljaja, 0.36% and 0.10%, respectively. Figure 2 shows the distribution of clinical cases of WNV infection and the epidemiological characteristics of the WNV human outbreaks in 2017, 2018, and 2019 in the South Banat District.3.1.2. SeasonalityThe distribution of human cases of WNV infection displayed a clear seasonal pattern (Figure 3). During the WNV transmission season in 2017, human cases of WNV infection were recorded between June to September, whereas in 2018 they occurred between June to October (Figure 3). The highest number of human cases were recorded in 23 August 2018 in total. In the following transmission season in 2019, although a significantly smaller number of cases of infection were registered, the appearance of the disease was also seasonal. Thus, human cases of WNV infection were recorded between July to August 2019, and the highest number of cases was recorded in July, i.e., seven in total (Figure 2 panel d).The average ambient and average ambient maximum temperature of the two weeks before the first WNV human case was detected in 2017 were 22.5 °C and 28.26 °C, respectively, in 2018 were 20.84 °C and 25.96 °C, respectively, and in 2019 were 23.35 °C and 26.61 °C, respectively. The average ambient and average ambient maximum temperature of the two weeks before human cases terminated ranged between 23.35 °C to 26.05 °C and 26.61 °C to 32.31 °C, respectively. The highest incidence was observed when the two-week average ambient and average ambient maximum temperature were 26.73 °C and 26.73 °C, respectively. Negative binomial regression analysis was used to estimate relationships between registered WNV human cases and monthly average ambient temperatures, i.e., minimum average temperatures, average temperatures, and maximum average temperatures. Regression analysis proved a positive association between environmental temperature, the number of WNV-positive mosquito pools, and the registered number of WNV human cases. The estimated negative binomial regression coefficients for the model predictors variables, i.e., average minimum temperature, and WNV-positive mosquito pools, were positive and significant, indicating that these factors are significant predictors of infection in humans. The negative binomial coefficients for the minimum average temperature and WNV-positive mosquito pools were βmin_t⁰ = 0.251 and βmosq.1 = 0.335, respectively, bound with 95% Wald confidence interval (CI) of 0.072 to 0.431 and 95% CI of 0.029 to 0.641, respectively. The estimated p-values were.006 and.032, respectively. The results showed that for a one-unit change in the predictor variables, the difference in the logs of expected counts of the WNV human cases is expected to change by the respective regression coefficient, provided that the other predictor variables in the model are held constant. Concerning the incidence rate ratio (IRR) the exponential value of the regression coefficient βmin_t⁰ for minimum average temperature indicates that every increase of average minimum temperature for one degree of celsius would increase the incidence rate of WNV infection in humans by a factor of 1.286 or 28.6% and for predictor variable WNV-positive mosquito pools the exponential value of the coefficient βmosq.1 indicates that every increase of registered WNV-positive mosquito pools by one would increase the incidence rate of WNV infection in humans by a factor of 1.398 or 39.8%, provided that the parameter of the minimum average temperature is unchanged. The negative binomial regression coefficients for the average temperature and WNV-positive mosquito pools were βaverge_t⁰ = 0.223 and βmosq.2 = 0.331, respectively, bound with 95% CI of 0.055 to 0.392 and CI of 0.028 to 0.635, respectively, while for the average maximum temperature and WNV-positive mosquito pools were βmax_t⁰ = 0.219 and βmosq.2 = 0.303, respectively, bound with 95% CI of 0.043 to 0.395 and 95% CI of 0.004 to 0.602, respectively.The results of the regression analysis show that fluctuations, ie changes in the water level of the Danube, affect the number of registered cases of WNV infection, both in humans and mosquitoes. The estimated negative binomial regression coefficients for the minimum, average, and maximum water level, were negative and significant, which practically means that as the river level decreases, the number of registered cases increases. The following values of regression coefficients concerning the variables minimum water level, average level and maximum water level were obtained: βwater_min = −0.021 (95% CI −0.038 to −0.003, p-value 0.02) βwater_average = −0.0019 (95% CI −0.034 to −0.005, p-value 0.009) and βwater_max = −0.013 (95% CI −0.023 to 0.003, p = 0.012).Regression analysis proved a positive association between environmental temperature and the number of WNV-positive mosquito pools. The estimated negative binomial regression coefficient was positive and significant, indicating that this factor is a significant predictor. However, the precipitation was negative and insignificant, so as such it was excluded from the model. Instead of this predictor, the model includes the water level of the Danube as a predictor. The following values of the coefficients were obtained by regression analysis: βmin_t⁰ = −0.184 (95% CI −0.005 to 0.364, p = 0.04) βwater_max = −0.013 (95% CI −0.026 to −0.001, p = 0.04). The estimated negative binomial regression coefficient for predictor—the water level of the Danube was negative and significant, indicating that this factor is a significant predictor.The model was validated by the cross-validation method with satisfactory results. For the sample test that included 45% of the data, Spearman’s rank correlation coefficient, ρ was 0.708 (p = 0.033), while Kendall’s compliance coefficient, τ was 0.591 (p = 0.045). For validation purposes, the data on the number of registered WNV human cases of infection in humans, data on WNV-positive mosquito pools, and temperature were used. For the training sample with 55% of the data the correlation coefficient ρ was 0.628 (p = 0.029) and Kendall’s τ was equal to 0.557 (p = 0.018). By comparing the actual values of the WNV human cases and the values predicted by the model were calculated bias, mean average error (MAE), normalised MAE, mean square error (MSE) I root mean square error (RMSE). The following values were obtained: bias = 1.53, MAE = 2.07, NormMAE =0.64, MSE = 16.03, and RMSE = 4.08.More information on the analysis of the effects of environmental temperatures on the risk of WNV transmission is available in the Supplementary Material (Supplementary Tables S1–S15).3.1.3. Geographical Distribution and Abundance of WNV-Positive MosquitoesData in Figure 4 displays the distribution of WNV-positive mosquito pools and the distribution of human cases between 2010 and 2019. Mosquitoes were surveilled for the presence of WNV in 2014, 2015, 2017, and 2018. During the 2017 and 2018 surveillance seasons, significantly higher numbers of WNV-positive mosquito pools were registered, especially in 2018. In 2017, the virus was detected in mosquitoes at a total of 5 out of 10 locations, while in 2018, circulation was proven at 9 out of 10 tested locations. The frequency of WNV-positive mosquito pools was used to create kernel density maps for each year (2017 and 2018) and these were compared with kernel density maps of WNV infection in humans and animals (Figure 4 panels (e) and (f)). By comparing clusters, we noticed a positive match between the clusters of WNV-positive mosquitoes and the clusters of registered human and animal cases of infection, shown in Figure 4 panels (c) and (d). Moreover, the clusters of WNV-positive mosquito pools, which were classified by the KDE technique into the category of very high-density clusters, coincided with the clusters of WNV infection in humans and animals, which were also categorized as high-density clusters or very high-density clusters. Matches were not observed only in those mosquito sampling localities where there were no domestic animals, and that were distant from populated areas. The increased higher numbers of registered human and animal cases correspond to the areas with a higher density of WNV-positive mosquito pools.3.1.4. The Water Level of the Danube and Tamiš RiversComparison of changes in water levels in the Danube and Tamis rivers and cases of WNV showed that changes in water levels in rivers were accompanied by a change in the number of registered clinical cases of WNV in humans. Except in 2016, every increase in water levels and a sharp decline in river levels was accompanied by a consequent rise in the number of sick people. An identical pattern was observed for cases of infection in tested animals (Figure 5). It was notable that animal cases of infection appear after the end of the period of decline of previously high water levels. (Figure 6).For more information on the analysis of effects of river water levels and precipitation on the risk of WNV transmission see Supplementary Material (Supplementary Tables S15–S19).3.1.5. Results of Seroconversion Tests in Domestic and Wild AnimalsThe study of the seroprevalence of WNV in domestic animals showed high prevalences of infection in domestic animals, especially those raised on rural holdings in an extensive manner (Figure 7). The highest level of seroprevalence was recorded in cattle in 2018, followed by pigs raised on rural holdings, 45.71%, and 40.74%, respectively. Pigs reared on rural holdings are 7.79 times more likely to be infected with WNV than pigs on industrial farms. This odds ratio (OR 7.79, 95% CI 2.12 to 28.57) means the type of production was recognized as an exposure risk factor, and this category of animals is more prone to infection. In 2018, chickens kept on rural holdings also had a high prevalence of infection (p = 39.29%), while no case of infection was detected in any chicken kept on industrial farms. Chickens reared on rural holdings are more likely to be infected with WNV than chickens on industrial farms.3.2. Results of Spatiotemporal Analysis3.2.1. Cluster AnalysisThe global spatial autocorrelation analysis of the average annual incidence of WNV infection in South Banat District in 2018 suggested that a significant positive spatial autocorrelation existed. By calculating the global autocorrelation, we determined the relevance of attributes at the level of the entire study area, whereby a perimeter of 23.3 km indicates distances where spatial processes that promote clustering are most pronounced. As shown in Figure 8 panel (a), the test results indicate that a statistically significant grouping of clusters occurs in a diameter of 23.3 km (Global Moran’s I test: the first peak of z-score 2.40, p-value of 0.016 for a distance of 23.3 km). The results of the hot spot analysis and purely spatial cluster analysis showed a variation in the spatial distribution of WNV cases in South Banat County, with most high-risk clusters located nearby water streams (Figure 8). Red, orange, and light orange dots on panels (b), (c), and (d) represent clusters that are were located as “hot” spots, i.e., they form a group of clusters in which grouping is not the result of a random event but occurs as a result of environmental risk factors influence. The level of statistical significance of cluster grouping was graded in the interval of 90–99%. Yellow dots represent clusters whose grouping is the result of a random event and cases are not interconnected.3.2.2. Space–Time AggregationKulldorff’s spatial cluster analysis indicated that the cases of WNV were not randomly distributed in space and time in 2017, 2018, or 2019. A total of six significant purely spatial clusters (Table 1), one purely temporal cluster (Table 2), and two significant and one secondary spatio-temporal clusters were discovered (Table 3). The most likely spatial clusters were mainly found in the southwestern part and central part of South Banat District, nearby main water streams and dense vegetation (the Danube and Tamiš rivers, and Ponjavica Nature Park), including 18 settlements. For more information on spatial cluster analysis see Supplementary Material (Supplementary Figures S4–S9).4. DiscussionThis research revealed the epidemiological characteristics of WNV infection in the South Banat District. We have analyzed the changes occurring concerning the spatial, temporal, and spatiotemporal trends during the study period between 2017–2019, using the GIS spatial analysis technique, scan statistics, descriptive analysis of WNV case data, and statistical modeling.Our study confirmed that the most important spatial risk factors for WNV infection in South Banat District were the nearness of waters (water streams, standing waters, and the network of irrigation canals), and proximity to WNV-positive mosquito sites. In addition, proximity to wild bird habitats and the presence of dense vegetation cover were important risk factors. The graphical representation of the land cover of the South Banat District is shown in Supplementary Materials (Supplementary Figures S8 and S9). Thematic maps, created using the above-mentioned risk factors, identified WNV hot spot locations in all of the analyses. Areas that were identified as primary and secondary hot spots were mostly located nearby water streams, in the southwestern or central parts of the South Banat District.We found the main risk factors of the epidemic in humans in 2018 were the high environmental temperatures and the sharp decline of previous maximum water levels in July and August, i.e., the peak of the epidemic correlated with the lowest water levels of the Danube and Tamis rivers (monthly average minimum and monthly average maximum temperature: 17.8 °C and 31.7 °C respectively, and river water levels of the Danube and Tamis rivers of 236 mm and 238 mm, respectively) [49].However, although a similar pattern of the interrelationship of these risk factors with the peak of the epidemic was observed in 2017 and 2019, the number of registered human cases was significantly lower than in 2018. At the peak of the epidemics in August 2017 and 2019, only three and five cases of the disease were registered, while in 2018, 23 cases were registered. In terms of hydrology, 2017 and 2019 differed from 2018 in that in 2017 and 2019, no extended periods of maximum water levels were registered, preceding dry summer periods with minimum water levels; this was not the case in 2018.The results of negative binomial regression analysis show that temperature is a significant risk factor, i.e., a predictor of the appearance of mosquitoes infected with WNV. At the time of the onset of the disease in 2018, the average ambient and average ambient maximum temperature ranged between 20.84 °C to 23.35 °C and 25.96 °C to 28.26 °C, respectively. The highest incidence was observed when the two-week average ambient and average ambient maximum temperature were 26.73 °C and 26.73 °C, respectively. It is important to note that 2018 and 2019 were the warmest years in the history of meteorological measurements in Serbia. The prolongation of the warmest summer period and the warmer autumn period coincides with the early spring migration of birds, thus extending the time duration for infection of mosquitoes with WNV, which increases the chances of transmitting the disease to humans and animals [50]. High environmental temperature increases the risk of WNV transmission [51,52]. High environmental temperature promotes higher growth rates of mosquitoes and makes the extrinsic incubation period shorter; the gonotrophic cycle (the time required to produce eggs after a blood meal) is also shorter [53,54]. Based on negative binomial regression analysis, we found that variations in the number of registered WNV human cases can be partially explained and related to temperature variations and the increased number o WNV-positive mosquito pools. These two factors have proven to be good predictors of the growth in the number of infected people. On the other hand, the declining level of the Danube also shows a positive association with the increase in the number of patients and the number of registered WNV-positive mosquito pools.For more information on water levels of the Danube and Tamis rivers, and temperature regimes see Supplementary Materials (Supplementary Tables S1–S4 and Figures S1 and S2).Besides climate factors and analysis of seasonality, in this study, the risk factors were also evaluated at spatial scales, based on the primal assumption that the flight range of mosquitoes is crucial for transmission and mosquitoes are the main transmission force. The landscape of the South Banat District, such as vegetation, soils, wetlands, and topography, provides favorable conditions for both mosquito and wild bird multiplication. The high-risk zones include plenty of agricultural areas, farms, rural holdings, wetlands [55], and urban areas with inland marshes and small forest zones [52].An important element for performing either risk analysis or hot spot analysis is determining the radius of the grouping, i.e., the boundaries (perimeter) of the geographical area on which the grouping of entities into clusters is analyzed and which determines their significance. The hot spot analysis for the numbers of cases indicated that the disease is most prevalent in locations nearby water streams. By comparing the grouping of clusters, we noted a positive match between the clusters of WNV-positive mosquitoes and the clusters of registered human and animal cases of infection. The clusters of WNV-positive mosquitoes, which were classified by the KDE technique into the category of very high-density clusters, coincided with the clusters of WNV infection in humans and animals, which were also categorized as very high-density clusters or high-density clusters.Possible factors that might influence such a wide grouping (grouping radius) of statistically significant clusters of WNV occurrence in South Banat District are the intensity of prevailing winds and the abundance of running and standing waters. Winds have a great effect on mosquito dispersion. Kay and Farrow reported flight distances of 648 km for C. annulirostris as a result of wind blow [56]. Other authors have reported the results of similar research where substantial distances were identified, i.e., 200 km for Culex tritaeniorhynchus [57], 280 km for Anopheles pharoensis [58], 500 km for C. tritae-niorhynchus [59], 740 km for A. vexans [60], and even 850 km for Cx. pipiens pipiens [61]. Verdonschota and Besse-Lototskaya reported mean flight distances of 10.97 km for genus Culex [62].South Banat District is flat and characterized by a high frequency of winds. The highest frequency of wind occurrence is the southeast wind (košava) which occurs at 306 ‰, followed by the northwest wind with 255‰, while the lowest frequency of occurrence is the north wind 48‰ and the northeast 44‰. The prevailing southeast wind most often occurs in autumn 368‰, and least often in summer 196‰. The highest frequency of silences (Calme) is 143‰ in May and the lowest in November 51‰. As for the wind speed, the highest average annual speed for the area has the wind that blows from the east-southeast direction 3.3 m/s, and the lowest south-southwest with a speed of 1.7 m/s. The graphical representation of the dominant winds is shown in Supplementary Materials (Supplementary Figure S3 and Table S20).In the study area, a total of six significant purely spatial clusters, one purely temporal cluster, and two significant spatio-temporal and one secondary spatio-temporal clusters were discovered. In most situations, except in 2017, the human cases were preceded by the detection of WNV in animals and mosquitoes. In 2017, the detection of WNV-positive mosquitoes preceded the registration of the first cases of infection in humans by 55 days, while in 2018, this period was 9 days. In sentinel animals in 2017, contact with the virus was detected three months and 3 days after the registration of the first human case. However, in 2018, contact with the virus was detected 9 days before human infection, and in 2019, a full two months earlier.To better understand the history of WNV epidemics in the study area, we additionally used the data provided by ECDC [9] and analyzed the number of infected people in the designated geographical area in the period between 2012–2019. It was notable that after the sudden appearance of the disease in 2012 and the further increase of the epidemic in 2013, there was a significant decrease in the registered cases in the following year, as was the case in 2019. The numbers of human cases per year were as follows: n = 10 in 2012, n = 32 in 2013, n = 19 in 2014, n = 7 in 2015, n = 4 in 2016, n = 5 in 2017, n = 51 in 2018 and n = 12 in 2019. The mean of the WNV case numbers was 17.5 cases per year in the period. The differences of the cases from the mean were: −7.5 in 2012, 14.5 in 2013, 1.5 in 2014, −10.5 in 2015, −13.5 in 2016, −12.5 in 2017, 33.5 in 2018, and −5.5 in 2019. This decline in human cases could be partly explained if many people had already been in contact with the virus and earned immunity, which made them less susceptible to the virus [21] but this assumption needs additional research. Variations that have emerged over the last decade might also be explained if birds re-introduce WNV into South Banat District from year to year.Our research indicated that in creating a surveillance model for WNF, the primary concern is to consider the behavior and the requirements of every element in the disease transmission chain. Climate factors might be the main determinant of WNV distribution and predictors of endemicity, but in other situations, climate factors are not sufficient to explain the observed distribution of WNV cases. Precipitation and water levels of main rivers have an important influence on the mosquito abundance on main rivers, local water streams, and especially standing waters. Water abundance also greatly affects the habitats of wild birds, which are important for maintaining the virus in nature. The epidemic process is conditioned by the action of several necessary factors in the disease. Winds were identified as a likely risk factor, taking into account wind effect on vector dispersion. However, each of these factors is not enough to lead to the disease on its own, so each risk assessment and prediction is conditioned by observing and researching all these risk factors together. Taking into account the substantial presence of domestic animals raised in extensive production mode, our study leads us to the conclusion that the WNV surveillance system, besides horses and chickens, should be augmented by surveilling the virus in domestic animals kept on rural holdings, particularly cattle, pigs, and chickens. Since these animals live in conditions of very low levels of biosecurity measures, they are significantly more exposed to the vectors and WNV than domestic animals on industrial farms. Animals on rural holdings are relatively easy to trace and sample and do not have high surveillance costs. WNV, despite the decline in the number of human cases in 2019, remains a threat to the human population in the South Banat District. For the disease to be successfully monitored and detected in time, it is necessary to conduct constant monitoring of the presence of the virus in natural reservoirs and sentinel species, monitor and analyze climatological risk factors and be especially focused on indicators of the presence of the virus in nature.5. ConclusionsWe found the main risk factors of the epidemic in 2018 were the high environmental temperatures and the sharp decline of previous maximum water levels in July and August. The temperature is a significant risk factor and predictor of WNV infection in humans and mosquitoes. We found that variations in the number of registered WNV human cases can be partially explained by temperature variations and associated with the increased number of WNV-positive mosquito pools, which proved to be very significant in this study and leads us to the conclusion that it is important, as part of the surveillance program in the South Banat District, to place special emphasis on monitoring the presence of WNV in mosquitoes. It is important to emphasize that almost all cases of infection in humans were preceded by the detection of WNV-positive mosquitoes. Using statistical models, such as regression models, it is possible to quantify risk factors, especially climatic factors, and identify those that are not significant as indicators to be monitored. The WNV high-risk zones include plenty of agricultural areas, farms, rural holdings, wetlands, and urban areas with inland marshes and small forest zones. The hot spot analysis for the numbers of cases indicated that the disease is most prevalent in locations nearby water streams. By comparing the grouping of clusters, we noted a positive match between the clusters of WNV-positive mosquitoes and the clusters of registered human and animal cases of infection. In most situations, the human cases were preceded by the detection of WNV in animals and mosquitoes. The epidemic process is conditioned by the action of several necessary factors in the disease. Winds were identified as a likely risk factor, taking into account wind effect on vector dispersion. However, each of these factors is not enough to lead to the disease on its own, so each risk assessment and prediction is conditioned by observing and researching all these risk factors together. Taking into account the substantial presence of domestic animals raised in an extensive production way, our study leads us to the conclusion that the WNV surveillance system, besides horses and chickens, should be augmented by surveilling the virus in domestic animals kept on rural holdings, particularly cattle, pigs, and chickens. For the disease to be detected in time, it is necessary to conduct constant monitoring of the presence of the virus in natural reservoirs and sentinel species. Climatic risk factors should be constantly monitored and analyzed.

animals : an open access journal from mdpi

10.3390/ani11040922

PMC8064100

Reproductive management of sheep for autumnal lambing often require induction and synchronization of oestrus and ovulation, either for natural mating or artificial insemination, by the use of pharmacological treatments. Such treatments are mostly based on the administration of progesterone followed by a single intramuscular dose of equine chorionic gonadotrophin (eCG) at progesterone withdrawal. However, repeated eCG treatments in consecutive mating seasons can result in the outbreak of resistance with a rise of anti-eCG antibodies. Furthermore, the future use and availability of eCG appears to be strongly challenged by the highly active animal-rights movement because the hormone is obtained from pregnant mares. The present study demonstrated that the administration of glycerol-based formulations to milked ewes is a valid alternative to eCG treatment in reproductive management protocols based on the induction of ovulation with progesterone-releasing devices at the end of anoestrus period. The glucogenic treatment administration to late lactation dairy ewes at the end of the anoestrus period improved their metabolism without harming animal production or animal welfare, thus promoting a sustainable reproductive management of dairy sheep.

This study investigated whether the administration of equine chorionic gonadotrophin (eCG) in a protocol to induce and synchronize ovulations before mating could be replaced by the administration of glycerol-based formulations in milked ewes at the end of their seasonal anoestrus. Forty-eight late-lactation dairy ewes of the Sarda breed were synchronized using sponges impregnated with progestogen and then joined with fertile rams (day (D) 0, ram introduction). From D−4 to D−1, the ewes received by gavage either 100 mL of a glucogenic mixture (70% glycerol, 20% propylene glycol and 10% water; GLU group; n = 24) or 100 mL of water (GON group; n = 24) twice daily. Moreover, on the day of sponge withdrawal (D−1), GON ewes received 200 IU of eCG. There were no differences in reproductive performances between groups. GLU ewes showed higher glycemia (p < 0.001), insulinemia (p < 0.05), plasma glycerol (p < 0.001), triglycerides (p < 0.001) and lower cholesterol (p < 0.001), non-esterified fatty acids (NEFA; p < 0.05) and urea (p < 0.001). Plasma osmolality was higher in GLU but only 4 h after dosing (p < 0.001). Milk yield and milk composition were not affected by the treatments with exception of milk glycerol (p < 0.001) and milk urea (p < 0.001), which were higher and lower in GLU than GON ewes, respectively. In conclusion, the administration of the glucogenic mixture to late lactation dairy ewes at the end of anoestrus period resulted in reproductive responses as good as the ones obtained by the eCG treatment, suggesting that the objective of a sustainable reproductive management of dairy sheep can be successfully pursued.

1. IntroductionThe reproductive management of sheep for out-of-season lambing may require induction and synchronization of oestrus and ovulation, either for natural mating or artificial insemination by means of pharmacological treatments, which are based mostly on the administration of progesterone or its analogues for mimicking the activity of the corpus luteum. Usually, a single intramuscular dose of equine chorionic gonadotrophin (eCG) is injected at progesterone withdrawal to induce oestrus and ovulation, to reduce variability in the onset of oestrus among treated animals and to increase ovulation as a consequence of a greater follicular development [1]. However, repeated eCG treatments in consecutive mating seasons can result in the outbreak of resistance with a rise of anti-eCG antibodies, lower responsiveness to the drug and hence lower reproductive performance [2]. Hence, there is a need for alternative synchronization protocols without eCG [1,3].We have recently reported that the administration of glucogenic glycerol-based formulation to Sarda ewes was able to elicit an ovulatory response which did not differ from that of eCG-treated ewes dosed with water [4]. However, the study was conducted during the natural breeding season using a PGF-based synchronization protocol without the administration of progesterone. The short-term use of glucogenic glycerol-based formulations in the ewe’s diet led to a positive metabolic chain of events, starting with a rise in glucose plasma levels [3] which, besides acting as metabolic fuel, also play a role as signalling molecule to stimulate folliculogenesis [1] and determine oocyte quality [5,6]. Previous studies also demonstrated that a single dose of glycerol given at the moment of progesterone-releasing device removal can increase ovulation rates [7,8]. The rise of glucose plasma levels leads in turn to increased circulating concentrations of insulin [3,9,10] and insulin-like growth factor (IGF-1) [10], which often act as signalling molecules between metabolism to fertility [11] by interacting with the reproductive axis at the level of the central nervous system as well as in the follicle. The treatment also lowers the plasma levels of non-esterified fatty acids (NEFA) [3,12,13] and urea [3]. High levels of NEFA and urea have been associated with lower fertility [14,15,16,17] due a change of their concentration in follicular fluid [18,19,20,21] or modification of the uterine environment [15,17]. Thus, the administration of glucogenic glycerol-based formulations at the end of luteal phase prior to ram introduction could stimulate follicular development, leading to ovulation and pregnancy rates comparable to those obtained by a hormonal treatment.However, a previous study showed that the short-term administration of a high dose of glucogenic mixture (23% on dry matter intake basis) to milked ewes resulted in a temporary reduction of milk yield and milk lactose content and an increase of milk protein and casein. Therefore, considering both these changes in milk yield and composition and the recently reported potential side effects of glycerol on red blood cell indices and plasma osmolality [4], the dose given should be carefully set. In a previous study we reported that medium doses of glucogenic mixtures (≈12% on dry matter intake basis) should be preferred for flushing dairy ewes, as they proved to be effective at metabolic level without causing alterations in RBC indices or possibly in their functionality [4].The present study arose from this finding. Keeping in mind the above premises, it was aimed at (1) investigating whether the administration of eCG in a standard synchronization protocol could be replaced by the administration of glycerol-based formulations in milked ewes at the end of their seasonal anoestrus; (2) monitoring the changes in lipid profile, liver and kidney function consequent to glycerol administration; (3) assessing whether the dose set in a previous study; and [4] preserving milk yield and composition. Moreover, given the high diffusibility of glycerol across plasma membranes, the concentrations of milk glycerol were also determined.2. Materials and Methods2.1. Animals and TreatmentsThe experiment was run during May–June 2019 at Bonassai research station of Agris Sardegna (40° N, 8° E, 32 m above sea level) at the end of the anoestrus season described for Sarda breed at this latitude. Forty-eight adult (mean ± s.d., 4.5 ± 1.4 years) and multiparous lactating Sarda dairy ewes were used. They were selected from the experimental flock after two consecutive ovarian ultrasound scans performed 8 days apart confirmed the absence of corpora lutea suggesting that ewes were in their seasonal anoestrus period. On the day of the second ultrasound examination, the ewes were divided into two experimental groups homogeneous for body weight (BW), body condition score (BCS), days in milk (DIM) and milk yield (MY; g/day): glucogenic treated group (GLU, n = 24; BW 44.7 ± 3.3 Kg; BCS 2.53 ± 0.08, DIM 177 ± 6.1; MY 806 ± 240 g); gonadotrophin treated group (GON, n = 24; BW 44.7 ± 4.1 Kg, BCS 2.53 ± 0.08; DIM 176.5 ± 5.6; MY 826 ± 222 g). Each group was further divided in two subgroups, which were used as replicates.The experimental protocol is depicted in Figure 1, where day 0 (D0) is the day of ram introduction. In brief, oestrus synchronization was induced in all the animals with the insertion (D−13) of one intravaginal progestagen-impregnated sponge (45 mg fluorogestone acetate, FGA, Chronogest; Intervet International, Boxmeer, the Netherlands) which remained in situ for 11 d. On the day of sponge withdrawal (D−2), GON ewes received an intramuscular injection of 200 IU of eCG (Folligon, MSD Animal Health Srl, Segrate, Italy). From D0 to D2, all ewes were joined with fertile rams (n = 10) fitted with crayon markers.From D−4 to D−1, the GLU group received 100 mL of a glucogenic mixture orally twice daily at 0800 h and at 1900 h. The dose of the glucogenic formulation was chosen on the basis of the results of a previous study that compared 8 different doses of glucogenic formulation [4].The glucogenic formulation administered contained 70% glycerol, 20% propylene glycol and 10% water. Glycerol and propylene glycol had a purity grade of 99.5–100% and complied with EU Reg. 231/2012 for food additives (E422 and E1520 for glycerol and propylene glycol, respectively; Farmalabor srl, Assago, Milano, Italy). The amount of energy supplied was 0.6 NEL Mcal/d, corresponding to ca. 12% of the expected intake of offered diet on DM basis. Gonadotrophin treated ewes received 100 mL of water twice daily simultaneously to treatment administration. Both the glucogenic formulation and the water were administered using a drench gun.From D−13 to D+2 (i.e., throughout the treatment periods), the ewes of both groups were kept indoors in separate pens and were machine-milked twice daily at 0700 h and 1500 h.Indoor daily feeding for both groups consisted of 600 g/head of a commercial pelleted feed divided into two equal meals at milkings fed individually in the milking parlour: 1000 g/head of dehydrated lucerne hay and 1000 g/head of a ryegrass hay fed in different troughs in two meals after each daily milking. On the blood sampling day, the morning meal based on the pelleted concentrate was administered immediately after the first blood sampling. Water and mineral-vitamin blocks were available ad libitum.2.2. Feedstuff CompositionSamples of the hays and the concentrate were collected every week and bulked to form two samples per feed for subsequent determinations. All these samples were oven-dried at 65 °C and subsequently ground to pass a 1 mm screen to determine the content of dry matter (DM, oven drying at 100 °C overnight), ash (ID#942.05), ether extract (EE, ID# 920.39) and CP (N × 6.25, [ID# 988.05]) according to AOAC [22]. NDF, acid detergent fiber on an ash-free basis (ADF) and acid detergent lignin (ADL) were also determined [23]. Finally, the in vitro dry matter digestibility (IVDMD) of all feed was measured by the pepsin–cellulase method [24] whereas the starch content of the concentrate was assessed by polarimetry. Net Energy (NEL, Mcal/kg DM) and metabolic protein content of feedstuffs and diet were calculated using the equations published by Cannas et al. [25]. Diet formulation was done using Small Ruminant Nutrition System package. Feedstuff chemical composition is shown in Table 1. Estimated average NEL were 1.60 Mcal/kg DM, (concentrate), 1.00 Mcal/kg DM (dehydrated lucerne), 0.80 Mcal/kg DM (grass hay), and 2.70 Mcal/kg DM (glucogenic mixture).2.3. Blood SamplingsBefore starting the glucogenic treatment (D−5 of the experimental period), blood samples were collected at fasting (0800 h). In addition, on the third day of glucogenic treatment administration (D−2 of the experimental period), two blood samples were collected, one at fasting immediately before the morning administration of the glucogenic formulation at (0800 h) and one four hours later (1200 h). Finally, three days after the end of the glucogenic treatment a final blood sample was collected at fasting (D2; 0800 h).At each sampling, from each ewe, two blood samples were collected: one using 2-mL vacuum collection tubes with glycolytic inhibitor (5.0 mg sodium fluoride, 4.0 mg potassium oxalate—Vacutainer Systems Europe; Becton Dickinson, Le Pont-de-Claix, France) for glucose assay; one using 2.0 mL vacuum collection whole blood tube with spray-coated K2EDTA (Vacutainer Systems Europe; Becton Dickinson, Le Pont-de-Claix, France) for metabolites and insulin quantification. Immediately after recovery, samples were cooled to 4 °C. Blood samples were centrifuged at 1500× g for 15 min at 4 °C degrees. Individual plasma was removed and stored in vials at −20 °C until assayed.2.4. Plasma Osmolality DeterminationPlasma osmolality (Osm/kg) was measured using a freezing point osmometer (Osmomat 030, Gonotec, Berlin, Germany).2.5. MetabolitesPlasma samples were measured in duplicate. Glycerol concentration was measured in a single assay by colorimetric method using a commercial free glycerol assay kit (Cell Biolabs, Inc., San Diego, CA, USA), with glycerol standards in the concentration range of 0–400 μM. The kit measures free, endogenous glycerol by a coupled enzymatic reaction system. The glycerol was phosphorylated and oxidized to produce hydrogen peroxide, which reacted with the kit’s colorimetric probe (absorbance maxima of 570 nm). The analytical detection limit was 5 μM.Glucose, NEFA, Urea, total cholesterol and triglycerides, creatinine, total protein, albumin, alanine aminotransferase (ALT) and aspartate aminotransferase (AST), were measured using commercial kits (Real Time Diagnostic Systems kits) and BS-200 Mindray (Adaltis, Milan, Italy) clinical chemistry analyzer. We used Serum I Normal (Wako) and Serum II Abnormal (Wako) as a multicontrol for each measured parameter. Glucose concentrations were determined in a single assay by the liquid enzymatic colorimetric method (GOD-POD) with a glucose standard of 100 mg/dL for calibration. Intra-assay CV value was 1.1%. NEFA, Urea, total cholesterol and triglyceride concentrations were measured in multiple assays by the enzymatic endpoint method with different standards for calibration: 1 mmol/L, 50 mg/dL, 200 mg/dL and 2.28 mol/L for NEFA, urea, total cholesterol and triglycerides respectively. The intra-assay and interassay CV values were 1.07 and 0.98% (for NEFA), 1.7 and 1.6% (for urea), 0.95 and 1.24% (for total cholesterol) and 0.99 and 1.05% (for triglycerides) respectively. Creatinine was measured in an assay by the increasing kinetics method with a standard of 2 mg/dL for calibration; its intra-assay and interassay CV values were 1.07 and 2.55%, respectively. Total protein and albumin concentrations were measured in multiple assays by spectrophotometric endpoint methods, with the following standards for calibration: 6 g/dL and 46.2 g/L of bovine albumin for total protein and albumin assays respectively. Their intra-assay and interassay CV values were 1.94 and 1.35% for total protein and 1.36 and 1.52 for albumin, respectively. ALT and AST concentrations were measured in multiple assays by kinetics UV method using an ALT standard solution of 97.65 U/L and AST standard of 102 U/L for calibration. Their intra-assay and interassay CV values were 2.07 and 2.24% (for ALT) and 2.42 and 2.29% (for AST) respectively.2.6. InsulinELISA assays were performed using the Personal Lab Adaltis (Adaltissrl, Rome, Italy), a tool that performs automated ELISA protocols. Insulin concentration was measured in duplicate using a commercial ovine insulin ELISA Kit (Mercodia developing diagnostics, Marburg, Germany) which is a solid-phase ELISA based on the direct sandwich technique. The kit is calibrated against an in-house reference preparation of ovine insulin, and it has been previously used for insulin determination in ovine plasma [3,4]. The mean ovine insulin concentrations of the six reference solutions were 0, 0.05, 0.15, 0.5, 1.5, and 3 mg/L. The recovery on addition was 94%–114% (mean 103%). The analytical sensitivity was 0.025 mg/L and the intra-assay and interassay CV values were <7%.2.7. Intake, Body Weight, Body Condition Score, Milk Yield and Milk CompositionFeed intake at group level was measured by weighing the feed offered and the corresponding orts after each meal for the concentrate and after 24 h for the hays. Body weight was measured before the morning meal using an electronic scale before (D−13) and after the glucogenic treatment (D2). On the same days, body condition score ranging from 1 (extremely thin) to 5 (obese) was estimated by two trained evaluators with an approximation of 0.25 BCS units [26]. Their scores were averaged prior to data analysis.Milk yield was measured by weighing the production of each ewe at two consecutive milkings at 0700 h and at 1500 h on two occasions before (D−10 evening and D−9 morning) and during (D−3 evening and D−2 morning) glucogenic treatment. Milk samples were collected and milk composition assayed on composite samples for fat, protein, casein and lactose using the Fourier-transformed infrared method (Milkoscan FT+, Foss Electric, Hillerød, Denmark) and milk urea concentration using an enzymatic colorimetric assay (Chem Spec 150; Bentley Instruments Inc., Chaska, MN, USA). The glycerol concentration in the milk was measured as above described for blood.2.8. Ultrasound ScanningOvarian ultrasound scanning was performed in the pre-experimental period and on D10 of the experimental period to determine ovulation rates by counting the corpora lutea present in each ovary. In both cases, the ovaries were examined by transrectal ultrasonography using a real-time, B-mode scanner (Aloka SSD 500; Aloka Co. Ltd., Tokyo, Japan) fitted to a 7.5 MHz linear-array probe (82 mm prostate transducer UST-660-7.5, Aloka Co.). In each observation, each ovary was scanned several times from different angles to determine the presence and the number of all corpora lutea. Pregnancy diagnosis was performed on D40 using transrectal ultrasonography (Aloka SSD 500, fitted to 82 mm prostate transducer UST-660-7.5, Aloka Co.). Pregnant sheep displayed enlargement of uterine horns and an embryo heartbeat was evident.2.9. Statistical AnalysesDifferences between groups in body weight and BCS and their changes during the glucogenic-treatment period were analysed by a mono-factorial general linear model (GLM).Mean plasma circulating concentration of metabolites and osmolality on day −2 (during treatment period), milk glycerol, milk yield and composition were analysed by GLM with time, treatment groups and their first-order interactions as fixed effects using Minitab 17 Statistical Software (2010, Minitab, Inc., State College, PA, USA). As post-hoc test, Tukey’s test was used to highlight differences within and between groups.A chi-square analysis was used to highlight the differences between groups for ovulation and pregnancy rates.Results are expressed as mean values (mean ± SE) and the differences were considered to be statistically significant at p < 0.05.3. ResultsAt the end of the treatment period, no differences were observed in body condition between the experimental groups (Supplementary Table S1). During the treatment period, the concentrate and the lucerne dehydrated hay were almost completely consumed in both groups (525 and 523 g DM/head, for the concentrate and 797, 795 g DM/head for the lucerne, in GON and GLU groups, respectively), with no evidence of abrupt changes during treatment period. In contrast, the intake of ryegrass hay (537 and 499 g DM/head) tended to decrease more in GLU than GON group in the 4 days of glucogenic administration as compared with the previous 4 days (−9% vs. −22%, in GON and GLU, respectively, one-tail student t test for the difference between GON and GLU after the beginning of the treatment, r = 8, T = 2.12, p = 0.07).Regarding the reproductive performances, no significant differences in ovulation and pregnancy rates were found between glycerol- and gonadotrophin-treated ewes (Table 2).The metabolic status of the GLU group was modified compared to the GON group, showing a more favourable energy balance. As expected, the glucogenic administration led to a rise in glycerol plasma levels on GLU group (p < 0.001; Table 3), and consequently of plasma osmolality (p < 0.01; Table 3) as shown in a previous study [4]. On the third day of glycerol-based formulation administration (D−2), GLU group showed higher glycemia 4 h after the morning treatment (p < 0.001; Table 3), and insulinemia at fasting and after treatment administration (p < 0.05; Table 3) compared to the GON group. Moreover, on the same day, these effects reflected on lower plasma levels of NEFA at fasting (p < 0.05; Table 3) and lower urea mean circulating concentrations (p < 0.001; Table 3). This metabolic shift included a significant rise in the circulating concentration of triglycerides (Table 3) in the GLU group, whose values were 15-fold higher than those found in gonadotrophin treated ewes (p < 0.001).Milk yield was not affected by the nutritional treatment (Table 4), while milk composition was modified in the urea concentration, which reflected the changes observed in the plasma (Table 4). It is noteworthy that glycerol proved to reach the mammary gland, as demonstrated by the significant higher glycerol milk concentration found in the GLU group compared to the GON one (6.480 ± 1.050 and 0.497 ± 0.193 mg/mL, in GON and GLU groups, respectively; p < 0.001).4. DiscussionResearch efforts are focusing on finding alternative treatments to eCG administration in protocols to induce oestrus and ovulation in small ruminants [4,27,28] considering the concerns linked to its production for brood mare health and welfare [29]. The reduction in the use of synthetic hormones in livestock farming would also meet consumer demand for “natural, green, and clean” methods [30].The present study demonstrated that ovulation and pregnancy rates did not differ in glucogenic- and eCG-treated ewes. Thus, the administration of a glycerol-based glucogenic formulation can replace hormonal stimulation with eCG in a standard protocol with progesterone-releasing devices for the induction of oestrus and ovulation in milked Sarda ewes at the end of their seasonal anoestrus. This finding confirms previous findings in dry ewes subjected to a prostaglandin-based synchronization period during the breeding season [4]. However, pregnancy rates were below expectations considering that rates as high as 70% are usually achieved [31]. A slightly better pregnancy rate (0.60) was achieved in a previous experiment in which a double dose of glucogenic formulation was administered to milked ewes [3], but in that study the ewes were in a better condition at mating (BCS = 2.7 vs. 2.5). Moreover, in this study as well as in the previous one, the sheep could have suffered from heat stress. Actually, just after mating there were at least two days in June with average THI > 72 (severe heat stress according to, Marai et al. [32]) with more than 8 h/daily above 30 °C. These conditions can favour embryo mortality [32]. Preliminary results of our laboratories suggest that high doses of glucogenic (i.e., the double of the one used in this experiment) can increase eye and rectal temperature by almost 1 C° in ewes experiencing severe heat stress [33]. This aspect warrants further investigation in the light of the increasing risks of dramatic heat waves as a consequence of the climate change.Glucogenic formulations can boost ovarian activity by creating a suitable systemic and intra-follicular metabolic milieu for the promotion of ovarian function [10]. The dose of the glycerol-based formulation used in the present study had been previously designed [4] to ensure this positive switch in ewe metabolic status without causing any significant alteration in red blood cell (RBC) indices. Higher doses (from 18.2 to 27.4% on DM basis) cause indeed significant changes in RBC indices consequent to glycerol diffusion across the RBC membrane, which may be associated with impaired RBC functionality [34,35]. However, our previous study [4] was carried out in dry ewes during their breeding season. The present study thus confirms that this dose was able to elicit a comparable metabolic status also in milked ewes at the end of their seasonal anoestrus although during lactation, lipolysis and proteolysis are increased to meet the energetic requirements of the mammary gland.Moreover, the more complete metabolic status assessment performed in the present study revealed that the administration of the glycerol-based formulation, while not significantly affecting liver and kidney functions, led to a significant rise in circulating triglyceride concentrations. Under negative energy balance (NEB), which is frequently found in ruminants in the peripartum period or during lactation [36,37], NEFAs were released from the triglycerides stored in adipose tissue by lipases that split them into NEFA and glycerol following a reduction in blood glycemia. Once in the hepatocytes, NEFAs enter the mitochondria and underwent β-oxidation to produce energy as an alternative energy source to glucose. However, when a positive shift in energy balance occurred and glycemia increased, there was a subsequent rise in insulinemia-stimulated glucose uptake and glycolysis by the hepatocytes. This reaction produced glycerol-3-phosphate, which acted as a substrate for the re-esterification of NEFAs to triglycerides [38]. Thus, the significant rise in circulating triglyceride concentrations, together with the significant decrease in circulating NEFA concentrations, in glycerol-treated ewes confirmed that the administration of the glycerol-based formulation led to a switch from fat oxidation to carbohydrate utilization [39].The significant rise in circulating triglyceride concentrations consequent to the administration of the glycerol-based formulation may have also been triggered by the increase in the plasma glycerol concentrations. When the circulating glycerol concentration exceeded its baseline level, its inclusion in hepatic gluconeogenesis and triglyceride synthesis increased [40,41]. In the GLU group glycerol metabolism shifted indeed to the liver synthesis of fatty acids and then triglycerides. In humans, treatments with glycerol have been associated with an increase in hypertriglyceridemia [42]. A similar effect was reported in rats [43]. On the other hand, the inclusion of crude glycerin in the diets for growing lambs caused no significant increase in triglyceride plasma concentrations [44,45]. This difference may be related to differences in metabolism between growing and lactating sheep and to the significant decrease in dry matter intake found in lambs following glycerol inclusion in the diet [44,45]. However, in the present study, mean cholesterol concentrations were lower in glucogenic-treated compared to eCG-treated ewes. Similar results (high triglyceride and low cholesterol) were found by Mazur et al. [46] in well-fed compared to underfed pregnant ewes. In sheep, as well as in other ruminants the level of lipoprotein in blood is usually lower than in monogastric mammals, with a lower secretion of VLDL-triglyceride in ruminant liver and hence higher risk of steatosis at the end of pregnancy [46]. However, in the case of this study, the rise was detectable only in the post-treatment blood sampling. Moreover, since the glucogenic treatment was limited to 4 days, this period was probably too short to cause the above-mentioned lipid alterations.The effect of high triglyceride concentrations on ovarian activity was poorly explored. However, triglycerides did not pass through the follicular membrane [47] and their intracellular concentration was mainly the result of local metabolic processes [19]. Therefore, an acute rise in triglycerides should not affect follicular development.The lower dose of glucogenic mixture used in the present study caused no significant reduction in milk production and lactose contents, thus backing the opportunity of reducing the dose administered. In contrast, the significant decrease in milk urea in glycerol-treated compared to eCG-treated ewes reflected the positive change in the metabolic status already pointed out. Moreover, high milk urea is a marker of reduced fertility in lactating cows [48] and ewes [49]. The significant higher milk glycerol concentration consequent to the administration of the glycerol-based formulation suggests that glycerol from the bloodstream can pass through the mammary gland. The presence of glycerol in milk could affect some physical and chemical characteristic [50]. Goats orally treated at increasing levels with bi-distilled glycerol [51] showed that glucogenic treatment could affect cheese softness. When the treatment is made on few animals (individual cows in the post-partum period) the quantity of glycerol on bulk milk can be negligible. However, in small ruminants the effect on milk quality should be taken into account when the treatment is administered to the entire flock during the mating season, particularly when higher doses of glycerol are given or the administration lasts for a longer time.5. ConclusionsIn conclusion, the present study demonstrated that the administration of glycerol-based formulations to milked ewes can be a valid alternative to eCG treatment in reproductive management protocols based on the induction of ovulation with progesterone-releasing devices at the end of anoestrus period at our latitudes. This finding confirmed previous results in dry ewes submitted to the same treatment during the breeding season, thus suggesting that the objective of meeting consumer demand of natural “green, clean and ethical products” [30] can be successfully pursued. However, given the diffusion of glycerol in the mammary gland, further investigations are needed to study the effects of the treatment on milk composition and the possible influence on cheesemaking. Moreover, more information should be gathered on the effects of glucogenic dosing when mating occurs under extreme heat conditions.

animals : an open access journal from mdpi

10.3390/ani11051282

PMC8146694

The color of fresh pork is a very important feature, both for the meat industry and especially for direct consumers. Consumers often relate color to the freshness and quality of pork. Quality evaluation of fresh pork intended for sale or processing seems to be a priority task for meat plants due to the expectations of modern consumers. Meat can be characterized as a mixture of different chemical components, each of which contributes to its quality either independently or in combination with other ingredients. The complexity of muscle structure and the tasks performed by individual muscles may contribute to uneven color distribution. The presented results are an important contribution to the rapid and precise instrumental evaluation of pH and color. The information may be particularly important for meat plants, which, by introducing such an evaluation, allow consumers to additionally verify the product.

The aim of this study was to determine the effect of 30 min bloom time and the type of muscle on pH and color parameters together with the possibility of estimating these measurements. The research material consisted of 270 samples from 6 muscle types: LD—Longissimus dorsi, LL—Longissimus lumborum, IL—Iliacus, SEM—Semimembranosus, CT—Cutaneous trunci, LTD—Latissimus dorsi. Measurements included pH and color of fresh pork at 0 min, and after 30 min bloom time. Bloom time influenced all analyzed parameters, although to a varying effect, depending on the muscle type. The lowest pH values were noted for dorsal-located muscles (LD, LL), then in the ham area (IL, SEM), and the highest values of the location on the side surface of the carcass (CT, LTD). The large increase in the proportion of L* and a* was observed for CT muscle (20–30%, the highest of all observed) and LTD (20–25%); for LD and LL the largest growth changes were observed for parameters b* (15–20%) and H* (20–30%). The lowest number of strong correlations was noted for LD and CT muscles, and the largest for SEM. A very good fit (R2 > 0.90) of regression equations was achieved in 7 cases. The presented results are an important contribution to the rapid and precise instrumental evaluation of pH and color.

1. IntroductionThe color of fresh pork is a very important feature, both for the meat industry and especially for direct consumers. Consumers often relate color to the freshness and quality of pork [1].However, the evaluation of color by consumers is usually based on visual evaluation with the naked eye; while for research, scientific or industrial needs, a precise instrumental evaluation is much more significant. It is worth remembering that the evaluation of quality by consumers has been proven to be not objective, burdened with an extremely large number of external and internal factors shaping the personal perception [2,3,4].The complexity of muscle structure and the tasks performed by individual muscles may contribute to uneven color distribution [5]. Pork half-carcass consists of approximately 100 striated muscles, and five of them make up almost a third of the musculature weight of the carcass [6]. LD—Longissimus dorsi, LL—Longissimus lumborum and SEM—Semimembranosus muscles have been evaluated most often, whereas muscles from the abdominal location have been subjected to evaluation much less frequently. However, due to the increasing commercial value of the belly, this is an element that should be taken into particular account [7]. The meat industry is interested in all the muscles but especially those included in the most valuable cuts [8]. Furthermore, anatomical and spatial imaging of changes gives an illustrative picture of the whole carcass.It is considered that after 24 h of cooling down, changes in color are completed [9], although bloom time is essential for precise measurement methodologies [10]. Bloom time can have a major impact on fresh instrumental color readings because the oxidation of myoglobin pigment in the muscle is responsible for creating a red color and does not occur immediately [11]. Blooming takes place on muscle’s cut surface during the action of the air and allows the reaction of reduced purple myoglobin and hemoglobin on the muscle surface with oxygen to transform it into bright red oxygenated forms [12,13]. Inactivation of oxygen-absorbing enzymes gradually reduces the rate of change in color [14]. When the meat pigment metmyoglobin (MetMb) is oxidized, it causes a grayish-brown color to appear on the surface of the meat. In such situations, consumers judge the meat as stale and do not purchase it [15].Research indicates that bloom time for meat should range from 20 min to 1–2 h [10,16,17], although in the case of pork, 30 min is deemed to be sufficient [12]. However, studies on the blooming time effect on pork were initially rare; most research activity in this area was conducted in 2001–2008. Even so, reviewing the research results brought ambiguous conclusions [5]. It seems that constant changes in the population of slaughtered pigs with regard to genetic and environmental factors causes the need to conduct ongoing research on this subject. Hence, the need to accurately evaluate the variability of pH and muscle color under industrial conditions.The aim of this study was to determine the effect of 30 min bloom time and the type of muscle on pH and color parameters as well as to explore the possibility of predicting this effect.2. Materials and Methods2.1. Design of the StudyThe research material consisted of: LD—Longissimus dorsi, LL—Longissimus lumborum, IL—Iliacus, SEM—Semimembranosus, CT—Cutaneous trunci and LTD—Latissimus dorsi, obtained in the same way from the same pig carcasses.The animals were transported in special livestock transport vehicles over a distance of max. 35 km. The animals were slaughtered at the age of 6.5–7 months and at a body weight of approx. 110 kg in a meat plant in Wielkopolskie province, using the electric stunning method. Pre-slaughter and slaughter time were similar for all the animals (the time of rest and fasting was about 4 h at a temperature of about 17 °C). The carcasses were bled, separated along the center line and deprived of tongue, bristle, hooves, genital organs, perirenal fat, kidneys, diaphragm, eyes, middle ear, brain and spinal cord.The muscles intended for further analysis were cut from right carcasses after 24 h of cooling, according to the production methodology in force at the slaughterhouse. The bloom time samples were collected immediately after dissection of the muscle in question, cut perpendicularly to the width of the muscle. The samples for testing were vacuum-packed and transported in stable cooling conditions directly to the meat evaluation laboratory at the Institute of Animal Science, Wroclaw University of Environmental and Life Sciences, where they were then subjected to further analysis.The cooling of carcasses was performed first by holding them for 2 h at −15 to −8 °C, followed by storage at 0–2 °C (8–12 h) until carcass temperature dropped to 5 °C. After 24 h of cooling, the abovementioned muscles were cut from the same right half-carcasses and marked. In total, 270 samples were obtained.2.2. Meat Quality AssessmentBloom time was determined using measurements at 0 and 30 min. The 0 min measurement was classified as the measurement of fresh meat; 30 min bloom time took place in dark and cool conditions (4 °C).pH measurements were conducted for each muscle location immediately after cross-division of the element with a Testo 205 pH meter with gel electrolyte (manufactured by Testo Sp. z o. o., Warsaw, Poland). Instrumental color evaluations were performed with a Minolta Chroma CR 400 device (Konica, Osaka, Japan) with an 11 mm diameter aperture, D65 illuminant, calibrated against a white tile. The measurements were made three times for each muscle. During instrumental color evaluation, the following parameters were determined: CIE—Commission Internationale de l’Eclairage L*, a* and b*—values representing lightness, redness and yellowness, respectively. Additionally, Chroma (C*) and Hue angle (H*) indicators were determined. Chroma described the color saturation in the CIE L*a*b* space and was calculated according to the formula C* = (a*2 + b*2)1/2. Hue angle determined the immaculateness of color and was calculated as H* = (tan−1 b*/a*) [18]. The percentage change in color over the 30 min bloom time period was determined using the formulas developed by [14].2.3. Statistical AnalysisThe data analysis was based on factor analysis in which two main effects of muscle and bloom time were observed. Additionally, differences between muscle samples at 0 and 30 min bloom time were separately compared. Correlation coefficients between qualitative parameters within each location were calculated using r-Pearson correlation. Models of regression equations estimating particular pH and color parameters at 30 min bloom time compared to measurements at 0 min were based on correlations with the dependent variable by the stepwise method to obtain the lowest estimation error. The accuracy of the estimation was determined by means of prediction error and variable matching. F, R2 determination coefficient, p-value and estimation error were used to describe the equations.3. ResultsThe changes in pH depending on the bloom time (0 and 30 min) are presented in Figure 1. All pH0min values were observed on the level of correct values for RFN pork (meat of normal quality—reddish-pink, firm, normal, non-exudative) after 24 h from slaughter. However, differences between the analyzed individual muscles were noted. Furthermore, the differences were also partially confirmed statistically, irrespective of the pH value measurement at 0 or 30 min. The lowest pH values were noted for dorsal-located muscles (LD, LL) and those located in the ham area (IL, SEM), while the highest pH values were for the location on the side surface of the carcass (CT, LTD). No statistically significant differences were found in the anatomically similar locations for pH30min, but for pH0min, differences were reported in ham and side locations.An increase in all pH values after 30 min bloom time was observed regardless of the muscle type, although the lowest increase was found for LD and LL below 0.11 (p ≤ 0.05), and the highest for IL and SEM over 0.20 values (p ≤ 0.01).Color changes with respect to bloom time and muscle type are presented in Table 1. The distribution of differences between muscles was not uniform, both within individual color parameters and for the two bloom time periods. Differences at a statistically confirmed level (p ≤ 0.05) after 30 min bloom time were observed only for LD muscle (all color parameters), LL (b*, C*, H*) and SEM (L*, a*, C*). Color measurements at 0 min differed (p ≤ 0.01) between muscles for L*, a*, C*, H*. Color measurements at 30 min differed (p ≤ 0.01) between muscles for L*, a*, b*, H*. Regardless of bloom time, always two LL and SEM muscles alternately obtained the highest and lowest values of the evaluated color parameters.The percentage share of color changes during 30 min bloom time is shown in Figure 2. A percentage increase of color parameters was noted for almost all the muscles (except SEM and partially IL), although to a varying degree. These changes were not homogeneous and were subject to individual variability. However, it should be noted that for LD and LL, the largest growth of changes was observed for parameters b* and H* (between 20–30% for the former muscle and 15–20% for the latter). The percentage decrease in the SEM muscle color value was at a level not exceeding 5%, with the dominant a* and C* parameters. As it was the case for LD and LL muscles, the highest variability of color in the IL muscle was observed for b* and H*, although it exceeded 5% only slightly. A large increase in the proportion of L* and a* was observed for CT muscle (20–30%, the highest of all observed) and LTD (20–25%).Figure 3 presents the dependence of percentage color change on pH parameters after 30 min bloom time without the division for muscles. A general tendency emerging from individual observations does not indicate any unambiguous dependence in this respect. Therefore, there seems to be no dependency between bloom time effect and the changing muscle pH, nor the interaction of these two main effects. The greatest dispersion was recorded for the b* and H* parameters. On the other hand, the highest stability was observed for the L* and a* parameters.The correlation coefficients for pH and color parameters for the individual muscle types at 0 min and 30 min bloom time are presented in Table 2. Numerous correlation coefficients were observed at moderate and strong levels, which were statistically confirmed (p ≤ 0.05, p ≤ 0.01). The lowest number of such correlations was noted for LD and CT muscles, and the highest for SEM.Regression equations estimating the pH values and color parameters after 30 min bloom time on the basis of measurements at 0 min were obtained by progressive regression method (Table 3). The accuracy of the obtained equations varied depending on the estimated parameter and muscle. An observation that connected all the equations was the use of more than one constituent parameter (except equations a* and b* for LD). A very good fit (R2 > 0.90) of regression equations was achieved in 7 cases (2 for LD—pH and L*; 1 for LL—H*; 1 for IL—pH; 1 for SEM—pH; 1 for LTD—pH; 1 for overall—pH). The adjustment at an unsatisfactory level (R2 < 0.50) was noted in 6 cases (1 for SEM—a*; 4 for LTD—a*, b*, C*, H*; 1 for overall—C*). Within the individual muscles, the equations for LTD were burdened with the biggest error of estimation, whereas those for LL were most precise.4. DiscussionWe found that the affinity of color variability due to bloom time is dependent on the type of muscle and location (especially in the case of dorsal and ventral muscles). The measurements were taken immediately after the cutting of the carcasses into material that was sold directly, in an unprocessed form, which undoubtedly influenced the practical aspect of the experiment. Our previous studies indicated that differences in carcass quality parameters may have a pronounced effect on the efficiency of meat production [19]. Meat plants are interested in the evaluation of the color of meat as a whole commercial cut [20]. Meat color is one of the main quality parameters, and as a sensory feature, affects the price of freshly cut parts and the final quality of processed pork products [21].SEM and IL muscles, located in the ham part, were characterized by the highest stability of color during 30 min bloom time. Other muscle types were characterized by similarly high variability, although it concerned various parameters. LD and LL muscles were found to change (increase) the proportion of the yellow color and H* the most, whereas CT and LTD muscles—the proportion of light and red color. High surface oxygen consumption by tissue is associated with poor oxygen penetration [22], which explains the clear disproportions of results for different muscle types. Since high variability of oxidative enzyme activity was observed for beef muscles [23], it may be presumed that similar dependences exist in the case of pork meat.The color of meat is determined by many factors, including the pH level that affects protein structure and moisture [24,25]. The connection between pH changes and stabilization and color variability was noted in earlier studies [26,27]. In our study, comparing the variability of color parameters with pH changes, the lowest average pH value was obtained for muscles located in the ham area. This is the opposite of what was reported by Zhu and Brewer [28], who pointed to the highest color variation for muscles with low pH, not those with high pH. However, it should be noted that our research was carried out on a research population with no quality defects. On the other hand, additional observations did indicate dependences in the change in pH over a 30 min period: SEM and IL muscles, in which the increase of pH was the highest, were characterized by the smallest variability of color (less than 5%) regardless of the tested parameters. Similarly, the muscles with the lowest increase of pH were characterized by the highest variability, although this concerned different color parameters.The distribution of individual observations indicates that pH determines the color although it depends on the specific color parameter. Most often, the best fit for regression equations was characterized by the equations developed for pH30min, always using the pH0min. The analysis of the developed regression equations indicates that the equations developed for the LTD muscle were most inaccurate. This emphasizes the need to pay special attention to changes occurring in this muscle during bloom time, because these changes cannot be easily predicted. Our previous studies clearly indicate that muscles from the belly cut are most difficult to evaluate qualitatively [29,30].The saturation index (C*), which refers to the intensity of color, in most cases was the only color parameter that did not significantly correlate with the L* parameter, which measures the lightness of the color compared to the white standard. Some authors indicate that this parameter shows the least variability during bloom time [10,12]. It is partially consistent with our results, but with the exception of CT and LTD muscles (again located abdominally). Škrlep and Čandek-Potokar [12] reported that the highest variability during bloom time was observed for b* parameter. We confirm this observation for dorsal location of muscles (LD and LL) and even IL muscle.Additionally, in the case of the loin muscle, the correlations demonstrated in this study were different from previous findings [12], and definitely with a different intensity. We showed a surprisingly large number of weak correlations for the tested pH parameters and color before and after bloom time for the LD muscle, which, after all, is treated as a model muscle for quality analysis. However, recent research clearly indicates that using the quality parameters of loin for the entire carcass is doubtful [30,31].As is the case with beef, we agree with the statement of Holman et al. [32], that there are still no objective instrumental criteria that could be used by consumers to evaluate pork color. The color difference in the case of meat with quality defects was the basis for a deeper visual analysis of fresh elements, hence the development of research in this area [10]. However, the distribution of meat quality defects in the pig carcass is not homogeneous for all skeletal muscles [33]. Our study focused exclusively on muscles without qualitative defects, which allowed us to visualize changes caused by 30 min bloom time. In a wider context, it should be considered that even muscles from the same carcass behave differently.5. ConclusionsAlthough bloom time influenced all the analyzed parameters, its effect was different depending on the muscle type. The presented results constitute an important contribution to the quick and precise instrumental evaluation of pH and color. The information may be of particular importance for slaughterhouses which, by introducing this assessment, could allow consumers to additionally verify the product. The contemporary consumer pays ever greater attention to the quality of consumed products and subjects, and evaluates them increasingly carefully. Only products characterized by appropriate quality and health safety are chosen. The quality of pig meat has long been of interest to both scientists and food technologists. We found that the affinity of color variation with bloom time was dependent on the type of muscle and its location (especially dorsal and abdominal muscles). SEM and IL muscles, located in the ham part, were characterized by the highest stability of color during 30 min bloom time. Other muscle types were characterized by similarly high variability, although it concerned various parameters. The quality of the finished product and the profitability of a meat processing plant depends, to a large extent, on the quality of raw materials. The nature of color variation can also be predicted from the regression equations proposed by us, which constitute a convenient computational tool. To recapitulate, we presented additional possibilities of quick and precise sorting of raw materials intended for consumer sale.

animals : an open access journal from mdpi

10.3390/ani11102856

PMC8532846

Parasitism with gastrointestinal nematodes represents a significant risk to the health of livestock populations. Besides the local oxidative damage caused by the parasite, the host reacts by increasing the production of oxidants. The study of thiol-disulphide homeostasis can be of help in the evaluation of the oxidative status of sheep during this type of parasitism. In this study, the thiol-disulphide homeostasis, together with other biomarkers of oxidative stress and inflammation, were assessed in the serum of lambs infected with gastrointestinal nematodes and were evaluated after 70 days of integrated crop-livestock system and anthelmintic treatment. This study showed that the thiol-disulphide balance was impaired in the infected lambs and the changes were correlated with the parasite load, which therefore could indicate their potential use as a tool to evaluate and monitor the disease in sheep.

This work aimed to evaluate the thiol-disulphide homeostasis in serum of lambs naturally infected by gastrointestinal nematodes presenting different levels of parasite load indirectly indicated by faecal worm egg counts (EPG). Furthermore, the possible changes in the thiol-disulphide dynamic after different procedures to reduce the parasitic charge, such as the integrated crop-livestock system or anthelmintic treatment, were assessed. The results were compared with a panel of various oxidative stress and inflammatory biomarkers. The lambs were divided into three groups: animals highly infected (EPG higher than 5000) and packed cell volume (PCV) lower than 24% (G1); animals highly infected (EPG higher than 5000) and normal PCV (>24%) (G2); and animals presenting EPG lower than 5000 and normal PCV (>24%) (G3). The highly infected lambs (G1 and G2) showed lower total thiol (TT) and native thiol (SH) (p ≤ 0.01) than those from G3. After treatment, TT and SH increased significantly in all groups (p ≤ 0.01), and the disulphide (SS)/TT and SS/SH ratios decreased significantly (p < 0.01) in G1 and G2. These results show that the thiol-disulphide balance was impaired in lambs infected by gastrointestinal nematodes and that it could be potentially used as a biomarker to monitor this disease.

1. IntroductionParasitism with gastrointestinal nematodes represents the main risk to livestock populations’ health, welfare, and productivity throughout the world. Particularly in sheep, it is an important cause of the drop in production [1,2]. This parasitic gastroenteritis is associated with different species of nematodes such as Haemonchus contortus, Teladorsagia circumcincta, and Trichostrongylus axei that mainly act in the abomasum; Trichostrongylus colubriformis, Cooperia curticei, and Nematodirus spathiger in the small intestine; and Oesophagostomum venulosum in the large intestine [3]. In tropical areas, the most predominant are H. contortus and T. colubriformis [4], which are considered the most important in small ruminants [5,6].Almost all sheep are infected with one or more of these nematodes in field conditions. The infection occurs when animals are grazing contaminated pasture, and the mix of larvae species (third-stage) ingested infects both the abomasum and intestine. The worm completes its development in adults in the gastrointestinal tract in approximately three weeks, and some species such as H. contortus may survive over a year inside a host sheep [7]. The gastrointestinal nematode parasitism may result in severe clinical disease, particularly related to the hematophagous H. contortus [8]. However, the disease is rarely due to just one parasitic species but rather to the cumulative effects of mixed parasitism [3]. Furthermore, the intensity of infection and clinical signs related to the disease could vary significantly depending on factors such as the severity of parasitism in the gastrointestinal tract, the general health and immunological status of the animal, type of climate and pasture, management, and diet [9]. Faecal worm egg counts (EPG) associated with the packed cell volume (PCV) are basic approaches that help to determine the severity of the infection [4,10,11].Currently, the integrated crop-livestock (ICL) system, consisting of a single farm or areas managed by integrating agricultural crops and livestock, is increasingly used to eliminate the free-living stages of gastrointestinal nematodes from the pasture [12]. However, in situations where the high parasitic load is associated with severe clinical signs, treatment using chemical compounds with a broad spectrum of action is still used [13].The inflammation in the abomasal mucosa due to the parasite attachment is an initial response to control the parasitic load. In addition, the parasites lead to local oxidative damage during the digestion of host tissue [14]. Furthermore, in an attempt to damage the parasite, the host reacts by increasing the production of reactive oxygen species (ROS), which, depending on the antioxidant system’s capacity, could produce a state of oxidative stress [15]. The local production and expression of oxidants and antioxidants in the gastrointestinal tract following a nematode challenge have been reported in sheep [16]. In addition, changes in serum biomarkers of oxidative stress have been described in lambs and sheep experimentally or naturally infected by gastrointestinal nematodes such as H. contortus [17,18,19,20,21].The study of thiol-disulphide homeostasis, which involves the analysis of serum total thiol (TT), native thiol (SH), and disulphide (SS) concentrations, is gaining interest in human medicine to evaluate oxidative status in different diseases [22,23,24,25]. Previous studies suggested that thiol-disulphide homeostasis can be used as a marker of oxidative stress in sheep during sarcoptic mange infection [26]; however, to the author’s knowledge, no information about the evaluation of these biomarkers in gastrointestinal parasitism in sheep has been reported.The main objective of this study was to determine serum levels of thiol-disulphide in lambs naturally infected by gastrointestinal nematodes presenting different levels of parasite load and clinical signs of the disease, in order to evaluate whether thiol-disulphide balance could be used as a new marker of oxidative stress in this disease. Furthermore, the second objective of this work was to evaluate the possible changes in thiol-disulphide levels after the reduction of the EPG and consequently the parasite burden, by different procedures such as ICL and anthelmintic treatment. 2. Materials and Methods2.1. AnimalsForty-one lambs of the breed Corriedale, aged approximately 90 days and originally from southern Brazil, were included in this study. The animals were raised under commercial and extensive breeding conditions, and before starting the experimental period, they received vitamin supplementation (Vitagold®) and a dose of a prophylactic vaccine for botulism, symptomatic carbuncle, gaseous gangrene, enterotoxaemia, and sudden death of ruminants (Poli-Star®, Vallée). They were then introduced in the ICL system, and during the first day of study (June 28), blood and faecal samples were obtained. Three groups were defined according to the cut-off value of 5000 EPG in agreement with a previous study [27] and the presence of decreased PCV to the species (reference intervals for the species is 24 to 50% [28]). The first group (G1) was integrated by 15 animals presenting EPG higher than 5000 and PCV lower than 24%; group 2 (G2) was integrated by 14 animals presenting EPG higher than 5000 and normal PCV (>24%), and group 3 (G3) was comprised by those lambs presenting EPG lower than 5000 and normal PCV (>24% (n = 12)). The study was developed in the experimental farm of the School of Veterinary Medicine and Animal Science, São Paulo State University, Botucatu, Brazil, in an experimental area in which an ICL system was planned to be implanted and conducted during winter, from 28 June to 6 September 2017, for 70 days as indicated previously [12]. Tropical conditions characterized the climate with dry winter. Daily at 6:00 a.m., all lambs were released from the shed to the paddocks where they grazed and had water ad libitum from drinking fountains. During the day, they were maintained in a pasture of black oats (Avena strigosa), overrun in the area after a harvest of maise intercropped with palisade grass (Urochloa brizantha, common name Marandu) to produce silage in an ICL system, which was offered to the animals. In the late afternoon, the lambs were housed in a covered shed where they were feed with maise silage and concentrate. Maise with marandu palisade grass was sown in the same area in the summer, and mechanical harvesting for ensilage was conducted at the beginning of the autumn. The roughage:concentrate ratio was 30:70. The diet of the lambs was formulated via the computer program Small Ruminant Nutrition System (SRNS) based on the Cornell Net Carbohydrate and Protein System (CNCPS) for sheep [29].As clinical anaemia is the determinant for treating the animals with anthelmintic drugs for gastrointestinal nematodes [30], the animals presenting a PCV lower than 24% (G1) received a treatment consisting of closantel, 10 mg/Kg live weight (Diantel®, RS-Brazil) [12]. The other two groups of sheep were just maintained in the ICL system following previous recommendations [12,31]. The Ethics Committee of the School of Veterinary Medicine and Animal Sciences (CEUA 0152/2017), São Paulo State University (FMVZ-UNESP), Brazil, approved the study following current legislation on animal protection.2.2. SamplingFaeces and blood samples were taken from each animal on the first day of the initiation of the ICL system and 70 days after. The sampling was made at 06:30 a.m. while they were still fasting and in their shed. Blood samples were collected from the jugular vein and placed in tubes with EDTA (BD Vacutainer® K2 EDTA Blood Collection Tube; Becton, Dickinson and Company, Franklin Lakes, NJ, USA) for haematology analysis and with a vacuum system and gel separator (BD Vacutainer® Blood Collection Tube; Becton, Dickinson and Company, NJ, USA) in others to obtain serum. The vacuum tubes were centrifuged at 1500× g for 5 min and the serum obtained was stored frozen (−80 °C) in Eppendorf tubes (Eppendorf, Hamburg, Germany) until analysis. Faeces were collected directly from the rectum, stored in labelled plastic bags at 4 °C, and taken to the Veterinary Helminthology Laboratory at the UNESP—Institute of Biosciences, Campus of Botucatu, Brazil, where they were analysed.2.3. Faecal ExaminationThe EPG was performed using the modified McMaster technique [32] following the specifications for small ruminants (2 g of faeces diluted in 58 mL of a hypersaturated solution of NaCl), where each egg counted represented 100 eggs/g of faeces. The larval culture [33] was done in a pool of samples from all the animals for identifying genera of third-stage infective larvae (L3) [34].2.4. Analysis2.4.1. Blood SamplesThe PCV was determined by the microhematocrit method [28].2.4.2. Thiol-Disulphide HomeostasisSerum TT, SH and SS concentrations were determined by using a colorimetric method [35]. In brief, dynamic SS bonds in the sample were reduced to functional thiol groups (SH) by sodium borohydride (NaBH4). The excess of NaBH4 was completely removed by formaldehyde, preventing the extra reduction of the 5,5′-dithiobis-2-nitrobenzoic acid (DTNB) and further reduction of the SS formed, which was produced after the DTNB reaction. The TT content of the sample was measured using a modified Ellman reagent. The SH content was subtracted from the TT content, and half of the obtained difference gave the SS bond amount. In addition, the SS/TT, SS/SH, and SH/TT ratios were calculated.The assays were performed using the Olympus AU400 (AU400 Automatic Chemistry Analyser, Olympus Europe GmbH, Hamburg, Germany) and showed lower than 15% imprecision.2.4.3. Oxidative Status BiomarkersThe results of the thiol-disulphide homeostasis were compared with a comprehensive panel of various automatic antioxidant (cupric reducing antioxidant capacity (CUPRAC), ferric reducing ability of plasma (FRAP), Trolox equivalent antioxidant capacity (TEAC), and uric acid) and oxidant (total oxidant status (TOS), ferrous oxidation-xylenol orange (FOX), reactive oxygen metabolites derived compounds (d-ROMs), and advanced oxidation protein products (AOPP)) biomarkers usually included in the profiles to evaluate oxidative stress.Briefly, the CUPRAC was measured following an assay previously described [36]. It is based on reducing cupric-bathocuproinedisulfonic acid to a cuprous-bathocuproinedisulfonic acid by the sample. The determination of TEAC was based on the reduction of the radical 2,2′-azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) pre-generated enzymatically with horseradish peroxidase (HRP) [37]. The measurement of FRAP was based on reducing the ferric-tripyridyltriazine to ferrous-tripyridyltriazine by the non-enzymatic antioxidants present in the sample [38]. Uric acid was determined by a commercially available spectrophotometric method (Uric Acid reagent OSR6698 Beckman Coulter AU analysers, Nyon, Switzerland). Serum PON1 was determined using a previously assay based on p-nitrophenyl acetate as substrate [39].Serum TOS was measured by a method [40] based on the oxidation of the ferrous ion–o-dianisidine complex. The FOX assay was determined based on the automatic determination of the ferrous oxidation by xylenol orange [41]. The d-ROMs assay was based on monitoring the N,N-Diethyl-p-phenylenediamine radical cation concentration as previously published [42]. The indirect evaluation of oxidative damage to proteins was done by the measurement of AOPP, which was based on the determination of di-tyrosine containing cross-linked proteins and oxidatively modified albumin as previously described [43].All assays were performed using the Olympus AU400 and showed imprecision lower than 15%.2.4.4. Inflammatory BiomarkersHaptoglobin (Hp) was determined by a colorimetric method (kit haptoglobin Tridelta phase range, Tridelta Development Ltd., Bray, Ireland) commercially available. Butyrylcholinesterase (BChE) was analysed according to a previous assay based on butyrylthiocholine iodide as substrate [44].Total protein and albumin concentrations were measured following the instructions of the manufacturer using Olympus commercial kits (Total protein OSR 6132, Albumin OSR 6102, Olympus Life and Material Science Europe GmbH, Hamburg, Germany). Globulin concentrations were calculated by the difference between total proteins and albumin concentrations. In addition, the albumin/globulin (AGR) ratio was calculated (AGR = Albumin/(Total protein − Albumin).2.4.5. Statistical AnalysisData analyses were performed using Excel (Microsoft, Redmond, WA, USA) and Graph Pad Software Inc. (GraphPad Prism, version 6 for Windows, Graph Pad Software Inc., San Diego, CA, USA). A value of p < 0.05 was considered significant. Descriptive statistics were used to describe the basic features. The different biomarker concentrations were evaluated for normality of distribution using the Shapiro–Wilk normality test. All results met the normal distribution criteria except for EPG values, which were log-transformed prior to the analysis. Data are presented as means ± standard deviation in bar graphs. Two-way analysis of variance (ANOVA) mixed model of repeated measured tracked by Sidak’s multiple comparison test was used to evaluate the results over time from each group of animals. Two-way ANOVA followed by the Tukey post hoc multiple comparison analysis was performed to compare the results of each biomarker over time (days 0 and 70) comparing between the groups. Spearman’s correlation coefficients were calculated to assess the relationship between all biomarkers and the EPG values.3. Results3.1. AnimalsResults of EPG and PCV in each group of sheep infected by gastrointestinal nematode are shown in Figure 1. When groups were compared at day 0, G1 and G2 presented higher EPG counts (p < 0.0001) than G3. On day 0, the third larval stage of Haemonchus and Trichostrongylus were identified in 94.5% and 5.5% of the animals, respectively. On day 70, Haemonchus was identified in 73.1%, Trichostrongylus in 23% and Cooperia in 3.8% of the animals. Due to the high percentage of Haemonchus at the beginning of the trial and the low PCV, the lambs from G1 received anthelmintic treatment with closantel, 10 mg/Kg live weight (Diantel®, RS-Brazil) [12] two days after the first sampling. The rest of the animals were not treated since it was expected that the ICL system would reduce their parasitic burden.After 70 days of being introduced in the ICL system and the closantel treatment, the EPG in G1 decreased significantly (p < 0.001). In the same way, the EPG in G2 decreased significantly at the end of the trial (p < 0.001). Regarding G3, although no statistical difference was observed (p > 0.05), the EPG values decreased compared to day 0 (Figure 1).When groups were compared at the beginning of the trial, the PCV was different between all of them (p < 0.05). The PCV increased at the end of the experiment in all groups compared with the first sampling (p < 0.01) (Figure 1).A significant interaction between time and group was observed in EPG (p < 0.0001) and PCV (p = 0.0002).3.2. Thiol-Disulphide HomeostasisThe results for the thiol-disulphide homeostasis system in sheep naturally infected by gastrointestinal nematodes are shown in Figure 2. On day 0, significant differences in thiol-disulphide analytes were found between the groups of sheep, with G1 and G2 showing lower values of TT and SH than G3 (p ≤ 0.01) (Figure 2). The SS/TT and SS/SH ratios were significantly higher (p < 0.05) in G2, and in G1 and G2, respectively, than G3 on the first day of the trial. On the same day, G1 and G2 also exhibited a lower SH/TT ratio than G3 (p < 0.05). When biomarkers were compared before and after treatment, the TT and SH increased significantly (p < 0.01) on day 70 in G1, G2 and G3. Higher SS levels (p < 0.01) were observed in G1 and G2 at the end of the trial. The SS/TT ratio in all groups and the SS/SH ratio in G1 and G2 were decreased (p < 0.05) on day 70 compared to day 0. On the other hand, the SH/TT ratio in G1, G2 and G3 were significantly higher (p < 0.05) at the end of the trial (Figure 2).Significant interaction between time and group was observed in TT (p = 0.001) and SH (p = 0.001). On the other hand, there was no significant interaction between time and group in SS (p = 0.074), SS/TT (p = 0.244), SS/SH (p = 0.088), SH/TT (p = 0.169).3.3. Antioxidant BiomarkersAt the first sampling, when the antioxidants were compared between the groups, CUPRAC and TEAC showed significantly lower concentrations in G1 and G2 than in G3 (p < 0.001). CUPRAC was also lower (p < 0.05) in G1 than G2. FRAP, uric acid and PON1 did not show significant changes between groups (p > 0.05) (Figure 3). After treatment, CUPRAC and TEAC increased significantly in G1 and G2 (p < 0.001) (Figure 3). FRAP decreased significantly (p < 0.05) at day 70 of G3. Uric acid and PON1 did not show significant changes between days (p > 0.05) (Figure 3). Significant interaction between time and group was observed in CUPRAC (p = 0.0005) and TEAC (p = 0.0001). On the other hand, there was no significant interaction between time and group in FRAP (p = 0.073), uric acid (p = 0.199) and PON1 (p = 0.305).3.4. Oxidant Biomarkers Results for TOS, FOX, d-ROMs and AOPP are shown in Figure 4. FOX on day 0 was significantly lower in G1 than G3 (p = 0.03) and increased significantly (p < 0.05) in G1 on day 70. TOS, d-ROMs and AOPP showed no significant changes throughout the study (p > 0.05) (Figure 4).There was a significant interaction between time and group in FOX (p = 0.018). There was no significant interaction between time and group in TOS (p = 0.987), d-ROMs (p = 0.568) and AOPP (p = 0.356).3.5. Inflammatory Biomarkers The results for inflammatory biomarkers in sheep infected by gastrointestinal nematodes are shown in Figure 5. When the groups were compared in the first sampling, total proteins, albumin, and globulins concentrations were lower in G1 and G2 than G3 (p < 0.05). The Hp and BChE results in serum were below the detection limit of the assays for all groups (data not shown). When treatment was evaluated, sheep from G1 and G2 showed increased total protein, albumin and globulins concentrations (p < 0.05) at day 70. The AGR was also higher on day 70 in G1 (p < 0.05) (Figure 5). A significant interaction was observed in total proteins (p = 0.008) and albumin (p = 0.004). On the other hand, there was no significant interaction between time and group in globulins (p = 0.074) and AGR (p = 0.683).3.6. Correlation StudyTable 1 shows Spearman correlation coefficients and significance between EPG and all biomarkers studied in the sheep with gastrointestinal nematodes during the 70 days of the ICL system. Regarding thiol-disulphide homeostasis, the analytes that showed the highest correlation coefficients with EPG were TT, SH, and SH/TT ratio, which correlated negatively (r > −0.72), and with SS/TT, SS/SH ratios that correlated positively (r > 0.71). Within antioxidant markers, the highest correlation with EPG was between TEAC (r = −0.68) and CUPRAC (r = −0.66). Among the inflammatory biomarkers, the highest coefficient correlation was found negatively between EPG and albumin and total proteins (r = −0.69).4. DiscussionThe study of thiol-disulphide homeostasis for the clinical evaluation of oxidative stress in different diseases is gaining interest in recent years, especially in human medicine [45,46]. The analytes used for the assessment of thiol-disulphide homeostasis are total thiol (TT), native thiol (SH) and disulphides (SS). The TT is integrated mainly by free cysteine, glutathione, cysteine residues and albumin and includes both reduced and oxidised thiols, whereas SH represents only the reduced thiol [47]. The thiols have a high vulnerability to oxidation [45], forming SS when oxidised [48,49]. Since albumin is the main component of the plasma thiol pool, the SS/TT, SS/SH and SH/TT ratios are used to avoid the possible effect of albumin [25]. In this study, the thiol-disulphide homeostasis system changed in sheep naturally infected by gastrointestinal nematodes, and in most analytes, this change was correlated with the parasite load. This correlation was negative for TT, SH, SS and SH/TT ratio and positive for SS/TT and SS/SH ratios. Furthermore, the animals infected with more than 5000 EPG, independent of having anaemia or not, showed lower TT and SH than animals having lower than 5000 EPG. This may indicate that the infection by gastrointestinal nematodes in sheep causes changes in the steady-state balance between reduced and oxidised thiol. These changes consist in the decrease in the levels of reduced forms and increasing the oxidised components and consequently the ratios among SS and thiol forms, suggesting a state of oxidative stress. In addition, the results reveal that the more intense the parasite burden the greater the oxidative response. Similar changes have been described in sheep with sarcoptic mange [26] and in other species and parasitic diseases such as cutaneous leishmaniasis in humans [50].This report also evaluated the changes after treatment in thiol-disulphide metabolism in sheep infected by gastrointestinal nematodes. Two different treatment methods were used in the animals of our study: anthelmintic closantel with ICL and only ICL. Although the practice of only using anthelmintics was common in the past, it is not sustainable and has led to resistance [51,52]. Therefore, an integrated approach (ICL) including environmental management and chemoprophylaxis to minimise the pressure for parasite adaptation and parasite control is frequently used nowadays [31]. In our study, the two different treatment methods used were efficient in decreasing parasite load and improving inflammation and oxidative status of the sheep. Our results are in line with a previous study in which the ICL associated with a good nutrition plan resulted in progressively declining degrees of gastrointestinal parasite infections and satisfactory performance of lambs [12].The improvement in the oxidative status after the treatment was reflected in the changes in the dynamic thiol-disulphide homeostasis since all animals showed a reduced SS/TT ratio after treatment. Those highly infected also exhibited decreased SS/SH ratio at the end of the trial. Furthermore, all of them showed increased TT and SH. In this context, decreased SS/TT and SS/SH levels suggest that the rate of thiol being oxidised has declined, and the redox state of the thiol-disulphide system has been re-established. When other analytes regarding oxidative stress were evaluated, CUPRAC and TEAC were also decreased in both groups of sheep with high parasite load on the day of diagnosis. In addition, both biomarkers increased at the end of the trial just as the parasite burden reduced. TEAC was decreased in sheep infected with H. contortus and T. circumcincta [18,53,54]. In this study, FRAP did not change between groups; however, it increased after 70 days of study in G3, which could be in line with a previous study [19] that related its elevation to cellular protection during the infection with H. contortus in lambs. The antioxidant system has a cellular protective action against oxidative stress of cells, organs and tissue damage that results from parasitic invasion [55]. The depletion of systemic antioxidant status may reflect the increase in oxidative stress, which in this study was more pronounced in G1 and G2. The presence of the parasite increases mucosal damage and local oxidative stress, which depending on the intensity may lead to systemic changes. In addition, the pathogenesis of most parasitic infections has been associated with lipid peroxidation [56]. The increase in FOX in G1 after the treatment could be related to possible oxidative damage caused during the infection. However, additional studies are necessary to confirm this result. In general, according to our results, just CUPRAC, TEAC and the dynamic thiol-disulphide homeostasis could be used as markers of oxidative stress, which is related to the parasite burden, and monitor treatment of sheep infected by gastrointestinal nematodes. However, the biomarkers of thiol-disulphide homeostasis such as TT, SH and SS/TT and SS/SH ratios, concurrently with CUPRAC, showed more significant changes between groups and during treatments and may indicate that they could be more sensitive oxidative stress biomarkers for the monitoring of the response to treatment against gastrointestinal parasitism in sheep.It is important to point out that the biomarkers of oxidative stress are nonspecific and cannot be used for diagnosis. Moreover, in agreement with a previous study [57], the use of various biomarkers should be strongly recommended for the evaluation of the redox balance in livestock species, in which reference intervals for each assay are still lacking. Additionally, the results of this study should be confirmed using a higher number of animals, and ideally, a healthy control group involving animals without parasites as well as a group of animals with more severe clinical signs of the disease should be included.5. ConclusionsIn our study conditions, the thiol-disulphide balance was impaired in lambs infected by gastrointestinal nematodes. The highly infected lambs showed lower TT and SH than those with a lower parasite burden. After treatment, these markers increased significantly in all groups of lambs, but the observed decreased SS/TT and SS/SH ratios were more intense in those highly infected animals. The results also indicated that the oxidised form of thiol decreased after reducing parasite load, where the thiol-disulphide homeostasis was more sensitive than other oxidative stress biomarkers in monitoring treatment. These results show that thiol-disulphide homeostasis is involved in parasitism by gastrointestinal nematodes in sheep and there is potential for using the biomarkers of this system to evaluate the severity of the disease and monitor treatment in sheep infected by gastrointestinal nematodes.

animals : an open access journal from mdpi

10.3390/ani11102973

PMC8532706

As we keep and use sheep, we need to be able to assess their welfare and deal with welfare problems as they arise. To assess welfare, a comprehensive protocol based on valid and feasible indicators is needed. The aim of this study was to review the scientific literature and identify protocols and indicators for assessing the welfare of sheep. We identified promising protocols, well-known and established indicators, such as lameness or body condition score, as well as novel indicators that still need to be evaluated to prove their validity, such as pruritic behaviour or resting time. This review provides a starting point for the development of valid and feasible on-farm protocols using animal-based indicators to assess sheep welfare.

The value society assigns to animal welfare in agricultural productions is increasing, resulting in ever-enhancing methods to assess the well-being of farm animals. The aim of this study was to review the scientific literature to obtain an overview of the current knowledge on welfare assessments for sheep and to extract animal-based welfare indicators as well as welfare protocols with animal-based indicators. By title and abstract screening, we identified five protocols and 53 potential indicators from 55 references. Three out of the five protocols include animal-based as well as resource-based indicators. All of them were assessed as being practicable on-farm but lacking reliability. Some of the single indicators are endorsed by the literature and widely used in the field like assessment of behaviour, lameness or body condition score. Others (e.g., Faffa Malan Chart FAMACHA©, dag score or pain assessment) are regularly mentioned in the literature, but their reliability and usefulness are still subject of discussion. Several indicators, such as pruritic behaviour, eye condition, lying time or tooth loss are relatively new in the literature and still lack evidence for their validity and usefulness. This literature review serves as a starting point for the development of valid and practicable welfare protocols for sheep.

1. IntroductionAnimal welfare has always been an issue of concern to varying degrees in our society and has evolved enormously over the years. As animals cannot express their needs directly, their welfare depends on our interest and understanding, as well as our diligence in measuring, respecting and improving the conditions for the animals we keep [1]. In order to get an impression of the live quality of stock animals, we need to be able to assess their welfare with practical and robust protocols and be able to address welfare problems as they occur [2]. To meet these expectations, it is imperative that a valid and understandable protocol, based on coherent indicators, must be developed to attest welfare.Animal welfare indicators can be sorted in three categories: (i) indicators assessed by observation or examination of animals (animal-based); (ii) indicators that assess animal-related provisions such as housing and grazing (resource-based); or (iii) indicators that relate to farmers’ policies and management practices (management-based) [3]. Animal-based indicators of sheep welfare selected for a welfare assessment must be valid (relevant to sheep welfare), reliable (produce consistent results when performed at different time points or by different assessors) and feasible (efficient in terms of time, staff and materials) [4].Several studies have identified the main welfare problems of sheep [5,6], and some studies have provided welfare protocols [5,7,8,9,10], or identified single welfare indicators. The aim of this review was to evaluate the possibilities to assess sheep welfare and to provide an overview of how appliccable these possibilites are considered to be on-farm. We review the scientific literature on either assessment protocols or single animal-based welfare indicators for sheep, and state their value in terms of validity (does the indicator reflect welfare?), reliability (how accurate is the indicator between observers and over time?) and/or feasibility (is the indicator considered practical in terms of time and resource consumption?). For this review, we chose to select animal-based welfare indicators since many experts consider them the most valid method to assess animal welfare. Such indicators provide a direct measurement of the welfare status of the animals and often reflect the outcome of resource inputs and management practices [11].2. Materials and Methods2.1. Literature SearchA search of scientific literature using assembled search terms accepted by consensus of experts was performed. Included were experimental and observational studies on sheep welfare referring to welfare assessments for adult sheep or lambs. For this purpose, five main terms (“ewe”, “lamb”, “ovine”, “sheep”, “small ruminants”,) were separately combined with twelve secondary terms out of four categories (“data”, “health”, “mortality”, “welfare”). The search was conducted in four search engines: PubMed [12], Science Direct [13], Scopus [14] and Web of Science [15]. We considered all articles published between the 1 anuary 1995 and the 4 March 2020. Any query that resulted in more than 500 articles was narrowed down with filters to reduce the number and improve the precision of the results. The used filters were set sequentially until fewer than 500 papers were listed. First, we chose filters to include only publications written in English, French or German, second, we used filters to exclude those related to human medicine (e.g., “animals”) and finally, we isolated papers specifically in the field of veterinary medicine (e.g., “veterinary sciences”). At the end, 24,675 papers were compiled. All papers were saved using the literature management software Zotero 5.0.93 (Vienna, VA, USA) [16].2.2. Title and Abstract ScreeningFor an evaluation of the content, the articles were transferred to DistillerSR (Ottawa, ON, Canada) [17] and checked for duplicates, allowing the removal of 9606 duplicates. The remaining 15,069 articles underwent a title screening by the first author to sort out all publications that did not match the purpose of this review. We excluded publications not written in English, German or French, if were not filtered during the search-result reduction step, and publications which were clearly not related to sheep welfare. This step reduced the number of articles to 749, which were then subject to abstract screening. For the subsequent data extraction step, only peer-reviewed journal articles that contained information on animal-based indicators of sheep welfare or protocols, defined as any procedure that includes information on the measurement of sheep welfare using multiple indicators, were forwarded. After this step a total of 51 references remained in the selection. 2.3. Data ExtractionThe remaining 51 articles were evaluated to identify animal-based indicators for sheep welfare. We included all articles with descriptions of welfare indicators or protocols and specifically assessed the authors’ discussion of these indicators in terms of validity, reliability and/or feasibility. We excluded 17 articles, as the discussions of the respective articles did not contain sufficient information on the authors’ appraisal of the presented indicators in terms of the selected quality features. From the remaining 34 papers, 21 additional articles mentioned in the references, were included using the same procedure as described in the previous sections. The extracted indicators and protocols were registered in a Microsoft Excel v2102 (Redmond, WA, USA) [18] spread sheet. Indicators with very similar definitions, e.g., fleece condition and fleece derangement, were grouped together, resulting in 53 individual indicators that were considered relevant to this review. We classified the indicators according to the number of publications in which they were mentioned and report and discuss them in the results section according to citation frequency in descending order. Additionally, there were five articles describing procedures that included multiple indicators united in an entire animal welfare assessment protocol. These articles are reported starting with the most comprehensive protocol.3. Results and Discussion3.1. Welfare Assessment ProtocolsEstablished protocols may be the easiest way to assess sheep welfare, as they cover more than one aspect of welfare using a set of different indicators. We identified five protocols from the scientific literature, which we subsequently denote after their authors or, if available, their name [5,7,8,9,10]. All protocols consist of animal-based indicators and were declared to be practicable on-farm by the respective authors.3.1.1. AWINWell-founded animal welfare protocols emerged from the Welfare Quality® (Leystad, Netherlands) [19] project established in 2004 for cattle, pigs and poultry, but none for small ruminants. In response, the Animal Welfare Indicators (AWIN) project was established in 2011 [9,20]. The AWIN project was developed with the aim of improving animal welfare and filling a gap in the Welfare Quality® project [9]. The AWIN protocol is based on a two-level approach, a prior herd-level approach and an in-depth individual-level assessment [20]. The indicators for each level are provided in Table 1. At the first level, a screening of the flock is carried out with robust and rapid animal-based indicators with no or minimal animal handling. Performing the second level assessment is recommended when there is a non-compliance with the current animal welfare legislation or if the assessment of a specific indicator results in the assessed farm belonging to the lowest 5% of the farms in the reference population. The second level consists of a more detailed and an in-depth assessment requiring restraining the animals and collecting individual data [20]. The two-level approach was chosen to reduce animal’s stress and the time needed for the assessment.3.1.2. Protocol of Napolitano Napolitano et al. elaborated a protocol in Italy [7]. The protocol is based on four categories derived from the Animal Needs Index [21] for cattle. These four categories focus mainly on resource-based parameters, whereas, a fifth category also includes animal-based parameters. The fifth category, which accounts for 36.6% of the total score, contains seven animal-based indicators presented in Table 1; body condition, integument alterations, animal dirtiness, hoof overgrowth, lameness, lesions and mutilations, such as de-horning and tail removal. In addition, body condition was chosen as an indicator of malnutrition and disease. According to the authors, the indicator body condition score could not be evaluated in practice because the sheep were not shorn as the farms were visited in winter. Therefore, thin and even emaciated sheep could not be identified, and body condition did not contribute to the evaluation of welfare even if the authors recommend this indicator for inclusion in a welfare protocol. Each measure results in a score depending on its prevalence in the herd. The more frequent violations for a specific indicator are observed within the herd, the lower the assigned score. The score “optimal” is given if 5% or fewer animals in the herd are observed with a violation, “good” for 10% or fewer, “medium” for either 50% or 25% or fewer, depending on the indicator, and poor for more than 50% or 25%, respectively. The separate indicator scores are then expressed as numeric values and summarized to a final score.To evaluate the protocol in terms of feasibility and inter-observer reliability, two trained observers carried out the protocol on ten organic and ten conventional sheep farms in Southern Italy. The average number of animals per farm was 350, and at least 20% of lactating animals were recorded on each farm. No sophisticated equipment was required and the average time to complete the assessment was 85 minutes per farm. The authors identified the lack of direct measurement of internal parasites as a weakness of the protocol and recommend the inclusion of parasite egg counts to increase the validity of the scheme, although the assessment time may increase. The authors also advised good training in lesions assessment and to visit farms soon after shearing to facilitate lesion detections and body condition assessment and to increase the reliability of interventions.Based on these findings, the protocol seems to be a practical tool for assessing the welfare of sheep on-farm. In addition, the protocol could provide farmers with recommendations on which management aspects need to be improved. However, further studies are needed to test the scheme on a larger sample size to assess its reliability [7].3.1.3. Protocol of StubsjøenStubsjøen et al. [8] also proposed to assess the welfare of sheep using animal- and resource-based measurements. They adapted the animal welfare protocol established for dairy cattle, based on the Five Freedoms [12] to sheep. Sixteen animal-based, 15 resource-based and three measurements on production records (slaughter weight, carcass classification and fat class) were selected. The animal-based indicators are presented in Table 1. The protocol consists of two parts; the animal- and resource-based measurements carried out during farm visits and the analysis of production data. The assessment starts with a flock observation to detect signs of clinical disease, lameness and coughing. Then, ten randomly selected animals undergo a clinical examination. Finally, in the animal-based measurements, the animal’s behaviour is observed to assess anxiety levels and the human–animal relationship.In relation to the indicator measuring fear, the authors used a modification of methods validated in Reference [22] to assess the ewes’ response to an unfamiliar person. In brief, the indicator counts how often a test person can walk up to and touch selected animals. To test the farmer-animal relationship, the person who interacted most with the animals was asked to enter different pens and tag randomly selected ewes in each pen. The ewe’s response was categorised into four groups, ranging from 3) “behaved calmly when approached” to 0) “attempts to escape by jumping out of the pen”. An average score was calculated for each farm. Finally, resource-based indicators such as relative humidity or temperature are measured three to 27 times, depending on the farm’s sizes, and an average is calculated. On average, three to five hours are needed to carry out the assessment and all the observations are carried out indoors. The second part, the analysis of production data, includes individual information on carcass weight, fat class and carcass classification. Regarding these three indicators, the authors could not find sufficient information on their value as indicators.To test the protocol in relation to inter-observer reliability, two observers with clinical experience from veterinary practice and one ethologist visited 36 farms in Norway and assessed ten randomly selected animals on each farm. The assessment took place during lambing season. The observer’s agreement was excellent, except for body condition score (BCS), callus on carpus and claws. Therefore, the scoring systems for these three measures need to be more clearly defined or the observers have to be trained in more detail. Furthermore, the reliability and feasibility of the selected parameters still need to be assessed [8].3.1.4. First and Second Protocol of MunozTwo protocols by Munoz et al., published in 2018 [5] and 2019 [10], include animal-based indicators and were developed for extensively managed sheep. The authors derived 17 indicators from a review of the relevant scientific literature (Table 1) and matched them to the five domains of welfare. Of the 17 measurements, eight were selected for their reported validity and their reliability and feasibility [5] to assess the welfare of extensively managed ewes.The first protocol was tested on 100 randomly selected ewes from a larger flock of about 3000 breeding ewes in Victoria, Australia. Each animal was studied at three key stages: pregnancy, lactation and weaning. The ewes were kept in four groups of 25 animals. First, a group flight distance test was conducted to observe the ewe’s response to an unfamiliar human. Then, the ewes were placed in a single row and were individually examined. The indicators included in this protocol were able to detect impaired welfare and welfare risks in extensively managed systems [5] but their reliability and feasibility need further research.The second protocol, by the same authors is an adaptation of the first one [5]. Of the eight animal-based welfare indicators, six were kept: body condition score (BCS), fleece condition, skin lesions, tail length, dag score and lameness (Table 1). In addition, the number of ewes that required further care, defined as sick or injured, was recorded.This protocol was tested on 32 commercial sheep farms in Victoria, Australia. For the protocol animal-based indicators were considered to be the most important, but the authors state that some relevant management- and resource-based indicators, such as nutrition management or shelter provision should also be included in future assessments. According to their judgement a combination of animal-, management- and resource-based information could lead to a better understanding of potential problems for sheep welfare and how they could be avoided or minimised best. This protocol seems as well to be able to identify and assess the main sheep welfare issues as the first one but with fewer indicators [10].3.2. Single IndicatorsThirty welfare indicators could be extracted from the scientific literature for which data on validity, reliability or feasibility exists in multiple articles. They are listed and discussed below from the most frequently cited to the least. Further 23 indicators were cited only once [23,24,25,26,27,28,29,30,31,32]. Thus, the amount of information is too small to judge the value of these indicators in terms of a general applicability. These indicators are listed in Table 2. Further research is needed to estimate the validity, reliablity and feasibility before these indicators can be recommended or rejected from inclusion in a general welfare assessment protocol for sheep.3.2.1. Behaviour AssessmentThe most suitable stress assessment for routine on-farm checks seems to be a behavioural observation [6,33]. For example, feeding or rumination behaviours have been suggested by experts [6] as good indicators of positive conditions in sheep [30]. However, neither the method of assessment nor its reliability have been described in these articles.Categorizing animals as “obviously sick” would allow an overall impression [23], as sheep suffering from welfare issues can be recognised through their dull, depressed demeanour [6,26]. Conducting this kind of observations, in sheep or in lambs, has proven to be feasible on-farm as the animals do not need to be gathered or handled [26,32] and showed a promising level of intra- and inter-observer reliability [24,32].To gain a systematically assessed insight into the animal quality of life, the qualitative behavioural assessment (QBA) chose to assess how an animal demonstrated a behaviour rather than the behaviour itself. The focus of this method is to keep the whole animal perspective, and to assess observed details of posture and behaviours in the light of the entire animal’s interaction with its environment. To do so, a list of characteristics, such as content, sociable, playful or irritable, can be prepared in advance or developed by the observers themselves [34].This assessment has been tested at many different levels and may be the most promising indicator for assessing positive emotional state in sheep, as it is considered both valid and feasible [4]. It has been used to assess sheep’s behaviour via video [35], on-farm [36,37] and during transport [38]. According to the studies, QBA has the potential to serve as a sensitive, meaningful indicator for assessing sheep welfare due to its feasibility, reliability and correlation with physiological responses. In addition to providing an overview of the animal’s behaviour, QBA appears to allow the identification and monitoring of sheep with intestinal worms and those requiring treatment [28]. The background information on the farm given to observers, did not substantially affect the relative rankings of animals on the main expressive dimensions (i.e., the pattern of interpretation), but did sensitise observers to certain aspects of the observed sheep’s expression. Therefore, in accordance with all the studies cited above, the need for good training for observers prior to the assessment was pointed out [39].3.2.2. LamenessLameness is a significant problem affecting young and growing lambs as well as adult ewes and rams. As any production group can be affected, the presence or absence of lameness seems to be a good indicator to include in an animal welfare protocol [6,40]. Several lameness and gait scoring systems have been developed using different categories [41,42]. Even if all scoring systems produce a fair to good level of inter- and intra-observer reliability [25,26], a binary scoring scale that rates the animal as “healthy” or “lame” appears to be the most reliable and practical method for sheep [24,43] and lambs [32]. When used on-farm, a group assessment appears to be more feasible and shows a slightly higher percentage of lameness detection [26,43]. Reference [43] suggests that the higher prevalence in group assessments originate from an increased difficulty in detecting lameness in sheep that are stressed from an inspection in isolation. In conclusion, each of the aforementioned studies concludes that this indicator is a robust and viable tool for on-farm assessment and recommend its inclusion in the animal welfare protocol. Lameness as a welfare indicator was also included in all the above discussed protocols. 3.2.3. Body Condition Score (BCS)The BCS is a scientific measure for assessing the degree of fatness or condition of the animal using a descriptive score from 0 to 5. It is easy to learn and use and requires no equipment [44]. Even if the assessment needs handling of the sheep, the BCS shows a good on-farm acceptance and may be the most direct method of assessing persistent hunger in sheep [4]. BCS varies throughout the production cycle and knowing how BCS changes during the shepherding year allows the identification of individual animals with welfare or health problems [40]. This indicator appears to be a valid quantitative predictor of animal welfare [45] as well as a monitoring tool for selective treatment of internal parasites as part of the Five Point Check© [27]. The method shows good inter-observer reliability [25,26], which could still be improved by simplifying the scale to a fit-fat-thin score [6,26]. Because the method is based on a subjective assessment [44], the need for good training is of vital importance [32]. In short, all studies recommend the inclusion of a BCS or a fit-fat-thin assessment in animal welfare protocols. 3.2.4. Faffa Malan Chart (FAMACHA©)The Faffa Malan Chart (FAMACHA©) system is a colour chart for the non-invasive detection of anaemia in small ruminants [46] and for targeted selective treatment of gastrointestinal parasites, as part of the Five Point Check© protocol [27]. The FAMACHA© chart shows no interrelationship with faecal egg count, but has been shown to correlate with haematocrit [47] and, therefore, seems to be a valid indicator for Haemonchus sp. [48] as well as adult Fasciola hepatica and could be used to identify sheep with high established fluke burden [49]. However, the chart shows low sensitivity in growing lambs, with an accuracy level of only 50% in identifying lambs in need of treatment [50,51] and should not be used alone to control haemonchosis in young animals [52,53]. In conclusion, the FAMACHA© chart correlates with haematocrit and could therefore be used as an indicator of anaemia in sheep. This method may be relevant to identify blood-feeding gastrointestinal parasites such as Haemonchus sp. and adult F. hepatica, but only in adult sheep.3.2.5. Ears PosturesThree pain assessments for sheep or lambs could be found in the literature: the Sheep Pain Facial Expression Scale (SPFES), the Sheep Grimace Scale (SGS) and the Lamb Grimace Scale (LGS). All scales assess expression in different facial areas that are rated in three categories of abnormal expression “absent”, “partially present” or “present”. The SPFES shows a high degree of accuracy in detecting suffering sheep. According to the observers, SPFES is easy to assess, and their study showed a high inter-rater reliability and high consistency. The SPFES seems to provide a reliable and effective method for assessing pain in sheep after minimal training [54]. The SGS was also shown to be a valid and reliable method for identifying distress in laboratory sheep [55]. In contrast, the LGS results should be taken with caution due to the small number of lambs (only nine) used in the study. Nevertheless, the LGS score increased significantly from before to after painful interventions had been carried out, while the score of the control lambs remained the same. These results suggest that trained human observers were able to apply the LGS and distinguish suffering lambs from control lambs [56].Both the LGS and the SPFES consider ears that are tense and pointed backwards or downwards as a reliable sign of pain in sheep [54,56]. Ears pointed backwards could, however, as well be a sign of an uncomfortable situation or fear [57]. The SGS describes a slightly different scale with erect ears as a sign of no pain, flattened ears as moderate indication of pain, and hanging ears as severe pain [55]. Yet, when sheep are being brushed, horizontal and backward ears with only few ear posture changes seem to reflect a neutral or even positive state [31,57,58,59]. The breed characteristics may also be an important factor in interpreting ear posture, as ear posture may vary between breeds [31]. Nevertheless, changes in ear position should remain the same [54]. Because of the conflicting reports and difficulty to interpret ear postures, further research is needed to determine its usefulness as an indicator for the wellbeing of sheep.3.2.6. Eye ApertureEye aperture or orbital tightening has been suggested as an indicator of positive emotions as well as pain in sheep and lamb depending on the situation in which it is observed. This feature has been observed during brushing and shows that sheep seem to close their eyes while experiencing positive emotions [30,31]. Eye opening correlates well with cardiac measures and would be readily applicable on the farm using descriptive categories, such as “wide open” and “half closed” eyes [58]. However, eye opening is also a component of three pain scales, namely the SPFES, SGS and LGS under the name “orbital tightening”. All three interpret the “squeezing” of the eye or the narrowing of the eye aperture as a sign of pain [54,55,56]. The eye aperture seems to be a valid component of the pain scales, but not on itself, as it can indicate a state of well-being as well as a state of pain.3.2.7. Comfort around HumansSheep’s alertness to approach in the field has been recognized as a potential welfare indicator [6]. A common assessment constitutes the human approach test. This test involves observing the animal’s reactions when approached by a human. Behaviours such as escape attempts or aggression are typically expressions of fear. These reactions are seen as possible indicators of discomfort around humans [60]. It is debatable, whether the approaching human should be familiar to the sheep (which might be more relevant in terms of welfare when sheep are in daily contact with their keeper) or not familiar (which might be better standardized across farms). Another way to assess the animal’s comfort around humans might be the fear test [4]. This test is based on observing the behaviour of sheep in the presence of an unmoving human [22]. It has been used to detect fear behaviours in lambs, such as inhibition of feeding, long distance from the frightening stimulus, frequent immobilizations, and numerous high-pitch bleats [61]. Both the human approach test and the fear test performed with indoor ewes have the potential to be used for on farm welfare assessment. However, both require further work to develop the details of the methods and to assess the reliability of the test [4].3.2.8. Fleece ConditionPresence or absence of wool loss and fleece condition have been suggested by stakeholders as indicators of well-being in sheep [6,26]. Fleece condition can be a strong early indicator of the presence of aphids [62] or ectoparasites such as Psoroptes ovis, which can have a significant negative impact on sheep welfare [24]. These indicators appear to be more reliable and easier to assess than pruritic behavior [62]. Furthermore, group assessment via fleece condition appears to be reliable, yet further research is required to determine the optimum group size, as closer observation of individual animals may be required to identify areas of wool loss [24]. The indicator shows high inter-observer reliability at every production stage [25]. Fleece condition assessment was judged to be feasible and suitable for inclusion in sheep welfare protocols by all the studies mentioned above.3.2.9. Faecal Soiling or Dag ScoreFaecal soiling may occur as a result of a complex interaction of factors, such as gastrointestinal infections [63] or high-quality spring grass [4]. The proportion of faecal soiling correlates with faecal egg counts and therefore with worm burden. The degree of faecal soiling can be assessed by scoring the animal according to the size of the region soiled around the breech; using a dag score between 0 and 5, where a score of 0 represented a clean breech region and 5 described a breech region where faeces adhered to more than two thirds of it [63]. The dag score is part of the Five Point Check© for selective treatment of internal parasites in small ruminants. South African farmers consider the dag score as understandable and useful for worm causing diarrhoea [27]. Depending on the study, its reliability varies from poor to high, but it has been recognized as rapid, non-invasive [63] and feasible [24] and should be included in animal welfare protocols [25] at least because faecal soiling is a risk factor for fly infestation and therefore remains relevant for sheep welfare [4].3.2.10. Skin LesionsSkin lesions or wounds are considered highly valid welfare indicators as they provide a direct assessment of the presence or absence of injuries [4,6]. Large skin lesions are easily observed, but small lesions are more difficult to identify [23] and may be hindered by the presence of wool. References [25,26] suggest skin lesions to be assessed on the entire body, even turning sheep over. In fully fleeced sheep, inspection is performed by parting the fleece and by palpating the skin. Therefore, handling of the animal is required to allow an efficient examination of the animals. Reference [23] states the possibility to identify ectoparasites from lesions, as the extensive scratching and biting of infected areas may cause wounds. However, the validity and reliability of such recognizing ectoparasites through lesions are, to date, unknown. The assessment can easily be performed [4] and seems to be reliable [25,26]. Based on these results, Reference [25] recommends the inclusion of skin lesion assessment in welfare protocols for sheep [25].3.2.11. Tail Docking and Tail LengthTail docking is considered a painful procedure [6] and risk factor for infections if the procedure is poorly performed [20]. Therefore, the tail length is a key welfare issue [6] and an indicator of preceding poor welfare [23]. In line with this conclusion, a group of experts suggested tail length as an indicator of sheep welfare [6]. This measure is feasible on-farm, where it has a good reliability and can be assessed with a binary score: 0 = tail covers the anus in males or vulva in females, 1 = tail is over-shortened [25,26]. In addition, experts suggested a management-based indicator to be more feasible by recording whether tail docking was practised, and if so, which method, analgesia and anaesthesia were used, rather than measuring the pain responses of the lambs [64]. The tail length seems to be a robust and feasible indicator to include in animal welfare protocols [25,26].3.2.12. Fleece CleanlinessFleece cleanliness measures the extent of soiling from external sources, such as rain, mud or dirty pens, whereas, faecal soiling should be assessed as a separate indicator (see Section 3.2.9). The fleece cleanliness seems to be a promising indicator of sheep’s environmental status that can be used in further animal welfare protocols [4,23]. It achieves a good level of inter- and intra-observer reliability [25,26]. As to how exactly fleece cleanliness is recorded, the available information is scarce. Reference [25] considers the whole body, using a 4-point visual assessment, whereas Reference [26] assessed the ventral abdomen with a 3-point visual scale. References [4,23] refer to fleece cleanliness in a more general nature. Given the few sources of research, it is not possible to adequately compare the practices regarding the ideal approach to assess cleanliness. Nevertheless, this measure is easily feasible because it does not require the animals to be gathered and handled and can be performed on undisturbed animals in their home environment [4].3.2.13. MastitisMastitis may be a useful indicator of ewe welfare and health and can be assessed using a variety of methods. For example, the California Mastitis Test is considered a good diagnostic technique [65]. Another example is udder palpation. Mammary glands can be palpated to identify areas of focal or diffuse thickening, swelling, heat, pain or discomfort. They can be scored as “no evidence of mastitis”, “one gland” or “both glands affected by mastitis”. This method achieved good inter-observer reliability and is considered feasible [26]. However, Reference [25] remarks that udder examination and milk samples collection are time-consuming and labour intensive, making them less attractive for on-farm use. An alternative proposed by Reference [66] would be to use altered lamb and ewe behaviours. On one side, lambs show a preference to suckle on the unaffected gland. On the other side, ewes show an increased vocalisation and prevent their lambs from suckling more frequently when affected by mastitis. This change in normal behavioural pattern could be observed as early as 3 days after infection.3.2.14. Pruritic BehaviourSelf-traumatising behaviours such as scratching and rubbing appear to be useful observations for assessing welfare in sheep infested with ectoparasites such as Psoroptes ovis [67] as well as Bovicola ovis [62]. The time sheep spend rubbing themselves correlates positively with the total lesion area and the number and age of lesions. The amount of rubbing behaviour increased with age and lesion size. However, larger lesions were associated with a decrease in the frequency of standing-up attempts followed by a rubbing attempt. This suggests that other factors associated with lesion development may affect rubbing behaviour. These factors include increased pain and discomfort, which may also interfere with the lying behaviour of infested sheep [68].3.2.15. Diarrhoea ScoreThe diarrhoea score (DISCO) is used to describe the sheep faeces with a score of 1 corresponding to normal sheep faeces in pellets, 2 for “soft” faeces (similar to cow pat) and 3 for diarrhoea (semi-liquid faeces) [69]. Presence of diarrhoea seems to be a valid indicator with a significant relationship to the intensity of intestinal parasite infestation in lambs. This score allowed to correctly identify 80% of the animals in need of treatment [50]. The DISCO score was lower in healthy animals or those infected only with nematodes than in sheep infected with cestodes. It also correlated with the number of cestode but not nematode eggs per gram of faeces (EPG) [48]. According to Ref. [47], DISCO should not be used to detect early infection with H. contortus as it does not reflect the intensity of infection nor is it consistent with faecal egg counts.3.2.16. Weight GainReduced weight gain can be associated with intestinal parasite infections. According to Reference [53] daily weight gain in lambs can be effectively used to identify lambs in need of treatment. In contrast, in Reference [50], reduced weight gain is described as not useful and without association to any other pathophysiological indicator relevant to the diagnosis of intestinal parasites. Reference [47] questioned the accuracy of weight gain reduction as it does not correlate with faecal egg count and cannot reflect the intensity of H. contortus infection in sheep. This measure needs further research to clarify its usefulness.3.2.17. Rumen FillA panel of experts identified rumen or abdominal fill as an animal-based measure of access to feed [6]. It was scored on-farm using a simple binary scale: 0 if the animal’s left-hand side was not sunken/or was convex between the hip bone and the ribs and 1 if the animals’ left-hand side was deeply sunken. The results showed a poor reliability, probably due to the difficulties to assess the rumen fill on sheep with a lot of fleece [25]. For lambs, the same indicator showed good inter-observer reliability, but due to the close observation required, 96% of lambs kept outdoors could not be scored [32]. Therefore, depending on the housing conditions the results from this indicator should be interpreted with caution.3.2.18. Excessive PantingExcessive panting has been identified as an animal-based, non-invasive and feasible indicator for use under farm conditions [6] to assess thermal comfort [4]. Excessive panting is defined as a rapid breathing with abdominal effort, with or without rasping noise or open-mouthed stance. This indicator could be assessed without gathering the sheep, making it easily feasible. Yet, the respective study could not investigate its reliability, as no sheep were showing this behaviour [24]. Therefore, the relevance of such an indicator is debatable, and a validation should occur with herds with known suboptimal thermal comfort. Excessive panting is a specific indicator for heat stress when measured in undisturbed animals. Under other conditions, increased respiration rate may be an indicator of distress [4]. In conclusion, the reliability of excessive panting as a welfare indicator still needs to be tested on-farm.3.2.19. Eye ConditionEye condition or abnormality has been suggested by a group of experts as health and welfare indicator for sheep [6]. An abnormal eye condition was deemed to be present if any one of the following signs was observed—blepharospasm, corneal opacity, abnormal ocular discharge, lacrimation with tear-staining of skin, conjunctivitis, or entropion. After an on-farm test and although the sheep had to be restrained for the evaluation, the assessment of eye condition was declared feasible. However, due to the small number of sheep involved, reliability could not be assessed [26]. Reference [32] uses the same indicator for lamb, but because of the close observation required for assessment, 96% of the lambs kept outdoors could not be assessed. Nevertheless, eye condition showed an excellent level of inter-observer reliability as well as a high sensitivity and specificity. Abnormalities were clearly identified. Therefore, the authors suggest that eye condition is a highly relevant indicator and should be included in future lamb health and welfare inspection tools. Eye condition seems to have an excellent level of sensitivity and specificity but needs to be tested on a larger sample size.3.2.20. VocalisationSheep vocalise during social isolation, depending on breed and age class [29] and remained silent while being brushed. Considering that brushing is perceived as a positive stimulus, vocalisation could be an indicator of negative welfare [31]. Vocalisations have been shown to be associated with negative emotional reactions and have a strong correlation with the sheep activity levels [70], which could make this indicator a good predictor of an active sheep reaction to an anxious situation. Although it needs further standardisation and validation, vocalisation would be easy to assess and seems to be a valid measure of animal welfare [29].3.2.21. Mouth FeaturesMouth features are used as pain indicator and are included in three different pain scales discussed before. In the SGS, pain is assessed using three levels: (i) “closed mouth” indicating absence of pain, (ii) “puckered lips” indicating moderate pain and (iii) “flehming” representing the higher level of pain. The validity and reliability of mouth characteristics were not assessed separately from the other indicators of the SGS, including orbital tension and ear and head position [55]. In the SPFES and the LGS, the indicator is defined as flattened and tight lips with straight or slightly ventrally rotated corners. The mouth features alone do not appear to be reliable as an indicator of pain due to the low observers agreement [54,56], but may be useful as part of the various pain scores.3.2.22. Cheek FlatteningSimilar to mouth characteristics, cheek tightening or flattening is included as an indicator of pain in the SPEFS as well as in the LGS. For the SPFES, Reference [54] defines cheek flattening as a more convex expression of the cheek in the region of the masseter muscle and zygomatic arch and scaled this characteristic as absent, partially present or present. Cheek tightening appeared to be relatively easy to score and showed a high inter-class correlation of 82%. Reference [56] characterises less bulging cheek area or, in obvious cases, a hollowed cheek as indicators of lambs in pain. According to their observers, cheek flattening was a difficult feature to assess due to differences in camera angle or lighting. This characteristic also had a low inter-observer reliability, suggesting that this action unit contributed little to the pain assessment and therefore could be excluded from the LGS. Therefore, cheek tightening may be a useful indicator within the SPFES to assess pain in sheep, but not in lambs.3.2.23. Nasal FeaturesThe last facial expression included in the SPFES and the LGS are nasal features. According to both, References [54,56], sheep or lambs in pain showed a tightening nose with a decrease in nostrils, resulting in a “V” shape. Although they agree on the validity of this indicator, the results of their study diverged on the reliability. McLennan et al. found that the nose features did not correlate strongly with the other areas of the face and that this indicator was less reliable between scorers than the other measures of the SPFES [54]. In contrast, Guesgen et al. showed a good inter-observer reliability for nose features in lambs. However, they pointed out that restraining lambs affected their facial expression and influenced the measure of that feature [56]. These differing opinions suggest that this indicator should be interpreted cautiously and needs to be confirmed in future studies.3.2.24. Hoof OvergrowthHoof overgrowth has been cited as an indicator of sheep welfare to assess ease of movement. However, currently, there are no studies, directly linking reduced movement to an increase in hoof overgrowth. This measure depends on other factors, such as the frequency with which hooves are trimmed and the ability of the animal to move if it suffers from lameness [4]. A recent study evaluated hoof overgrowth in terms of ease of application, but found poor reliability and low feasibility, likely due to the difficulty of the assessment. According to the observers, assessing the hoof overgrowth was time-consuming and not easy to do, because ewes would not stand still. The authors suggest to use broader measures, such as lameness scoring (see Section 3.2.2), which may be more relevant [25].3.2.25. Nasal DischargeNasal discharge has been suggested by a group of experts as a non-invasive and practicable animal-based indicator for use under farm conditions [6]. The measure is part of the Five Point Check© protocol developed for the selective treatment of internal parasites in small ruminants. In this protocol nasal discharge serves as an indicator of nasal bots such as nasal botfly or lungworms. Note that nasal discharge can also be an indicator of pneumonia or other diseases [27].3.2.26. Tail FeaturesTail wagging and raised up tail are controversially discussed indicators. On one hand, because they are rarely observed in sheep, especially with tail docking, and on the other hand, because scientists do not agree on their meaning. Lambs raise and wag their tail while suckling and being brushed. Assuming that both are positive stimuli for sheep, they may be important indicators of positive states in sheep [30]. However, lambs seem to show raised tails during separation with their ewe, which might indicate that this behaviour occurs during intense negative emotional states. According to this contradictory information, the raised tail may be shown during a negative or positive emotional state, which would render this indicator useless for discriminating emotional valence [59].3.2.27. Lying TimeThe assessment of lying time for individuals was proposed to measure either the comfort of the resting places or an infestation with ectoparasites (e.g., Psoroptes ovis). With the aim of measuring the comfort of the resting areas, Reference [4] concludes, based on the available literature, that the measure was difficult to apply in the field. The study authors mentioned that the lying synchrony of the sheep would provide sufficient information in a simpler way. The possibility of all sheep lying at the same time can be easily assessed without disturbing the animals. However, the reliability of sheep lying synchrony has not been assessed yet. The lying time may as well be an indicator of ectoparasites as sheep infested with P. ovis spend less time lying down at the expense of rubbing time. The development of lesions (e.g., secondary bacterial infections) may also influence lying behaviour [68].3.2.28. ShiveringShivering is known to be a sign of cold, which would make it a potential welfare indicator for thermal comfort. Two studies have tried to use it in both sheep and lambs and came to the same conclusions: shivering had a very low prevalence and showed a low level of inter-observer reliability, possibly due to the presence of fleece, which makes it difficult to assess. Authors of both studies considered this measure unfeasible for sheep [4,32].3.2.29. Rectal TemperatureRectal temperature is commonly used in clinical examinations and provides useful information about the animal’s health status. This measure has been proposed as an indicator for positive welfare in sheep, but has been discarded due to the lack of significant matches with positive states [31]. Handling is required to take the rectal temperature of a sheep, which could cause stress and stress-induced hyperthermia, which could ultimately bias the results. Moreover, the invasive nature of this measure may compromise biosecurity [4]. In conclusion no study supports the use of rectal temperature as a welfare indicator for sheep.3.2.30. Tooth LossAssessing tooth loss or dental abnormality could give an indication of the sheep’s ability to feed and could allow animals at risk to be identified earlier. Even if the assessment requires handling of sheep, this procedure is quick, simple and frequently performed on-farm, suggesting good feasibility. The reliability of this measure has not been tested [4] but the assessment of tooth loss or dental abnormalities was found to be feasible [26].4. ConclusionsThe aim of this study was to review the scientific literature published from January 1995 to March 2020 to obtain an overview of the articles available linked to sheep’s welfare and to extract animal-based welfare indicators as well as already established welfare protocols. For this review, a total of five protocols and 53 indicators were identified. All the protocols include animal-based indicators validated in the literature and seem feasible on-farm, that is, they need limited resources, effort and can be applied with little disturbance for the animals. However, all of them have yet to be tested on a larger scale and bigger sample sizes to be able to affirm their reliability for providing a consistent and truthful reproduction of the status of sheep welfare.For individual indicators, the amount of data is greater than for entire protocols. This is owed to the fact, that most protocols relied on expert and stakeholder opinion to determine the included indicators. This practice to determine indicators through expert and stakeholder opinion in turn directly resulted from the high variation in the availability of research for the different indicators. Some indicators, such as behaviour assessment, lameness, BCS, fleece condition or skin lesions are frequently addressed in the literature and have acquired a status to be useful indicators for measuring sheep welfare. Others such as FAMACHA©, dag score, DISCO or the various pain assessments and their components are regularly mentioned in the literature, but opinions differ on their validity or feasibility. Finally, some of the indicators mentioned in this review, such as pruritic behaviour, eye condition, lying time or tooth loss, are relatively new and seem feasible, but their validity and repeatability has not yet been assessed in-depth. It may be possible to derive some priming information for these indicators from established indicators in other ruminants. Rumen fill and rumination behavior for example is rather well studied in cattle, but less in other ruminants. Although a direct comparison between sheep and cattle is not possible, research to establish welfare indicators may profit from prior knowledge as to which parameters to look for and which methodologies may be practical.In our search terms we also specifically included terms for data-based indicators. Nevertheless, our literature search found no studies explicitly investigating data-based indicators, and only one assessment protocol includes three measurements on production records, yet, without clear results or a discussion on their quality. Given the increasing efforts for simplified welfare assessments, more research should be directed towards identifying useful, reliable and feasible methods to indicate the status of animal welfare from the ever growing stack of available data. For the time being, this literature review should serve as a starting point for further development of comprehensive, valid and practicable on-farm welfare protocols for sheep, which could also be used to validate future implementations of data-based indicators.

animals : an open access journal from mdpi

10.3390/ani13081329

PMC10135104

Racehorse welfare has generated public attention for both on-track treatment and practices, as well as the fate of racehorses upon completion of their racing careers. As such, the successful transition from racing into sport horse and other disciplines is paramount to the welfare of retired racehorses whose racing career lasts, on average, just 4.5 years. Thoroughbreds are frequently used in sport horse disciplines, and the consistent supply of retiring racehorses provide an affordable and athletic breed option for equestrians. This study analyzes demand for thoroughbreds offered for sale through online, sporthorse auctions. The results show buyer preferences for coat color, age, sex, and those registered with USEF, USHJA, or USEA. These findings support the value of thoroughbreds to buyers interested in sport horse disciplines. The results from this study may help non-profit, off-the-track thoroughbred rehoming facilities, reduce the number of unwanted thoroughbreds, and improve the public image of welfare in the racing industry.

Racehorse welfare is a prominent, public issue which affects nearly every aspect of the racing industry. Thoroughbred care after race career completion has garnered increasing attention from the equine industry, general public, and animal welfare groups alike. As the average racehorse’s career lasts just 4.5 years, owner demand for thoroughbreds is essential for post-race careers and acceptable welfare standards. This study utilized data from and hedonic pricing models to analyze buyer demand for thoroughbreds sold in online auctions held from 2012 to 2020. The results indicate buyer preferences for age (p < 0.01), sex (p < 0.05), and organization registration (p < 0.05), with bid price premiums for age and registration status (USEF, USEA, USHJA, etc.) and price discounts for mares compared to geldings and horses listed for non-competition careers (trail, p < 0.01). The results of this study confirm and quantify the value potential buyers place on thoroughbreds offered for sale in sport disciplines. The findings may help non-profit organizations charged with rehoming off the track thoroughbreds and reduce the number of unwanted thoroughbreds by illustrating the desired traits and skills within the equine market, thus improving welfare optics overall.

1. IntroductionHorse racing is prevalent within the United States, and given the breed restrictions within racing varieties, many purebred thoroughbreds are initially destined for a racing career. In 2018, the American Horse Council estimated that of the 1.2 million horses involved within the racing sector of the United States Equine Industry, over 500,000 of them were thoroughbreds [1]. Furthermore, between 2016 and 2021 The Jockey Club registered an average of 21,000 thoroughbred foals each year and reported nearly 50,000 actively racing thoroughbreds within the United States [2]. A thoroughbred racehorse’s career averages approximately 4.5 years [3], while the average lifespan of a horse is between 20 and 30 years [4]. The actual number of thoroughbreds retiring from the racetrack not selected for racing breeding programs is not known. The Equine Welfare Data Collective aggregates data from self-reported rescue, rehabilitation, and rehoming agencies, but has not yet released reports by breed [5]. For perspective, CANTER has facilitated the rehoming of over 25,000 retired thoroughbreds through either CANTER owned horses or by facilitating trainer/owner horse sale listings [6]. Current industry circumstances, including the restriction of the USDA inspection of equine processing facilities, high numbers of unwanted horses, and scrutiny of equine welfare from the general public and industry members alike [7] have increased the need for marketable skills and options for thoroughbreds that are not or are no longer utilized for racing. Furthermore, the restriction of USDA inspection of equine processing facilities and the prevalence of thoroughbreds entering the slaughter pipeline can complicate welfare implications for horses in post racing careers, despite efforts from tracks and racing jurisdictions to prevent horses from being sold for processing purposes [7,8,9,10].Attention on the post-racing welfare of thoroughbreds has increased and created various outlets for placement. The scrutiny of horse racing itself poses unique challenges for second career placement. In a poll assessing welfare, injury incidence, fatalities, and other pressing issues affecting the racing industry, 31% of United States residents likely to vote in upcoming elections felt that racehorses are not treated in humane ways [11]. Securing care for post-racing Thoroughbreds can be challenging due to the number of Thoroughbreds leaving active racing each year, available adoption/rehoming programs, and industry demand for purchasing horses for alternative careers. Furthermore, horses that are not marketable within available outlets may face poor welfare situations, which complicates the tenuous nature of the perception of horse racing [12]. The Jockey Club encourages the adoption of retired racehorses who are not selected to enter breeding programs and endorses their suitability for second careers (i.e., sport disciplines, barrel racing). Various placement or rehoming programs have incorporated training, certifications, or incentives for individuals and organizations who work with off-the-track thoroughbreds (OTTB), such as The Thoroughbred Aftercare Alliance, Thoroughbred Charities of America, and the Thoroughbred Incentive Program [13,14,15]. For example, the Thoroughbred Incentive Program offers thoroughbred only competitions to encourage equestrians to purchase and retrain retired racehorses, even hosting championships in disciplines such as hunter, jumper, dressage, and western. The addition of marketable skills through alternative discipline training is useful for increasing desirability; however, the value of additive skills is largely unknown.While existing research has investigated the demand for retired/OTTB offered through adoption facilities [16], there is limited information on price determinants for thoroughbreds sold outside of race-specific auctions (such as Fasiq-Tipton and Keeneland) and sales [17,18,19,20]. Understanding market demands can aid in identifying the desired characteristics for thoroughbreds entering alternative disciplines, which may in turn improve welfare for this breed sector by better aligning buyer demand with available market supply [7]. Thus, the objective of this study was to assess buyer demand and price determinants for thoroughbreds offered for sale through non-racing online auctions.2. Materials and Methods2.1. Data CollectionData on thoroughbreds was collected from 39 online auctions offered through an online auction service [21]. Auctions were held between 2012 and 2020, and data were accessed between April and December 2020 through publicly available archives. All auctions within this date range were examined for the listing and successful sale of thoroughbred horses, with the exception of breed specific sales that would preclude the entry of thoroughbreds (such as the Chincoteague pony or Rancho Corazón auctions). The thoroughbreds represented in these auctions comprised all age ranges, including both thoroughbreds transitioning into second careers and those who have already transitioned. It was not possible to determine which category each horse represented and when each horse retired from the racetrack based on the reported sale information. Data were collected to create a dataset on equid and auction characteristics including age, sex (gelding or mare, no stallions listed), colors recognized by the Jockey Club (bay/brown/black, chestnut and bray/roan), height, binaries for pre-sale veterinary information, and for any health concerns including major surgery, past lameness, other defects/abnormalities/blemishes, and cribber/weaver. Insufficient observations for dressage level and jumping height suffered from low reporting, and as a result were not included in the model. Information on equid training was collected and included professional memberships with the Jockey Club, the United States Equestrian Federation (USEF), the United States Hunter Jumper Association (USHJA), and the United States Eventing Association (USEA) (Jockey Club registration and USEF/USHJA/USEA membership) as well as disciplinary training (Dressage, Jumper, Hunter Under Saddle, Trail, and English General Riding). In some instances, more than one membership or disciplinary training classification was listed, and thus was quantified accordingly. Insufficient observations for dressage level and jumping height suffered from low reporting, and as a result were not included in the model. Memberships provide information on second career skills the horse may already have been introduced to or trained for, and could augment specifically listed disciplines in the sale advertisement.In addition to horse characteristics, information presented on the sale advertisement is an important component for sellers in online auctions [22]. Online auctions may not provide the buyers the opportunity to view horses in person, so sellers attempt to overcome these limitations by providing images, video, and detailed information regarding the horse. Auction images were classified as Non-Action Shot, stationary photographs of showing the equid’s conformation through front, back, side, and hoof pictures, or Action Shot, depicting the horse moving or working. Quantities of each image classification were obtained. Additionally, video content within the sale advertisement was assessed, and when a video(s) was present, the quantity of videos was recorded. Finally, advertisements containing descriptions or notes were quantified by counting the number of lines of notes, which could include information such as racing history, show experience, athletic capabilities, or demeanor. Screen resolution was kept consistent among data collection to ensure the consistent quantification of notes between auctions and displays.Listings were accessed and available data was recorded for initial inclusion screening. Specifically, for a listing to be included in study analyses, it must have met the following criteria: (1) The breed was listed as thoroughbred only, with no crossed or mixed breeds, (2) a successful sale was indicated and a final sale price was listed in the advertisement, and (3) the horses sold were located within the United States. No restrictions were placed on age or sex for study inclusion.2.2. Statistical AnalysisAll statistical analysis including summary statistics and model results were calculated using Stata 17 software [23]. Specifically, summary statistics included calculations of means, minimums, maximums, and standard deviations for each variable. To understand the market for sport discipline thoroughbreds and provide the market with realistic value expectations to help the transition to a second career, a hedonic pricing model was used in accordance with previously published methods [22,24,25,26,27], where models are used for decomposing prices into individual attributes to help future market participants with regard to better price expectations. The empirical model used in this analysis is described in equation 1 below as:(Bid)= α + βHorse Characteristics + γ Auction Characteristics + ε,(1) where the bid price is regressed on a constant term (α), a series of horse characteristics and auction characteristics with vectors of respective coefficients (β and γ). Horse and auction characteristics are described in the data section and in Table 1. The independent error term is represented by ε. The models were tested for homoscedasticity and then remodeled with robust standard errors to address heteroskedasticity [28]. Significance was detected when p < 0.05, and trends were identified when p < 0.1. Where appropriate, data are reported as the mean ± SD unless otherwise stated. To avoid multicollinearity for binary variables in the model, Gelding, bay/black/brown, and English General Riding were excluded from the model. The excluded variables serve to provide base comparisons for sex, coat color, and discipline-independent binary variables.3. ResultsOf the 245 thoroughbred breeds listed for sale between 2012 and 2020, 57 did not have a successful sale bid price, 11 did not have a height listed, 5 were from Canada or had an unlisted location, and 2 were dropped because they were limited sample outliers. Thus, a total of 170 sale advertisements for thoroughbred horses met the inclusion criteria and comprised the study population.The thoroughbreds sold varied in physical characteristics and sale listing information (Table 1). Horses (n = 170) ranged in age (8.64 ± 3.68 years, min = 1 year, max = 20 years, n = 170) and were mostly mares (107; 63%). Bay/brown/black was the predominant color identified (n = 105, 62%), followed by chestnut (n = 44, 26%). Thoroughbreds in this study stood 164.48 cm ± 5.47 tall (min =147.32, max = 180.34). All horses offered for sale (n = 170) had at least one picture presented on the sale ad. Only 16% (n = 27) of the horses provided pre-sale veterinary information. In terms of second careers or experience potential, Jumper (n = 63) was most commonly reported, followed by Hunter Under Saddle (n = 54) and Trail (n = 43). The hunter jumping height was reported on 30 listings, where jump height ranged from 68 to 106.68 cm, whereas jumper height was reported on 17 listings and spanned from 70 cm to 142.2 cm. The dressage level was only reported in six advertisements, ranging from training, first, and second levels. Due to the low sample size within each jumping height or dressage level, these variables were not included in the hedonic model. Horse registration was indicated through The Jockey Club (n = 59) and USEF/USHJA/USEA (n = 17) membership. For this analysis, pictures were categorized into action pictures and non-action pictures, where at least one horse had no posted non-action or no posted action picture. Non-action pictures were used in 120 listings displaying 1.96 ± 1.71 pictures (min = 0 picture, max = 11 pictures). Action pictures were used in 140 listings which contained 0.82 ± 0.38 pictures (min = 0 pictures, max = 6 pictures). Videos were included in 154 listings, with some having as many as four videos. All sale ads included notes which ranged widely between a minimum of three to a maximum of 72 lines (16.62 ± 11.63 lines).Hedonic Model ResultsThe hedonic model results are presented in Table 2. The age of the horse when auctioned was statistically significant (p < 0.01) and had a positive effect on bid price (USD 430.50); however, the squared age term indicates that this increase is at a decreasing rate (USD −22.94, p < 0.01). Holding other horse characteristics constant, a horse’s sale price increases until price peaks at age 9 and then begins to decrease thereafter. Gender influenced bid price (p < 0.05), where mares were found to be valued an average of USD 924.08 less than geldings. A trend for chestnut and gray or roan horses to be valued over USD 1000 (p < 0.10) more than bay, brown, or black horses was detected.Registration memberships to The Jockey Club and USEF/USHJA/USEA help inform owners of the horse’s current or past experience. Membership to the USEF, USHJA, or USEA contributed to Thoroughbred value, adding USD 1967.33 ± USD 820.24 to the bid price (p < 0.05).The type of disciplines the horses were either trained in or were prospects for included dressage, jumper, hunter under saddle, trail, and English general riding. When compared to English general riding, trail was the only variable found to significantly (p < 0.01) influence final sale price, where horses with trail listed were valued USD 1599.46 less than English general riding horses on average.4. DiscussionGiven the scrutiny surrounding Thoroughbred racing and the Social License to Operate, any improvement on horse welfare, marketable skill increase, and long term placement of thoroughbred horses may help improve the overall perception of the racing industry sector [7]. While not all thoroughbreds within the United States have actively raced, much of the population may have been involved with the industry through initial breeding or production [1,2]. Even without a racing background, the thoroughbred breed is so closely linked to racing that negative welfare or publicity could be associated with the racing sector. Attention to aftercare of OTTBs through sales, rescue, retraining, and or retirement has increased in recent years, where factors relating to length of stay in rescue/retraining facilities and long term satisfaction with thoroughbred ownership have been evaluated [7,16,29]. This study assessed buyer demand and price determinants for thoroughbreds offered for sale through non-racing online auctions. A further understanding of the factors relating to owner demand, skills, and overall desire for thoroughbreds within the equine market plays an important role in understanding how horses can be marketed and sold or successfully placed and maintained long-term within new careers.Hedonic pricing models have been used in the literature extensively and specifically for equine market analyses. For example, within the thoroughbred industry, Hansen and Stowe [24] determined the impact of thoroughbred weanling characteristics on price when sold at auction, Neibergs [25] investigated the demand characteristics of broodmares, Stoeppel and Maynard [26] looked specifically at thoroughbred broodmares in foal and how their characteristics affected their auction price, and Stowe and Kibler [16] investigated the characteristics of thoroughbred racehorses adopted from nonprofit organizations. Studies pertaining to non-thoroughbreds at auction include Kibler and Thompson’s [22] analysis of the price determinates for stock-type horses and Taylor et al.’s [27] calculation of the price determinants of show quality quarter horses.While numerous factors can be considered important to individual buyers, this study determined age, sex, color, registration, and discipline to be significant factors in thoroughbreds’ online auction price. Buyers place increasing value on a horse as it ages though at a decreasing rate, indicating that as the horse ages, there will be a turning point, or an age, where value begins to decrease as the remaining years a horse can be used for competition becomes shorter [27]. Parameter estimates indicate this turning point to be at age 9, and horse value declines each additional year afterward. Price decline with age has been documented in other breeds, and displayed a similar peak in mare and gelding prices for quarter horses sold in quality auctions and horses sold at a Minnesota auction [27,30]. Based on hedonic modeling results by age 19, horse value is negative where, with other horse characteristics held constant, a horse would have to be given away. However, in many cases, a horse’s value will continue to be greater than zero, as other characteristics of the horse have value to horse owners (training, experience, etc.). This places the average racehorse retirement age at approximately 4.5 years, the halfway point to the optimal value based on age to potential buyers and, when time is spent retraining these horses for second careers, the horse is still considered to be at prime age. Findings from previous studies show the time until adoption of off-the-track thoroughbreds indicates that older horses spend more time at adoption facilities than younger horses, highlighting the importance of thoroughbreds retiring and transitioning to second careers at a younger age in order to attract potential buyers [16].The buyers showed a preference for geldings over mares, where mares were valued USD 924.08 less than geldings. This result contradicts findings from stock and quarter horse studies, where buyers placed premiums on mares compared to stallions and geldings [22,27], perhaps indicating a heterogeneity among segments of the equine sales market. There was a larger number of mares in this study, which is a function of the sex of the horses that sellers were offering for sale in this format over the study period. Coat colors of chestnut and gray/roan were preferred (USD 1071.07 and USD 1054.05, respectively) to bay/brown horses. Interestingly, gray thoroughbreds placed for adoption off-the-track were found to spend less time at an adoption facility, but were more likely to be returned to the adoption facility though the authors did not determine a cause for this occurrence [16]. This result is not surprising, as chestnut and gray or roan horses are not as common as bay, brown, or black horses. In this study, approximately 26% of the horses were chestnut and 12% of the horses were gray or roan, making up less than half of the total sample. Existing studies have found similar results in buyer preference for less common coat colors and/or patterns for stock type horses and wild mustangs [22,31], though color preferences for stock type horses contributed to price discounts for gray and sorrel horses [22], perhaps supporting a preference difference between disciplines.Horses registered with national breed or discipline registration are preferred by buyers [22]. While registration for The Jockey Club did not affect the bid price, registration within USEF, USHJA, or USEA was found to bring in a premium bid. This indicates that buyers place a premium on thoroughbreds with showing or competition experience, and registration with these organizations affords buyers the opportunity to look up past competition experience and placings. The lack of significance for thoroughbreds with Jockey Club registration indicates neither a preference for or against Thoroughbreds who were bred for racing, which may illustrate that buyers viewed all thoroughbreds similarly despite prior racing history or association with The Jockey Club.Buyers have shown a variance in how they value horses suitable for different disciplines [22,27]. In an assessment of sport horse sale outcomes, professional training did not increase the likelihood of a successful sale, but did increase the commanded price [32]. Furthermore, in the same study, horses with less experience reported in the sale advertisement were more likely to sell compared to horses with more experience reported [32]. Results from the current analysis indicate that thoroughbred horses listed with experience or potential in trail were found to be valued less than horses listed as English general riding, suggesting that auction site bidders in these online auctions could be searching for a more competitive horse, or that trail riding was viewed as a discipline that required less formalized training than other types. Thoroughbred horses with trail riding listed in sale advertisements also brought lower sale prices compared to other disciplines for use in adult recreational riding markets [33]. Additionally, lower market values were observed for stock horses sold in online auctions, where on average trail horses yielded nearly USD 1600 less than horses with other disciplines listed [22]. This connection is of particular merit, as thoroughbreds retiring from the racetrack often retire with an existing injury [34]. Injuries limiting the discipline or upper-level suitability for a particular discipline (eventing or jumping, for example) have shown to be less desirable to potential adopters, as these horses experience longer stays at adoption facilities [16]. As the auction site is named ‘SportHorseAuctions.com’, it is possible that buyers using this site sought horses with specific sport horse training, and this may have also influenced the bid price for horses listed with trail experience.Competitive dressage, hunter jumper, and jumper disciplines inherently contain distinctive levels or jumping requirements within shows and other event types. Unfortunately, within this study, insufficient observations prevented jumping height and dressage level from being analyzed for independent effects on price premiums; however, it is possible that these variables may influence overall price. The continuous nature of the jumping and dressage levels coupled with the limited sample with the information listed provide too few observations for an analytical estimation of its hedonic value. While we captured the effects of binarily reporting dressage, which was insignificant, future work could try to collect more complete disciplinary performance metrics to better understand their value. In other breeds, the number of championships or top placing at the American Quarter Horse Association World Show and number of futurity championships or placings yielded a higher sale price [27]. Similarly, for each additional point earned within registered shows during the same year as the sale increased the price premium by 2% [27]. Logically, as a horse’s ability to compete in dressage levels and jumping height progresses, the amount of training would consequently increase. Similarly, given the increased nature of horses in post-racing careers to develop performance limiting conditions, it may be of value to assess the performance level across competitive disciplines to determine how the competitive level affects premiums (eventing level, dressage level, jumping height, etc.) [22,29].The inclusion of photos, videos and lines of notes or descriptions were not found to be influential in the sale price; however, previous studies found lines of notes, back pictures, side pictures, action shots, and videos to be significant [22]. Of the equids listed for sale in the study, every sale ad included at least one photo, and the mean number of note lines was 16.62. Furthermore, the size of the advertisement with specifications to eighth or quarter page ads was an influencing factor to increase the sale price of horses sold for recreational disciplines [33]. It is possible that the results in this study could imply that videos, pictures, and descriptions are standard types of advertising information within online auctions, or play a larger role within different targeted disciplines within this venue.Pre-sale veterinary information could include health information along with the horse’s soundness, and can inform the buyer of any relevant health concerns [34]. When a seller discloses this type of information, the buyer can make a more informed decision pertaining to the horse’s health and the potential financial responsibility and management needs for that specific animal. While pre-sale veterinary information was insignificant in the current study, previously completed work has illustrated the differential impacts of health-related disclosures on sale prices. In thoroughbred horses with previous hospitalization, no differences were detected in racing sale price from foal through 2 years of age [35]. Conversely, the disclosed health information for thoroughbred yearlings influenced the price for horses that sold at lower prices [36]. Furthermore, individuals that adopted OTTBs reported increased musculoskeletal injuries and gastrointestinal, behavioral, and foot/hoof issues than non-OTTBs [29]. Within the reported sale advertisements assessed in the current study, a minimal amount of horses were listed with health concerns (16%, n = 27). The sellers may have elected not to disclose health information due to apprehension of potential sale impact; however, health disclosure is an important factor for buyers to fully understand the anticipated financial burden for known health problems.Given small sample size, individual health concerns (cribber, weaver, past surgeries, etc.) listed for each horse were categorized as one binary variable, thus, it is possible that buyers have preferences for avoiding specific health concerns or value horses with these attributes less, but individual impacts could not be investigated. In an assessment of online auction sale outcomes, negative horse traits (stereotypic behaviors, cribbing, weaving, stall walking, etc.) did not influence buyer demand for stock type horses [22]. Similarly, the incidence of stall walking or weaving did not influence the performance in active racing Thoroughbred horses [37]. While stereotypic behaviors are often regarded as undesirable traits within the equine industry and frequently result from poor welfare situations, additional work to investigate the connection between thoroughbred welfare, stereotypic behaviors, and non- or post-racing careers is warranted [37,38].5. ConclusionsThe successful transition of retired thoroughbred racehorses is paramount to the image and social license to operate for the thoroughbred racing industry. While successful transitions may be measured in various ways (e.g., the number of retired racehorses placed in non-racing homes), buyer demand for thoroughbreds who have already begun second careers may indicate the long-term successful placements of OTTBs and non-racing thoroughbreds. This study found that thoroughbred buyers have preferences for coat color, age, sex, riding discipline, and discipline registry status (USEF, etc.). Future works should consider the demand for thoroughbreds in second careers within the broader context of all breeds and disciplines (e.g., sport horses) and the implications of perceived welfare conditions. Further investigations of the length of racing careers and/or racing industry connections, individual health concerns, jumping height, and training or discipline level on buyer demand is warranted. It is not known how prior racing history, including length of racing career, racing industry connection, race earnings, etc., may impact buyer demand for OTTBs, and therefore future work could expand the scope of data collection to all information, not just what is listed on a sale ad.

animals : an open access journal from mdpi

10.3390/ani11082248

PMC8388356

The water buffalo is found in many tropical countries worldwide. In the current world scenario, where meeting the protein requirements of the population is one of the biggest future challenges, buffalo meat could be a good source of protein and other nutrients. Currently, very little information is available regarding buffalo meat quality attributes. Therefore, this study was designed to evaluate the effects of aging time and muscle type on meat quality attributes (pH, color, tenderness, water holding capacity, and sensory acceptance) of buffalo meat. The results showed that color, tenderness, and sensory attributes were improved with aging time; the suitable aging time required to enhance meat quality attributes in Longissimus lumborum and Gluteus medius muscles is 28 and 21 days, respectively.

The present study aimed to investigate the effect of wet aging on meat quality characteristics of Longissimus lumborum (LL) and Gluteus medius (GM) muscles of buffalo bulls. Meat samples from six aging periods, i.e., 0 day (d) = control, 7 d, 14 d, 21 d, 28 d, and 35 d, were evaluated for pH, color, metmyoglobin content (MetMb%), cooking loss, water holding capacity (WHC), myofibrillar fragmentation index (MFI), Warner–Bratzler shear force (WBSF), and sensory evaluation. The pH, instrumental color redness (a *), yellowness (b *), chroma (C *), and MetMb% values were increased, while the lightness (L *) and hue angle (h *) values showed non-significant (p > 0.05) differences in both LL and GM muscles in all aging periods. The cooking loss increased while WHC decreased till 35 days of aging. MFI values significantly (p < 0.05) increased, while WBSF values decreased; in addition, sensory characteristics were improved with the increase in the aging period. Overall, the color, tenderness, and sensory characteristics were improved in LL and GM muscles until 28 and 21 days of aging, respectively. Based on the evaluated meat characteristics, 28 days of aging is required to improve the meat quality characteristics of LL, whereas 21 days of aging is suitable for GM muscle.

1. IntroductionThe water buffalo (Bubalus bubalis) is found in many tropical countries globally, particularly in the Southeast Asia region [1]. These buffalo are primarily kept for milk purposes; however, they have excellent potential for meat production and play a vital role in the agricultural economies of many developing countries [2]. The water buffalo is an excellent converter of low-quality forage into good-quality meat, is resistant to many bovine diseases, and has excellent body weight gain, making it easier to manage them on locally available roughages [3]. Buffalo meat is almost similar to cattle meat; moreover, its meat has many superior characteristics, including higher protein, low fat, and cholesterol content [4]. In the current world scenario, where 1.4 billion people are protein deficient [5], buffalo meat could be a good source of protein and other nutrients, and it can serve a crucial part towards achieving the Sustainable Developmental Goal of the United Nations for food security for all.Meat quality is described by many attributes, including its color, tenderness, juiciness, flavor, and palatability. Among all these attributes, tenderness is considered the primary quality characteristic regarding eating satisfaction [6,7]. This is evidenced by consumers’ willingness to pay the extra price for guaranteed tender beef [8]. Meat juiciness positively correlates with tenderness and depends on moisture retention during meat processing [9]. Inconsistency in meat tenderness has always been a significant concern for the beef industry. One reason for this inconsistency is the individual variations among different beef muscles [10,11]. Meat color is another critical quality attribute, as consumers associate bright cherry-red color with the freshness and wholesomeness of meat [12,13].Researchers have developed various techniques to improve meat quality [14,15]. Post-slaughter interventions such as suspension methods, blade tenderization, exogenous enzymes, and conventional aging procedures are used to improve meat tenderness [16]. Among all these methods, postmortem aging is widely used by the industry to improve beef quality, particularly tenderness and palatability [8]. During postmortem aging, actin-myosin cross-linkages are broken down by different proteolytic enzymes that result in tender meat [17]. The aging process also improves meat color. The reduction in oxygen consumption rates with an enhanced storage period leads to the bright red color of meat [18]. Similarly, early activation of calpain-2 during the postmortem aging period coincides with improved meat’s juiciness [9]. Previous studies have performed the meat aging maximum until 28 days, and it is known that aging improves the tenderness during the first 14 days postmortem of cattle and lamb [19,20]. However, this study used the extended aging time (up to 35 days) because the proteolytic enzymes’ action is slower in buffalo than in cattle [7].Among aging methods, wet aging has an advantage over dry-aging due to less weight loss, juicier meat, and little or no shrinkage of meat at storage temperature. This is also an inexpensive and less time-consuming method [18,21,22]. Being high value meat muscles among consumers, Longissimus lumborum (LL) and Gluteus medius (GM) were selected for the current study; as these muscles differ in their biochemical properties and tenderness, they respond differently to postmortem aging [8,23,24]. Therefore, the present study was designed to investigate the effect of wet aging on the meat quality characteristics of Longissimus lumborum (LL) and Gluteus medius (GM) muscles of buffalo bulls.2. Materials and Methods2.1. Ethics StatementAll studies related to animals were pre-approved (vide letter No. DR/567) by the Institutional Ethical Review Committee, Office of Research Innovation and Commercialization (ORIC), the University of Veterinary and Animal Sciences (UVAS), Lahore, Pakistan.2.2. Animals SlaughteringBuffalo bulls (n = 18) of the Nili-Ravi breed were procured from Research & Development Farm, Big Feed (Pvt.) Ltd., Lahore, Pakistan. Animals were 24 months of age (average body weight 290 ± 10 kg), reared under similar management and feeding conditions. All the animals were transported to the lairage facility of the University of Veterinary and Animal Sciences, Lahore, Pakistan. The animals were off-feed for 12 h before the slaughtering to ensure hygienic processing with free access to water. The animals were randomly slaughtered at University commercial slaughterhouse facility following the Halal slaughtering guidelines mentioned in Pakistan Halal Standards PS3733. The average hot carcass weights were 140 kg (SD = 10), and carcasses were chilled at 0–2 °C for 24 h.2.3. Muscle Sampling and Aging ProcessingFollowing chilling, carcasses were deboned at 10 ± 1 °C, and both-sided Longissimus lumborum (LL) and Gluteus medius (GM) muscles were removed. Every LL and GM was cut into steaks (n = 9), each steak with 2.5 cm thickness, and tagged for identification [7]. All LL (n = 18) steaks and GM steaks (n = 18) of each carcass were mixed separately for randomization. Three steaks were selected for each aging period (0 d = control, 7 d, 14 d, 21 d, 28 d, and 35 d) from both muscles of the carcass. One of the steaks was utilized to estimate pH, color, aging loss, cooking loss, tenderness, and water holding capacity (WHC), one was used for the evaluation of metmyoglobin and myofibrillar fragmentation index (MFI), and the remaining one steak was used for sensory evaluation by a sensory panel. Except for the control, all other steaks were vacuum packed in polyethylene bags (150 × 200, PA/PE 90) by a vacuum packing machine (Multivac® Baseline, C300 twin, Wolfertschwenden, Germany) and stored at 0–2 °C for processing at 7 d, 14 d, 21 d, 28 d, and 35 d storage periods.2.4. Study Parameters2.4.1. Instrumental ColorInstrumental color values of meat samples were measured using Minolta Chroma meter (Konica Minolta® CR-410, Tokyo, Japan), provided with C illuminant, 2o standard observer, and 50 mm aperture, calibrated with a white tile each time before taking measurements [25]. Meat sample was removed from vacuum packs in the meat processing hall maintained at 10 ± 1 °C and bloomed for one hour in horizontal display chiller (ALVO, Model MD-12, size: 72″ × 42″ × 48″ by Technosight, Lahore, Pakistan) working at 0–4 °C [26]. Following the CIELAB color system, color values were measured, i.e., lightness (L *), redness (a *), yellowness (b *), hue angle (h), and chroma (C *). Three values from each steak were taken by placing the chroma meter at three different locations, and the mean value was calculated.2.4.2. pH MeasurementThe pH of meat samples was measured by pH meter (WTW, pH 3210 SET 2, Wissenschaftlich-Technische Werkstätten GmbH, Weilheim, Germany), calibrated at pH 4 and 7. The pH meter probe was inserted at three different sites in the steak, and the mean value was calculated.2.4.3. Metmyoglobin ContentMetmyoglobin content (MetMb%) was determined as described by [27]. A meat sample of 5 g was homogenized with 25 mL of 40 mM phosphate buffer (pH 6.8) for 10 s and then refrigerated at 4 °C for one hour. Afterward, the sample was centrifuged at 4500× g for 30 min, and the supernatant was filtered by a Whatman filter paper No. 1, using a spectrophotometer. The sample’s absorption values were observed at wavelengths of 525, 545, 565, and 572 nm. The MetMb% of the meat sample was calculated with the given formula. MetMb% = [−2.51 (A572/A525) + 0.777(A565/A525) + 0.8(A545/A525) +1.098] × 100(1)In the above formula, A represents the absorbance value with respect to wavelengths.2.4.4. Water Holding Capacity (WHC)Water holding capacity was measured with a compression machine (YYW-2, Nanjing Soil Instrument, Nanjing, China). A raw meat sample of 5 g was weighed and folded in filter papers. Then, it was compressed with a force of 373 N for 5 min. The weight of meat samples before and after the compression was recorded by the digital compact weighing balance (SF-400, 7000 g ± 1 g). The expressible water (%) was calculated using the following formula [28]. Expressible water (%) = w1 − w2/w1 × 100(2) where w1 = initial weight of meat sample, and w2 = final weight of meat sample.2.4.5. Cooking LossFor measuring cooking loss, samples were packed in polyethylene bags (150 × 200, PA/PE 90) and cooked in a waterbath (WNB45, Memmert GmbH + Co. KG, Schwabach, Germany) maintained at 82 °C until the core temperature of meat samples reached up to 72 °C [29] as recorded by digital food thermometer (TP101; temperature range of −50 °C to 300 °C). The weight of the initial raw meat sample and the cooked sample was recorded by digital compact weighing balance (SF-400, 7000 g ± 1 g, Zhejiang Tiansheng Electronic Co., Ltd., Zhejiang, China); the cooking loss was calculated using the following formula [30]. Cooking Loss (%) = w1 − w2/w1 × 100(3) where w1 = weight of steak before cooking, and w2 = weight of steak after cooking.2.4.6. Warner–Bratzler Shear ForceThe tenderness of meat samples was determined in terms of Warner–Bratzler shear force (WBSF) values. First, the cooked meat samples were chilled overnight at 0–4 °C [31]. Then, 6 cm long × 1 cm high × 1 cm wide strips of muscles were cut parallel to the direction of muscle fibers and sheared under a V-shaped stainless-steel blade, while the meat strips were placed with fibers oriented perpendicular to the Warner–Bratzler shear force device coupled with a texture analyzer (TA.XT plus® texture analyzer, Stable Micro Systems Ltd., Godalming, UK). The WBSF value was recorded in newton per centimeter square (N/cm2). Three readings for each sample were taken. Furthermore, the mean value was calculated.2.4.7. Myofibrillar Fragmentation Index (MFI)To estimate myofibrillar fragmentation index, a 50 mL solution was prepared containing cold sucrose (0.24 M) and potassium chloride (0.02 M). A minced meat sample of 10 g was added to the prepared solution and blended for 40 s at high speed in a tissue homogenizer. This homogenate was filtered using a cheesecloth, and after this, we also recorded the weight of the cheesecloth separately. The MFI was calculated as the residue’s weight in grams multiplied by 100 [32]. MFI = residue (grams) × 100(4)2.4.8. Sensory AnalysisThe meat samples from different storage durations were taken and cooked on a hot plate to attain the core temperature of 80 °C. Each meat sample was divided into equal cubes and served to trained panelists comprised of 8 members, including postgraduate students, a lecturer, and an assistant professor from the University. The sensory analysis was performed in the sensory analysis lab at the Department of Meat Sciences UVAS Lahore, followed by the mouth rinse method to avoid the carry-over taste effect. To avoid biases, the tagging of samples was non-familiar to the sensory panel. Panelists judged the meat samples for overall tenderness, juiciness, flavor intensity, and overall acceptability on an 8-point hedonic scale, where 8 = extremely beef-like odor, extreme flavor intensity, extremely tender, extremely juicy, and high overall acceptance, while 1 = extremely non-beef-like odor, extremely weak flavor intensity, extremely tough, extremely dry, and extremely low overall acceptance [33].2.5. Statistical AnalysisData regarding the effect of aging period on meat color, tenderness, and sensory analysis were analyzed under a complete randomized design by using two-way ANOVA through SAS software (version 9.1). The significant difference was considered at p < 0.05. The means were compared using DMR Test [34].The following mathematical model was used:Y = μ + F1i + F2j + (F1 F2)ij + εij(5) where Y = response variable, μ = overall population mean, F1i = fixed effect of muscle type, F2j = fixed effect of aging period, (F1 F2)ij = interaction effect of muscle type and aging period, and εij = random error.3. Results3.1. Instrumental ColorThe results of the instrumental color analysis after aging for 0, 7, 14, 21, 28, and 35 days are presented in Table 1.Both muscle type and aging time significantly (p ≤ 0.05) influenced the CIE a * values; GM muscle had higher a * values than LL muscle. During the aging time, a significant (p < 0.05) increase in the CIE a * values was observed from 0 to 7 days, and the highest values were witnessed on days 7 and 14. Additionally, significant (p < 0.05) differences were found in muscle type and aging times’ interactions. Similarly, muscle type and aging time significantly (p < 0.05) affected the chroma values. Furthermore, muscle type and aging time interactions also showed significant (p < 0.05) differences in chroma values. The CIE b * values were significantly (p < 0.05) affected by the aging time and increased from day 0 to 35 days of the storage period. Non-significant differences were found in the values of CIE L * among the two muscles, the aging time, and their interactions. Likewise, no significant (p > 0.05) difference was observed in the hue values of the aging time and its interactions with muscle type; however, significant (p < 0.05) differences were found in the main effects of muscle type, and GM muscle had higher hue values than LL muscle.3.2. pHThe results of this analysis are summarized in Table 1. There was no significant (p > 0.05) difference in pH of LL and GM muscles; however, the pH was significantly increased with the aging time. Muscle type and aging time interactions showed a non-significant effect on pH values.3.3. Water Holding Capacity (WHC)There were no significant (p > 0.05) differences in WHC of the two muscle types; however, aging time significantly (p < 0.05) affected the WHC% of both muscles. The WHC% decreased significantly with the increase in aging time (Figure 1). 3.4. Cooking LossThere was no significant (p > 0.05) interaction between the muscle type and aging time in cooking loss. Additionally, cooking loss values were independent of muscle type. However, significant (p < 0.05) differences were found in cooking loss values of aging times that increased significantly (p < 0.05) throughout the aging time (Table 2).3.5. Warner–Bratzler Shear ForceAs shown in Table 2, significant (p < 0.05) differences in shear force values were found in the main effect and interaction of muscle type and aging time. GM had lower shear force values than LL. The shear force values significantly (p < 0.05) decreased with an increase in aging time (0, 7, 14, 21, 28, and 35 days). The lowest shear force values were observed on day 35 in both GM and LL muscles.3.6. Myofibrillar Fragmentation Index (MFI)There was a significant (p < 0.05) difference in the main effect of the aging time, although a non-significant (p > 0.05) difference in MFI values of different muscles. The interactions between muscle type and aging time showed significant influence on MFI; both muscles presented an increasing trend with the aging time (Table 2).3.7. Metmyoglobin (%)As shown in Figure 2, non-significant (p > 0.05) differences were observed in the main effects of muscle type. At the same time, the metmyoglobin (%) values of these two muscles increased significantly (p < 0.05) with increased aging time. Similarly, the interactions between muscle type and aging period had significant (p < 0.05) effects on metmyoglobin (%), which increased with increasing aging duration in both muscle types.3.8. Sensory EvaluationThe effect of muscle type, aging time, and their interactions on sensory evaluation is shown in Table 3. The LL muscle showed significantly (p ˂ 0.05) better overall acceptability than the GM muscle. However, the panelists observed non-significant differences in odor, flavor, tenderness, and juiciness between the two muscles. The meat texture and overall acceptability were significantly improved throughout the aging time. The meat texture and overall acceptability also showed significant interactions between muscle type and aging time. The LL muscle showed the highest texture characteristics and overall acceptability at 35 days of the storage period. 4. DiscussionThe present study aimed to evaluate the muscle-specific effects of aging time on the color, tenderness, and sensory attributes of buffalo meat. Generally, the L * values should be increased with increasing aging time due to protein denaturation with endogenous proteases, leading to the weakening of protein structure, resulting in more light scattering [7]. Similarly, the L * values of the LL and GM muscles increased linearly with an increase in aging time in the present study; however, this increase in L * values was statistically non-significant. Likewise, a * and C * values increased with aging time in both LL and GM muscles. The a * values of both muscles significantly increased up to 28 days of aging, while the C * values of LL and GM muscles increased up to 28 and 35 days, respectively. These results are inconsistent with the findings as reported by Stanišić et al. [35]. The authors found a significant increase in a * and C * values of LL and GM muscles during the aging periods.Similarly, Stanišić et al. and Vitale et al. [35,36] found that the a * value of LL and GM muscles improved in the aged meat compared with non-aged meat. The improvement in meat color characteristics is attributed to the enhanced blooming ability of aged meat. Tang et al. and Jacob et al. [37,38] concluded that the difference between the blooming ability of the fresh un-aged meat cut surface and aged meat is due to a decrease in oxygen-consuming respiratory enzymes within the mitochondria of aged meat. This reduction allows more oxygen to bind with myoglobin, forming a thicker layer of oxymyoglobin [39]. Another possible reason for the increased a * value from 7 to 14 days of aging might be the exposure of the meat surface to oxygen, which penetrates the meat leading to the conversion of deoxymyoglobin to oxymyoglobin.The pH values were gradually increased from 5.64 to 5.85 throughout the aging durations. In a similar study, Daszkiewicz et al. [40] also reported an increase in the pH of meat samples up to 14 days of aging. This increase in meat pH during aging could be attributed to the conformational changes linked with the proteolytic degradation of the muscle fibers [41] and hydrolysis of meat proteins into amino acids [42,43]. Likewise, Luz et al. [29] reported an increase in pH values up to 14 days of aging; however, afterward, the pH values decreased until 21 days of aging. This decrease in pH values between 14 and 21 days of aging might be due to a decrease in enzymatic activity with an extended aging time.The WHC decreased with an increase in the aging time, while the cooking loss initially increased up to 7 days of aging and then decreased gradually throughout aging times. In their studies, Purslow et al. and Cho et al. [44,45] reported decreased WHC and cooking losses during different aging durations. The decrease in WHC with increasing aging duration is mainly caused by proteolysis and small protein fragments, leading to more fluid losses from muscle structure. When both muscles’ interactive effect was compared, LL had more drip losses than GM; this increase in drip losses could be due to more enzymatic reactions within the muscle structure [46]. Similarly, in the present study, an increase in the cooking loss up to 7 days of aging could be due to higher proteolysis during the earlier aging time. The tenderness of the meat is improved due to the proteolysis of myofibrillar proteins during aging [47,48,49]. Between two muscles, LL showed higher WBSF values than GM muscles; however, both muscles showed a significant decrease in the WBSF values and an increase in myofibrillar fragmentation index (MFI) values during an extended aging period. Similarly, various studies (Vaskoska et al., Shanks et al., Monsón et al., Koohmaraie et al., and Koohmaraie et al. [50,51,52,53,54]) reported a decrease in WBSF values and an increase in the MFI with the extended aging durations. The decrease in WBSF and increase in MFI values in GM muscle could be due to having low connective tissue content and more proteolytic breakdown of cross-linkages between muscle proteins in comparison to LL muscle [55]. In their study, Irurueta et al. [56] also noted a significant decrease in buffalo meat shear force values under vacuum storage of 25 days. This increase in the values of MFI over periods of aging is due to enhanced proteolysis by calpains resulting in myofibril fragmentation and leading to increased tenderness. This gives a fair indication that MFI can be used as an indicator of tenderness which could be used for selecting tender meat for processing and retail sale.Similar to the present study, Mancini et al. [57] reported an increase in the metmyoglobin concentration with the increase in the aging periods. However, metmyoglobin is the indicator of the shelf life of meat color, and more metmyoglobin formation (brownish color meat) leads to a lower acceptance of the meat by the consumer. In the current study, although the metmyoglobin was increased with an increase in aging time, the improvement in the color characteristics with an increase in the aging period could be attributed to increased color intensities during the blooming process. To reduce metmyoglobin production, one should maintain the storage temperature (0–4 °C) and proper meat handling during transportation [58].The sensory evaluation showed a better overall acceptability of LL in comparison GM muscle. This could be due to the comparatively better texture and juiciness of the LL than the GM muscle. The texture and overall acceptability were improved throughout the aging periods. Similarly, Monsón et al. [59] evaluated the effect of aging on the sensory characteristics of beef and observed better tenderness and overall acceptability with an aging period. The improvement in the sensory characteristics with aging time could be attributed to increased proteolysis and MFI with aging durations. Between the two muscles, LL showed the highest sensory texture and overall acceptability score throughout aging durations. This could be due to the highest MFI of LL throughout different aging durations, as reported in the current study.5. ConclusionsWet aging improved both muscles’ color attributes, tenderness, and sensory characteristics till 35 days of aging. Both LL and GM muscles responded differently to different aging periods. Based on the evaluated meat characteristics, overall, 28 days of aging is required to improve the meat quality of LL, while 21 days of aging is enough for improving the meat quality of GM muscle. However, further studies are needed to explore the effect of wet aging for extended periods on the microbial count, protein oxidation, and volatile compounds, along with the color stability and sensory attributes.

animals : an open access journal from mdpi

10.3390/ani11092649

PMC8468282

Peste des petits ruminants (PPR), also known as sheep and goat plague, is a highly contagious animal disease affecting small ruminants and camels. It is caused by a virus belonging to the genus Morbillivirus, family Paramixoviridae. Once newly introduced, the virus can infect up to 90 percent of an animal herd. A PPR outbreak is an emergency due to its rapid spread and high animal mortality rate. This study simulated three control strategies of PPR spread among animals in the United Arab Emirates. These strategies include implementing mass vaccination, ring vaccination and ceased vaccination strategies, combined with or without strict animal movement control simultaneously. The simulation results compared the level of the effectiveness and direct government costs for each of the three strategies. Such results aid the decision-makers in the country and globally in line with the World Animal Health Organization’s goal to eradicate the disease by 2030.

Peste des petits ruminants (PPR) is an important infectious viral disease of domestic small ruminants that threatens the food security and sustainable livelihood of farmers across Middle East, Africa, and Asia. The objective of this research is to analyze the disease’s spread and its impacts on direct government costs through conducting three simulations of different control strategies to reduce and quickly eradicate PPR from the United Arab Emirates in the near future. A Modified Animal Disease Spread Model was developed in this study to suit the conditions of the United Arab Emirates. The initial scenario represents when mass vaccination is ceased, and moderate movement restrictions are applied. The second scenario is based on mass vaccination and stamping out the disease, whereas the third simulation scenario assumes mass and ring vaccination when needed, very strict movement control, and stamping out. This study found that the third scenario is the most effective in controlling and eradicating PPR from the UAE. The outbreak duration in days was reduced by 57% and the number of infected animals by 77% when compared to the other scenarios. These results are valuable to the country’s animal health decision-makers and the government’s efforts to report to the World Animal Health Organization (OIE) regarding the progress made towards declaration of the disease’s eradication. They are also useful to other concerned entities in other Middle Eastern, North African, and Asian countries where the disease is spreading.

1. BackgroundPeste des petits ruminants (PPR) is a viral disease, caused by a morbillivirus closely related to rinderpest virus, which mainly affects goats, sheep, and some wild relatives of domesticated small ruminants, as well as camels. Wild ruminants may play an important epidemiological role as a virus source for domestic small ruminants [1]. PPR was first reported in Ivory Coast in 1942. Today, more than 70 countries have confirmed PPR within their borders, and many countries are at risk of the disease being introduced. PPR is characterized by high morbidity and mortality rates of up to 90% [2]. Affected animals present with high fever, depression, along with eye and nose discharges, severe pneumonia and diarrhea. The PPR virus (PPRV) does not cross from animals to infect humans [3]. The availability of effective and safe live attenuated PPR vaccines in sheep and goats has boosted the control program in some developing countries, and currently a total of 59 countries are recognized by the OIE as being free from the disease (Figure 1). Other developing countries are currently unable to develop and apply an effective strategy to control and eradicate the PPR virus (Figure 1).The Food and Agriculture Organization (FAO) reported that PPR affects 30 million animals across 70 countries around the world. Sixty percent of these countries are in the African continent, and the other 40% are in the Middle East and Asia. (FAO, OIE Report [3]). The disease causes annual economic losses of up to USD 2.1 billion each year. Looking beyond this figure, 300 million families are at risk of losing their livelihoods, food security, and employment opportunities. The inability of families, communities, and institutions to anticipate, absorb, or recover from PPR can compromise national and regional development efforts, and turn back the clock on decades of progress.In the United Arab Emirates (UAE), PPR was first reported in wildlife in 1986. To date, a total of fifty outbreaks in both wildlife and small ruminants have been reported to the OIE. After implementing the UAE national animal health plan in 2016, which adopted a mass vaccination strategy to control and eradicate PPR, outbreaks have sharply decreased to two to three outbreaks per annum. The UAE is currently at stage 2 of the five stages of the progressive step-wise approach for the prevention and control of PPR. Despite the progress achieved in controlling PPR in the country, no simulations or trials to assess its spread and the effectiveness of different control strategies have been carried out [3,4], Figure 2 and Figure 3 (Supplementary Table).Few studies have been reported in developing mathematical models to simulate and assess the effectiveness of various PPR control strategies. Mitchell et al. (2017) [6] noticed a lack of empirical studies on PPR. Such models provide policy-makers with a tool to develop powerful containment strategies for handling out-breaks of PPR by understanding how it spreads through herds, the uncertainty of the disease parameters, and the impact of each herd’s configuration on the disease’s spread. Therefore, the mathematical model, which can be used to configure any infectious disease, led to the finding that lowering the amount of time from the first identification of PPR in a herd to vaccination will significantly reduce the number of deaths that result from PPR [7].Lyons et al. (2019) [8] estimated various cost components in pastoral and mixed-crop livestock systems in four PPR vaccination campaigns in Ethiopia. The cost of the overall vaccination process for mixed-crop livestock systems is approximately double the price in pastoral areas. Due to the lack of knowledge about the transmission potential of the PPRV, Fournié et al. (2017) [9] used a dynamic model of transmission and elimination of PPR in Ethiopia. The outcome was an estimation for the vaccination coverage required for elimination and the level of viral transmission in an endemic setting.Thus, this study aimed to develop a model for PPR spread in the UAE, simulate possible control and eradication strategies, and to evaluate the effectiveness and direct government cost of such strategies. To achieve this, three scenarios of PPR spread in UAE with a corresponding different control strategy were simulated and discussed using The North American Animal Disease Spread Model (NAADSM) [10].2. Global Strategy to Eradicate Peste des Petits Ruminants (PPR)PPR can be eradicated worldwide by 2030. It can be readily and cost-effectively diagnosed, and a reliable, inexpensive and high-quality vaccine is available that confers lifelong immunity to vaccinated animals after a single dose. The virus also has a relatively short infectious phase and does not survive for long outside a host, making it an ideal candidate for a concerted eradication effort. Strengthening the capacities of national veterinary services to control and eradicate this disease will also generate wide-ranging benefits in the fight against other animal diseases [7].The growing international consensus and political support for the eradication of PPR, technical feasibility, high rates of return on investment that span generations, and the proven FAO–OIE partnership in successfully eradicating transboundary animal diseases—such as rinderpest—are all strong guarantees for the success of the global PPR eradication program Figure 4. The stages of this program range from stage 1, when the epidemiological situation is assessed, to stage 4, when the country can provide evidence that there is no virus in circulation, and it is ready to apply for the OIE official status of freedom from PPR. A country is “below stage 1” if there is no epidemiological information available, and “beyond stage 4” if the OIE official status recognition has been concluded. Currently, the United Arab Emirates (UAE) shows evidence of stage 2 to stage 3 being achieved [7].The eradication of PPR in the UAE is one of the most important goals for relevant government and non-government institutions in the country. The UAE is implementing an on-going strategy to eradicate PPR. The progress achieved by June 2021 toward the country’s PPR eradication strategy following the OIE guidelines described in the previous section of this study is assessed to be stage 3.3. Objective and Simulation ScenariosThe objective of this study is to analyze the impacts (in terms reducing the epidemiological impact as well as the estimated direct government cost) of control strategies through simulations to limit and rapidly eradicate PPR from UAE.4. Data and MethodThis section includes a detailed description of the study data and the method applied in this study.4.1. DataData on the population of livestock were obtained from the Abu Dhabi Agriculture and Food Safety Authority (ADAFSA) Animal Health Division. Data were collected on animals from 24,836 holdings and farms. The total population of all of the animals considered in this study was nearly 3 million heads. In this model, a cluster of animals was called a “unit”, which was the basis of the simulation. A unit had a product type, a number of animals, a point location (expressed in terms of longitude and latitude), and a disease state. All small farms had 30 animals or less of one type or a mixture of various species (sheep, goats, and camels). The majority, 14,914 farms, representing 60% of the farms in the country, harbored either sheep or goats or both sheep and goats in this study, and these were considered as one type of production due to the epidemiological similarity between the two species when it comes to PPR spread. There were 8410 small mixed sheep and goats farms, representing 34% of the total farms. The camel farms included in this study numbered 1484 farms, representing 6% of the total farms. Camels were included and assumed to be able to be infected because there is evidence of PPR outbreaks in the region [11,12]. Dairy cattle, which are mostly commercial large operations farms in the country, are found on 26 farms containing 39,750 heads of milking cows, calves, and heifers. Dairy cattle farms are part of the animal population database. However, this study does not assume any PPR infections to occur at the dairy cattle farms.4.2. Methodology and Simulation ScenariosTo analyze the impacts on PPR (in terms reducing epidemiology as well as estimated direct government cost) of control strategies to eradicate it from the UAE in a timely manner, three different simulation scenarios (A, B and C) were carried out using the NAADSM as shown below:This scenario illustrates the case where there is no virus circulation either at the zonal or national level and the country is ready to apply for official OIE recognition of PPR freedom. This scenario represents a strategy that recognizes the initial situations when the country, here the UAE, has already controlled the disease in areas where it is highly endemic, mass vaccination is ceased (no vaccination), and moderate movement restrictions are applied.This scenario represents the case where efforts and the control strategy are based on mass vaccination and stamping out the disease to reduce the impact of PPR at the national level to the minimum level possible.The full eradication scenario represents a strategy based on mass and ring vaccination when needed (vaccinating detected herds when they occur in the vaccination ring), very strict movement controls, and stamping out the disease.Epidemiologic modeling is a common tool used to simulate and develop possible scenarios to estimate the potential impact of outbreaks of contagious diseases, such as PPR, in populations of domesticated animals. The information generated from these scenarios is very useful and can be used by policymakers to control diseases and plan for early and on-going efforts and responses to possible outbreaks for disease management and eradication. Several spatially explicit stochastic epidemic simulation models have been developed for estimating the spread of highly contagious animal diseases and simulate outbreaks of diseases (e.g., Bates et al., 2003 [13]; Garner and Beckett, 2005 [14]). The North American Animal Disease Spread Model (NAADSM) was the adopted framework and software used in this study. The software was designed to simulate the spread and control of foreign animal diseases in a population of susceptible livestock herds (Harvey et al., 2007 [15]).5. PPR Epidemiology and CostThe following are the epidemiological features of PPR spread that were carefully modelled using NAADSM:The morbidity rate in susceptible populations can reach 90–100%;Mortality rates vary among susceptible animals, but can reach 50–100% in more severe instances;Both morbidity and mortality rates are lower in endemic areas and in adult animals when compared to the young;The latent period is considered 3 days minimum, 7 days mode, and 10 days maximum using the Beta-PERT probability function;The period required for herd-level natural immunity to be achieved, either through infection and recovery or vaccination, is considered to be 1460 days (4 years). Natural immunity was modeled as having a normal distribution;The probability of an infected animal dying from the disease is considered to be 90% because of the high mortality rate of the disease [2,16].5.1. PPR Detection and Clinical DiagnosisThe probability that the owner or veterinarian will observe the death form the suspected PPR infection and report it is 100%, as the UAE government considers the disease as being mandatorily notifiable to the relevant authorities. Diagnostic testing’s sensitivity in this study was assumed be 95%. The delay in obtaining testing results was assumed to be from no delay to 3 days maximum.5.2. PPR Disease StatesSeven discrete disease states were used in the NAADSM, as shown in Figure 5. These states are susceptible, latent, sub-clinically infectious, clinically infectious, naturally immune, vaccine immune, and dead from the disease. When a susceptible unit is infected, the natural progression is to become latent, unless a disease control action is taken. Moreover, an infected unit will proceed naturally from its latent state to a sub-clinically infectious state, unless a disease control action is taken. The disease transition will follow the natural progression, as shown in the outer loop. Implementation of any type of disease control may alter the natural disease cycle, as shown inside the loop.5.3. Animals Movements and Indirect ContactsThe direct and indirect contact parameters used in this study that impact PPR spread in the UAE are summarized below (Table 1).5.3.1. Direct Contact Spread ParametersMovement or shipment of animals among units/herds Mean rate of animal shipments (number of recipient units per source unit per day);Movement distance (probability density function per km);Shipping delay (probability density function per days);The probability of infection of the recipient unit, given exposure to an infected unit;Movement rate multiplier (scalar value as a function of the number of days since the first detection of the outbreak).5.3.2. Indirect Contact Spread ParametersIndirect contacts may include the movement of people, equipment and materials, vehicles, animal products, etc., among units, and are simulated in the same manner as direct contact, except that only sub-clinically infectious and clinically infectious units, not latent units, can act as the source of infection. The parameters for indirect contact are similar to, but independent of those for direct contacts.6. Peste des Petits Ruminants (PPR) Control StrategiesThe NAADSM uses the following measures to control the disease: zoning and tracing, quarantine, depopulation, and vaccination parameters.6.1. Zoning and TracingTrace out investigations by NAADSM are simulated after an infected unit is detected by zoning into two zones, surveillance low-risk zones and restricted high-risk zones. Tracing is one-level forward. The probability of successful tracing is assumed to be 80%. The maximum delay in tracing forward is assumed to be 3 days.6.2. Isolation and Quarantine ParametersIn the model, units are quarantined for one or more of the following reasons: an infected unit is immediately quarantined following detection, units traced out are also quarantined, and lastly when units are selected and waiting for destruction.6.3. Destruction (Depopulation) ParametersThe following are the depopulation parameters and triggers used:Delay to beginning a destruction program (fixed integer value per days)Destruction capacity (relational function: number of units that can be destroyed as a function of the number of days since the first detection of an outbreak);Destruction priorities (rank order of reasons for unit destruction, as described in the text);The radius of the destruction ring, if a ring is triggered;Destruction capacity is assumed to be 100 units/herd per week6.4. Vaccination and Immunity ParametersThe following are the vaccination parameters and triggers used:Number of units that must be detected before vaccination begins (number of detected units), which is assumed to be one unit/herd detected;Vaccination capacity (relational function: number of units that can be vaccinated as a function of the number of days since the first detection of an outbreak). The vaccination capacity in the model is assumed to 100 units/herd per week;Vaccination priorities (Rank order of reasons for unit vaccination)—vaccinate detected unit when they occur in the vaccination rings;The radius of the vaccination ring, if a ring is triggered and other parameters, which is assumed to be 3 km.7. Costs Estimation ParametersThe direct costs associated with destruction and vaccination can be calculated in the model to compare the costs of different control measures. Below are some cost input parameters used in NAADSM for determining the direct costs associated with disease control—see Appendix A on the cost parameters used to estimate direct government costs for the following items:Parameters associated with destruction:Appraisal;Cleaning and disinfection;Euthanasia;Indemnification;Carcass disposal;Parameters associated with cost of vaccination;Number of animals that can be vaccinated at the baseline cost;The baseline cost of vaccination;Additional cost incurred when the number of animals vaccinated exceeds the threshold set;Cost of vaccination site set-up.8. Peste des Petits Ruminants (PPR) Simulation Results and DiscussionSimulation results are illustrated in Figure 6 and Figure 7, and in Table 2 below. The results indicate that the outbreak durations are 171, 148, and 73 days for scenarios A, B, and C, respectively. The total cumulative number of infected animals was reported to be 1505, 1481, and 327 animals for the three scenarios A, B, and C, respectively.The number of susceptible animals was reported to be 2,954,213, 2,954,213, and 2,954,213 animals for the three scenarios A, B, and C, respectively, while the number of latent animals was reported to be 1327, 1316, and 315 animals in the three scenarios A, B, and C, respectively. The number of animals showing subclinical signs was reported to be 887, 889, and 271 animals for the three scenarios A, B, and C, respectively. Moreover, the number of animals showing clinical signs was reported to be 690, 697, and 255 animals for the three scenarios A, B, and C, respectively.Tracing had also been addressed as number of animals exposed to any infected herd were reported to be 2454, 2476, 227 animals of the three scenarios A, B, and C, respectively. While number of animals directly exposed that could possibly have been traced forward were reported to be 1833, 1837, 71 animals of the three scenarios A, B, and C, respectively. Total number of animals in units successfully identified by tracing (either forward or back) after direct contact 1638, 1644, 63 animals of the three scenarios A, B, and C, respectively. Number of animals in units successfully identified by tracing (either forward or back) after contact (either direct or indirect) were reported to be 2220, 2248, 208 animals of the three scenarios A, B, and C, respectively.Regarding diagnostic testing, data showed that the number of animals subjected to diagnostic testing after a trace forward or trace back after direct contact was reported to be 1444, 1461, and 56 animals for the three scenarios A, B, and C, respectively. Meanwhile, the number of animals subjected to diagnostic testing after a trace forward or trace back after (either direct or indirect) contact was reported to be 1962, 2007, and 421 animals for the three scenarios A, B, and C, respectively. Additionally, the number of animals in the tested units with a true negative diagnostic test result was reported to be 1046, 1111, and 1 animals in the three scenarios A, B, and C, respectively.The number of animals that are destroyed was reported to be 1384, 1370, and 321 animals in the three scenarios A, B, and C, respectively.Niu et al. (2017) [17], based on the global effort to combat PPRV, presented a global online prediction system by adopting correlational analysis, based on collected data from 2977 cases from 2009 to 2018, and showed that PPR has a severe impact on people depending on the livestock production system as a means to generate income to reduce poverty.Controlling PPR is essential for poverty alleviation, especially in Africa, the Middle East, and South Asia. Additionally, the model results show that the outbreaks were concentrated in the continents of Asia and Africa, and widely spreading in the Middle East region [18].Cameron (2019) [19] found that a more sustainable option for PPR eradication could be adopting guerrilla tactics, where the primary weapon is information and understanding PPR. This tactic can be divided into four main phases—the foundation, planning, implementation and demonstration of global freedom. The author asserted that continuous real-time information in the form of guerilla tactics should be the primary tool for disease eradication, not long-term vaccination. This will also optimize the use of available sources and minimize the disruption related to managing the movement of animals from infected to uninfected areas. We herein developed a model for the disease spread of PPR, simulated possible control and eradication strategies, and evaluated the effectiveness and direct government cost of such strategies.These results show that control strategies that intensify vaccination compared to no vaccination and with no strict and intensive animal movement controls moderately reduce the outbreak duration by 57% and the spread of PPR by 78% when compared to vaccination with stamping out (B) or ceased vaccination (A) strategies. However, an integrated and targeted eradication and control strategy that applies all possible effective measures of both triggered ring vaccination and animal movement controls reduced the disease’s outbreak duration in scenario C by 57% and the total cumulative number of infected animals by 78% compared to the initial eradication scenario A (without vaccination and lesser restriction of animals’ movement).Scenario B address the impact of vaccination. Efficient PPR vaccines are available and can induce life-long protective immunity in vaccinated animals.The total number of susceptible animals is nearly 3 million animals in the UAE. The simulation scenarios detailed the control strategies that can be applied against PPR, and we discuss the outcomes and perform a comparison between the three scenarios’ outcomes, as illustrated in Table 2 below. The comparison between scenario A and scenario B shows the changes due to the introduction of the triggered vaccination as the disease control strategy. Due to vaccination, the outbreak duration on average of the model’s 1000 iterations reduced from 171 days to 141 days, or by −13%. The number of infected animals slightly reduced from 1103 to 1092, or by −1%. Similarly, all other parameters regarding disease spread and the impact of the vaccination strategy showed changes that varied from 0% (no change) to −1%. This indicates that a reliance on vaccination measures alone (e.g., not accompanied by restrictions on animals’ movement) would not contribute enough to achieve PPR eradication. However, the most effective scenario, scenario C, showed a reduction in the disease’s spread or the outcome of the control strategy in the range from 57% for the disease duration in days to as high as 96% for the number of animals directly exposed that could have been traced forward. The number of animals depopulated reduced from 1384 animals to only 324 animals, or by 77% in scenario C compared to scenario A. The number of vaccinated animals was also reduced from 35,061 to 2704. These results, compared to the progress the country has achieved (reaching stage 2 of PPR eradication) as of 2021, applying PPR national eradication strategy, indicate that a successful eradication strategy may apply a very intensive and strict movement control strategy integrated with vaccination and depopulation only when such measures are triggered. Such observations are in line with the global PPR control and eradication strategy, where it is stated that prevention and control measures are a combination of different tools, which can include vaccination, improved biosecurity, animal identification, movement control, quarantine and stamping out. These individual tools are likely to be applied at different levels of intensity as an individual country moves along the pathway. (FAO OIE Global Strategy for the Control and Eradication of PPR. Paris: OIE and FAO; (2015).9. Direct Government Costs Comparison of PPR Eradication Scenarios ResultsTable 3 below shows a comparison of the various scenarios’ direct government cost estimates. The scenario simulation results show the direct government cost are found to be of 1.1, 1.5, and 0.2 million US Dollars in scenario A, B, and C for each to the three scenarios, respectively. The most expensive item in all three scenarios is the cost of vaccination (about 0.5 million US dollars), followed by the cost of depopulation indemnification, which is compensation provided to the farmers when their animals are depopulated (0.4 million US dollars) and the the cost of cleaning and disinfecting the depopulated farms (0.4 million US dollars). These results indicate the presence of a trade-off between applying a vaccination strategy with less strict animal movement controls versus vaccination along with more strict animal movement controls. The latter strategy, strategy C, is more effective from a government costs perspective; mass and ring vaccination, very strict movement controls, and stamping out measures will collectively lead to the faster eradication of PPR (FAO OIE Global Strategy for the Control and Eradication of PPR. Paris: OIE and FAO; (2015)).10. ConclusionsThe inter-regional consultative meeting on the FMD and PPR situation, labelled as an OIE/FAO GF TADs event and held in Jordan in 2014, considered PPR’s spread across the Middle East and North Africa and other regions in Asia, and pointed to vaccination as a very effective tool, but it should not be considered as the only tool, and instead should be combined with other tools, in particular the control of the movement of live animals.The meeting recommended the preparation of specific tools, including a monitoring and evaluation tool, a post-vaccination monitoring tool, a PPR Global Research and Expertise Network (PPR-GREN) to assess the socioeconomic impact of PPR in livestock production, livelihoods and food security, and cost/benefit analysis of PPR control options used when preparing national control programs and financial project proposals [20].The UAE aims to report the progress of its PPR control efforts to the OIE [9] in order to be recognized as a country free from PPR. The parameters used for the disease spread were localized to UAE settings as described. The total animal population considered in this study is nearly 3 million heads from 24,836 holdings and farms. The majority (60%) of the farms in the country are sheep and goat farms, mixed small farms comprise 34% of the farms, while camel farms represent 6% of the total farms. Eradication control strategies, including zoning and tracing, vaccination, movement control, and stamping out, are all effective strategies. The triggered vaccination control strategy was necessary to reach stage 4 of the global PPR control strategy. However, comparison between scenario A and B did not show high effectiveness for triggered vaccination. Movement control was found to be more effective in reducing the number of infected animals and disease duration. To move to stage 4 and beyond stage 4 (free status), it is recommended to apply triggered vaccination, movement control and stamping out animals as needed. The analysis of the various control strategies must include assessment of the costs and resources needed for successful eradication strategies. Business losses due to restricted movement control and so businesses disruption due to movement control have to be taken into consideration when the country implements disease eradication strategies. The simulation scenarios presented here are applicable to other countries in the Middle East, Africa, and Asia, where several countries are joining the global effort by the World Animal Health Organization (OIE) and its strategy to eradicate PPR by 2030.11. Limitations and Future ResearchThe limitations of this model include modelling within-herd, local area and airborne disease spread. Furthermore, stochastic methods for unexpected low-probability high-impact shocks should be added to the modeling, accounting for the private business losses due to the movement control and restrictions. Future research may expand the modelling and control strategies considering within-herd dynamics and the reaction of market prices to the spread of PPR.

animals : an open access journal from mdpi

10.3390/ani12030230

PMC8833676

Online dating applications offer new ways for people to search for social contacts. While previous studies have indicated that the inclusion of animals in profiles can increase users’ dating success rates, the question of how many users display animals, and what kinds of animal are shown on dating profiles, has not yet been empirically investigated. Using a structured observational study of profiles in Vienna and Tokyo on a popular online dating app (n = 2400), we examined how many profiles show animals and what kinds of animal are shown. Further, we investigated whether cultural background (Vienna versus Tokyo), gender, age and sexual orientation affect the use of animal pictures. Approximately 16% of investigated profiles had at least one photo showing an animal. In both cities, dogs were the most frequently shown animal, followed by cats. Further, results indicate that users in Vienna, women and older users were more likely to present animals on their profiles.

Online dating applications offer new ways for people to search for social contacts. While previous studies have indicated that the inclusion of animals in profiles can increase users’ dating success rates, the question of how many users display animals, and what kinds of animals are shown on dating profiles, has not yet been empirically investigated. Using a structured observational study of profiles in Vienna and Tokyo on a popular online dating app (n = 2400), we therefore looked at how many profiles show animals and what kinds of animals are shown. We found that 15.5% of the investigated profiles had at least one photo showing an animal. In both cities, dogs were the most frequently shown animal. Taking the cities together, they appeared in 46.4% of the animal pictures, as compared with cats at 25.7%. Other animals such as exotic animals (9.9%), farm animals (6.4%) or horses (4.6%) played a minor role. Users were significantly more likely to show cats in Tokyo (35.8%) than they were in Vienna (18.0%). We found that users in Vienna; women; and older adults were more likely to present animals on their profiles than were users in Tokyo; men; and younger users. Sexual orientation showed no significant differences in the analyses.

1. IntroductionIn recent years, the increasing use of online dating platforms and applications (from here on “apps”) has led to at least two decisive developments: first, a change in the way in which people search for social contacts, casual encounters or committed relationships; and second, a new opportunity for users of the apps to present themselves in ways that make a positive impression. Against this background, the present study combines two themes of growing significance in contemporary society: online dating apps, and the role and integration of animals in human life.In 2015, 185 million people used dating apps; in 2020, the number of global users rose to 270 million, and the number is continuously increasing [1]. In this process, various dating apps have been established, all of which have the main goal of connecting people in a simple, convenient way. With an average of 66 million active users a month [2], Tinder® is currently considered the most popular app, and it dominates the dating market worldwide [1,2]. Founded in 2012, this app has revolutionised the dating industry with its simple-to-use design. By using smartphone GPS functionalities, it searches for people who are located in the same region and suggests potentially suitable dating candidates based on gender, sexual orientation and age. In addition, users are able to upload several profile photos or write short profile texts about themselves. With this information, the user can make a decision: swiping right means “like”, swiping left means “pass”. When two users “like” each other, this produces a “match”, meaning that they are able to chat with each other from then on via the app. Although one can add a short, written self-description, the mechanism and design of the app emphasises visual impression—it encourages users to focus on profile photos [3,4]. This visual focus is reflected in the way people actually make use of the app. For instance, previous studies revealed that profile texts are rarely, if ever, considered by users, who mainly, or exclusively, make their decisions on the basis of the visual impression given by the (first) profile photo(s) [5,6].Although online dating apps are sometimes called “hook-up” or “sex” apps [1], the motives of those who use them seem to be more diverse. As research indicates, these apps are used not only as a searching tool for casual sex but also by those looking for relationships, entertainment, thrills and excitement, new social contacts, an ego-boost or just to be trendy [5,7,8,9]. However, the aims to control the way one is represented to others [10,11] and convey a positive impression unite all users [5]. Although it can be argued that users of online dating profiles tend toward overstatement, research indicates that users look for a balance between an ideal and authentic self-presentation [5,10]. Degen and Kleeberg-Niepage [4] and Ward [5] have emphasised that users’ self-presentation is not a spontaneous act but a rather conscious choice they make. So, users of online dating apps think about how they are going to present themselves, and in particular about what photos to upload, to make a lasting and positive impression on other users. As Jagger noted more generally: “Dating advertisements are a revealing site for examining the social construction of identities—identities that are deemed desirable and marketable in a specific cultural context” [12] (p. 90).With this in mind, users are encouraged to make the right visual impressions in order to present themselves favourably. For instance, users with an interest in sports or traveling will probably make these visible by showing themselves doing sports or traveling in foreign countries. Besides hobbies, a further interest or passion can be easily shown via profile photos: namely, the user’s interest in, and passion for, animals. Undoubtedly, in recent decades, especially, the role and perception of animals has changed, both in society and science. In particular, our handling of companion animals, and the integration of them in our lives, signal the changing human–animal relationship [13]. Animals such as dogs and cats are increasingly viewed and treated as family members, and tight emotional relationships between animals and their owners, or people in general, can be observed [13,14]. Relatedly, the term animal turn indicates the increased scholarly interest in animals in various academic disciplines. For instance, the field of human–animal studies investigates multiple and complex relationship forms between humans and animals in an interdisciplinary way. It includes disciplines such as history, philosophy, literary studies, communications sciences and sociology [15].Against this background, it comes as no surprise that our spatial proximity and emotional closeness to companion animals is also evident in the special situation of dating. Guéguen and Ciccotti [16] have shown that men are more successful in obtaining women’s phone numbers in encounters on the street when they are accompanied by a dog than they are without one. Since dogs are usually perceived as social animals that need the attention and care of their owner, they seem to reflect owner characteristics such as being social, reliable or empathic, which in turn have a positive effect on dating [17]. In addition, research indicates differences between animal species [18,19,20]. For instance, heterosexual men labelled as “dog persons” may be perceived as more masculine and attractive than they are when labelled “cat persons” [19,20]. This does not necessarily mean that photos with animals always improve the chances of success. For example, Kogan and Volsche [18] surveyed 708 women aged 18–24 concerning the impact of cats on the dateability of men. They found that men with cats were perceived as less masculine by female respondents, representing a higher level of neuroticism, and hence were perceived as less dateable for both short- and long-term pairing [18]. Whereas previous publications have focused on why animals appear on dating profiles, and the extent animals to which they increase the chances of success [16,17,19,20], the question of how many users of online dating platforms display animals, and of what kind of animals are shown, has not, as far as we know, been empirically investigated. Given this, the overall aim of the present observational study of profiles on a popular online dating app in Vienna and Tokyo was to examine the prevalence of users showing animals on their dating profile and to investigate whether differences in socio-demographic factors, including different cultural backgrounds, can be overserved.A German study showed that around one quarter (26%) of animal owners are 60 years or older, whereas only 15% are younger than 30 years [21]. In addition, differences between men and women were found. There was evidence that women are more likely to keep animals than men [22]. Assuming that people who own animals are more likely to present them on their profiles than those that do not, it is to be expected that more women, and older users, will picture animals such as dogs and cats on their profiles than men and younger users. Whereas these German studies [21,22] build an appropriate basis for the development of hypotheses for the Austrian context, these numbers might differ in Japan. However, to the authors’ best knowledge, studies on this issue are not available for the Japanese context. Further, it is noticeable that most existing research into the use of dating apps and related issues has focused on heterosexual relationships [16,17,19,20]. The present study attempts to overcome this limitation by considering both male and female heterosexual and homosexual users.To determine whether cultural differences in the presentation of animals on the selected app exist, profiles were analysed from Vienna, exemplifying a metropolitan area in Central Europe, and Tokyo, a metropolitan area in Asia. These cities were chosen because they may be considered culturally distinct in their social norms [23] and because their citizens have rather different possibilities of integrating animals in their lives. In 2020, for instance, 55,649 dogs were registered in Vienna, and their numbers are rising [24]. The provision of dog zones in every district, and individual sections in public parks designated for dog owners with their animals, makes Vienna especially attractive for dog owners [24]. By contrast, in Tokyo dog numbers are falling somewhat [25]. Strict regulation in public areas as well as the quite common prohibition requiring owners to keep their animals in their apartments make it less attractive to keep animals in this city. It is likely that these different circumstances would lead to users in Vienna showing a dog or a cat on their profile photos more often than users in Tokyo would. On the other hand, a major trend in Tokyo is the establishment of cat cafes. It is estimated that there are 40 such cafes in the city. Although the number of animals might be lower in Tokyo, it is likely that cats, especially, as opposed to dogs, will be shown in online dating profile photos there. Based on data from an observational study of profiles from Vienna and Tokyo on the selected online dating app, it is the overall purpose of this paper to empirically address the following research questions. (i) How many profiles show animal photos, and do socio-demographic factors explain differences between profiles with and without animals? (ii) Where animals are pictured, what animal species are shown, and are there differences here between users from Vienna and Tokyo? Finally, (iii) to what extent do socio-demographic factors and the total number of profile photos have an impact on the likelihood that users are (a) including pictures of animals in general, (b) using a photo with/of an animal as their first profile picture and (c) whether a cat or dog is displayed on one of the profile photos?2. Material and Methods2.1. Cross-Sectional Structured Observational Study and the Role of the ObserverThe cross-sectional structured observational study [26] aimed to gather quantitative data on the presentation of animal photos on Tinder® (Match Group, Inc., Dallas, TX, USA) profiles. The researcher (C.D.) had the status of an independent scientific observer who collected data on socio-demographic factors of app-users and the presentation of animal photos on user profiles. This was an observation with a low degree of participation since the researcher took a passive role without making any contact with the other users at any point in the study. Data were collected based on information from profiles only. At no point did the researcher “like” profiles or answer received “likes” from other users. In addition, the observation was covert, as the observed users were unaware that the researcher profile was created by a research group.The study was approved by the head of the Ethics Committee of the Medical University of Vienna (Ref: 1698_001) and the chief privacy officer of the Vetmeduni, Vienna. 2.2. Research Account and Set-Up of Online Dating ProfilesA “Tinder® plus” account was set up for data collection. This type of account allows users to search globally for other users regardless of where they are located. In total, eight chronologically consecutive profiles were set up for data collection: four for Vienna and four for Tokyo. Table 1 shows characteristics of the profiles designed for this study, including city, name of designed profile, gender, sexual orientation and age (see Table 1). Since the selected app captures gender and sexual orientation in binaries, both variables were gathered as dichotomous binaries. Since the selected dating app put an emphasis on profile photos, the same photo was uploaded for each research profile displaying Saint Stephen’s Cathedral in Vienna. This rather neutral photo was chosen to avoid giving false impression and false expectations to other users who came across the research profile while using the dating app.2.3. Selection Process and Search Criteria for ObservationThe target and source population for this observational study were users of the selected online dating app, and the users eligible for inclusion in the study needed to have at least one uploaded profile photo [27]. In total, 2400 Tinder® profiles were examined (300 profiles per research profile). Based on the set-up options provided by the app, gender (female/male), sexual orientation (heterosexual/homosexual), age (in years) and distance (in km) were the relevant criteria for the stratified sampling. Regardless of whether or not these profiles showed an animal, socio-demographic data were collected, and the numbers of profile photos were noted. However, the main target group of participants was those users who displayed animals and hence would contribute data on the investigated research questions. The data were collected in the period 16–22 March 2021. The eight research profiles set up reflected differences in gender, sexual orientation, age and search distance in km (see Table 2).2.4. Development of the Structured Observational StudyA set of questions was developed based on a preliminary investigation of the profiles with the aim of gaining quantitative data on the prevalence of animal photos on the selected online dating app. This preliminary investigation was of relevance to identify what information is provided on the app, since this has an impact on what questions, in relation to socio-demographic aspects or photos, can be answered in the course of the study. As a consequence, questions and codes were created related to the information provided by profiles. The sets of questions and codes were piloted against ten profiles showing animals. This pilot served to check (i) that the sequence of questions was practicable; (ii) that the animals were categorised in the same way by both researchers (C.D., S.S.), ensuring inter-rater reliability; and (iii) that the structure and content of questions were suitable for the observational study. Based on results from the pilot, minor adjustments and changes were made in order to improve the practicability of the questionnaire and the content of the questions. In case of uncertainty in relation to the categorisation of animals, the responsible researcher for data collection (C.D.) was in exchange with the second researcher involved in this project (S.S.) in order to ensure consistent and appropriate categorisation of animals during the entire period of data collection. 2.5. Study Design and MeasurementsThe questionnaire for study consisted of three sections. Section A included three closed-ended questions about sexual orientation, age and number of uploaded profile photos. Section B requested information on the animals presented in the profile photos (see Table 3).To ensure the investigated profiles in Vienna and Tokyo were comparable, the categorisation into farm animals, wild animals and exotic animals was always based on the Austrian context. Thus, for instance, regardless of whether an eagle appeared on an Austrian or Japanese profile, in Vienna or in Tokyo, the animal was categorised as a wild animal since eagles are native to Austria’s wilderness. Whales, in contrast, were always categorised—again, based on the Austrian context—as an exotic animal, even though they are considered wild animals in Japan. The third section, C, focused on the written profile texts of the users. The data from section C are not relevant to the research questions considered in this paper and are therefore not presented here. In addition, a research diary [28] was used in which the observer (C.D.) recorded notes during data collection. The aim of the diary was to collect impressions of investigated profile photos that would not be revealed by the set of closed-ended questions. Hence, these notes provided a qualitative view of the data going beyond the quantitative findings, which may serve as a basis for the development of hypotheses for future research studies.2.6. Data Collection and AnalysesThe questionnaire was designed using the survey software Alchemer® (Alchemer®, Louisville, CO, USA). Univariate descriptive statistics were presented in tables, figures or text. For bivariate analysis, χ2-tests were conducted to assess whether the frequency distribution of gender and city, “Animal(s) on profile photo(s)”, “Animal(s) on the first profile photo” and “Classified animal species” differed in the sub-populations in Vienna and Tokyo. Mann–Whitney-U Tests were conducted with the variable age, number of profile photos (continuous variables) and age groups (ordinal scaled variable) to detect differences between (a) Vienna and Tokyo and (b) profiles with animals and without animals. Spearman’s rho correlations were calculated to identify whether associations between the (i) number of profile photos, (ii) age of user and (iii) number of photos displaying animals were associated.We then pooled the data from both cities and ran three binary logistic regression models to understand how changes in values of the socio-demographic factors were associated with changes in the probability of the following outcomes: (a) users are displaying animals in general on photos, (b) users are displaying animals on their first profile photo, and (c) a cat or dog is displayed on a photo. The dependent variables were inserted on a dichotomous scale. The scale for (a) and (b) was: 1 = yes; 2 = no, and the scale for (c) was: 1 = dog; 2 = cat. Categorical predictor variables inserted in the regression analyses were gender (1 = male, 2 = female), city (1 = Vienna, 2 = Tokyo) and sexual orientation (1 = heterosexual; 2 = homosexual). Age and number of profile photos were included as continuous predictor variables. IBM® SPSS® Statistics version 26.0 (IBM® SPSS® Statistics, Chicago, IL, USA) was used in all analyses. The significance level was 0.05.3. Results3.1. Overview of Socio-Demographic ResultsIn total, 2400 profiles of the selected online dating app were included in the study. Using the research profiles, we created (see Table 1) we collected data from 300 heterosexual and homosexual men, and 300 heterosexual and homosexual women, in Vienna and Tokyo. With a mean age of 28.6 ± 4.5 years, users in Vienna were significantly older than users in Tokyo (27.0 ± 5.1; p < 0.001). Further, the users of the investigated profiles in Vienna showed significantly more photos on their profiles than those in Tokyo (p < 0.001). In Tokyo, a positive correlation was identified between user age and number of profile photos (rs = 0.134, p < 0.001) (see Table 4).3.2. Comparison between Profiles with Animals and Profiles without AnimalsOf the 2400 investigated profiles, 373 (15.5%) displayed at least one animal photo. In both cities, we found a positive correlation between the number of profile photos and the number of profile photos showing animals (Vienna: rs = 0.184; p = 0.008 | Tokyo: rs = 0.206; p = 0.009). Comparison of the users who displayed animal photos on their profile and the users who did not do so resulted in the following significant differences (see Table 5). On the selected analysed dating app, significantly more women than men (p = 0.049) present animal photos on their profiles. Further, significantly more users in Vienna (p = 0.006), and significantly more older users (p = 0.019), have profiles with animal photos as compared with users in Tokyo and younger users. In addition, users who display an animal photo on their profile post, on average, display one more photo than users who do not do so (p < 0.001). No significant differences between heterosexual and homosexual users of the analysed app were identified (p = 0.639) (see Table 5).3.3. Prevalence and Categorisation of Animals Displayed on ProfilesA further aim of the study was to determine how many profiles showed animals and what kinds of animal were presented. In general, significantly more users in Vienna (211; 17.6%) show animals on their profile than users in Tokyo (162; 13.5%) (χ2(1) = 7.622; p = 0.006). Most of the profiles—i.e., 77.7% in Vienna and 76.5% in Tokyo—showed the animal, or animals, on just one profile photo. In a smaller proportion of cases—i.e., 22.3% in Vienna and 23.5% in Tokyo—the users had more than one photo showing the animal, or animals, in their profile.3.3.1. Presentation of Animals in the First Profile PhotoOf the 373 users choosing to include animal photos, 73 (19.6%) displayed the animals on their first profile photo. Here, comparison of users in Vienna and Tokyo revealed significant differences as 65.9% users in Vienna showed a dog on the first photo as compared with 31.3% of users in Tokyo (χ2(1) = 8.610, p = 0.003). In addition to that, only profiles in Vienna (12.2%) have shown farm animals on the first profile photo. This resulted in a significant difference to users in Tokyo (χ2(1) = 4.189, p = 0.041). We also found that more users in Tokyo displayed cats (40.6%) and exotic animals (15.6%) in their first profile photos than users in Vienna (cats = 2.4%; exotic animals = 0.0%) (cats: χ2(1) = 7.819, p = 0.005; exotic animals: χ2(1) = 6.877, p = 0.009).3.3.2. Presentation of Animals in All Profile Photos (Including the First Profile Photo)Figure 1 shows the percentages of various animal species shown on the analysed profiles. Again, comparison between the profiles in Vienna and Tokyo revealed significant differences here. Users in Tokyo were significantly more likely to show cats (35.8%) and small animals (6.8%) than users in Vienna (cats = 18.0%; small animals = 0.0%) (cats: χ2(1) = 15.179, p < 0.001; small animals: χ2(1) = 14.763, p < 0.001). The Viennese profiles included farm animal (10.9%) and horse (7.1%) photos significantly more often than the profiles in Tokyo (farm animals = 0.6%; horses = 1.2%) (farm animals: χ2(1) = 16.096, p < 0.001; horses: χ2(1) = 7.270, p = 0.007) (see Figure 1).3.4. Effects of Socio-Demographic Factors and Number of Profile Photos on Posting Animal PhotosWe pooled data from Vienna and Tokyo and conducted three binary logistic regression models in order to determine whether socio-demographic factors or total numbers of profile photos were associated with the decision of the user (i) to show an animal, or animals, at all; (ii) to show an animal, or animals, on the first photo; or (iii) to show a dog rather than a cat (or vice versa) (see Table S1). We found that users of the analysed app with several profile photos, and women, are more likely to display an animal on their profile photo than users with fewer photos (χ2(1) = 86.798; p < 0.001) and men (χ2(1) = 3.942; p = 0.047). Additionally, being a younger user (χ2 (1) = 9.315, p = 0.002) and having a higher number of profile photos (χ2(1) = 17.462; p < 0.001) increased the probability of the animal being shown in the first profile photo. Significantly more often than users in Tokyo, Viennese users displayed photos of dogs, but where cats were concerned the reverse was found to be the case (χ2(1) = 14.281; p < 0.001).3.5. Qualitative Notes in Research DiaryNotes were taken with the aim of gaining qualitative insights into the data, and specifically developed preliminary conclusions about how animals are presented on the investigated profiles. The notes suggested that two typical types of presentation of animals (including sub-categories) could be identified. These types were not mutually exclusive. However, they provide an initial overview of the ways in which animals are presented on the analysed online dating app, and they offer a basis on which to develop category systems for future investigations.(i)The animal as close friend and/or family member: this form of presentation includes photos showing a portrait of an animal. The animal is prominently displayed, and the impression given is that the animal is as “deserving” of its own photo in the gallery of the user as are children or close friends. Its inclusion is part of describing one’s life and important companions in photos.(ii)The animal as a mirror of the user’s personal character traits: this type of photo is not about the animal per se. Instead, the form of presentation gives certain insights into the life and character traits of the user. In total, four sub-categories of this form of presentation were identified: (a)The animal as a part of an emotional encounter: A typical photo shows a user cuddling with a dog or cat in bed.(b)The animal as part of an active and/or sporty lifestyle: here, animals are used to indicate activity and a healthy, sporty and active lifestyle through their presentation in the context of outdoor activities. A typical photo shows a user jogging with a dog or looking at a cow while hiking.(c)The animal as part of aesthetics: in this context, animals are used to communicate the user’s fashion or housing style. A typical photo shows a white dog with a turquoise hair band on the couch in a flat where all the design elements are white or turquoise.(d)The animal as part of humour: here, animals appear humorously staged. A typical photo shows a cat with an amusing hairstyle or with its eyes wide open as if amazed at what it sees. Photos in this category are intended not so much to show the user’s own animals but rather as internet memes.4. DiscussionThe present study provides the first systematically collected data on the prevalence of animal photos on profiles on a popular online dating app. The data indicated that around 16% of the investigated profiles displayed at least one animal. Due to a lack of data on this issue, it is difficult to assess this number against the background of previous studies. However, in recent years, a lively debate about the presentation of animals on online dating apps has taken place in media such as newspapers and online blogs [29,30,31,32]. In these media, reports and articles discuss the role of animals in online dating and ask specifically whether animals improve dating success rates, as Guéguen and Ciccotti [16] note. Based on the high level of media interest in this topic, the percentage of animals in profile photos might have been expected to be higher than 16%. Relatedly, Watson [31] uses the term “dogfishing” to mark the fact that it is not only dog owners who pose on their online dating profile with animals but also users who do not own a dog but believe that showing one will attract more matches. However, given the relatively low prevalence of animal photos on the analysed profiles we found, it seems questionable whether this occurs often in reality or is rather something hyped up by the media. If the phenomenon were frequent, we would surely expect the number of photos displaying animals to be higher.Turning to the question of what kinds of animal appear on the investigated online dating profiles, our evidence confirms that it is mainly dogs and cats that are shown. Here, differences between Tokyo and Vienna were identified. At approximately 70%, dogs and cats were the most frequently presented animals on the analysed profiles in the two cities taken together. Gray et al. [17] emphasise that dogs are particularly popular on online dating apps because they are more social and require more care than cats, meaning that they underline the social skills of the user. Comparison of the cities indicates that more dogs are shown in Vienna than are shown in Tokyo and that more cats are shown in Tokyo than are shown in Vienna. This result is in line with our expectations and echoes other study results. In one such study, based on data gained from 22 Mio Tinder® profiles in 16 countries, it was found that Austrian users presented dogs significantly more often they presented cats [33]. However, this data has to be interpretated with caution since it involved automised analysis using artificial intelligence software.In general, the number of dogs and cats in modern societies’ households, as well as the intensification of relationships between these pets and their owners, is mirrored in the high prevalence of dogs and cats displayed on profiles showing animals. One explanation for this could be that our data were collected in urban areas, and hence the users had little direct contact with farm animals or wild animals. Perhaps in rural regions, the prevalence of animals beyond dogs and cats would rise in comparison to urban areas. Profiles that display particularly “unusual” animals such as wildlife or zoo animals, which might represent the more extraordinary life, play a rather vanishing role. With this in mind, it seems fair to say that the animals presented on profiles on the analysed app are mostly those with whom users frequently have direct and/or emotionally close contact. These offer insights into the everyday life of the user.Turning to the relationship between socio-demographic factors and whether or not online dating users show animals on their profiles, our results indicate that gender, age and users’ cultural background (Tokyo as compared with Vienna) all have an impact on the probability of animals being shown. Thus, we found that female and older users are more likely than men and younger users to show animals on their profile photos. Although some of the animals (most obviously, wild creatures) would not have been owned, in general we did not gather information allowing us to determine whether the displayed animals (principally, the dogs and cats) were actually owned by the users. However, our findings are in line with other study results. For instance, a European study of pet owners has found that women and older adults are more likely than men and younger people to keep animals [21,22]. A possible explanation for this is that older adults often maintain lifestyles that make keeping animals easier. Younger people are more likely to want, or need, to have the flexibility to, for example, move between different cities. They may live in small apartments together with other people, or have no regular income to shoulder the costs of keeping a dog or cat. Sexual orientation did not show any significant differences in the analyses we conducted.As regards possible cultural differences between users in Tokyo and Vienna, we found that significantly more users in Vienna than users in Tokyo had profiles with animals. Another significant difference was that only the Viennese profiles showed farm animals on the first profile photo. These differences can be plausibly explained by the urban structure of the cities. With around 10 million people, Tokyo has far more inhabitants than Vienna, which has around 2 million. It is also larger geographically (628 km2, as against Vienna’s 415 km2), which means that users in Tokyo tend to be further away from rural areas. Again, the population density in Tokyo is far higher (15,351/km2, as against Vienna’s 43,261/km2), suggesting that housing conditions there are more cramped and less pet-friendly. This last point may also help to explain why users in Tokyo are more likely than users in Vienna to display photos of cats, while dogs are more popular on Viennese profiles than they are on the profiles of users in Tokyo. Vienna—as we touched upon in the introduction—is the more dog-friendly of the two cities. However, the urban structure of Vienna and Japan may not be the only explanation for this difference, but rather general cultural differences between the two cities are decisive for the different prevalence of dogs and cats in profiles. For instance, it might be that dogs are generally more popular in Austrian culture or are ascribed a higher social status than in Japan. Nevertheless, where socio-demographic factors are concerned, we see that choices about the presentation of animals on the analysed app have close connections with everyday human–animal relationships. Online dating seems to be like a mirror of everyday human–animal relationships.Self-presentation in online-dating is rarely spontaneous. In most cases it is a consciously planned activity [4,5]. It seems reasonable to infer that in most cases the inclusion of an animal photo is considered carefully the user. We might say that it is part of the user’s “impression construction” [5], by which we mean that users think about what impression they want to make and how they can achieve it. However, it is not only the fact that an animal is shown that seems to be relevant in this context, but also how users present animals. Our qualitative notes provide exploratory insights into this issue. In particular, they indicate an interweaving of “real life” and the use of the investigated dating app that manifests itself in various ways. Thus, the two typical presentations of animals (including the sub-categories) that we set out above in Section 3.5 correspond with the two key types of photo elaborated by Degen and Kleeberg-Niepage [4] in their qualitative, serial analysis of Tinder® profile photos. According to the authors, so-called “informative type” photos offer insights into hobbies, activities and lifestyle. “Type of sociability and enjoyment” photos, by contrast, show the user interacting or in a social context [4]. In both cases, animals seem to be a great vehicle for conveying something of the user’s lifestyle and/or personality. With this mind, we might say that the displaying of animal photos serves to communicate aspects of the user—such as being empathic, kind, social, active, healthy, sporty, aesthetic or humorous—that play an important role in online dating for users, whether those users have long-term or short-term dating goals [34,35].We believe that the present observational study provides useful preliminary insights into the ways different species of animal are displayed on the selected online dating app, and how these animals are presented on profile photos. However, we also recognise that the study has several limitations. We considered a sample size of a total of 2400 profiles including 300 profiles per research profile as sufficient in order to conduct analyses and provide first results on the outlined research questions. However, we suggest that future research may increase the number of profiles that can be investigated by research profiles. Further, the eight research profiles we set up meant that the study included only users of a specific app of a certain age group (20–40 years). Younger, as well as older, users as well users of other online dating apps were not investigated. Further, we decided to examine possible cultural differences by looking at just two cities—Vienna and Tokyo. We are aware that these two cities do not permit overall conclusions to be drawn about cultural differences more generally. In addition, the data were, of course, collected in urban regions. Hence, profiles in sub-urban as well as rural areas are yet to be investigated. As a consequence, we strongly recommend that future studies on this issue should, if possible, examine cities with different cultural backgrounds (e.g., in the US or Africa), consider not only urban, but also sub-urban and rural areas, and include profiles of younger (under 20 years) and older users (over 40 years) as well as users of other online dating apps. In addition, there may exist further aspects that influence the sampling of profiles that go beyond the defined search criteria used in this study. For instance, the number of “likes” a user receive can play a role. This aspect may result in a potential bias and sampling error of profiles investigated in this observational study. 5. ConclusionsAnimals do play a role in online-dating. Around 16% of the investigated profiles on the selected app displayed at least one animal. However, based on the high level of media interest in this topic, the percentage of animals in profile photos might have been expected to be higher. In both investigated cities (Tokyo and Vienna), dogs were the most frequently shown animal, followed by cats. Users in Vienna, women and older adults were more likely to present animals on their profiles than were users in Tokyo, men and younger users. Sexual orientation showed no significant differences in the analyses. We conclude that the animals most likely to be shown are those in a close and frequent contact with users (i.e., certain kinds of companion animal). Given that self-presentation in online-dating is usually not a spontaneous act but a consciously planned one, future studies could explore in depth the ways in which animals are represented and arranged on profile photos.

animals : an open access journal from mdpi

10.3390/ani12010065

PMC8749870

Mastitis is the most common inflammatory disease of economic and animal welfare concern in dairy animals. The present study was designed to evaluate the gene expression and epigenetic modifications in cattle with mastitis and healthy cows. The CpG islands in the promoter regions of the JAK2 and the STAT5A revealed hypo-methylation levels and higher gene expression in cows with mastitis compared to the healthy control, and vice versa in those with the CD4 gene. DNA methylation was negatively correlated with gene expression in the JAK2 and CD4 genes. Findings of the current study showed that aberrant DNA methylation due to mastitis in the promoter region of the three genes under study could be used as potential epigenetic markers to predict the mastitis susceptibility in dairy cattle.

The present study was designed to evaluate the gene expression and DNA methylation level in the promoter region of the CD4 and the JAK-STAT-pathway-related genes. A total of 24 samples were deployed in the gene expression and 118 samples were used in the DNA methylation study. Student’s t-tests were used to analyze the gene expression and DNA methylation. The evaluation of DNA methylation in promoter regions of JAK2 and STAT5A revealed hypo-methylation levels of CpG sites and higher gene expression in cows diagnosed with mastitis as compared to the healthy control, and vice versa in those with CD4. DNA methylation was negatively correlated with gene expression in JAK2, STAT5A, and CD4 genes. Six, two, and four active transcription factors were identified on the CpG sites in the promoter regions of JAK2, STAT5A, and CD4 genes, respectively. Regarding correlation analysis, the DNA methylation levels of CD4 showed significantly higher positive correlations with somatic cell counts (p < 0.05). Findings of the current study inferred that aberrant DNA methylation in the CpG sites at the 1 kb promoter region in JAK2, STAT5A, and CD4 genes due to mastitis in cows can be used as potential epigenetic markers to estimate bovine mastitis susceptibility in dairy cattle.

1. IntroductionMastitis is an inflammation of the mammary gland and is characterized by pathological, physiological and bacteriological changes in the udder which affect the quality and quantity of the milk [1]. Mastitis is the most common inflammatory disease of economic and animal welfare concern in dairy animals. Globally, published estimates of the economic losses of clinical mastitis per cow on a farm range from EUR 61 to EUR 97 [2]. In addition to economic importance, bovine mastitis also carries public health significance. Milk and other dairy products are often reported to be contaminated with S. aureus and E. coli. Milk from cows with sub-clinical mastitis accidentally mixed into bulk milk enters food chain, thus causing a great threat to human health [3]. The California mastitis test and somatic cell count are tests routinely performed for the detection of mastitis in dairy animals [4]. Large numbers of pathogens causing mastitis; therefore, successful vaccination to control the disease is not effective. Generally, antimicrobial agents and management strategies are used to combat the disease in dairy cattle [5]. However, marker-assisted selection using genetics and epigenetics approaches is considered an appropriate strategy to minimize the incidence of mastitis in dairy cattle [6,7].DNA methylation is a biochemical process in which a methyl group is added to the 5′-carbon of cytosine in the CpG dinucleotide sequence of DNA by the catalytic activity of the DNA methyltransferases, i.e., DNMT1, DNMT3A and DNMT3B [8]. The epigenetic markers of DNA are heritable during cell division, and do not alter the DNA sequence. Aberrant DNA methylation at CpG islands in the gene promoter region results in transcriptional silencing and is associated with many disease conditions, i.e., cancer formation and tumor progression [9]. Aberrant DNA methylation suppresses the transcription by inhibiting the binding of specific TFs [10]. Hypomethylation of global DNA methylation induced by S. aureus infection suppressed DNA methyltransferase activity in bovine mammary epithelial cells [11].The JAK-STAT pathway mediates signal transduction between nucleus and the cell surface receptors [12], and any disturbance in this inflammatory signaling pathway can result in various immune disorders such as immune deficiency syndromes, various cancer conditions, and mastitis in dairy animals [13,14]. CpG sites in JAK2 were hyper-methylated in myeloproliferative neoplasms compared with the healthy control, and 87.5% of the hyper-methylated CpG sites were located in the CpG island [15]. Our previous study demonstrated higher methylation levels of the CD4 promoter region and lower gene expression in clinically mastitic cows compared with healthy controls using pyrosequencing assays and qRT-PCR [16]. CpG islands are CpG-rich areas in the promoter region of highly expressed genes and are a common site of methylation. Day and Bianco-Miotto (2013) [17] reported that CpG islands in around 55% of cases form clusters and result in the inhibition or activation of transcription depending on the status of methylation. DNA methylation is the most understood mechanism amongst all epigenetic mechanisms. DNA methylation markings have been suggested to be relatively stable over time in adult individuals, showing average heritability of 0.19 [18], and are reported to play key roles in the regulation of gene activity, DNA repair, recombination, replication, and the establishment and maintenance of cellular identity [18]. Studies on mouse models have shown that the risk of tumor incidence can be transmitted across generations (up to three generations) through aberrant DNA methylation [19,20]. Due to the stable nature of DNA methylation across generations and its average heritability of 0.18–0.19 [21], it has been suggested to consider it as a potentially important epigenetic marker for selection in breeding programs. Therefore, we designed this study to evaluate the role of DNA methylation levels at the 1 kb promoter region in CD4 and JAK-STAT-pathway-related genes (JAK2, STAT5A, and STAT5B) in mastitic and healthy cows.2. Materials and Methods2.1. Sample Collection and DNA ExtractionBlood and milk samples were randomly collected from 118 lactating cattle (clinical mastitis, n = 58; healthy control, n = 60) from three Chinese Holstein dairy cattle farms (Qiqihar, Tianjin, and Shanxi) located in the northwest of China for DNA methylation analysis. The blood samples were collected in the morning between 9 a.m. and 11 a.m., whereas milk samples were collected at the routine milking time of 4 p.m. in the afternoon. A subset of 24 samples (of the total 118 samples) was used to evaluate the mRNA expression in the genes under study in mastitic and healthy cattle. We used routine dairy herd improvement (DHI) data to classify the different categories of lactating animal health status based on the SCC, i.e., healthy cows with SCC < 200,000/mL, subclinical mastitis cows with SCC ranging between 200,000 and 500,000/mL, and cows with clinical mastitis had SCC > 500,000/mL. The term “clinical mastitis” is used for cows with SCC > 500,000/mL based on the DHI data; however, these animals were apparently healthy and did not exhibit any clinical signs and symptoms. The healthy control represents lactating cows with healthy udders, without any history of mastitis in the last month, and SCC lower than 200,000/mL. The cows had different parities, ranging from parity one to five, and were milked three times per day (Table 1). The cattle were fed on a lactation diet as recommended by the Dairy Association of China for lactating cows (NRC2001). Milk samples were collected from all four quarters in sterile 50 mL falcon tubes and the SCC was analyzed at the official Dairy Center of China (Beijing, China). Blood samples were collected from the caudal vein of cows in three 9 mL tubes—one for serum isolation, the second for DNA extraction and the third for RNA extraction (the tubes contain pre-added TRIzol)—which were then immersed in liquid nitrogen at −196 °C to avoid any damage to the RNA. Serum was isolated from blood samples by incubating the samples at room temperature for 30 min followed by centrifugation at 3000× g for 10 min. The serum samples were then sent to Beijing Huaying Biological Technology Research Institute to evaluate serum biochemical assays of IL-4, IL-6, IL-10, IL-17, TNF-α, and IFN-γ using radioimmunoassay a technique, as described by Usman et al. (2017) [22].2.2. DNA and RNA Extraction, cDNA Synthesis, and Real-Time Quantitative PCRGenomic DNA and RNA extraction, reverse transcription of mRNA to cDNA, and real-time quantitative PCR were performed as described in our previous paper [14]. Primers were designed for JAK2, STAT5A, CD4, and GAPDH genes in Oligo 6.0 software, based on the golden rules for real-time PCR (Supplementary Materials Table S1) [23].2.3. Bisulfite Treatment of Extracted DNA and Hot Start PCRIn order to evaluate the DNA methylation in JAK2, STAT5A, STAT5B, and CD4 genes, we first checked the presence of CpG islands in the promoter region of these genes using a genome browser (UCSC BAU 6.0). We found a CpG island located in the 1 kb promoter region of all genes except for STAT5B. Sequencing primers of the JAK2, STAT5A, and CD4 genes were designed with Oligo 6 software for evaluating DNA methylation in the genes under study (Supplementary Materials Table S2). Genomic DNA (1 μg) of each sample was treated for sodium bisulfite conversion using the EZ DNA Methylation Golden kit, following the manufacturer’s protocol (ZYMO Research, Irvine, CA, USA). A volume of 20 μL elution buffer was used to elute the bisulfite-converted DNA (ZYMO Research). In our study, we used a biotin-labeled universal primer (5′-GGGACACCGCTGATCGTTTA-3′), as mentioned in our previous research [23]. A hot-start PCR was performed to amplify the target sequence in a 25 μL volume, including 12.5 μL hot-start PCR premix (ZYMO Research), 0.5 μM forward primer, 0.05 μM reverse primer labeled with universal tail, 0.45 μM biotin-labeled universal primer, and 20 ng bisulfite-converted DNA. The PCR protocol was 95 °C for 10 min, 94 °C for 30 s, 50 to 60 °C for 45 s, and 72 °C for 45 s, and a final extension at 72 °C for 10 min for a total of 45 cycles. The PCR products were checked by running them on 2% agarose gel stained with ethidium bromide for visualization.2.4. Quantitative DNA Methylation Evaluation Using PyrosequencingPyrosequencing assays were conducted to quantitatively examine the methylation levels in the promoters of JAK2, STAT5A, and CD4 genes in clinically mastitic cows and healthy controls. The DNA methylation level was tested for the nine CpG sites in the JAK2 gene (USCS database: Bau6.0, Chr8: 39751199~39750253), seven CpG sites in the STAT5A gene (USCS database: Bau6.0, Chr19: 43033111~43034137), and five CpG sites located in the promoter region of the bovine CD4 gene (USCS database: Bau6.0, Chr5: 104015622~104015994). Pyrosequencing techniques were used to evaluate the promoter DNA methylation levels using a Pyro Q-CpG system (Qiagen), following the manufacturer’s instructions. The high-quality purified DNA and the PCR products with high concentration (about 25 to 35 μL) were considered for pyrosequencing assays [24,25]. Streptavidin Sepharose High Performance (GE Healthycare) was utilized with the PCR products. The PCR products labeled with biotin and attached to Sepharose beads were distilled in 70% ethanol for 30 s, denatured for 30 s in denature buffer (0.2 M NaOH), then washed for 45 s with washing buffer with the Pyrosequencing Vacuum Prep Tool (Qiagen). Subsequently, 0.5 μM pyrosequencing primers of each gene was mixed with annealing buffer (Qiagen) in order to purify the PCR product. The methylation levels at each CpG site in the promoter region were revealed as the percentage [(mC/(mC + C)) × 100]. Here, mC indicates methylated cytosine and C denotes unmethylated cytosine. In order to verify bisulfite conversion, non-CpG cytosine residues were used as internal controls.2.5. Statistical AnalysisStudent’s t-tests were used for analyzing the mRNA expression (values from RT-qPCR) and DNA methylation between mastitic cows and healthy controls. Pearson’s correlations of DNA methylation levels with mastitis-related traits in the genes under study were evaluated using the SAS (9.1) package. TFSEARCH software was used to predict the transcription factor binding sites in the promoter regions of the study genes. 3. Results3.1. CpG Sites Methylation in the Genes under StudyThe methylation statuses of CpG sites in the CpG island of the 1 kb promoter region in all genes under study were analyzed in mastitic cows and healthy controls using a pyrosequencing assay, except for the STAT5B gene, where no CpG island was present. The results of the pyrosequencing assay are shown below as methylation pyrogram (Figure 1). The pyrosequencing assay revealed aberrant methylation in almost all of the CpG sites in the genes under study (Table 2).3.2. Predicted Binding Sites of TFs in the Promoter Regions of the Studied Genes The results showed six active transcription factors (TFs) (c-Myb, HSF, SRY, MZF1, ADR1, and Sp1) on the CpG sites of the JAK2 gene, two TFs (ADR1 and Ik-2) on the STAT5A gene, and four active TFs (cap, Sp1, GATA-1, and GATA-2) were predicted on the promoter region of the CD4 gene (Figure 2).3.3. Relationship of DNA Methylation with mRNA Expression To evaluate the results of DNA methylation with the transcription levels of the genes under study, we performed real-time quantitative PCR to estimate the mRNA expression levels. The results showed that the mRNA expression was significantly higher in clinically mastitic cows compare to the healthy controls in the JAK2 gene, whereas in the STAT5A and CD4 genes, the mRNA expression of mastitic cows was significantly lower than the healthy controls. Moreover, the mRNA expression and DNA methylation were negatively associated with JAK2 and CD4 genes, i.e., a higher methylation level was associated with a lower mRNA expression (Figure 3). DNA methylation in the figure below was calculated for the same cows for which mRNA expressions were analyzed.3.4. Correlation among the CpG Sites in the Genes under StudyWe analyzed the correlation among all of the CpG sites in CpG island of the genes under study. The results showed that almost all of the CpG sites were highly significantly correlated with each other in the JAK2, STAT5A, and CD4 genes (Supplementary Materials Tables S3–S5).3.5. Methylation Correlation with Mastitis Traits in the Studied GenesThe results of methylation correlations (Spearman) with mastitis indicator traits in JAK2, STAT5A, and CD4 genes are shown in Table 3. The results indicated that the methylation levels of the JAK2 gene exhibited a significant association with IL-4; STAT5A was significantly associated with IL-4 and Il-17; and the methylation level of the CD4 gene was highly significantly associated with SCC, SCS, mastitis status, and IFN-γ (p < 0.01).4. DiscussionUnlike genetic variants, which cause irreparable changes in the gene and can potentially result in gene activation or other effects, epigenetic modifications are known to be reversible changes which influence gene expression whilst keeping the DNA sequence unaltered [8]. DNA methylation is the most understood mechanism amongst all of the epigenetic mechanisms and its role is well known in mediating the gene expression [18]. DNA methylation is known to be an important epigenetic modification that potentially suppresses gene expression; thus, it can play a vital role in inflammatory conditions [17,18,19]. In recent years, the locus-specific methylation levels in peripheral leukocytes have emerged as suitable epigenetic markers in breast cancer [26] and in various other inflammatory conditions [17,18,19]. It has been reported that DNA methylation patterns can serve as a stable epigenetic marker of gene silencing memory in animals [19]. One recent study suggested potential regulatory roles of DNA methylation in bovine mammary glands during S. aureus-induced mastitis [27]. Song et al. (2016) reported aberrant DNA methylation in 1078 genes in cows with subclinical mastitis vs. healthy controls; most of these genes were associated with inflammation and ErbB signaling pathway [28]. A study in mice showed that DNA methylation is quite stable across generations and the epigenetic information can be passed on for up to three generations [19,20]. Due to the stable nature of DNA methylation across generations and its average heritability of 0.18–0.19 [21], it is suggested to consider it as an important epigenetic marker for selection in breeding programs. Based on the important function of DNA methylation in various inflammatory conditions in different species, its role as a potential epigenetic marker in the 1 kb promoter region of genes under study in the mastitis resistance of dairy cattle has been evaluated.We found that the methylation levels at CpG sites in the JAK2 gene in mastitic cows was significantly low and the mRNA expression was significantly higher, and vice versa in healthy controls (p < 0.05). This revealed that a lower methylation status of the JAK2 gene in mastitic cows possibly induces enhanced gene expression. Perez et al. (2013) reported that CpG sites in JAK2 were aberrantly methylated in myeloproliferaitve neoplasms compared with healthy controls, and 87.5% of the differentially methylated CpG sites were located in the CpG island [15]. JAK2 plays an essential role in the activation of the lmo2 leukemogenic gene through phosphorylation of the H3Y41 histone [29]. The results of the present study are in line with a previous study which reported that the methylation patterns of CpG sites in the JAK2 gene were hypomethylated after 24 h in E. coli-challenged porcine mammary epithelial cells compared with the control cells [30]. It is well documented that hyper-methylation of the CpG island is associated with decreased mRNA expression because hyper-methylated CpG sites can block the transcription factor (TF) motifs [31,32]. The presence of six active TFs in the CpG island of the JAK2 gene in the present study shows that these TFs could have a potential role in gene activation and silencing; thus, they can play important roles in mastitis resistance. The TFs revealed in a recent study were reported to be involved in transcription activation and gene regulation [33]. The variable DNA methylation levels of CpG sites in the JAK2 promoter region reveal that methylation at these sites could be a potential epigenetic marker for mastitis resistance. We suggest that further studies should be carried on epigenetic modifications (DNA methylation) of the JAK2 gene in dairy cattle with clinical mastitis in a larger population. In the STAT5A promoter region, the CpG sites exhibited higher methylation in healthy controls than mastitic cows, and the association between mRNA expression and methylation was non-significant. Moreover, the presence of the two active transcription factors in the CpG sites of the STAT5A gene manifest that these TFs can potentially play an important role in gene switch-on or switch-off. Stefanowicz et al. (2012) reported variable methylation between airway epithelial cells and the peripheral blood mononuclear cells [34]. DNA methylation selectively inhibits gene expression in the STAT5A gene of oncogenic cells [35]. To the best of our knowledge, this is the first epigenetic study focused on DNA methylation level evaluations in mastitic cows and healthy controls in the STAT5A gene. The results of the different DNA methylation levels, although non-significant, between the mastitic cows and the healthy controls, suggest that STAT5A could be considered in future epigenetic studies on mastitis resistance in dairy cattle.The findings of the present study are in complete agreement with our previous study that reported hyper-methylation and lower mRNA expression in the promoter region of the CD4 gene in mastitic cows compared with healthy controls [16]. CpG island methylation in the CD4 promoter region was downregulated in a line of chickens susceptible to MDV, whereas the gene expression was upregulated in the spleen of that particular line of chicken [36]. Many studies have reported that DNA methylation influences CD4 gene silencing and plays an important role in inflammatory conditions, causing the development and differentiation of CD4+ T cells [37]. Li et al. (2010) reported that the hyper-methylation of CpG sites can suppress gene expression by inactivating the transcription factor binding sites [31]. The presence of four active TFs at CpG sites of the CD4 gene indicates that these TFs might be involved in the switching on/off of the CD4 gene [31]. Hence, these results are promising, suggesting that DNA methylation at the promoter region of CD4 can be considered as a potential epigenetic marker for mastitis resistance. We recommend carrying out further research to verify the functional role of DNA methylation in promoter CpG islands of the CD4 gene in mastitis resistance studies.The results of the present study are in line with our previous study [16], although the time, individuals’ samples, and the environmental conditions were totally different between the two studies. These results indicate that CD4 DNA methylation at the promoter region can be used as a potential biomarker in mastitis resistance studies. Similarly, the study of JAK2 and STAT5A methylation status, and mRNA expression in mastitic versus healthy controls, indicated the potential role of the CpG island at the 1 kb promoter region in these genes in hypo-methylation, and higher mRNA expression in mastitic cows compared to healthy controls in the JAK2 gene. The results from TFSEARCH software revealed active transcription factors in the CpG sites of these genes, which indicates that variation in the TF methylation during mastitis can affect the switch-on or switch-off of the gene, thus modifying gene action.5. ConclusionsThe aberrant DNA methylation levels in JAK2 and CD4 genes between mastitic cows and healthy controls and their inverse relationship with gene expression suggests that these genes could be potential candidate genes of epigenetic importance. The CpG sites at the CpG island in the 1 kb promoter region of the studied genes showing peculiar methylation patterns are recommended to be studied in other populations at larger scale to validate their roles as potential biomarkers in future epigenetic studies on mastitis resistance in dairy cattle.

animals : an open access journal from mdpi

10.3390/ani11113119

PMC8614255

This study’s data suggest that under the projected scenarios of ocean acidification by 2100 and beyond, significant negative impacts on growth, health, and meat quality are expected, particularly on black sea bream, and will be susceptible to the scientifically approved fish having a weaker resistance to diseases and environmental changes if CO2 emissions in the atmosphere are not curbed. Knowing the expected consequences, mitigation measures are urgently needed.

Acidification (OA), a global threat to the world’s oceans, is projected to significantly grow if CO2 continues to be emitted into the atmosphere at high levels. This will result in a slight decrease in pH. Since the latter is a logarithmic scale of acidity, the higher acidic seawater is expected to have a tremendous impact on marine living resources in the long-term. An 8-week laboratory experiment was designed to assess the impact of the projected pH in 2100 and beyond on fish survival, health, growth, and fish meat quality. Two projected scenarios were simulated with the control treatment, in triplicates. The control treatment had a pH of 8.10, corresponding to a pCO2 of 321.37 ± 11.48 µatm. The two projected scenarios, named Predict_A and Predict_B, had pH values of 7.80-pCO2 = 749.12 ± 27.03 and 7.40-pCO2 = 321.37 ± 11.48 µatm, respectively. The experiment was preceded by 2 weeks of acclimation. After the acclimation, 20 juvenile black sea breams (Acanthopagrus schlegelii) of 2.72 ± 0.01 g were used per tank. This species has been selected mainly due to its very high resistance to diseases and environmental changes, assuming that a weaker fish resistance will also be susceptibly affected. In all tanks, the fish were fed with the same commercial diet. The seawater’s physicochemical parameters were measured daily. Fish samples were subjected to physiological, histological, and biochemical analyses. Fish growth, feeding efficiency, protein efficiency ratio, and crude protein content were significantly decreased with a lower pH. Scanning electron microscopy revealed multiple atrophies of microvilli throughout the small intestine’s brush border in samples from Predict_A and Predict_B. This significantly reduced nutrient absorption, resulting in significantly lower feed efficiency, lower fish growth, and lower meat quality. As a result of an elevated pCO2 in seawater, the fish eat more than normal but grow less than normal. Liver observation showed blood congestion, hemorrhage, necrosis, vacuolation of hepatocytes, and an increased number of Kupffer cells, which characterize liver damage. Transmission electron microscopy revealed an elongated and angular shape of the mitochondrion in the liver cell, with an abundance of peroxisomes, symptomatic of metabolic acidosis.

1. IntroductionThe full report of the Intergovernmental Panel on Climate Change (IPCC) published in 2019 [1] informed that from the pre-industrial period (1850–1900) until the present time, 2021, the Industrial Revolution involving the anthropogenic emission of CO2 into the atmosphere has led to climate change, causing ocean warming and acidification [2,3,4,5,6,7,8,9,10]. Considering the land-based greenhouse gas (GHGs), the fifth assessment report on climate change (AR5) conveyed with medium confidence that the annual net CO2 emissions from anthropogenic land-use change were 0.9 (0.1–1.7) gigaton of carbon per year (GtC∙yr–1), on average, from 2002 to 2011. From 1750 to 2011, CO2 emissions from fossil fuel combustion have released a mean value estimated as 375 (345–405) GtC to the atmosphere, while deforestation and other land-use changes have released an estimated mean value of 180 (100–260) GtC. Of these cumulative anthropogenic CO2 emissions, a mean of 240 (230–250) GtC have accumulated in the atmosphere, 155 (125–185) GtC have been taken up by the ocean, and a mean value of 160 (70–250) GtC have accumulated in terrestrial ecosystems [1,11]. Currently, industrialization is exponentially higher than before. As a result, the projected Ocean Acidification (OA) in the coming century is assumed to negatively impact aquatic ecosystems, with increased risks of environmental pollution and a significant threat to its living resources [12,13,14,15,16,17,18].As a consequence of the industrial revolution, the atmospheric concentration of carbon dioxide (CO2) is continually increasing in our oceans [19], thereby increasing the concentration of hydrogen ions in seawater. Seawater consequently shifts into the acidic direction, decreasing the ocean’s environmental pH. According to the IPCC [1], global warming resulting in OA has caused a 0.1 reduction in ocean pH. Remarkably, this slight decrease in the ocean’s pH (i.e., ∆ pH = 0.1) from pre-industrial times until the present corresponds to a 26% to 30% increase in water acidity due to its logarithmic scale, pH =−log [H+] (i.e., [H+] = 10^(-pH)) [20]. This slight increase of 0.1 in the pH value is significant when considering the pollution and other negative effects on the marine environment and marine species [19,20,21,22,23,24,25,26,27,28,29]. Furthermore, a model projection scenarios indicate that by 2100 and beyond (23rd century), the oceans will warm up by about 3 to 4 °C, leading to an additional decrease of pH estimated from 0.3 to 0.4 in 2100 and 0.7 to 0.8 in the 23rd century, depending on the region, habitat, and emission scenario [13,30,31,32]. It is estimated that the ocean pH in the pre-industrial period was about 8.18. Nowadays, this pH has dropped to around 8.08 [1,25,33]. An important number of articles have reported that such a slight decrease of ∆pH = 0.1, corresponding to 26% increase of acidity had a significant negative impact on the marine environment and species [13,30,31,32]. How much impact will the projected decrease of 0.3 in pH, which is an increase of 99.53% in water acidity, have at the end of the 21st century? And how much impact will a further decrease of 0.7 in pH, which is an increase of about 401.19%, have? We have designed this study to measure the impacts on marine fish and see what mitigating measures are needed.Considering the great benefit of black sea bream (Acanthopagrus schlegelii) to the world fishing economy (6703 tons in 2014) and aquaculture industries (154,389 tons in 2014 including mariculture) [34,35], there is a need to study and understand the risks of OA on marine fish to take preventives measures. The present research intends to evaluate the negative effects of OA on fish survival, health, growth, and fish meat quality. Our targeted values correspond to the projected acidification pH values of 7.80 and 7.40 projected in 2100 and beyond, respectively. It is worth mentioning that Han et al. [3] used these same projected pH values and explained how OA might threaten the population recruitment of broadcast spawning marine species. Likewise, Zhao et al. [36] demonstrated how the pH values 7.80 and 7.40 affected mussels, damaging their shell structure and reducing their shell strength and closure strength. Cao et al. [12] confirmed that due to OA, an elevated pCO2 impairs pacific oysters’ immune function, increasing the risk of enhanced disease of marine mollusks. Furthermore, Araujo et al. [20] reported the negative synergistic impacts of ocean warming and acidification on the survival and proteome of gilthead seabream (Sparus aurata). The black sea bream inhabits shallow waters of with 1 to 50 m of depth [37], and has an excellent meat quality and a fast growth rate. This fish species is also known for its efficient feed conversion, high market value, and particularly its high resistance to diseases and environmental changes [38,39,40]. The Data from Nansei National Fisheries Research Institute, published by Sako [41], reported a trend of cultured fish diseases in Japan. They investigated the resistance to diseases of more than thirteen species of fish, including black sea bream. In their results, black sea bream was classified as the most resistant species, having the lowest case (17) of diseases reported, followed by the second more resistant fish, dark-banded rockfish (Sebastes inermis), with 21 cases, up to the less resistant one, yellowtail (Seriola quinqueradiata), with 1428 cases of diseases [41]. All these characteristics qualify the black sea bream as a potentially essential aquaculture species for offshore culture in the Yellow Sea, the Bohai Sea, the East China Sea, and the South China Sea, also affected by OA [42]. It is assumed that if black sea bream, with a high disease resistance, is affected by OA, other fish species with weaker resistance will be highly susceptible. To the best of our knowledge, no previous studies have considered the effect of the rising OA on the growth and survival of black sea bream, an important aquaculture candidate in Asian countries, as mentioned earlier [34,35]. The present experimental study seeks to understand the implications of the projected 7.80 and 7.40 pH values, respectively, on the survival, growth, health, and meat quality of black sea bream (Acanthopagrus schlegelii).2. Materials and Methods2.1. Experimental Design and Procedure, Seawater ParametersThree different pH values were used as three treatments, each treatment having three replicates for statistical analysis. The seawater source supply was pumped from the sea, decanted, and well filtered by sand. That seawater supply, with its initial physicochemical characteristics, was directly used as our control group and had a constant pH of 8.10 ± 0.01 during the experiment. This first group corresponds to the present condition of the marine environment. The second and third groups, respectively Predict_A and Predict_B, are the future projection values of pH, respectively pH = 7.80 and pH = 7.40, projected to occur in the year 2100 and beyond. These last two values of pH, 7.80 and 7.40, were obtained by a mixture of dry air and pure CO2 and kept stable through low-pressure flow controllers LZM-6T and 100–1000 mL/min oxygen air gas flowmeter with a control valve to adjust CO2, working at a pressure p = 1atm = 101,325 Pa [3,29,36]. Each tank was filled with 200 L of seawater, with a system of a constant water inflow/outflow of 2 L/min. The pH values in each tank were checked daily by the pH meter Hach SC200 Universal Controller. Before utilization, the pH meter Hach SC200 was calibrated with standard NBS buffers. The pH meter Hach was put inside the seawater tank for 1 h to make sure the evolution of the pH curve over time stayed constant. The constancy of the pH curve over time and the CO2 flow controller valve were adjusted only once at the end of the week according to the slight increase of the regression line displayed (Figure 1). The total alkalinity (TA) was assessed through potentiometric titration with an SM-Titrino 702 automatic titrator system [36]. The carbonate chemistry parameters, pCO2 (µatm), bicarbonate ions (HCO3−) (µmol/kg), carbonate ions (CO32−) (µmol/kg), dissolved inorganic carbon (DIC) (µmol∙kg−1), and the saturation state of aragonite and calcite, respectively Ωara and Ωcal, were all calculated from the measured salinity, measured pH, measured temperature, and measured total alkalinity TA, at an atmospheric pressure of 1atm = 1013.25 kilobar, using the open-source software “CO2cal 1.2.0”. CO2cal 1.2.0 is the updated version of CO2SYS, previously used by many other articles [3,29,36].pH Stabilization CurveThe following picture displays the stabilization of the pH value in treatment 3, in the Predict_B group, where the desired pH value is 7.40. For the sake of a high and stable temperature, the experiment was conducted in the summer, from June to August, at the Marine Fisheries Research Institute of Zhejiang Province, Zhoushan City, China, located in Xixuan Island, to take advantage of the high and stable temperature of 27.53 ± 0.85 °C. Black sea bream fingerlings with an initial weight of 1.52 ± 0.03 g were obtained from a private fish farm and transported by boat to the laboratory. Before initiating the main experiment, all fish were first stocked in a single, very large tank, with a capacity of about 14,824 L, and the following dimensions: 4.62 m × 2.63 m × 1.22 m. It was opaque, blue, and made of fiberglass. All fish were fed with a commercial diet (42% crude protein, Ming-Hui Feed Co. Ltd., Zhejiang, China) for 2 weeks of acclimation, to get used to the environmental conditions of Xixuan Island and to overcome the stress caused by the transportation [39,43,44]. Only the surviving and healthy fish remaining after acclimation were selected for the experiment. After the 2 weeks of acclimation, 20 healthy fish, black sea bream fingerlings with an initial weight of 2.72 ± 0.01 g (mean ± SD), were carefully selected and stocked per tank. Each of the three groups had three replicates for a total of nine tanks. Each of the nine smaller tanks had a diameter Φ = 80 cm and a height of 65 cm, with a cone-shaped bottom also made of opaque blue fiberglass with a capacity of 350 L. The fish were maintained under a natural photoperiod. The temperature, the ammonia-nitrogen, and the salinity of the seawater in the tanks were, respectively, 27.53 ± 0.85 °C, 0.02–0.04 mg∙L−1, and 27.13 ± 1.06 g∙L−1. The dissolved oxygen concentration was maintained above the safe value of 5.0 mg∙L−1—more precisely, at 7.05 ± 0.51 mg∙L−1—at any point during the experiment by continuous aeration with air-stones [39,43,44]. For the first two weeks of the experiment, fish were fed three times a day (08:00, 12:00, and 16:00), and then, for the remaining 6 weeks, twice a day (08:00 and 16:00). The fish were hand-fed little by little until apparent satiation, for the efficacy of feeding and to prevent uneaten feed waste. To avoid feed nutrient leaching and to have a very stable feed in the seawater, we used an expanded diet, which means a floating diet that is a professional commercial feed. The fish would swim to the water surface to ingest the feed pellets because the expanded feed floats in water. As long as the fish were fed to apparent satiation, they would not come up to the water surface again. Hence, their apparent satiation could be judged visually [39,43]. The experiment lasted for 56 days (8 weeks), and feed consumption was recorded daily. The tanks were thoroughly cleaned, and the mortality was also checked daily. 2.2. Sampling for Growth Parameters, Proximate Composition, Histological StudiesThe growth performance of the fish was evaluated in each treatment on the standard feeding basis of 8 weeks. Eight weeks is the international duration norm for fish growth trials, especially when evaluating the growth performance [43,45,46]. After the 56 days of the experiment, the survival percentage was recorded in each tank. On the 57th day, the fish were first starved for the whole day before sampling [43,46,47]. Then, on the 58th day, the fish were anesthetized with Tricaine methane-sulfonate MS-222, 60 mg∙L−1. After anesthesia, the first two parameters, final weight (g) and total length, were recorded for all fish. Among the 20 fish in each tank, 17 specimens were carefully dissected, and other parameters, such as liver weight and visceral weight, were also recorded. Moreover, from the 17 fish dissected, the gill, skin, dorsal muscle, whole intestine, foregut, midgut, and hindgut were sampled. Some were fixed in a solution of glutaraldehyde in a phosphate buffer (0.1 M, pH 7.0) for Hematoxylin & Eosin (H&E), and another sample was fixed in a formaldehyde solution and prepared for Scanning Electron Microscope (SEM) and Transmission Electron (TEM) microscope observation. For the dorsal muscle proximate composition analysis, a considerable amount, around 100 g, of the dorsal muscle was pulled from each of the 17 fish put together in plastic bags and immediately stored at −20 °C. The remaining three fish were not dissected but instead kept in the fridge at −20 °C, for the whole-body proximate composition analysis. 2.3. Method for the Proximate Composition of the Whole-Body and Dorsal MuscleMoisture, ash, crude proteins, and crude lipids were assessed following the method of the Association of Official Analytical Chemists [48]. Moisture was determined by drying ground samples in a forced-air oven at 105 °C for 24 h. Ash content was determined by incinerating samples at 600 °C for 24 h in a muffle furnace. Crude proteins were evaluated as Kjeldahl-nitrogen using a factor of 6.25, and crude lipids were analyzed by Soxhlet extraction with petroleum ether [49,50].2.4. Histological Studies: H&E, SEM & TEM, Sample PreparationThe histological studies achieved in the present work include three different sample preparations. The first preparation is the H&E stain with light microscope observation. The second and third treatments were done for SEM observation and TEM observation, respectively. SEM gives us a 3D image with more details. The TEM observation was done to observe the liver cell’s inner structure for any histological difference or organelles pathology, focusing on the mitochondria structure shape, cristae arrangement, and the abundance and size of peroxisomes to identify if the long-term exposure to a lower pH could cause metabolic acidosis.2.4.1. Sample Preparation H&E (Hematoxylin and Eosin) Stain and Light Microscope Observation of Gills, Liver, Skin, Foregut, Midgut, and HindgutAfter the fish dissection, gill, liver, skin, foregut/duodenum, midgut/jejunum, and hindgut/ileum samples were collected. Each sample was fixed immediately after dissection in 10% formalin for 24 h. Later, in the laboratory, the samples were decalcified in 10% nitric acid, dehydrated in increasing alcohol concentrations, cleared in xylene, and impregnated and embedded in paraffin. Thin sample sections were stained with hematoxylin and eosin for histological description according to the standard method described by Munro [51] and used by Titford [52]. The gill, liver, skin, foregut midgut, and hindgut tissues were stained with H&E, put on small rectangle glass slides, and observed with the light microscope (Axiocam 506 color, ZEISS) [53,54,55]. All images were obtained in high-definition.2.4.2. Foregut, Midgut, and Hindgut Preparation for SEM ObservationsFor the first-step treatment called double fixation, the foregut was first fixed with 2.5% glutaraldehyde in a phosphate buffer (0.1 M, pH 7.0) for more than 4 h; then, it was washed three times in the phosphate buffer (0.1 M, pH 7.0) for 15 min at each step; then, it was post-fixed with 1% OsO4 in a phosphate buffer for 2 h and washed three times in a phosphate buffer (0.1 M, pH 7.0) for 15 min at each step. Then, the second step was a two-stage dehydration. In the first stage, it was dehydrated by a graded series of ethanol (30%, 50%, 70%, 80%, 90%, and 95%) for about 15 min at each step; then, it was dehydrated two times by alcohol for 20 min at each step, or stored in alcohol. In the second stage, the sample was dehydrated in a Hitachi Model HCP-2 critical point dryer. On the third and last step, the sample was coated with gold-palladium in a Hitachi Model E-1010 ion sputter for 5 min and observed in the Hitachi Model SU-8010 SEM. The image obtained here is a 3D image.2.4.3. Liver Sample Treatment and TEM ObservationAfter repeating the double fixation and the first stage dehydration used for SEM mentioned above, we did the infiltration process. Upon infiltration, the liver sample was placed in a 1:1 mixture of absolute acetone, and the final Spurr in a resin mixture for 1 h at room temperature, then transferred to a 1:3 mixture of absolute acetone; to the final resin mixture for 3 h, and finally to the Spurr resin mixture overnight. The sample was placed in Eppendorf-containing Spurr resin and heated at 70 °C for more than 9 h and was finally sectioned in a LEICA EM UC7 ultra-tome. Sections were stained by uranyl acetate and alkaline lead citrate for 5–10 min each and observed in a Hitachi Model H-7650 TEM. The Figure obtained here is a 2D view. 2.5. Statistical AnalysisStatistical analyses were performed using the Software IMB SPSS Statistics 23.0.0.0. All data were tested for normality and homogeneity of variances by Kolmogorov–Smirnov and Levene’s tests, respectively. The data were subjected to a one-way analysis of variance (ANOVA), and a Tukey’s HSD test was used to compare significant differences between means at (p < 0.05) (n = 3 replicates). All the quantitative data are presented as mean ± SD (standard deviation).3. Results3.1. Seawater Physicochemical ParametersThe results of the measured seawater physicochemical parameters are presented in the Table 1.3.2. Growth ParametersThe following Table 2 display the fish growth performance parameters after 8 weeks of growth trial, fed with the same diet.Despite stocking juveniles with the same initial weight and feeding them with the same diet throughout the experimental period, the results revealed that the three treatments had varying growth parameters with respect to pH. Here, the data demonstrate that fish growth, fish weight, and fish length significantly decreased with decreasing pH. Regarding fish survival, there was no significant difference with p > 0.05 for all treatments. For the feed intake (FI) analysis, we observed that the fish in the Predict_A group had the highest values. However, while the feed intake increased from the Control to Predict groups, the decrease of the specific growth rate (SGR) was surprising. Statistically, there was no significant difference in the hepatosomatic index (HSI) or the condition factor (CF). Finally, considering the protein efficiency ratio (PER), a significant decrease was observed from the control group to the two predict groups. 3.3. Proximate Composition of the Fish SamplesTable 3 shows the proximate composition of the whole body and the dorsal muscle. The Control group (pH of 8.10), with superscript “a”, shows significant differences compared to the Predict_B group (pH of 7.40), with superscript “b”. The Predict_A group, with superscript “ab”, is a middle value belonging to both the Control and Predict_B groups. Predict_A’s crude protein value explains the decrease, though it is not significantly different (p > 0.05) from either the Control or Predict_B groups.Considering both the whole-body and dorsal muscle proximate composition, the moisture, crude lipid, and ash, having no superscript, indicate no significant difference between treatments. However, the crude protein content showed a decrease according to the lower pH. The decrease is significant between the Control group (pH of 8.10) and the Predict_B group (pH of 7.40). However, though the decrease is not significant from the Control to the Predict_A group, we can see from the superscript “ab” that this value can be classified to the significance of both the Control (not significant, superscript “a”) and the Predict_B group (significant, superscript “b”). 3.4. SEM Observation of the Foregut TissueFigure 2 is a 3D capture, obtained from SEM, and the Control group (1-a, 1-b, 1-c) displays a very regular tissue distribution, with almost no microvilli (air-like structure) atrophies over the brush border, suggesting that the level of seawater pH of 8.10 has a negligible effect on the intestine tissue. However, Predict_A (2-a, 2-b, 2-c) and Predict_B (3-a, 3-b, 3-c) showed multiple atrophies (black arrows) of the microvilli all along the brush border, signifying that seawater pH has a significant effect on the tissue. The Control group shows a regular structure of brush border without microvilli atrophy (1-c). In Predict_A (2-b, 2-c), and Predict_B (3-b, 3-c), the black arrows indicate multiple atrophies of the microvilli over the brush border.3.5. H&E Stain: Observation of Foregut, Midgut, and HindgutIn the Control group (Figure 3(1)), the villi’s brush border (red ellipse) has a regular plan surface, signifying a normal tissue distribution with almost no atrophies of the microvilli all along the surface. On the contrary, in the Predict_A and Predict_B groups (Figure 3(2),(3)), the surface of the villi is sawtooth-like, emphasizing the atrophy of the microvilli due to a lower pH or a higher acidification. A graphic of what we mean by regular surface and sawtooth-like surface has been presented (Figure 3 and Figure 4).3.6. H&E Stain: Light Microscope Observation of the Liver In the Control group (Figure 4(1-a–c)) with normal seawater pH of 8.10 ± 0.01, a healthy liver’s characteristics are observed. The liver’s general structure is standard, with almost no blood congestion (B.c.) in all hepatopancreas (1-a, 1-c). There is no trace of cell necrosis in hepatocytes and no blood congestion inside the hepatopancreas. Moreover, Kupffer cells that could indicate infection are hardly found. In this Control group, only the hepatocytes’ vacuolation (V.H.) is seen in 1-b, in the black circle. In the Predict_A group with a pH of 7.80 ± 0.02, we also observe the hepatocytes’ vacuolation (V.H.) (2-b). Though we have no blood congestion in the hepatopancreas (2-c), we can identify blood congestion inside the blood vessel (2-d). At last, we identify the cell necrosis (2-e). Concerning the Predict_B group with a pH of 7.40 ± 0.02, we can observe a bloody liver with a hemorrhage emphasized with blood coloration (3-a to 3-e). It can be observed (3-a) that all hepatopancreas have blood congestion, magnified in (3-c). Two areas of hemorrhage (Hr) can be noticed (3-b). Blood congestion is also present in the central vein (3-d). The cell necrosis (N) can also be identified, with an emphasis on the abundance of Kupffer cells (K) (3-e).3.7. H&E Stain: Light Microscope Observation of SkinIn the three treatments, there was no histological atrophy related to the pH influence in the skin tissue (Figure 5). The main parts of the skin are identifiable. In the stratum spongiosum (SP) and stratum compactum (SC) of the dermis, we observed a normal skin with a well-defined stratified squamous epithelium with scattered mucous (MC), alarm cells (AC), epithelial cells (EC), and fibroblasts (F). Each of them has a standard structure distribution without any atrophies.3.8. H&E Stain: Light Microscope Observation of the GillThe phenomenon of hemorrhage in the gill’s filament appears in the Predict_A group (Figure 6(2-b)) and also in the Predict_B group (3-b). The blood congestion (B.c.) is observed at the inner structure of the lamellae in Predict_A (2-a). The lamellae inflammation called Aneurysm (A) is present in the Predict_A (2-c) and Predict_B groups (3-c). All three groups showed the presence of a parasitic cyst (PC) (1-b, 2-b, 3-d) and lamellae fusion (Lf) (1-c, 1-d, 2-d, 3-a). 3.9. TEM Observation of Liver Cells (Inner Structure)The first anomaly noticeable was a difference in the nucleus (N) shape (Figure 7(1-b) vs. Figure 7(2-b)/Figure 7(3-b)). The conventional round shape of the nucleus is observed in the Control group (1-b), while the Predict_A and Predict_B groups, on the contrary, present a crenated nucleus (2-b, 3-b). The absence of nucleolus is noticeable (3-b). Interestingly, Predict_B displays a greater abundance of peroxisomes (p) (3-c), which are hardly noticeable in the two other groups. Furthermore, a difference in the shape and size of the mitochondria is noticeable among the three groups; most of the mitochondria in the Control group have a conventional round shape, denoted with a single arrowhead (1-c). On the contrary, samples from Predict_A display some mitochondria with a very elongated structure, identified with a double head arrow (2-c). Furthermore, the samples from Predict_B also show mitochondria with an elongated and angular shape, also indicated by a double head arrow (3-d).4. DiscussionOur results illustrate the significant negative effect of OA on the health, growth, and meat quality of the black sea bream (Acanthopagrus schlegelii) from the end of the 21st century. If we consider the single effect of ocean acidification, the fish are susceptible to survive up to that period. However, as the phenomenon of acidification is always coupled with warming, it is reported that a threat to the survival of black sea bream is expected, as well as to the growth health and the meat quality of this species by the end of 2100 and beyond [20]. 4.1. Water Physicochemical Parameters The projected scenario represents a very significant decrease in pH, as can be seen by our measured pH. Consequently, there will be a significant reduction of the concentration of carbonate ions in the water [CO32−], and also a significant decrease of the state of aragonite and calcite (Ωara and Ωcal). The more significant the decrease in the concentration of carbonate ions [CO32−], the more significant the increase in the concentration of dissolved inorganic carbon in the seawater.In case the projected ocean pH values effectively reach 7.80 and 7.40 by 2100 and beyond, as projected by IPCC [1], the present study has provided in Table 1 an overview of the expected seawater physicochemical parameters. From the measured parameters (S, T°, pH, TA), we were able to calculate others (pCO2, HCO3−, CO32−, DIC, Ωara, Ωcal) using the software CO2cal 1.2.0. From the equation pH =-log [H+] [12,36], the acidity concentration can be deduced as [H+] = 10(−pH). This means that the projected pH values 8.10, 7.80, and 7.40 give acidity values of 7.94 × 10−9, 15.85 × 10−9, and 39.81 × 10−9 in Control, Predict_A, and Predict_B, respectively. Compared to the Control group, Predict_A and Predict_B represent a 99.53% and a 401.19% increase in acidity, respectively, which is highly significant.4.2. Proximate Composition Showing a Reduction in Meat QualityExcept for crude proteins, there was no significant difference in moisture, crude lipids, and ash content among the treatments (Table 3). For both the whole body and the dorsal muscle, a significant decrease was observed in the crude protein content from the Control to the Predict_B groups. This significant decrease implies a reduction in the meat’s nutritional quality. Nevertheless, the decrease was not significant from the Control to the Predict_A groups, but we can see from the superscript “ab” that this value can be classified according to the significance levels of both the Control group (not significant) and the Predict_B group (significant). Another parameter is therefore necessary to confirm which conclusion should be considered. The PER values in the growth parameters shown in Table 2 confirmed this reduction in meat quality, because we see the same significant decrease of protein content from the Control to the Predict_B groups. However, in contrast to the case of crude proteins, the PER displayed a significant decrease from the Control to Predict_A groups. The best approach to explain this slight difference is to calculate the PER and evaluate the crude protein content. There formula used to calculate the PER—PER = Weight Gain (g)/Protein intake on a dry basis (g)—is very precise. The crude protein content is evaluated through chemistry titration, following the international reference Kjeldahl method [56], which is not as precise as the PER value calculation. It is therefore possible that during the three replicates evaluation of the crude protein content, the resulting standard deviation was larger and, consequently, the significance decreased. At least the decreasing trend is observed in both the PER and the crude protein, which are both significant in the Predict_B group. When scientifically studying the risks, it is always better to consider the worst case. Consequently, there may be the risk of a significant decrease in the meat quality of black sea bream from the end of the 21st century. Studying the impact of OA on mussels, Martin et al. [57] reported a negative effect on meat quality. Furthermore, the studies of Rossoll et al. [58] and Jin et al. [59] also reported a decrease in fish samples’ meat quality due to OA. A possible reason is that the atrophy of the microvilli in the small intestine will reduce the amount of nutrients absorbed. This explains why, in the Predict_A and Predict_B groups, the microvilli atrophy in the duodenum jejunum and ileum decreases the nutrient absorption, affecting the crude protein gain and resulting in slow growth, lower weight gained, lower total length, and, consequently, a reduction in fish meat quality.4.3. Growth Parameters Revealing a Significant Low Growth and Confirming the Reduction in Meat QualityThere was no significant difference in the Hepatosomatic index (HSI), and the Condition factor (CF) indicated the same feeding intensity. Although the fish in all the treatments were fed with the same feeding intensity, the data revealed a low growth rate in Predict_A and Predict_B, despite the higher feed intake (FI) in these Predict groups than in the Control group. In simple terms, the fish ate more than normal but still grew lesser than they should, which was observed earlier by Zhang et al. [60] while studying subtidal scavengers (Nassarius conoidalis). The growth parameter data in Table 2, which include the final length (FL), the weight gained (WG), and the specific growth rate (SGR), show a significant decrease (p < 0.05). This indicates that a decrease in seawater pH led to a significant decrease in fish growth—the more the seawater decreases the more the fish growth is affected. The significant decrease of the feed efficiency (FE) and the protein efficiency ratio (PER) for the Control group to the Predict_A and Predict_B groups indicate that the feed consumed is less efficient in the two prediction groups than in the Control group. The possible reason may be the lower pH, causing an atrophy of the microvilli that will lead to less nutrient absorption in the small intestine. Inversely to the significant decrease of FE, we see a significant increase in the feed conversion ratio (FCR) in the Control group vs. the Predict_A and Predict_B groups, because the formula of the FE is inversely proportional to that of the FCR (see the FE and FCR formula under Table 2). The FCR is defined as the amount of feed necessary to produce a unit of animal weight. This indicates that more feed is necessary to produce a unit of animal weight in the Predict_A and Predict_B groups than in the Control group. This latter statement can also be explained by the loss of nutrients through the small intestine. Research studies on Sparus aurata and Nassarius conoidalis published by Araujo et al. [20] and Zhang et al. [60], respectively, highlighted that OA decreased fish survival, boosted the energy demand, and reduced the physiology and scope for growth. 4.4. The Survival Rate and the Histology of Liver and Gill Tissue Showing Degrading Effects on Fish HealthThe results on survival and the histological studies of liver and gills tissues have revealed a considerable impact on fish health. The survival rate (SR) had no significant difference (p > 0.05) among treatments, suggesting that considering the only factor of pH decrease in the 21st and 23rd centuries, the fish may be able to survive. Nevertheless, it has been reported that if we add temperature as a second factor, the fish survival will be affected from the 21st century and well beyond [20]. Concerning the fish health evaluation, we noticed the damage of the fish liver in Predict_A and much more in the Predict_B group. Regarding the liver tissue observation after the H&E stain shown (Figure 4), the light microscope revealed necrosis, blood congestion, vacuolation of the hepatocytes, hemorrhage, and an increased number of Kupffer cells on Predict_A and Predict_B. The anomalies mentioned above are signs of liver damage, leading to the perturbation of the fish liver’s optimal function [17,61].Gill histology showed a slight difference among treatments where signs of blood congestion were observed in the Predict_A (2-b, 2-d) and Predict_B groups (3-d) shown in Figure 6. Except for the blood congestion, each of the three groups showed the same anomalies: the absence of lamellae, lamellae fusion, aneurysm, and parasitic cyst. Therefore, they cannot be considered an effect of low pH [62,63,64].4.5. Histology Showing No Atrophy on Fish Skin Tissue Interestingly, no observable effect of acidification was noticed on the fish skin (Figure 5). This suggests that OA’s future prediction at the end of the 21st century and beyond will have minimal or no effect on black sea bream skin. This is possible because the ocean pH would still be within the acceptable range for fish skin to endure [65,66]. 4.6. TEM Observation of Fish Liver Cell’s Inner Structure, Revealing Signs of Metabolic AcidosisThe TEM observation revealed an elongated and angular shape of mitochondria in the Predict_B group, a condition known as lactic acidosis manifestation [67]. Lactic acidosis is a form of metabolic acidosis. The abundance of peroxisomes was also noticed in this 3rd group, with the highest acidification value [67]. This result suggests that through fish respiration, when the blood gains oxygen, CO2 also gets into the blood through the gill and comes into direct contact with the acidic seawater [36,68,69]. This result suggests that when the blood gains dissolved oxygen during respiration, the dissolved CO2 also gets into the blood through the gill because of its direct contact with the acidic seawater [36,68,69]. Not being able to regulate the excess of dissolved CO2 in the blood, the fish will experience a decrease in the blood pH value, causing acidosis. This inefficiency of the acid–base regulation is due to the long-term exposure to the acidic environment. [70] demonstrated earlier that a value of pCO2 = 1000 µatm could lead to metabolic acidosis. Since the present study recorded a high pCO2 value of 1993.71 ± 102.12 µatm in Predict_B, metabolic acidosis was expected.4.7. Histological Studies of Small Intestine Revealing Microvilli AtrophyThe small intestine’s wall contains many different anatomical structures, functioning either to provide a defense for the small intestine tissues or to absorb nutrients from food. The microvilli are the small, hair-like projections that extend outward from the small intestine wall [71], whose function is to absorb nutrients from food [72]. Our results showed that both the light microscope and the SEM observation of samples of the small intestine in the Control, Predict_A, and Predict_B groups (Figure 2, Figure 3) showed the same results. The 3D pictures (Figure 2) provide details of (Figure 3) the black arrows shown (Figure 2), obtained after SEM, and explain that the sawtooth-like structure observed (Figure 3) is actually due to multiple atrophies of the microvilli throughout the brush border of the duodenum. These atrophies suggest that the number of microvilli responsible for absorption is significantly reduced, resulting in nutrient loss. Though the image presented (Figure 2 and Figure 3) is just the small intestine’s foregut, the analysis and observation were also conducted on the midgut and hindgut. Similar results regarding the atrophy of the microvilli were obtained. To the best of our knowledge, no previous study has reported atrophies of the microvilli on the small intestine under the effect of OA, thus making our result a novel discovery. However, a broader look revealed that the microvilli atrophy observed (Figure 2 and Figure 3) is similar to that of a well-known human disease called microvilli inclusion disease (MVID). In his publication, Sidhaye, et al. [73] discovered that zebrafish exhibited cellular attributes of human microvillus inclusion disease. The MVID investigated by researchers revealed microvilli atrophy, metabolic acidosis, and diarrhea [71,74,75,76,77,78,79,80,81]. Likewise, in our experiment, samples from the Predict_A and Predict_B groups revealed the same anomalies of microvilli atrophy, signs of metabolic acidosis, and nutrient loss. This nutrient loss can be considered as “diarrhea” in humans. [78] stated that many patients with MVID experience liver dysfunction, supporting our discussion in Section 4.4. 5. ConclusionsThis research found that pH values of 7.80 at the end of the 21st century and 7.40 in the 23rd century would significantly decrease the crude protein content of black sea bream and the protein efficiency ratio. Under the projected OA pH, the growth performance data revealed that black sea bream eat more than normal but grow less than normal. Histological studies revealed that OA had caused multiple atrophies on small intestine microvilli, causing a significant decrease in nutrient absorption, resulting in a lower weight gain and a lower specific growth rate. The observation of samples on a scanning electron microscope, a transmission electron microscope, and an optical microscope demonstrated liver and gill anomalies, suggesting a negative effect on the health of black sea bream. Given that black sea bream is one of the most resistant fish to disease and environmental change, it can be deduced that much more varieties of fish—such as yellowtail (Seriola quinqueradiata), flounder (Paralichthys olivaceus), purplish amberjack (Seriola purpurascens), red sea bream (Pagrus major), puffer (Takifugu rubripes), striped jack (Pseudocaranx dentex), sea perch (Lateolabrax spp.), goldstriped amberjack (Seriola aureovittata), Schlegel’s black rockfish (Sebastes schlegelii), striped beak perch (Oplegnathus fasciatus), or dark-banded rockfish (Sebastes inermis), just to cite these, all having a lower resistance—may be affected [41]. Consequently, it can be presumed that if mitigation measures are not taken to curb the emission of CO2 in the atmosphere, OA is expected to cause a significant degrading effect on fish health, fish growth, and fish meat quality. Fish survival could also be threatened beyond 2100 if we consider the effect of both pH and temperature. The impact on fish health and, possibly, on the survival rate will affect the fish population, leading to a decline in marine resources. Given the likely risks of future OA in black sea bream, and the even greater risks in more sensitive species, CO2 mitigation is needed. For example, we highly recommend and encourage the use of green energy, via electric vehicles or the planting of trees, especially in the most polluted environments, to absorb the excess CO2 released into the atmosphere. Furthermore, this study helps aquaculture practices by demonstrating the implications of low pH on fish health, survival, meat quality, and feeding efficiency. Further analyses are being carried out on this same experiment in the current year (2021), one on brain transcriptomics and another on fish behavior assessment by computer vision.

animals : an open access journal from mdpi

10.3390/ani11113122

PMC8614348

The effects of grazing dairy cows in mountainous areas for 12 and 24 h compared with keeping the cows under confined settings were evaluated by examining the overall levels of fatty acids and cortisol in milk as an indicator of stress. Our findings revealed favorable changes in cows grazing for 12 h, with significantly improved milk fatty acid profile and decreased milk cortisol content. Additionally, we found no significant changes in the milk yield of 12 h grazed cows compared to the control group of confined cows, while milk fat and protein showed higher values in 12 h grazed cows. Grazing for 24 h caused a significant decrease in the milk yield compared to the two other groups. Overall, grazing for 12 h in a mountainous area is suggested to be beneficial for animal welfare, with positive improvements in milk fatty acids and lessening of stress levels without adverse effects on milk yield.

The effects of grazing lactating cows in mountainous areas for 12 and 24 h compared with the confined indoor system were evaluated by examining the overall milk fatty acid and cortisol. Twenty-one dairy cows were allocated to three treatment groups: (1) control (confined management system in a free-stall barn; TMR based), (2) grazing for 12 h (12 hG; TMR plus grazing pasture), and (3) grazing for 24 h (24 hG; pasture-based feeding system). Dry matter intake was higher in the control and 12 hG groups than in the 24 hG group. The yields of total milk and the 3.5% fat-corrected milk were the lowest in the 24 hG group. Milk fat was the highest in the 24 hG group and higher in 12 hG compared with the control group. Milk protein and lactose levels were the highest in the 12 hG group. The highest somatic cell count was observed in the 24 hG group. The saturated fatty acid levels were higher in the control group compared with the 12 hG and 24 hG groups. There was no difference in overall mono-unsaturated fatty acids between 12 hG and 24 hG groups. Poly-unsaturated fatty acids were higher in the 12 hG group compared with the control and 24 hG groups. There was no difference in omega-6 (ω-6) fatty acids among the groups, and omega-3 fatty acids were higher in the 12 hG group than in the control group. Milk cortisol was the highest in the 24 hG group and higher in the control group compared with the 12 hG group. Taken together, grazing for 12 h is advisable for farms that have access to mountainous areas to improve the milk fatty acid profile and decrease the stress levels in high-yielding Holstein lactating cows.

1. IntroductionGrazing is an important management practice in terms of, for example, the nutrition of fresh forage intake, animal health [1], animal well-being, natural behavior [2], landscape values, and grassland biodiversity [1,3,4]. Grazing dairy cows has been practiced for a long time as a strategy for improving animal welfare and milk quality [5], lessening oxidative stress [2], and economic purposes [6,7].One indicator to measure animal welfare is cortisol concentrations in different matrices, including milk. Measuring cortisol in milk is becoming a promising indicator for evaluating the daily or short-term stress conditions and well-being of dairy animals [8,9] and humans [10]. Measuring plasma cortisol as a primary stress indicator is problematic because the animal needs to be restrained for invasive blood collection [11,12], and measuring cortisol in milk has been suggested as alternative due to its accessibility regardless of the management system together with the ease of collecting samples in a manner that does not cause stress [9]. Therefore, to support the hypothesis on the potential of grazing animals in decreasing stress, particularly in dairy cows [7], we further tested the hypothesis of whether grazing cows have decreased levels of milk cortisol compared with indoor animals in this study.Korea is one of the countries that need to improve land use efficiency because it has only 16% arable land [7] that can potentially be used for agricultural and livestock purposes using smart farming. It should be noted that Korea imports 75% of its compound feed and 96.4% of its feed crops, which has become a matter of concern among Korean livestock industry participants and the Korean government [12,13]. In addition to its economic advantages (free pasture, decreasing cost of human resources, etc.), grazing pasture allowance can reduce the negative effects of high stock density faced by farms of the cow industry in Korea. Given that 70–80% of Korea is mountainous [7], the application of smart farming is necessary for using the accessible land beside livestock farms with controlled grazing using electric fences for containment.Many factors influence the compositions of milk including but not limited to the nutritional values of feed, nutrition level, feeding management, stress source and severity, genetic, environment, season, stage of lactation, reproductive and productive status of the animals, age, and BW of the animals, etc. Improving milk quality has the dual purpose of increasing both the nutritional value and the commercial value of the milk (i.e., low worker and feed costs). In this regard, increasing the levels of poly-unsaturated fatty acids (PUFAs), particularly omega-3 PUFA, in milk is desirable in dairy animals [1,6]. Although previous studies have investigated the effects of grazing on the welfare [4,14,15] and milk of dairy animals [16,17], to what extent grazing can positively impact milk characteristics without negatively influencing production performance in high-yielding lactating cows needs further exploration. Therefore, this study was designed to evaluate the effect of allowing daytime grazing for 12 h and full-time grazing of 24 h compared with the indoor rearing of animals on the milk yield and characteristics, such as milk unsaturated fatty acid and cortisol contents, in high-yielding lactating cows.2. Materials and MethodsAll experimental procedures involving animals were approved by the Animal Welfare and Ethics Committee of Kangwon National University, Chuncheon, Republic of Korea (KIACUC-16–0098). This study was carried out in spring at the Sky Ranch farm (Daegwallyeong, Pyeongchang, Gangwon Province, Korea) 920 m above sea level.2.1. Experimental Animals and Treatment ApplicationTwenty-one dairy cows (age = 36 ± 4.5 months) were used in this study. The specifications of BW, milk yield, and parity of the experimental animals are presented in Table 1.The control group (indoor rearing) was fed total mixed ration (TMR) plus additional commercial concentrate feed during milking. The maintenance system was indoor (confined management system). The TMR was consisted of forage (alfalfa hay, timothy hay, tall fescue straw, oat straw), silage (reed canary grass silage), concentrate feed (commercial concentrate), ad vitamin-mineral mix plus protected fat. The ratio of forage to concentrate was 42:58. The indoor facility was equipped with free stalls bedded with river dry sand. The pasture encompassed a mixture of timothy and reed canary grass. The pastureland was cultivated in 2014 and planted by timothy grass for the purpose of using the land for grazing cows. Electric fences were used to divide paddocks, and grazing was allowed when the grass was approximately 30 cm tall. TMR samples were collected for composition analysis at three times during the study: the beginning, middle, and end. The chemical compositions of the TMR and commercial concentrate feed that were used are provided in Table 2. Water was provided using a water trough in the barn for the control group and using connecting hoses in the pasture for the grazing cows.The grazing cows were divided into two groups: the 12 h grazing (12 hG) and 24 h grazing (24 hG) groups. The 12 hG group was not only fed TMR while indoors (18:00–06:00 h) but also allowed to graze during the daytime (06:00–18:00 h). The maintenance system for 12 hG group was half indoor (nighttime; the same as the control, and half (daytime) outdoor grazing, while the 24 hG group was outdoor grazing except the time of milking. A rotational grazing system was applied using four different areas adjacent to the farm each delineated by electric fences. Pasture sampling was conducted biweekly (total 4 times) to analyze chemical compositions. The experiment lasted 5 weeks (mid-late spring; 15 May, 22 June), of which the first week was used for grazing adaptation and conduction and the remaining 4 weeks for the main experiment. For adaptation, grazing animals were grazed daily for 3 h, increasing the time of grazing to meet the grouping criteria. The chemical compositions of the pastures used by the 12 hG and 24 hG groups are illustrated in Table 3.Before starting the actual experiment, the animals underwent 7 days of training to graze in the pastures after morning milking. For the study of feed intake, the amount of grass before and after grazing was investigated on weekly basis and the difference was calculated as the intake.2.2. AnalysesFeed analyses included chemical compositions of TMR (Table 2) and pastures (Table 3). The analyses included milk yield, milk composition (Table 4), milk FAs (Table 5), and milk cortisol (Figures 1–3).2.2.1. Feed Samples and AnalysisThe feed included TMR and pastures. A TMR sample (approximately 1 kg) was collected from the manger at the beginning, middle, and end of the experiment, and the pasture collection method [7] was performed four times (weekly, samples 1 to 4). The pasture sampling was conducted in 5 places in each paddock using 1 m × 1 m quadrat. The samples were stored for chemical analysis. The average chemical compositions of TMR samples are reported because there were no variations in the TMR ingredients. However, we decided to report the chemical composition of the pasture for each sampling time separately as the contents of pasture may possibly differ at each time point. The pasture and TMR samples were dried at 65 °C for 72 h, milled, and stored before further analysis. The chemical compositions of TMR and pastures were measured using the method of the AOAC (1990).2.2.2. Milk Collection and AnalysesThe experiment lasted 5 weeks, and all the cows in this study (control and grazing groups) were milked twice a day, at 06:00 and 18:00 h, using robotic milking machines (VMSTM V300; DeLaval International AB, Tumba, Sweden). Milk samples for component and cortisol analyses were collected at the beginning (15–17 May; for homogenized grouping purpose only prior to the start of the experiment), middle (third week, 5–7 June), and end (fifth week, 19–21 June) of the experiment. In doing so, and to obtain the highest possible accuracy, we collected the milk on 3 consecutive days at each time point to make sure that daily variations in milk components did not affect the outcome of the measures. We then pooled the morning and evening milk samples for all the consecutive days and made one sample for each animal to measure the milk composition, fatty acid profile, and cortisol content.2.2.3. Milk CompositionMorning and evening milk samples (15 mL) from each cow were collected in 45 mL sterile tubes (Wisd, Seoul, Korea) at each milking time for 3 consecutive days during the experimental period. The pooled composited milk samples were analyzed for concentrations of fat and protein, lactose percentages, and somatic cell count using a Foss 4000 milko Scan (Foss Electronic, Hilleroed, Denmark).2.2.4. Fatty Acid Profiles of Milk SamplesThe fatty acid profiles of the milk samples were analyzed using an Agilent Intelligent Gas Chromatography (GC) system (7890B FID GC, Agilent Technologies, Santa Clara, CA, USA). In brief, for fat extraction of milk samples, the sample was first homogenized. Then, a test solution was prepared to check for interference with undecanoic acid used as an internal standard. We accurately weighed the homogenized sample such that it contained about 100 to 200 mg of fat, placed it in a Mojonnier tube, added about 100 mg of pyrogallol, and then added 2 mL of internal standard solution. Then, we boiled the Mojonnier tube, added 2 mL of ethanol, and waited until the entire sample was thoroughly mixed. Afterward, we added 4 mL of water, stirred well, added 2 mL of ammonium hydroxide, stirred, and sealed the cap of the Mojonnier tube with a rubber band or Teflon tape for 10 min. We mixed the sample with a stirrer every 5 min so that the particles on the wall of the Mojonnier tube could be thoroughly mixed in. After sample decomposition, we added a few drops of phenolphthalein solution and then ammonium hydroxide to maintain the basic (pink) color. Finally, the lower part of the Mojonnier tube was filled with ethanol to allow for easy separation during ether extraction. Saturated fatty acids (SFAs), mono-unsaturated fatty acids (MUFAs), and poly-unsaturated (PUFAs) fatty acids were analyzed by calculating the sum of individual fatty acids and the sum of representative fatty acids.2.2.5. Milk Cortisol AnalysisThe same milk samples that were used for milk composition and fatty acid analyses were subsampled and used for cortisol analysis using the Bovine Cortisol ELISA kit (cat. no: MBS028594, MyBioSource, Inc. San Diego, CA, USA). After subsampling the milk collection, the samples were centrifugated at 1000× g (or 3000 rpm) for approximately 20 min. Then, the supernatant was separately and carefully collected from each sample. The assay was either conducted immediately or the sample stored at −20 or −80 °C for later analysis. Both the intra-assay coefficient of variation (CV = 7.9%) and the inter-assay CV (5.2%) were less than 15% (CV% = SD/mean × 100). All the CV% values were compared using concentration and not optical density (OD) values.2.3. Statistical AnalysisFeed intake and milk yield were analyzed by repeated-measures analysis using the MIXED procedure of SAS (ver. 9.01; SAS institute Inc. Cary, NC, USA), and the mean values were compared for significance using Tukey’s t-test at p < 0.05. The initial values of the days in milk, milk yield. body weight, and the parity were tested as covariate and since they were not significant at the beginning of the study (Table 1), we did not use them as covariates. We have tested several variance and covariance assumption structures including AR(1), UN, CS, ANTE(1), TOEPH, and ARH(1), and the covariance structure that resulted in the lowest values of the Akaike information criterion (AIC), herein AR(1), was chosen for the final analysis as it was the best fit to our design. The least-squares means of each group were calculated, and the differences between means were tested using the PDIFF option with Tukey’s adjustment. The data of non-repeated measures were analyzed using the GLM procedure of SAS. The mean comparisons between the groups were evaluated using Tukey’s test. Different levels of significance were declared as significant at p < 0.05, highly significant at p < 0.01, and very highly significant at p < 0.001.3. Results3.1. Dry Matter Intake, Milk Yield, and Milk CompositionThe dry matter (DM) intake was significantly higher (p < 0.05) in the control group and 12 hG groups compared with the 24 hG group, whereas no difference (p > 0.05) was observed between the control and 12 hG groups. The total milk and the 3.5% fat-corrected milk (FCM) yields were the lowest (p < 0.01) in the 24 hG group, while no difference (p > 0.05) was observed between the control and 12 hG groups (Table 4).Milk fat levels were the highest in the 24 hG group and higher in the 12 hG group compared with the control group (p < 0.001). Milk protein levels were higher in the 12 hG and control groups compared with the 24 hG group (p < 0.01). There was no significant difference in milk protein levels between the control and the 12 hG group (p > 0.05). Lactose levels were the highest in the 12 hG and higher in the control group compared with the 24 hG group (p < 0.05). The highest somatic cell count (SCC; p < 0.05) was observed in the 24 hG group, while the lowest (p < 0.05) SCC was in the control group (Table 4).3.2. Milk Fatty AcidsSaturated fatty acid (SFA) levels were higher (p < 0.05) in the control group compared with the 12 hG and 24 hG groups, while no difference between the 24 hG and 12 hG groups was observed (p > 0.05). There was no difference (p > 0.05) in mono-unsaturated fatty acid (MUFA) levels between the 12 hG and 24 hG groups, while MUFA levels were the lowest (p < 0.05) in the control group. Poly-unsaturated fatty acid (PUFA) levels were higher in the 12 hG group compared with the control and 24 hG groups, whereas no difference was observed between the 24 hG and control groups (p < 0.05). Total UFA levels were the highest in the 12 hG group, while they were higher in the 24 hG group compared with the control group (p < 0.05). The SFA/UFA ratio was the highest (p < 0.01) in the control group, while no significant difference (p > 0.05) was observed between the 12 hG and 24 hG groups (Table 5).Levels of omega-6 (ω-6) fatty acid, including γ-linolenic acid and arachidonic acid, were not significantly different (p > 0.05) among the groups, while linoleic acid levels were significantly higher (p < 0.05) in the 12 hG and 24 hG groups compared with the control group. α-Linolenic acid levels were the highest (p < 0.001) in the 12 hG group, whereas no significant difference (p > 0.05) was observed between the control and 24 hG groups. Eicosapentaenoic and docosahexaenoic acids levels were higher (p < 0.01) in the 12 hG group compared with the control group, while there was no difference (p > 0.05) between the 12 hG and 24 hG groups and between the control and 24 hG groups. Omega-3 (ω-3) fatty acid content was the highest in the 12 hG group, whereas no significant difference was observed between the control and 24 hG groups (Table 5).At the beginning of the study, there were no significant differences (p < 0.05) in milk cortisol levels among the groups (Figure 1). In the middle of the experiment, milk cortisol levels were the highest (p < 0.05) in the 24 hG group and higher (p < 0.05) in the control group compared with the 12 hG group (Figure 2). At the end of study, milk cortisol levels were lower (p < 0.05) in the 12 hG group compared with the control and 24 hG groups and lower (p < 0.05) in the control group compared with the 24 hG group (Figure 3).4. DiscussionIndependently of the type of grazing management (strip or rotational), when using high-quality pastures, the pasture use per unit area and pasture intake per cow are major factors determining milk production in grazing dairy cows, both being primarily controlled by pasture allowance [18,19,20]. Several studies have reported increased DMI due to the balanced protein and dietary energy of TMR, which increases the net energy supply and thus milk yield compared with the DMI in the grazing pasture system [21,22,23,24]. Consistent with the abovementioned studies, in this study, the decrease in the dry matter intake (DMI) in the 24 hG group could simply be explained by the fact that this group did not have access to TMR unlike the control (TMR based) and 12 hG (in addition to pasture) groups, which had access to TMR. The reason for no difference in the DMI between the control and 12 hG groups could be attributed to the access of the cows in the 12 hG group to TMR during the 12 h of non-grazing (indoor). The significant decrease in the DMI of the 24 hG group was reflected in the drastic decrease in the total milk yield and the 3.5% FCM compared with the other groups. The decrease in the milk production of the 24 hG group to approximately 20 kg/day indicates that the actual grazing grass intake did not meet the expected grazing grass intake and the energy–protein balance requirements needed for milk production. Méndez et al. [25] and Kolver and Muller [26] have stated that pastures, in general, have relatively lower DM and metabolizable energy (ME) content, resulting in lower DMI and milk production in high or high-medium dairy cows. Additionally, the growth rates of pastures and the variations in the environment could act as a limiting factor in meeting the DM and energy requirements for grazing cows [27]. Hence, the lower DMI and consequently lower milk yield in the 24 hG group observed in this study were arguably expected. Another reason for the lower DMI and consequently lower milk yield of dairy cows in the 24 hG group could be attributed to the higher content of NDF and ADF and lower CP in pastures (Table 3) compared to TMR (Table 2) provided for the other groups. It should be noted that all cows were homogenized when grouping and had no significant difference in milk yield at the beginning of the study (average 32.5 ± 1.35 kg/day). This result is corroborated by similar studies that have reported lower milk yield in the pasture-based feeding system compared with the confined indoor feeding system [21,26,28]. Overall, it is assumed that nutritional management, such as providing TMR for 24 h grazing, should be considered while maintaining grazing pastures to reduce factors that cause a sudden decrease in the production of grazing cows. The amount of grass intake fluctuates greatly and, thus, the management method must be improved. Consistent with these results in the 12 hG group compared with the control group, Di Grigoli et al. [2] demonstrated that pasture grazing and short daily grazing can cause comparable milk production compared to cows permanently kept indoors. A similar phenomenon was observed by Chapinal et al. [29] and Kennedy et al. [30] when comparing grazing cows with confined cows. Contrarily, some studies have reported higher milk yield in free-stalled or confined cows compared with grazing cows [31,32].Characteristics of the basal diet, including the TMR, concentrate allowance, forage type, forage-to-concentrate (F:C) ratio, fifiber content, and starch content, have a profound influence on fatty acid synthesis in the rumen and, consequently, on the milk fat level and fatty acid profile [1]. Higher milk fat in the 24 hG group compared with the other groups and higher milk fat in the 12 hG group compared with the control group could be attributed to the higher pasture intake having a higher F:C ratio in the 24 hG group compared with the 12 hG group and having access to a pasture resulting in higher fresh fiber intake in both 12 hG and 24 hG groups compared with the control group. The reason for higher milk fat in the 12 hG and 24 hG groups is speculated to be higher acetic acid production in the rumen of grazing cows due to higher fiber intake. Bath [33] revealed that ration with a higher fiber fraction induces the production of acetate as a main building block of milk fat. Thus, as the time of grazing increases, we expect higher milk fat, which is consistent with the obtained result. This result is also in line with a report [2] that showed higher fat content in the milk of daily grazed cows. Protein synthesis requires constituents of the protein synthesis machinery as well as the adequate amino acid availability and a large energy supply [34,35]. The high energy requirement of protein synthesis and turnover is evidenced not only by the reduction in protein synthesis and ion transport [34,35] but also by the decrease in overall protein synthesis as a consequence of caloric restriction [34,35]. The lower milk protein content in the 24 hG group compared with the 12 hG and control groups is likely attributed to the lower protein intake induced by the lower DMI of the corresponding cows on the pasture-based ration compared to those having the balanced protein energy intake of TMR (control) or TMR plus pasture (12 hG). In line with this result, there have been reports showing increased protein in the milk of cows under TMR feeding systems [36,37]. By contrast, Couvreur et al. [38] revealed an increase in milk protein with increased pasture grazing among cows [5]. Lactose, as the main carbohydrate in bovine milk, can be affected by udder health and the energy balance of the diet [39]. Lactose is basically synthesized from glucose that diffuses from blood into mammary epithelial cells. Thus, a ration with balanced energy–protein has a higher chance of producing more lactose compared with a ration with imbalanced energy–protein or is low in either energy or protein [39]. It can be postulated that in this study, the latter scenario was the case for the 24 hG group compared with the other groups. Fox et al. [40] demonstrated that the amount of water diffusing into the alveoli is determined by lactose [39], and as soon as lactose is synthesized, it is packed into secretory vesicles. In fact, the uptake of the lactose precursor (glucose) directly affects milk synthesis [41]. Having said that, the higher milk lactose in the control and 12 hG groups compared with the 24 hG group could be explained by the abovementioned reason. Moreover, Antanaitis et al., [42] have reported that the increase in milk lactose has a strong negative linear relationship with SCC. They underlined the importance of milk lactose as an indicator of subclinical mastitis showing that the increase in lactose levels in cow’s milk was most closely associated with a decrease in the prevalence of subclinical mastitis pathogens such as S. agalactiae, S. aureus and other Streptococci. Given this finding, our result indicated that cows in 24 hG group had higher SCC coinciding with lower milk lactose (and observation of mastitis in this group) compared with the other groups. The relationship between subclinical mastitis, lactose, and grazing pasture allowance should be further studied. However, O’Callaghan et al. [5] did not find any difference in the lactose concentrations of milk from groups under various feeding systems. They found lower milk lactose levels in cows during late lactation and reported higher milk lactose levels during early lactation. The SCC in grazing cows was significantly higher compared with the control group. The higher SCC in grazing cows could be due to mastitis. Another reason for a higher SCC in the 24 hG group could be the wet pasture on many of the days the cows rested in the area. It was observed that the udder area of cows in the 24 hG group was dirtier compared to the control and 12 hG groups, which could explain the higher SCC and incidence of mastitis in the 12 hG group compared with the control group. These results are in contrast with the report by Di Grigoli et al. [2], who found a lower SCC in the milk of grazing cows compared with cows kept in free-stall barns [28,31]. It is worth noting that their cows grazed for a short time (5 h/day). In this regard, the management quality of grazing cows and the environment (i.e., rain) during grazing could be influential factors that need to be considered when interpreting the results of different studies. Another reason for the lower SCC in short-term grazing cows in the abovementioned studies was the lower incidence of subclinical mastitis and greater hygiene of the udder. In this study, the higher SCC in grazing groups was considered a negative effect of grazing compared with the confined indoor system. However, can be prevented through higher quality management.Milk fatty acids originate from two sources: (a) blood lipids (approximately 60%) intermediated by non-esterified fatty acid (NEFA) or triglyceride-rich lipoproteins and (b) de novo (C4–C16) synthesis from acetate and butyrate in the mammary glands [43,44]. Di Grigoli et al. [2] indicated the beneficial effect of grazing on the quality of milk, by inducing unsaturated fatty acids (FA), including ω-3 fatty acids [2,45], which eventually improve the health of humans as the final consumer. O’Callaghan et al. [5], in corroboration with Baltušnikienė et al. [46], revealed that pasture-fed cows produce lower saturated fatty acids (SFAs) compared with cows fed TMR, which supports the findings of our study. Lower SFA levels would be more beneficial for human health because of the potential adverse effects of higher SFA levels for humans, such as cardiac issues [47,48]. The health benefits of milk with higher unsaturated fatty acids (UFAs), including mono- and poly-UFAs, have been previously discussed and established [2,5]. A PUFA is a biological membrane structural material, and as a precursor of prostaglandin, it is involved in maintaining homeostasis of the living body, such as through its roles in blood coagulation and various allergic reactions. In line with a previous study [2], in this study, we found that the feeding system influences milk fatty acid levels. By contrast, Di Grigoli et al. [2] reported higher SFA levels in cows grazed for a short time (5 h/day). They speculated that the short time of daily grazing could be a reason for this phenomenon, in which grazing cows are unable to ingest a sufficient mass of green pasture. Consequently, the SFA/UFA ratio decreased in grazing groups compared with the control group due to lower SFA and higher UFA levels in the corresponding groups. Elgersma et al. [1] explained the mechanism behind the higher UFA (MUFA and PUFA and total UFA) levels in grazing grass-fed cows compared with TMR-fed cows that resulted in higher SFA levels in milk fat. Elgersma et al. [1] documented that pasture allowance can induce the production of α-linolenic acid in milk fat. Consistently, we found higher α-linolenic acid levels in the 12 hG group, whereas the reason for the lower α-linolenic acid levels in the 24 hG group remained unclear. Grassland species diversity may improve PUFA transfer efficiency from feed to milk [1]. Because grazed grass is often younger than the forage ingredients of TMR and also because grazing animals prefer leaf blades to stems [1], these fifindings imply that the fatty acid intake and the proportional intake of α-linolenic acid is higher for cows in grazing pastures compared with confined cows. Green pastures are particularly rich in α-linolenic acid. A higher amount of α-linolenic acid in the milk can therefore be attributed to the higher intake of pasture in grazing cows [1], herein the 12 hG group. The lower ω-3 contents of milk in the 24 hG group could be explained by the lower intake of the group, as discussed earlier, and the imbalanced nutritional value of the intake. Omega-6 encompasses linoleic acid, γ-linolenic acid, and arachidonic acid, whereas omega-3 derivatives consist of α-linolenic acid, eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA). EPA and DHA have several health benefits for humans and, thus, increasing their levels and that of α-linolenic acid is desirable in the milk of dairy cows, resulting in healthier consumers. Moreover, the ratio of ω-3 to ω-6 FA in dietary products is as important as the dietary proportions of SFAs, MUFAs, PUFAs, and total fat [49]. Similar to this study, Di Grigoli et al. [2] also found higher concentrations of EPA and DHA in the milk of grazing cows compared to the free-stall building system. Lower total ω-3 in the 24 hG group compared with the 12 hG group could be attributed to the important ω-3 components, α-linolenic acid, and lower numerical values of EPA and DHA in the corresponding group. Some of the fatty acid results reported in this study are not in line with those of previous studies when comparing the confined management feeding system and the pasture-based grazing system. Some of the factors that could influence the profiles and values of the fatty acids and explain these discrepancies are (1) time allowance for grazing; (2) quality and maturity of the pasture and TMR quality; (3) duration of grazing; (4) physiological status of grazing cows (lactation phase, parity, DIM, etc.); (5) milking interval (two or three times a day) and method (automatic, semi-automatic); (6) animal breed; (7) farm management quality, including the use of smart farming systems and water-providing systems; and (8) environmental and seasonal factors. All these factors should be considered when interpreting the results.Cortisol is a primary hormone that is released into the blood not only in response to stress stimuli [12,50] but also in response to decreased feed intake and impaired animal behavior [51,52]. The process includes hypothalamic–pituitary–adrenocortical axis (HPA) activation, which is paramount to the physiological endocrine response of the cows to stress. Any change in blood cortisol levels can be reflected by a similar pattern of change in milk since cortisol diffuses into the milk via blood. Milk sample collection is also compatible with animal welfare protocols and would thus be a good approach to measure stress levels [53]. Thus, among the various stress matrices and because of the limitations associated with blood collection (mainly invasive sample collection), milk is considered a preferential matrix in dairy cows to indicate short-term stimulation of the HPA axis, due to the ease of sample collection and the non-stressing sampling procedure. In this study, we measured milk cortisol levels prior to grouping the cows to homogenize the groups having the same average milk cortisol levels. This helped us to use these data not only for pre-test evaluation but also as the baseline range. Data from the middle and end of the experiment regarding milk cortisol levels supported the hypothesis and revealed a similar pattern of lower milk cortisol levels in the 12 hG group. This phenomenon could be attributed to the role of the natural environment of the pasture area the grazing cows are exposed to, as previously described [15]. Corroborating the result of this study, in an earlier study, we showed lower hair cortisol levels in cows that grazed in pastures with a natural environment compared to the control indoor system during hot and humid summers [7]. The lower levels of cortisol in the 12 hG group in the middle of the experiment were confirmed by the same measure at the end of the experiment. Unexpectedly, we found higher levels of milk cortisol levels in the 24 hG group compared with the 12 hG and control groups. We actually expected a decrease in the cortisol levels as the time of grazing increased. One hypothesis is that during nighttime rest, the 24 hG group was uneasy throughout the experiment due to lots of rainy days in the pasture area (average rainy days: 3–4). This discomfort among the cows was indicated by the fact that the cows were dirtier, particularly in the udder area, because they lay down in the pasture, which was confirmed by the higher SCC in the same group of cows. Consistent with this claim, Díaz et al. [50] listed various factors that can alter plasma cortisol levels, including but not limited to the management system, such as the type of housing [54] and transport [55], lactation stage [56], and milk yield [57]. Another factor could be the lower feed intake in the 24 hG group compared with the 12 hG group, which can be considered a source of stress, as stated by Gellrich et al. [51] and Fisher et al. [52]. They reported that feed restriction during early lactation, when production is the highest, can cause imbalanced energy levels, leading to activation of the HPA axis [58] and thus increasing cortisol levels in the blood. This is mainly associated with the physiological stage of the cow and its attempt to maintain homeostasis in using body reserves to compensate for the negative energy balance. However, Gellrich et al. [51] reported that the feed intake reduction over the short period in their study did not lead to a stress response involving HPA axis activation and increased cortisol secretion. The cows in this study were still in the high constant lactation period, after the peak, and could be influenced by any stressor, such as lowered feed intake. Overall, the results obtained in milk production and quality in the 24 hG group were expected as well as the high levels of cortisol which is probably associated with a restricted intake as one influential factor. The reasons for lower feed intake in the 24 hG group have been discussed earlier in this article and may explain this phenomenon. Moreover, fighting behavior could be a source of stress in animals [59], however, to what extent the activity of the animals can cause higher activation of HPA activity is not clear. It is because the higher activity levels may increase cortisol concentration in the blood and possibly the milk of animals. Thus, cautions should be taken when interpreting the result of cortisol in active and non-active animals. Bring this to an end, although there are studies in human subjects that fighting behavior in athletes (e.g., martial art) [60] and moderate to high intensity exercise [61] can provoke increases in cortisol concentration, however, further studies in domesticated animals (particularly during grazing) are suggested to investigate the effect of activity level on cortisol concentration.5. ConclusionsOur findings indicate that a 12 h grazing allowance in addition to a conventional confined system, herein the free-stall system along with the TMR-feed-based system, is preferable to only confined management or 24 h grazing allowance with respect to a higher favorable milk fatty acid and lower cortisol levels without adversely affecting milk yield. In this study, the assumption of linear stress reduction with increased grazing time was refuted because of the higher milk cortisol levels observed in the 24 hG group. Thus, grazing is an important influencing factor that can improve animal welfare in addition to other factors such as the quality of management (e.g., use of shade in the pasture area for rainy days; use of smart farming protocols, including using cameras to observe cows in the pasture; providing enriched feed in the pasture to prevent DMI reduction, etc.). Regarding lowering the cost of dairy production systems, which is supported by efficient pasture use in diets, a 12 h grazing allowance in addition to the confined system is suggested to provide the most efficient management system. Grazing dairy cows for 24 h, although economically beneficial, should be performed with cautious or better management practice (e.g., providing TMR or additional concentrate in the grazing area) to avoid low DMI and milk yield and impaired compositions. Further study on grazing allowances for varying amounts of time is needed to identify the most suitable grazing time supporting the highest productivity, in terms of both milk enrichment and economics. Based on the findings of this study, we also suggest further study to find possible relationship between subclinical mastitis, lactose, and grazing pasture allowance.

animals : an open access journal from mdpi

10.3390/ani12030325

PMC8833469

At present, climate warming is a very serious environmental problem. A sudden and large increase or decrease in temperature is likely to cause stress response in animals. Rainbow trout is a kind of cultured cold-water fish, which is very sensitive to high temperature. Therefore, it is very vulnerable to heat waves during production. The current study found that the behavior, antioxidant capacity, and natural immune function of rainbow trout under acute heat stress were significantly enhanced in the early stages of stress response, but its anti-stress ability decreased with an increase in stress intensity and duration. Transcriptome sequencing and bioinformatics analysis showed that some non-coding RNAs could competitively bind to target genes, and jointly participate in metabolism, apoptosis, and the immune regulation of rainbow trout under stress environments. In conclusion, our study can lay a theoretical foundation for the breeding of heat-resistant rainbow trout varieties.

As the global climate warms, more creatures are threatened by high temperatures, especially cold-water fish such as rainbow trout. Evidence has demonstrated that long noncoding RNAs (lncRNAs) play a pivotal role in regulating heat stress in animals, but we have little understanding of this regulatory mechanism. The present study aimed to identify potential key lncRNAs involved in regulating acute heat stress in rainbow trout. lncRNA and mRNA expression profiles of rainbow trout head kidney were analyzed via high-throughput RNA sequencing, which exhibited that 1256 lncRNAs (802 up-regulation, 454 down-regulation) and 604 mRNAs (353 up-regulation, 251 down-regulation) were differentially expressed. These differentially expressed genes were confirmed to be primarily associated with immune regulation, apoptosis, and metabolic process signaling pathways through Gene Ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis and coding-noncoding co-expression network analysis. These results suggested that 18 key lncRNA-mRNA pairs are essential in regulating acute heat stress in rainbow trout. Overall, these analyses showed the effects of heat stress on various physiological functions in rainbow trout at the transcriptome level, providing a theoretical basis for improving the production and breeding of rainbow trout and the selection of new heat-resistant varieties.

1. IntroductionRainbow trout are cold-water fish that are widely farmed worldwide. They have no obvious lower limit of temperature, 0–18 °C is their optimum growth temperature, and they are extremely sensitive to high temperature. When the water temperature exceeds 20 °C, it will cause a series of obvious stress reactions, such as reduced or decreased intake, abnormal behaviors, growth and reproduction disorders, and immunosuppression [1]. In water, their body temperature is close to the temperature of the surrounding water. When they cannot stand the high temperature of the surrounding environment, some fish can migrate to more profound and colder water, which can help the animals to a certain extent. However, for most fish, high temperatures mean misfortune. In extreme weather, shallow water will heat up rapidly (especially in intensive artificial environments), making the ambient temperature of rainbow trout close to the extreme temperature they can withstand. This makes the rainbow trout an ideal test species for studying whether physiological regulation can keep up with rising temperatures.The head kidney is an organ similar to the mammalian adrenal gland that synthesizes and releases hormones and produces lymphocytes [2,3]. In terms of function, the head kidney, on the one hand, produces lymphocytes and can synthesize and release cytokines, which has the function of the immune system; on the other hand, endocrine organs that produce and release hormones participate in the regulation of stress and physiological activities by secreting cortisol, catecholamine, and thyroid hormone. The unique function and action form of head kidney makes the bi-directional information exchange and transmission between immunity and endocrine become more close and complicated [2]. Continuous and intense stress stimulation will cause the animal’s body to produce many free radicals and reactive oxygen species (ROS), which damage lipid substances in the cell membrane. Some lipid decomposition products will seriously affect the metabolism and functional activities of normal cells, resulting in cell dysfunction, and then cell damage. The detection of antioxidant factors in rainbow trout serum, including superoxide dismutase (SOD), malondialdehyde (MDA), total antioxidant capacity (T-AOC), and Na+-K+-ATPase, can directly reflect the state of the immune and antioxidant systems of the animal and can also indirectly reflect the degree of peroxidation damage in the body. Under heat stress, macrophages in the head kidney are activated and secrete a large number of pro-inflammatory cytokines, including lysozyme (LYZ), tumor necrosis factor alpha (TNF-α), interleukin 1β (IL-1β), and interleukin 6 (IL-6). In stress response, the release of pro-inflammatory cytokines is an important cause of stress injury and inflammatory response [2,4].Transcriptomics are the collection of genes studied at the RNA level in a particular tissue or a population of cells at a specific developmental stage or functional state. The transcriptome is essentially different from the genome. Under the influence of internal and external environment, transcriptomes can reflect the functional state and phenotype of animal cells in real time and accurately. Therefore, transcriptomics can be used to explore the real-time dynamic linkages between animal genomes and environmental interactions [5]. Noncoding RNA (ncRNA) is usually not directly involved in protein translation and is merely a functional RNA molecule transcribed from DNA [6,7]. ncRNAs are subdivided into several subclasses, including microRNAs (miRNAs) and long noncoding RNA (lncRNAs) [8]. These ncRNAs can competitively bind to mRNA and directly or indirectly regulate the biological function of mRNA [9]. lncRNA sequencing uses specific methods to study the sequencing of noncoding RNAs greater than 200 nt in samples, so as to quickly, comprehensively, and accurately obtain the regulation information of the relevant lncRNAs on target gene mRNAs in the animal body under a specific state. lncRNAs usually act as a signal molecule, inducer molecule, lead molecule, and scaffold molecule when involved in the regulation of gene expression. Transcriptome sequencing was used to mine and analyze the genetic information of rainbow trout, and to screen key non-coding genes (lncRNAs, miRNAs) and coding genes (mRNAs) involved in the regulation of acute heat stress for combined analysis. The present study can explore the molecular regulation mechanism of rainbow trout in response to high temperatures, and provide a theoretical basis for the breeding of heat-resistant rainbow trout varieties in the future.2. Materials and Methods2.1. AnimalsNorwegian rainbow trout (425-day-old, average weight of 118 ± 5 g) with the same genetic background were selected for the experimental material. The experiment was performed in an indoor recirculating aquaculture system under natural light, and the ammonia nitrogen content in the water was no more than 0.03 mg∙L−1, the dissolved oxygen content was no less than 7 mg∙L−1, and the pH value was controlled within 7.4 ± 0.2. We used three oxygento pumps (power is 45 watts) to continuously pump air into each tank, increasing the dissolved oxygen in the water. The experiment was conducted after 2 weeks of domestication at 16 °C water temperature, and fasting was performed 24 h before the experiment. The control group (CO) was set based on the optimal growth temperature of rainbow trout being 16 °C. According to our previous research results [1], the 48-h median lethal temperature of rainbow trout is 22.5 °C, set as the treatment group (LTS). We used three sets of electric heating rods and an automatic temperature controller to raise the water temperature to the target temperature within 2 h. After reaching the target temperature (22.5 °C), the acute heat stress model of rainbow trout was established by continuous stress for 24 h. Three biological replications (20 fish per tank) were conducted in both the LTS and CO groups. That is, 60 fish were used in the control group in three different tanks with 20 fish in each tank (n = 3); another 60 fish were also assigned to the acute stress group, again in three tanks, each containing 20 fish (n = 3). At the end of the experiment, rainbow trout were anesthetized using MS-222 (Silver Chemical Laboratories, Redmond, WA, USA). Blood was collected through the caudal vein and serum was prepared. ELISA kit was used to detect various physiological indices in rainbow trout serum. All test kits were provided by Meimian Industrial Co., Ltd. (Yancheng, China). The head kidney tissue was collected for subsequent tests.2.2. Ethics StatementThis study was conducted in strict compliance with animal ethics and ethics, and was reviewed and approved by the Ethics Committee of School of Life Sciences, Lanzhou University (Approval no. EAF2021042).2.3. Transcriptome Sequencing and AnalysisTo obtain high-quality sequencing data, we conducted a strict quality control at each step. Total RNA extraction from rainbow trout head kidney was conducted using the mirVana™ miRNA Isolation Kit (Ambion, Austin, TX, USA). The obtained RNA (28S/18S ≥ 1.5 and RNA integrity number > 8.0) was evaluated using an Agilent 2100 (Agilent Technologies, Santa Clara, CA, USA) bioanalyzer. The sequencing of lncRNAs was completed by the Illumina HiSeq platform (Wuhan, China). The construction of cDNA libraries and high-throughput sequencing work refers to our previously published research [10]. Clean reads were obtained by filtering raw data, including the removal of rRNA, low-quality reads, contaminated joints, and reads with high N content in unknown bases were then compared to the reference genome of rainbow trout for transcription assembly. CPC (http://cpc.cbi.pku.edu.cn, Accessed on 29 August 2018), txCdsPredict (http://hgdownload.soe.ucsc.edu/admin/jksrc.zip, Accessed on 29 August 2018), CNCI (https://github.com/www-bioinfo-org/CNCI, Accessed on 29 August 2018) software, and pfam (http://pfam.xfam.org/, Accessed on 29 August 2018) database were used to predict the coding ability of the new transcripts to distinguish between mRNAs and lncRNAs accurately. For quantitative analysis, Bowtie2 (http://bowtie-bio.sourceforge.net/bowtie2/index.shtml, Accessed on 29 August 2018) was used to compare clean reads to the reference sequence, and then RSEM (http://deweylab.biostat.wisc.edu/rsem, Accessed on 29 August 2018) was used to calculate the expression levels of genes and transcripts. The main tool for screening and analyzing differentially expressed genes (DEGs) in current work is the DEGseq software [11]. The significance analysis conditions of the DEGs were fold change ≥ 2 and adjusted p-value ≤ 0.001. Cluster analysis was conducted on the expression trend of DEGs using pheatmap software (https://cran.r-project.org/web/packages/pheatmap/index.html, Accessed on 29 August 2018). The Gene Ontology (GO, http://www.geneontology.org, Accessed on 29 August 2018) and the Kyoto Encyclopedia of Genes and Genomes Pathway (KEGG, http://www.genome.jp/kegg/pathway.html, Accessed on 29 August 2018) databases were used to perform functional enrichment analysis on the DEGs, which were divided into three functional modules: biological process (BP); cell component (CC); and molecular function (MF). KEGG analysis identified the DEGs in the signaling pathways. Finally, the Cytoscape software (https://cytoscape.org/, Accessed on 29 August 2018) was used for the joint analysis of the candidate DEGs (lncRNAs, miRNAs, and mRNAs) to obtain the lncRNAs-miRNAs-mRNAs interaction network involved in the regulation of heat stress in rainbow trout.3. Results3.1. Behavior EvaluationDuring the experiment, differences in the indices of fish behavior between the control group and different stress temperature groups were observed (Figure S1). When the water temperature was below 22 °C, no fish died in 48 h. When the water temperature reached 22 °C, the fish began to die after 36 h of stress. As the water temperature increased, so did the cumulative number of dead fish. When the water temperature reached 24 °C, the cumulative death rate of rainbow trout in 48 h was almost 100%. After water temperature exceeded 25 °C, the rainbow trout died within hours. This indicates that 22 °C is the limit temperature for rainbow trout to survive under acute temperature rise. At a suitable water temperature (12–18 °C), rainbow trout always swam in neat arrangements, slowly, and in the same direction. In the early stages of heat stress (when the water temperature exceeded 22 °C, within 6 h before the stress), the fish were in an excited state and moved in the upper layer of the tank, showing fidgety actions, increased alertness, and a faster swimming speed. In the middle stages of heat stress (approximately 24 h), the fish were in a state of inhibition, most of them sank to the tank bottom showing reduced activity, and some of the fish lost balance. In the late period of heat stress (approximately 36 h), the fish began to show abdominal upward, increased activity, sometimes jumping out of the water, and even died.3.2. Oxidative Stress Index and Pro-Inflammatory Cytokine IndexIn the 48-h period post-22.5 °C stress, compared with the CO group, there were significant differences in the SOD activity, MDA content, T-AOC, Na+-K+-ATPase, LYZ activity, IL-1β, IL-6, and TNF-α in the serum (Figure 1A–H). The SOD activity in the LT group increased significantly from 923.88 to 1565.00 U∙L−1 at 0 h and then stayed above 1385.13 U∙L−1 for 48 h. The MDA content in the LT group increased significantly from 4.36 nmol∙L−1 at 0 h to a maximum of 5.07 nmol∙L−1 at 36 h and stayed above 4.33 nmol∙L−1 for 48 h. The T-AOC level in the LT group increased significantly from 16.55 U∙mL−1 at 0 h to a maximum of 21.32 U∙mL−1 after 6 h and stayed above 16.10 U∙mL−1 for 48 h. The Na+-K+-ATPase activity in the LT group increased significantly from 15.00 μmol∙L−1 at 0 h to a maximum of 15.26 μmol∙L−1 at 48 h and stayed above 13.67 μmol∙L−1 for 48 h. The LYZ activity in the LT group increased significantly from 20.62 μg∙L−1 at 0 h to a maximum of 25.39 μg∙L−1 at 48 h and stayed above 19.74 μg∙L−1 for 48 h. The IL-1β level in the LT group increased significantly from 37.02 ng∙L−1 at 0 h to a maximum of 43.55 ng∙L−1 at 36 h and stayed above 36.93 ng∙L−1 for 48 h. The IL-6 level in the LT group increased significantly from 160.73 ng∙L−1 at 0 h to a maximum of 190.83 ng∙L−1 at 36 h and stayed above 141.86 ng∙L−1 for 48 h. The TNF-α level in the LT group increased significantly and reached the peak value of 501.26 ng∙L−1 at 0 h and stayed above 444.10 ng∙L−1 for 48 h. These values were higher in the LT group than in the CO group at every time point. The significant increase in these biochemical indices indicated oxidative damage in rainbow trout under acute heat stress.3.3. DEGsAs stated in our previous study [1], 22.5 °C is the 48-h median lethal temperature (48 h-LT50) of rainbow trout under acute heat stress. After 24-h post-22.5 °C heat stress, the head kidney tissue of the rainbow trout was examined to study the molecular regulation mechanism of the heat stress. Six libraries were constructed in this experiment, namely, the control groups (CO-1, CO-2, and CO-3) and the treatment groups (LTS-1, LTS-2, and LTS-3). We used three software predictions (CPC, txCdsPredict, and CNCI) and a protein database (pfam) to predict the coding ability of new transcripts, and identified lncRNAs and mRNAs (Figure 2A,D). According to different differential multiples, the number of selected differential genes was shown in a histogram (Figure 2B,E). Cluster analysis was conducted on the top 20 DEGs up-regulated and down-regulated in lncRNAs and mRNAs (Figure 2C,F). The number of DEGs is shown as follows: there were 55 DE lncRNAs (up-regulated 35, down-regulated 20) obtained when the fold change was ≥7 or ≤-7; there were 46 DE mRNAs (up-regulated 27, down-regulated 19) obtained when the fold change was ≥10 or ≤-10. We presented the information on the top 20 genes in which the DE lncRNAs and mRNAs were up-regulated and down-regulated (Tables S1 and S2). These DEGs may play an essential role in regulating heat stress in rainbow trout, so we focused on their analysis and verification.3.4. Analysis of DEGsAccording to the result of GO functional enrichment, the function and number of DEGs were statistically displayed in the form of a histogram. As shown in Figure 3, DEGs (lncRNAs) in rainbow trout head kidney under acute heat stress were significantly enriched in BP, CC, and MF functional groups. In the BP group, 1084 DEGs enriched in the “metabolic process” were significantly up-regulated and 1034 DEGs were significantly down-regulated; 638 DEGs enriched in “response to stimulus” were significantly up-regulated, and 673 DEGs were significantly down-regulated; 94 DEGs enriched in the “immune system process” were significantly up-regulated, and 77 DEGs were significantly down-regulated. In CC group, 1472 DEGs enriched in “membrane” were significantly up-regulated, and 1614 were significantly down-regulated; 182 DEGs enriched in “extracellular region” were significantly up-regulated, and 150 DEGs were significantly down-regulated; 31 DEGs enriched in “cell junction” were significantly up-regulated, and 62 DEGs were significantly down-regulated. In the MF group, 2779 DEGs enriched in “binding” were significantly up-regulated, and 2579 are significantly down-regulated; 270 DEGs enriched in “molecular function regulator” were significantly up-regulated, and 278 DEGs were significantly down-regulated; 8 DEGs enriched in “antioxidant activity” were significantly up-regulated, and 5 DEGs were significantly down-regulated.Pathway enrichment analysis is a method to classify and analyze multiple genes with a certain biological function according to the form of signaling pathway. In the current study, the top 20 pathways of KEGG enrichment of DEGs in rainbow trout under acute heat stress were presented in the form of scatter plots (Figure 4). The DE lncRNAs in the head kidney of rainbow trout were significantly enriched in the following signaling pathways: allograft rejection, RNA degradation, pyruvate metabolism, biosynthesis of amino acids, the pentose phosphate pathway, RNA transport, circadian rhythm-fly, phenylalanine, tyrosine and tryptophan biosynthesis, tyrosine metabolism, protein processing in the endoplasmic reticulum (ER), circadian rhythm, steroid hormone biosynthesis, influenza A, viral myocarditis, spliceosome, phenylalanine metabolism, ECM-receptor interaction, hematopoietic cell lineage, phagosome, and antigen processing and presentation. These results were similar to enriched pathways, which were analyzed with predicted genes in the heat stress regulation of rainbow trout.3.5. Regulatory Network of AnalysisIn line with previous miRNA sequencing data [1], the lncRNA-miRNA-mRNA regulatory network was constructed to conduct interaction association analysis on the major DEGs involved in regulating heat stress in rainbow trout. The results showed that there were 18 lncRNAs, 38 miRNAs, and 29 mRNAs DEGs co-expressed, among which there were 88 miRNA-mRNA interaction pairs and 18 lncRNA-mRNA interaction pairs (Figure 5).4. DiscussionGlobal warming and intensification have begun to affect the survival and reproduction of an increasing number of animals, especially cold-water fish such as rainbow trout. The stress of the high temperature environment on animals has a cascade effect from molecules, cells, and organs to the whole body, which destroys the homeostasis of the internal environment of the body and leads to the death of the fish [12]. The damage caused by non-biological stress is procedural. Studies have shown that when stress intensity and duration exceed the threshold of animal tolerance, “metabolic compensation strategy” will be transformed into “metabolic conservation strategy”, resulting in stress damage to some organs, this is to endanger the organs that will not affect life activities temporarily [13]. It is an indisputable fact that high temperatures can damage the internal environment of fish. Although many studies reported that heat stress affects this behavior, energy, metabolism, growth, reproduction, immunity, disease resistance, and other functions of fish, the molecular regulation mechanism of heat stress on fish is still not clear. In the present study, when exposed to 22.5 °C for 48 h, the levels of antioxidant factors (including SOD, MDA, T-AOC, and Na+-K+-ATPase) and inflammatory cytokines (including LYZ, TNF-α, IL-1β, and IL-6) in serum were significantly increased compared with the control group. These results indicated that although the innate immune function of rainbow trout might be enhanced in the early stages of acute heat stress, lipid metabolism would be impaired with the accumulation of many peroxides, which leads to oxidative damage. Transcriptome sequencing studies showed that acute heat stress had the most pronounced effects on the organism’s immune, metabolic, and apoptotic functions (Figure 6).4.1. Immune RegulationIn the current study, transcriptome study and analysis were performed on the head kidney of rainbow trout under acute heat stress, and comprehensive bioinformatics analysis was performed on differentially expressed lncRNAs and mRNAs. Bioinformatics analysis shows that immune regulation plays an important role in rainbow trout resistance to acute heat stress. Meanwhile, DEGs-related to immune regulation were enriched in antigen processing and presentation, and toll-like receptor signaling pathways. Under acute heat stress, transcriptional sequencing showed that DEGs lncRNAs (LTCONS_00051111) and miRNAs (let-7g-3p_2, novel_mir103, and novel_mir58) were significantly down-regulated in the head kidney of rainbow trout, and were competitively bound with mRNAs and negatively correlated with the regulation target gene Shc3 that was significantly up-regulated. Shc proteins are involved in adaptive immune responses [14]. Studies have shown that the expression of the Shc protein is up-regulated in LPS-induced inflammatory response, thereby preventing the secretion and release of pro-inflammatory cytokines (including IL-6 and IL-12) [14]. For example, Shc protein can play an important regulatory role in the production of pro-inflammatory cytokines and anti-inflammatory cytokines in dendritic cells [14]. The dedicator of cytokinesis 11 (DOCK11), also known as Zizimin 2, is a guanine nucleotide exchange factor that activates the Rho family of GTPase cell division 42 (CDC42) leading to cytoskeleton recombination [15]. As DOCK11 is similar to CDC42, it plays an essential role in humoral immune responses [16,17,18,19]. Sakamoto et al. [20] showed that the lack of DOCK11 in B cells decreased the frequency of antigen-specific germinal center B cells along with an increase in apoptosis during immunization. In the current study, the DEGs lncRNAs (LTCONS_00087205) and miRNAs (let-7g-3p_2, and novel_mir159) were significantly down-regulated in the head kidney of rainbow trout, were competitively bound with mRNAs, and negatively correlated with the regulation of the target gene DOCK11 that was significantly up-regulated.As a highly conserved molecular chaperone, heat shock proteins (HSPs) play an essential role in cell function protection, repair, enhancement of stress tolerance and the regulation of innate immune function [12,21,22,23]. In the present study, transcriptome analysis showed that Hsp90 was highly regulated in the acute heat stress of rainbow trout. The DEGs lncRNAs (LTCONS_00017856) and miRNAs (novel_mir111, and miR-301_1) were significantly up-regulated, competitively bound with mRNAs, and positively correlated with the regulation target gene HSP90bb that was significantly up-regulated. The result indicated that in the heat stress environment, the rainbow trout body was induced to produce more HSPs, which may be closely related to the key stress signals and apoptosis molecules, and jointly participate in the stress response and immune defense process. In related studies, Hsp90 is often considered as a marker of cellular stress and endogenous protective proteins, and has the function of molecular chaperone [24]. Other studies showed that in the early stage of heat stress (4 h), the expression of HSPs in rainbow trout erythrocytes was up-regulated, which could prevent cell apoptosis, and promote cell survival [25]. However, when the intensity and duration of stress increases, HSPs can no longer repair the damage and maintain the stability of the intracellular environment, and therefore promote cell death through necrosis or apoptosis. There are many similar results in fish studies. The expression of Hsp90 mRNA is significantly up-regulated in various organs under heat stress [12,21]. These results indicate that the Hsp90 gene is not only a marker of heat-stress response, but also plays an important role in anti-stress damage and immune regulation.4.2. ApoptosisSignal transducer and activator transcription (STAT) play essential roles in cell growth, differentiation, proliferation and apoptosis [26]. STAT binds to its receptor, activates Janus kinase (JAK), and transits cell signals from extracellular to intracellular via the JAK-STAT pathway, activating relevant genes and realizing the regulation of cell growth and differentiation [26]. Ouchi et al. [27] believed that most of the activity of IFN-γ was the result of the transcriptional response mediated by STAT1; under the synergistic effect of BRCA1 tumor suppressor and STAT1, the transcription of the IFN-γ target gene subsets could be differentially activated, and this cytokine could mediate growth inhibition. In the present study, the DEGs lncRNAs (LTCONS_00078080) and miRNAs (miR-205a-5p and novel_mir156) were significantly up-regulated, and were competitively bound with mRNAs and negatively correlated with the regulation target gene Stat1-1 that was significantly down-regulated.Protein kinase R (PKR) is considered to be an interferon-induced protein that plays an important role in antiviral, antitumor and anticellular activities, as well as in the regulation of immunity and apoptosis [28]. PKR can be activated by factors such as HSPs, growth factors, and heparin, and can also be activated in response to various insults, including non-viral pathogens, nutrient or excess energy, cytokines, calcium, ROS, radiation and multiple stressors resulting from the presence of a large quantity of unfolded proteins [28,29,30,31,32]. Similarly, PKR is not only involved in the activation of several transcription factors, but also participates in stimulus-induced apoptosis, including LPS and cytokines [28]. Therefore, PKR, as a central hub of cell stress signal detection and response, plays an important role in signal transmission and regulation of cell functions during the stress response [33]. In our study, the DEGs lncRNAs (LTCONS_00068949, LTCONS_00068954, and LTCONS_00068957) and miRNAs (miR-205a-5p and novel_mir158) were significantly down-regulated, and were competitively bound with mRNAs and negatively correlated with the regulation target gene PKR that was significantly up-regulated. Thus, the expression of some apoptosis-inducing genes under heat stress is regulated by PKR. These results indicate that PKR may be a key gene involved in the regulation of apoptosis under heat stress, which needs to be focused and verified.Activating transcription factor 6 (ATF6) is a member of the leucine zipper family of transcription factors and ER-locator protein [34]. ATF6 can transduce stress signals to the ER and be a major regulator of organogenesis and tissue homeostasis [34]. When additional protein synthesis is triggered in the ER, ATF6, with the ER stress transduction factor inositol required enzyme 1 and PKR-like endoplasmic reticulum kinase, can slow protein translation and induce stress to reduce chaperone and folding enzyme production [34]. In the current study, the DEGs lncRNAs (LTCONS_00023039) and miRNAs (miR-122_1 and novel_mir223) were significantly down-regulated, and were competitively bound with mRNAs and negatively correlated with the regulation target gene ATF6 that was significantly up-regulated.TIPIN protein is a conserved replication-related protein, which can prevent the collapse of the replication fork and play a role in slowing the replication process [35]. TIPIN and its partner TIM (Timeless) are part of the fork protection complex and participate in normal DNA replication to maintain genomic stability [36]. In the current study, the DEGs lncRNAs (LTCONS_00025273) and miRNAs (miR-214-5p and novel_mir157) were significantly up-regulated, and were competitively bound with mRNAs and negatively correlated with the regulation target gene TIPIN that was significantly down-regulated. Chou et al. [37] showed that TIPIN is a nuclear protein that associates with the replicative helicase and protects cells against genotoxic agents. In the studies related to DNA damage, TIPIN is considered to be important in the cell cycle arrest and maintenance of DNA replication, and the depletion of TIPIN can lead to apoptosis in breast cancer cells [38]. Current studies suggest that apoptosis-related DEGs play an important role in the regulation of acute heat stress, which requires further verification and analysis.4.3. MetabolicSome studies suggest that the effect of heat stress on the metabolic function of animals has nothing to do with the decrease in food intake, but is closely related to the metabolic changes of carbohydrates, lipids, and proteins after absorption [39]. Zhao et al. [40] found that six key genes, including lysophosphatidylcholine acyltransferase (Lpcat2), ethanolamine kinase 1 (Etnk1), tafazzin (Taz), sterol carrier protein 2 (Scp2), cholesterol 25-hydroxylase-like protein (Ch25hl), and glycerol-3-phosphate dehydrogenase (Gpd1l), were related to lipid metabolism regulation in turbot kidney under heat stress through integrated analysis of metabolome and transcriptome. A disintegrin and metalloproteinases (ADAMs) are proteolytic enzymes that regulate cell phenotypes by affecting cell adhesion, migration, proteolysis, and signal transduction [41]. Recent studies have found that the elevated expression of ADAM19 is related to the occurrence of metabolic syndrome, and that it is expected to become a new target for treating metabolic syndrome in humans and mice [42]. In the current study, the DEGs lncRNAs (XR_002470621.1) and miRNAs (miR-125c, miR-133-3p, miR-133a-3p_1, and miR-214-5p) were significantly up-regulated, were competitively bound with mRNAs, and negatively correlated with the regulation target gene ADAM19 that was significantly down-regulated. PMS1 is a mismatch repair gene that plays an important role in cancer and some genetic diseases [43]. Mismatch repair is a highly conserved DNA mismatch repair protein that occurs frequently during DNA replication, gene recombination, and some damage, leading to the recognition and repair of mismatched bases [44]. As one of the mismatch repair genes, PMS1 has been widely studied in repairing DNA damage and carcinogenesis, but seldom in stress regulation and metabolism. In the current study, the DEGs lncRNAs (LTCONS_00031115) and miRNAs (miR-212a-5p and novel_mir137) were significantly up-regulated in the head kidney of rainbow trout, and were competitively bound with mRNAs and negatively correlated with the regulation target gene PMS1, which was significantly down-regulated. In conclusion, ncRNA and mRNA competitively bind in correlation analysis, and jointly regulate target genes to participate in the regulation of acute heat stress in rainbow trout.5. ConclusionsConclusively, the present study provides a systematic description of the changes in lncRNAs and mRNAs in rainbow trout under acute heat stress conditions. Within 48 h of acute heat stress, oxidative damage was observed in rainbow trout at the physiological level due to lipid metabolism disorders and the excessive release of pro-inflammatory cytokines. At the transcriptome level, the DEGs lncRNAs and miRNAs competitively bind to mRNAs. The target genes are mainly involved in physiological functions such as immune regulation, apoptosis, and metabolic processes in rainbow trout. Furthermore, the data obtained represent a resource for further investigations of the function of some of these ncRNAs, as it could provide the basic information required to elucidate the mechanisms associated with regulating acute heat stress in rainbow trout at the molecular level.

animals : an open access journal from mdpi

10.3390/ani11102869

PMC8532621

Fasciolosis is a parasitic disease of livestock causing important economic losses worldwide and it is also a zoonosis. Current therapy relies on the use of anthelmintic drugs, which is no longer sustainable due to the increase of anthelmintic resistance and the risk of drug residues in food. A deep understanding of the host-parasite interaction is required to develop protective vaccines for the control of fasciolosis. The aim of the present study is to evaluate the hepatic lesions in sheep vaccinated with a partly protective vaccine for F. hepatica, a non-protective vaccine and an infected control group. The protective vaccine showed less severe hepatic lesions than the infected control group. In addition, in the protective vaccine group dead flukes surrounded by a severe granulomatous inflammation were observed, which taken together with the lower fluke burden, suggests that the host response induced by the partially protective vaccine may have been involved in the death of adult flukes of F. hepatica. This is the first study reporting the presence of degenerated flukes associated to a severe granulomatous inflammation in bile ducts in a vaccine trial, a finding that would be useful for improving vaccine efficacy in future trials.

Fasciolosis is an important economic disease of livestock. There is a global interest in the development of protective vaccines since current anthelmintic therapy is no longer sustainable. A better knowledge of the host-parasite interaction is needed for the design of effective vaccines. The present study evaluates the microscopical hepatic lesions in sheep immunized with a partially protective vaccine (VAC1), a non-protective vaccine (VAC2), and an infected control group (IC). The nature of granulomatous inflammation associated with degeneration of adult flukes found in the VAC1 group was characterized by immunohistochemistry. Hepatic lesions (fibrous perihepatitis, chronic tracts, bile duct hyperplasia, infiltration of eosinophils and lymphocytes and plasma cells) were significantly less severe in the VAC1 group than in the IC group. Dead adult flukes within bile ducts were observed only in the VAC1 group and were surrounded by a severe granulomatous inflammation composed by macrophages and multinucleate giant cells with a high expression of lysozyme, CD163 and S100 markers, and a low expression of CD68. Numerous CD3+ T lymphocytes and scarce infiltrate of FoxP3+ Treg and CD208+ dendritic cells were present. This is the first report describing degenerated flukes associated to a severe granulomatous inflammation in bile ducts in a F. hepatica vaccine trial.

1. IntroductionFasciolosis is a zoonosis caused by the helminth F. hepatica with a significant economic [1] public health importance all over the world and it is considered by the WHO as a re-emerging neglected tropical disease [2]. After the ingestion of the infective form called metacercariae, the pathogenesis of F. hepatica involves a pre-hepatic stage starting with the penetration of newly excysted juveniles (NEJs) through the host intestine wall and its migration within the peritoneal cavity. In the hepatic stage of pathogenesis, the NEJs reach and penetrate the liver capsule, and in the parenchyma the parasites move randomly, forming characteristic transects/tunnels and feed on the hepatic cellular components and blood making NEJs growing and developing rapidly. Lately, the parasites enter the bile ducts where they develop into their adult form and start releasing up to 20,000–24,000 eggs per fluke per day. The feeding and migratory activities cause tissue perforation meanwhile the presence of adult flukes within the bile ducts induces a severe chronic cholangitis with erosion and hyperplasia of the biliary epithelium leading to extensive tissue damage [3].Nowadays, the control of the disease is based on the use of drugs, particularly triclabendazole, which is effective against multiple parasite stages. However, the growing resistance of the parasite to the chemical products and the concerns about chemical residues in food with their detrimental impact on the environment make the development of novel strategies critical that are more effective and sustainable. Thus, vaccines have been highlighted as the most suitable option to deal with the detrimental effects found with the current used chemical therapies [3,4].However, over the last three decades there has been proposed numerous vaccine candidates and assessed in several animal models such as rats, mice, and rabbits with non-consistent results in ruminants [5,6]. The immunomodulatory capacity exerted by F. hepatica is claimed to be the main obstacle to produce an effective vaccine, some new strategies in vaccine development include the identification of protective peptides by mapping B-cell epitopes of immunodominant F. hepatica antigens, which has been recently reported in sheep [7] and cattle [8] with promising results.Moreover, the assessment of vaccine protection is mainly based on parasitological and systemic immunological parameters. However, up to the date, the liver damage and local immune response is not a relevant parameter to assess the effectiveness of the vaccines against liver flukes and it could be a potent tool to complement the parasitological and serological studies carried out routinely in these vaccine trials. Furthermore, few studies have described the liver pathology in vaccine trials conducted in the natural hosts [9,10].Based on these premises, we consider the reduction in the liver damage as a crucial feature to evaluate the vaccine effectiveness against F. hepatica taking into account that a reduction of the liver damage can be beneficial for the animal welfare and the improvement of the animal production, hence reducing the economic impact of the disease. The aim of the present study was to evaluate the microscopical hepatic damage from sheep immunized with a partly protective vaccine composed of four recombinant proteins from F. hepatica in adjuvant Montanide 61 VG, a non-protective vaccine composed of the same antigens in adjuvant Alhydrogel and an infected control group of sheep experimentally infected with F. hepatica. Granulomatous cholangitis associated to dead parasites found in sheep immunized with the protective vaccine were characterized using immunohistochemistry.2. Materials and Methods2.1. Experimental DesignThirty-seven 8-month-old male Merino-breed sheep obtained from a liver fluke-free farm were used in this study. Before starting the experiment, the animals were treated with fenbendazole and subsequently confirmed to be negative for parasite eggs by fecal zinc-sulphate base flotation technique, with no eggs detected. Additionally, all animals were tested for serum IgG specific antibodies for F. hepatica cathepsin L1 (FhCL1) by ELISA, obtaining negative results in all cases. During the experience, animals were housed indoors in the experimental farm of the University of Cordoba and fed with hay and commercial pellet.Sheep were randomly distributed in four groups called vaccine 1 (VAC1), vaccine 2 (VAC2), infected controls (IC) and uninfected controls (UC). Groups VAC1 (n = 10) and VAC2 (n = 10) were immunized subcutaneously with two doses, 4 weeks apart, of a multivalent vaccine. The formulation of the two vaccines assessed in this study were published previously by [11] finding a reduction of fluke burden (37.2%) and egg output (28.71%) in comparison to IC group. Briefly, each vaccine dose (2 mL) contained a cocktail of F. hepatica recombinant proteins including cathepsin L1 (rFhCL1), peroxiredoxin (rFhPrx), helminth defence molecules (rFhHDM), and leucine aminopeptidase (rFhLAP) at a concentration of 100 μg per antigen emulsified in two different adjuvants, Montanide ISA 61 VG (Seppic, Puteaux, France) and Alhydrogel® adjuvant 2% (InvivoGen, San Diego, CA, USA), respectively. The F. hepatica recombinant proteins were obtained as described [11]. Group IC (n = 10) was unimmunized and infected; and group UC (n = 7) was unimmunized and uninfected. Eight weeks after the first immunization, groups VAC1, VAC2, and IC were infected orally with 150 metarcercariae of the South Gloucester strain of F. hepatica (Ridgeway Research Ltd., UK) administered in gelatin capsules with a dosing gun. At 15 weeks post-infection (wpi), all animals were culled in batches of six per day by intravenous injection of a proper dose of T61® (MSD Animal Health, Salamanca, Spain) according to manufacturer’s instruction. The experiment was approved by the Bioethics Committee of the University of Cordoba (No. 1118, date 11 January 2016) and was performed considering European (2010/63/UE) and Spanish (L32/2007 and RD53/2013) directives on animal experimentation. Fluke burden and gross pathology of the liver were reported by [11].2.2. Liver PathologyDuring necropsy, livers were removed and a total of six hepatic samples per liver were collected from hepatic lesions from both the right and left hepatic lobes of each animal. In the UC group, liver samples were randomly collected from the left and right hepatic lobes. Next, all the samples were fixed in 10% buffered formalin for 24 h and routinely processed and embedded in paraffin wax. Tissue sections (4 µm thick) were stained with hematoxylin and eosin (H&E) and evaluated independently by two pathologists to assess the severity of the hepatic lesions per animal and group, as follows: 0, absent; 1, mild; 2, moderate; 3, severe; 4, very severe. The parameters scored were related to chronic stages of the infection and immunized groups (VAC1 and VAC2) were compared to the unimmunized and infected group (IC). The pathological changes studied were fibrous perihepatitis, chronic migratory tracts, bile duct hyperplasia, periportal fibrosis, granulomas, eosinophilic and lymphoplasmacytic infiltrates, globule leukocytes and parasite eggs within the bile ducts or in the hepatic parenchyma.2.3. ImmunohistochemistryThe avidin-biotin-peroxidase (ABC) method was used on paraffin wax liver sections of 3-μm thick as previously described [12]. Briefly, tissue sections were dewaxed, rehydrated and endogenous peroxidase activity was exhausted by incubation with 0.3% hydrogen peroxide in methanol for 30 min at room temperature (RT). Tissue sections were incubated in different retrieval antigen pre-treatments based on the primary antibody used (Table 1). After three rinses in phosphate-buffered saline (PBS, pH 7.2), tissue sections were incubated with 20% normal goat serum (MP Biomedicals) for 30 min at RT. Endogenous liver biotin was blocked using the Avidin/Biotin blocking kit (Vector Laboratories) following the manufacturer instructions. A panel of primary antibodies were diluted in PBS containing 10% normal goat serum (Table 1) and incubated overnight at 4 °C. Following washing in PBS, biotinylated goat anti-rabbit or anti-mouse secondary antibodies (Dako, Agilent, E0432 and E0433, respectively) diluted 1:200 and 1:50, respectively, were applied correspondingly for 30 min at RT. After washing in PBS, the sections were incubated with the ABC complex (Vectastain ABC Elite Kit) for 1 h at RT in darkness, washed in 0.05 M Tris buffered saline (pH 7.6), and then incubated in the chromogen solution (Vector NovaRED Peroxidase Substrate Kit). Finally, the sections were counterstained with Harris’ hematoxylin and mounted in Eukitt quick-hardening mounting medium. Tissue sections in which the specific primary antibodies were replaced by non-immune isotype antibody were used as negative controls.2.4. Cell CountingImmunostained cells with the different antibodies used in the study were counted in five areas of 0.08 µm2, randomly selected areas in granulomatous lesions associated to degenerated flukes. Results were expressed as mild: <10 immunostained cells per field; moderate: 10–30 immunostained cells per field; severe: 30–50 immunostained cells per field; and very severe: >50 immunostained cells per field.2.5. Statistical AnalysisStatistical analysis was carried out for the histopathological hepatic lesions using the Graphpad Prism 7.0 software package (Graphpad Software, Inc., San Diego, CA, USA). A non-parametrical one-way ANOVA Kruskal–Wallis test with Dunn’s post-hoc test was carried out. p values < 0.01 were considered very statistically significant and p values < 0.05 were considered statistically significant.3. Results3.1. Histopathological Evaluation of Hepatic LesionsThe livers of the UC group showed no histopathological changes, portal areas showed scarce connective tissue with occasional lymphocytes. The scores of the microscopical hepatic lesions of the three infected groups (VAC1, VAC2, and IC) are summarized in Figure 1.Fibrous perihepatitis consists of focal fibrosis with or without infiltrate of lymphocytes and plasma cells in the Glisson’s capsule coinciding with the healing of the lesions induced by migrating larvae when they penetrate in the liver or the healing of superficial migratory tracts (Figure 2a,c,e). The severity of this lesion was significantly lower (p < 0.05) in the VAC1 group compared with the IC group, while in theVAC2 group it was slightly lower than in the IC group but without significant differences (Figure 1). Chronic tracts are composed of fibrosis and macrophages loaded with hemosiderin pigment (Figure 2b,d,f) as a consequence of the healing migratory tunnels with hemorrhages induced by migrating larvae. The VAC1 group presented significantly lower chronic tracts (p < 0.05) than the IC group whereas VAC2 groups showed slightly lower severity of chronic tracts than IC group but with no significant differences (Figure 1).Bile duct hyperplasia consists of enlargement of bile ducts and presence of papilla to increase the surface of epithelium (Figure 3a,c,e) and it is due to the presence of adult F. hepatica in the lumen of bile duct causing mechanical irritation with the spines and oral sucker and chemical irritation due to the excretory secretory (ES) product that the adult parasite release. The VAC1 group showed significantly (p < 0.01) lower severity bile duct hyperplasia than the IC group (Figure 1) while no significant differences were found between VAC2 and IC groups. Portal fibrosis is another common hepatic lesion of chronic fasciolosis and consist of fibrosis in the portal spaces, particularly in those showing bile duct hyperplasia (Figure 3a,c,e). The VAC1 group showed significantly (p < 0.01) lower portal fibrosis than the IC group (Figure 1) but no significant differences were recorded for portal fibrosis between the VAC2 and IC groups (Figure 1).Granulomas composed of a necrotic center surrounded by a core of epithelioid macrophages and in some cases multinucleate giant cells and outer inflammatory infiltrate of eosinophils and lymphocytes (Figure 3b,d,f) were found in the two vaccinated groups and in the IC group, but the individual variability of this lesion was high in the three groups and there were no significant differences between vaccinated groups and the IC group (Figure 1). Inflammatory infiltrate of eosinophils was common in portal areas (Figure 4a,c,e), in the periphery of granulomas and surrounding parasite eggs that reached the hepatic parenchyma. The group VAC1 showed a significantly (p < 0.01) less severe infiltration of eosinophils than group IC, whereas the group VAC2 presented highly individual variability and a slightly lower infiltration than group IC but without significant differences (Figure 1). Inflammatory infiltrate of lymphocytes and plasma cells was also common in portal areas (Figure 4a,c,e) and in the periphery of granulomas, in some cases this infiltrate was arranged forming lymphoid follicles. The group VAC1 showed a less severe (p < 0.05) infiltration of lymphocytes and plasma cells than group IC, among the group VAC2 and IC no significant differences were recorded although the infiltration of eosinophils tends to be more pronounced in the group IC (Figure 1). Globule leukocytes with wide cytoplasm containing large eosinophilic granules were observed in the epithelium of some bile ducts, particularly in hyperplastic ducts. The score for globule leukocytes was similar in the three infected groups without significant differences between them (Figure 1). F. hepatica eggs were observed within some bile ducts (Figure 4b,d,f) and in some cases in the hepatic parenchyma associated to a severe inflammatory response with abundant eosinophils. The score for presence of parasite eggs was higher in the IC group than in VAC1 group, without significant differences between the IC and VAC2 groups (Figure 1).3.2. Evaluation of F. Hepatica MorphologyDuring chronic fasciolosis the parasites are allocated within the bile ducts as adults and as consequence of their hermaphrodite reproductive strategy they are able to shed thousands of eggs per day. In this study conducted at 15 wpi, F. hepatica adults were allocated within the hyperplasic bile ducts surrounded by a thick fibrous capsule (Figure 5a). The anatomy of these adults showed the gastrointestinal and reproductive tracts intact with the normal morphology (Figure 5b,c) as has been described previously [13,14]. Moreover, the tegument presented well-developed eosinophilic spines (Figure 5b) which cause mechanical damage of the bile duct epithelium. These adults were reproductively viable confirmed by the high number of eggs found within the reproductive tract of the parasites observed and the eggs found within the lumen of the bile ducts (Figure 5c). Apparently, these eggs were viable containing cells which correspond to the fertilized ovum and vitelline cells enclosed in a yellowish capsule.Unexpectedly, four sheep from the VAC1 group showed occasional degenerated forms of F. hepatica (Figure 5d–f) within bile ducts that were surrounded by a thick band of fibrous tissue and by an unusual severe outermost granulomatous inflammatory infiltrate with the presence of lymphoid follicles (Figure 5d). The majority of epithelium of bile ducts containing degenerated flukes was desquamated but it was possible to identify some epithelial bile duct cells using the cytokeratin AE1/AE3 antibody as marker for this cell type (Figure 6d). Occasionally, necrotic tissue was found surrounding degenerated parasites. Compared with the viable parasites, degenerated flukes showed a marked homogeneous eosinophilia lacking cell nuclei. Moreover, the tegument displayed a loss of spines and its morphology changes drastically toward a waving structure shedding acidophilic tissue associated and the reproductive and digestive systems could not be identified (Figure 5e). Additionally, it was also possible to observe degenerated eggs which could be identified in some cases according to their morphological features and the sclerotin-based eggshell (Figure 5f). These degenerated flukes were considered dead flukes and the bile ducts containing them were associated to a severe granulomatous reaction characterized by a layer of large epithelioid macrophages and multinucleate giant cells surrounding the dead flukes and necrotic tissue, a middle layer mainly composed of lymphocytes with lymphoid follicle formation, some of them showing large germinal center, and an outer layer of fibrous connective tissue.The inflammatory infiltrates associated to degenerated flukes from the group VAC1 were characterized by immunohistochemistry using primary antibodies described in Table 1 and scored in base of the number of positive immunostained cells (Table 2). The immunohistochemical study revealed that the inflammatory infiltrates surrounding degenerated flukes from in the VAC1 group contained a high number of macrophages alternatively activated-M2 (CD163+) (Figure 7a), as well as a high number of macrophages and multinucleate giant cells strongly labelling by the lysozyme and S100 antibodies (Figure 7b,c, respectively). The CD68 antibody was expressed weakly by epithelioid macrophages and multinucleate giant cells in the inner layer of granulomas while it was expressed strongly by circulating monocytes, Kupffer cells, and macrophages located at the periphery of granulomas and in portal spaces. CD163+ cells were also present in the outermost inflammatory infiltrate of granulomas but they were not found within the lymphoid follicles. The CD163 stain is strictly membranous displaying a strong intensity. Thus, the morphology of the CD163+ stained cells are compatible with macrophages. However, lysozyme and S100 antibodies performed a strong cytoplasmic stain that immunolabeled the epithelioid macrophages highlighting the presence of multinucleated giant cells. They were also located in the middle area between the fibrous tissue and the outermost inflammatory infiltrate with no positivity within the lymphoid follicles.The outer inflammatory infiltrate was composed by abundant lymphocytes, the majority of them reacted with the CD3 antibody (Figure 7d), whereas the presence of B cells (IgG lambda light chain+) was moderate (Figure 6e) and the FoxP3+-expressing regulatory T cells was scarce (Figure 7f). The cells CD3+ and B cells were localized in larger quantities in the inflammatory infiltrates outside the lymphoid follicles being present to a lesser extent within them. The staining pattern of both CD3 and IgG-lambda light chain antibodies was cytoplasmic with membranous strong intensity showing a morphology fully compatible with lymphocytes. The Foxp3+ cells showed nuclear and cytoplasmic staining pattern with a strong intensity, being localized dispersed within the outermost inflammatory infiltrate. Moreover, these infiltrates presented a high number of antigen presenting cells positive to MHC-II (HLA-DR+) marker (Figure 6a) displaying a cytoplasmic stain with moderate intensity and morphologically compatible with lymphocyte and macrophages. This positivity was present mainly within the lymphoid follicles but also dispersed along the inflammatory infiltrate area. However, the number of mature dendritic cells (CD208+) (Figure 6b) and proliferating cells (Ki-67+) (Figure 6c) was very scarce. The CD208 positivity was membranous with moderate intensity immunolabelling a very low number of cells in the outermost inflammatory infiltrate localized mainly in the surroundings of the lymphoid follicles. The Ki-67 marker was intranuclear with a strong intensity being localized randomly dispersed within the outermost inflammatory infiltrate.4. DiscussionThe VAC1 included in the present study was considered partially protective since it induced a fluke burden reduction of 37.2% (p = 0.002) and a significant reduction (p = 0.03) of gross hepatic lesion compared to the IC group. However, the VAC2 was considered non protective since it did not induce significant reduction of fluke burdens and gross hepatic lesions with respect to the IC group [11]. The results of the present study confirm the protective nature of VAC1 in term of histopathological hepatic lesions since it induced significant reduction of the majority of microscopical lesions scored (fibrous perihepatitis, chronic tracts, bile duct hyperplasia, portal fibrosis, inflammatory infiltrate of eosinophils, lymphocytes and plasma cells and presence of parasite eggs in bile ducts and hepatic parenchyma) in the VAC1 group with respect to the IC group. The only two lesions evaluated that showed no significant differences between the VAC1 group and the IC group were granulomas and infiltrate of globule leukocytes. On the other hand, the results of the present study also confirm that VAC2 was non protective since none of the microscopical hepatic lesions evaluated showed significant differences between the VAC2 and the IC groups. Among a variety of hematological, serological, and immunological parameters, the hepatic damage score is considered the best single index indicator of fluke burden, as has been previously reported in concern to F. gigantica by [15]. However, not much attention has been paid to evaluating hepatic damage as an additional indicator of protection in vaccine trials and only some studies have taken it into consideration [9,11,16,17,18,19].Apart from the reduction of the above mentioned histopathological hepatic lesions, liver samples from the VAC1 group showed degenerated adults F. hepatica specimens within bile ducts, surrounded by a severe granulomatous inflammatory response. Degenerated flukes could be easily distinguished from viable flukes because they did not display the normal internal structures including reproductive and digestive tracts described previously [13,20,21,22,23] and, conversely, the internal body presented a homogenous acidophilic material that could be compatible with proteinaceous material after the denaturation of the proteins. Degenerated flukes have not been reported in previous pathological studies in experimental vaccine trials against F. hepatica [11,16,17,19,24] or in experimental F. hepatica infections in sheep [25,26] and goats [27]. Furthermore, in rats the effective host response against F. hepatica is supposed to occur during early stages of infection at the intestinal lamina propria or peritoneal cavity while the lumen of the bile ducts where adult flukes are located is considered a relatively immunologically safe environment [28,29]. It has been reported that some migrating flukes may die during the migratory stage and become encapsulated in the hepatic parenchyma surrounded by a fibrous connective tissue capsule forming a cyst that may caseate and become mineralized [30]. In the present study degenerated flukes were found within enlarged bile ducts, moreover, some of them showed degenerated eggs which confirm they had reached maturity before they died. This is the first study reporting degeneration of adult flukes within bile ducts in a vaccine trial for F. hepatica. This finding may explain the lower reduction of fecal egg counts (FEC) than fluke burdens in the VAC1 group, since the degenerated flukes contained degenerated eggs, they could have been producing viable eggs during some weeks and became degenerated later.The mechanism implicated in the occurrence of this aberrant morphology of the adult parasites after vaccination is certainly unknown, but they were only observed in the VAC1 group, which also showed a significantly fluke burden reduction with respect to the infected control group, while VAC2 in which degenerated flukes were not found, did not induce significant fluke burden reduction with respect to the IC group. It is possible that the host immune response generated by VAC1 may have induced an effect on the survival of adult flukes causing the degeneration of some of them, or the blockage of some of the four proteins used in vaccine formulation (FhCL1, FhLAP, FhHDM, and FhPrx) could have affected the viability of adult flukes in the VAC1 group. The adjuvant used in the vaccine formulation may also have influenced this effect since the only difference between the two vaccinated groups was the adjuvant. The presence of a very severe granulomatous reaction with the organization of lymphoid follicles surrounding the fibrous tissue close to the lumen of the bile ducts containing dead flukes was similar to that reported in granulomas associated to disintegration in porcine cysticercosis [31], human neurocysticercosis [32], and in other parasitic infections such as schistomiasis or leishmaniasis [33] and in granulomatous cholangitis associated to the trematode Campula spp. in cetaceans [34]. The presence of a well-formed granuloma-like entity surrounding the degenerated adult flukes has not been described before in vaccination studies in fasciolosis.The characterization of the local cellular immune response induced by the vaccine and/or the parasite infection may provide key information regarding the type of cell subsets involved in it and consequently, help to better comprehend the mechanism triggered during the pathogenesis of F. hepatica infection. In this study, the immunological characterization of this inflammatory infiltrates revealed the presence of epithelioid macrophages and multinucleate giant cells strongly expressing both S100 and lysozyme antibodies forming an inner band surrounding degenerated flukes and necrotic tissue. Epithelioid cells are modified tissue macrophages composing certain granulomas associated with intense immunological activity and that can be induced by several infectious agents [35,36,37]. Lysozyme and S100 antibodies that are specific of phagocytes are claimed to immunolabel the epithelioid macrophages as well [38,39]. Moreover, there was a very high number of CD163+ macrophages which corresponds to M2-macrophages possessing mainly a homeostatic anti-inflammatory and tissue-repair function after a severe tissue injury, as occurs in human or cattle [40,41,42,43,44]. In the present study, M2-macrophages may have played a role in the formation of the abundant connective band of the middle band of the granulomas.CD3+ T lymphocytes constituted the major populations of the outer layer of the granulomas, this result is in agreement with the high number of this cell type in granulomas associated to remnants of the dead parasite Hypoderma lineatum but they were scarce in the periphery of viable larvae [45] and in the periphery of granulomatous cholangitis associated to the trematode Campula spp. in cetaceans [34]. This suggest that CD3+ T lymphocytes play a role in the local host response, either by producing cytokines to induce macrophage activation or tissue repair. B lymphocytes and plasma cells expressing IgG-lambda light chain were also abundant in the peripheral layer of the granulomas, suggesting a strong local humoral response, also reported in other parasitic granulomas [34,45].The high number of HLA-DR+ cells located specifically within the lymphoid follicles which can be B cells, dendritic cells, and/or other antigen presenting cells, suggest that the role of these cell types is important in parasitic granulomas [33,46,47]. These HLA-DR+ cells are involved in the MHC class II-restricted antigen presentation that is essential for CD4+ T cell-dependent immune responses suggesting an efficient adaptive immune response against F. hepatica in the VAC1 group through an active presentation of antigens within the parasite-induced lymph follicles. Additionally, the low presence of CD208+ cells which is a marker expressed in human and ruminants by dendritic cells upon activation [43,48,49] indicates a less relevant role of this cell population at this location.The low number of Treg cells (FoxP3+) in the granulomas associated to dead flukes contrasts with the expansion of FoxP3+ cells in the liver of experimentally infected sheep and goats, particularly at the periphery of enlarged bile ducts [50,51]. The expansion of this cell type in helminth infections is considered an important immunomodulatory effect to facilitate parasite survival [52]. It is possible that VAC1 has had some effect blocking the expansion of Foxp3 cells, which may have caused the local immune response to be more effective against flukes than in the IC and VAC2 groups. The low number of Ki-64+ cells in the granulomas indicates a low proliferation of cells at this late granuloma-like entity as has been reported by other authors in granulomatous lesions indicating the recruitment of cells mainly from blood instead of multiplying in situ [44].5. ConclusionsThe present study reports for the first time the presence of degenerated flukes associated to a severe granulomatous inflammation in bile ducts from sheep immunized with a partially protective vaccine (VAC1) composed of four recombinant proteins from F. hepatica in adjuvant Montanide 61 VG. The characterization of the granulomas associated to degenerated flukes revealed a high number of type 2 macrophages (CD163+) and CD3+ T lymphocytes with a low population of Foxp3+ Treg cells. Since sheep from the VAC1 group showed a significantly lower fluke burden and hepatic lesions than those from the infected control group, it is feasible that VAC1 may have induced certain effective host immune responses and at least, part of this response could have occurred against adult flukes allocated within the bile ducts.

animals : an open access journal from mdpi