Heat stress-associated changes in the intestinal barrier, inflammatory signals, and microbiome communities in dairy calves

Recent studies indicate that heat stress pathophysiology is associated with intestinal barrier dysfunction, local and systemic inflammation, and gut dysbiosis. However, inconclusive results and a poor description of tissue specific changes must be addressed to identify potential intervention targets against heat stress illness in growing calves. Therefore, the objective of this study was to evaluate components of the intestinal barrier, pro-and anti-inflammatory signals, and microbiota community composition in Holstein bull calves exposed to heat stress. Animals (mean age = 12-week-old, mean body weight = 122 kg) penned individually in temperature-controlled rooms were assigned to 1) thermoneutral conditions (constant room temperature at 19.5°C) and restricted offer of feed (TNR, n = 8), or, 2) heat stress conditions (cycles of room temperatures ranging from 20 to 37.8°C) along with ad libitum offer of feed (HS, n = 8) for 7 d. Upon treatment completion, sections of the jejunum, ileum, and colon were collected and snap-frozen immediately to evaluate gene and protein expression, cytokine concentrations, and myeloperoxidase (MPO) activity. Digesta aliquots of the ileum, colon, and rectum were collected to assess bacterial communities. Plasma was harvested on d 2, 5, and 7 to determine cytokine concentrations. Overall, results showed a section-specific impact of HS on intestinal integrity. Jejunal mRNA expression of TJP1 was decreased by 70% in HS relative to TNR calves. In agreement, jejunal expression of heat shock transcription factor-1 protein (HSF-1), a known tight junction protein expression regulator, decreased by 48% in HS calves. Jejunal analyses showed that HS decreased concentrations of interleukin-1 α by 36.6% and tended to decrease the concentration of interleukin-17A. Conversely, HS elicited a 3.5-fold increase in jejunal concentration of anti-inflammatory interleukin-36 receptor antagonist. Plasma analysis of pro-inflammatory cytokines showed that interleukin-6 decreased by 51% in HS relative to TNR calves. Heat stress alteration of the large intestine bacterial communities was characterized by increased genus Butyrivibrio_3, a known butyrate-producing organism, and changes in bacteria metabolism of energy and amino acids. A strong positive correlation between the rectal temperature and pro-inflammatory Eggerthii spp. was detected in HS calves. In conclusion, this work indicates that HS impairs the intestinal barrier function of jejunum. The pro-and anti-inflammatory signal changes may be part of a broader response to restore intestinal homeostasis in jejunum. The changes in large intestine bacterial communities favoring butyrate-producing organisms e.g., Butyrivibrio spp. may be part of a successful response to maintain the integrity of the colonic mucosa of HS calves. The alteration of intestinal homeostasis should be the target for heat stress therapies to restore biological functions, and, thus highlights the relevance of this work.


INTRODUCTION
Livestock producers lose about 2 billion dollars to heat stress each year including $900 million losses in the dairy industry.Reduced feed efficiency is a major factor associated with economic loss as well as reduced dry matter intake (DMI), milk production and body weight gain, and increased deaths and reproductive culls (St-Pierre et al., 2003).Dairy cattle expose to summer ambient temperature and relative humidity in the US are particularly prone to heat stress (Kaufman et al., 2021).To increase heat dissipation through the skin, heat-stressed cattle redistribute blood flow from splanchnic to peripheral tissues, decreasing oxygen and nutrients reaching the gastrointestinal tract.Subsequently, the reduced blood flow limits the supply of oxygen and nutrients, blunts intestinal mucosa development, increases epithelial stress, and alters intestinal barrier function (Hall et al., 2001).The impairment of the intestinal barrier integrity leads to bacterial infiltration and penetration of antigens causing the activation of the innate immune system.Consequently, heat stress leads to local and systemic activation of inflammatory pathways shifting nutrient availability and utilization away from anabolic functions such as skeletal muscle synthesis in growing calves and milk synthesis in lactating cows (Hall et al., 2001, Koch et al., 2019, Fontoura et al., 2022, Ríus et al., 2022).
Direct and indirect evidence of impaired intestinal barrier function and homeostasis caused directly by heat stress has been reported in cattle and other species (Kaufman et al., 2020, Cantet et al., 2021, Kaufman et al., 2021).Studies in rodents (Hall et al., 2001, Lambert et al., 2002) and lactating cows (Koch et al., 2019) indicate that impairment of intestinal barrier function is associated with dysfunctional tight junction proteins between adjacent epithelial cells as well as shifted immune responses to the altered local environment.The balance of pro-and anti-inflammatory cytokines produced by epithelial and immune cells modulate the host response against lumen antigens by altering tight junction protein expression and intestinal barrier function (Dokladny et al., 2008).Altered expression of tight junction proteins and inflammatory signals may be part of the deleterious consequences of heat stress in cattle.Therefore, the study of local mechanisms associated with altered homeostasis in cattle should enhance our understanding of heat stress pathophysiology.
Intestinal microorganisms have a symbiotic relationship with the host and the disruption of host-microbial homeostasis, i.e., dysbiosis, has been linked to intestinal barrier dysfunction.Hooper et al., (2001) showed that gut colonization with commensal Bacteroides thetaiotaomicron elicited the upregulation of 7 genes linked to mucosal barrier function and the upregulation of several genes involved in the cytoskeleton and extracellular matrix, suggesting the stimulation of the mucosal barrier integrity.In chickens, heat stress affected the composition of intestinal microbiota and reduced viable counts of beneficial Lactobacillus and Bifidobacterium, resulting in dysbiosis (Al-Fataftah and Abdelqader, 2014, Song et al., 2014, Kers et al., 2018).Studies conducted on sheep and heifers showed that heat stress negatively affected rumen microbiota composition (Tajima et al., 2007, Duffy et al., 2018).Although heat stress had little effect on the α diversity in rumen samples, alterations at the phylum and genus levels were observed in the rumen of goats (Zhong et al., 2019).Our group showed that feeding a postbiotic from Aspergillus oryzae to heat-stressed lactating cows and calves reduced the pro-inflammatory state and improved intestinal barrier function and nutrient metabolism (Kaufman et al., 2021, Ríus et al., 2022).This suggests that the beneficial effects of the Aspergillus oryzae postbiotic resulted from improvements in gut microbiota composition.Heat stress alterations of intestinal microbial communities in cattle are not well described; however, this information may facilitate the development of prophylactic interventions to maintain gut health in heat-stressed cattle.
A comprehensive study was performed to understand the direct effects of heat stress on intestinal homeostasis despite the expected reduction in DMI.In this regard, this work aimed to evaluate the components of the intestinal barrier and mediators of the inflammatory response by measuring gene and protein expression and the composition of intestinal bacterial communities in dairy calves.We hypothesized that the heat stressassociated alterations of intestinal barrier function will be accompanied by a downregulation in the expression of tight junction proteins, an increase in pro-and antiinflammatory signals, and intestinal dysbiosis.A better understanding of the events altering intestinal homeostasis would illuminate ways for developing strategies to combat the negative consequences of heat stress.

MATERIALS AND METHODS
This study was approved by the Institutional Animal Care and Use Committee of the University of Tennessee (IACUC, protocol #2665-0219) following animal ethics approval and regulations.Full details regarding study design, housing, management, and animal performance were described previously in a companion paper (Ríus et al., 2022).For this work, we used a subset of animals to study the direct effect of heat stress, despite the associated dissimilar DMI.

Animal management and experimental design
A total of 16 Holstein bull calves (body weight = 122.2± 8.8 kg, 12-week-old) were divided into 2 cohorts of 8 calves.Each calf was housed in an individual pen in a temperature-controlled room in the Johnson Research and Teaching Unit building (East Tennessee Research and Education Center, Knoxville, TN) for a total of 10 d.Animals were allowed to acclimate for 3 d Yu et al.: Heat stress-associated changes… and then randomly assigned to one of 2 treatments for 7 d: 1) thermoneutral conditions (constant room temperature at 19.5°C) and ~8% restricted offer of starter (TNR, n = 8), and 2) heat stress conditions (10 h of room temperature at 37.8°C followed by 14 h of room temperature at 20°C, relative humidity ranged from 33% to 81%) and fed ad libitum amounts of starter (HS, n = 8) for 7 d.A commercial milk replacer was fed once daily to all animals in both groups (340 g DM/calf/day), and calf starter was provided 4 times daily.On a dry matter basis, the chemical composition of the milk replacer contained 26% crude protein, 20% crude fat, 0.15% crude fiber, minerals, and vitamins.The calf starter contained 18% crude protein, 2% crude fat, 21% acid detergent fiber, minerals, and vitamins.The restriction of starter intake in the TNR treatment was designed to account for the effect of the expected dissimilar nutrient consumption between heat-stressed and thermoneutral animals (Kim et al., 2021;Fontoura et al., 2022).A 7-d treatment period was used based on previous heat stress work in pigs (Pearce et al., 2013).During the 3-d adaptation period all animals were fed ad libitum calf starter and individual animal DMI was recorded.
On d 2, 5, and 7 of the experiment, blood samples were collected from the jugular vein using sodium heparin tubes (BD and Co., Franklin Lakes, NJ).Blood samples were centrifuged at 1,200 × g for 12 min at 4°C and plasma was harvested and stored at −80°C until further analysis.At the end of the treatment period, calves were euthanized by administration of pentobarbital, and tissue samples were collected and processed within 15 min of auscultated cardiac arrest.A 40-cm piece of whole jejunum was collected approximately 10 m distal from the descending duodenum.A piece of the whole ileum was collected approximately 1 m proximal from the ileum-cecum anastomosis.A piece the of whole colon was collected approximately 0.5 m proximal from the transverse colon.Samples of digesta content from the ileum, colon, and rectum segments were collected, snap-frozen, and stored at −80°C until bacteria community analyses were conducted.Total digesta samples were collected based on prior heat stress work reporting intestinal injury and microbiome changes in pigs (Pearce et al., 2013b, Xiong et al., 2020) and poultry (He et al., 2019).Tissue samples of jejunum, ileum, and colon were collected, rinsed with 1 × phosphate buffer saline, snap-frozen, and stored at −80°C.Tissue samples of jejunum and ileum were chosen due to the importance of these segments in nutrient absorption.Additionally, jejunum and ileum have shown high sensitivity to hypoxia and heat stress in rodents, poultry, and swine research models.Colon samples were collected due to the importance of this segment in water absorption and the alteration of the mucosal barrier in heat-stressed mice (Hall et al., 2001, Lambert et al., 2002) and pigs (Pearce et al., 2013b, Pearce et al., 2014).

Isolation of RNA and performance of quantitative PCR
Isolation of jejunum, ileum, and colon mRNA was conducted according to the protocol provided by Direct-zol RNA MiniPrep kit (Zymo Research, Irvine; catalog No. R2050).The procedure eliminated the potential action of inhibitors during isolation of mRNA.In addition, the mRNA quality was evaluated using Nanodrop (Thermo Scientific, Waltham, MA) and agarose gel electrophoresis.Eleven target genes were evaluated using reverse transcription (Table A1; SuperScript First-Strand Synthesis System; Eppendorf, Louis, MO).The qPCR procedure included 40 cycles, each cycle consisted of 4 periods (95°C for 15 min, 95°C for 10 s, 55°C for 30 s, and 72°C for 30 s).Genes that had one melting curve and amplification efficiency between 90% -125.5% were used in this study (Figure A1.Hellemans and Barbara, 2010).Hypoxanthine phosphoribosyltransferase 1 (HPRT1), ribosomal protein S9 (RPS9), ribosomal protein L0 (RPL0), β-actin (ACTB), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) were tested as reference genes.Student's t-test analysis on delta CT (threshold cycle) results of references genes (Schmittgen and Livak, 2008) determined that RPS9, RPL0, and ACTB were deemed reference genes.The 2-delta CT method was used to analyze the results of target genes (Schmittgen and Livak, 2008).Each gene was analyzed in triplicate and the CT for the gene of interest was calculated using the arithmetic mean and the CT for reference genes was calculated using the arithmetic means of each reference gene to compute the geometric mean.
gene of interest geometric mean of reference genes ( )

Tissue protein analysis
Heat shock transcription factor-1 (HSF-1) regulates the expression of occludin in vitro in caco-cells (Dokladny et al., 2008) and in vivo in mucosal cells of pigs (Yi et al., 2020), chicken (Song et al., 2014) and rodents (Huang et al., 2020).Occludin is an integral transmembrane protein and is known for its protective role against a dysfunctional tight junction.To determine the abundance of β-actin, HSF-1, and occludin in jejunum, WES system technology was used follow-Yu et al.: Heat stress-associated changes… ing manufacturer recommendations (ProteinSimple, San Jose, CA, USA).Briefly, sample lysates described above were diluted with sample buffer, then mixed with fluorescent master mix and denatured under 95°C for 5 min.Next, the following was added: 3 µL lysate, 10 µL antibody diluent, 10 µL primary antibody, 10 µL secondary antibody, and 15 µL luminol-peroxide to each well, and the plate was centrifuged at 2,500 rpm for 5 min.The electrophoresis run (30 min) at 475 V followed by blocking (5 min), primary and secondary antibody (30 min,) and luminol/peroxide chemiluminescence detection (15 min).Exposure signals, baseline, peaks, and quality control analysis were monitored to ensure adequate data quality.The area under the curve was calculated using commercial software (Compass for SW Version 3.1.7;Estoppey et al., 2021).Target protein results were normalized to β-actin results presented as the ratio of each target protein to β-actin * 1,000.

Myeloperoxidase analysis
The neutrophil infiltration marker, myeloperoxidase (MPO) activity, was determined to investigate its relationship with inflammatory pathways in the gastrointestinal tract.The MPO activity was measured based on the protocol established by (Suzuki et al., 1983) as modified by (Nieto-Veloza et al., 2019).About 20 ± 5 mg of tissue (jejunum, ileum, and colon) was mixed with 500 µL of hexadecyltrimethylammonium buffer (TCI, Montgomeryville, PA), in a 2 mL tube with 0.2 mg beads and homogenized with a tissue homogenizer for 4 min at 30 Hz and centrifuged for 6 min at 13,400 × g and 4°C (Qiagen, Germantown, MD).The supernatant was collected and transferred to a 1.5 mL tube.The supernatant was mixed with 7 µL of homogenate and 200 µL of the o-dianisidine dihydrochloride solution (TCI, Montgomeryville, PA) then transferred to a 96-well clear plate and read every 60 s for 10 min at 450 nm using a spectrophotometer (Biotek, Santa Clara, CA).Units of MPO in each sample were determined as the change in absorbance following the formula: where ∆A t t 2 1 − ( ) stands for the absorbance margins between 2 adjacent readings; w stands for tissue weight.Protein concentration was quantified using a commercial kit and spectrophotometry (Pierce 660 nm Protein Assay ThermoFisher, Waltham, MA; catalog No. 22662).Protein concentrations were used to normalize the results of MPO activity and reported.
Plasma interleukin-6 (IL-6) and IL-1β concentrations were detected using enzyme-linked immunoassay reagent kits (Invitrogen, Waltham, MA; catalog No. ESS0027 and ESS0029).Briefly, plasma was added to each well in a 96-well microplate (Corning, Tewksbury, MA, catalog No. 9018) incubated for 1 h at room temperature, detection antibody was added, and incubated for another hour at room temperature.Streptavidin-HRP reagent was added followed by a 30 min incubation at room temperature.Substrate solution was added and incubated in dark conditions for 20 min at room temperature followed by stop solution.Readings were conducted at 450 nm using a spectrophotometer (Biotek, Santa Clara, CA) following manufacturer recommendations.

DNA isolation and 16S rRNA amplicon library preparation and sequencing
The establishment of gut microbiota in calves is a dynamic, gradual transition from colonization to stability during the first 9 weeks of age (Du et al., 2023).To determine gut microbial communities, 12-week-old calves were used in our study.Previous research has shown that heat stress affects the microbiome community, in Yu et al.: Heat stress-associated changes… the ileum, colon, and feces causing gut dysbiosis in pigs and broilers (Zhu et al., 2019, Liu et al., 2020, Xia et al., 2022), hence we targeted the sequencing work in ileum, colon, and rectum contents.Genomic DNA extraction and quality control were performed as described by (De la Guardia-Hidrogo and Paz, 2021) and amplicon libraries of the V4 region from the 16S rRNA gene were prepared as described by Paz et al., (2018).Following amplification, PCR products were analyzed on a 2% gel to verify the correct product size.Then, amplicons from each sample were normalized (1 to 2 ng/µL) using the SequalPrep Normalization Plate Kit (Invitrogen, Carlsbad, CA, USA).Normalized samples were pooled by plate and subsequently purified using the MinElute PCR Purification Kit (Qiagen, Valencia, CA, USA).The resulting libraries were quality controlled using the Agilent BioAnalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA) and quantified using the DeNovix QFX Fluorometer (Wilmington, DE, USA).Libraries were paired-end sequenced (2 × 250) using the Illumina Miseq System (Illumina, San Diego, CA, USA) according to the manufacturer protocol.Bioinformatics analyses were performed using QIIME 2 v2019.7 (Bolyen et al., 2019).Denoising was done with an initial quality filtering (Bokulich et al., 2013) using q2quality-filter followed by the Deblur algorithm (Amir et al., 2017).Amplicon sequence variants (ASV) were aligned with MAFTT (Katoh et al., 2002) and then a phylogenic tree was generated with FastTree (Price et al., 2010).Samples were rarefied to an even depth (5168 quality-filtered reads) and rarefaction curves and Good's coverage index (Good, 1953) were used to evaluate sequencing depth (Figure A2).Alpha and β diversity metrics were computed using q2-diversity.The principal coordinate analysis (PCoA) based on the weighted Unifrac distances (Lozupone et al., 2011) was used to visualize clustering among samples.Representative sequences were assigned taxonomy using Naives Bayes classifier trained on the Greengenes operational taxonomic units' reference (McDonald et al., 2012).Heatmaps were generated in R v4.0.3 (RCoreTeam, 2020) using the heatmap.2function from the gplots package (Figure A3, Warnes et al., 2020).Predictions of functional profiles were done using the Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) software via q2-picrust (Douglas et al., 2019).A phylogenetic tree was generated by The Molecular Evolutionary Genetics Analysis (MEGA) software.

Statistical analyses
Tissue and plasma results were analyzed to determine the normal distribution of residuals and homosce-dasticity using Shapiro-Wilk's, and Levine's tests in SAS v. 9.4 (SAS Inst.Inc., Cary, NC).Results that did not meet the above criteria were log-transformed before conducting statistical analysis.Results were analyzed using mixed-effect models (SAS, version 9.4, SAS Institute Inc., Cary, NC) that included the overall mean, the fixed effect of treatment and day (i.e., plasma cytokines), the random effect of calf, and the random error.Least squares means and SEM were reported, and effects were declared significant at P ≤ 0.05, and trends were discussed for P < 0.10.Statistical analyses of microbiota results were performed in R v4.0.3 (RCoreTeam, 2020).Differences in α diversity metrics were evaluated using the Kruskall-Wallis test, while differences in β diversity were evaluated using the permutational multivariate ANOVA (Permanova, permutations = 999).Pairwise comparisons for functional profiles were done using the statistical analysis of taxonomic and functional profiles (STAMP) software v2.1.3 (Parks et al., 2014).The P-values were corrected for multiple testing using the false discovery rate (FDR) method (Benjamini and Hochberg, 1995) and statistical significance was declared at adjusted P ≤ 0.05.

Intestinal barrier, cell stress response, and inflammation
To assess local changes in components of the intestinal barrier and inflammation, gene and protein expression were determined in the intestine of calves (Table 1).The HS treatment elicited a 2.4-fold decrease (P < 0.05) in mRNA abundance encoding the tight junction protein zonula occludens-1 (i.e., TJP1) in the jejunum but not in the ileum and colon (Table 1).In the jejunum, ileum, and colon, the mRNA abundance of other major tight junction proteins with functional and regulatory roles was not affected.Protein abundance results in jejunum showed that treatments did not change the expression of tight junction protein occludin (P > 0.05, Figure 1).
Concerning markers of cellular stress, HS tended (P = 0.098) to increase the mRNA abundance of haptoglobin (HP) in jejunum but not in ileum and colon (Table 2).The expression of transcription factor and regulator of the stress response HSF-1 in the jejunum was lower Yu et al.: Heat stress-associated changes… in the HS compare with TNR group (P < 0.05, Figure 1).Gene expression of key inflammatory markers in the jejunum, ileum, and colon were not affected by treatments (Table 2).

Pro-and anti-inflammatory cytokine concentrations and MPO activity
The jejunal analysis of major pro-inflammatory cytokines showed that HS calves had a 36% reduction in IL-1α concentration, compared with their TNR counterparts (P < 0.05, Table 3).In line with this result, HS tended to decline jejunal IL-17A concentrations (P = 0.0734).Furthermore, HS resulted in a 1.5-fold decline in plasma IL-6 concentration (P < 0.05, Table 3).The analysis of anti-inflammatory cytokines showed that jejunal IL-36RA concentration had a 3.8-fold increase in HS animals (P < 0.05).The activity of neutrophil infiltration marker MPO was assessed in intestinal samples as a possible mediator in the pathogenesis of heat stress, but the MPO activity was not altered (P > 0.10, Table 4).

Microbiome analyses
To investigate changes in the intestinal microbiota and the possible role in the pathogenesis of heat stress, analyses of bacterial communities in the small and large intestines were performed.Statistics assessing α diversity indicated that community richness and evenness were lower in the ileum than colon and rectum digesta whereas dominance was greater in the ileum than colon and rectum digesta (Table 5).Taxonomic profiling indicated a diverse microbial community in the intestine of TNR and HS calves (Figures 2, 3, and 4).In the small intestine, Firmicutes was the predominant phylum with a small presence of Actinobacteria.In the large intestine, however, Firmicutes and Bacteroidetes shared similar proportions and constitute the main phyla (Figures 3  and 4).The results of the α diversity indexes showed that the homogeneity of samples in each group was high and Shannon diversity was affected by the intestinal segment but was not affected by treatment (Tables 5 and 6).The permanova analysis showed differences in the bacterial communities between the ileum and large intestine (P = 0.01; Figure A3).Ileum bacterial communities were not affected by treatment (P = 0.79).
The treatments affected bacterial communities in the large intestine (colon and rectum, P = 0.05) and the LEfSe analysis identified 22 discriminative species between groups.Compared with TNR calves, the species enriched with an LDA score > 2 were numerically lower in HS.Differentially abundant species in the colon and rectum microbiota of TNR calves were dominated by the genus Butyrivibrio_1 while that in HS calves was dominated by the genus Butyrivibrio_3 (Figure 6).The phylogenetic tree identified that Butyrivibrio fibrisolvens is the closest reference strain with Butyrivibrio_3 (Figure 7).The functional profile of ileum and colon microbiome was assessed, and results suggested that functional genes associated with energy and amino acids metabolism and DNA synthesis were affected by treatment (Figures 8 and 9).Spearman correlation coefficients were calculated to reveal relationships between all bacteria abundance in the large intestine with rectal temperature and only significant correlations were presented.Our data indicated that rectal temperature was negatively correlated with Cylindroides in both groups (TNR r = −0.61(P = 0.01), HS r = −0.57(P = 0.03), Figure 10), and Eggerthii was positively correlated with rectal temperature in HS calves (HS r = 0.64 (P = 0.01), Figure 10).

DISCUSSION
Heat stress disrupts homeostasis triggering a response characterized by changes in physiology and metabolism that impair animal productivity.The current view that heat-stressed animals display intestinal and systemic inflammation has been found in rodents, swine, poultry, and bovine models (Lambert et al., 2002, Pearce et al., 2014, Koch et al., 2019, Cantet et al., 2021), but the alteration of homeostasis has been described poorly in heat-stressed dairy cattle.We suggest that this shortcoming needs to be addressed to understand and alleviate health and production concerns.By design, our heat stress protocol resulted in marked hyperthermia (Ríus et al., 2022).The study discussed herein was conducted to determine the direct effect of heat stress (Koch et al., 2019) while the indirect effect of dissimilar intake of nutrient was not the focus of this work.It is worth noting that the TNR calves showed a 6% drop in DMI   (Schmittgen and Livak, 2008), Each gene was analyzed in triplicate and the CT for the gene of interest was calculated using the arithmetic mean and the CT for reference genes was calculated using the arithmetic means of each reference gene to compute the geometric mean. 2 HP: haptoglobin. 3HSP70: heat shock protein 70. 4 INOS: inducible nitric oxide synthase. 5LCN2: lipocalin 2. 6 TLR4: toll-like receptor 4. 7 TNF-α: tumor necrosis factor-α.compared with the HS counterparts (discussed in Ríus et al., 2022).To this point, the reduced metabolic efficiency and elevated intestinal permeability observed in HS but not in TNR calves suggested that the small difference in feed intake had negligible effects on nutrient metabolism (discussed in Ríus et al., 2022).Therefore, the results presented in the present study should be regarded as heat stress-mediated changes in pro-and anti-inflammatory pathways, intestinal integrity, and microbial community.Our analysis of jejunum tissue showed that HS reduced TJP1 abundance.In comparison, jejunum TJP1 abundance increased in lactating dairy cows and growing pigs exposed to constant heat stress for 4 and 7 d (Pearce et al., 2013a, Koch et al., 2019).We found no changes in ileum and colon mRNA abundance of tight junction proteins and these results are consistent with those found in the ileum of growing pigs exposed to heat stress for 6 h (Pearce et al., 2014).The segmentspecific changes in the intestine may be associated with variations in mucus thickness.Compared with the duodenum, ileum, and large intestine, the jejunum has the thinnest mucus layer.The mucus layer in the ileum and colon is 2-4 times thicker than that in jejunum and duodenum (Atuma et al., 2001).Gut mucus functions as a physical barrier that protects the intestinal mucosa from mechanical stress and the infiltration of pathogens (Johansson et al., 2013).A thin mucus layer in the jejunum of calves might be associated with increased stress susceptibility.Further to this point, studies in broilers and mice showed that heat stress reduced the expression of jejunum MUC2 (Goel et al., 2023, Lu et al., 2023).In addition, protein junction ZO-1 is the product of TJP1 expression and a regulator of paracellular permeability between adjacent enterocytes.Zonula occludens-1 is an essential structural protein as well as a gene expression regulator and a component of signal transduction pathways.Although ZO-1 was not measured in this study, jejunal downregulation of TJP1 suggests a disruption in tight junction and, subsequently, augmented para- One unit of MPO activity was expressed as the amount of MPO needed to degrade 1 mmol of hydrogen peroxide/min/mL.cellular permeability.Alternatively, increased intestinal permeability was associated with cellular redistribution of occludin and ZO-1 from membrane to cytosolic location in stressed rats (Mazzon et al., 2002) and might offer a possible explanation for our results.Together, the alterations in tight junction appear to be limited to TJP1, segment-and species-specific, and dependent on heat stress insult characteristics.Tight junction changes associated with HS were paralleled with an increase in intestinal permeability (Ríus et al., 2022) and, thus, this supports the notion of tight junction   In our study, we observed a reduction in the expression of the transcriptional regulator of the heat shock protein pathway HSF-1 in the jejunum of HS animals.In comparison, work conducted in pigs showed that heat stress did not affect jejunum HSF1 mRNA abundance (Pearce et al., 2014).Specific to thermal biology, heat shock factors are rapidly upregulated within minutes to a few hours after initial exposure to thermal insults (Takii et al., 2010).This response induces a transcriptional response of heat shock proteins (HSP) to cope with the deleterious effects of stress on the synthesis, folding, and translocation of proteins as well as to prevent the aggregation of misfolded proteins (Hartl, 1996).A key feature of HSF-1 activation is the transient nature of the transcriptional response, a consequence of both the intensity and the duration of exposure to the stress.Upon recovery from thermal stress, HSF-1 rapidly associates with HSP70 to decrease HSF-1 expression and transcriptional activity (Abravaya et al., 1992, Satyal et al., 1998).Together, in our study, the HSF-1 results appear to be linked to the downregulation of tight junctions and impairment of mucosal barrier function.
We observed heat stress-associated alterations in the expression of inflammatory signals in jejunum, in agreement with results reported in heat-stressed lactating cows (Koch et al.,2019).In addition, heat stressassociated gut mucosa alterations have been described in the literature; however, our group highlighted that the jejunum of cattle may have greater susceptibility than other segments of the gastrointestinal tract (Cantet et al., 2021).Subsequently, we measured cytokine concentrations in jejunum samples only.The pro-inflammatory cytokines IL-1α and IL-17A were reduced in the jejunum of HS calves, which suggests a modulation in inflammatory tone, transitioning from a pro-inflammatory to an anti-inflammatory state, to enable intestinal healing and restoration of homeostasis.In support of this view, mice deficient in IL-1α were protected from enteritis and induced inflammation inflicted by the administration of dextran sodium sulfate (Bersudsky et al., 2014, Malik et al., 2016).In addition, the activation of IL-1α signaling can lead to a selfperpetuating inflammatory state (i.e., dysregulation of inflammation) unless a potent anti-inflammatory signal suppresses the IL-1α signaling.
In our study, the increase in jejunal IL-36RA concentration suggests the downregulation of pro-inflammatory pathways and this is required for tissue regeneration and to restore homeostasis.This notion is supported by data in vivo and in vitro indicating that IL-36RA binds to the IL-36 receptor with higher affinity than IL-36 agonists resulting in the downregulation of inflammation (Gunther and Sundberg, 2014).In support of this view, IL-36RA gene knockout resulted in hypersensitivity to dextran sodium sulfate-induced enteritis (Yang et al., 2022).Intriguingly, IL-36 receptor ligands are overexpressed in animal colitis models and may play both pathogenic and protective roles, depending on the context.Together, it is possible that an increase in jejunal IL-36RA concentration regulated the maintenance and restoration of the intestinal structure in heat-stressed calves.
Observed MPO results suggest that a neutrophilderived response may not be related to the changes observed in the jejunum on d 7 of the study.In comparison, pigs exposed to heat stress for 24 h displayed enhanced MPO activity in the ileum (Pearce et al., 2013b).Neutrophils elicit an early response and produce and release MPO to combat the translocation of antigens from the intestinal lumen to the gut mucosa and submucosa.Thus, the lack of changes in MPO activity in the current study suggests the negligible involvement of neutrophils in preventing the translo- Gut microorganisms influence the host's homeostasis and exert strong influences on physiology, regulating metabolism and immune function, as well as complex animal behaviors (Lynch and Hsiao, 2019).In our study, the general characterization of the core microbial communities shows similar intestinal microbial community composition relative to those found in the intestine of pigs (Xiong et al., 2020).Our analysis showed that treatment explained 44% of the variation in microbial communities in the large intestine.The microbiome analysis showed that Firmicutes were the dominant phylum in all 3 intestinal segments.Bacteroidetes were predominating in the colon and rectum but not in the ileum, and these results are in line with previous reports (Jami et al., 2013).In our study, the genus Butyrivibrio was affected by treatment.Butyrivibrio fibrisolvens (cluster XIVa) is the main butyrate-producing bacteria (Louis and Flint, 2009) and is one of the most common in the rumen (Merchen, 2002).Previous work showed that butyrate stimulated the production of tight junction proteins and mucosal epithelium in the large intestine (Peng et al., 2009).Butyrate-producing bacteria may be responsible for the lack of molecular changes in the colon samples analyzed in our study.Also, butyrate promoted the differentiation and abundance of regulatory T cells and modulated the inflammatory response in intestinal segments in mice models of enteritis (Josefowicz et al., 2012, Arpaia et al., 2013).Although microbial metabolites were not measured in our study, the enrichment of a specific population of butyrate-producing bacteria (i.e., B3) may be related to an increase in the anti-inflammatory signal suggested by the augmented IL-36RA concentration in the jejunum, thus, the enrichment of butyrate-producing bacteria could aid in restoring gut homeostasis in heatstressed calves.
A correlation heatmap analysis was conducted to determine relationships between the relative abundance of large intestine bacteria and rectal temperature in TNR and HS calves.A negative correlation between the rectal temperature and the abundance of Cylindroides was found in TNR and HS calves.This correlation might be associated with a reduction in short-chain fatty acid production and the protective role against gut inflammation.This is because there is a strong positive correlation between the abundance of Eubacterium spp.and short-chain fatty acid concentrations (Mukherjee et al., 2020), which is beneficial for the treatment of enteritis in rodents and humans (Smith et al., 2013, Morrison and Preston, 2016, Venegas et al., 2019).In HS calves only, there was a strong positive correlation between the rectal temperature and Eggerthii spp.While B. eggerthii has not been studied extensively, it was shown to be abundant in people with type 2 diabetes (Medina-Vera et al., 2019).Bacteroides eggerthii was identified as colitis-promoting species in mice with undisturbed intestinal microflora and in mice with antibiotic-depleted intestinal microflora (Dziarski et al., 2016).These results suggest that HS may have created an environment favoring B. eggerthii colonization in the large intestine, but the nature of this relationship remains unknown.Collectively, our results indicate that heat stress plays a role in altering butyrate-producing bacteria, which might contribute to the inflammatory response in the intestine in calves.

CONCLUSION
Heat stress-associated alterations of intestinal barrier function appear to be segment specific and accompanied by changes in the expression of tight junction proteins, pro-and anti-inflammatory signals, and alterations of bacterial communities.The low abundance of TJP1 and HSF-1 in the jejunum may be associated with the impairment of the intestinal barrier during heat stress.Additionally, changes in pro-and anti-inflammatory cytokines may be part of a broader response to restore intestinal homeostasis during heat stress.The changes in colonic bacterial communities in favor of butyrateproducing bacteria in the HS group may be a contributing factor to maintaining the integrity of the intestinal mucosa, particularly in the large intestine.The alteration of intestinal homeostasis should be the target for heat stress therapies to restore biological functions, and, thus highlights the relevance of this work.

Figure 2 .
Figure 2. The taxonomic profiles of the ileum microbiota at phylum level in thermoneutral feed-restricted (TNR, panel A) and heat stress (HS, panel B) dairy calves.The taxonomic composition was compared between 2 groups based on relative abundance (reads of a taxon/total reads in a sample).

Figure 3 .
Figure 3.The taxonomic profiles of the colon microbiota at phylum level in thermoneutral feed-restricted (TNR, panel A) and heat stress (HS, panel B) dairy calves.The taxonomic composition was compared between 2 groups based on relative abundance (reads of a taxon/total reads in a sample).

Figure 4 .
Figure 4.The taxonomic profiles of rectum microbiota at phylum level in thermoneutral feed-restricted (TNR, panel A) and heat stress (HS, panel B) dairy calves.The taxonomic composition was compared between 2 groups based on relative abundance (reads of a taxon/total reads in a sample).

Figure 5 .
Figure5.Beta diversity analysis of large intestine microbiome from thermoneutral feed-restricted (TNR) or heat stress (HS) dairy calves.Results were obtained using principal coordinate analysis (PCoA) based on the weighted Unifrac distances.Differences in diversity were evaluated using the permutational multivariate ANOVA (PERMANOVA).Ellipses represent 95% confidence intervals.

Figure 6 .
Figure 6.The LEfSe analysis identified the biomarker bacterial species in the large intestine of thermoneutral feed-restricted (TNR) and heat stress (HS) dairy calves.The linear discriminant analysis (LDA) scores represent the effect size of each abundant species.Species enriched in each group with an LDA score > 2 are considered.Prefix f represents family, g represents genera; suffix 1,2,3 represent different strains.

Figure 7 .
Figure 7. Maximum-likelihood tree showing specific Butyrivibrio detected by LEFSe's analysis of reference strains, obtained from Basic Local Alignment Search Tool (BLAST).GenBank accession numbers of the reference strains are shown together with the name of the strains.The scale bar represents the number of substitutions per sequence position.

Figure 10 .
Figure 10.Heatmap of Spearman correlation between rectal temperature (RT) and specific bacteria species in the large intestine of calves exposed to thermoneutral (TNR, A) or heat stress (HS, B).All correlations listed in the heatmap are significant (P < 0.05).

Table 1 .
Yu et al.: Heat stress-associated changes… Expression of tight junction genes in calves exposed to heat stress (HS) or thermoneutral feedrestricted (TNR) conditions (n = 8/treatment)

Table 2 .
Expression of inflammation-related genes in the intestinal mucosa of calves exposed to heat stress (HS) or thermoneutral feed-restricted (TNR) conditions (n = 8/treatment) 1Computed using 2-ΔCT method

Table 5 .
Alpha diversity analysis of the bacterial communities in ileum, colon, and rectum digesta of dairy calves abValues within the same row with different superscripts indicate significant differences (P < 0.05). 1 ASV: Amplicon sequence variants.

Table 6 .
Alpha diversity analysis of the bacterial communities in dairy calves assigned to thermoneutral feed-restricted (TNR) or heat stress (HS), values shown in the table are the mean of three intestinal segments(ileum, colon, rectum)