Changes in Rumen Epithelial Morphology and Transcriptome, Rumen Metabolome, and Blood Biochemical Parameters in Lactating Dairy Cows with Subacute Rumen Acidosis Following Rumen Content Transplantation

Interventions targeting the gut microbiota, such as fecal microbiota transplantation, prove effective in repairing the intestinal barrier and facilitating the recovery of its function and metabolism. However, the regulatory mechanisms governing the remodeling of rumen epithelial morphology and function, rumen metabolism, and host metabolism in cows of subacute ruminal acidosis (SARA) remain poorly understood. Here, we explored the changes in rumen epithelial morphology and transcriptome, rumen metabolome, and blood biochemical parameters in SARA cows following rumen content transplantation (RCT). The entire experiment consisted of 2 periods: the SARA induction period and the RCT period. During the SARA induction period, 12 ruminally cannulated lactating Holstein cows were randomly allocated into 2 groups, fed either a conventional diet [CON; n = 4; 40% concentrate, dry matter (DM) basis] or a high-grain diet (HG; n = 8; 60% concentrate, DM basis). Following the SARA induction period, the RCT period started. The HG cows were randomly assigned to 2 groups: the donor-recipient (DR) group and the self-recipient (SR) group. Rumen contents were entirely removed from both groups before RCT. For the DR group, cows were administered 70% rumen content from the CON cows, paired based on comparable body weight; for the SR group, each cow received 70% self-derived rumen content. The results revealed no significant differences in the thicknesses of the stratum corneum, granulosum, and spinosum/basale layers, as well as the total depth of the epithelium be-tween the SR and DR groups. All these measurements exhibited a decreasing trend and fluctuations over time after the transfer. Notably, these fluctuations tended to stabilize at 13 or 16 d after RCT in the SR group, whereas they tended to stabilize after 8 or 13 d of transfer for the DR group. Transcriptome sequencing revealed that a total of 277 differentially expressed genes (DEGs) were identified between the 2 groups. Enrichment analysis showed that the DEGs were significantly enriched in 11 Gene Ontology biological processes and 14 KEGG pathways. The DEGs corresponding to almost any of these 11 biological process terms and 14 pathways showed mixed up-or downregulation following RCT. Metabolomics analysis indicated that a total of 33 differential metabolites were detected between the SR and DR groups, mainly enriched in 5 key metabolic pathways, including plant polysaccharides and starch degradation, lipid metabolism, amino sugar and nucleotide metabolism, purine metabolism, and Krebs cycle. Among them, the levels of differential metabolites associated with the degradation of plant polysaccharides and starches, metabolism of amino sugars and nucleotides


INTRODUCTION
Subacute ruminal acidosis (SARA) is a prevalent metabolic disease in lactating dairy cows caused by the excessive consumption of high-grain diets in pursuit of ever-increasing milk yield (Enemark, 2008;Millen et al., 2016).Numerous studies have explored the impact of SARA on rumen health (Kleen and Cannizzo, 2012;Abdela, 2016).In addition to well-documented changes in rumen fermentation and rumen microbial structure and function (Khafipour et al., 2009;Mao et al., 2013;Mu et al., 2021b), the SARA challenge also induces alterations in the rumen metabolic profile.These alterations include elevated concentrations of trans fatty acids resulting from shifts in biohydrogenation pathways within the rumen microbial ecosystem (Sterk et al., 2011;Kleen and Cannizzo, 2012), increased accumulation of biogenic amines due to enhanced abundances of microbiota taxa with amino acid decarboxylase activity (Saleem et al., 2012;Zhang et al., 2017), and an enrichment of degradation products (xanthine, hypoxanthine, lipopolysaccharide and so on) from rumen microbiota owing to the death and lysis of microorganisms intolerant to the low pH (Ametaj et al., 2010;Saleem et al., 2012;Zhang et al., 2017).Alterations in rumen fermentation and metabolism further impair the structural integrity and function of the rumen epithelium, resulting in a thickened rumen epithelial papilla and triggering epithelial inflammation (Liu et al., 2013;Xu et al., 2018).All those variations in rumen microbiota, rumen metabolic profile, and rumen epithelium collectively disrupt rumen homeostasis, ultimately exerting a negative impact on host metabolism and cow health (Mu et al., 2022).Therefore, it is of vital importance to investigate the regulatory strategies and their corresponding mechanisms for the effective prevention and treatment of SARA.
Extensive research in medicine has indicated that the gut microbiome of the hosts with metabolic disorders can be remodeled through microbiota intervention methods, such as fecal microbiota transplantation (Borody and Khoruts, 2012;De Groot et al., 2017;Imdad et al., 2018).Concurrently, for ruminants, microbiota intervention can be accomplished through the transplantation of rumen liquid or content, garnering increased attention in recent years (Jasmin et al., 2011;Weimer et al., 2017;Belanche et al., 2020).In comparison to rumen liquid transplantation, rumen content transplantation (RCT) might be a preferable choice, as it allows the transfer of a larger and more comprehensive amount of rumen microbiota from the donors in a single procedure.Furthermore, RCT has shown some beneficial effects on the modulation of rumen microflora and the improvement of host production performance (Cole, 1991;Weimer et al., 2017).
Therefore, we conducted an experiment to investigate whether RCT from healthy donors could restore rumen homeostasis of the SARA cows, along with elucidating its underlying microbiological mechanism (Mu et al., 2021a).The results indicated that RCT contributes to the restoration of rumen bacterial homeostasis and rumen fermentation in cows suffering from SARA without affecting the core microbiome.Studies on humans and mice suggested that fecal microbiota transplantation cannot only facilitate the repopulating of intestinal flora but is also effective in repairing the intestinal barrier and recovering the gastrointestinal function and metabolism, which finally rectifies the metabolism disorder of the hosts (De Groot et al., 2017;Rao et al., 2021).So, the objective of the present study was to investigate the changes in rumen epithelial morphology and function, rumen metabolism, and blood biochemical parameters in lactating dairy cows with SARA following RCT.The study findings provide some suggestive and comprehensive insights into the regulation of rumen homeostasis in cows with SARA.

MATERIALS AND METHODS
All the procedures in this experiment were conducted according to the Animal Protection Law based on the Guide for the Care and Use of Laboratory Animals approved by the Ethics Committee of Nanjing Agricultural University (Nanjing, China).

Animals, Diets, and Experimental Design
Twelve ruminally cannulated (internal diameter, 10 cm) multiparous and clinically healthy (supervised by the herd veterinary technician) lactating Holstein cows (582 ± 50kg BW; 134 ± 5 DIM; 18.2 ± 2.66 kg/d milk yield) were selected for the experiment.Feeding and management were described previously (Mu et al., 2021a).Briefly, the whole experiment was comprised of 2 periods: the SARA induction period and the RCT period, with each period lasting for 21 d.During the SARA induction period, cows were randomly divided into 2 groups: one group was fed a conventional diet (CON; n = 4; 40% concentrate, DM basis), and the other group was fed a high-grain diet (HG; n = 8; 60% concentrate, DM basis) (Supplemental Table S1, https: / / doi .org/ 10 .6084/m9 .figshare.25027175.v2).Through the SARA induction period, we successfully established a SARA model of lactating dairy cows with a high-grain diet.During the RCT period, the SARA cows were further randomly allocated into 2 groups: the donor-recipient group (DR) and the self-recipient group (SR).The RCT was conducted 1 h before the morning feeding.First, the rumen contents of all 12 cows were entirely extracted and placed into separate large plastic garbage cans that had been cleaned previously.Afterward, the rumen wall was thoroughly rinsed with sterile prewarmed PBS (pH 6.8) to effectively eliminate any residue, excluding the cows in the CON group.For the DR group, cows were administered 70% of the rumen contents from the CON cows, paired based on comparable body weights.For the SR group, each cow was administered 70% of self-derived rumen contents (Figure 1).After the transfer, the diets for all 12 cows were switched to the CON diet.The day of transfer was considered time point 0 d.

Rumen Epithelial Papilla Biopsies and Microscopic Study
Epithelial papillae from the ventral sac were biopsied before evening feeding one day before the transfer, and on d 1, 2, 4, 6, 8, 13, 16, and 20 after the transfer, as described previously (Mu et al., 2022).The obtained papillae (approximately 200 mg) were quickly washed multiple times with ice-cold phosphate-buffered saline to remove feed particles.Then, the papillae were stored in 4% paraformaldehyde or liquid nitrogen for subsequent morphological observation or transcriptome analysis, respectively.
Five papillae per time point of each cow were used for light microscopy histomorphometric observation with methods described previously (Steele et al., 2015).The PFA-fixed papillae were sequentially processed by dehydration, paraffin embedding, sectioning, and staining with hematoxylin and eosin before being mounted for analysis.For each rumen papilla, 3 images were captured, including the base, middle, and tip, and a total of 15 replicates were taken per time point per cow.The thickness of each stratum (corneum, granulosum, and spinosum/ basale) was measured at a magnification of 40x using Image Pro Plus software (Media Cybernetics, Bethesda, MD, USA) according to previously published protocols (Steele et al., 2011).The mean value of all rumen papillae histomorphometric measurements for each cow at each time point was used in the statistical analysis.

Blood Sampling and Analyses
Blood samples were taken from the tail vein into evacuated K 2 -EDTA (anticoagulation) tubes 6 h after the morning feeding one day before transfer, and on d 1, 2, 4, 6, 13, and 20 after transfer.The plasma was separated immediately through centrifugation at 3,000 * g for 15 min at 4°C and stored at −20°C until analysis.The concentrations of glucose, nonesterified fatty acids (NEFAs), triglyceride, cholesterol, and BHBA in plasma were determined in duplicate using standard procedures and commercial kits (Nanjing Jiancheng Biology Research Institute, Nanjing, China).

Epithelial RNA Extraction and Sequencing
Total RNA was extracted using TRIzol (Takara Bio, Otsu, Japan) from rumen epithelial papillae of a total of 32 samples (DR*16 and SR*16), collected before evening feed delivery on d 2, 4, 8, and 16 after transfer (Chomczynski and Sacchi, 1987).The RNA concentration was determined using a NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).RNA quality was measured using an Agilent 2100 bioanalyzer (Agilent Technologies, San Diego, CA, USA), and the RNA integrity number was at least 8.5 for all samples.The sequencing libraries were constructed with 1 mg of high-quality total RNA using an Illumina TruSeq RNA sample preparation kit (Illumina, San Diego, CA, USA) to enrich poly(A)-tailed host mRNA with oligo(dT) beads.After quantification with a TBS 380 fluorometer (Turner Biosystems, Sunnyvale, CA, USA), the pairedend libraries (2 * 150 bp) were subjected to sequencing with Illumina NovaSeq 6000.
Principal-component analysis was conducted using the prcomp function of R (version 4.0.5).Gene Ontology (GO) enrichment and KEGG analyses were employed using the R package clusterProfiler v3.18.1 (Yu et al., 2012).Biological categories were deemed significant at a Q value of < 0.05.

Rumen Content Sampling and Metabolome Analyses
The ruminal content samples were obtained from the ventral sac of the rumen at 0 h before the morning feeding on d 1, 2, 4, 6, 13, and 20 after transfer via the rumen fistula.After collection, the ruminal contents were immediately strained through 4 layers of sterile cheesecloth to obtain rumen fluid and then frozen in liquid nitrogen until subsequent rumen metabolome analysis.
For the metabolome analysis, all 48 rumen fluid samples collected above from the DR and SR cows (Figure 1) were first slowly thawed at 4°C, and a 100 μL aliquot from each sample was taken to mix with 300 μL methanol and 10 μL internal standard (2.8 mg/mL, L-2-Chlorophenylalanine).The mixture was vortexed for 30 s and then kept at −20°C for 1 h, followed by centrifugation at 13,800 × g for 10 min at 4°C.The resulting supernatant was then collected for the liquid chromatographymass spectrometry (LC-MS) analysis using an Ultimate 3000LC-Q-Exactive instrument (Thermo, California, USA) incorporating a Hyper gold C18 column (Thermo; 100 mm × 2.1 mm, 1.9 μm).The mobile phase consisted of mobile phase A [water + 5% (vol/vol) acetonitrile + 0.1% (vol/vol) formic acid] and mobile phase B [ace-tonitrile+0.1% (vol/vol) formic acid], flowing at a rate of 0.3 mL/min.The column temperature was maintained at 40°C.The injection volume was 10 μL and the autosampler was kept at 4°C.The elution procedure was as follows: 5% mobile phase B for 0-1 min, 5% to 95% mobile phase B from 1 to 11 min, and 95% to 5% mobile phase B from 11 to 19.5 min.The mass spectrometric settings for positive/negative ion modes were as follows: the heater temperature was set at 300°C, the sheath gas flow rate was 45 arb, the auxiliary gas flow rate was 15 arb, the sweep gas flow rate was 1 arb, the spray voltage was 3.0 kV/3.2 kV, the capillary temperature was set at 350°C, and the S-lens radio frequency level was 30%/60%, respectively.
The raw data was subjected to feature extraction and preprocessing using Compound Discoverer 2.0 software (Thermo Scientific).Ion peak data present in <50% of the samples were excluded.The primary parameters set were: an intensity threshold of 300,000, an m/z range value of 70 to 1,050, an m/z width of 5 ppm, a frame time width of 0.2 min, and retention time start and end values of 0.01 and 19.5 min, respectively.The data were then normalized based on the interior label and post-edited in Excel 2010 software.The molecular mass data was aligned to identify metabolites using the online Human Metabolome Database (https: / / hmdb .ca)and KEGG database (https: / / www .genome.jp/kegg/ ).The metabolites were reported only if the discrepancy between their theoretical mass and observed mass was within 20 ppm, with further validation by isotopic distribution measurement.The principal component analysis (PCA), orthogonal partial least squares discriminant analysis (OPLS-DA), and loading plots were performed with SIMCA-P software (version 13.0, Umetrics, Umea, Sweden).The validation of OPLS-DA models was conducted based on the variation interpretation (R 2 Y) and predictability (Q 2 ) of the model in cross-validation and permutation tests with 200 iterations.Metabolites with the importance in projection (VIP) score of > 1 and a Benjamini-Hochberg-adjusted P value (Q) of < 0.05 were considered statistically significant between groups.When the metabolites were both identified in positive and negative ion modes, data from the mode with the lower Q value were kept.KEGG pathway analyses were processed using MetaboAnalyst 6.0 (https: / / www .metaboanalyst.ca/).

Statistical Analyses
Power analyses conducted before the experiment determined that a minimum sample size of 4 cows per group was required.Each group was designed to achieve a power of at least 80%, considering an effect size of 0.25 and a type I error of 5%, utilizing G*Power 3.1.9.6 (Faul et al., 2007) based on F-test of repeated measures within-between interaction ANOVA.
Analyses of ruminal epithelium thickness and blood biochemical parameters were performed with the linear mixed-effects models (MIXED) procedure of IBM SPSS statistics V25.0 (IBM Corp., Armonk, NY, USA).The treatment (DR or SR), day, and their interaction were treated as fixed factors.The cow was considered a random effect.If the model indicated a significant difference in sampling time, Tukey's honest significant difference test (R v. 3.6.3)was used for subsequent multiple comparisons.When the model indicated a significant difference in treatment (DR or SR), the t-test (R v. 3.6.3)was conducted to compare the 2 groups at the same sampling time point.Effects were deemed significant when P < 0.05.
The data of rumen fluid metabolome were analyzed using the non-parametric Scheirer-Ray-Hare extension of the Kruskal-Wallis test (Sokal and Rohlf, 1995), which is a non-parametric analog of ANOVA based on ranked variates with 2 independent factors (treatment and day) plus their interactions.

Data Availability
Raw reads of rumen epithelium transcriptome sequencing were deposited in the NCBI SRA database under accession number PRJNA1066254.

Effects of RCT on Morphological Parameters of Rumen Epithelial Papillae
As shown in Table 1, microscopic examination of the rumen epithelial papillae revealed that no significant difference was observed in the thickness of the stratum corneum, stratum granulosum, stratum spinosum/basale, and the total depth of the epithelium between the SR and DR groups (Table 1; P < 0.05).However, sampling day and the interaction between treatment and sampling day had a significant impact on the thickness of each layer and the total layers of the rumen epithelium (P > 0.05).Further Tukey's multiple comparison test between different sampling days showed that the rumen epithelium thickness of each layer and the total layers were significantly reduced in both SR and DR groups after RCT (Table 2; P < 0.05), all of which showed dynamic fluctuation over time.Almost all changes in the epithelium thickness tended to stabilize 13 or 16 d after transplantation for the SR cows, whereas they tended to stabilize after 8 or 13 d of transplantation for the DR cows (Table 2).

Effects of RCT on Blood Biochemical Parameters
Compared with the SR group, the plasma triglyceride concentration was significantly increased in the DR group (Table 1; P < 0.05), which was equivalent to the level in the CON group during the SARA induction period (1.33 vs. 1.30mmol/L).Results of the t-test between the 2 groups on the same sampling day indicated that the increase was detected on d 4 after transplantation (Table 3; P < 0.05).In contrast, the DR group showed a decreased trend in the cholesterol level than the SR group (Table 1; P < 0.10), and a significant decrease in the DR cows was observed on d 13 and 20 after transplantation (Table 3; P < 0.05).The day of sampling affected the levels of plasma NEFAs, triglyceride, and BHBA (P < 0.05), all of which exhibited variable fluctuations across different sampling days after RCT (Table 3).The interaction between treatment and sampling day had an apparent effect on plasma glucose concentration (Table 1; P < 0.05), without affecting other blood biochemical parameters.

Effects of RCT on Transcriptional Profiles of Rumen Epithelial Papillae
A total of 1,344,133,540 high-quality paired reads were produced through transcriptome sequencing of 32 rumen epithelial papilla samples, with an average number of 42,004,173 reads per sample.Among them, 1,185,558,399 reads were aligned to the Bos taurus reference genome ARS-UCD1.2,resulting in a mapping rate of 88.20% (Supplemental Table S3, https: / / doi .org/ 10 .6084/m9 .figshare.25523779.v2).The PCA results showed that the transcriptional profiles of the SR and the DR groups were not significantly separated (PER-MANOVA, P = 0.487; Supplemental Figure S1, https: / / doi .org/ 10 .6084/m9 .figshare.25027175.v2).We obtained 277 differentially expressed genes (DEGs) between the 2 groups in total.Out of these, 113 genes were upregulated and 164 genes were downregulated in the DR cows.

Mu et al.: COW RESPONSE TO RUMEN CONTENT TRANSPLANTATION
The results of the Gene Ontology (GO) enrichment of biological processes indicated that the DEGs were significantly enriched in 11 differential terms of organic acid metabolic process, small molecule metabolic process, carboxylic acid metabolic process, oxoacid metabolic process, small molecule catabolic process, monocarboxylic acid metabolic process, multicellular organismal water homeostasis, growth, water homeostasis, small molecule biosynthetic process, and biological adhesion (Figure 2A; Q < 0.05).The KEGG analysis revealed that the DEGs did not exhibit significant enrichment in any metabolic pathways at a Q value of < 0.05.However, the DEGs were significantly enriched in 14 pathways at a P value of < 0.05, including Th17 cell differentiation, glutathione metabolism, Th1 and Th2 cell differentiation, glycerophospholipid metabolism, biosynthesis of amino acids, various types of N-glycan biosynthesis, linoleic acid metabolism, glycosphingolipid biosynthesis -globo and isoglobo series, arginine and proline metabolism, taurine and hypotaurine metabolism, histidine metabolism, glycosaminoglycan degradation, endocytosis, and vitamin digestion and absorption (Figure 2B; P < 0.05).Despite that, the DEGs corresponding to almost any of these 11 biological process terms and 14 metabolic pathways showed mixed up-or downregulation following RCT (Supplemental Figure S2, https: / / doi .org/ 10 .6084/m9 .figshare.25027175.v2).

Effects of RCT on Rumen Metabolome Profiles
Based on the LC-MS analyses, we identified a total of 155 and 130 metabolites in the positive and negative ion modes from the rumen fluid, respectively.After removing duplicates from the 2 modes, a total of 236 metabolites were obtained, predominantly comprised of organic acids, amino acids, lipids, and fatty acids.
The 3-dimensional PCA result showed that the metabolic profiles of the SR and DR cows were not significantly separated (Figure 3A and C; PERMANOVA test, P > 0.05).However, from the 2-dimensional PCA plot, we found that the samples of the 2 groups within 1 to 2 d were distinctly separated along axis 2, and these samples tended to cluster together after 4 d following RCT (Figure 3B and D).Further OPLS-DA results also revealed clear separations between the 2 groups (Supplemental Figure S3A and B).The cumulative values of R2 Y and Q 2 in the positive (0.726 and 0.586) and negative ion (0.838 and 0.477) modes were all above 0.40, which indicated the stability and reliability of the model.The Q 2 intercept values in the positive and negative ion modes were both less than 0.05 (Supplemental Figure S3, https: / / doi .org/ 10 .6084/m9 .figshare.25027175.v2),signifying that there was no overfitting.
Based on the criteria of a VIP score of > 1 and a Q value of < 0.05, we identified 33 differential metabolites in the positive and negative ion modes in total.Among these, the levels of 19 metabolites exhibited an elevation, whereas the levels of the other 14 metabolites demonstrated a reduction in the SR group compared with the DR group (Table 4).Pathway analysis showed that these differential metabolites were mainly enriched in 9 metabolic pathways (Supplemental Figure S4, https: / / doi .org/ 10 .6084/m9 .figshare.25027175.v2).We then meticulously amalgamated the acquired results to construct a comprehensive metabolic network map manually (Figure 4), which effectively encapsulated the pivotal rumen metabolic changes of the lactating dairy cows with SARA following RCT.It depicted 5 key metabolic pathways, encompassing plant polysaccharides and starch degradation, lipid metabolism, amino sugar and nucleotide metabolism, purine metabolism, and Krebs cycle, with involvement of 10 differential metabolites.e Within a row, means a significant difference (P < 0.05) between the SR and DR groups at the same sampling time.

DISCUSSION
Our previous study revealed that the SARA challenge in dairy cows led to a thickened rumen epithelium, with a concurrent thickening observed in each layer (Mu et al., 2022).In the present study, we conducted RCT with the SARA cows and found that no significant differences were observed in the epithelium thickness of each layer and the total layers between the SR and DR groups, all of which showed decrease and fluctuated over time after transfer.Notably, these fluctuations tended to stabilize at 13 or 16 d after RCT in the SR group, whereas they tended to stabilize after 8 or 13 d of transfer for the DR group.In our prior research, which focused on the impact of RCT on rumen fermentation and bacterial community, we found that all fluctuations in the rumen fermentation parameters for both the SR and DR cows stabilized after 6 d of transfer (Mu et al., 2021a).Considered altogether, these results suggested that the recovery of rumen epithelial morphology was about 1 week slower than the recovery of rumen fermentation parameters.Additionally, RCT accelerated the recovery of rumen epithelial morphological structure, which potentially might be attributed to its positive contribution to the restoration of rumen bacterial homeostasis and rumen fermentation (Mu et al., 2021a).
Results for the rumen epithelial transcriptome demonstrated that the DEGs between the SR and DR groups were significantly enriched in the GO biological processes of cellular processes (organic acid metabolic process, oxoacid metabolic process, carboxylic acid metabolic process, and monocarboxylic acid metabolic process) and metabolic processes (small molecule metabolic process, small molecule catabolic process, small molecule biosynthetic process, water homeostasis, and multicellular organismal water homeostasis) as well as the KEGG pathway of glycan biosynthesis and metabolism (various types of N-glycan biosynthesis, glycosphingolipid biosynthesis -globo and isoglobo series, and glycosaminoglycan degradation), amino acid metabolism (arginine and proline metabolism, and histidine metabolism), lipid metabolism (glycerophospholipid metabolism and linoleic acid metabolism), and immune system (Th17 cell differentiation and Th1 and Th2 cell differentiation).Meanwhile, the corresponding DEGs did not present treatment-dependent up-or downregulated in nearly any of these biological processes or pathways.Thus, we speculated that the impact of HG-induced SARA on ruminal epithelial function was enduring and sustained, and RCT could not facilitate its recovery.
In the current study, a total of 236 metabolites were identified based on the LC-MS metabolomics analysis.These metabolites were mainly organic acids, amino acids, lipids, and fatty acids, which was consistent with the previous studies (Saleem et al., 2013;Zhang et al., 2017).The PCA results revealed distinct rumen metabolic profiles between the SR and the DR groups within 0 to 2 d after transfer, but the samples of the 2 groups began to converge after 4 d of transfer.This observation was in accordance with the principal coordinate analysis results of the rumen bacterial community after RCT (Mu et al., 2021a), suggesting that changes in rumen bacterial structure were accompanied by variations in rumen metabolism.
In total, we identified 33 differential metabolites between the 2 groups, which were involved in 5 key metabolic pathways.Among them, the increased levels of D-glucose from plant polysaccharides and starch degradation, and N-acetyl-D-glucosamine from amino sugar and nucleotide metabolism together resulted in an increased level of β-D-fructose-6P in the DR cows, which ultimately led to an elevated rumen acetate level through pyruvate metabolism.We also found increased levels of inosine and guanosine in the DR group, which could generate NH 3 through purine metabolism (Rodwell, 2003;Moudříková et al., 2017).The NH 3 could be further used to synthesize arginine (Huang et al., 2016) and potentially contribute to the production of valerate through its degradation (Andries et al., 1987).Collectively, those findings provided explanations for the elevated acetate and valerate levels from a metabolic perspective following RCT and parallelly echoed the outcomes of rumen bacterial analyses, all reported in our earlier study (Mu et al., 2021a).Besides, our data revealed decreased levels of a cluster of metabolites associated with lipid metabolism (butyrylcarnitine, MG (22:0/0:0/0:0), L-hexanoylcarnitine, and MG (0:0/20:0/0:0)) for the DR cows; this might further lead to a declined acetyl-CoA level and thereby affected Krebs cycle, which eventually caused a lessened rumen citrate level.
Previous studies showed that the SARA challenge induced by HG diets caused perturbations in plasma biochemical parameters associated with carbohydrate and lipid metabolism (Zebeli et al., 2011;Mu et al., 2022).In the present study, we found no difference in the plasma metabolites allied to carbohydrate metabolism.As for lipid metabolism, we observed a notable increase in plasma triglyceride concentration in the DR cows, comparable to the level observed in the CON cows during the SARA induction period.Conversely, the plasma cholesterol level exhibited a declining trend in the DR group compared with the SR group.Cholesterol is the precursor for bile acid synthesis, and bile acids play a crucial role in lipid metabolism by promoting lipid emulsification and absorption (Javitt, 1994;Chiang, 2009).Therefore, the decreased trend in the cholesterol level in the DR group might be a response to the increased concentration of plasma triglyceride.Moreover, for the DR cows, an apparent increase in plasma triglyceride concentration was detected on d 4 after transfer, whereas a significant decrease in plasma cholesterol concentration was observed on d 13 and 20 after transfer.This might be attributed to an increased utilization of cholesterol for bile acid synthesis from d 4 to 20 following the elevation of triglyceride concentration on d 4 after the transfer, ultimately leading to a reduction in plasma triglyceride level.Overall, our results suggest that RCT might positively regulate the lipid metabolism of the cows and promote the recovery of plasma metabolites related to lipid metabolism.
Finally, it is essential to emphasize that although RCT demonstrated certain positive effects on the recovery of rumen fermentation (Mu et al., 2021a), rumen bacterial homeostasis (Mu et al., 2021a), rumen epithelial morphological structure, rumen metabolism, and blood biochemical parameters in cows suffering from SARA, conducting RCT is a technically demanding and laborious task in practical dairy production.So, in our future studies, we will continue to enhance our work with experimental verification in vitro using corresponding microbial isolates and metabolites and further develop these strains and metabolites for potential application in production, rather than pursuing direct RCT.

CONCLUSIONS
A schematic diagram depicting the alterations of rumen epithelial morphology and function, rumen metabolism, and blood biochemical parameters in SARA cows following RCT is presented in Figure 5.In general, RCT expedited the recuperation of rumen epithelial morphological structure, trailing about 1 week behind the recovery of rumen fermentation parameters.Nevertheless, RCT did not contribute to the restoration of rumen epithelial function.A total of 33 differential metabolites were identified between the SR and the DR groups, predominantly enriched in 5 key metabolic pathways, including plant polysaccharides and starch degradation, lipid metabolism, amino sugar and nucleotide metabolism, purine metabolism, and Krebs cycle.Notably, the levels of differential metabolites associated with the degradation of plant polysaccharides and starches, metabolism of amino sugars and nucleotides, and purine metabolism pathways were significantly elevated in the DR cows.Moreover, RCT promoted the recovery of plasma metabolites linked to lipid metabolism, aligning them with levels comparable to those in the CON group during the SARA induction period.
Figure 1.Experimental layout.(A) Visualization of rumen content transplantation (RCT).After the transplantation, the diets for all 12 cows were switched to the CON diet.(B) Overview of sample collection during the RCT period.The transfer was regarded as time point 0 d.CON = conventional diet; HG = high-grain diet.

Figure 2 .
Figure 2. Functional enrichment analysis of differentially expressed genes (DEGs) in rumen epithelial papilla between the self-recipient (SR) and the donor-recipient (DR) groups.(A) The significantly enriched Gene Ontology biological process terms of DEGs.(B) The enriched KEGG pathways of DEGs at a P value of < 0.05.Q-value represents the Benjamini-Hochberg adjusted P value.

Figure 3 .
Figure 3.The principal component analysis (PCA) of rumen metabolites between the self-recipient (SR) and the donor-recipient (DR) groups.The (A) 3-dimensional and (B) 2-dimensional PCA score plot in positive ion mode.The (C) 3-dimensional and (D) 2-dimensional PCA score plot in negative ion mode.PERMANOVA results with 999 permutations are shown.
Mu et al.: COW RESPONSE TO RUMEN CONTENT TRANSPLANTATION Mu et al.: COW RESPONSE TO RUMEN CONTENT TRANSPLANTATION

Table 1 .
Mu et al.: COW RESPONSE TO RUMEN CONTENT TRANSPLANTATION Comparison of ruminal epithelium thicknesses and blood biochemical parameters between the selfrecipient (SR) and the donor-recipient (DR) groups

Table 2 .
Dynamic changes in ruminal epithelium thicknesses in cows of the self-recipient (SR) and the donor-recipient (DR) groups Within a row, values with different superscripts indicate a statistical difference (P < 0.05) at different sampling times in SR or DR groups.

Table 3 .
Dynamic changes in blood biochemical parameters in cows of the self-recipient (SR) and the donor-recipient (DR) groups Item Within a row, values with different superscripts indicate a statistical difference (P < 0.05) at different sampling time in SR or DR groups.

Table 4 .
The differential metabolites identified in rumen fluid between the self-recipient (SR) and donor-recipient (DR) groupsMetabolitesVIP 1 Q-value 2 Fold change (DR/SR)