Ability of the Chinese herbal residue to alleviate short-distance transportation stress in sheep through the remodeling of the rumen microbiome–metabolism axis
Jing Li, Jianrong Ren, Jiawen Xu, Jinhui He, Jingyi Xu, Qingyan Yin, Junhu Yao, Shengru Wu
TL;DR
Chinese herbal residue helps reduce stress in transported sheep by improving their gut microbes and metabolism.
Contribution
The study reveals a novel mechanism by which Chinese herbal residue alleviates transport stress in sheep via the rumen microbiome–metabolism axis.
Findings
Chinese herbal residue supplementation reduced oxidative stress markers like ROS and lactate dehydrogenase in sheep.
The herbal residue increased the abundance of Selenomonas ruminantium and specific CAZymes in the rumen microbiome.
Glycerophospholipid metabolism and related pathways were enriched and correlated with antioxidant effects.
Abstract
Transportation is a common stressor in sheep production that is capable of inducing oxidative stress and impairing sheep health and production performance. This study aimed to investigate the alleviating effects of the traditional formula Siji Antiviral Mixture residue after water extraction, which still contains active ingredients, including fiber, polyphenols, and flavonoids, on short-distance transport stress in sheep, as well as its mechanism of action in regulating oxidative stress through the rumen microbiota‒metabolism axis. Twenty first-lambing East Friesian × Hu sheep hybrids weighing 54.49 ± 7.94 kg were randomly assigned to a control group (CON, basal diet) or a Chinese herbal residue group (CMR, basal diet + 50 g/d CMR) feeding at 4 h after approximately 300 km of short-distance transport. Results indicated that 4 h of short-distance transport significantly elevated serum…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —National Center of Technology Innovation for Dairy
- —National Natural Science Foundation of China
- —Shaanxi Province's Elite Recruitment Initiative: The Three Qin Talents Program - Regional Young Talent Project
- —Fundamental Research Funds for the Central Universities
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsRuminant Nutrition and Digestive Physiology · Animal health and immunology · Effects of Environmental Stressors on Livestock
Introduction
With the development of intensive sheep farming, cross-regional transportation has become increasingly common and is now a key factor that can induce oxidative stress. Factors such as jostling, fasting, and dehydration during transport can induce reactive oxygen species (ROS) accumulation, leading to lipid peroxidation, cell membrane damage, protein inactivation, and DNA damage [1–3]. Based on duration and distance, transport stress is typically categorized into long-distance and short-distance types [4–6]. Research has focused predominantly on long-distance transport, which has more pronounced negative effects on animal health and production performance, with longer-lasting detrimental impacts on animals [7–10]. In contrast, there are fewer relevant studies on short-distance transportation stress. Only Odore et al. [11] reported that even short-distance transit can elicit significant stress responses in cattle, as indicated by increased blood progesterone levels. However, there is a lack of systematic studies on whether short-distance transportation similarly induces ROS accumulation and tissue damage.
The rumen microbiota plays a central role in ruminant nutrition and homeostasis maintenance. Studies indicate that long-distance transport significantly alters the rumen microbial composition and function [12] and that the microbiota is highly sensitive to stress—even brief exposure to stressors can disrupt its ecological balance [13, 14]. Stress-induced ROS can directly or indirectly affect microbial structure, promoting pathogenic bacterial proliferation and exacerbating digestive dysfunction [15, 16]. Recent studies have further clarified the dynamic changes in the rumen microbiota during long-distance transport stress and proposed interventions during critical windows to alleviate stress [17]. However, it remains unknown whether ruminants would experience similar changes in the rumen microbiota after short-distance transportation and how the microbial activity could affect host’s response.
Leveraging the modifiable nature of the rumen microbiota and reshaping microbial communities through functional feed additives has emerged as a novel strategy to mitigate transport stress. Traditional Chinese medicines (TCMs) and their residues are rich in polyphenols, flavonoids, and degradable fibres, offering notable anti-inflammatory and antioxidant properties [18]. As a traditional Chinese medicinal preparation, the residue from Siji Antiviral Mixture after water extraction remains rich in degradable fibre and active components such as polyphenols and flavonoids, potentially sustaining anti-inflammatory, antioxidant, and gut microbiota regulatory effects. Extensive research has demonstrated the antioxidant and immune-enhancing effects of herbs in the Siji Antiviral Mixture, such as Houttuynia cordata, Perilla frutescens leaf, and Forsythia suspensa [19–21]. On the basis of these findings, we hypothesize that the residue from the Siji Antiviral Mixture may mediate antioxidant effects by regulating the “microbiota–metabolism axis”. We employed combined rumen metagenomics and metabolomics analyses to elucidate the mechanism of action of this herbal residue.
Materials and methods
Animal ethics
The animal management and experimental procedures in this study were performed in accordance with the Guidelines for Care and Use of Laboratory Animals and were approved by the Animal Ethics and Welfare Committee of Northwest A&F University (approval number: NWAFU20230354).
Experimental design and sample collection
Twenty first-lambing crossbred sheep (East Friesian × Hu sheep), with an average body weight of 54.49 ± 7.94 kg, were selected for the experiment. After 4 h of short-distance transportation covering approximately 300 km to the Animal Science Teaching and Research Base of Northwest A&F University, the 20 sheep were randomly divided into a control group (CON group) and a Chinese herbal residue group (CMR group): the sheep in the CON group were fed a basal TMR diet only, and the sheep in the CMR group were fed a basal TMR diet + 50 g/d Chinese herbal residue. The herbal residue originated from the water extraction process used to produce the Siji Antiviral Mixture, which contains 11 herbal ingredients, including mint, forsythia, and chrysanthemum. The dosage of Chinese herbal residue was determined by referring to previous literature [22–25]. The feed formulations are listed in Table 1. Table 1. Composition and nutrient levels of the basal dietItemBasal dietIngredients, % DM Corn silage39.55 Alfalfa hay20.73 Crushed corn22.00 Rice bran meal3.92 Corn germ meal3.52 Soybean meal2.36 Spout corn husks3.56 Expanded urea1.45 Stone powder1.82 Calcium hydrogen phosphate0.22 Salt0.44 Plant oil0.07 Premix^1^0.36Nutrient composition^2^, % DM Dry matter (DM)49.88 Crude protein (CP)14.65 Neutral detergent fibre (NDF)29.57 Starch29.85 Calcium1.13 Phosphorus0.44 Gross energy (GE)16.70^1^Each kilogram of premix contained 800 mg of vitamin E, 45,000 IU of vitamin D and 120,000 IU of vitamin A, 450 mg of niacin, 600 mg of manganese, 950 mg of zinc, 430 mg of iron, 650 mg of copper, 30 mg of selenium, 45 mg of iodine and 20 mg of cobalt^2^The nutritional levels of the feeds were all measured
Before transportation, and before morning feeding at 1 d, 3 d, and 7 d after transportation, blood was collected in a procoagulant tube to prepare serum samples, and blood was collected in an anticoagulant tube to prepare plasma samples. All the samples were subsequently centrifuged at 1,000 × g for 10 min at 4 °C. The supernatant was aliquoted into 2-mL centrifuge tubes and stored at −80 °C until further analysis. Rumen fluid was collected via oral stomach collection tubes. Before sampling, the tubes were flushed, and the first 30 mL of rumen fluid was discarded to minimize saliva contamination. The pH of each sample was measured immediately with a portable pH meter. The samples were subsequently aliquoted into 5-mL sterile cryovials, frozen in liquid nitrogen, and stored at −80 °C until analysis.
Determination of feed nutritional value
Feed samples were collected, dried in an oven at 65 °C, ground through a 1-mm sieve, and analyzed for nutrient content. Dry matter (DM), ash, and crude protein (CP) were determined according to AOAC (2023) methods [26]. Neutral detergent fiber (NDF) content was analyzed using an Ankom A200I fiber analyzer (ANKOM Technology, Macedon, NY, USA) following the methods of Mertens and AOAC [26, 27]. Starch content was measured using a commercial assay kit (A148-1-1, Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Calcium content in the feed was determined by potassium permanganate titration (GB/T 6436-2018) [28], and phosphorus content was measured by the molybdenum yellow colorimetric method (GB/T 6437-2018) [29]. Gross energy (GE) content was determined using an oxygen bomb calorimeter (Model 6100, Parr Instrument Company, Moline, IL, USA).
Determination of volatile fatty acid (VFA) and ammonia nitrogen (NH3-N) contents in the rumen fluid
After the rumen fluid samples were thawed at room temperature, they were centrifuged at 1,000 × g for 10 min at 4 °C to separate the solids from the liquids. Two hundred microlitres of 25% (w/v) pyrophosphate solution and 1,000 μL of the supernatant were transferred to a 2-mL centrifuge tube. The mixture was mixed thoroughly and incubated at 4 °C for 3–4 h to precipitate the proteins. The mixture was subsequently centrifuged at 17,000 × g for 15 min at 4 °C to remove proteins and impurities. Then 200 μL of 28.45 mmol/L crotonic acid solution and 500 μL of the supernatant were transferred to a 1.5-mL centrifuge tube, mixed well, and incubated at 4 °C for 0.5–1 h. Finally, the mixture was filtered through a 0.22-μm membrane, and the filtrate was transferred to a gas chromatography vial. Analysis was performed via an Agilent 6850 gas chromatograph (Agilent Technologies Inc., Santa Clara, CA, USA) equipped with a polar capillary column (HP-FFAP, 30 m × 0.25 mm, 0.25 μm) and a flame ionization detector, as described previously [30, 31]. As previously described, the pupa phenol-hypochlorite assay was employed to determine ruminal ammonia nitrogen concentrations [32, 33].
Serum biochemical marker assays
Serum biochemical indicators such as glucose (GLU-YA), triglyceride (TG), alanine aminotransferase (ALT), aspartate aminotransferase (AST), lactate dehydrogenase (LDH), albumin (ALB), alkaline phosphatase (ALP), total cholesterol (CHO), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), total protein (TP), and non-esterified fatty acid (NEFA) contents were measured via a fully automated biochemical analyser (Hitachi 7600; Hitachi, Tokyo, Japan). Total antioxidant capacity (T-AOC, A015-3-1), superoxide dismutase (SOD, A001-3-2), glutathione peroxidase (GSH-PX, A005-1-2), and malondialdehyde (MDA, A003-1-2) levels were measured via commercial kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) in accordance with the manufacturer’s instructions. The contents of reactive oxygen species (ROS, ml026288), immunoglobulin G (IgG, ml025387), and tumor necrosis factor-alpha (TNF-α, ml022565) were determined via enzyme-linked immunosorbent assay (ELISA) kits (Cloud-Clone Corporation, Houston, USA) in accordance with the manufacturer’s instructions.
Rumen fluid metagenomics
Five sheep from both the CON and CMR groups were randomly selected for rumen fluid metagenomic sequencing. DNA was extracted from the samples via the E.Z.N.A.^®^ Soil DNA Kit (Omega Biotek, USA), and the DNA concentration, purity, and integrity were determined. DNA was fragmented with a Covaris M220 (Gene Inc., China) and screened for fragments with a truncated length of approximately 350 bp to construct a PE library. After bridge PCR, metagenomic sequencing was performed via the Illumina NovaSeq sequencing platform (Illumina, USA).
Quality control of the raw data was performed via Fastp (https://github.com/OpenGene/fastp) [34]. BWA was used to align and filter host sequences [35]. The data were assembled via Megahit (https://github.com/voutcn/megahit) [36]. Open reading frames (ORFs) were predicted from overlapping contigs > 300 bp via MetaGene (https://metagene.nig.ac.jp) [37]. A non-redundant dataset was created via CD-HIT (https://www.bioinformatics.org/cd-hit/) [38]. SOAPaligner (95% identity; https://github.com/ShujiaHuang/SOAPaligner) was used to map and calculate the abundance of microbial genes. Using BLASTP, the amino acid sequences of the nonredundant gene sets were compared with those in the NR, Kyoto Encyclopedia of Genes and Genomes (KEGG), and CAZymes databases to obtain the corresponding species and functional annotation information [39].
Rumen fluid and plasma LC–MS/MS analysis
Five sheep from both the CON and CMR groups were randomly selected for rumen fluid and plasma collecting to conduct untargeted metabolomics analysis. The quantitative rumen fluid and plasma samples were collected in 1.5-mL centrifuge tubes. Metabolite extraction was performed by adding an extraction solvent (acetonitrile:methanol = 1:1). After vortexing for 30 s, the samples were subjected to low-temperature ultrasonic extraction for 30 min at 5 °C (40 kHz). The samples were placed at −20 °C for 30 min. Then, they were centrifuged at 4 °C and 13,000 × g for 15 min. The supernatant was transferred, evaporated to dryness under nitrogen, and resuspended in resuspension buffer (acetonitrile:water = 1:1, v/v). Low-temperature ultrasonic extraction was repeated for 5 min at 5 °C (40 kHz), followed by centrifugation at 4 °C and 13,000 × g for 5 min. The supernatant was then transferred to an injection vial with an insert tube for instrumental analysis. LC‒MS/MS analysis was performed using ultrahigh-performance liquid chromatography system (Vanquish, Thermo Fisher Scientific) coupled with a Q Exactive HF‒X Mass Spectrometer and equipped an ACQUITY HSS T3 column (100 mm × 2.1 mm i.d., 1.8 μm; Waters, USA). The raw data were preprocessed and matched with public metabolite databases, including HMDB (http://www.hmdb.ca/) [40], Metlin (https://metlin.scripps.edu/) [41], and Meiji’s self-built database, to identify metabolites. The data matrix was cleaned via the 80% rule, retaining variables with nonzero values in at least 80% of the samples and removing missing values. Missing values were imputed using the minimum value from the original matrix. To minimize errors from sample preparation and instrument instability, mass spectrometry peak intensities were normalized via sum normalization, generating a normalized data matrix. Variables with relative standard deviations (RSDs) > 30% in the QC samples were removed, and log10 transformation was applied to the final data matrix for subsequent analysis. Orthogonal proportional least squares discriminant analysis (OPLS-DA) was conducted via the R package ropls (version 1.6.2), and model stability was assessed via sevenfold cross-validation. Differentially abundant metabolites were mapped to metabolic pathways via the KEGG database (https://www.kegg.jp/kegg/pathway.html) [42]. Principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), OPLS-DA, KEGG pathway enrichment, and data visualization were performed via the Majorbio Cloud platform (cloud.majorbio.com) [43].
Statistical analysis
SPSS software (version 21.0) was used to perform Student's t-test on the serum biochemical indicators and oxidative stress-related indicators, and visualization was performed via GraphPad Prism 9. The metagenomic and metabolomic data were analysed via the Wilcoxon rank-sum test. Microbial α diversity was calculated via the Chao1 index (for abundance) and the Shannon index (for diversity). PCoA was used for β diversity analysis based on the Bray‒Curtis distance matrix. The intergroup difference test was performed via ANOSIM (999 permutations). Differentially abundant microbes and functional features were identified via the Wilcoxon rank-sum test and visualized with appropriate R packages. For the metabolomic data, OPLS-DA was performed via the ropls package (v1.6.2). The differential microorganisms and their functions were identified via the Wilcoxon rank sum test and visualized via the R package. The metabolome data were analysed via the OPLS-DA model (ropls package, v1.6.2). The criteria for screening differentially abundant metabolites were a VIP > 1.0 and P < 0.05. Pathway enrichment analysis was performed via the KEGG database. The significance of the pathway was subjected to a hypergeometric test, and FDR correction was performed [43]. Finally, a microbial–metabolite–phenotype correlation network was constructed via Spearman correlation analysis (|r| > 0.6, P < 0.05), visualized with the ComplexHeatmap R package [44]. Significant differences: ^^P < 0.001; ^^P < 0.01; ^^P < 0.05.
Results
Effects of Chinese herbal medicine residues on biochemical indicators, antioxidant capacity, and immunity in sheep serum after transportation
In this study, we first evaluated the effects of short-distance transport on sheep and the mitigating role of traditional Chinese medicine residues. Compared with pretransportation levels, serum SOD and ROS concentrations significantly increased on d 1 post-transportation (P < 0.05, Fig. 1A). As an oxidative stress marker, the significant elevation of ROS on d 1 post-transportation indicates the occurrence of transportation stress. Further comparison of oxidative stress-related indicators between the CON group and the CMR group was conducted on d 3 post-transportation. Compared with the CON group, at 3 d post-short-distance transportation, the CMR group presented significantly elevated GSH-Px levels (P < 0.05) and significantly reduced ROS levels (P < 0.05) in the serum, whereas T-AOC, SOD, and MDA levels were not significantly different between the groups (Fig. 1B). On d 7 post-short-distance transportation, no significant differences were observed in the serum T-AOC, SOD, GSH-PX, MDA, or ROS levels between the two groups (P > 0.05, Additional Fig. S1A). On the basis of these interim findings, subsequent studies focused on the d 3 post-transport time point to further investigate the mechanism by which supplemental Chinese herbal residue alleviates oxidative stress in sheep. On d 3 post-transport, serum immune and liver function indicators were compared between the CON group and the CMR group. Compared with the CON group, the CMR group presented significantly increased serum IgG levels (P < 0.05) and significantly decreased LDH levels (P < 0.05). However, no significant differences were observed in the serum ALB, ALT, AST, TNF-α, or ALP levels between the CON group and the CMR group (P > 0.05, Fig. 1C).Fig. 1. Differences in oxidative stress-related indicators before and after transportation and differences in oxidative stress-related indicators and liver function indicators between the CON group and CMR group after supplementation with Chinese herbal residue. A Changes in oxidative stress-related indicators before and after transportation. B Intergroup differences in oxidative stress-related indicators between the CON group and CMR group on d 3 after supplementation with Chinese herbal residue. C Intergroup differences in liver function and immunity-related indicators between the CON group and CMR group after supplementation with Chinese herbal residue. ^*^P < 0.05
Further investigation of the effects of traditional Chinese medicine residue supplementation on health status and rumen fermentation in sheep following transport stress was performed. On d 3 post-transport, the serum CHO, GLU, LDL-C, HDL-C, TG, and NEFA concentrations did not significantly differ between the CMR group and the CON group (P > 0.05, Additional Fig. S1B). Moreover, there was no significant difference in the abundance of VFAs or NH_3_-N between the two groups on d 3 post-transportation (P > 0.05, Additional Fig. S1C, D).
Differences in plasma metabolites between the CON group and the CMR group
After stringent quality screening and identification, a total of 1,312 metabolites were identified from the two groups of sheep plasma samples. As shown in Fig. 2, all samples in the OPLS-DA score plot fell within the 95% Hotelling T^2^ ellipse, with a significant separation trend between the two groups (Fig. 2A and B). Following t-tests and variable importance in projection (VIP) screening of blood metabolite relative concentrations (P < 0.05 and VIP > 1), 54 differentially expressed metabolites were identified between the CON group and the CMR group (P < 0.05, Fig. 2C). These included 34 upregulated metabolites (e.g., Elatoside G, Phosphorylcholine, Norgalanthamine) and 20 downregulated metabolites (e.g., Pelargonidin 3-sophoroside, Isoleucyl-Isoleucine, Leucylleucine). Classification annotation via the HMDB grouped these metabolites into 9 superclass categories: lipids and lipid-like molecules: 23 (44.23%), benzenoids: 7 (13.46%), organic nitrogen compounds: 6 (11.54%), organic acids and derivatives: 5 (9.62%), organic oxygen compounds: 3 (5.77%), organoheterocyclic compounds: 3 (5.77%), phenylpropanoids and polyketides: 3 (5.77%), alkaloids and derivatives: 1 (1.92%), and nucleosides, nucleotides, and analogues: 1 (1.92%) (Fig. 2D). To better explore the mechanisms underlying the metabolic pathway changes between the two groups, pathway enrichment analysis was performed on the 54 differentially expressed metabolites. The differentially expressed metabolites from both groups were enriched in eight distinct metabolic pathways: glycerophospholipid metabolism, alpha-linolenic acid metabolism, linoleic acid metabolism, caffeine metabolism, arachidonic acid metabolism, purine metabolism, nucleotide metabolism, and drug metabolism–cytochrome P450. Among these, PC (15:0/18:1(9Z)) was significantly enriched in glycerophospholipid metabolism, alpha-linolenic acid metabolism, linoleic acid metabolism, and arachidonic acid metabolism. Phosphorylcholine (3Z,6Z)-3,6-nonadienal, and xanthosine were significantly enriched in glycerophospholipid metabolism, alpha-linolenic acid metabolism, and caffeine metabolism, respectively (Fig. 2E).Fig. 2. Metabolomics analysis of sheep plasma from the CON group and CMR group and characteristics of differential metabolites. A OPLS-DA cation score plot of sheep plasma samples from the two groups. B OPLS-DA anion score plot of sheep plasma samples from the two groups. C The Volcano plot of differential metabolites between the two groups. D Proportion of differential metabolites by category. E KEGG pathway enrichment analysis of differential metabolites and corresponding Sankey diagram
Identification and functional prediction of the active components of TCM residues
On the basis of data from the ETCM database (ETCM v2.0, http://www.tcmip.cn/ETCM2/front/#/), as shown in Fig. 3A, the herbal residue of the Siji Antiviral Mixture comprises 11 herbal ingredients, Houttuynia cordata, Platycodon grandiflorus, Morus alba leaf, Forsythia suspensa, Mentha haplocalyx, Schizonepeta tenuifolia, and the Phragmites communis rhizome. These herbs collectively contain 1,592 known active compounds, such as coflodiol, afzelin, platyeoside E, D-mandelonitrile, and liquoridin. In total, 508 associated target proteins, including CAT, NFKB1, TJP1, STAT3, and TLR4, were identified (Fig. 3A). KEGG pathway enrichment analysis of these target proteins revealed that these target proteins were enriched in multiple metabolic pathways, including arachidonic acid metabolism, arginine and proline metabolism, arginine biosynthesis, ascorbate and aldarate metabolism, and carbon metabolism (Fig. 3B). Furthermore, a comparison between the metabolic pathways enriched in the plasma metabolome (Fig. 2D) and those targeted by herbal residue components (Fig. 3B) revealed that both the active compounds in the herbal residue and the plasma metabolome were enriched in the arachidonic acid metabolism and drug metabolism–cytochrome P450 pathways (Fig. 3C).Fig. 3. Analysis of the active components, targets, and pathways of the Chinese herbal residue of the Siji Antiviral Mixture and plasma metabolites correlation analysis. A Core component and target protein interaction network diagram. B Functional prediction analysis and corresponding relationship Sankey diagram. C CMR and plasma metabolomics pathway Venn diagram
Differences in the microbial composition of sheep rumen fluid between the CON group and the CMR group
To investigate the changes in the rumen microbiome during sheep transportation stress following supplementation with Chinese herbal residue, we performed metagenomic sequencing on rumen fluid samples collected 3 d after transport.
As shown in Fig. 4A, B and Supplementary Fig. S2C, D, the richness (Chao1 index) and diversity (Shannon index) of the rumen bacteria, fungi, and archaea did not significantly differ between the two groups. PCoA at the species level revealed that there was a significant difference in the β diversity of the rumen bacteria between the two groups (Fig. 4C), whereas no significant separation was detected for rumen fungi or archaea (Additional Fig. S2E, F). Therefore, subsequent analyses focused on rumen bacteria. At the genus level, the five most abundant rumen bacterial genera were Prevotella, unclassified Bacteroidales, unclassified Clostridia, unclassified Lachnospiraceae and unclassified Lachnospiraceae. At the species level, the five most abundant rumen bacterial species were Bacteroidales bacterium, Prevotella sp., Clostridia bacterium, Lachnospiraceae bacterium and Oscillospiraceae bacterium (Fig. 4D). The Wilcoxon signed-rank test identified three genera and seven bacterial species with significantly different abundances between the two groups: Schwartzia, Selenomonas, Anaerovibrio, Parabacteroides, *Schwartzia *sp., *Selenomonas *sp., Schwartzia succinivorans, Selenomonas ruminantium, Anaerovibrio sp., Anaerovibrio lipolyticus and *uncultured Anaerovibrio *sp. (P < 0.05, Fig. 4E, Fig. S2B).Fig. 4. The effect of Chinese herbal residue supplementation on the rumen bacterial community structure of transportation-stressed sheep. A Chao1 index at the bacterial species level between the two groups. B Shannon index at the bacterial species level between the two groups. C β-diversity PCoA analysis of bacterial community structure between the two groups. D Top ten bacterial species by relative abundance. E Differing bacterial species between the two groups
Differences in the functional composition of the rumen microbiome
The results revealed that there were no significant differences in the richness (Chao1 index) or diversity (Shannon index) of KOs between the two groups (P > 0.05, Fig. 5A). PcoA analysis of the KO levels between the two groups revealed that there was no significant difference in the β diversity of the KOs between the two groups (P > 0.05, Fig. 5B). Wilcoxon rank-sum tests identified 538 differentially expressed KOs (P < 0.05), including K01139 (spoT), K00059 (fabG), K01154 (hsdS), K09809 (tagF), K02335 (polA), K07485, K00876 (udk), K03695 (clpB), K03168 (topA), and K03495 (gidA) (Fig. 5C). Additionally, 20 metabolic pathways, such as 2-oxocarboxylic acid metabolism, porphyrin metabolism, butanoate metabolism, histidine metabolism, and valine, leucine, and isoleucine biosynthesis, were differentially regulated (P < 0.05, Fig. 5D).Fig. 5. Regulatory effects of supplementation with Chinese herbal residues on rumen microbial functional pathways and CAZymes activity in sheep. A Differences in the KEGG functional Chao1 index and Shannon index between the two groups. B PCoA of differences in bacterial community functional β diversity. C Differences in the KO between the two groups. D Differences in pathway level 3 between the two groups. E Differences in the Chao1 index and Shannon index at the CAZymes family level between the two groups. F PCoA of β diversity in CAZymes between the two groups. G Differences in CAZymes between the two groups. H Analysis of the contribution of different bacterial species to CAZymes activity
A total of 533 CAZyme-encoding genes were annotated in the rumen, which were classified into 16 auxiliary activities (AAs), 64 carbohydrate-binding modules (CBMs), 15 carbohydrate esterases (CEs), 268 glycoside hydrolases (GHs), 92 glycosyltransferases (GTs), and 77 polysaccharide lyases (PLs). α-Diversity analysis (Fig. 5E) revealed a significant difference in CAZyme diversity (Shannon index) at the family level between the two groups (P < 0.05), whereas richness (Chao1 index) was not significantly different (P > 0.05). PCoA analysis (Fig. 5F) revealed no significant difference in CAZyme β diversity (P > 0.05). On the basis of the Wilcoxon rank sum test, a total of 21 family-level differential Cazymes were identified (P < 0.05), including 5 GTs, 9 GHs, 4 CBMs, 2 PLs, and 1 AA, such as GT2_Glycos_transf_2, GT4, GT2_Glyco_tranf_2_3, GH13_14, GT32, AA7, GH113, GH84, CBM62, and GT110 (Fig. 5G). Analysis of species and functional contributions (Fig. 5H) revealed that among the identified differential bacteria, *Schwartzia *sp., *Selenomonas *sp., Schwartzia succinivorans, and Selenomonas ruminantium contributed relatively large amounts to the changes in the relative abundances of GTs, CEs, GHs and AAs and may be the driving force. The main reasons for the differential changes in CAZymes include GT2_Glycos_transf_2, GT4 and GT2_Glyco_tranf_2_3.
Differences in rumen fluid metabolites between the CON group and the CMR group
After stringent quality screening and identification, a total of 2,774 metabolites were identified from the rumen fluid samples of the two sheep groups. All samples in the OPLS-DA score plot fell within the 95% Hotelling T^2^ ellipse, with a clear separation trend between the two groups (Fig. 6A and B). Following t-tests and variable importance in projection (VIP) screening of rumen fluid metabolite relative concentrations (P < 0.05 and VIP > 1), 260 differentially expressed metabolites (P < 0.05) were identified between the CMR and CON groups (Fig. 6C), including 107 upregulated metabolites, such as (R)-glabridin, diosmin, and hesperidin, and 153 downregulated metabolites, including adipamide, lupenone, and aminopropylcadaverine. Classification annotation via the HMDB revealed that these metabolites were annotated into 9 superclass categories: lipids and lipid-like molecules: 22 (37.93%); organoheterocyclic compounds: 14 (24.14%); benzenoids: 9 (15.52%); organic acids and derivatives: 5 (8.62%); organic oxygen compounds: 3 (5.17%); nucleosides, nucleotides, and analogues: 2 (3.45%); organic nitrogen compounds: 2 (3.45%); and phenylpropanoids and polyketides: 1 (1.72%) (Fig. 6D). To better explore the mechanisms underlying the metabolic pathway changes between the two groups, pathway enrichment analysis was performed on the differentially expressed metabolites. Among the two groups, the differentially expressed metabolites were enriched in 27 distinct metabolic pathways. The five pathways with the greatest importance were alpha-linolenic acid metabolism; valine, leucine and isoleucine metabolism; sphingolipid metabolism; glycerophospholipid metabolism; and glycerolipid metabolism. Among these, 13(S)-hydroperoxylinolenic acid was significantly enriched in alpha-linolenic acid metabolism; (S)-2-aceto-2-hydroxybutanoic acid was significantly enriched in valine, leucine, and isoleucine metabolism; dihydroceramide was significantly enriched in sphingolipid metabolism; and GPEtn (16:1/18:1) and LysoPC (16:1(9Z)/0:0) were significantly enriched in glycerophospholipid metabolism. TG (10:0/8:0/i-17:0) and glyceric acid were significantly enriched in triacylglycerol metabolism (Fig. 6E).Fig. 6. Metabolome analysis and differential metabolite characteristics of rumen fluid from the CON and CMR groups of sheep. A OPLS-DA cation score plot of rumen fluid samples from the two groups. B OPLS-DA anion score plot. C Volcano plot of differential metabolites between the two groups. D Proportion of differential metabolite categories. E KEGG pathway enrichment analysis of differential metabolites and the corresponding Sankey diagram
Differences in metabolic enzymes enriched in the CON group and the CMR group
Further investigation into the antioxidant stress mechanism of the residual herbal material from the Siji Antiviral Mixture is needed. As shown in Fig. 7, five metabolic pathways—glycerophospholipid metabolism, alpha-linolenic acid metabolism, drug metabolism–cytochrome P450, nucleotide metabolism, and purine metabolism—were enriched in both the rumen fluid and the plasma metabolome (Fig. 7A). The sheep rumen microbiome can synthesize and interconvert the rumen metabolites LysoPC (16:1(9Z)/0:0) and GPEtn (16:1/18:1) and the blood metabolite phosphorylcholine (PC) (15:0/18:1(9Z)) during glycerophospholipid metabolism. Differential analysis revealed that among the enzymes involved in glycerophospholipid metabolism, ten enzymes were enriched during conversion: EC 2.1.1.103, EC 2.7.1.32 (CKI1), EC 2.7.1.82 (ETNK), EC 2.7.7.15 (PCYT1), EC 3.1.1.4 (PLA2G), EC 3.1.1.5 (LYPLA1), EC 3.1.1.7 (ACHE), EC 3.1.1.32 (PLA2G16), EC 3.1.4.3 (PLC), and EC 3.1.4.46 (GDE1), among which EC 3.1.4.3 (PLC) was significantly upregulated in the CMR group (P < 0.05, Fig. 7B). In α-linolenic acid metabolism, four enzymes were enriched during the conversion of the rumen metabolite 13(S)-hydroperoxylinolenic acid and the plasma metabolite (3Z,6Z)-3,6-nonadienal: EC 1.3.1.42 (OPR), EC 2.3.1.16 (fadA), EC 3.1.1.32 (PLA2G16), and EC 3.1.1.4 (PLA2G). Among these genes, EC 3.1.1.32 (PLA2G16) was significantly downregulated in the CMR group (P < 0.05, Fig. 7C). In the drug metabolism–cytochrome P450 pathway, three enzymes were enriched: EC 1.1.1.1 (ADH), EC 1.2.1.5 (ALDH3), and EC 2.5.1.18 (GST) (Fig. 7D).Fig. 7. Metabolic pathways mediated by rumen microorganisms in Chinese herbal residues. A Venn diagram showing the intersection of the rumen fluid and plasma metabolic pathways. B Metabolite conversion process and related enzyme differences in the glycerophospholipid metabolism pathway. C Metabolite conversion processes and associated enzyme differences in the α-linolenic acid pathway. D Metabolite conversion processes and associated enzyme differences in drug metabolism and P450 metabolism. ^*^P < 0.05
Construction of an association network of rumen microbiome metabolism-related indicators in the CON group and the CMR group
Spearman correlation network analysis (Fig. 8A) revealed that ROS were significantly negatively correlated with the blood metabolite phosphorylcholine, the rumen metabolite LysoPC (16:1(9Z)/0:0), and bacterial species, including Selenomonas sp., Schwartzia succinivorans, and Selenomonas ruminantium. Selenomonas sp., Schwartzia succinivorans, and Selenomonas ruminantium were significantly positively correlated with the rumen metabolite phosphorylcholine and the blood metabolite LysoPC (16:1(9Z)/0:0). To further elucidate the mechanism by which Chinese herbal residue alleviates transportation stress, Spearman correlation analysis between differentially abundant species and annotated enzymes revealed that Selenomonas ruminantium had a significant positive correlation of EC 3.1.4.3 (PLC) and that Schwartzia succinivorans had a significant positive correlation with EC 2.7.32 (CKI1) (P < 0.05, Fig. 8B). However, no significant correlations were detected between these bacterial species and enzymes related to α-linolenic acid metabolism or drug metabolism, such as cytochrome P450. Further Spearman correlation analysis revealed significant positive correlations between EC glycerophospholipid metabolism EC 3.1.4.3 (PLC) and drug metabolism–cytochrome P450 EC 1.1.1.1 (ADH), as well as between arachidonic acid metabolism EC 3.3.2.10 (EPHX2). Simultaneously, EC 1.1.1.1 (ADH) in drug metabolism–cytochrome P450 revealed significant positive correlations with EC 3.1.1.5 (LYPLA1) and EC 3.1.4.46 (GDE1) in glycerophospholipid metabolism, indicating an association between glycerophospholipid metabolism and both arachidonic acid metabolism and drug metabolism–cytochrome P450 (P < 0.05, Fig. 8C). Furthermore, random forest analysis identified key microbial species. Species importance statistics revealed that Selenomonas ruminantium most effectively distinguished glycerophospholipid metabolism differences between the CON and CMR sheep groups and was a significantly different microorganism between the two groups (Fig. 8D).Fig. 8. Host-microbial multi-omics interaction network analysis. A Correlation network analysis. B Correlation network between microbes and key metabolic pathways. C Correlation analysis of enzymes across metabolic pathways. D Random forest analysis identified the key differential bacteria that effectively distinguished glycerophospholipid metabolism differences between the CON and CMR sheep groups
Discussion
With the development of the sheep industry, the stress response caused by sheep transportation has become an important issue that restricts industrial efficiency and animal welfare. This study systematically revealed the molecular mechanism underlying the effective alleviation of short-distance transport stress in sheep through remodelling of the rumen microbiome–host metabolic axis. This study confirmed that short-distance transportation for 4 h could induce oxidative stress in sheep, which manifested as significantly increased serum ROS levels. Dietary supplementation with Siji Antiviral Mixture herbal residue can effectively relieve the stress state and enhance immune function. Mechanistically, microbial fermentation of herbal residues reshaped the rumen microbiome composition. The resulting microbial metabolites activate the glycerophospholipid metabolic pathway, which in turn modulates host arachidonic acid metabolism and drug metabolism–cytochrome P450—via the "microbiome–metabolite axis", thereby exerting antioxidant effects. Through the establishment of a “microbe–metabolite–phenotype” association network, the strong correlation of Selenomonas ruminantium with glycerophospholipid-metabolizing enzymes (ACHE, PLC) and antioxidant phenotypes was proven, providing new insights for the development of anti-stress feeds based on microbiome regulation.
A previous study revealed that transportation stress can significantly increase serum ROS in goats and ROS not cleared in time can lead to disorders of the antioxidant system and thus induce oxidative stress [2]. Danyer et al. [45] reported increased serum activities of GSH-PX and SOD in sheep after short-distance transportation, indicating the activation of systemic antioxidative responses. According to Pleńkowsk et al. [46] and Jomova et al. [47], under oxidative stress, cells usually increase the activity of SOD in response to the excessive generation of ROS. Moreover, numerous studies have shown that supplementation with Chinese herbal residues in the diet can effectively relieve stress-induced tissue damage and improve the immunity of animals [25]. Consistent with these reports, our results revealed that short-distance transport significantly increased serum ROS and SOD in sheep, indicating the initiation of oxidative stress and a protective antioxidant response. On d 3 post-transport, the serum ROS levels were significantly reduced in the CMR group, whereas the antioxidant GSH-PX levels were markedly elevated, demonstrating that supplementation with Chinese herbal residues alleviates transportation stress in sheep. Additionally, we observed that supplementation with Chinese herbal residue significantly reduced serum LDH levels and markedly increased IgG levels in sheep. IgG plays a crucial role in the immune system and reflects an individual's immune status. LDH level serve as an indicator of tissue damage caused by adverse conditions, with increased LDH activity typically reflecting tissue injury induced by stress responses [48]. However, our results did not reveal significant changes in the serum T-AOC or MDA levels. This may be attributed to the relatively weaker impact of short-distance transportation stress than of long-distance transportation stress, which fails to fully activate the antioxidant system. Consequently, the degree of lipid peroxidation was lower, not yet leading to significant alterations in MDA levels [49].
The Siji Antiviral Mixture as a traditional Chinese herbal formulation, is an anti-inflammatory, anti-bacterial, and anti-viral herbal mixture [50]. Many studies have reported that the active components of Siji Antiviral Mixture, such as chlorogenic acid and rutin, exhibit significant antioxidant and anti-inflammatory effects [51, 52]. Water-extracted herbal residues still contain high levels of active components, such as degradable fibres, polyphenols, and flavonoids, although their specific mechanisms of action remain unclear. Nontargeted metabolomics serves as a powerful tool for detecting dynamic changes in metabolic profiles, providing a direct, rapid, and objective reflection of alterations in host metabolism [53]. Therefore, by comparing the targeted functional pathways predicted from the active components of herbal residue with the enriched pathways from plasma metabolomics in this study, both pathways were found to be enriched in arachidonic acid metabolism and drug metabolism–cytochrome P450. Arachidonic acid metabolites exhibit significant anti-inflammatory and antioxidant stress effects and are capable of inhibiting ROS production, scavenging free radicals, and enhancing endogenous antioxidant enzyme expression [54]. Previous studies have shown that the stress state has a major effect on drug metabolism-active enzymes in the liver and changes the expression of cytochrome P450, thereby affecting drug treatment efficacy and toxicity [55]. When free arachidonic acid is metabolized by the cytochrome P450 pathway, CYP450 cyclooxygenase generates epoxydocosatetraenoic acid (EET), and CYP450 ω-hydroxylase produces 20-hydroxy-eicosatetraenoic acid (20-HETE) and further mediates antioxidant stress responses [56, 57].
Metabolomic analysis of both plasma and rumen fluid revealed significant enrichment in several metabolic pathways, including glycerophospholipid metabolism, α-linolenic acid metabolism, drug metabolism–cytochrome P450, nucleotide metabolism, and purine metabolism. Glycerophospholipid metabolism plays crucial roles in cellular signalling, membrane fluidity, and cell‒cell interactions. Previous studies have shown that oxidative stress can modify intermediates in glycerophospholipid metabolism, leading to the oxidation of glycerophospholipids and thereby affecting membrane integrity and function [58–60]. Kovaničová et al. [61] and Putera et al. [62] suggested that intermediate metabolites of α-linolenic acid metabolism may serve as potential molecular markers of cold stress, whereas supplementation with conjugated linoleic acid (CLA) can increase oxidative stress. In this study, the rumen metabolite LysoPC (16:1(9Z)/0:0) and the plasma metabolite phosphorylcholine—key interconvertible metabolites in the glycerophospholipid pathway—were both significantly upregulated in the CMR group. More importantly, Spearman correlation analysis revealed that both LysoPC (16:1(9Z)/0:0) and phosphorylcholine were significantly negatively correlated with the serum ROS level. Previous studies have reported that LysoPC (16:1(9Z)/0:0) and phosphorylcholine possess known membrane-protective and anti-inflammatory functions. LysoPC maintains cell membrane integrity, and its downregulation is correlated with heat stress damage. Concurrently, phosphorylcholine, a choline precursor, is involved in the synthesis of the antioxidant glutathione [63, 64]. Research indicates that glycerophospholipids serve as the primary source of arachidonic acid on cell membranes. Three major phospholipases (PLA2, PLC, and PLD) can release esterified arachidonic acid, providing the basis for the antioxidant effects of arachidonic acid [65, 66]. This study also revealed significant positive correlations between EC 3.1.4.3 (PLC) in glycerophospholipid metabolism, EC 1.1.1.1 (ADH) in drug metabolism P450, and EC 3.3.2.10 (EPHX2) in arachidonic acid metabolism. In contrast, EC 1.1.1.1 (ADH), which is involved in drug metabolism, and P450 were significantly positively correlated with EC 3.1.1.5 (LYPLA1) and EC 3.1.4.46 (GDE1), which are involved in glycerophospholipid metabolism.
CAZymes are essential for the survival and proliferation of microbial communities. The complexity of polysaccharide structures dictates the diversity of CAZymes required for their degradation [67]. Among CAZymes, GHs hydrolyse glycosidic bonds; GTs catalyze glycosidic bond formation with specific substrates, playing vital roles in carbohydrate synthesis and microbial adaptation [68, 69]; CEs primarily remove polysaccharide ester groups, participating in carbohydrate side-chain degradation and synergistically promoting glycosidic bond cleavage by GHs and PLs; PLs can disrupt glycosidic bonds through complex mechanisms; AAs can participate in lignin modification, breaking down plant antidegradation barriers and accelerating substrate hydrolysis; CBMs lack catalytic activity but anchor CAZymes to substrate surfaces, aiding GHs, PLs, and other hydrolytic enzymes [70–74]. This study revealed that supplementation with medicinal residue from the Siji Antiviral Mixture significantly increased rumen microbial CAZymes diversity, as evidenced by the significantly greater Shannon index for CAZymes in the CMR group than in the control group. Compared with those in the CON group, the abundances of enzymes such as GT2_Glycos_transf_2, GT4, and GH13_14 were significantly lower in the CMR group, whereas the abundances of enzymes such as GH84, GH113, and AA7 were significantly greater. Metagenomic analysis revealed species-level differences, indicating significant alterations in the rumen bacterial composition between the two groups. Compared with those in the CON group, the abundances of seven differentially abundant bacteria, including Selenomonas sp., Schwartzia succinivorans, Selenomonas ruminantium, and *Schwartzia *sp., significantly increased in the CMR group. Notably, the differentially abundant species *Selenomonas *sp., Schwartzia succinivorans, Selenomonas ruminantium, and *Schwartzia *sp. significantly contributed to changes in GT/CE/GH-type CAzyme abundance. This likely explains the substantial alterations in the abundances of enzymes, such as GT2_Glycos_transf_2, GT4, and GH13_14. The increased abundance of these species suggests that these differential species efficiently utilize fibre and polyphenolic compounds in herbal residue, thereby promoting their proliferation within the rumen. Spearman correlation analysis revealed that Selenomonas ruminantium was significantly negatively correlated with serum ROS but positively correlated with LysoPC (16:1(9Z)/0:0) and phosphorylcholine and significantly positively correlated with the key glycerophospholipid metabolism enzymes EC 3.1.1.7 (ACHE) and EC 3.1.4.3 (PLC). Furthermore, random forest analysis revealed Selenomonas ruminantium as a key microorganism for distinguishing glycerophospholipid metabolism differences. Selenomonas ruminantium is a functionally diverse bacterial species in the rumen. Numerous previous studies have shown that it can ferment lactic acid to generate propionate [75]. Selenomonas ruminantium works together with other rumen microbes, and this synergy may further affect the efficiency of glycerophospholipid metabolism. Carbohydrate metabolites from the rumen microbiome may provide precursors for glycerophospholipid synthesis [76]. Liu et al. [77] reported that the abundance of Selenomonas ruminantium in the rumen significantly increased in a treatment group following heat stress in dairy cows. Xia et al. [78] reported that β-sitosterol is a naturally active plant compound. Supplementation with β-sitosterol under high-grain diets could relieve chronic inflammatory stress caused by high-grain feeding, and the proportion of Selenomonas ruminantium in the rumen significantly increased. This may have occurred because supplementation with TCM residues changed the diversity of CAZymes by releasing active molecules and promoting the proliferation of Selenomonas ruminaria, thereby maintaining rumen microbial homeostasis and accelerating the full recovery of metabolic function in the rumen.
This study innovatively revealed the mechanism by which the herbal residue of Siji Antiviral Mixture alleviates transportation stress in sheep by reshaping the rumen microbiome–metabolism axis through multi-omics integration analysis. However, the study is limited because few time points are available for observation, the effectiveness of Selenomonas ruminantium for single-bacteria transplantation is not known, and the biotransformation process of the active components of TCM residues in the rumen is not fully understood. In the future, the stability of the microbiome‒metabolite axis in relieving transport stress can be further evaluated, the function of Selenomonas ruminantium can be verified via FMT or single-bacteria transplantation, the microbial metabolites of TCM residues can be analysed, and key active molecules can be identified.
Conclusion
This study demonstrated that 4 h of short-distance transport can successfully induce oxidative stress in sheep, as manifested by a significant increase in ROS levels. Supplementing diets with Siji Antiviral Mixture herbal residues effectively alleviated transport stress and enhanced immune function, as evidenced by significantly reduced serum ROS and LDH levels alongside markedly elevated GSH-PX and IgG levels. The mechanism involves rumen microbial processing of herbal residues, which reshapes the composition of the rumen microbiota by substantially increasing the abundance of bacteria, including Selenomonas ruminantium. Metabolites produced by these microbial communities drive glycerophospholipid metabolic pathways, indirectly regulating host arachidonic acid metabolism and drug metabolism via the cytochrome P450 pathway through the “rumen microbiota–metabolism axis”, which collectively exerts antioxidant effects.
Supplementary Information
Additional file 1: Fig. S1. Analysis of differences in oxidative stress-related indicators on day 7 after sheep transportation, blood routine indicators on day 3 in the CON and CMR groups, and rumen fermentation parameters. (A) Intergroup differences in oxidative stress-related indicators between the CON group and CMR group on day 7 after supplementation with Chinese herbal residue. (B) Blood routine indicators (CHO, GLU, HDL-C, LDL-C, TG, TP, NEFA). (C) Proportion of rumen fermentation parameters. (D) Ammoniacal nitrogen content. Fig. S2. Differences in fungal and archaeal species levels between the CON group and the CMR group in sheep rumen. (A) Genus-level species composition. (B) Differential bacterial genus. (C) Fungal Chao index and Shannon index between the two groups. (D) Archaeal Chao index and Shannon index between the two groups. (E) Fungal PCA analysis. (F) Archaeal PCA analysis.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Malmuthuge N, Howell A, Arsic N, Prysliak T, Perez-Casal J, Griebel P. Effect of maternal separation and transportation stress on the bovine upper respiratory tract microbiome and the immune response to resident opportunistic pathogens. Anim Microbiome. 2021;3:62. 10.1186/s 42523-021-00123-2.10.1186/s 42523-021-00123-2PMC 845107834538279 · doi ↗ · pubmed ↗
- 2Guo KJ, Xu SF, Yin P, Wang W, Song XZ, Liu FH, et al. Active components of common traditional Chinese medicine decoctions have antioxidant functions. J Anim Sci. 2011;89(10):3107–15. 10.2527/jas.2010-3831.10.2527/jas.2010-383121571894 · doi ↗ · pubmed ↗
- 3Long S, Piao X. Effects of dietary Forsythia suspensa extract supplementation to lactating sows and nursery pigs on post-weaning performance, antioxidant capacity, nutrient digestibility, immunoglobulins, and intestinal health. J Anim Sci. 2021;99(8):skab 142. 10.1093/jas/skab 142.10.1093/jas/skab 142PMC 837204634014312 · doi ↗ · pubmed ↗
- 4Wishart DS, Guo A, Oler E, Wang F, Anjum A, Peters H, et al. HMDB 5.0: The human metabolome database for 2022. Nucleic Acids Res. 2022;50(D 1):D 622–D 631. 10.1093/nar/gkab 1062.10.1093/nar/gkab 1062 PMC 872813834986597 · doi ↗ · pubmed ↗
- 5Antoun J, Goulitquer S, Amet Y, Dreano Y, Salaun JP, Corcos L, et al. CYP 4F 3B is induced by PGA 1 in human liver cells: a regulation of the 20-HETE synthesis. J lipid Res. 2008;49(10):2135–41.10.1194/jlr.M 800043-JLR 20018566475 · doi ↗ · pubmed ↗