https://orcid.org/0000-0002-4466-9246
Figures
Abstract
Holstein Friesian is the most popular breed of dairy cows worldwide due to its exceptional milk production capabilities. In dairy cow management, the body condition score (BCS) is a useful tool, serving as a reliable indicator of a cow’s nutritional status and overall health. It is determined via a subjective visual and tactile assessment of fat cover and muscle mass. A low BCS is associated with decreased milk production and fertility. While genetic and nutritional factors have previously been associated with BCS, their effects are often moderate. In this study, we compared the fecal microbiome and the untargeted fecal metabolome of normal (BCS ≥ 3, n = 16) and thin (BCS < 3, n = 16) Holstein Friesian dairy cows. The 16S rRNA gene-based metagenomic analysis revealed that thin cows had significantly higher levels of Clostridiaceae, Erysipelotrichales, Erysipelotrichaceae, and Turicibacter, while normal cows had higher levels of Clostridiales_vadinBB60_group, UCG-010, Bacteroidaceae, Ruminococcaceae, Paludibacteraceae, Alistipes, and Bacteroides. The fecal metabolomic analysis showed that key signaling pathways, including the mechanistic target of rapamycin (mTOR), phosphatidylinositol 3-kinase (PI3K)-Akt, and AMP-activated protein kinase (AMPK) pathways, were enriched in thin cows. In addition, a significant correlation was observed between differential microbial taxa and metabolites. Notably, Clostridiaceae and Erysipelotrichaceae species are linked to inflammation, infectious diseases, and conditions such as ruminal acidosis. Additionally, the mTOR, PI3K-Akt, and AMPK pathways are known to be activated by both nutrient deficiencies and inflammation. We propose that, in addition to genetic and nutritional factors, gut microbiome dysbiosis may contribute to subclinical health conditions, such as chronic inflammation and acidosis, which indirectly affect the cow’s BCS. These findings are guiding our ongoing research on the underlying health conditions in thin cows to better understand the role that the gut microbiome plays in the regulation of the body condition.
Citation: Saraphol B, Hinthong W, Chienwichai P, Pumipuntu N, Reamtong O, Srisook T, et al. (2025) Analysis of the fecal microbiome and metabolome in dairy cows with different body condition scores. PLoS ONE 20(3): e0319461. https://doi.org/10.1371/journal.pone.0319461
Editor: Vinod Kumar Yata, Central University of Andhra Pradesh, INDIA
Received: December 4, 2024; Accepted: February 4, 2025; Published: March 10, 2025
Copyright: © 2025 Saraphol et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All sequenced reads generated during this study are available in the Sequence Read Archive (SRA) of the National Center for Biotechnology Information (NCBI) under the accession number PRJNA1173238.
Funding: This work was funded by Chulabhorn Royal Academy (https://www.cra.ac.th/). Fundamental Fund: fiscal year 2022 by National Science Research and Innovation Fund, grant number FRB650039/0240, project number 165430, contract number FF-65/008. The funder had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Holstein Friesian is the most widely recognized and popular breed of dairy cattle in the world. This breed plays a pivotal role in the global dairy industry due to its exceptional milk production capabilities. With careful management and attention to health and nutrition, Holstein Friesian cattle can thrive and contribute significantly to dairy operations [1,2]. Established breeding programs continue to enhance the traits of this breed, making it a cornerstone of modern dairy farming [3–5].
The body condition score (BCS) is a critical tool used in dairy cattle management. It is a reliable indicator of a cow’s nutritional status and overall well-being. The BCS is determined by performing a subjective visual and tactile evaluation of the fat cover and muscle mass on a cow’s body, and it is typically quantified using a scale of 1 to 5 (1 = extremely thin, 5 = excessively fat) [6–8]. Maintaining an optimal BCS is essential for maximizing both production and reproductive performance; a low BCS can lead to reduced milk production and fertility, while an excessively high BCS can increase the risk of metabolic disorders [6,7]. By effectively managing each cow’s BCS throughout the lactation cycle, producers can enhance their herd’s overall efficiency and profitability [6,9,10]. Various factors influence or are associated with the BCS in dairy cows. Several studies have demonstrated moderate heritability of BCS values in Holstein Friesian cows; however, genetic correlation estimates differ across studies [11–14]. Management practices and nutrition also play significant roles in determining a cow’s BCS [8,15,16].
The microbiome and metabolome of dairy cows have gained significant attention in recent years due to their potential effects on animal health, nutrition, and productivity [17–19]. The microbiome and metabolome are closely interconnected. Microorganisms produce a variety of metabolites that can influence the host’s physiology and metabolism [17–19]. In turn, the host’s diet, genetics, and health status can impact the composition and function of the microbiome [18,20,21]. To date, limited comparative studies have been conducted on the microbiome and metabolome of dairy cows with different BCS values. In a study that examined the relationship between the rumen microbiome and metabolome and the BCS in prepartum Holstein Friesian cows, it was found that certain microbial taxa were significantly more abundant in cows with a high BCS compared to those with a low BCS [22]. It was also found that the citrate cycle was significantly enriched, with a notable increase in citrate levels, and this increase was strongly linked to the bacterial genera that were more abundant in cows with high BCS values.
The presence of a balanced and effective rumen microbiome is essential for the health and productivity of dairy cows. By breaking down plant material and producing nutrients, these microbial communities contribute significantly to the host’s BCS [23]. However, the gut microbiome can also influence dairy cow health beyond nutrient consumption. Imbalances in the gut microbiome have been linked to various diseases, including intestinal ulcers, metritis, diarrhea, reticuloperitonitis, and ruminal acidosis, which can negatively impact a cow’s health and overall well-being [19,24].
In this study, we aimed to investigate the fecal microbiome and metabolome of Holstein Friesian cows with normal and low BCSs using 16S rRNA gene sequencing and an untargeted metabolomic approach. By doing so, we sought to extend the analysis of the factors that influence the BCS and thus obtain a deeper understanding and broader perspective of the contributing factors in this breed.
Materials and methods
Sample collection
A total of 32 fecal samples were collected from Holstein Friesian cows across 8 dairy farms in Maha Sarakham province, Thailand. The fecal samples were divided into two groups: those from cows with a low BCS ( < 3.0; n = 16; the thin group), and those from cows with a normal BCS (3.0; n = 16; the normal group). All farms involved in this study were members of a cooperative and followed similar feeding practices. For heifers, the feed consisted of a roughage-to-concentrate ratio of 60:40. The roughages included cassava peels, corn silage, grass, and straw from local areas, while the concentrates were supplied by the cooperative and contained more than 16% protein. All participating farms were certified with Good Agricultural Practices for Dairy Cattle Farms by the National Bureau of Agricultural Commodity and Food Standards, ensuring compliance with uniform feeding and management guidelines. All cows included in this study were non-pregnant heifers aged over one year. The numbers of normal and thin cows from each farm are provided in S1 Table. The sample collection was approved by the Institutional Animal Care and Use Committee of Mahasarakham University (license number: IACUC-MSU-27/2023). The fecal samples were collected in sterile bags by a veterinarian. All samples were stored in an ice box immediately after collection, transferred to the laboratory, and stored at − 80oC until further experimentation. The experimental design is illustrated in Fig 1.
Extraction of DNA from fecal samples
The total DNA of the fecal samples was extracted using a QIAamp PowerFecal Pro DNA Kit (Qiagen, Germany) according to the manufacturer’s instructions. Briefly, each fecal sample was added to a bead beating tube. Buffer was then added, and the tube was vortexed to lyse the cells. DNA was captured using a silica membrane spin column and then washed and eluted. The DNA concentration and purity were measured using a Nanodrop (Thermo Scientific, USA). Samples with a concentration > 50 ng/μL and an OD 260/280 ratio of 1.8–2.0 were selected for sequencing. All DNA samples were stored at − 80oC until sequencing.
Sequencing and bioinformatics analysis
The sequencing of the 16S rRNA gene and bioinformatics analysis were performed using the Illumina Hiseq 2500 platform (Novogene, China). The DNA samples were used as templates for the amplification of the V3–V4 hypervariable region of the 16S rRNA gene by polymerase chain reaction (PCR) using the barcoded 16S rRNA universal primer, 341F primer 5’-CCTAYGGGRBGCASCAG-3’ and 806R primer 5’-GGACTACNNGGGTATCTAAT-3’. The PCR product was ligated with an Illumina adapter, and the libraries were sequenced. The paired-end reads from the sequencing were merged using FLASH (V1.2.11, http://ccb.jhu.edu/software/FLASH/) [25], and fastp software (version 0.20.0) [26] was used to filter the raw tags and obtain clean tags, which were then compared with the references in the Silva database (version 138.1, https://www.arb-silva.de/) [27]. VSEARCH (version 2.15.09) [28] was used to detect and remove chimeric sequences.
Amplicon sequence variants (ASVs) were denoised using DADA2 in QIIME2 (version QIIME2-202005) to obtain the initial ASVs [29], and ASVs with an abundance value < 5 were discarded from the analysis. Species annotation was performed using QIIME2 with the Silva database. Two-sample t-tests and Wilcoxon rank-sum tests were conducted using R software (version 2.15.3) to compare the relative abundances of individual taxa in normal and thin cows, with statistical significance determined at a p-value of 0.05.
Analysis of species richness and diversity
Alpha and beta diversities were calculated using QIIME2. The alpha diversity results were used to determine the bacterial community in each group, with reference to the Shannon and Simpson indices, and the taxonomic richness was indicated by the Chao1 value. Beta diversity was visualized by nonmetric multidimensional scaling (NMDS) based on the Bray–Curtis dissimilarity of the ASV composition. The differences in the bacterial community among the samples from each cow group were calculated by conducting an analysis of similarity using ADONIS functions (p < 0.05) in QIIME2.
Extraction of metabolites from fecal samples
Metabolites were extracted from the fecal samples using a previously described method, with some modifications [30]. Briefly, approximately 100 mg of feces was mixed with ice-cold isopropanol at a 1:4 ratio (w/v). To ensure the complete dissolution of metabolites, the mixture was vigorously shaken for 2 min and sonicated on ice for 5 min. The supernatant was collected after centrifugation (14,700 g, 15 min at 4°C), transferred to a new tube, and dried using a speed vacuum machine (Tomy Digital Biology, Tokyo, Japan).
Metabolite identification
Metabolites were identified using an ultra-high performance liquid chromatography (UHPLC) instrument (Agilent 1260 Quaternary Pump, Agilent 1260 High Performance Autosampler, and Agilent 1290 Thermostatted Column Compartment SL, Agilent Technologies, USA) coupled to a quadrupole time-of-flight mass spectrometry (QTOF-MS) instrument (TripleTOF 5600 + , SCIEX, USA) with DuoSpray ion source electrospray ionization (ESI). For the chromatographic separation, a 1:1 (v/v) mixture of 0.1% formic acid in water (mobile phase A) and 0.1% formic acid in acetonitrile (mobile phase B) was used to resuspend the dried metabolite samples. The samples were kept at 6˚C and automatically injected into the UHPLC system, which was equipped with a C18 reversed-phase column (ACQUITY UPLC BEH, 2.1 × 100 mm, 1.7 µ M, Waters, USA) and operated at a flow rate of 0.3 mL/min at 40°C. Analyst Software (version 1.7, SCIEX) was used to acquire the mass ion chromatograms and mass spectra from the QTOF-MS instrument in both the positive (+ESI) and negative (−ESI) ESI modes. An information-dependent data acquisition process was used to acquire a TOF-MS scan and 10 dependent product ion scans using the high sensitivity mode with dynamic background subtraction. The mass range of the TOF-MS scan was m/z 100–1,000, and the product ion scan was set to m/z 50 − 1,000. Quality control (QC) samples, which were created by pooling equal aliquots of each metabolite sample, were injected before, during (after every fourth sample), and after sample analysis to evaluate the system’s performance.
Metabolite annotation
Metabolites were annotated using the XCMS online platform (version 3.7.1, https://xcmsonline.scripps.edu/landing_page.php?pgcontent=mainPage) [31]. The raw.wiff and.wiff.scan files were uploaded to the XCMS server and later analyzed in a pairwise manner. Regarding feature extraction, the polarity was selected as either positive or negative, depending on the type of data. The parameters for maximal tolerated m/z deviation, second peak width, signal/noise threshold, and minimum difference in m/z were set to 15, 5–20, 6, and 0.01, respectively. For the alignment processes, the allowable retention time duration was set to 5 s, with a 0.5 minimum fraction and a 0.015 width of overlapping m/z. The annotation parameters included a 5 ppm error, a 0.01 m/z absolute error, and an isotopic search for the features and their adduct formations. Regarding the identification processes, the METLIN database [32]was searched, considering common adducts with a 5-ppm error tolerance. The raw metabolomic data are shown in S2 Table.
Metabolomic data preprocessing and analysis
The m/z and peak intensity data of each metabolomic feature were exported from the XCMS platform and statistically analyzed using the MetaboAnalyst online platform (version 5.0, https://www.metaboanalyst.ca/) [33]. Using the Statistical Analysis [one factor] module, the median intensity of each peak was preprocessed with interquartile range filtration, quantile normalization, cube root transformation, and range scaling. In addition, the principal component analysis (PCA) algorithm was used to evaluate the reliability of the metabolite identification system by analyzing all the metabolomic data together with the data from the QC samples. The acceptance criterion for high-quality data was that the QC samples should be clustered in the middle of the PCA plot [34]. Datasets that met this criterion were further analyzed. The metabolomic data were visualized using a partial least squares discriminant analysis (PLS-DA) score plot, a hierarchical clustering heatmap, and a volcano plot. In the multivariate PLS-DA scores plot, the 95% confidence regions are displayed to enable the evaluation of the separation of the metabolomic data across the groups. To further investigate the metabolite patterns in normal and thin cows, a hierarchical clustering heatmap showing the top 50 metabolites was generated with the Euclidean distance measure and Ward clustering. Moreover, a volcano plot was generated to pinpoint the significantly different features (i.e., those with a t-test p-value < 0.05 and fold change value ≥ 2).
Pathway analysis
A pathway analysis of the significantly different metabolites was performed using the STITCH platform (version 5.0, http://stitch.embl.de/) [35]. All putative metabolites with a p-value < 0.05 and a fold change ≥ 2 were added to the server with the “Bos taurus” organism criteria. The condition “medium confidence (0.400)” was selected for the minimum required interaction score, and no more than 10 interactions were shown for the first and second shells.
Correlation analysis
Spearman’s rank correlation test was performed using PAST software [36] to assess the correlation between significantly different microbial taxa and metabolites. Results were considered significant when the p-value was < 0.05.
Results
The fecal microbiomes of normal and thin cows differed
The 16S rRNA gene sequencing generated 6,477,917 raw reads, and there was an average of 202,435 reads per sample. After QC and filtering, 12,031 ASVs were identified. The alpha diversity analysis conducted at the ASV level revealed no significant differences between the samples from normal (BCS ≥ 3) and thin (BCS < 3) Holstein Friesian cows in terms of the Chao1, Shannon, and Simpson indices (p > 0.05) (Fig 2A). However, the beta diversity analysis based on the Bray–Curtis dissimilarity indicated significant differences in the fecal microbiota structure between the two groups (Fig 2B). All sequenced reads generated during this study have been deposited in the Sequence Read Archive (SRA) of the National Center for Biotechnology Information (NCBI) under accession number PRJNA1173238.
(A) Box plots representing the alpha diversity indices (Chao1, Shannon, and Simpson indices). (B) Bray–Curtis nonmetric multidimensional scaling (NMDS) plot based on amplicon sequence variant (ASV) composition.
This was confirmed by ADONIS analysis, which revealed a significant difference in the fecal microbiota structure of dairy cows with different BCS values (p = 0.035). The predominant phyla in the fecal microbiota of both normal and thin cows were Firmicutes, Bacteroidota, Euryarchaeota, and Spirochaetota (Fig 3A). The overall taxonomic composition at the order, family, and genus levels differed between the two groups (Figs 3B and 3D). The two-sample t-tests or Wilcoxon rank-sum tests revealed significant differences in the abundance of certain taxa between normal and thin cows (p < 0.05).
(A) Top 10 most abundant phyla. (B) Top 10 most abundant orders. (C) Top 10 most abundant families. (D) Top 30 most abundant genera.
Thin cows had significantly higher levels of Clostridiaceae, Erysipelotrichales, Erysipelotrichaceae, and Turicibacter compared to normal cows. Conversely, normal cows had significantly higher levels of Clostridiales_vadinBB60_group, UCG-010, Bacteroidaceae, Ruminococcaceae, Paludibacteraceae, Alistipes, and Bacteroides (Fig 4). Taxa that had a relative abundance of less than 0.1% were considered “low abundance” and were not included in the analysis.
Asterisks (*) indicate significant differences based on t-test or Wilcoxon rank-sum tests results (p < 0.05).
The fecal metabolites of normal and thin cows differed
The metabolomic data were subjected to a quality assurance process to ensure the reliability of the MS-based identification of the metabolites. The predefined criterion for this quality assurance was that the QC samples had to cluster together in the PCA plot. As shown in S1 Fig, the QC samples were located close to each other in the middle-left area of the PCA plot. This finding indicated that the metabolite identification procedure was reliable and consistent. After confirming the quality of the procedure, we performed the data analysis.
The PLS-DA score plot showed that fecal metabolites in the samples from the normal group were distinct from those in the samples from the thin group (Fig 5A). This result suggested that the fecal metabolites of the two groups of cows were different. We also generated a hierarchical clustering heatmap of the top 50 metabolites to explore and compare the metabolite profiles of the individual cows in the two groups (Fig 5B). The results showed that there were similarities within the groups and differences between the groups.
(A) Partial least squares discriminant analysis (PLS-DA) score plot showing the separation of the samples from the normal (green dots) and thin (red dots) cows. The ellipses represent the 95% confidence regions. (B) Hierarchical clustering heatmap of the top 50 metabolites. Red indicates features with increased intensity, and blue indicates features with decreased intensity. The purple and green bars indicate the data from normal and thin cows, respectively.
Of the 12,175 features identified using both positive and negative modes, 110 features were considered to be significantly different. Approximately 41% (n = 45) of the features were assigned as putative metabolites (Fig 6A). We then generated a volcano plot to screen for metabolites present at significantly different levels in the two groups. Despite the fact that many features could not be matched to existing compounds in the METLIN database, several interesting putative metabolites were identified, for example, cholacalcioic acid and ikarisoside A (Fig 6B).
(A) The number of metabolomic features identified in the metabolomic analysis of the cow fecal samples. (B) Volcano plot highlighting the significantly different metabolomic features. The red dots represent the features with significantly increased intensity, and the purple dots represent the features with significantly decreased intensity. The dashed lines indicate the fold change and p-value cutoffs.
The top 10 putative metabolites found to be at higher levels in the fecal samples from thin cows compared to those from normal cows are shown in Table 1. Similarly, the top 10 putative metabolites found to be at lower levels in the fecal samples from thin cows compared to those from normal cows are also shown in Table 2.
Signaling pathways were perturbed in thin cows
The significantly different putative metabolites were analyzed further to identify the pathways that were affected in thin cows. Using the STITCH platform, we determined that the identified compounds were enriched in 15 pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database [37] (S3 Table). The protein–protein enrichment p-value was 0.000463, indicating the significance of the pathway analysis. Among the enriched pathways, the top three were signaling pathways: the mechanistic target of the rapamycin (mTOR) signaling pathway, the phosphatidylinositol 3-kinase (PI3K)-Akt signaling pathway, and the AMP-activated protein kinase (AMPK) signaling pathway (Fig 7).
The blue, red, and yellow circles show nodes in the mechanistic target of rapamycin (mTOR), phosphatidylinositol 3-kinase (PI3K)-Akt, and AMP-activated protein kinase (AMPK) signaling pathways, respectively.
Correlation between the fecal microbiota and metabolites
Spearman’s rank correlation test was used to assess the correlation between significantly different microbial taxa and metabolites. A substantial portion (64.5%, 71/110) of the fecal metabolites exhibited significant correlations with the fecal microbial taxa (p < 0.05) (Fig 8). Notably, taxa that were significantly more abundant in thin cows compared to normal cows exhibited inverse correlations with many metabolites. However, the majority of the significantly correlated metabolites remain unidentified, which limited our ability to reconstruct the metabolic networks or pathways from the data.
The rectangular boxes indicate significance (p < 0.05). Blue and red represent positive and negative correlations, respectively.
Discussion
Holstein Friesians are widely recognized for their exceptional milk production, underscoring their importance in global dairy farming [1,2]. The BCS is a key tool for assessing the nutritional status and well-being of dairy cows and ensuring optimal animal health and productivity [6,7]. This study uniquely contributes to understanding BCS variations in Holstein Friesians by highlighting the role of the gut microbiome and metabolome. By identifying significant differences in microbial taxa and metabolic pathways between cows with normal and low BCS, this research provides evidence that microbial dysbiosis and altered metabolic signaling pathways are critical factors influencing BCS. These findings deepen our understanding of the biological processes underlying BCS variations and provide new insights for addressing health challenges in dairy herds. In this study, we analyzed the fecal microbiome and metabolome of normal (BCS = 3) and thin (BCS < 3) Holstein Friesian dairy cows. Alpha diversity which reflects species richness and diversity within each group and beta diversity which assesses differences in community composition between groups were used to evaluate the bacterial community diversity in normal and thin cows. While no significant differences were observed in the alpha diversity, the beta diversity analysis revealed that the microbial community structure significantly differed between the two groups. Compared to normal cows, thin cows had significantly higher levels of Clostridiaceae, Erysipelotrichales, Erysipelotrichaceae, and Turicibacter.
Previous studies have shown that various Clostridiaceae species can cause gastrointestinal diseases in cattle, including diarrhea, enteritis, abomasitis, and hemorrhagic bowel syndrome [38–40]. Furthermore, Clostridiaceae species were found to be more abundant in the rumen of dairy cows with lower milk protein yields and negatively associated with milk protein content [41]. Addressing these microbial imbalances may involve targeted interventions, such as dietary adjustments to promote a healthy gut microbiome or the use of probiotics to reduce the abundance of pathogenic species. For example, incorporating high-fiber feed or specific prebiotics could help create a favorable environment for beneficial microbes, potentially mitigating the negative impacts of Clostridiaceae overgrowth [42–44].
Within the Firmicutes phylum, members of the Erysipelotrichaceae family are commonly found in the cattle digestive tract, and dietary variations have been shown to have a significant impact on the abundance of Erysipelotrichaceae species within the cow’s gut microbiome [45–47]. Most Erysipelotrichaceae species can ferment a diverse array of sugars to produce lactic acid [46,48]. Notably, an increased abundance of Erysipelotrichaceae species has been linked to a higher risk of ruminal acidosis in dairy cows [49], which is caused by substantial increases in lactic acid in the rumen. It is not classified as a distinct disease but rather as a spectrum of ruminal acidity levels. A drop in pH below the normal range significantly affects microbial activity, rumen function, and overall animal productivity and health [50]. This aligns with the observation that thin cows may have experienced dietary imbalances, potentially including high levels of readily fermentable carbohydrates, which could exacerbate the growth of these bacteria and contribute to ruminal acidosis [51,52].
Turicibacter species (phylum Bacillota) reside in the digestive tract of many mammals, including cattle [53,54]. However, the role that Turicibacter species play in cattle remains uncharacterized. A previous study demonstrated the ability of Turicibacter strains to modify bile acids and host lipids; mice inoculated with Turicibacter strains exhibited decreased levels of serum cholesterol, triglycerides, and adipose tissue mass [54]. Therefore, the relatively higher levels of Turicibacter species observed in thin dairy cows in this study may have contributed to altered lipid metabolism, potentially influencing their lower BCS. Furthermore, a previous study identified a correlation between Turicibacter species and increased somatic cell count (SCC) in milk, which serves as a marker of mammary gland inflammation and subclinical mastitis in dairy cows [53,55]. These findings suggest a possible link between Turicibacter abundance and systemic metabolic and inflammatory processes in thin cows, warranting further research to clarify its role and potential impact on dairy cow health [53,55].
Normal cows demonstrated significantly higher levels of Clostridiales_vadinBB60_group, UCG-010, Bacteroidaceae, Ruminococcaceae, Paludibacteraceae, Alistipes, and Bacteroides compared to thin cows. Microbial taxa such as Clostridiales_vadinBB60_group and Ruminococcaceae have been associated with anti-inflammatory effects and healthy metabolic states in various studies. For example, Clostridiales_vadinBB60_group species, though their role in cattle remains unclear, tend to be more abundant in healthy control groups across different species. In mice, higher levels of these bacteria have been linked to reduced obesity, dyslipidemia, and insulin resistance [56]. Similarly, in broilers, these bacteria have been associated with decreased inflammation [57]. In humans, increased levels of Clostridiales_vadinBB60_group have been correlated with lower incidences of coronary artery disease and primary sclerosing cholangitis [56,58].
Other taxa, such as UCG-010, Bacteroidaceae, and Paludibacteraceae, are commonly found in the cattle gut microbiome and are influenced by dietary factors and animal age. These microbes are thought to contribute to healthy digestion and nutrient metabolism [59–65]. By grouping these microbial taxa based on their potential functional roles, this highlights their relevance to anti-inflammatory effects and healthy metabolic states, offering insight into their possible contribution to the better body condition observed in normal cows. Additionally, Alistipes species, which belong to the Rikenellaceae family, have been identified in the fecal microbiota of Holstein cows [66,67]. However, the specific role these species play in dairy cows remains unknown.
Our metabolomic analysis of the fecal samples revealed that the mTOR, PI3K-Akt, and AMPK signaling pathways were enriched in thin cows. The nutrient-sensing mTOR signaling pathway is a conserved mammalian pathway that plays a crucial role in the regulation of various cellular processes, including cell growth, proliferation, and metabolism. It acts as a central hub in a complex signaling network, integrating signals from various sources, such as nutrients, hormones, growth factors, and stress [68]. Previous studies conducted on dairy cows have also highlighted that the mTOR signaling pathway is involved in the modulation of the immune response [69,70], suggesting that this pathway may contribute to nutrient deficiencies and inflammation observed in thin cows. Modulating the mTOR pathway could potentially serve as a biomarker or therapeutic target to enhance nutritional status and immune function, thereby contributing to improved body condition [71,70].
The PI3K-Akt pathway is a crucial signaling pathway that regulates essential cellular functions, including gene expression, protein synthesis, and cell growth and survival. It is activated by various cellular signals and stressors [72]. Notably, research conducted with dairy cows has shown that viral and bacterial infections can trigger the activation of the PI3K/Akt/mTOR pathway [73,74]. Targeting this pathway might provide opportunities to manage infections or inflammation that indirectly affect BCS.
AMPK is a key energy sensor responsible for maintaining cellular homeostasis in eukaryotes. When activated, AMPK stimulates ATP synthesis while suppressing ATP-consuming processes, helping maintain the cellular energy balance. Nutrient deficits and various physiological stressors can elevate AMP/ADP levels, leading to AMPK activation [75]. In dairy cows, AMPK activation results in the oxidative degradation of lipids and the inhibition of fatty acid synthesis [76,77]. Furthermore, the AMPK signaling pathway is involved in alleviating inflammation in dairy cows [78]. Given its role in energy regulation and anti-inflammatory effects, AMPK activation could be explored as a biomarker for identifying cows at risk of poor body condition or as a target for nutritional or pharmaceutical interventions aimed at improving BCS [78,79].
Collectively, the signaling pathways found to be enriched in thin cows play crucial roles in cellular physiological processes and functions and are activated by nutrient deficiencies, which potentially stem from nutrient digestion and absorption issues. These findings suggest that a cow’s BCS may be closely linked to the gut microbial and metabolomic profiles. Notably, all the enriched pathways in thin cows, including mTOR, PI3K-Akt, and AMPK, are associated with immune responses and inflammation. This aligns with the observed higher levels of bacterial taxa in thin cows, such as Clostridiaceae and Erysipelotrichaceae, which are known to trigger inflammatory processes and contribute to conditions like ruminal acidosis. The connection between these microbial taxa and pathways highlights the potential role of inflammatory markers, such as interleukins and cytokines, in mediating the observed physiological changes in thin cows. Future studies could further elucidate this relationship by directly measuring these markers to validate the links between microbial imbalances, pathway activation, and systemic inflammation.
We hypothesize that dysbiosis of the cow’s gut microbiome contributes to subclinical conditions, which indirectly affect BCS. Unlike traditional models of BCS determination that primarily focus on genetic predisposition and dietary intake [14–16], these findings highlight the novel integration of microbiome and metabolome data as critical components in understanding the underlying mechanisms regulating BCS. This approach provides a more comprehensive framework for addressing subclinical health issues and improving body condition in dairy cows. Future research should focus on comparing the BCS, fecal microbiome composition, fecal and serum metabolome, and inflammatory markers (e.g., interleukins and cytokines) of normal and thin cows. Longitudinal studies would be particularly effective in capturing temporal changes and causative relationships between microbiome composition, metabolomic shifts, and inflammatory markers. Additionally, intervention-based studies, such as dietary trials or microbiome transplants, could provide direct evidence of how targeted modifications influence BCS and underlying health conditions. The resultant knowledge will enhance our understanding of the gut microbiome’s role in the development and progression of underlying health conditions and in the regulation of the cow’s body condition.
Conclusions
In this study, we investigated and compared the fecal microbiome and metabolome in normal (BCS = 3) and thin (BCS < 3) Holstein Friesian dairy cows. The results showed that thin cows had significantly higher levels of Clostridiaceae, Erysipelotrichales, Erysipelotrichaceae, and Turicibacter. Notably, Clostridiaceae and Erysipelotrichaceae have been linked to inflammation, infectious diseases, and other health conditions (e.g., acidosis). Our metabolomic analysis revealed that key signaling pathways, including the mTOR, PI3K-Akt, and AMPK signaling pathways, were enriched in thin cows. Activation of these pathways is often associated with nutrient deficiencies and inflammation. We propose that, in addition to genetic and nutritional factors, gut microbiome dysbiosis may contribute to subclinical health conditions that lead to lower BCS values. These findings are guiding our ongoing research on the underlying health conditions in thin cows to better understand the gut microbiome’s influence on the BCS.
Supporting Information
S1 Table. The numbers of normal and thin cows from each farm.
https://doi.org/10.1371/journal.pone.0319461.s001
(XLSX)
S3 Table. The pathway analysis of fecal metabolomic data using STITCH platform and KEGG database.
https://doi.org/10.1371/journal.pone.0319461.s003
(DOCX)
S1 Fig. The principal component analysis (PCA) plot utilized in the quality assurance process.
The quality control (QC) samples (purple dots) clustered tightly together, indicating the quality of the metabolite identification process. The green dots represent data from the normal group, and the red dots represent data from the thin group.
https://doi.org/10.1371/journal.pone.0319461.s004
(TIF)
Acknowledgments
The authors thank all the farms that participated in this project. The authors also thank Kristen Sadler, PhD, from Scribendi (www.scribendi.com) for editing a draft of this manuscript.
References
- 1. McGuffey RK, Shirley JE. Introduction | History of Dairy Farming. Encyclopedia of Dairy Sciences. Elsevier; 2011. pp. 2–11.
- 2. Coffey EL, Horan B, Evans RD, Berry DP. Milk production and fertility performance of Holstein, Friesian, and Jersey purebred cows and their respective crosses in seasonal-calving commercial farms. J Dairy Sci. 2016;99(7):5681–9. pmid:27132092
- 3.
Houghton FL. Holstein Friesian Cattle: A History of the Breed and Its Development in America. CreateSpace Independent Publishing Platform; 2017.
- 4. O’Sullivan M, Butler ST, Pierce KM, Crowe MA, O’Sullivan K, Fitzgerald R, et al. Reproductive efficiency and survival of Holstein-Friesian cows of divergent Economic Breeding Index, evaluated under seasonal calving pasture-based management. J Dairy Sci. 2020;103(2):1685–700. pmid:31837792
- 5. O’Sullivan M, Shalloo L, Pierce KM, Buckley F. Economic assessment of Holstein-Friesian dairy cows of divergent Economic Breeding Index evaluated under seasonal calving pasture-based management. J Dairy Sci. 2020;103(11):10311–20. pmid:32952014
- 6. Coelho K, Machado P, Coldebella A, Meyer P, Cassoli L, Rodrigues P. Factors affecting milk yield at peak and during current lactation of Holstein cows. J Anim Feed Sci. 2004;13(Suppl. 1):475–8.
- 7. Berry DP, Macdonald KA, Penno JW, Roche JR. Association between body condition score and live weight in pasture-based Holstein-Friesian dairy cows. J Dairy Res. 2006;73(4):487–91. pmid:16827949
- 8. Roche JR, Friggens NC, Kay JK, Fisher MW, Stafford KJ, Berry DP. Invited review: Body condition score and its association with dairy cow productivity, health, and welfare. J Dairy Sci. 2009;92(12):5769–801. pmid:19923585
- 9. Curone G, Filipe J, Cremonesi P, Trevisi E, Amadori M, Pollera C, et al. What we have lost: Mastitis resistance in Holstein Friesians and in a local cattle breed. Res Vet Sci. 2018;116:88–98. pmid:29223308
- 10. Erickson PS, Kalscheur KF. Nutrition and feeding of dairy cattle. Animal Agriculture. 2020:157–80.
- 11. Berry DP, Buckley F, Dillon P, Evans RD, Rath M, Veerkamp RF. Genetic relationships among body condition score, body weight, milk yield, and fertility in dairy cows. J Dairy Sci. 2003;86(6):2193–204. pmid:12836956
- 12. Dechow CD, Rogers GW, Klei L, Lawlor TJ. Heritabilities and correlations among body condition score, dairy form and selected linear type traits. J Dairy Sci. 2003;86(6):2236–42. pmid:12836961
- 13. Pryce JE, Harris BL. Genetics of body condition score in New Zealand dairy cows. J Dairy Sci. 2006;89(11):4424–32. pmid:17033031
- 14. Jagusiak W, Ptak E, Otwinowska-Mindur A, Zarnecki A. Genetic relationships of body condition score and locomotion with production, type and fertility traits in Holstein-Friesian cows. Animal. 2023;17(6):100816. pmid:37172357
- 15. Bewley JM, Schutz MM. An Interdisciplinary Review of Body Condition Scoring for Dairy Cattle. The Professional Animal Scientist. 2008;24(6):507–29.
- 16. da Costa ANL, de Araujo AA, Valmir J. Particularities of Bovine Artificial Insemination. Artificial Insemination in Farm Animals. 2011.
- 17. Xue M-Y, Sun H-Z, Wu X-H, Liu J-X, Guan LL. Multi-omics reveals that the rumen microbiome and its metabolome together with the host metabolome contribute to individualized dairy cow performance. Microbiome. 2020;8(1):64. pmid:32398126
- 18. Russo N, Floridia V, D’Alessandro E, Lopreiato V, Pino A, Chiofalo V, et al. Influence of olive cake dietary supplementation on fecal microbiota of dairy cows. Front Microbiol. 2023;14:1137452. pmid:37206333
- 19. Casaro S, Prim JG, Gonzalez TD, Cunha F, Bisinotto RS, Chebel RC, et al. Integrating uterine microbiome and metabolome to advance the understanding of the uterine environment in dairy cows with metritis. Anim Microbiome. 2024;6(1):30. pmid:38802977
- 20. Li Y, Ma N, Ren L, Wang M, Hu L, Shen Y, et al. Microbiome-Metabolome Responses in Ruminal Content and Feces of Lactating Dairy Cows With N-Carbamylglutamate Supplementation Under Heat Stress. Front Vet Sci. 2022;9:902001. pmid:35812889
- 21. Wang Q, Cui Y, Indugu N, Loor JJ, Jiang Q, Yu Z, et al. Integrated meta-omics analyses reveal a role of ruminal microorganisms in ketone body accumulation and ketosis in lactating dairy cows. J Dairy Sci. 2023;106(7):4906–17. pmid:37296048
- 22. Du D, Wang Y, Gao Y, Feng L, Zhang Z, Hu Z. Analysis of differences in the rumen microbiome and metabolic function in prepartum dairy cows with different body condition scores. Anim Microbiome. 2024;6(1):35. pmid:38915057
- 23. Mizrahi I, Wallace RJ, Moraïs S. The rumen microbiome: balancing food security and environmental impacts. Nat Rev Microbiol. 2021;19(9):553–66. pmid:33981031
- 24. Xu Q, Qiao Q, Gao Y, Hou J, Hu M, Du Y, et al. Gut Microbiota and Their Role in Health and Metabolic Disease of Dairy Cow. Front Nutr. 2021;8:701511. pmid:34422882
- 25. Magoč T, Salzberg SL. FLASH: fast length adjustment of short reads to improve genome assemblies. Bioinformatics. 2011;27(21):2957–63. pmid:21903629
- 26. Chen S, Zhou Y, Chen Y, Gu J. fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics. 2018;34(17):i884–90. pmid:30423086
- 27. Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590-6. pmid:23193283
- 28. Rognes T, Flouri T, Nichols B, Quince C, Mahé F. VSEARCH: a versatile open source tool for metagenomics. PeerJ. 2016;4:e2584. pmid:27781170
- 29. Bolyen E, Rideout JR, Dillon MR, Bokulich NA, Abnet CC, Al-Ghalith GA, et al. Reproducible, interactive, scalable and extensible microbiome data science using QIIME 2. Nat Biotechnol. 2019;37(8):852–7. pmid:31341288
- 30. Erben V, Poschet G, Schrotz-King P, Brenner H. Evaluation of different stool extraction methods for metabolomics measurements in human faecal samples. BMJ Nutr Prev Health. 2021;4(2):374–84. pmid:35028509
- 31. Gowda H, Ivanisevic J, Johnson CH, Kurczy ME, Benton HP, Rinehart D, et al. Interactive XCMS Online: simplifying advanced metabolomic data processing and subsequent statistical analyses. Anal Chem. 2014;86(14):6931–9. pmid:24934772
- 32. Smith CA, O’Maille G, Want EJ, Qin C, Trauger SA, Brandon TR, et al. METLIN: a metabolite mass spectral database. Ther Drug Monit. 2005;27(6):747–51. pmid:16404815
- 33. Pang Z, Chong J, Zhou G, de Lima Morais DA, Chang L, Barrette M, et al. MetaboAnalyst 5.0: narrowing the gap between raw spectra and functional insights. Nucleic Acids Res. 2021;49(W1):W388–96. pmid:34019663
- 34. Godzien J, Alonso-Herranz V, Barbas C, Armitage EG. Controlling the quality of metabolomics data: new strategies to get the best out of the QC sample. Metabolomics. 2014;11(3):518–28.
- 35. Szklarczyk D, Santos A, von Mering C, Jensen LJ, Bork P, Kuhn M. STITCH 5: augmenting protein-chemical interaction networks with tissue and affinity data. Nucleic Acids Res. 2016;44(D1):D380-4. pmid:26590256
- 36.
Hammer DAT, Ryan PD, Hammer Ø, Harper DAT. Past: Paleontological Statistics Software Package for Education and Data Analysis. Palaeontologia Electronica. 2001. Available: http://palaeo-electronica.orghttp://palaeo-electronica.org/2001_1/past/issue1_01.htm.
- 37. Kanehisa M, Goto S. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 2000;28(1):27–30. pmid:10592173
- 38. Cho Y-I, Yoon K-J. An overview of calf diarrhea - infectious etiology, diagnosis, and intervention. J Vet Sci. 2014;15(1):1–17. pmid:24378583
- 39. Simpson KM, Callan RJ, Van Metre DC. Clostridial Abomasitis and Enteritis in Ruminants. Vet Clin North Am Food Anim Pract. 2018;34(1):155–84. pmid:29421028
- 40. De Jonge B, Pardon B, Goossens E, Van Immerseel F, Vereecke N, Pas ML, et al. Hemorrhagic bowel syndrome in dairy cattle: Gross, histological, and microbiological characterization. Vet Pathol. 2023;60(2):235–44. pmid:36601786
- 41. Xue M-Y, Xie Y-Y, Zhong Y-F, Liu J-X, Guan LL, Sun H-Z. Ruminal resistome of dairy cattle is individualized and the resistotypes are associated with milking traits. Anim Microbiome. 2021;3(1):18. pmid:33568223
- 42. Cull C, Singu VK, Cull BJ, Lechtenberg KF, Amachawadi RG, Schutz JS, et al. Efficacy of Two Probiotic Products Fed Daily to Reduce Clostridium perfringens-Based Adverse Health and Performance Effects in Dairy Calves. Antibiotics (Basel). 2022;11(11):1513. pmid:36358168
- 43. Lu W, Sun H, Xu Z-M, Du Z, Si L, Yuan S, et al. Diagnostic and therapeutic strategy for Clostridium perfringens infection in postpartum dairy cows: a report of 14 cases. Journal of Applied Animal Research. 2022;50(1):350–4.
- 44. Nekrasov RV, Lozovanu MI, Laptev GY, Ilina LA, Yildirim EA, Tyurina DG, et al. Bioactive Feed Additive for the Prevention of Clostridial Disease in High-Yielding Dairy Cattle. Agriculture. 2023;13(4):786.
- 45. de Menezes AB, Lewis E, O’Donovan M, O’Neill BF, Clipson N, Doyle EM. Microbiome analysis of dairy cows fed pasture or total mixed ration diets. FEMS Microbiol Ecol. 2011;78(2):256–65. pmid:21671962
- 46. Deusch S, Camarinha-Silva A, Conrad J, Beifuss U, Rodehutscord M, Seifert J. A Structural and Functional Elucidation of the Rumen Microbiome Influenced by Various Diets and Microenvironments. Front Microbiol. 2017;8:1605. pmid:28883813
- 47. Zhang J, Shang J, Hao Y, Wang Y, Cao Z, Yang H, et al. Growth performance, blood metabolites, ruminal fermentation, and bacterial community in preweaning dairy calves fed corn silage-included starter and total mixed ration. J Dairy Sci. 2023;106(7):4545–58. pmid:37164844
- 48. Yunusbaeva M, Borodina L, Terentyeva D, Bogdanova A, Zakirova A, Bulatov S, et al. Excess fermentation and lactic acidosis as detrimental functions of the gut microbes in treatment-naive TB patients. Front Cell Infect Microbiol. 2024;14:1331521. pmid:38440790
- 49. Golder HM, Rehberger J, Smith AH, Block E, Lean IJ. Ruminal bacterial communities differ in early-lactation dairy cows with differing risk of ruminal acidosis. Front Microbiomes. 2023;2.
- 50. Hernández J, Benedito JL, Abuelo A, Castillo C. Ruminal acidosis in feedlot: from aetiology to prevention. ScientificWorldJournal. 2014;2014:702572. pmid:25489604
- 51. Abdela N. Sub-acute Ruminal Acidosis (SARA) and its Consequence in Dairy Cattle: A Review of Past and Recent Research at Global Prospective. Achievements in the Life Sciences. 2016;10(2):187–96.
- 52. Li W, Larsen A, Fregulia P. Investigating the impact of feed-induced, subacute ruminal acidosis on rumen epimural transcriptome and metatranscriptome in young calves at 8- and 17-week of age. Front Vet Sci. 2024;11:1328539. pmid:38455258
- 53. Liu J, Liu H, Cao G, Cui Y, Wang H, Chen X, et al. Microbiota Characterization of the Cow Mammary Gland Microenvironment and Its Association with Somatic Cell Count. Vet Sci. 2023;10(12):699. pmid:38133250
- 54. Lynch JB, Gonzalez EL, Choy K, Faull KF, Jewell T, Arellano A, et al. Gut microbiota Turicibacter strains differentially modify bile acids and host lipids. Nat Commun. 2023;14(1):3669. pmid:37339963
- 55. Green MJ, Green LE, Schukken YH, Bradley AJ, Peeler EJ, Barkema HW, et al. Somatic cell count distributions during lactation predict clinical mastitis. J Dairy Sci. 2004;87(5):1256–64. pmid:15290974
- 56. Sawicka-Smiarowska E, Bondarczuk K, Bauer W, Niemira M, Szalkowska A, Raczkowska J, et al. Gut Microbiome in Chronic Coronary Syndrome Patients. J Clin Med. 2021;10(21):5074. pmid:34768594
- 57. Liu H, Li X, Zhang K, Lv X, Zhang Q, Chen P, et al. Integrated multi-omics reveals the beneficial role of chlorogenic acid in improving the growth performance and immune function of immunologically stressed broilers. Anim Nutr. 2023;14:383–402. pmid:37635925
- 58. Fu Y, Li J, Zhu Y, Chen C, Liu J, Gu S, et al. Causal effects of gut microbiome on autoimmune liver disease: a two-sample Mendelian randomization study. BMC Med Genomics. 2023;16(1):232. pmid:37789337
- 59. Chuang S-T, Ho S-T, Tu P-W, Li K-Y, Kuo Y-L, Shiu J-S, et al. The Rumen Specific Bacteriome in Dry Dairy Cows and Its Possible Relationship with Phenotypes. Animals (Basel). 2020;10(10):1791. pmid:33019774
- 60. Beaver A, Petersen C, Weary DM, Finlay BB, von Keyserlingk MAG. Differences in the fecal microbiota of dairy calves reared with differing sources of milk and levels of maternal contact. JDS Commun. 2021;2(4):200–6. pmid:36338447
- 61. Hartinger T, Pacífico C, Poier G, Terler G, Klevenhusen F, Zebeli Q. Shift of dietary carbohydrate source from milk to various solid feeds reshapes the rumen and fecal microbiome in calves. Sci Rep. 2022;12(1):12383. pmid:35858964
- 62. Liu K, Zhang Y, Huang G, Zheng N, Zhao S, Wang J. Ruminal bacterial community is associated with the variations of total milk solid content in Holstein lactating cows. Anim Nutr. 2022;9:175–83. pmid:35573096
- 63. Williamson JR, Callaway TR, Lourenco JM, Ryman VE. Characterization of rumen, fecal, and milk microbiota in lactating dairy cows. Front Microbiol. 2022;13:984119. pmid:36225385
- 64. Brulin L, Ducrocq S, Even G, Sanchez MP, Martel S, Merlin S, et al. Characterization of bovine vaginal microbiota using 16S rRNA sequencing: associations with host fertility, longevity, health, and production. Sci Rep. 2024;14(1):19277. pmid:39164272
- 65. Estrada R, Romero Y, Figueroa D, Coila P, Hañari-Quispe RD, Aliaga M, et al. Effects of Age in Fecal Microbiota and Correlations with Blood Parameters in Genetic Nucleus of Cattle. Microorganisms. 2024;12(7):1331. pmid:39065099
- 66. Zhang J, Xu C, Huo D, Hu Q, Peng Q. Comparative study of the gut microbiome potentially related to milk protein in Murrah buffaloes (Bubalus bubalis) and Chinese Holstein cattle. Sci Rep. 2017;7:42189. pmid:28176851
- 67. Brulin L, Ducrocq S, Estellé J, Even G, Martel S, Merlin S, et al. The fecal microbiota of Holstein cows is heritable and genetically correlated to dairy performances. J Dairy Sci. 2024;107(12):11254–68. pmid:39245169
- 68. Li B, Khan MZ, Khan IM, Ullah Q, Cisang Z-M, Zhang N, et al. Genetics, environmental stress, and amino acid supplementation affect lactational performance via mTOR signaling pathway in bovine mammary epithelial cells. Front Genet. 2023;14:1195774. pmid:37636261
- 69. Sipka AS, Chandler TL, Weichhart T, Schuberth H-J, Mann S. Inhibition of mTOR in bovine monocyte derived macrophages and dendritic cells provides a potential mechanism for postpartum immune dysfunction in dairy cows. Sci Rep. 2022;12(1):15084. pmid:36064574
- 70. Sipka A, Weichhart T, Mann S. Pharmacological inhibition of the mTOR pathway alters phenotype and cytokine expression in bovine monocyte-derived dendritic cells. Vet Immunol Immunopathol. 2022;249:110441. pmid:35597229
- 71. Coleman DN, Totakul P, Onjai-Uea N, Aboragah A, Jiang Q, Vailati-Riboni M, et al. Rumen-protected methionine during heat stress alters mTOR, insulin signaling, and 1-carbon metabolism protein abundance in liver, and whole-blood transsulfuration pathway genes in Holstein cows. J Dairy Sci. 2022;105(9):7787–804. pmid:35879168
- 72. Glaviano A, Foo ASC, Lam HY, Yap KCH, Jacot W, Jones RH, et al. PI3K/AKT/mTOR signaling transduction pathway and targeted therapies in cancer. Mol Cancer. 2023;22(1):138. pmid:37596643
- 73. Liu Y, Liu S, Wu C, Huang W, Xu B, Lian S, et al. PD-1-Mediated PI3K/Akt/mTOR, Caspase 9/Caspase 3 and ERK Pathways Are Involved in Regulating the Apoptosis and Proliferation of CD4+ and CD8+ T Cells During BVDV Infection in vitro. Front Immunol. 2020;11.
- 74. Qi M, Geng H, Geng N, Cui Y, Qi C, Cheng G, et al. Streptococcus agalactiae-induced autophagy of bovine mammary epithelial cell via PI3K/AKT/mTOR pathway. J Dairy Res. 2022:1–7. pmid:35388773
- 75. Wu Z, Tian M, Heng J, Chen J, Chen F, Guan W, et al. Current Evidences and Future Perspectives for AMPK in the Regulation of Milk Production and Mammary Gland Biology. Front Cell Dev Biol. 2020;8:530. pmid:32671074
- 76. McFadden JW, Corl BA. Activation of AMP-activated protein kinase (AMPK) inhibits fatty acid synthesis in bovine mammary epithelial cells. Biochem Biophys Res Commun. 2009;390(3):388–93. pmid:19747896
- 77. Li Y, Ding H, Dong J, Ur Rahman S, Feng S, Wang X, et al. Glucagon attenuates lipid accumulation in cow hepatocytes through AMPK signaling pathway activation. J Cell Physiol. 2019;234(5):6054–66. pmid:30478902
- 78. Xu T, Wu X, Lu X, Liang Y, Mao Y, Loor JJ, et al. Metformin activated AMPK signaling contributes to the alleviation of LPS-induced inflammatory responses in bovine mammary epithelial cells. BMC Vet Res. 2021;17(1):97. pmid:33648513
- 79. Li X, Li Y, Ding H, Dong J, Zhang R, Huang D, et al. Insulin suppresses the AMPK signaling pathway to regulate lipid metabolism in primary cultured hepatocytes of dairy cows. J Dairy Res. 2018;85(2):157–62. pmid:29785900