A plethora of studies have contributed to the exploration of the genetic architecture underlying bovine mastitis [12, 45–48], however, much of the intricate mechanisms still remain unexplained. One possible reason is that most previous studies considered single data types, such as RNA transcripts [46, 47], DNA sequence variants [49, 50], non-coding RNA transcripts [51, 52], and DNA methylation [18, 53]. These single-data based studies provide limited insights and assembled only fragmented pieces of the puzzle; which does not consider the interconnectedness and synergistic effects that arise from the integration of multiple data layers. Recently, multi-omics approaches that integrate and analyze diverse genomic data sets simultaneously have been identified as imperative to fully unravel the complexities of the molecular mechanisms of disease conditions and the discovery of reliable biomarkers [54–56]. Hence, this study employed different multi-omics approaches to integrate four omics data sets, including the DNA methylation profile and the transcriptomes of mRNA, lncRNA and miRNA, to deepen our understanding of the molecular architecture of subclinical mastitis. To our best knowledge, this study is the first to integrate more than three omics data sets to unravel the genomic and epigenomic basis of mastitis. We provide deeper insights into the multi-level signatures simultaneously involved in the regulation of mammary gland response to subclinical mastitis. Our data is consistent with previous studies about bovine mastitis that focused on the integration of two omics data sets [26–31, 48] and at the same time provided deeper insights into the molecular signatures of subclinical mastitis.
This study identified the altered activities of biological processes and pathways with the involvement of DNA methylation signatures, genes, lncRNAs and miRNAs, during subclinical mastitis. These multi-omics signatures served as reference for the construction of comprehensive regulatory networks of the crucial biological processes required for mammary gland defense against subclinical mastitis. For example, the “Staphylococcus aureus infection” pathway has been identified through many single-omics studies as a pathway of interest for subclinical mastitis caused by S. aureus [15, 57]. The integration of epigenomic signatures (dMHBs, DELs and DEMs) with DEGs involved in this pathway were revealed in this study, which could provide more information to better understand the underlying regulatory mechanisms (Fig. 7). For instance, the hypo-methylated dMHB chr18:54314202: 54314226 is found at the regulatory region of the upregulated C5αR1, which plays important roles in the inhibition of chemotaxis and phagocyte activation. Moreover, a hypo-methylated dMHB chr16:4555707:4555760 overlapped with the promoter of up-regulated IL10, which is involved in the secretion of cytokine and synthesis of interferon in the inflammation process. The demethylation at regulatory regions has previously been identified as an important regulatory mechanism of the activation of corresponding genes [58–60]. Therefore, the hypo-methylation detected at the regulatory region of these genes (C5AR1 and IL10) may serve as a possible regulatory mechanism of their activation and thereby the corresponding immune processes. Besides, IL10 was also potentially targeted by three differentially expressed non-coding RNAs, including the down-regulated bta.miR.378 and the up-regulated bta-miR.6119p and LOC104974370 (lncRNA) which may post-transcriptionally modulate the expression of IL10 during subclinical mastitis. Notably, the altered expression of IL10 has been associated with subclinical mastitis in previous studies [19, 61, 62], highlighting its crucial role in regulating mammary gland inflammation. Therefore, the epigenetic signatures associated with IL10 may represent the molecular mechanisms regulating its altered expression in response to subclinical mastitis.
Unlike clinical mastitis, where the immune response triggers a visible inflammatory reaction, subclinical mastitis elicits a more subdued and prolonged immune response. The immune system upon recognizing the presence of the pathogen triggers the innate immune response by recruiting immune-related cells, particularly macrophages and neutrophils, into the mammary gland [63, 64]. Consistent with this immune response, many GO terms and KEGG pathways related to leukocyte activities, such as “Leukocyte cell cell adhesion”, “Neutrophil migration”, “Myeloid leukocyte migration”, “Leukocyte mediated immunity” and others (Table S1), were detected with a higher overall active level in subclinical mastitis group compared to healthy control group. However, due to the subdued immune response of the mammary gland and the development of biofilm which provides protection from host immune responses, pathogenic bacteria such as S. aureus are capable of multiplying and colonizing mammary gland tissue leading to long-term infection with no visible symptoms but elevated somatic cell count in milk [65, 66]. Also, when the innate immune response fails to eradicate the invading pathogens, the adaptive immune responses, such as antigen presentation, T cell and B cell activation, and antibody secretion, are activated for further defense of pathogen invasion [67]. Consistent with this, we found that the alterations of multi-omics signatures related to subclinical mastitis were enriched in these processes, such as “T cell activation”, “Regulation of T cell activation”, “Adaptive immune response” and “B cell differentiation” among others (Table S4). It is worth noting that the multi-omics signatures showed concordant contributions to the up-regulation of most immune-related processes and pathways in subclinical mastitis group (Fig. 5). This suggests that the hypo-methylation and enhanced DELs may mediate the up-regulated expression of DEGs thereby contribute to regulate the immune response to subclinical mastitis. The association of epigenetic signatures to these important processes in this study provides a more holistic view of their regulatory networks. For instance, hypo-methylated dMHBs (chr3:8902020–8902136 and chr3:8916776:8916809) and the down-regulated DEMs (bta.miR.3431, bta.miR.2285t, and bta.miR.455.3p) may mediate the up-regulated expression of CD48 which is involved in the communication between target cells and natural killer cell in the “natural killer cell mediated cytotoxicity” pathway (Fig. 6). Moreover, down-regulated bta.miR.143 and bta.miR.205 potentially played roles in the up-regulated expressions of INFG and FAS, respectively, which are involved in the cytokine-cytokine receptor interaction pathway. Consistent with our results, the altered expression of bta.miR.143 and bta.miR.205 have been previously associated with bovine mastitis [52, 68], further suggesting their regulatory roles in subclinical mastitis. Besides, the abundant hyper-methylated dMHBs at the regulatory region of PYK2 and multiple DELs targeting FAS could potentially open new avenues for investigating the underlying regulatory mechanisms governing the abnormal expression of these genes and thereby the active level of the “natural killer cell mediated cytotoxicity” pathway during subclinical mastitis. In a similar manner, we believe that the involvement of multi-omics signatures could provide valuable information to enhance our understanding of regulatory mechanisms underlying the important biological processes and pathways required for the mammary gland defense against subclinical mastitis.
Furthermore, mammary epithelial cells play a role in defending against invading pathogens causing subclinical mastitis by producing antimicrobial peptides and shedding off infected cells [67, 69]. Consistently, we found that the biological processes related to epithelium development, such as “positive regulation of mammary gland epithelial cell proliferation”, “epithelial cell maturation”, “mammary gland epithelium development” and “epithelial structure maintenance” among others (Table S1) showed general higher activity in subclinical mastitis group according to all multi-omics signatures, which may contribute to the epithelial defense against the invasion of S. aureus and S. chromogenes. Nevertheless, we also observed the significant involvement of down-regulated DEGs and DELs which negatively associated with Factor 1 and hyper-methylated dMHBs which negatively associated with Factor 2 in processes related to epithelium development, which may explain the impaired mammary gland homeostasis and functions during the long-term course of subclinical mastitis.
Consistently, the overall lower active level of some metabolic processes, such as “fatty acid catabolic process”, “fatty acid derivative metabolic process” and “long chain fatty acid transport”, etc. (Table S1), were observed in cows with subclinical mastitis compared to healthy cows. This disparity in metabolic activity underscores the metabolic alterations that occur in cows experiencing subclinical mastitis. Moreover, some of the metabolic processes related to milk synthesis, such as “fatty acid metabolic process”, “lipid oxidation”, “fatty acid biosynthetic process” were also found to be significantly enriched by down-regulated DEGs that were positively associated with Factor 1 (Table S4A). This intriguing finding suggests a potential interplay between suppressed gene expression and the modulation of these critical pathways, possibly influencing the reduced milk production observed in cows with subclinical mastitis. Adding a nuanced perspective to this scenario, the negative regulation of certain metabolic processes has been found to exhibit relatively higher activity levels in cows with subclinical mastitis, as highlighted by processes like “negative regulation of lipid metabolic process”, “negative regulation of calcium ion transport”, and “negative regulation of fat cell differentiation” (Table S1). These elevated negative regulatory processes could indicate a concerted effort by the cow's system to counterbalance the metabolic disruptions brought about by the subclinical mastitis condition. Collectively, these observations unveil potential underlying regulatory mechanisms responsible for the compromised milk production performance in cows with subclinical mastitis. The evident altered active level of key metabolic pathways associated with milk synthesis, suggests a complex interplay of molecular events that contribute to the reduced milk production observed in subclinical mastitic cows. By elucidating the involvement of multi-omics signatures in these regulatory dynamics, we may gain a deeper understanding of the genetic and epigenetic basis behind the milk production decline in cows affected by subclinical mastitis, paving the way for targeted interventions and management strategies to mitigate these effects and enhance overall dairy productivity.
Moreover, we report a high number of multi-omics signatures with significant alterations associated with subclinical mastitis, a small subset of which are identified as candidate discriminant signatures of subclinical mastitis. The discriminant signatures described the most variation between subclinical mastitis group and healthy control group and have the potential to be used as biomarkers for improving mastitis control strategies. This study identified five Factors driving the principle variance from one or more omics related to subclinical mastitis. Interestingly, the significant signatures associated with Factors 1 and 2 are involved in biological processes and pathways related to the mammary gland response to subclinical mastitis, highlighting the regulatory roles of Factors 1 and 2. For instance, we observed the significant involvement of the up-regulated DEGs and DELs negatively correlated with Factor 1 and hypo-methylated dMHBs negatively associated with Factor 2 in the immune-related processes. In addition, the combination of Factors 1 and 2 clustered the cows into subgroups consistent with their mammary gland health condition (subclinical mastitis or healthy) (Fig. 5-M). This indicates that the combination of Factors 1 and 2 has the potential to predict if the cows have subclinical mastitis, and could be used to improve the genomic selection program to further improve dairy cows’ resistance to mastitis. The discriminant signatures − 14 dMHBs, 25 DEGs, 18 DELs and 5 DEMs - are highly correlated and able to distinguish healthy cows from sick cows with subclinical mastitis (Fig. 8). Amongst the identified dMHBs signatures, the hypo-methylated dMHB chr3:8902020:8902136 harbored the TSS of up-regulated CD48, whose up-regulated expression has also been detected in E. coli infected mammary gland tissue [70]. The CD48 is an important number of the signaling lymphocyte activation molecule family and participates in the adhesion and activation of immune-related cells [71] involved in communications between target cells and natural killer cells in the “natural killer cell mediated cytotoxicity” pathway. DNA methylation near TSS has been revealed to block the expression of corresponding genes [58], therefore, the hypomethylation of chr3:8902020:8902136 may activate the expression of CD48 during subclinical mastitis. Other identified dMHBs candidates which overlapped with the first intron of genes (TNFAIP8L2, CLIC5, LRRFIP1, TMEM229B, PPP1R12A, IDH2, CORO7, USP47, PSAP, KCTD1, FNBP1, RBMS1, and EHF), are being associated with bovine mastitis for the first time. Three discriminant DEG candidates, including TMEM53, ACOT4 and HDDC3, have also been identified as discriminant signatures for S. aureus subclinical mastitis [15, 32]. Consistent with our finding, ACOT4 has also been identified as a significant gene for chronic subclinical mastitis in Norwegian Red cattle [72]. The up-regulated expression of TMEM53 has been identified in the blood of goats with a potential role in Small ruminant lentiviruses infection [73]. The involvement of these DEGs in subclinical mastitis or other infectious diseases suggests their possible roles in immune responses and highlights their potential as biomarkers for subclinical mastitis which deserves further investigation. Among the five DEM discriminant candidates, bta-miR-99a-5p and bta-miR-499 have been associated with S. aureus mammary infection [74, 75], revealing their roles in subclinical mastitis. To the best of our knowledge, the discriminant signatures identified in our study are being associated to bovine mastitis for the first time. However, through our comprehensive multi-omics integration, we have highlighted their potential utility as biomarkers for enhancing mastitis control strategies. These findings warrant further investigation and validation to ascertain their clinical relevance and applicability in improving dairy cow health and management practices.
Furthermore, we acknowledge the limitations of our study and propose potential solutions and further study directions to address these limitations. The limited sample size for certain omics data pose a constraint, potentially impacting statistical power and generalizability. For instance, miRNA profiles were only available for five out of the ten healthy control samples, while other omics datasets included ten samples for the healthy group. To strengthen the robustness and reproducibility of our findings, our next step is to replicate the study with a larger cohort of cows, incorporating more healthy control cows. By expanding the sample size, we can enhance the reliability and validity of our observations. In addition, while we successfully integrated RNA sequencing data, small RNA sequencing data, and DNA methylation sequencing data from milk somatic cells, the absence of other crucial omics datasets, such as DNA sequence (genome), proteome, metabolome, and microbiome data, limits a comprehensive understanding of subclinical mastitis. To fully unravel the complexities of the disease, future studies that include these additional omics datasets will be indispensable. Furthermore, we recognize the significance of functional studies and validation experiments to bridge the gap between our observational findings and the mechanistic aspects of the immune response to subclinical mastitis. Despite these challenges, our investigation offers valuable insights into the genomic and epigenomic signatures of subclinical mastitis in milk somatic cells. Thus, more extensive studies and functional analyses, are necessary to contribute to a better understanding of the regulatory mechanisms underlying subclinical mastitis and its potential application in dairy cattle management. Our study therefore lays a foundation for future research on the genomic and epigenomic aspects of subclinical mastitis, emphasizing the importance of addressing the limitations and pursuing further investigations to advance our understanding of this important dairy cow disease.