Mastitis is a common and frequently occurring disease that places a substantial economic burden on dairy farmers. This disease can cause serious harm to dairy cows and reduce the quality of milk (Firth et al., Reference Firth, Kasbohrer, Schleicher, Fuchs, Egger-Danner, Mayerhofer, Schobesberger, Kofer and Obritzhauser2017). Staphylococcus aureus is one of the major pathogens causing mastitis in cows (Anderson and Lyman, Reference Anderson and Lyman2006). S. aureus attaches to host cells and tissues and can evade host immune responses and release toxins that destroy tissue structures (Zbinden et al., Reference Zbinden, Pilo, Frey, Bruckmaier and Wellnitz2015). S. aureus is sensitive to a variety of antibiotics but can also acquire resistance to drugs. Therefore, the study of S. aureus is of profound importance to the pathogenicity of cow mastitis. Despite a substantial amount of research in the past few decades on the mechanism of action of S. aureus-induced mastitis in dairy cows, most recent research has mainly focused on single gene analysis, or the verification of gene function. Thus, comprehensive research utilizing multi-omics analysis and molecular regulation are lacking.
Gene chip technology allows for the simultaneous analysis of expression patterns of tens of thousands of genes in specific tissues or cells, hence providing a broader picture of cellular mechanisms (Piechota et al., Reference Piechota, Korostynski, Ficek, Tomski and Przewlocki2016; Li et al., Reference Li, Ding, Liu and Zhu2017). In this study, a S. aureus mastitis challenge was used to study gene expression profiling using Affymetrix bovine genome expression microarray technology. In addition, Solexa sequencing was used to establish patterns of affected microRNA (miRNA). Subsequently, a miRNA–mRNA regulatory network map was constructed to provide an experimental basis for future studies of mastitis resistance in cows.
Materials and methods
Ethics statement
This project was approved by the Animal Care and Use Committee of the College of Animal Science and Technology, Yangzhou University, China. All experiments were conducted according to the Regulations for the Administration of Affairs Concerning Experimental Animals published by the Ministry of Science and Technology, China, in 2004.
Establishment of S. aureus-induced mastitis
Three healthy mild-lactation Chinese Holstein cows from the dairy farm at Yangzhou University were used. A single colony of S. aureus (number: ATCC29213, gift from the College of Veterinary Medicine, Yangzhou University) was grown to a density of 1 × 107 CFU/ml S. aureus prior to challenge. One mammary gland (quarter) of each cow was inoculated with 5 ml of this suspension xx whilst a contralateral quarter received 5 ml of phosphate buffered saline as control (Pu et al., Reference Pu, Li, Zhang, Chen, Liao, Zhu, Geng, Ji, Mao, Gong and Yang2017).
Pathological test of experimental mastitis
Mammary gland tissue was collected after 24 h. After anesthesia, duplicate 0.5 cm3 biopsies of mammary tissue were acquired form each cow and transferred to 10% formalin, then 1 mm3 pieces of the mammary tissue were transferred into 2.5% glutaraldehyde. The mammary tissue was dewaxed. HE staining was conducted 15 min after washing the samples. The stained sections were removed to ensure that the xylene was not dry. The sections were observed after solidification (Wang et al., Reference Wang, Ma, Gao, Mao, Zheng, Sun and Liu2011).
RNA extraction
Tissue extraction of the total RNA and DNase treatment was done in accordance with manufacturer's instructions using Trizol reagent (Invitrogen Corp., Carlsbad, CA). RNA samples complying with standardized criteria were used for subsequent experimentation. The full procedure is described in Chen et al. (Reference Chen, Luo, Zhang, Ma, Sun, Zhang and Loor2018a, Reference Chen, Xia, Shen, Xu, Arbab, Li, Zhang, Mao and Yang2018b).
Sequencing analysis
The quality control qualified data (MAS 5.0) was normalized using Gene Spring Software 11.0. The differentially expressed genes were analyzed separately using a t-test analysis, pathway analysis, and hierarchical clustering analysis. The aim was to screen and explore S. aureus-associated mastitis-related genes and their biological significance (Liu et al., Reference Liu, Yang, Fu, Zhang, Tang, Guy, Hu, Guo, Xu and Zhang2016; Bayega et al., Reference Bayega, Fahiminiya, Oikonomopoulos and Ragoussis2018).
The miRNA sequencing analysis was done using the single-end 50 bp sequencing format in the Illumina Hiseq 2500 sequencing platform. After annotating all the small RNA fragments, novel uncommented fragments were used for prediction of novel miRNAs and base editing prediction of known miRNAs (Nozaki et al., Reference Nozaki, Sasaki, Fukuda, Isumi, Nakamoto, Onodera and Masutani2018).
RT-qPCR and western blot analysis
The mature miRNA expression level was determined using the S-Poly(T) assay (online Supplementary Table S7). For mRNA, 0.5 µg of the total RNA was synthesized into cDNA using the Prime Script® RT Reagent Kit (Perfect Real time, Takara, Japan). The sequence of primers is listed in online Supplementary Table S8. Relative expression was calculated using 2−△△Ct (Chen et al., Reference Chen, Luo, Zhang, Ma, Sun, Zhang and Loor2018a, Reference Chen, Xia, Shen, Xu, Arbab, Li, Zhang, Mao and Yang2018b).
Cells were obtained and lysed in RIPA buffer (Solarbio, China) prior to western blot. Proteins extracted from cells were separated by SDS-PAGE, transferred to nitrocellulose membranes (Millipore, USA) and probed with the primary monoclonal rabbit anti- IRAK2 (Cell Signaling Technology Kit #4367), and monoclonal mouse anti-β-actin (Proteintech Gronup, 66009-1-IG, China) antibodies, respectively. Signals were detected with the chemiluminescent ECL Western blot system (Pierce, USA).
Cell culture and transfection
Bovine mammary epithelial cells (BMECs) were cultured and fractionated using the procedures previously used for goat mammary cell culture (Chen et al., Reference Chen, Luo, Zhang, Ma, Sun, Zhang and Loor2018a, Reference Chen, Xia, Shen, Xu, Arbab, Li, Zhang, Mao and Yang2018b). The cells were transfected with either the miR-15a mimic (60 nM) or the inhibitor (60 nM) (Invitrogen, USA) using Lipofectamine™ RNAMAX (Invitrogen, USA) according to manufacturer's instructions. The cells were harvested after 48 h of transfection.
Luciferase reporter assay
To generate a reporter construct luciferase assay, a segment containing a miRNA target site in the 3′-UTR of IRAK2 was inserted into the psi-CHECK-2 vector (Promega, USA) between the NotI and Xho sites immediately downstream of the Renilla luciferase gene. The sequence of primers is listed in online Supplementary Table S9. All the constructs were verified by sequencing. (Chen et al., Reference Chen, Luo, Zhang, Ma, Sun, Zhang and Loor2018a, Reference Chen, Xia, Shen, Xu, Arbab, Li, Zhang, Mao and Yang2018b).
Statistical analysis
Statistical analyses were performed with the SPSS 18.0 statistics software package. Data are presented as the mean ± sd (standard deviation) of three independent experiments. Significant differences between the groups were determined using ANOVA, taking *P < 0.05, **P < 0.01 as significant differences.
Results
Pathological changes and histology
After the experimental infection of S. aureus in the mammary glands of the dairy cows, the mammary gland tissues of the control and inoculated quarters were taken for tissue sectioning and HE staining. Microscopic examination of mammary tissue after 400× magnification revealed an intact acinar cavity filled with milk in control quarters. The tissue between acinar cavities was thick and complete, and the milk ducts were normal (online Supplementary Fig. 1a). In contrast, examination of mammary tissue from S. aureus-challenged cows revealed accumulation of large numbers of inflammatory granulocytes in the acinar cavity. The walls of the acinus were thin, even broken or severely deformed, and severe congestion occurred in the lactating ducts underscoring the clinical mastitis condition (online Supplementary Fig. 1b).
Transcriptome analysis
We examined the success of chip hybridization and scanning. The fluorescence intensity images obtained after the hybridization of the chip approximate the hybridization status of the chip: the higher the brightness, the clearer, denser, and better are the hybridization conditions. The fluorescence intensity map of this experiment is high in brightness, dense and uniform, and has good definition. This confirmed that the experimental requirements were achieved and that the hybridization effect is valid (online Supplementary Fig. 2).
Next, we validated the processing of Gene Chip Data. In the case of biological reproducibility (i.e., using statistical analysis to analyze differential expression), volcano maps can be used to visually show genes that are differentially expressed between two groups of samples. These maps display two important metrics in one graph (fold change/P-value). t-test was used to identify genes that were significantly differently expressed between the S. aureus inoculated quarters and control quarters (online Supplementary Fig. 3). In this study, a total of 1230 genes, with a multiple of more than 2-fold, were screened (online Supplementary Table S4), of which 763 genes were upregulated and 537 genes were downregulated (Fc > 2, P < 0.05).
We conducted differential gene expression pathway analysis, using the functional annotation tool to perform pathway analysis of differentially expressed genes. We screened a total of 15 significant pathways related to immunity (online Supplementary Table 1). These pathways were derived from genes with significant differences. The pathways are good functional indicators that can be further studied (online Supplementary Fig. 4).
Next, we used Q-PCR to validate the differentially expressed genes. Eight genes from the differentially expressed gene group were selected for fluorescent quantitative PCR analysis. The results showed that the differential expression of these genes was significant (P < 0.05), and the trend was consistent with the chip results (Fig. 1). The results show that the results of gene chip analysis are credible.
Validation of genes differentially expressed between control quarters and S. aureus inoculated quarters of cows using q-PCR. Eight genes from the differentially expressed gene group were selected for fluorescent quantitative PCR analysis. The results showed that the differential expression of these genes was significant (P < 0.05), and the trend (up-regulation/down-regulation) was consistent with the chip results. Black bars represent the negative control; red bars represent the S. aureus inoculated quarters.
miRNA sequencing analysis
A total of 21.29 M of total reads were obtained from mammary tissue in the non-challenged control quarters. After the high-quality sequences were de-linked and decontaminated, a clean sequence number of 20.85 M was obtained. Most of the sequence lengths were concentrated at 20–24 nt, with a peak length distribution of 49.32% at 21 nt. A total of 18.59 M total reads were obtained from the S. aureus-inoculated quarters After the high-quality sequences were decontaminated, a clean sequence number of 18.50 M (clean reads, online Supplementary Table S10) was obtained. Most of the sequence lengths were concentrated at 21–24 nt, of which the peak length distribution was 45.47% at 22 nt (online Supplementary Table S2).
The expression of miRNAs in control and challenged cows was obtained via comprehensive comparison of the precursor and mature body using the bovine miRbase database (online Supplementary Table S3). In general, the first base pair U at the 5′ end of the miRNA has a strong preference and is resistant to G. The second to fourth bases generally lack U; the other bases usually lack C. Comparing differences between the control and S. aureus-inoculated quarters, a total of 52 differentially-expressed miRNA were detected (Fig. 2, online Supplementary Table S2).
Differentially-expressed miRNA comparing control and S. aureus-inoculated quarters. 52 differentially-expressed miRNA were detected (points above horizontal cut-off).
A subset of 10 miRNAs were selected for verification via Q-PCR technology. Results indicated that the expressions of miR-223, miR-184, and miR-132 was upregulated, and the expressions of miR-196a, miR-205, miR-200b, miR-31, and miR-145 was downregulated (Fig. 3, online Supplementary Table S6).
A subset of 10 differentially-expressed miRNAs were selected for verification using q-PCR technology. Results indicated that the expression of miR-223, miR-184, and miR-132 was upregulated, and the expression of miR-196a, miR-205, miR-200b, miR-31 and miR-145 was downregulated. Black bars represent the negative control quarters; red bars represent the S. aureus inoculated quarters.
Association analysis of miRNA–mRNA expression
miRNAs are a class of evolutionarily conserved non-coding small RNAs that function to regulate gene expression at the translational level. miRNAs play an important role in many signaling pathways, such as development, differentiation, proliferation, and apoptosis, by acting on multiple target genes (Chen et al., Reference Chen, Luo, Zhang, Ma, Sun, Zhang and Loor2018a, Reference Chen, Xia, Shen, Xu, Arbab, Li, Zhang, Mao and Yang2018b). miRNAs can bind to mRNA, and then degrade mRNA or inhibit mRNA's translational expression. Correlation analysis of the mRNA data and miRNA sequencing results revealed a total of 329 pairs of negatively correlated miRNA/mRNA. There were 31 pairs of downregulated/upregulated miRNA/mRNA pairs, and 298 pairs of upregulated/downregulated miRNA/mRNA pair (Fig. 4, online Supplementary Table S5). These differential miRNA–gene pairs may play an important role in S. aureus-induced cow mastitis.
Association-correlation analysis of the mRNA data and miRNA sequencing results revealed a total of 329 pairs of negatively correlated miRNA/mRNA. There were 31 pairs of downregulated/upregulated miRNA/mRNA pairs, and 298 pairs of upregulated/downregulated miRNA/mRNA pairs.
miR-15a specific targeting of IRAK2
miR-15a and IRAK2 were selected for verification of their regulatory relationships. The software Targetscan uses the 3′-UTR mutual assignment, to identify potential regulatory miRNA. We found that IRAK2 has a miR-15a binding site in the 3′-UTR, hence, we speculate that IRAK2 may be a target gene of miR-15a (Fig. 5b). Results revealed that IRAK2 expression was upregulated by miR-15a inhibition, whereas IRAK2 expression was downregulated in cells overexpressing miR-15a (Fig. 5a and d). To demonstrate that miR-15a directly targets IRAK2, we synthesized a 3′-UTR fragment containing the miR-15a targeting site IRAK2. This fragment was cloned into a psi-CHECK2 vector to construct a 3′-UTR reporter plasmid. Luciferase reporter assays revealed that miR-15a overexpression reduced the relative luciferase activity of the wild-type reporter 3′-UTR. However, there was no significant difference in the relative luciferase activity of the mutant reporter gene (Fig. 5c). These results indicate that miR-15a directly targets the IRAK2 mRNA site and plays a negative regulatory role.
miR-15a and IRAK2 were selected for verification of their regulatory relationships. Results revealed that IRAK2 expression was upregulated by miR-15a inhibition, whereas IRAK2 expression was downregulated in cells overexpressing miR-15a. (a) CMECs are transfected with the miR-15a mimic or the inhibitor for 48 h; IRAK2 expression level is quantified by RT-qPCR (n = 6). Red bars represent the negative control; black bars represent the miR-15a mimic or inhibitor. (b) and (c) Target site of miR-15a in IRAK2′-UTR and the construction of the luciferase (Luc) expression vector fuses with the IRAK2 3′-UTR. WT represents the Luc reporter vector with the WT IRAK2 3′-UTR (684 to 691); MU represents Luc reporter vector with the mutation at miR-15a site in IRAK2 3′-UTR. (d) Western blot analysis of IRAK2 expression in the miR-5a mimic and NC treatment experiments. The effect of miR-15a mimics and Inhibitor on IRAK2 protein expression is evaluated by western blot analysis. Total protein is harvested 48 h post-transfection, respectively.
Discussion
Inoculation of individual quarters with pathogens such as S. aureus can be an effective model for the study of mastitis (Allard et al., Reference Allard, Ster, Jacob, Scholl, Diarra, Lacasse and Malouin2013). To ensure that our mastitis model could be successfully constructed and met statistical criteria, we selected three cows from the same dairy farm, under the same feeding and management conditions, of the same age, parity, and weight, with similar milk production and in good body condition. Furthermore, the cows had to have SCC below 100 000/ml and culture negative for bacteria in milk. We determined that 5 ml of a 1 × 107 CFU/ml concentration of bacteriostatic solution was effective for the bacterial challenge.
Our objective was to clarify some of the relevant control mechanisms of mastitis immune responses at the molecular level. miRNAs are known to coordinate some aspects of responses to disease through their action on specific target genes (Jin et al., Reference Jin, Ibeagha-Awemu and Liang2014; Jia et al., Reference Jia, Zhou, Huang, Xu, Jia, Wu, Yao, Wu and Wang2018; Luo et al., Reference Luo, Wang, Mei and Zan2018a). Some miRNAs have an important role in immunoregulation in response to bacterial infections (Lewandowska et al., Reference Lewandowska-Sabat, Boman, Downing, Talbot, Storset and Olsaker2013; Sun et al., Reference Sun, Aswath, Schroeder, Lippolis, Reinhardt and Sonstegard2015; Kung et al., Reference Kung, Ravi, Rowley, Bell, Little and Bateman2017; Luo et al., Reference Luo, Wang and Zan2018b). In the present study, using the S. aureus model we were able to construct a small RNA library from mammary tissue to identify profiles of miRNA that respond to mastitis. Subsequent quantitative analysis of bta-miR-196a, bta-miR-205, bta-miR-200b, bta-miR-31, bta-miR-145, bta-miR-223, bta-miR-184, and bta-miR-132 by RT-PCR confirmed the sequencing results. Thus, we speculate that these differentially expressed miRNAs can be used as markers of mastitis caused by S. aureus and can provide a reference for future studies.
Previous miRNA sequencing investigation of S. aureus mastitis has used between-cow analysis comparing inoculated and control groups (Li et al., Reference Li, Zhang, Liao, Chen, Wang, Zhu, Geng, Ji, Mao, Gong and Yang2015). Instead, we used a paired analysis that provided increased statistical power. Li et al. (Reference Li, Zhang, Liao, Chen, Wang, Zhu, Geng, Ji, Mao, Gong and Yang2015) used the Audic_Claverie (Tino, Reference Tino2009) formula to calculate the P-value. Differential expression miRNAs were selected using P-value <0.05 and TPM (transcripts per kilobase million) difference multiples >2 as threshold. Our research used the paired test algorithm of DESeq2 (Love et al., Reference Love, Huber and Anders2014) software according to the sample pairing test analysis. The trial used one-to-one pairing for differential screening. The selected miRNAs were P-value <0.05 and the TPM difference multiples >2. Also, the previous study (Li et al., Reference Li, Zhang, Liao, Chen, Wang, Zhu, Geng, Ji, Mao, Gong and Yang2015) only tested the expression level of mRNA and screened differentially expressed mRNA. Our research combined the analysis of miRNA and transcripts. Based on the above, we can get more reliable and useful information.
The biological analysis of miRNAs and their target genes can provide a good way for us to study the immunoregulatory process of mastitis caused by the infection of mammary tissue by S. aureus (Zhang et al., Reference Zhang, Li, Liu, Li, Bi and Fang2016). The comparison of miRNA and mRNA expression profiles in S. aureus-inoculated vs. control tissue allowed us to identify key miRNAs and mRNAs pairs (Ramirez et al., Reference Ramirez, Pastar, Jozic, Stojadinovic, Stone, Ojeh, Gil, Davis, Kirsner and Tomic-Canic2018) associated with development of mastitis in dairy cows. The resulting miRNA/mRNA pairs are involved in widely differing biological functions (Zhang et al., Reference Zhang, Li, Xing, Liu, Chen, Yang and Li2018). The miRNA- and transcriptome regulatory association analysis was based on the predicted interaction between the miRNA and the target transcript. We obtained a negative regulatory relationship between miRNAs and specific transcripts and constructed a microRNA negative regulatory network map. Comparing the differences between the control and the S. aureus inoculated quarters, a total of 1230 genes associated with the 52 differentially-expressed mRNA that showed a 2-fold or greater difference in expression were screened (Fc > 2, P < 0.05). We paired the differentially expressed miRNAs with mRNA using miranda software. Among them, 763 genes were upregulated and 537 genes were downregulated. These genes contain the most studied mastitis resistance genes: interleukins and interleukin receptor genes IL6, IL17A, and CXCR1. IL-6 is a multifunctional inflammatory cytokine and is a key member of the body's cytokine network (Medina-Estrada et al., Reference Medina-Estrada, Lopez-Meza and Ochoa-Zarzosa2016). This cytokine plays an important role in the inflammatory response and is a glycoprotein composed of 212 amino acids. IL-6 has a variety of biological functions and plays an important role in lymphocyte, inflammatory and immune responses (Medina-Estrada et al., Reference Medina-Estrada, Lopez-Meza and Ochoa-Zarzosa2016). In this experiment, IL-6 was also significantly upregulated, and it can be speculated that it plays an immunomodulatory role in the course of inflammation in cows infected with S. aureus. CXCR1 is mainly expressed in neutrophils and monocytes and mediates the activation of neutrophils. This cytokine promotes the release of storage enzymes, enhances phagocytosis, and causes the body's local inflammatory response, thereby playing a role in killing of pathogenic bacteria (Mao et al., Reference Mao, Zhu, Li, Chen, Xin, Zhu, Liao, Wang, Zhang and Yang2015). In the present study, the expression of CXCR1 in the S. aureus inoculated quarters was significantly higher compared with the controls. Also confirmed in this study was the previous correlation between CXCR1 and cow mastitis (Pokorska et al., Reference Pokorska, Dusza, Kulaj, Zukowski and Makulska2016). CXCR1 may be an important candidate gene for mastitis resistance in dairy cows.
The IRAK2 gene is a member of the interleukin-1 receptor-associated kinases (IRAKs) family of serine/threonine kinases. Its main role is to bind to interleukin receptor (IL-1R) upon upregulation of the inflammatory cytokine interleukin-1 (IL-1). IRAK2, IRAKM, IRAK-lb, IRAK1c are negatively regulated in the IRAK family for toll-like receptor signaling pathways. Studies have shown that the IRAK2 has a significant negative regulatory function and can reduce the apoptosis of mast cells induced by lipopolysaccharide infusion (Liu et al., Reference Liu, Chen, Huang, Liu, Ji, Hu, Zhu, Zhang and Dong2018). Similarly, miR-15a is involved in a variety of functions that regulate inflammation and immune cell differentiation (Calin et al., Reference Calin, Amelia, Muller, Manuela, Sylwia, Masayoshi, Cristian, Nicola, Ramiro, Rami, Hansjuerg, Stefano, Laura, Liu, Liu, Thomas, Massimo and Carlo2008; Moon et al., Reference Moon, Yang, Zheng and Jin2014). Results from the luciferase reporter assay indicate miR-15a significantly inhibits luciferase activity, suggesting that it acts through the 3′-UTR of IRAK2 to inhibit reporter gene expression. In addition, mutating the potential binding site of miR-15a in the 3′-UTR abolished the inhibitory effect of miR-15a on the 3′-UTR of IRAK2. Together, the data lead us to hypothesize that miR-15a could target and repress IRAK2. As such, it provides a tool for future studies of bovine mastitis.
In conclusion, the miRNA–mRNA regulatory pairs obtained in this study may play an important role in S. aureus-induced cow mastitis. Hence, future studies can be tailored for evaluating the immune response and related control mechanisms of mastitis caused by bacterial infections. That information could be valuable for developing strategies for the treatment of mastitis in dairy cows.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0022029919000785.
Financial support
This research was supported by the National Natural Science Foundation of China (Grant No. 31472067, 31802035, 31872324), China Postdoctoral Science Foundation (Grant No. 2017M621841), Agricultural Science and Technology Independent Innovation Project of Jiangsu Province, China (Grant No. CX (17) 1005).
