MiRNAs in milk can be used towards early prediction of mammary gland inflammation in cattle

Abstract

Considering the importance of early disease detection for reducing the huge financial and animal welfare impact of bovine mastitis globally, improved tools are urgently needed that can accurately detect early mammary inflammation. MiRNAs have demonstrated value as disease biomarkers, however, their potential for accurately detecting early mammary inflammation has not been examined in detail. To address this, we investigated the association between levels of four inflammation-associated miRNAs (bta-miR-26a, bta-miR-142-5p, bta-miR-146a and bta-miR-223) and CMT scores (0 to 3) obtained from a large number of individual quarter milk samples (n = 236) collected from dairy cows at different lactations (1 to 4). Initial analyses (n = 21 samples) confirmed that the levels of each of bta-miR-142-5p, bta-miR-146a and bta-miR-223 in whole milk were significantly correlated with mRNA levels of known inflammatory markers (HP, TNF, CXCL8 and IL1B) in milk cells (Rho ≥ 0.49, P < 0.005). Subsequent analyses (n = 215 samples) revealed a significant effect of CMT score on each of the four miRNAs analysed (P < 0.0001), characterised by a progressive increase in miRNA levels in milk as CMT score increase from 0 to > 1. Moreover, a significant effect of lactation number (P < 0.01) for bta-miR-26a, bta-miR-142-5p and bta-miR-146a was attributed to higher miRNA levels during lactation 1 than later lactations. Finally, by generating ROC curves we showed that bta-miR-223 and bta-miR-142-5p levels could identify early inflammatory changes in individual quarter milk samples (CMT1) with high accuracy (100% sensitivity, > 81% specificity). Our results provide novel proof of the value of miRNAs as early diagnostic biomarkers of bovine mastitis.

Figure 1

Relative expression levels (Mean ± SE) of bta-miR-26a, bta-miR-142, bta-miR-146a and bta-miR-223 in milk samples with CMT score 0, 1 or > 1 that were collected from individual mammary quarters of cows at different lactations (1, 2, and 3/4). An effect of lactation number was significant for miR-26, miR-142 and miR-146a (P < 0.005), for which miRNA levels were higher during lactation 1 than during subsequent lactations (P < 0.05). For miR-142 and miR-146a, a CMT x lactation number interaction was also detected (P < 0.01); significant differences within CMT scores that were responsible for such interactions are indicated by different letters (a, b).

Figure 2

ROC curves obtained from the best performing miRNA predictor for each of the comparisons CMT 0 vs 1 (miR-223) and CMT 0 vs > 1 (miR-146a).