Deep phenotyping of dairy cows with different body condition for characterizing their adaptability to the onset of lactation

Abstract

Metabolic and endocrine functions of dairy cows under non- optimal body condition (underconditioned or over-conditioned) exhibit marked abnormalities that result in a significant decrease in milk production. Underconditioned cows (although not specifically addressed in this paper) may not be able to successfully make the transition from late gestation to lactation, resulting in health problems. Underconditioned cows may not be able to mobilize enough energy for maximum lactation and have greater mobilization of muscle mass after calving than optimally conditioned cows. Over-conditioned cows lose a relatively large amount of body fat around calving, have higher circulating concentrations of free FA, BHB, and AcylCN in their circulation, and experience a more pronounced and prolonged NEB in early lactation than optimally conditioned cows. <br/> Based on deep phenotyping, chapter 2 of this thesis deals with metabolic phenotyping of dairy cows with different body condition. Results for BA, AA, and AcylCN derived from targeted serum metabolomics are described in Publications I - IV, along with a detailed phenotype comparing optimally conditioned and over-conditioned dairy cows to elucidate the pathophysiological reasons for increasing metabolic dysfunction due to overconditioning. By comparing the metabolic profiles of optimally conditioned cows with those of over-conditioned cows, we identified important metabolites associated with overconditioning. Machine learning allowed us to move from descriptive studies to a better understanding of metabolic processes. Given the central role of the liver in metabolism, we also studied hepatic mRNA expression of genes involved in FA metabolism (especially FA oxidation) during transition in dairy cows to investigate the molecular mechanisms of overconditioning around parturition. <br/> In the following study, for a deeper understanding of the pathophysiology of overconditioning, a quantitative proteomics approach was used together with bioinformatics analyzes to investigate the changes in the plasma proteome of optimally and over-conditioned dairy cows during the transition period. A large portion of the plasma proteome was strongly influenced by lactation stage and exhibited longitudinal changes, with inflammation, immune responses, and acute phase response being most affected. Overconditioning around calving was found to be associated with changes in signaling pathways related to the acute inflammatory167 response and regulation of complement and coagulation cascades in transition cows. Our results are a first step toward understanding the complexity of longitudinal and individual variations in the plasma proteome of over-conditioned cows. We also compared the serum miRNA profile of optimal and over-conditioned dairy cows during the transition period and performed pathway enrichment analyzes. We found only subtle differences in circulating miRNA in optimally or overconditioned cows, so we could not derive potential biomarkers. However, we observed a longitudinal shift in serum miRNA profiles, independent of body condition. Our knowledge of miRNA pathways was extended to a higher level by identifying the most enriched pathways associated with cell cycle and insulin signaling, glucose and lipid metabolism, and many other pathways involving DE-miRNA. <br/> In Chapter 3, different machine learning approaches were used to analyze metabolic phenotypes and body condition variability in periparturient dairy cows. Accordingly, in one publication, data from a large cohort of multiparous Holstein cows were subjected to cluster analysis to characterize interindividual differences in the relationship between BFT ap and subsequent BFT losses during early lactation. Specifically, a lack of consistency in adaptation responses during lactation was observed in over-conditioned cows, and the over-conditioned cows with the least BFT loss produced less milk than over-conditioned cows with greater BFT loss. In another study, we used various machine learning algorithms and serum metabolomics data to see whether we could distinguish divergent metabolic patterns from apparently similar BCS phenotypes. According to the study, HBCS-PN cows consumed more feed and energy than HBCSPH cows, but produced milk with a higher protein content, resulting in a lower NEB. The results suggest that determining overconditioning based on BCS and BFT alone may not be sufficient to adequately address the needs of all individuals, such as management and feeding.