Metabolomic Analysis of Multiple Biological Specimens (Feces, Serum, and Urine) by 1H-NMR Spectroscopy from Dairy Cows with Clinical Mastitis

Abstract

Due to huge economic losses to the dairy industry worldwide, mastitis can be considered as one of the most common diseases in dairy cows. This work aimed to study this disease by comparing multiple biological specimens (feces, serum, and urine) from individuals with or without clinical mastitis. This was performed by a single analytical platform, namely ¹H-NMR, through a multi-matrix strategy. Thanks to the high reproducibility of ¹H-NMR, we could characterize 120 molecules across dairy cow feces, serum, and urine. Among them, 23 molecules were in common across the three biofluids. By integrating the results of multi-matrix metabolomics, several pathways pertaining to energy metabolism and amino acid metabolism appeared to be affected by clinical mastitis. The present work wished to deepen the understanding of dairy cow mastitis in its clinical form. Simultaneous analysis of metabolome changes across several key biofluids could facilitate knowledge discovery and the reliable identification of potential biomarkers, which could be, in turn, used to shed light on the early diagnosis of dairy cow mastitis in its subclinical form.

Keywords: 1H-NMR; clinical mastitis; dairy cows; pathway analysis; quantitative metabolome.

Figure 1

Venn diagram showing unique and shared metabolites among dairy cow feces, serum, and urine. Each number represents a zone of intersection, while the numbers in brackets indicate the number of metabolites comprised in the specified zone (see Table S1 for the corresponding list of molecules). The list of core metabolites (zone 5) is shown on the right, divided in amino acids, peptides and analogues (AA), organic acids and derivates (OA), carbohydrates and derivates (CA), nucleosides, nucleotides and analogues (NU), and miscellaneous metabolites (MI).

Figure 2

rPCA model was set up on the basis of the molecules whose concentrations showed significantly distinct between groups. In the scoreplot (a), we use squares (Healthy) and circles (Clinical Mastitis) to show the samples from the two groups, respectively. The median of each samples’ group is indicated by wide, empty circles. Boxplot (b) highlights the position of the samples along PC 1. The loading plot (c) shows the significant relationships between the concentration of each molecule and its importance over PC 1 (p < 0.05).

Figure 3

rPCA model was set up on the basis of the molecules whose concentrations showed significantly distinct between groups. In the scoreplot (a), we use squares (Healthy) and circles (Clinical Mastitis) to show the samples from the two groups, respectively. The median of each samples’ group is indicated by wide, empty circles. Boxplot (b) highlights the position of the samples along PC 1. The loading plot (c) shows the significant relationships between the concentration of each molecule and its importance over PC 1 (p < 0.05).

Figure 4

Volcano plot indicating the changes in the concentrations of metabolites in urine samples from healthy and clinical mastitic dairy cows.

Figure 5

Enrichment analysis based on the biofluids’ metabolites whose concentrations showed significantly different between groups (the cut-off of impact value was above 0.2).