Multi Platforms Strategies and Metabolomics Approaches for the Investigation of Comprehensive Metabolite Profile in Dogs with Babesia canis Infection

Abstract

Canine babesiosis is an important tick-borne disease worldwide, caused by parasites of the Babesia genus. Although the disease process primarily affects erythrocytes, it may also have multisystemic consequences. The goal of this study was to explore and characterize the serum metabolome, by identifying potential metabolites and metabolic pathways in dogs naturally infected with Babesia canis using liquid and gas chromatography coupled to mass spectrometry. The study included 12 dogs naturally infected with B. canis and 12 healthy dogs. By combining three different analytical platforms using untargeted and targeted approaches, 295 metabolites were detected. The untargeted ultra-high performance liquid chromatography-tandem mass spectrometry (UHPLC-MS/MS) metabolomics approach identified 64 metabolites, the targeted UHPLC-MS/MS metabolomics approach identified 205 metabolites, and the GC-MS metabolomics approach identified 26 metabolites. Biological functions of differentially abundant metabolites indicate the involvement of various pathways in canine babesiosis including the following: glutathione metabolism; alanine, aspartate, and glutamate metabolism; glyoxylate and dicarboxylate metabolism; cysteine and methionine metabolism; and phenylalanine, tyrosine, and tryptophan biosynthesis. This study confirmed that host-pathogen interactions could be studied by metabolomics to assess chemical changes in the host, such that the differences in serum metabolome between dogs with B. canis infection and healthy dogs can be detected with liquid chromatography-mass spectrometry (LC-MS) and gas chromatography-mass spectrometry (GC-MS) methods. Our study provides novel insight into pathophysiological mechanisms of B. canis infection.

Keywords: Babesia canis; chromatography; mass spectrometry; metabolomics; serum.

Figure 1

Venn diagram showing the number of metabolites identified in serum samples of dogs infected with Babesia canis and healthy dogs. The untargeted liquid chromatography-mass spectrometry (LC-MS) metabolomics approach identified 64 metabolites, the targeted LC-MS metabolomics approach (Biocrates analysis) identified 205 metabolites, and 26 metabolites were identified by gas chromatography-mass spectrometry (GC-MS) metabolomics approach. A total of 295 metabolites were detected using all three platforms.

Figure 2

Intensity of metabolites significantly altered in dogs infected with B. canis and analyzed by the untargeted LC-MS metabolomics approach. Data are presented as box and whiskers plot (mean ± SD). All metabolites shown demonstrated a statistically significant difference between groups by t-test (p < 0.05). Blue-control samples, red-disease samples.

Figure 3

Concentrations of metabolites significantly altered in dogs infected with B. canis determined by the targeted LC-MS metabolomics approach. Data are presented as box and whiskers plot (mean ± SD). All metabolites shown demonstrated a statistically significant difference between groups by t-test (p < 0.05). Blue-control samples, red-disease samples.

Figure 4

Concentrations of metabolites significantly altered in dogs infected with B. canis and generated by the GC-MS metabolomics approach. Data are presented as box and whiskers plot (mean ± SD). All metabolites shown demonstrated a statistically significant difference between groups by t-test (p < 0.05). Blue-control samples, red-disease samples.

Figure 5

Principal component analysis (PCA) score plots were performed using 24 serum samples of dogs by untargeted LC-MS metabolomics analysis (a), targeted LC-MS metabolomics analysis (b), and GC-MS metabolomics analysis (c). The first two principal components (PCs) explained 426% for the untargeted LC-MS metabolomics, 387% for the targeted LC-MS metabolomics, and 451% of the total variance for the GC-MS metabolomics.

Figure 5

Principal component analysis (PCA) score plots were performed using 24 serum samples of dogs by untargeted LC-MS metabolomics analysis (a), targeted LC-MS metabolomics analysis (b), and GC-MS metabolomics analysis (c). The first two principal components (PCs) explained 426% for the untargeted LC-MS metabolomics, 387% for the targeted LC-MS metabolomics, and 451% of the total variance for the GC-MS metabolomics.

Figure 6

The heat map of the metabolic features in the untargeted LC-MS metabolomics (a), targeted LC-MS metabolomics (b), and GC-MS metabolomics (c). Red color represents the increased level of each metabolite, while blue color represents the decreased level of each metabolite in dogs infected with B. canis versus control dogs. Blue panel-control samples, red panel-disease samples.

Figure 6

The heat map of the metabolic features in the untargeted LC-MS metabolomics (a), targeted LC-MS metabolomics (b), and GC-MS metabolomics (c). Red color represents the increased level of each metabolite, while blue color represents the decreased level of each metabolite in dogs infected with B. canis versus control dogs. Blue panel-control samples, red panel-disease samples.

Figure 7

Partial least squares-discriminant analysis (PLS-DA) score plots were performed in 24 analyzed serum samples from control group and group of dogs infected with B. canis by untargeted LC-MS metabolomics analysis (a), targeted LC-MS metabolomics analysis (b), and GC-MS metabolomics analysis (c) (left panels). The list of the 15 important compounds/metabolites was identified by PLS-DA according to the variable importance on projection (VIP) score (right panels). The intensity of the colored boxes on the right represents the relative abundance of the corresponding metabolite in each group (blue-control, red-disease).

Figure 7

Partial least squares-discriminant analysis (PLS-DA) score plots were performed in 24 analyzed serum samples from control group and group of dogs infected with B. canis by untargeted LC-MS metabolomics analysis (a), targeted LC-MS metabolomics analysis (b), and GC-MS metabolomics analysis (c) (left panels). The list of the 15 important compounds/metabolites was identified by PLS-DA according to the variable importance on projection (VIP) score (right panels). The intensity of the colored boxes on the right represents the relative abundance of the corresponding metabolite in each group (blue-control, red-disease).

Figure 8

Pathways’ analysis plot of the disturbed metabolic pathways in dogs infected with B. canis compared to healthy dogs (a). Glutathione metabolism, as well as alanine, aspartate, and glutamate metabolism, glyoxylate and dicarboxylate metabolism, cysteine and methionine metabolism, arginine and proline metabolism, arginine biosynthesis, citrate cycle and phenylalanine, tyrosine, and tryptophan biosynthesis were significant pathways with a p value of <0.05 and impact of >0.1. Biological pathways network and summary plot derived from metabolite set enrichment analysis for prediction of pathways associated serum metabolites sets (b). The colors and the bar length represent the metabolites with different levels of significance for enrichment analysis.