Metabolomic analyses reveal lipid abnormalities and hepatic dysfunction in non-human primate model for Yersinia pestis.

ABSTRACT

INTRODUCTION: Pneumonic plague is caused by the aerosolized form of Yersinia pestis and is a highly virulent infection with complex clinical consequences, and without treatment, the fatality rate approaches 100%. The exact mechanisms of disease progression are unclear, with limited work done using metabolite profiling to study disease progression. OBJECTIVE: The aim of this pilot study was to profile the plasma metabolomics in an animal model of Y. pestis infection. METHODS: In this study, African Green monkeys were challenged with the highly virulent, aerosolized Y. pestis strain CO92, and untargeted metabolomics profiling of plasma was performed using liquid and gas chromatography with mass spectrometry. RESULTS: At early time points post-exposure, we found significant increases in polyunsaturated, long chain fatty acid metabolites with p values ranging from as low as 0.000001 (ratio = 1.94) for the metabolite eicosapentaenoate to 0.04 (ratio = 1.36) for the metabolite adrenate when compared to time-matched controls. Multiple acyl carnitines metabolites were increased at earlier time points and could be a result of fatty acid oxidation defects with p values ranging from as low as 0.00001 (ratio = 2.95) for the metabolite octanoylcarnitine to 0.04 (ratio = 1.33) for metabolite deoxycarnitine when compared to time-matched controls. Dicarboxylic acids are important metabolic products of fatty acids oxidation, and when compared to time matched controls, were higher at earlier time points where metabolite tetradecanedioate has a ratio of 4.09 with significant p value of 0.000002 and adipate with a ratio of 1.12 and p value of 0.004. The metabolites from lysolipids (with significant p values ranging from 0.00006 for 1-oleoylglycerophosphoethanolamine to 0.04 for 1-stearoylglycerophosphoethanolamine and a ratio of 0.47 and 0.78, respectively) and bile acid metabolism (with significant p values ranging from 0.02 for cholate to 0.04 for deoxycholate and a ratio of 0.39 and 0.66, respectively) pathways were significantly lower compared to their time-matched controls during the entire course of infection. Metabolite levels from amino acid pathways were disrupted, and a few from the leucine, isoleucine and valine pathway were significantly higher (p values ranging from 0.002 to 0.04 and ratios ranging from 1.3 to 1.5, respectively), whereas metabolites from the urea cycle, arginine and proline pathways were significantly lower (p values ranging from 0.00008 to 0.02 and ratios ranging from 0.5 to 0.7, respectively) during the course of infection. CONCLUSIONS: The involvement of several lipid pathways post-infection suggested activation of pathways linked to inflammation and oxidative stress. Metabolite data further showed increased energy demand, and multiple metabolites indicated potential hepatic dysfunction. Integration of blood metabolomics and transcriptomics data identified linoleate as a core metabolite with cross-talk with multiple genes from various time points. Collectively, the data from this study provided new insights into the mechanisms of Y. pestis pathogenesis that may aid in development of therapeutics.