Coccidioidomycosis Detection Using Targeted Plasma and Urine Metabolic Profiling

Abstract

Coccidioidomycosis, or Valley Fever (VF), is a potentially lethal fungal infection that results in more than 200 deaths/year in the United States. Differential diagnosis and timely treatment of VF is made difficult due to presentation of vague symptoms that often mimic other pulmonary conditions. Accurate diagnosis therefore relies on a combination of clinical presentation, serology, radiography, histology and culture. These diagnostic methods are either costly, time‐consuming, invasive, or indeterminate. Previous studies have used MS‐based methods for the accurate diagnosis of similar fungal diseases as well as characterization of host‐fungal interactions. Herein, we present the first targeted plasma and urine liquid chromatography tandem mass spectrometry (LC‐MS/MS) profiling approach for the relatively non‐invasive, rapid and accurate detection of VF.In this targeted approach, 207 plasma metabolites and 231 urine metabolites were reliably detected using LC‐MS/MS. These metabolites, which are representative of more than 35 metabolic pathways of potential biological relevance, were monitored in 59 plasma samples and 88 urine samples acquired from excess clinical specimens taken from two groups of patients (48 VF patients and 99 non‐VF controls) at Mayo Clinic Arizona. Univariate and multivariate statistical analyses were performed to develop predictive plasma and urine biomarker panels. Pathway and enrichment analyses were performed to elucidate potential pathogenic mechanisms and therapeutic targets.Results of our univariate significance testing and multivariate model estimation informed the construction of a 3‐metabolite panel of plasma biomarkers (q < 0.05 and VIP > 2; Fig 1) and a 9‐metabolite panel of urinary biomarkers (q < 0.05 and VIP > 1; Fig 2). Receiver operating characteristic (ROC) curves (Fig 3) generated based on enhanced orthogonal partial least squares‐discriminant analysis (OPLS‐DA) models (Fig 4) showed excellent overall classification performance (99.5%), with 94.4% sensitivity and 97.6% specificity for plasma metabolites. Although urine metabolites were less accurate (92.5%), demonstrating 89.7% sensitivity and 88.1% specificity, they represent a substantial increase in diagnostic accuracy over currently available urinary antigen tests, which only report sensitivity of around 70%. Pathway analysis (Fig 5) and enrichment analysis (Fig 6) ubiquitously showed alterations in metabolic pathways related to fungal virulence, DNA/RNA synthesis and energy utilization, and host immune escape mechanisms. These results potentially reveal pathways or markers that could be targeted therapeutically.This study is the first‐ever MS‐based metabolomics approach for rapid and highly accurate diagnosis of VF.Support or Funding InformationSupport from the College of Health Solutions at Arizona State University, Arizona Biomedical Research Council (ABRC) grant# 16–162513 and American Achievement Rewards for College Scientists (ARCS) Foundation Award is gratefully acknowledged.This abstract is from the Experimental Biology 2019 Meeting. There is no full text article associated with this abstract published in The FASEB Journal.