P041 Tandem mass tag-based quantitative proteomic profiling identifies novel putative serum biomarkers for the diagnosis of drug-induced liver injury in patients

Abstract

<h3>Background</h3> Diagnosis of drug-induced liver injury (DILI) remains a significant challenge in interventional clinical trials during drug development and in clinical practice. We aimed to discover novel diagnostic biomarkers using proteomics and to evaluate their performance characteristics in distinguishing DILI from acute and chronic liver injury from other etiologies. <h3>Methods</h3> Adults with idiosyncratic DILI and acute non-DILI liver disease controls were recruited to the Prospective European DILI Registry. Case definitions and causality assessment were performed according to European Association for the Study of the Liver recommendations. Tandem Mass Tag (TMT) labelled quantitative proteomics was performed in serum samples from a discovery set of healthy volunteers (HV, n=10), patients with DILI at onset (DO) and recovery (DR) (n=10 each); acute non-DILI controls at onset (NDO) and recovery (NDR) (n=5 each) and patients with non-alcoholic fatty liver disease (n=10). The performance of top biomarker candidates was assessed in a confirmatory cohort with HV (n=60), DO (n=83), DR (n=85), NDO (n=35) and NDR (n=21) using peptide-based targeted mass spectrometry. Novel biomarkers and patient status were evaluated using univariate and multivariate models. <h3>Results</h3> Global proteomic profiling of the discovery cohort revealed 2323 proteins, of which 12 novel biomarkers were chosen based on differential expression, liver specificity, and mechanistic relevance to liver biology/pathogenesis. In the confirmatory analysis, serum concentration of the novel biomarkers was significantly different at DO and returned to levels comparable to HV during recovery (DR). When the diagnostic performance of these 12 biomarkers was assessed individually, along with the promising biomarker Cytokeratin 18 (CK18, Area Under the Receiver Operator Characteristic (AUROC) Curve = 0.96), the following biomarkers demonstrated the highest sensitivity/specificity (DILI vs HV): Cytoplasmic aconitate hydratase (ACO1, 0.99), Argininosuccinate synthase (ASS1, 0.98), Fumarylacetoacetase (FAH, 0.98), Carbamoyl-phosphate synthase (CPS1, 0.96), Fructose-bisphosphate aldolase B (ALDOB, 0.94), 4-Hydroxyphenylpyruvate dioxygenase (HPD, 0.94), Ornithine carbamoyltransferase (OTC, 0.92). <h3>Conclusion</h3> A TMT-based quantitative proteomic profiling approach identified novel putative biomarkers, which are candidates for further evaluation of their role in the diagnosis and prognostication of DILI.