Metabolome and transcriptome integration reveals metabolic profile of hepatocellular carcinoma

Abstract Background and Aim Accumulated evidence highlights the role of metabolites in cancer diagnosis. However, the diagnosis of hepatocellular carcinoma (HCC), especially its early diagnosis, is still very difficult. The main purposes of the study are to explore the comprehensive characteristic metabolites of HCC through an integrated nontargeted metabolomics and transcriptomics approach and evaluate the diagnostic value of some metabolic changes in HCC. Methods Dysregulated metabolites and pathways in HCC were identified by nontargeted metabolomics analysis of 72 pairs of matched liver tissues, including HCC tissue (HCT) and adjacent noncancerous tissue (ANT). Meanwhile, to ensure the reliability of the results, metabolic enzymes were quantified at the mRNA level by RNA sequencing. To facilitate the utilization of this information, a diagnostic model was developed based on binary logistic regression using 63 HCC serum samples collected from the aforementioned 72 patients and 40 noncancer serum samples. Results The results showed that 267 metabolites were significantly altered in HCT. These differential metabolites binding to related differential metabolic enzyme genes were enriched in 14 metabolic pathways. And combination of 5‐oxoproline, taurocholenate sulfate, and maltose could be used as a novel candidate early serum diagnostic marker for HCC. Conclusions We profiled the metabolic features of HCC and found global biochemical pathway aberration. The diagnostic potential of differential metabolites found in serum tissues, further validated in liver samples, showed that 5‐oxoproline, taurocholenate sulfate, and maltose combination had a high accuracy for hepatocellular carcinoma detection, especially for alpha fetoprotein negative patients.