Identification and Validation of a Nine-Gene Amino Acid Metabolism-Related Risk Signature in HCC.

ABSTRACT

Background: Hepatocellular carcinoma (HCC) is the world's second most deadly cancer, and metabolic reprogramming is its distinguishing feature. Among metabolite profiling, variation in amino acid metabolism supports tumor proliferation and metastasis to the most extent, yet a systematic study on the role of amino acid metabolism-related genes in HCC is still lacking. An effective amino acid metabolism-related prediction signature is urgently needed to assess the prognosis of HCC patients for individualized treatment. Materials and Methods: RNA-seq data of HCC from the TCGA-LIHC and GSE14520 (GPL3921) datasets were defined as the training set and validation set, respectively. Amino acid metabolic genes were extracted from the Molecular Signature Database. Univariate Cox and LASSO regression analyses were performed to build a predictive risk signature. K-M curves, ROC curves, and univariate and multivariate Cox regression were conducted to evaluate the predictive value of this risk signature. Functional enrichment was analyzed by GSEA and CIBERSORTx software. Results: A nine-gene amino acid metabolism-related risk signature including B3GAT3, B4GALT2, CYB5R3, GNPDA1, GOT2, HEXB, HMGCS2, PLOD2, and SEPHS1 was constructed to predict the overall survival (OS) of HCC patients. Patients were separated into high-risk and low-risk groups based on risk scores and low-risk patients had lower risk scores and longer survival time. Univariate and multivariate Cox regression verified that this signature was an independent risk factor for HCC. ROC curves showed that this risk signature can effectively predict the 1-, 2-, 3- and 5-year survival times of patients with HCC. Additionally, prognostic nomograms were established based on the training set and validation set. These genes were closely correlated with the immune regulation. Conclusion: Our study identified a nine-gene amino acid metabolism-related risk signature and built predictive nomograms for OS in HCC. These findings will help us to personalize the treatment of liver cancer patients.