Table_2.docx

Abstract

Background: Accumulating studies have demonstrated the abnormal expressions and prognostic values of certain members of the TRIM family in diverse cancers. However, comprehensive prognostic values of the TRIM family in hepatocellular carcinoma (HCC) are yet to be clearly defined. Methods: The prognostic values of the TRIM family were evaluated by survival analysis and univariate Cox regression analysis based on gene expression data and clinical data of HCC from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) databases. The expression profiles, protein-protein interaction among the TRIM family, prediction of transcription factors (TFs) or miRNAs, genetic alterations, correlations with the hallmarks of cancer and immune infiltrates and pathway enrichment analysis were explored by multiple public databases. Further, we built a TRIM family gene-based signature for predicting overall survival (OS) in HCC by using least absolute shrinkage and selection operator (LASSO) regression. We used TCGA-LIHC cohort as training set and GSE76427 was used for external validation. Time-dependent receiver operating characteristic (ROC) and survival analysis were used to estimate the signature. Finally, a nomogram combining the TRIM family risk score and clinical parameters was established. Results: High expression of TRIM family members including TRIM3, TRIM5, MID1, TRIM21, TRIM27, TRIM32, TRIM44, TRIM47 and TRIM72 were significantly associated with HCC patients’ poor overall survival. A novel TRIM family gene-based signature (including TRIM5, MID1, TRIM21, TRIM32, TRIM44 and TRIM47) was built for OS prediction in HCC. ROC curves suggested the signature’s good performance in OS prediction. HCC patients in high-risk group had poorer OS compared to the low-risk patients based on the signature. A nomogram integrating the TRIM family risk score, age and TNM stage was established. The ROC curves suggested that it presented better discrimination than the similar model without the TRIM family risk score. Conclusion: Our study identified the potential application values of TRIM family for outcome prediction in HCC.