Proteo-genomic characterization of virus-associated liver cancers reveals potential subtypes and therapeutic targets.

Primary liver cancer is a heterogeneous disease in terms of its etiology, histology, and therapeutic response. Concurrent proteomic and genomic characterization of a large set of clinical liver cancer samples can help elucidate the molecular basis of heterogeneity and thus serve as a valuable resource for personalized liver cancer treatment. In this study, we perform proteomic profiling of ~300 proteins on 259 primary liver cancer tissues with reverse-phase protein arrays, mutational analysis using whole genome sequencing and transcriptional analysis with RNA-Seq. Patients are of Japanese ethnic background and mainly HBV or HCV positive, providing insight into this important liver cancer subtype. Unsupervised classification of tumors based on protein expression profiles reveal three proteomic subclasses R1, R2, and R3. The R1 subclass is immunologically hot and demonstrated a good prognosis. R2 contains advanced proliferative tumor with TP53 mutations, high expression of VEGF receptor 2 and the worst prognosis. R3 is enriched with CTNNB1 mutations and elevated mTOR signaling pathway activity. Twenty-two proteins, including CDK1 and CDKN2A, are identified as potential prognostic markers. The proteomic classification presented in this study can help guide therapeutic decision making for liver cancer treatment.

Molecular Classification and Tumor Microenvironment Characterization of Gallbladder Cancer by Comprehensive Genomic and Transcriptomic Analysis.

Gallbladder cancer (GBC), a rare but lethal disease, is often diagnosed at advanced stages. So far, molecular characterization of GBC is insufficient, and a comprehensive molecular portrait is warranted to uncover new targets and classify GBC. We performed a transcriptome analysis of both coding and non-coding RNAs from 36 GBC fresh-frozen samples. The results were integrated with those of comprehensive mutation profiling based on whole-genome or exome sequencing. The clustering analysis of RNA-seq data facilitated the classification of GBCs into two subclasses, characterized by high or low expression levels of TME (tumor microenvironment) genes. A correlation was observed between gene expression and pathological immunostaining. TME-rich tumors showed significantly poor prognosis and higher recurrence rate than TME-poor tumors. TME-rich tumors showed overexpression of genes involved in epithelial-to-mesenchymal transition (EMT) and inflammation or immune suppression, which was validated by immunostaining. One non-coding RNA, miR125B1, exhibited elevated expression in stroma-rich tumors, and miR125B1 knockout in GBC cell lines decreased its invasion ability and altered the EMT pathway. Mutation profiles revealed TP53 (47%) as the most commonly mutated gene, followed by ELF3 (13%) and ARID1A (11%). Mutations of ARID1A, ERBB3, and the genes related to the TGF-ß signaling pathway were enriched in TME-rich tumors. This comprehensive analysis demonstrated that TME, EMT, and TGF-ß pathway alterations are the main drivers of GBC and provides a new classification of GBCs that may be useful for therapeutic decision-making.