Abstract LB-321: An integrated somatic mutation-miRNA-transcriptome network analysis identifies prognostic subgroups in advanced gastric cancer

Abstract

Abstract Gastric cancer (GC) is the second leading cause of cancer-related fatality worldwide. Although radical surgical resection with adjuvant chemotherapy is the mainstay of treatment, overall clinical outcome remains unsatisfactory. The development of genomic signatures to predict clinical outcome in GC patients can help to develop personalized cancer treatment to improve overall survival and clinical benefits with chemotherapy in GC patients. The aim of this study is to develop prognostic signatures in GC, and investigate the clinical significance of the signatures for the molecular stratification of GC and the therapeutic implications. In this study, we collected 571 patients with GC from Yonsei University College of Medicine (Yonsei Gastric Cancer cohort, YGC), Seoul, Korea, between 1999 and 2006, and generated genome-wide transcriptome and miRNA expression data. All samples, fresh frozen at the time of surgical resection, and clinical information were obtained in accordance with Institutional Review Board guidance and patient informed consent. Samples were reviewed by an expert pathologist and then subjected to RNA and protein extraction. We developed a novel computational algorithm to integrate somatic mutation with protein-protein interaction networks and pathway database to discover driver pathways associated with clinical outcomes in GC patients. We applied our method to 105 GC patients’ data from The Cancer Genome Atlas Project (TCGA), and found pathway signatures altered by somatic mutation in TCGA GC patients, which are likely essential driver pathways in GC progression. The signatures are consisted of 30 genes including well-known tumor suppressor or oncogenes such as BRCA1, CREBBP, TP53, GNL3, and MSH6. We applied consensus clustering to perform patient stratification for YGC cohort with the signatures obtained from TCGA GC data. We also performed an integrated miRNA-mRNA analysis to identify a novel miRNA-gene regulatory network linked to clinical outcome in YGC cohort. We found patient subgroups having significant different clinical outcomes in YGC cohort. Especially, we identified patient subgroup with Tumor stage IIIA having significant survival difference. High-risk group have shorter survival (48 months) than low-risk group (83 months), corresponding to a HR of 3.12 (95% CI 1.36-7.85, p = 0.00381). We found that DNA-repair genes such as BRCA1, MSH6, GNL3 are often downregulated in high-risk group and this might bring new prospects for developing new gastric cancer therapies. In addition, an integrative network analysis of miRNA and gene expression data in high risk group showed a novel miRNA-gene regulatory network including miRNA 200, 106 and 125 families that are previously known for poor prognosis subtype of GC as well as tumorigenesis. Our findings identify pathway signature-defined molecular subgroups of GC with distinct clinical outcomes, and provide novel insights of the miRNA-gene regulatory network in aggressive GC. Identification of the molecular subtypes may be useful to assess new drugs in preclinical trials and can influence treatment decisions. Development of clinically applicable assay based on our signatures for GC subtype classification may also contribute to clinical decision making in personalized management of GC. Citation Format: Jae-Ho Cheong, Sun Ho Park, Eun Sung Park, Hyun Ki Kim, Kyung Ho Pak, Sung Hoon Noh, Tae Hyun Hwang. An integrated somatic mutation-miRNA-transcriptome network analysis identifies prognostic subgroups in advanced gastric cancer. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr LB-321. doi:10.1158/1538-7445.AM2014-LB-321