Integrated DNA and RNA sequencing to reveal early drivers of metastasis in gastric cancer.

e16096 Background: Gastric cancer (GC) is the second cause of cancer-related death and metastasis is an important cause of death. Considering difficulties in searching for metastatic driver mutations, we tried a novel strategy to identify early drivers in primary gastric tumors. In this study, we conducted an integrative genomic analysis on GC patients and identified early drivers that lead to metastases. Methods: Tumors and matched normal tissues from 400 Chinese GC patients were characterized by whole-exome sequencing (WES), transcriptome sequencing and targeted exome sequencing (TES), especially the comparative analysis results between higher metastatic potential (HMP) group with T1 stage and lymph-node metastases, and lower metastatic potential (LMP) group without lymph-nodes or distant metastases. Functional experiments on candidate driver mutations were applied in GC cell lines to explore their roles in metastasis and immune escape. Results: HMP group presented higher mutation load and heterogeneity, enrichment in immunosuppressive signaling, lower tumor purity, more stromal and immune cell infiltration than LMP group. We identified 262 differentially expressed mRNA, lncRNA and miRNA between HMP and LMP group (with more than twice proportion of 2-fold increased or decreased tumor than normal tissue in HMP than LMP group, and with significantly 2-fold higher or lower in HMP tumors than LMP tumors). 168 candidate prometastatic mutations were found by WES and 8 were selected for following TES to detect their association with distant metastasis. We found that four candidate mutations were related to distant metastasis and mutated TP53 and MADCAM1 were significantly associated with poor metastasis-free survival. It’s worth noting that though with low mutational rate, MADCAM1 mutations were only observed in patients with metastasis within three years after surgery or at diagnosis. We demonstrated that MADCAM1-mutated proteins act as GOF mutants and could not only directly promote cancer cells migration, but also could trigger tumor metastasis by establishing complicated immunosuppressive microenvironment, including promoting PD-L1-mediated immune escape and reprogramming tumor-associated macrophages (TAMs). Conclusions: Our results showed that the GCs with different metastatic potential are distinguishable at the genetic level. A number of potential metastatic driver mutations was revealed in this study including known cancer-associated genes TP53. We also discovered some novel driver mutations, such as MADCAM1. The results indicate driver mutations in early-onset metastatic GC could promote metastasis not only by directly empower cancer cells to disseminate but also by establishing an immunosuppressive microenvironment. This study increased the understanding of molecular landscape of GC patients and provided possibility for future target therapy.