Immune subtype of mismatch repair deficient colorectal carcinoma.

Abstract

e15194 Background: Half of microsatellite instability-high (MSI-H) colorectal carcinoma (CRCs) patients do not response to checkpoint blockade (ICB) immunotherapy. We aimed to unleash the underlying mechanism of distinct efficacy to ICB by sophisticated depicting the immune microenvironment of MSI-H CRC. Methods: Public omics and clinical data of patients with MSI labels were collected from TCGA data portal (364 samples) and Gene Expression Omnibus (GSE39582, GSE13294, GSE13067, GSE35896, GSE24514, GSE25071, GSE79959, GSE103340, GSE29638, GSE33113 cohorts, 1119 samples). GEO datasets were used as discovering sets, and TCGA dataset was used as validating sets. The immune cell infiltrations and signaling pathways in the tumor microenvironment were quantified by single sample Gene Set Enrichment Analysis (ssGSEA), and the bio-similarities were estimated by t-Distributed Stochastic Neighbor Embedding (tSNE). Clustering was identified by k-mean unsupervised strategy with optimal k. Statistical analysis and data visualization were all carried out using the R. Results: Based on the whole transcriptome and immune cell infiltration profiling, both GEO and TCGA cohorts showed that MSI-H patients harbored unique pattern which significantly differed from the MSS counterpart. Unsupervised clustering further classified the MSI-H patients into “Hot” and “Cold” subgroups with remarkable different IFNG-related signatures, inflammatory response score, interferon-γ signaling score, and survival outcome. Screening assay identified 23 genes ( ADAMTSL3, DLGAP4, DNAH11, FFAR3, GABRA4, KCNA6, KCTD9, KIAA1033, KLHL6, LARGE, MORC1, NDST3, OR5H14, PAX5, PHF3, PIK3R5, RNF25, SPATA20, STAT4, TLR9, TMC3, TTC12, ZNF790) enriched and 5 genes ( CELSR3, DGKD, HECTD4, RECQL4, SLC9A4) spared in immune “hot” tumors. Moreover, “hot” tumor showed less WNT signaling pathway score, which was negatively correlated with the IFNG signaling and inflammatory response within the MSI-H CRC tumors. Conclusions: We firstly identified immune hot and cold groups from MSI-H CRCs, which may be related to different immunotherapy response. This result may help further understanding the anti-immunotherapy of MSI-H CRCs.