Abstract 4143: Characterizing the immunogenicity of gastric cancer by transcriptomic expression based immune phenotyping

Abstract

Abstract One of the main challenges in immuno-oncology is to quantify different types of immune cells in the tumor microenvironment, as such data can greatly facilitate our understanding of the mechanism of action of, and response to cancer immunotherapies. Traditionally immune profiling is performed by immunohistochemistry or flow cytometry experiments using surface markers specific to each immune cell type. However, both approaches suffer from practical limitations such as reagent availability and real world specimen conditions. Recently several computational deconvolution approaches using expression profiles of the bulk tumor tissue samples have been developed. This alternative approach utilizes existing database of purified immune cell gene signatures and relies on a rigorous mathematical framework of signal deconvolution. One of such algorithms, called CIBERSORT (Newman et al. 2015), has been demonstrated to perform robustly in deconvoluting the relative fractions of 22 human leukocyte subsets in solid tumor tissue samples, and benchmarked favorably against FACS analysis. In this work, we performed immune phenotyping by the CIBERSORT-based expression deconvolution on a cohort of 100 gastric cancer (GC) patients. We found that the fractions of activated CD4 memory T cells, rather than CD8+ T cell, are most significantly correlated with tumor neo-antigen load, whereas the latter is most significantly associated with the expression of a “Cytolytic Activity” metagene and an interferon gamma signature. There is stronger pre-existing host immune response against the MSI and EBV subtypes of GCs mediated by either cytotoxic T cells or Natural Killer cells. Lastly, we found a high M2/M1 tumor associated Macrophages (TAMs) ratio is strongly associated with poor GC prognosis, corroborating recent reports on the role of a TAM and stromal response in gastric cancer angiogenesis. Taken together, we demonstrated in this work that critical information about the tumor infiltrating immune cells can be gleaned from the computational deconvolution of bulk tumor gene expression profiling data. The comprehensiveness and cost-effectiveness of this approach can complement other immune profiling techniques in characterizing a wide variety of tumor specimens under various treatment conditions. We foresee its application in characterizing pharmacodynamics, understanding immunological mechanism of action and monitoring treatment response and disease progression for cancer immunotherapies and their combinations in both pre-clinical and clinical settings. Citation Format: Kai Wang, Hoicheong Siu, Shibing Deng, Jadwiga R. Bienkowska, Suet Yi Leung. Characterizing the immunogenicity of gastric cancer by transcriptomic expression based immune phenotyping. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4143.