Abstract 6526: Spatial characterisation of the tumour immune microenvironment in ARID1A-deficient gastric cancer

Abstract

Abstract AT-Rich Interacting Domain 1A (ARID1A) is an important component in SWI/SNF Complex, a chromatin remodeller and an established tumor suppressor gene that is mutated in a variety of cancer types. This includes endometrial (~45%), bladder (~20%), and gastric (~15%). ARID1A mutations usually lead to a loss of protein expression or function, which results in tumor progression. The loss of ARID1A has been linked to genomic instability, potentially making the tumor more immunogenic/amenable to immune checkpoint inhibitor therapy. However, there is a discrepancy between clinical and mechanistic studies in determining the consequence of ARID1A-loss on the tumor immune microenvironment (TIME): clinical studies have shown that ARID1A mutation is a potential predictive biomarker for immunotherapy success, whereas another mechanistic study found that ARID1A-loss results in a cold TIME that is less amenable to immunotherapy. Considering this controversy, our overarching objective is to use multiple high-dimensional microscopy-based techniques to robustly and comprehensively characterize the TIME sculpted by ARID1A-loss. We study this in the context of gastric cancer, a disease which has heterogenous responses to PD-1 based immunotherapy and also shows frequent (10-25%) mutations in ARID1A. We looked at two retrospective cohorts with 250 patients each, one from UK (Leeds) and one from Japan (KCCH). To ascertain ARID1A-loss in these patients, we have set up an IHC-based assay comparing ARID1A expression between tumor and stroma regions. From this, we derive a ratio. In both studies, we were able to clearly stratify the cohort into ARID1A-loss and ARID1A-wildtype. Leeds had about 23 cases (10%) of ARID1A-loss, while KCCH had 39 cases (17.9%). Looking at the clinical parameters, we realized that ARID1A-loss does not stratify for survival nor staging of the tumor, which is also seen in ARID1A-mutation studies. Using multiplex immunohistochemistry (IHC), we saw differences in the immune cell population. ARID1A-loss patients seem to have a greater proportion of CD3+ cells in the Leeds (1423 vs 666 cells/mm2, p = 0.0269) and KCCH cohorts (1457 vs 1268 cells/mm2, p = 0.1745). Interestingly, CD68+ cells also show an increase in the ARID1A-loss population in the Leeds (802.1 cells/mm2 vs 580.3 cells/mm2, p = 0.0649) and KCCH cohorts (659 vs 623.2 cells/mm2). However, CD20+ cells show no changes in abundance for either cohort (249.1 vs 193.2 cells/mm2, p = 0.4985; and 361.6 vs 357.5 cells/mm2, p = 0.9634) respectively. Overall, there seems to be an increase in the T-cell population in ARID1A-loss tumors and an increase in the macrophage population in stroma regions of ARID1A-loss tumors. The B-cell population remains unchanged. To complement our preliminary findings, we will leverage hyperplex and spatial transcriptome techniques to uncover the interactions between tumor cells and the TIME. Citation Format: Norbert Sheng Cong Tay, Yan Fen Peng, Heike Grabsch, Anand D. Jeyasekharan. Spatial characterisation of the tumour immune microenvironment in ARID1A-deficient gastric cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6526.