Tumors in the distal esophagus (EAC), gastro-esophageal junction including cardia (GEJAC), and stomach (GAC) develop in close proximity and show strong similarities on a molecular and cellular level. However, recent clinical data showed that the effectiveness of chemo-immunotherapy is limited to a subset of GEAC patients and that EACs and GEJACs generally benefit less from checkpoint inhibition compared to GACs. As the composition of the tumor immune microenvironment drives response to (immuno)therapy we here performed a detailed immune analysis of a large series of GEACs to facilitate the development of a more individualized immunomodulatory strategy.
Extensive immunophenotyping was performed by 14-color flow cytometry in a prospective study to detail the immune composition of untreated gastro-esophageal cancers (n=104) using fresh tumor biopsies of 35 EACs, 38 GEJACs and 31 GACs. The immune cell composition of GEACs was characterized and correlated with clinicopathologic features such as tumor location, MSI and HER2 status. The spatial immune architecture of a subset of tumors (n=30) was evaluated using multiplex immunohistochemistry (mIHC) which allowed us to determine the tumor infiltration status of CD3+, CD8+, FoxP3+, CD163+ and Ki67+ cells.
Immunophenotyping revealed that the tumor immune microenvironment of GEACs is heterogeneous and that immune suppressive cell populations such as monocytic myeloid-derived suppressor cells (mMDSC) are more abundant in EACs compared to GACs (p<0.001). In contrast, GACs indicated a proinflammatory microenvironment with elevated frequencies of proliferating (Ki67+) CD4 Th cells (p<0.001), Ki67+ CD8 T cells (p=0.002), and CD8 effector memory-T cells (p=0.024). Differences between EACs and GACs were confirmed by mIHC analyses showing lower densities of tumor- and stroma-infiltrating Ki67+ CD8 T cells in EAC compared to GAC (both p=0.021).
This comprehensive immune phenotype study of a large series of untreated GEACs, identified that tumors with an esophageal tumor location have more immune suppressive features compared to tumors in the gastro-esophageal junction or stomach which might explain the location-specific responses to checkpoint inhibitors in this disease. These findings provide an important rationale for stratification according to tumor location in clinical studies and the development of location-dependent immunomodulatory treatment approaches.