Analyzing the relationship between glutaminase expression and features of the immune tumor microenvironment in epithelial ovarian cancer using imaging mass cytometry.

5560 Background: Immunotherapy has thus far shown limited efficacy in epithelial ovarian cancer, likely driven in part by an immune suppressive tumor microenvironment (TME). Increased glutamine metabolism by cancer cells via upregulation of the drug-targetable enzyme glutaminase (GLS) may contribute to an immune suppressive TME. Inhibiting GLS may not only inhibit tumor growth but also enhance the anti-tumor immune response in patients with epithelial ovarian cancer. We investigated the relationship between GLS expression and features of the immune TME in epithelial ovarian cancer using imaging mass cytometry. Methods: Tissue microarrays constructed from 41 epithelial ovarian cancer (38 high-grade serous, 3 low-grade serous) surgical specimens were stained by immunohistochemistry for GLS, which was quantified using modified histologic score (H-score). Imaging mass cytometry was then performed on tissue microarrays using a panel of 43 channels, including markers for delineating tissue architecture and assessing lymphoid cells, myeloid cells, and stromal fibroblasts. Resulting multiplexed images were segmented into a single-cell dataset to quantitatively compare between specimens, according to GLS H-score, abundances of cell types and average shortest distances between cell types. Results: Median GLS H-score was 150. Compared to GLS-low (H-score 0-150) specimens, GLS-high specimens (H-score >150) demonstrated lower T cell abundance (9.18% vs. 16.84% of cells; p=0.010) and lower B cell abundance (1.09% vs. 6.66% of cells; p=0.022). Compared to GLS-low (H-score <130) and GLS-medium (H-score 130-150) specimens, GLS-high specimens (H-score >150) demonstrated the lowest T cell abundance (high: 9.18% < medium: 10.46% < low: 21.28% of cells; p=0.0015) with subtype analysis showing the lowest effector helper T cell abundance (high: 0.11% < medium: 0.25% < low: 3.39% of cells; p=0.040). On spatial analysis, based on average shortest distances between cell types, GLS-high specimens demonstrated longer distances between tumor cells and lymphoid/myeloid immune cells and shorter distances between tumor cells and stromal cells than GLS-low/medium specimens. Conclusions: In this cohort of patients with epithelial ovarian cancer, higher levels of GLS expression were associated with several features of an immune suppressive TME, including lower T cell abundance with lower effector helper T cell abundance on subtype analysis, lower B cell abundance, and decreased proximity between tumor cells and immune cells. Further clinical studies investigating the use of GLS inhibitors to modulate the immune TME in patients with epithelial ovarian cancer are warranted.