Spatial transcriptomic profiling of the tumor microenvironment in EGFR and KRAS mutant non-small cell lung cancer.

Abstract

8539 Background: Immune checkpoint inhibitors have greatly improved outcomes in advanced non-small cell lung cancer (NSCLC). However, patients with EGFR mutant NSCLC have a poor response to immune checkpoint inhibitor therapy. Emerging evidence suggests that an immunosuppressive tumor microenvironment plays an important role in this setting, however, we still lack fundamental knowledge about tumor endothelial cell biology. We hypothesized that oncogene specific changes in the expression of immune-related genes in tumor endothelial cells account for differences in the tumor microenvironment and efficacy of immune checkpoint inhibitor therapy. Methods: We utilized spatial transcriptomics (GeoMx Digital Spatial Profiling) on resected tumor tissue of EGFR mutant (n = 5) and KRAS mutant (n = 5) NSCLC patients to investigate the transcriptional signature of tumor, endothelial and stromal cells. Additionally, we used NicheNet to explore intercellular communication between tumor, endothelial, and stromal cells. Immune gene set enrichment analysis scores were calculated using ESTIMATE. Immune cell type proportions were estimated using CIBERSORT. Results: Using spatial transcriptomics, we dissected the tumor microenvironment into tumor, stromal and endothelial compartments. By analyzing predicted cellular communication, we found that tumor and stromal cells primarily affect an interferon-related gene signature in tumor endothelial cells. Notable differentially expressed interferon-related genes, that were significantly up-regulated in KRAS mutant and down-regulated in EGFR mutant NSCLC patients, included CXCL9, STAT1, WARS1, IRF1 and ICAM1. In the stromal compartment, immune gene set enrichment analysis scores were significantly lower in EGFR mutant than KRAS mutant NSCLC (median, 355 vs. 713, P = 0.026) indicating an immunosuppressive tumor microenvironment. We observed substantial heterogeneity while exploring the cellular landscape of the stromal compartment in EGFR mutant and KRAS mutant NSCLC. Notably, we found a significantly decreased proportion of pro-inflammatory macrophages in the stromal compartment of EGFR compared to KRAS mutant NSCLC (P = 0.006). Conclusions: We identified distinct interferon-related gene signatures in tumor endothelial cells of patients with EGFR and KRAS mutantNSCLC associated with variations in the cellular composition of the tumor microenvironment. This may provide a better understanding for the development of spatial biomarkers to identify which oncogene-driven NSCLC patients are most likely to benefit from immune checkpoint inhibitor therapy.