Abstract 4622: Tumor microenvironment differences between lung cancer subtypes revealed by spatial transcriptomics

Abstract

Abstract Lung cancer is a leading cause of cancer deaths worldwide and has complex underlying genetic drivers, subtypes and immune cell types. Molecular analysis of bulk tumors has repeatedly identified key somatic driver genes and subtypes in lung cancer. However, these key molecular strata of bulk lung tumors still contain significant heterogeneity which if characterized in finer detail may reveal new tumor microenvironment factors and lead to improved patient prognostication and therapy options. Here, we sought to compare the tumor microenvironments of lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) through spatial transcriptomics. Using four frozen lung tumors and three with replicated sections (n = 7), we sequenced spatial transcriptomes using the 10X Visium platform and illumina sequencing to measure up to 5,000 latticed-spots throughout 6.5 mm2 of the tumor surface area. Data analysis revealed a large number of latticed-spots per tumor (median 3,486) with a large number of genes detected per median spot (tumor median 4,427). Through unsupervised clustering of spots in each tumor, we found between 8 and 10 clusters per tumor with distinct pathway activities, including multiple immune-enriched clusters per tumor (range: 3-5). Immune-enriched clusters in one LUAD tumor displayed a spatial shape consistent with tertiary lymphoid structures (TLS). Concordantly, this cluster overexpressed both B cell and T cell pathways and as well as a TLS signature from liver cancer. Interestingly by bulk tumor RNA analysis, this TLS+ tumor was classified to be in the terminal respiratory unit expression subtype, which is an immune-mild bulk subtype. This supports that the TLS signal can be a unique property of spatial expression in lung cancer that may be unobservable by bulk tumor RNA sequencing. Then to compare global spatial heterogeneity among tumors, we calculated an index of expression spatial continuity and found LUSC tumors to have more contiguous expression patterns than LUAD tumors (mean 0.60 vs 0.54). We also quantified expression diversity across all tumor latticed-spots and found that LUSC tumors had greater values compared to LUAD tumors (mean 0.37 vs 0.27). Together, our results suggest that LUSC has a more contiguous and heterogeneous tumor expression microenvironment than LUAD. TLS are predictive of immune checkpoint inhibitor response in many other tumor types, and our results suggest that spatial transcriptomics may also identify this responsiveness in lung cancer. Future, larger cohorts of lung tumors are needed to determine recurrent spatial properties associated with patient outcome and treatment response. The views expressed in this abstract are solely of the authors and do not reflect the official policy of the Departments of Army/Navy/Air Force, Department of Defense, USUHS, HJF, or U.S. Government. Citation Format: Matthew D. Wilkerson, Savannah Kounelis-Wuillaume, Camille Alba, Teri J. Franks, Martin L. Doughty, Robert L. Kortum, Robert F. Browning, Clifton L. Dalgard, Craig D. Shriver. Tumor microenvironment differences between lung cancer subtypes revealed by spatial transcriptomics. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4622.