Abstract 3156: Meta-analysis of lung cancer single-cell RNA data reveals novel gene modules associated with lineage, stage, and genetic alterations

Abstract

Abstract Objective: Combinations of somatic alterations, epigenetic changes, and interactions with cells in the tumor microenvironment can induce gene expression changes in cancer cells. Single-cell RNA-seq can provide valuable insights into the heterogeneity of cancer cells within and across tumor samples. We performed a meta-analysis of single-cell RNA-seq (scRNA-seq) datasets containing primary tumor samples from patients with non-small-cell-lung cancer (NSCLC) to characterize modules of co-expressed genes and determine associations with histological, genetic, and immune features. Methods: scRNA-seq data of 116,260 cells from 127 NSCLC samples including lung adenocarcinomas (LUADs) and lung squamous cell carcinomas (LUSCs) were collected and merged from 12 datasets. Cancer cells were identified by inferring copy number variations (CNVs). Celda was used to identify 163 gene co-expression modules across cancer cells. MAST was used to assess association of each module with histology, stage, and study. CaDrA was used to identify combinations of genomic features associated with gene modules projected into TCGA bulk RNA-seq. Result: Broad biological category of modules included those related to lineage, cell cycle and proliferation, extracellular matrix production (ECM), epithelial-to-mesenchymal transition (EMT), antigen presentation, xenobiotic metabolism, stress response, chemokine signaling, interferon signaling, housekeeping, and other general cellular processes. The modules most strongly associated with study-specific effects were enriched in housekeeping and ribosomal genes. Several ECM and cell-cycling modules were induced in advanced LUADs while MHC class II modules were reduced in late stage disease. Several modules containing different surfactant proteins or secretoglobins were identified and displayed heterogeneity within and across early-stage LUADs. Three modules in LUSCs were enriched for detoxification enzymes and associated with somatic mutations in different members of the Nrf2 pathway. A module of genes involved in keratinization displayed intra- and inter-tumoral heterogeneity across LUSCs and was associated with neutrophil infiltration scores in samples from TCGA. Conclusion: We employed a novel meta-analysis approach to study gene co-expression patterns that are active in lung cancer cells profiled with single-cell RNA-seq. Characterization of these modules revealed the biological processes and cancer hallmarks active within and across lung tumors. Citation Format: Rui Hong, Sarah Mazzilli, Stefano Monti, Avrum Spira, Joshua D. Campbell. Meta-analysis of lung cancer single-cell RNA data reveals novel gene modules associated with lineage, stage, and genetic alterations [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 3156.