Abstract 1518: A single-cell transcriptomic atlas of lung adenocarcinoma and adjacent normal-appearing tissue

Lung adenocarcinoma (LUAD) is the most frequently diagnosed histological subtype of lung cancer and accounts for most smoking-related cancer deaths, warranting strategies for early intervention. Earlier work revealed genome-wide aberrations in LUADs and the adjacent premalignant field, known as “field carcinogenesis”, that are pertinent to LUAD pathogenesis. Yet, we still poorly understand the cellular and molecular architecture of LUAD and its nearby “field”. To fill this void, we performed, using the 10X Genomics system and NovaSeq 6000 platform, single-cell RNA sequencing (scRNA-seq) analysis of 4 early-stage resected LUADs as well as 11 matched normal lung tissues with differing spatial proximity from the tumors. We first analyzed 15,739 cells from a LUAD and matched tumor-adjacent and -distant normal tissues from patient 1. The resulting fraction of EPCAM+ cells from this analysis was approximately 4%. For deeper single-cell resolution of LUAD, we thus performed scRNA-seq of separately sorted/enriched epithelial cells (EPCAM+; n = 50,883) and non-epithelial cells (EPCAM-; n= 91,093) from patients 2, 3 and 4, each with an early-stage LUAD and three matched spatially-distributed normal-appearing tissues. Overall, we achieved on average approximately 150,000 reads and 1,956 genes detected per cell. Divergent populations of malignant and non-malignant cells and multiple epithelial and immune subsets clustered in an overall spatially modulated pattern according to proximity to the tumor. Hierarchical clustering revealed multiple distinct populations of airway lineage cells, including alveolar type 1 (AT1), alveolar type 2 (AT2), AT1/AT2 bipotential cells, club, goblet, basal and ciliated cells, with cells from the resected LUAD present in some but not all clusters. By analysis of epithelial-enriched fractions, we were also able to identify and interrogate rare cell subpopulations including CFTR-expressing ionocytes. By inference of copy number variation along with analysis of key oncogene expression patterns, we also identified epithelial cells that represent potential tumor cells-of-origin. Spatial reprogramming of field carcinogenesis encompassed immune cell populations including tumor-exclusive FOXP3+ regulatory T cells with marked over-expression of the major immune checkpoints CTLA4 and TIGIT. In contrast, for at least some of the cases, tumors were nearly devoid of natural killer cells, and abundance of CD8+ T cells dampened with increasing proximity to the tumor. Our single-cell surveys offer insights into novel cues in LUAD pathogenesis. Efforts are underway to interrogate the single-cell epithelial and immune landscape of additional LUADs and matched nearby normal-appearing lung, including those from non-smokers, to better understand the evolution of the disease and, thus, identify a low-hanging fruit of targets for early management of this fatal malignancy. Citation Format: Ansam Sinjab, Guangchun Han, Warapen Treekitkarnmongkol, Patrick Brennan, Kieko Hara, Kyle Chang, Elena Bogatenkova, Beatriz Sanchez-Espiridion, Carmen Behrens, Jianjun Zhang, Boris Sepesi, Tina Cascone, Don L. Gibbons, Jichao Chen, George Eapen, Edwin J. Ostrin, Junya Fujimoto, Avrum E. Spira, Paul A. Scheet, Ignacio I. Wistuba, Linghua Wang, Humam Kadara. A single-cell transcriptomic atlas of lung adenocarcinoma and adjacent normal-appearing tissue [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1518.