Deciphering cell lineage specification of human lung adenocarcinoma with single-cell RNA sequencing

Abstract Lung adenocarcinomas (LUAD) start as precancerous lesions such as atypical adenomatous hyperplasia (AAH), develop stepwise into adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA), then eventually progress toward invasive adenocarcinoma (IA). To date the cellular heterogeneity across these distinct clinical stages and the underlying molecular events driving tumor progression remain largely unclear. In this study, we performed single-cell RNA sequencing on 52 specimens from 25 patients spanning the four clinical stages. By assessing the expression pattern of marker genes among 268,471 cells, we identified 16 major cell types. We demonstrated that AT2 feature cell types (AT2-like cells) were associated with malignant composition. AT2-like subcluster emerged first in AAH and partially lost AT2 cell transcriptional identity, accompanied with a gain of stemness during cell transition. In addition, genes related to energy metabolism, ribosome synthesis were upregulated in the early stage of LUAD, leading us to identify new markers including miRNA10 and β-hydroxybutyric acid to diagnose early-stage LUAD noninvasively in the blood. We also identified MDK and TIMP1 as potential biomarkers to facilitate our understanding of LUAD pathogenesis. Taken together, our data identified a new mechanism in LUAD evolution, and provided a robust basis for diagnosis and treatment of LUAD.