An integrated single-cell lung cancer atlas reveals distinct fibroblast phenotypes between adenocarcinoma and squamous cell carcinomas.

The tumor microenvironment (TME) not only influences cancer progression but also has been shown to have a significant effect on the prognosis. One of the major TME components is cancer-associated fibroblasts (CAFs), of which several subtypes have been identified in tumor tissues. Although single-cell RNA sequencing (scRNA-seq) technology is a powerful tool for investigating the proportions and composition of cells in tissues, large-scale datasets are needed to analyze small cell populations. Here, we constructed an integrated scRNA-seq dataset for non-small cell lung cancer (NSCLC) by compiling publicly available data and observed differences in TME heterogeneity and cellular composition between lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC). In particular, we identified significant differences in CAF subtypes between LUAD and LUSC. Inflammatory CAFs (iCAFs) were predominantly found in LUAD, whereas myofibroblastic CAFs (mCAFs) were more common in LUSC. A coculture analysis of lung fibroblasts confirmed that LUAD cells induced the iCAF phenotype and, in contrast, that LUSC cells promoted myofibroblastic differentiation. A correlation analysis with the prognosis identified mCAFs as poor prognostic factors in both LUAD and LUSC, but iCAFs were found to be a poor prognostic factor only in LUSC, with the opposite pattern observed in LUAD. This work highlights important considerations regarding the CAF subtype dominance in LUAD and LUSC, which may be related to the prognosis.