Single-cell profiling and clinical characteristics analysis of lung squamous carcinoma

Lung squamous carcinoma (LUSC) is a highly heterogeneous disease. However, the tumor microenvironment (TME) landscape and clinical characteristics for LUSC have not yet been elucidated. To map the TME and clinical characteristics of LUSC, we performed single-cell RNA sequencing for 504 LUSC samples on basis of TCGA and Gene Expression Omnibus. We introduced the computational algorithms "ESTIMATE" and "CIBERSORT" to analyze immune cell infiltration and immune-checkpoint-related gene signatures in various LUSC clusters. Weighted gene co-expression network analysis was used to explore the connections between molecular characteristics and clinical traits in LUSC. A prognostic model was constructed by performing multivariate COX. Two gene clusters exhibiting disparate immune and clinical characteristics were identified. Our findings indicate that patients in cluster 2, who have a more favorable prognosis, exhibit immune characteristics such as elevated levels of immunosuppression-associated M2 macrophages, resting memory CD4 T cells, resting dendritic cells (DC), and TNFRSF4, alongside reduced infiltration of activated DC and lower expression of TNFRSF18.Whereafter, the Risk Score model was built on basis of 3-DEGs signature consisted of cystatin C (CST3), transglutaminase type 2 (TGM2), JUN, which were proved by q-PCR and immunofluorescence. Besides, high-Risk Score may be responsible for poor prognosis in LUSC patients. Our study identified that tumor-infiltrating immune cell subtypes and the Risk Score model might shed light on the heterogeneity in LUSC patients. The TME, three DEGs and Risk Score can effectively serve as biomarkers to elucidate the immune landscape and predict prognosis in LUSC patients. They may provide insights to the investigations on therapeutic strategies for LUSC.