Abstract
Lung cancer is one of the most common malignant solid tumors and the leading cause of cancer-associated mortality worldwide. Endocytosis is an essential physiological activity for cells to maintain membrane homeostasis, and has been reported to serve an important role in tumorigenesis and progression. In the present study, the aim was to construct a prognostic prediction model of endocytosis-associated genes for patients with lung adenocarcinoma (LUAD). The endocytosis-associated gene signature was established using Lasso Cox regression analysis using the training set of the LUAD cohort from The Cancer Genome Atlas (TCGA) database, and verified using two datasets from the Gene Expression Omnibus (GEO) database. Kaplan-Meier survival curves were used to evaluate the effectiveness of the prognostic evaluation of patients with LUAD. Differentially expressed genes were screened in the tumor tissue of patients compared with paired paracancerous tissues. A series of candidate genes associated to the prognosis of patients with LUAD was obtained using univariate Cox's regression analysis. Using the Lasso Cox regression analysis, an appropriate risk model with 18 endocytosis-associated genes was established. A high-risk score was positively correlated with a higher tumor stage and pathologic grade. Patients with LUAD and high-risk scores had shorter survival times, increased intratumor heterogeneities and immune cell infiltration into tumor tissues, compared with those patients with LUAD and low-risk scores. The endocytosis inhibitor chloroquine could repress proliferation and increase the apoptosis of lung cancer cells. In summary, a novel endocytosis-associated gene signature was constructed using TCGA and GEO datasets. Patients with LUAD and high-risk scores, as calculated by the signature, had a poor prognosis and short survival time.