Abstract
Molecules involved in crosstalk between tumor cells and fibroblasts play vital roles in tumor progression. Extracellular matrix proteins, whose abundance is altered after being affected by tumor-derived exosomes, possess considerable promise as biomarkers for diagnosis or prognosis. In this study, quantitative proteomics is employed to determine the abundance of proteins secreted by normal fibroblasts and exosome-activated fibroblasts, which first identify differentially secreted proteins affected by lung cancer cell-derived exosomes. Based on the differentially secreted proteins and multiple independent datasets comprising 1897 patient samples with non-small cell lung carcinoma or other lung diseases, a diagnostic marker is identified that can effectively distinguish tumor tissues from normal tissue, as well as tumor-associated stroma from normal stroma, and a five-gene prognostic signature is presented with independent prognostic impact to identify patients who may require further adjuvant therapy after surgical resection. In addition, the secretome provides novel potential targets for clinical treatment.