Abstract LB-52: Integrative protein and transcriptomic analysis identified an EMT-associated secretory phenotype that correlates with metastasis and predicts survival in lung adenocarcinoma patients

Abstract

Abstract The epithelial-mesenchymal transition (EMT) confers many clinicaly relevant properties to cancer cells, including migratory and invasive capacity, resistance to apoptosis, drug resistance, evasion of host immune surveillance, and tumor stem cell traits. Therefore, we reasoned that characterizing the secretome of cultured cells in EMT may identify biomarkers which allow monitoring of EMT in tumors and provide a signature to predict metastasis and survival in patients with cancers. Utilizing aTGF-β-induced EMT model with A549 lung adenocarcinoma cells, we harvested serum-free conditioned media from control and EMT cells (5 ng/ml TGF-β, 72h) to quantitatively profile differentially expressed proteins by GeLC-MS/MS and spectral counting. We identified >2000 proteins, >70% of which were putative secretory proteins. Corresponding transcriptomic analysis was performed by Affymetrix U133-plus microarray which identified ∼3,500 genes as differentially expressed. By integrating these two data sets we derived an EMT-associated secretory phenotype (EASP) comprising 97 genes of secretory proteins that were differentially up-regulated by at least two-fold during EMT both for protein and mRNA levels. Several of these proteins were validated by MRM. Analysis of EASP expression in a gene expression dataset from 443 human lung adenocarcinoma tumors revealed strong positive correlations with lymph node metastasis (p<0.05), advanced tumor stage (p=0.006) and histological grade (p<0.00001). After splitting the 443 tumors into a training (n=254) and test (n=189) sets, random forest survival analysis, including age, sex and stage as variables, stratified the test-set into low, medium and high-risk groups with significant differences in overall survival (p=0.0002). For further validation we tested EASP in two independent histologically distinguished lung cancer data sets. Interestingly, EASP predicted survival only in the dataset with adenocarcinomas (n=111, p=0.03), not in the dataset with squamous tumors (n=130). This result indicates that integrative analysis of relevant biological processes in cell-culture models may provide mechanism-related, tumor type-specific, clinically-relevant biomarkers with significant prognostic value. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr LB-52. doi:10.1158/1538-7445.AM2011-LB-52