Gene identification for risk of relapse in stage I lung adenocarcinoma patients: a combined methodology of gene expression profiling and computational gene network analysis

Abstract

// Vienna Ludovini 1, * , Fortunato Bianconi 2, * , Annamaria Siggillino 1 , Danilo Piobbico 2 , Jacopo Vannucci 3 , Giulio Metro 1 , Rita Chiari 1 , Guido Bellezza 4 , Francesco Puma 3 , Maria Agnese Della Fazia 2 , Giuseppe Servillo 2 , Lucio Crinò 1 1 Medical Oncology, S. Maria Della Misericordia Hospital, Perugia, Italy 2 Department of Experimental Medicine, University of Perugia, Perugia, Italy 3 Department of Surgical and Biomedical Science, University of Perugia, Perugia, Italy 4 Department of Experimental Medicine, Section of Anatomic Pathology and Histology, Perugia, Italy * These authors have contributed equally to this work Correspondence to: Vienna Ludovini, email: oncolab@hotmail.com Fortunato Bianconi, email: fortunato.bianconi@gmail.com Keywords: lung adenocarcinoma, gene expression profiling, gene networks, computational biology, cancer systems biology Received: December 17, 2015     Accepted: March 28, 2016     Published: April 13, 2016 ABSTRACT Risk assessment and treatment choice remains a challenge in early non-small-cell lung cancer (NSCLC). The aim of this study was to identify novel genes involved in the risk of early relapse (ER) compared to no relapse (NR) in resected lung adenocarcinoma (AD) patients using a combination of high throughput technology and computational analysis. We identified 18 patients (n.13 NR and n.5 ER) with stage I AD. Frozen samples of patients in ER, NR and corresponding normal lung (NL) were subjected to Microarray technology and quantitative-PCR (Q-PCR). A gene network computational analysis was performed to select predictive genes. An independent set of 79 ADs stage I samples was used to validate selected genes by Q-PCR. From microarray analysis we selected 50 genes, using the fold change ratio of ER versus NR. They were validated both in pool and individually in patient samples (ER and NR) by Q-PCR. Fourteen increased and 25 decreased genes showed a concordance between two methods. They were used to perform a computational gene network analysis that identified 4 increased (HOXA10, CLCA2, AKR1B10, FABP3) and 6 decreased (SCGB1A1, PGC, TFF1, PSCA, SPRR1B and PRSS1) genes. Moreover, in an independent dataset of ADs samples, we showed that both high FABP3 expression and low SCGB1A1 expression was associated with a worse disease-free survival (DFS). Our results indicate that it is possible to define, through gene expression and computational analysis, a characteristic gene profiling of patients with an increased risk of relapse that may become a tool for patient selection for adjuvant therapy.