Expression signatures in lung cancer reveal a profile for EGFR-mutant tumours and identify selective PIK3CA overexpression by gene amplification.

The development of targeted therapies creates a need to discriminate tumours accurately by their histological and genetic characteristics. Here, we aim to identify gene expression profiles and single markers that recapitulate the pathological and genetic background of non-small cell lung cancer (NSCLC). We performed cDNA microarray analysis on a series of 69 NSCLCs, with known mutation status for important genes, and six normal lung tissues. Unsupervised cluster analysis segregated normal lungs from lung tumours and lung tumours according to their histopathology and the presence of EGFR mutations. Several transcripts were highly overexpressed (by approximately 20 times) in squamous cell carcinomas (SCCs) relative to adenocarcinomas (ACs) and confirmed by immunohistochemistry in an independent cohort of 75 lung tumours. Expression of 13 genes constituted the most prominent hallmarks of EGFR-mutant tumours, including increased levels of proline dehydrogenase (PRODH) and down-regulation of X-box binding protein 1 (XBP1). No genes were differentially expressed, with a fold change >or= 4 or <or=0.25 and a significance level of 5% false-discovery rate, in tumours carrying mutations of TP53 or KRAS. In addition, we organized gene expression data by the position of each gene in the chromosome and observed a cluster of highly expressed genes in chromosome 3q, including PIK3CA, that was characteristic of SCCs. FISH analysis demonstrated a strong statistically significant association between increased levels of PIK3CA expression in these tumours and gene amplification (p < 0.0001; t-test). In conclusion, histopathological phenotypes and, likely, the presence of EGFR mutations confer lung tumours with a marked pattern of gene expression. Moreover, our cDNA microarray analysis identified increased PIK3CA expression due to gene amplification in lung squamous cell carcinomas: this may represent a marker of sensitivity to therapy with PI3K inhibitors. Copyright (c) 2007 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.