Genomic copy number alterations in non-small cell lung cancers identified using CBS and MSA

Abstract

7695 Background: Lung cancer is the leading cause of cancer death in both men and women with a 5-yr survival rate of 15.5%. Previous studies have begun to characterized genomic copy number alterations of non-small-cell lung CA using array CGH and SNP arrays. We have used aCGH using our 1MB BAC arrays and two algorthims for making copy number alteration (CNA) determinations. We have also pursued the exact copy number gains and losses of several genes using Q-PCR. Methods: Genomic DNA from fresh frozen tumors of 27 patients with NSCLC. We performed aCGH using 1MB Arrays. We used CBS, and MSA to identify regions of CNA. We further pursued several genes of interest (including HRAS, CRK, and CDC42) identified using Q-PCR. Unsupervised hierarchical clustering was performed to look for distinct subgroups. Significant Analysis of Microarrays (SAM) was applied to identify the association between CNAs clinical parameters including tumor subtype, gender, lymph node involvement, tumor stage, and overall survival. Results: 240 regions of amplification and 181 regions of deletions were found, and included all previously published regions implicated in lung cancer. CNAs in > 70% of tumors included amplifications in 1q, 3q, 5p, 6p, 11p, 16q, 20q, and Xq, and deletions in 1p, 8p and 13q. We verified CNAs of HRAS, CRK and CDC42 using Q-PCR. Hierarchical clustering revealed 2 subgroups: one with amplifications in 2q, 4p, 4q, 8q, 21q, 15q, and 16p, and the other with amplifications in 3q, and 5q. These were confirmed by supervised SAM analysis. Using SAM we found that gain of 2q, 4p and 10q, and loss of 16p and 19q were significantly present in adenocarcinomas. (q = 0, FDR = 0%). Gain of 10q, and loss of 6p and 14q were associated with female gender. (q = 0, FDR = 0%). Conclusions: We used aCGH to identify CNAs that characterize non-small cell lung CA tumors with the aim of finding key regions which may harbor important oncogenes and tumor suppressors. Several regions of CNA have been identified, several of which have been associated with clinical parameters. Because much heterogeneity exists in non-small-cell lung tumors, we have demonstrated that clustering analysis is useful in identifying subtypes which may possess prognostic and therapeutic significance. No significant financial relationships to disclose.