To identify new potential diagnostic markers for lung cancer, the expression profiles of 37 lung tumours were analysed using cDNA arrays. Seven samples were from small-cell lung cancer (SCLC), two from large-cell neuroendocrine tumours (LCNEC), and 28 from other non-small-cell lung cancers (mainly squamous cell cancer and adenocarcinoma). Principal component analysis and the permutation test were used to detect differences in the gene expression profiles and a set of genes was found that distinguished high-grade neuroendocrine carcinomas (SCLC and LCNEC) from other lung cancers. In addition, several genes, such as caveolin-1 (CAV1) and caveolin-2 (CAV2), were constantly deregulated in all types of tumour sample, compared with normal tissue. The expression of these two genes was investigated further at the protein level on a tissue microarray containing tumours from 161 patients and normal tissues. Immunostaining for CAV1 was negative in 48% of tumours, whereas 28% of the tumours did not express CAV2. Lack of CAV1 protein expression was not caused by methylation or mutation. In stage I adenocarcinomas, CAV2 protein expression correlated with shorter survival. In conclusion, the present study was able to identify genes that have not previously been implicated in lung cancer by the combined use of two different array techniques. Some of these genes may provide novel diagnostic markers for lung cancer.
Copyright 2004 Pathological Society of Great Britain and Ireland. Published by John Wiley & Sons, Ltd.