非小細胞肺癌のゲフィチニブ感受性予測診断法の構築と個別化治療への応用

摘要

Gefitinib (Iressa, ZD1839), an inhibitor of epidermal growth factor receptor-tyrosine kinase (EGFR-TK), has shown potent anti-tumor effects and improved symptom and quality-of-life of a subset of patients with advanced non-small cell lung cancer (NSCLC). However, a large proportion of the patients did not respond to this agent. To establish a method to predict the response of NSCLC patients to gefitinib and to develop useful protein markers for practical clinical tests, we initially used a genome-wide cDNA microarray to analyze 33 biopsy samples of advanced NSCLC from patients who had been treated with an identical protocol of second- to 7th-line gefitinib monotherapy. We identified 51 genes whose expression differed significantly between 7 responders and 10 non-responders to the drug. We selected the 12 genes that showed the most significant differences which enabled us to establish a numerical scoring system (GRS, gefitinib response score), for predicting response to gefitinib treatment. The GRS system clearly separated the two groups without any overlap, and accurately predicted responses to the drug in sixteen additional NSCLC cases. The system was further validated by the semi-quantitative RT-PCR for the 12 genes. Furthermore, we validated the effectiveness of antibodies for an immunohistochemical test of 5 biomarkers (AREG, TGFA, ADAM9, CD9, and OSMR), all of which were reported to be cell-surface or secretory proteins and involved in the ligand-EGFR signaling, and whose expression were significantly higher in non-responders than in responders. We were additionally able to detect serum TGFA protein in NSCLC patients by ELISA; about 60% of serum samples from non-responders were positive for TGFA and all samples from responders were negative, strongly implying a possible application of our system to clinical tests. Our results suggested that sensitivity of a given NSCLC to gefitinib can be predicted according to the expression levels of a defined set of genes/proteins that may biologically affect drug sensitivity and the survival of lung-cancer cells. Our scoring system and its development as a routinely available and less invasive diagnostic tool might eventually lead to the achievement of personalized therapy for NSCLC patients.