Prediction of prognosis for small cell lung cancer based on genome-wide methylation analyses with surgical materials and robust clustering methods

Methylation is closely involved in the development of various carcinomas. However, little datasets are available for small cell lung carcinoma (SCLC) due to the scarcity of fresh tumor samples. The aim of this study is to investigate the comprehensive genome-wide methylation profile of SCLC to predict the prognosis after surgical treatment. We investigated the high DNA methylated and low gene expression sites using 25 SCLC tumor tissues. First, we selected most differentially methylated CpG sites across the tumor tissues. Following hierarchical clustering (HC) and non-negative matrix factorization (NMF), gene ontology analysis was performed using DAVID software. Clustering of SCLC tumors led to the important identification of a CpG island methylator phenotype (CIMP) of SCLC, and showed that CIMP-high tumors had a significantly poorer prognosis (p = 0.001). Multivariate analysis revealed that postoperative chemotherapy, low neuroendocrine expression and the CIMP-low state were significantly good prognostic factors. Association analyses of methylation and gene expression provided 46 genes with significant correlation. Ontology studies to these genes showed that genes involved in extrinsic apoptosis pathway were suppressed, including TNFRSF1A, TNFRSF10A and TRADD, in CIMP-high tumors, prognosis of which was poorer. By comprehensive DNA methylation profiling, two distinct subgroups were identified to evoke a CIMP of SCLC as a useful marker for determination of treatment. Delineation of this phenotype may also be useful for the development of novel apoptosis-related chemotherapeutic agents for the treatment of an aggressive subtype of SCLC. Comprehensive genome-wide methylation analyses