Abstract C24: LRpath analysis reveals common pathways dysregulated via DNA methylation across cancer types

Abstract

Abstract Since the introduction of the Illumina HumanMethylation27 BeadArray platform, which assesses the % methylation of over 27,000 sites across the genome, several studies have been published testing for genes with aberrant methylation in their promoter regions. Interestingly, a majority of these publicly available datasets are studying cancer, making the time ripe to perform an integrative analysis of these results. Our hypothesis is that during the pathogenesis of cancer, certain pathways or biological gene groups are commonly dysregulated via DNA methylation across cancer types. The ability of our logistic regression-based gene set enrichment method (LRpath) to implicate important biological pathways and groupings has previously been demonstrated and published. Recently, we developed a web-based application for LRpath with greatly expanded and novel gene set annotations, including transcription factor and microRNA target sets not yet published, and clustering analysis functionality, allowing one to identify and compare biological concept signatures across multiple studies. We employed LRpath and clustering analysis to unravel the commonly altered pathways and other biological concepts across multiple cancer studies using DNA methylation data profiled using the Illumina Infinium HumanMethylation27 BeadArray. Using six independent cancer datasets (five solid tumors including breast, colorectal, glioblastoma, lung, and prostate, and one myeloma), we identified genes harboring abberant promoter methylation between normal and cancer samples using an empirical Bayes method. Each gene was assigned a p-value and change in % methylation indicating the level of hyper- or hypomethylation. The gene list in each study was then subjected to LRpath enrichment analysis using multiple annotation databases. To identify the biological concepts commonly altered across various types of cancer, concepts satisfying defined significance criteria were subjected to the clustering analysis using Euclidean distance matrix and ward linkage method. Clustering analysis performed on the results of LRpath identified tightly clustered hypermethylated concepts involved in neurogenesis and epidermis development. It has been previously reported that the group of genes regulating early development is occupied by polycomb repressive complex 2 subunit SUZ12 in human embryonic stem cells. Another tightly clustered group involves voltage-gated potassium channels, which play a role in cell proliferation processes. This result supports previously reported promoter DNA methylation events in breast and pancreas adenocarcinomas. The majority of hypomethylated concepts identified across various cancer studies were related to immune response concepts identified by GO terms and KEGG pathways. These immune-related concepts range from chemokine and cytokine activity, responses to stimulus and inflammation, to receptor binding activities. The DNA hypomethylation events often result in gene upregulation. Thus, this result suggests elevated immune responses as a commonly affected mechanism across multiple cancer types. Interestingly, cell cycle activity, one of the most commonly affected pathways in cancer development, was found to be significantly depleted in both hyper- and hypomethylation across cancer types. Based on this observation, we speculate that genes involved in cell cycle tend to be dysregulated by alternative mechanisms such as genomic aberrations, mutations, and histone modifications, rather than DNA methylation. Here, we performed an integrative analysis of commonly affected biological concepts across six different cancer types. We were able to clearly identify concepts that are affected in cancer regardless of its type and further support biologically important findings. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the Second AACR International Conference on Frontiers in Basic Cancer Research; 2011 Sep 14-18; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2011;71(18 Suppl):Abstract nr C24.