Epigenetic and genetic alterations and their influence on gene regulation in chronic lymphocytic leukemia

Di Huang, Ivan Ovcharenko

To understand the changes of gene regulation in carcinogenesis, we explored signals of DNA methylation - a stable epigenetic mark of gene regulatory elements - and designed a computational model to profile loss and gain of regulatory elements (REs) during carcinogenesis. We also utilized sequencing data to analyze the allele frequency of single nucleotide polymorphisms (SNPs) and detected the cancer-associated SNPs, i.e., the SNPs displaying the significant allele frequency difference between cancer and normal samples. After applying this model to chronic lymphocytic leukemia (CLL) data, we identified REs differentially activated (dREs) between normal and CLL cells, consisting of 6,802 dREs gained and 4,606 dREs lost in CLL. The identified regulatory perturbations coincide with changes in the expression of target genes. In particular, the genes encoding DNA methyltransferases harbor multiple lost-in-cancer dREs and zero gained-in-cancer dREs, indicating that the damaged regulation of these genes might be one of the key causes of tumor formation. dREs display a significantly elevated density of the genome-wide association study (GWAS) SNPs associated with CLL and CLL-related traits. We observed that most of dRE GWAS SNPs associated with CLL and CLL-related traits (83%) display a significant haplotype association among the identified cancer-associated alleles and the risk alleles that have been reported in GWAS. Also dREs are enriched for the binding sites of the well-established B-cell and CLL transcription factors (TFs) NF-kB, AP2, P53, E2F1, PAX5, and SP1. We also identified CLL-associated SNPs and demonstrated that the mutations at these SNPs change the binding sites of key TFs much more frequently than expected. Through exploring sequencing data measuring DNA methylation, we identified the epigenetic alterations (more specifically, DNA methylation) and genetic mutations along non-coding genomic regions CLL, and demonstrated that these changes play a critical role in carcinogenesis through damaging the regulation of key genes and alternating the binding of key TFs in B and CLL cells.