Enhancer Landscapes Reveal Transcription Factor Network Dependencies in Chronic Lymphocytic Leukemia

Genome sequencing efforts of chronic lymphocytic leukemia have revealed mutations that disrupt protein-coding elements of the genome (Puente et al, 2011; Wang et al, 2011; Landau et al, 2013). Recently, comprehensive whole-genome sequencing efforts have begun to reveal the genetic aberrations that occur outside of protein-coding exons, many that may perturb gene regulatory sites (Puente et al, 2015). These include enhancer elements that make physical contact with gene promoters to regulate gene expression in a cell-type specific manner. While mutations certainly promote CLL leukemogenesis, epigenomic alterations may also play an important role in facilitating disease progression and maintenance by inducing the gene expression aberrations that have long been observed in CLL. Epigenomic alterations include chromatin structure changes that facilitate altered transcription and chromatin factor recruitment to regulatory elements. While comprehensive genome-wide DNA methylation studies have been performed on human cancers and normal cell counterparts including CLL, other comprehensive studies of cancer epigenomes have been lacking. We have completed an analysis of chromatin structures in a cohort of primary chronic lymphocytic leukemia (CLL) samples with comparisons to normal CD19+ B lymphocytes (n = 18 CLL samples, n = 5 normal B lymphocyte samples). We used chromatin accessibility assays (ATAC-seq) and genome-wide enhancer mapping (H3K27ac ChIP-seq) to comprehensively define the transcriptionally active chromatin landscape of CLL. We have discovered greater than 15,000 novel regulatory elements when compared to previously annotated regulatory elements. Moreover, sites within the loci of several hundred genes were found to have large regions of gained chromatin accessibility and H3K27 acetylation, revealing the appearance of aberrant enhancer activity. These gained enhancer elements correspond with increased gene expression and are found at gene loci such as LEF1, PLCG1, CTLA4, and ITGB1. We have also systematically identified the super-enhancers of CLL – large complex regulatory regions that possess unique tissue-specific regulatory capabilities. Many of these super-enhancers are found in normal B lymphocytes, yet the super-enhancer at the ITGB1 and LEF1 loci are CLL-specific and may be considered to facilitate leukemia-specific expression. We have found CLL-specific enhancers are also significantly associated with annotated CLL risk variants, and have identified enhancer-associated SNPs found within CLL-risk loci predicted to disrupt transcription factor binding sites. These include SNPs at the IRF8 and LEF1 loci that lead to the creation and destruction of SMAD4 and RXRA binding sites, respectively. Additionally, we have analyzed whole-genome sequencing data from a subset of our sample cohort. Mutational hotspots in the CXCR4 and BACH2 promoters occur within open, acetylated regions. Moreover, we discover recurrent mutations in enhancers of the ETS1 and ST6GAL1 locus that have not been previously annotated. Using a transcription factor network modeling approach, we used these global chromatin structure characteristics to determine networks that are highly active in CLL. We find that transcription factors such as NFATc1, E2F5, and NR3C2 are among the most interconnected transcription factors of the CLL genome, and their connectivity is significantly higher in CLL cells compared to normal B cells. In contrast, network profiling of CLL cells predicts loss of MXI1 connectivity, a negative regulator of the MYC oncogene. By treating cells with specific pharmacological inhibitors of NFAT family members including cyclosporin and FK506, we are able to reduce NFAT-mediated network connectivity, resulting in a selective loss of NFAT-bound enhancers. This leads to CLL cell death in vitro of both cell lines and primary CLL patient samples. Our results reveal the unique chromatin structure landscape of CLL for the first time, and identify the CLL-specific enhancer elements that confer the transcriptional dysregulation that has long been observed in this disease. Use of these chromatin structure analyses and enhancer landscapes has allowed us to construct the intrinsic transcription factor network of CLL, and determine a particular dependency on NFAT signaling for cell survival.