Differentially Methylated Super-Enhancers Regulate Target Gene Expression in Human Cancer

Emily L Flam,Joseph A Califano,Elena Stavrovskaya,Alexander Favorov,Jiang Qian,Ludmila Danilova,Theresa Guo,Dylan Z Kelley,Daria A Gaykalova,Michael Considine,Elana J Fertig

doi:10.1038/s41598-019-51018-x

Abstract

Current literature suggests that epigenetically regulated super-enhancers (SEs) are drivers of aberrant gene expression in cancers. Many tumor types are still missing chromatin data to define cancer-specific SEs and their role in carcinogenesis. In this work, we develop a simple pipeline, which can utilize chromatin data from etiologically similar tumors to discover tissue-specific SEs and their target genes using gene expression and DNA methylation data. As an example, we applied our pipeline to human papillomavirus-related oropharyngeal squamous cell carcinoma (HPV + OPSCC). This tumor type is characterized by abundant gene expression changes, which cannot be explained by genetic alterations alone. Chromatin data are still limited for this disease, so we used 3627 SE elements from public domain data for closely related tissues, including normal and tumor lung, and cervical cancer cell lines. We integrated the available DNA methylation and gene expression data for HPV + OPSCC samples to filter the candidate SEs to identify functional SEs and their affected targets, which are essential for cancer development. Overall, we found 159 differentially methylated SEs, including 87 SEs that actively regulate expression of 150 nearby genes (211 SE-gene pairs) in HPV + OPSCC. Of these, 132 SE-gene pairs were validated in a related TCGA cohort. Pathway analysis revealed that the SE-regulated genes were associated with pathways known to regulate nasopharyngeal, breast, melanoma, and bladder carcinogenesis and are regulated by the epigenetic landscape in those cancers. Thus, we propose that gene expression in HPV + OPSCC may be controlled by epigenetic alterations in SE elements, which are common between related tissues. Our pipeline can utilize a diversity of data inputs and can be further adapted to SE analysis of diseased and non-diseased tissues from different organisms.

Highlights

Current literature suggests that epigenetically regulated super-enhancers (SEs) are drivers of aberrant gene expression in cancers
We developed a pipeline to study the role of SE methylation on target gene expression
While we used HPV + OPSCC as an example throughout the manuscript, the pipeline can be applied to any cohort that has DNA methylation and gene expression data available for the same samples

Summary

Introduction

Current literature suggests that epigenetically regulated super-enhancers (SEs) are drivers of aberrant gene expression in cancers. We develop a simple pipeline, which can utilize chromatin data from etiologically similar tumors to discover tissue-specific SEs and their target genes using gene expression and DNA methylation data. We applied our pipeline to human papillomavirus-related oropharyngeal squamous cell carcinoma (HPV + OPSCC) This tumor type is characterized by abundant gene expression changes, which cannot be explained by genetic alterations alone. Super-enhancers (SEs) are tissue- and disease-specific regulatory genomic elements related to chromatin that drive cell-specific gene expression changes in development, differentiation, and disease progression, including cancer[1,2]. While SEs are currently recognized as one of the main drivers of carcinogenesis, many tumor types are still missing information about SE locations This lack of information can be explained by the challenges of chromatin analysis procedures on clinical biopsy tissues[20]. We built our pipeline under the assumption that maximum gene expression occurs under the condition that both the gene promoter region and nearby SE region are hypomethylated, while hypermethylation signal from either the promoter or the SE region could diminish target gene expression (Fig. 1 and refs[8,15,19])

Methods

Results

Conclusion