Comprehensive investigation of chromatin regulators as master regulators in human cancer

Abstract

Chromatin regulators (CRs), a class of chromatin modifying enzymes, can shape the epigenetic state to control chromatin structure and function in a cell context-dependent fashion. Somatic mutation or dysregulation of CRs has been reported to be associated with a wide range of human diseases, especially cancer. The current study aims to comprehensively investigate CRs at pan-cancer level, which would help to not only elucidate the pathogenic mechanism of malignant tumors but also design novel diagnostic and therapeutic strategies.

Firstly, an annotation and visualization database CR2Cancer for CRs in cancer was built through high throughput data analysis and literature mining. It contains: (1) genomic, transcriptomic, proteomic, clinical and functional information on over 400 CRs across multiple cancer types; (2) diverse types of CR-associated relations, including cancer type dependent (CR-target and miRNA-CR) and independent (protein-protein interaction and drug-target) ones; (3) around 6000 items of aberrant molecular change (mutation and dysregulation) and interactions of CRs in tumorigenesis manually curated from 5007 publications. The CR2Cancer database provides a web interface for users to conveniently browse, search and download data of interest.

Secondly, we integrated multi-dimensional omics data to identify aberrant CRs affected by genomic alterations across eight cancer types. Master regulator analysis was further applied to prioritize the likely driver CRs responsible for cancer initiation and progression in each cancer type. We have taken liver cancer as an example to prove that 14 (82.4%) of the top ranked 17 driver CRs were critical for liver cancer development by publication or experiment validation. Moreover, we developed an R package based on differential connectivity to explore the modulation role of CRs in transcriptional regulation during tumorigenesis. Our results revealed that connectivity between transcription factors and their targets tends to be disrupted in the patients who have a high expression of oncogenic CRs or low expression of tumor-suppressive CRs.

Lastly, we streamlined cancer master regulator (MR) analysis pipeline, and developed a web server MR4Cancer to identify and prioritize MRs driving a cancer-related phenotypic divergence. Based on the input gene list (e.g. differentially expressed genes) or expression profile with two groups, MR4Cancer outputs ranked MRs by enrichment testing. Gene Ontology and canonical pathway analyses are also conducted to elucidate the function of likely MRs. Moreover, our server provides dynamic network visualization tool for MR-target relations, and users can interactively interrogate the network to generate new hypotheses and high-quality figures for their publication.

In summary, the current study explores the oncogenic role of CRs as well as their effectors during cancer development, and provides two valuable online tools, CR2Cancer and MR4Cancer, for biological community.