The Role of ARID1A in Oncogenic Transcriptional (de)Regulation in Colorectal Cancer
by Madhobi Sen
Date of Examination:2019-01-29
Date of issue:2019-07-23
Advisor:Prof. Dr. Steven A. Johnsen
Referee:Prof. Dr. Steven A. Johnsen
Referee:Prof. Dr. Matthias Dobbelstein
Files in this item
Name:Madhobi Sen_PhD thesis_2018.pdf
Format:PDFDescription:PhD dissertation on the tumor promoting role of the chromatin remodeler ARID1A in colorectal cancer by facilitation of the MEK/ERK pathway.
EnglishThe deregulation of epigenetic modulation has been well established as a common occurrence in cancer. However, the extent of its involvement in the development and progression of cancer was underscored by genome- and exome-wide sequencing studies over the last several years, which revealed a close association between the epigenome and the pathogenesis of cancer. A class of chromatin regulators, the ATP-dependent chromatin remodellers (particularly subunits of the mammalian BAF complex) are among the most frequently mutated genes across cancer types. These factors showed an alteration frequency in over 20% of all cancers. Among the BAF complex subunits, defects in ARID1A (AT-rich interactive domain-containing protein 1A) are the most frequently found and widespread across many cancer types. Mutations in ARID1A are usually frameshift or nonsense mutations which lead to a loss of its expression in these tumors. In fact, ARID1A is also one of the most frequently mutated chromatin regulators in human colorectal cancer (CRC). Several studies in cell culture and mouse models have shown that loss of ARID1A leads to increased proliferation and tumorigenesis in several cancer types, indicating a tumor suppressive function. Very interestingly, a study described the pivotal role of ARID1A in driving CRC in which its inactivation alone led to the formation of invasive adenocarcinomas in mice. Surprisingly, in contrast to the expected tumor suppressive role of ARID1A in CRC, we observed that the knockout (KO) of ARID1A in CRC cell lines leads to impaired proliferation. Moreover, subcutaneous xenografts in SCID mice using human ARID1A KO CRC cells did not form more aggressive tumors than their wildtype counterparts. Also, the generation of several mouse models in the literature of Arid1a deletion revealed that it can have oncogenic functions. These results indicate a context-dependent role of ARID1A in cancer. We observed an impairment in proliferation in two of the four cell lines in which we performed ARID1A KO. Strikingly, these cell lines harbor the KRASG13D mutation. Therefore, we sought to explore the transcriptional role that ARID1A plays downstream of this pathway. To uncover this, we utilized several publicly available ChIP-seq, mRNA-seq and ATAC-seq datasets and generated our own ChIP-seq dataset for ARID1A in the CRC cell line HCT116. We observed a substantial co-localization of the BAF complex with AP1 transcription factors, such as JunD, that act downstream of the MEK/ERK signaling pathway, suggesting cooperation between these factors. Analysis of the sites at which ARID1A binds showed an enrichment of AP1 transcription factor binding sequences. Most sites co-occupied by ARID1A/AP1 are distal to transcription start sites. Therefore, it is likely that these transcriptional regulators functionally interact at enhancers to elicit gene expression changes in CRC. To examine this, we next explored the MEK/ERK pathway. We identified some targets that are co-localized at distal regulatory sites for genes and are downregulated by ARID1A KO, Trametinib (a MEK/ERK pathway inhibitor) treatment and depletion of JunD. Strikingly, the occupancy of JunD and the acetylation of H3K27 (often an active enhancer mark) was also reduced at these distal regulatory sites upon Trametinib treatment and ARID1A KO. Conversely, the occupancy of ARID1A was reduced upon Trametinib treatment or JunD depletion. Thus, these regulatory regions are targets of the MEK/ERK pathway (through AP1) and are dependent on ARID1A as a co-factor. This effect does not seem to be mediated by the known chromatin remodeller functions of ARID1A since the accessibility of chromatin is not affected upon its loss. Thus, we were able to show that ARID1A is required for regulation of KRAS mutation-driven CRC by acting as a co-factor with AP1 transcription factors (TFs) which are downstream of the MEK/ERK pathway at distal regulatory elements. Importantly, this enables the identification of a strategy to stratify CRC by KRAS mutation status and to target the BAF complex in CRCs that are particularly dependent on this pathway.
Keywords: ARID1A, KRAS mutation, MEK/ERK pathway, BAF complex, AP1, Enhancers, Transcriptional regulation, Colorectal cancer