The role of SWI/SNF chromatin remodelers in the repair of DNA double strand breaks and cancer therapy

Abstract

Switch/Sucrose non-fermenting (SWI/SNF) chromatin remodelers hydrolyze ATP to push and slide nucleosomes along the DNA thus modulating access to various genomic loci. These complexes are the most frequently mutated epigenetic regulators in human cancers. SWI/SNF complexes are well known for their function in transcription regulation, but more recent work has uncovered a role for these complexes in the repair of DNA double strand breaks (DSBs). As radiotherapy and most chemotherapeutic agents kill cancer cells by inducing double strand breaks, by identifying a role for these complexes in double strand break repair we are also identifying a DNA repair vulnerability that can be exploited therapeutically in the treatment of SWI/SNF-mutated cancers. In this review we summarize work describing the function of various SWI/SNF subunits in the repair of double strand breaks with a focus on homologous recombination repair and discuss the implication for the treatment of cancers with SWI/SNF mutations.

Keywords: DNA end resection; PARP inhibitors; SWI/SNF; cancer therapy; chromatin remodelers; double strand break repair; homologous recombination.

FIGURE 1

Switch/Sucrose non-fermenting (SWI/SNF) chromatin remodelers. There are three forms of the SWI/SNF chromatin remodelers in somatic mammalian cells. The canonical BAF (cBAF, BRG1-associated factors), the polybromodomain BAF (PBAF), and the non-canonical BAF (ncBAF). These complexes can contain either the BRG1 or BRM ATPase (red). cBAF and PBAF complexes contain a group of core subunits (BAF47, BAF57, BAF60, BAF155, BAF170, orange) and a number of accessory factors of unknown function (yellow). The ncBAF only contains BAF60 and BAF155 core subunits. PBAF contains the polybromodomain protein PBRM1 (BAF180) and BRD7, while ncBAF contains BRD9. Subunits marked with an asterisk (*) are known to be mutated in a variety of cancers.

FIGURE 2

SWI/SNF subunits are mutated in cancer. A heatmap representing the mutation frequency for individual SWI/SNF subunits across different cancer types. The data was obtained from The Cancer Genome Atlas (TCGA Pan-Cancer Atlas study, n = 10,967 samples) through cBioPortal (Cerami et al., 2012; Gao et al., 2013). The top four cancer types with a mutation frequency higher than 2% were selected for each subunit. Supplementary Table S1 provides the mutation frequencies.

FIGURE 3

BRG1 and BRM domain structures. BRG1 and BRM are the sole catalytic subunits within these complexes and are mutually exclusive within the complex. Each ATPase contains a QLQ protein interaction domain (QLQ, blue), a helicase SANT-associated domain (HSA, red), the ATPase/helicase domain (yellow), a SNF2 ATP-coupling (SnAC, orange) domain, two A-T hook motifs (light green), and a bromodomain (Bromo, green). Domain structure information obtained from NCBI and cBioPortal.

FIGURE 4

AT rich interacting domain (ARID)-containing proteins domain structures: ARID1A, ARID1B, ARID2. ARID1A contains four LXXLL domains (orange) and an ARID domain (blue). ARID1B contains two LXXLL domains (orange) and one ARID domain (blue) and ARID2 contains one ARID domain (blue), one LXXLL domain (orange), and a regulatory factor binding to the X-box (RFX, purple) DNA binding domain. Domain structure information obtained from NCBI and cBioPortal.

FIGURE 5

Domain structure of bromodomain-containing proteins: PBRM1, BRD7, BRD9. PBRM1 (BAF180) contains six bromodomains (green), two bromo-adjacent homology domains (BAH, yellow), and one high mobility group (HMG, blue) DNA binding domain. BRD7 and BRD9 contain one bromodomain each (green). Domain structure information obtained from NCBI and cBioPortal.

FIGURE 6

Proposed model for the role of SWI/SNF complexes in HR. After the recognition of the DSB by the MRN complex and the ATM kinase, DNA damage signaling is initiated by ATM and there is a chromatin remodeling step mediated by a BRG1-containing SWI/SNF complex that reduces nucleosome density at the DSB (Shen et al., 2015; Velez-Cruz et al., 2016; Chen et al., 2019; Hays et al., 2020). We propose that this step likely results in the eviction of these nucleosomes at the DSB, which likely contain γH2AX (red) and that this chromatin remodeling step stimulates DNA end resection by stimulating or stabilizing the recruitment of the CtIP nuclease to the break site. We also propose that the ARID1A subunit is likely responsible for anchoring the SWI/SNF complex at the DSB, as inactivation of this subunit results in very similar repair defects as those observed upon the inactivation of BRG1 (Shen et al., 2015; Hays et al., 2020). This model is in agreement with the fact that ATM signaling is not affected by the absence of BRG1, but ATR signaling is attenuated upon BRG1 inactivation due to the defect in DNA end resection. This model is also in agreement with work showing that nucleosomes block DNA end resection and that the processes of resection and the removal of nucleosomes are coupled (Mimitou et al., 2017; Wiest et al., 2017; Peritore et al., 2021).