The Many Roles of BAF (mSWI/SNF) and PBAF Complexes in Cancer

Abstract

During the last decade, a host of epigenetic mechanisms were found to contribute to cancer and other human diseases. Several genomic studies have revealed that ∼20% of malignancies have alterations of the subunits of polymorphic BRG-/BRM-associated factor (BAF) and Polybromo-associated BAF (PBAF) complexes, making them among the most frequently mutated complexes in cancer. Recurrent mutations arise in genes encoding several BAF/PBAF subunits, including ARID1A, ARID2, PBRM1, SMARCA4, and SMARCB1 These subunits share some degree of conservation with subunits from related adenosine triphosphate (ATP)-dependent chromatin remodeling complexes in model organisms, in which a large body of work provides insight into their roles in cancer. Here, we review the roles of BAF- and PBAF-like complexes in these organisms, and relate these findings to recent discoveries in cancer epigenomics. We review several roles of BAF and PBAF complexes in cancer, including transcriptional regulation, DNA repair, and regulation of chromatin architecture and topology. More recent results highlight the need for new techniques to study these complexes.

Figure 1.

Homology between BAF and PBAF-like remodelers throughout evolution. (A) BAF and PBAF complexes in mammals share several features with Brahma-associated proteins (BAPs) and Polybromo-associated BAP (PBAP) complexes (Drosophila melanogaster), and SWI/SNF and RSC complexes (Saccharomyces cerevisiae), respectively. The similarities and differences between these complexes throughout evolution provide insight into their biological regulation and their roles in cancer. BAF/PBAF subunits labeled in a boldface white font have important roles in malignancy. Time since species divergence was estimated using TimeTree (Hedges et al. 2006), and plotted as a function of time, millions of years ago (Mya). (B) Summary of BAF and PBAF subunits and alternative names used in the text. In some cases, abbreviated names rather than the official human genome organization (HUGO) symbols are used in the text because of space constraints.

Figure 2.

The family of human Snf2-like ATPases and their differing roles in cancer. (A) Human Snf2-like ATPases cluster into groups based on degree of sequence similarity. The chromatin remodelers from model organisms are shown near these groups in bold. Radial dendrogram constructed using TreeDyn (Chevenet et al. 2006). (B) Human Snf2-like ATPases are mutated at different frequencies across all cancer types. The total number of mutations appearing in cBioPortal (including public datasets from The Cancer Genome Atlas (TCGA), Cancer Cell Line Encyclopedia (CCLE), and others cited in the text) is summed for each gene and presented by the type of mutation. Missense mutations predicted to have neutral, low, or medium functional impact are not shown because of the unknown nature of their effects and increased likelihood to be background mutations. (C) The number of mutations of each BAF/PBAF subunit is presented as in B. ARID1A and PBRM1 frequently undergo truncating mutation, but BRG (SMARCA4) frequently has missense mutations with high functional impact.

Figure 3.

Mutations of BAF and PBAF subunits occur in subunit-specific patterns in cancer. (A) (Left panel) Illustration of the different types of genetic and epigenetic disruptions that affect BAF/PBAF subunits in cancer. Deletion of chromosome arms or foci leads to loss of a subunit allele, point mutations alter coding sequence, gene fusions lead to altered function, and hypermethylation of promoters associated with loss of expression. (Right panel) Mechanisms leading to altered BAF/PBAF subunit expression in cancer may also include mutations in enhancers, loss of insulated neighborhoods leading to spreading of heterochromatin over BAF/PBAF genes, enhancer hijacking, and antisense silencing. (B) Heat map of the frequency of subunit alterations across cancer types (frequency includes all nonsilent mutations, biallelic deletions, and gene fusions). Mutation frequencies for malignant rhabdoid tumor (MRT) and synovial sarcoma are inferred from available cytogenetic and mutation data, as described by works cited in the main text. All other data obtained from studies cited in the main text. PCNSL, Primary central nervous system lymphoma; DLBCL, diffuse large B-cell lymphoma; AML, acute myeloid leukemia.

Figure 4.

Cancer mutations of BAF/PBAF subunits arise in characteristic patterns. ARID1A, ARID2, and PBRM1 are primarily affected by truncating mutations. The ATPases BRG and BRM show a high tendency for missense mutations at the conserved Snf2-like ATPase domains. Missense mutations predicted to have neutral, low, or medium functional impact are not shown because of the unknown nature of their effects and increased likelihood to be background mutations. N-term, Amino terminal; C-term, carboxy terminal.

Figure 5.

The effects of BAF and PBAF dysfunction in cancer. Dysfunctional BAF/PBAF complexes have been shown to deregulate Polycomb silencing of key tumor suppressors and oncogenes. In model systems, disruption of BAF- and PBAF-like complexes also affects DNA accessibility for transcription and other regulatory factors, and impacts splicing patterns. Given the conserved regulatory roles for BAF- and PBAF-like remodelers in DNA repair, maintenance of chromatin topology and 3D architecture, we anticipate that whole-genome sequencing and new techniques to examine 3D-chromatin architecture may reveal new roles for the complex in addition to its well-defined role as a transcriptional regulator.