ABSTRACT: The BAF (mammalian SWI/SNF) chromatin remodeling complexes play a fundamental role in nucleosome positioning and transcription factor binding. Despite the well-established role of the three BAF complexes in chromatin remodeling (cBAF, PBAF and ncBAF), no common mechanism has been identified that would explain widespread occurrence of BAF complex loss-of-function mutations across cancers and genetic disorders in diverse tissue types. Given that BAF complex regulates accessibility of both ubiquitous and tissue-specific transcription factors, we set out to determine whether a unifying chromatin regulatory mechanism of BAF remodeling across tissues exists. To investigate this, we profiled chromatin accessibility after inhibiting all BAF complexes across a panel of cancer-derived and non-cancer cell lines, encompassing lymphoid, myeloid, neuronal, epithelial, embryonic, and other lineages. We found that the only commonality between tissues was a striking increase in chromatin accessibility at CTCF sites, observed in most tissue types. This effect was most pronounced in lymphoid and myeloid tissues, tissues that also exhibited the highest baseline levels of CTCF. Chromatin immunoprecipitation in these tissues revealed increased CTCF and cohesin (SMC3) binding, demonstrating a previously unknown antagonistic binding relationship between cBAF and CTCF-cohesin complexes on chromatin. Genetic deletion of the cBAF-specific subunit ARID1A reproduced the chromatin accessibility changes observed upon pan-BAF complex inhibition, demonstrating that cBAF loss is the principal driver of increased CTCF accessibility. To determine the functional consequences of the observed CTCF gain, we investigated its impact on topologically associated domains (TADs). High-resolution chromatin capture (Micro-C) showed strengthening of existing TAD boundaries, increasing intra-domain interactions, and splitting larger domains into smaller, more insulated structures, demonstrating how BAF loss reshapes 3D chromatin architecture. To identify remodelers that help CTCF recruitment to chromatin in absence of cBAF, we identified CTCF interacting complexes and found that CTCF directly interacts with fully assembled PBAF, but not cBAF, stabilizing CTCF binding in the absence of cBAF. Furthermore, we show that upon cBAF loss, the PBAF complex is recruited to sites of increased CTCF occupancy, showing for the first time a functional shift in chromatin regulation by the antagonism of these remodelers. We further determined the functional consequence of this antagonism in vivo. Conditional Arid1a deletion in primary mouse B cells led to the expansion of highly proliferative cells with increased CTCF binding. To determine whether CTCF drives this phenotype, we generated a dual knockout model by crossing Arid1a- and Ctcf-conditional knockout mice. Strikingly, Ctcf deletion fully reversed the proliferative phenotype of Arid1a loss, establishing a direct functional in vivo link between ARID1A loss and CTCF gain. Finally, linking this phenotype to 3D chromatin structure, we found that loss of cBAF rewires chromatin looping by converting distal enhancer–promoter interactions into shorter, promoter-proximal loops via increased CTCF binding. This shift in loop anchors alters transcriptional programs, underscoring a unified mechanism by which cBAF loss redefines 3D chromatin organization and gene regulation
