SMARCB1 loss activates patient-specific distal oncogenic enhancers in malignant rhabdoid tumors

The study investigates how loss of the BAF complex subunit SMARCB1 drives malignant rhabdoid tumor (MRT) tumorigenesis through epigenetic mechanisms rather than additional recurrent genetic alterations. MRTs, aggressive pediatric cancers with low mutational burden, are largely defined by bi-allelic SMARCB1 inactivation, implicating chromatin remodeling defects as central drivers. Given prior evidence that reconstituting SMARCB1 can normalize MRT cell states, the authors hypothesize that SMARCB1 loss reprograms enhancer landscapes and long-range chromatin interactions to aberrantly activate oncogenes such as MYC. Using patient-derived organoids (PDOs) that preserve tumor epigenetic features, the study aims to define patient-specific enhancer reorganization, its impact on 3D genome topology, and functional consequences for MYC regulation and tumor growth.

• BAF/SWI–SNF chromatin remodeling complexes (CBAF, PBAF, ncBAF) regulate transcription via nucleosome positioning; SMARCB1 is present in CBAF/PBAF but absent in ncBAF. Mutations in BAF components are frequent across cancers (≈20–25%).
• MRTs are uniquely characterized by bi-allelic SMARCB1 loss as the dominant genetic event, suggesting epigenetic mechanisms drive oncogenesis.
• Prior studies implicate enhancer reprogramming in cancer progression and show that organoids can retain source tissue epigenetic states over passages, enabling personalized epigenomic studies.
• MYC overexpression is a hallmark of MRTs with SMARCB1-dependent expression signatures; however, mechanisms of MYC dysregulation in MRT have been unclear.
• ncBAF (via BRD9) aberrant occupancy at super-enhancers is a proposed vulnerability in SMARCB1-deficient cancers; BRD9 inhibition shows therapeutic promise.
• Super-enhancers can form insulation boundaries and are cohesin-associated; enhancer–promoter looping (including CTCF-independent cohesin-mediated loops) is a key mechanism of distal regulation, including at the MYC locus in other tumor contexts.

• Models: Three MRT patient-derived organoid (PDO) lines (P103, P78, P60). SMARCB1 reconstitution via lentiviral transduction (pLKO.1-UbC-hSMARCB1-blast) versus control (luciferase). Selection with blasticidin; downstream assays post-selection.
• Chromatin accessibility and occupancy: ATAC-seq (triplicates per condition); ChIP-seq for CTCF, RAD21, H3K27ac (generally n=1); double-crosslinked ChIP for SMARCB1; CUT&RUN for BRD9 (ncBAF) and SS18 (CBAF/ncBAF).
• 3D genome: High-resolution in situ Hi-C in P103 (control vs SMARCB1+); loop calling with HICCUPS; compartment and TAD analyses; 4C-seq with MYC promoter viewpoint in P78 and P60 (n=2 biological replicates per line).
• Transcriptomics and function: Bulk RNA-seq (n=2 per condition) to assess MYC and cell-cycle programs; MYC knockdown using shRNA (pLKO.1-puro) and cell viability (CellTiter-Glo). eRNA detection from RNA-seq at enhancer loci.
• Single-cell tumor tissues: Single-cell Multiome (scATAC + scRNA) on 7 MRT patient tissues (post-QC total 14,799 cells; median ~2,040 per tumor). Cell type assignment by marker genes and absence of SMARCB1. Pseudo-bulk scATAC profiles at MYC and distal enhancers.
• Pharmacology: ncBAF inhibition with BRD9 inhibitor I-BRD9 (10 μM) in PDOs; morphological assessment, RT-qPCR for MYC, RNA-seq, CUT&RUN for BRD9 and SS18 post-treatment.
• Computational analyses: Standard pipelines (TopHat2/DESeq2 for RNA-seq; BWA/Bowtie2, deepTools for ATAC/ChIP/CUT&RUN; HiC-Pro/GENOVA for Hi-C; peakC for 4C). Motif enrichment with GimmeMotifs; GREAT for functional annotation. K-means clustering of lost OCRs to define clusters (K1–K3).

• SMARCB1 reconstitution in MRT PDO P103 remodeled chromatin accessibility: 7,941 gained open chromatin regions (OCRs) enriched for SMARCC1/2 and AP-1 motifs; 1,211 lost OCRs enriched for CTCF motifs.
• Lost OCRs showed decreased CTCF and cohesin (RAD21) binding and reduced H3K27ac; gained OCRs showed increased H3K27ac and RAD21 occupancy without CTCF enrichment, indicating enhancer activation and cohesin association.
• Global 3D genome architecture (TADs, A/B compartments) remained largely unchanged upon SMARCB1 reconstitution; however, chromatin loop landscapes changed markedly (131 control-specific loops; 1,164 SMARCB1+-specific loops; ≥1.5-fold contact change).
• A prominent control-specific loop connects the MYC promoter to a distal +1.1 Mb super-enhancer (RhOME2); this loop was strongly reduced after SMARCB1 reconstitution. Both the MYC promoter and RhOME2 lost accessibility and BRD9/SS18 occupancy upon SMARCB1 restoration.
• MYC dependency: SMARCB1 reconstitution downregulated MYC mRNA/protein and cell-cycle gene expression with a proliferation arrest signature. shRNA knockdown of MYC significantly reduced PDO proliferation (CellTiter-Glo).
• Patient specificity of MYC enhancers: 4C-seq in P78 replicated MYC–RhOME2 (+1.1 Mb) looping loss after SMARCB1 reconstitution; P60 lacked this interaction but showed two distinct distal loops, also diminished upon SMARCB1 reconstitution. ATAC-seq indicated high accessibility at interacting regions that was lost after SMARCB1 restoration.
• The study defines a set of Rhabdoid Oncogenic MYC Enhancers (RhOMEs): RhOME1 (PCAT1 region, −0.9 Mb), RhOME2 (+1.1 Mb), and RhOME3 (CCDC26 region, +1.9 Mb). eRNAs were specifically expressed in PDOs where the corresponding RhOMEs were active and interacting with MYC.
• K-means clustering of P103 control-specific OCRs identified: K1 (non-enhancer CTCF sites), K2 (weak enhancers), and K3 (CTCF-independent super-enhancers). K3 showed highest ncBAF (BRD9) occupancy in control that was dramatically reduced upon SMARCB1 restoration. Hi-C showed K3 super-enhancers formed insulation boundaries in control that were abolished after SMARCB1 reconstitution.
• MRT tissues recapitulated intertumoral heterogeneity: Public H3K27ac ChIP-seq showed variable super-enhancer activity at RhOME1–3 across patients, with consistently active MYC promoters. scMultiome from 7 patients revealed tumor-specific combinations of accessible RhOMEs (e.g., P156/P168 at RhOME2; P041/P052 at RhOME1 and RhOME3; P166 at RhOME3), while normal cells lacked RhOME accessibility.
• ncBAF dependence: Pharmacologic BRD9 inhibition (I-BRD9) phenocopied SMARCB1 reconstitution—reduced MYC expression (RT-qPCR), decreased proliferation, downregulation of MYC targets and E2F/G2M Hallmarks, and upregulation of differentiation programs. CUT&RUN confirmed loss of BRD9/SS18 binding at MYC promoter and RhOME2/3 and genome-wide after I-BRD9.
• Mechanistic model: SMARCB1 loss shifts BAF complex balance toward ncBAF occupancy at distal super-enhancers (RhOMEs), recruiting cohesin and promoting long-range interactions with MYC promoter to drive oncogenic transcription. SMARCB1 restoration reduces ncBAF binding, enhancer activity, cohesin occupancy, and associated looping at MYC while leaving most CTCF-anchored architecture intact.

The findings demonstrate that SMARCB1 loss reprograms enhancer landscapes in MRT to activate MYC via long-range, patient-specific enhancer–promoter contacts mediated by ncBAF and cohesin. Despite limited global changes in 3D genome architecture, key local loops that connect MYC to distal super-enhancers (RhOME1–3) are established in SMARCB1-deficient states and collapse upon SMARCB1 reconstitution. This suggests that SMARCB1 is critical for preventing formation and maintenance of ectopic, super-enhancer-driven insulation boundaries and loops that sustain oncogene expression. The intertumoral heterogeneity in RhOME usage, recapitulated in both PDOs and primary tissues, implies that enhancer selection is context-dependent, potentially reflecting neural crest lineage states during development. Normal cells lack accessibility at RhOMEs, underscoring their tumor specificity. Functionally, MRT growth depends on MYC, and both genetic restoration of SMARCB1 and pharmacologic inhibition of ncBAF (BRD9) converge on suppressing MYC programs and cell-cycle progression, highlighting ncBAF as a therapeutic target. Cohesin association at RhOMEs supports a model where cohesin-mediated loop extrusion facilitates enhancer–promoter proximity independent of CTCF at many sites, while SMARCB1 restoration reduces ncBAF and cohesin binding at these enhancers to dismantle oncogenic loops.

This study identifies patient-specific distal super-enhancers (RhOME1–3) that drive MYC overexpression in SMARCB1-deficient malignant rhabdoid tumors through ncBAF- and cohesin-associated long-range interactions. SMARCB1 reconstitution reverses these enhancer activities, reduces ncBAF occupancy, and disrupts MYC looping and expression, suppressing proliferative programs. Single-cell profiling of MRT tissues confirms intertumoral heterogeneity in enhancer utilization, while normal cells lack RhOME accessibility, indicating tumor specificity. Therapeutically, BRD9 inhibition phenocopies SMARCB1 restoration, supporting ncBAF as a target. Future work should mechanistically dissect enhancer selection and the regulomes of individual MYC–RhOME loops, quantify intratumoral heterogeneity and clonal enhancer usage, and extend this blueprint to other BAF-mutant cancers to inform precision epigenetic therapies.

• Limited number of PDO models (three) and Hi-C performed in one PDO (P103) may constrain generalizability of 3D genome findings.
• Some chromatin assays have low replicate numbers (e.g., ChIP-seq n=1 for certain marks), limiting statistical power at individual loci.
• Functional validation of RhOME elements in tumors (e.g., CRISPR perturbation of enhancers) was not performed; causality is inferred from correlative chromatin changes and MYC dependence.
• scMultiome data show intertumoral heterogeneity, but the extent and dynamics of intratumoral enhancer heterogeneity remain to be quantified in larger cohorts.
• Not all tumors exhibited RhOME1–3 activity, indicating alternative regulatory elements may drive MYC in some cases and were not fully characterized here.