SWI/SNF complexes govern ontology-specific transcription factor function in MYC-subtype atypical teratoid rhabdoid tumor

Abstract

Background: Atypical teratoid rhabdoid tumor (ATRT) is a deadly central nervous system embryonal tumor caused by loss of SMARCB1, a core subunit of SWItch/Sucrose Non-Fermentable (SWI/SNF) chromatin remodeling complexes. SMARCB1-deficient cancers are defined by loss of cell differentiation-associated enhancers, but how SWI/SNF interacts with other arbiters of cell differentiation (specifically lineage-specific transcription factors [TFs]) remains poorly understood.

Methods: We leveraged a multi-omics approach, patient-derived ATRT cells, and patient-derived orthotopic xenografts to investigate the interplay of SWI/SNF with lineage-specific TFs in a clinically relevant setting.

Results: We observe that an activating protein 1 (AP-1)-dependent transcriptional regulatory network is lost in ATRT, and AP-1 and lineage-specific TFs TEAD1 and ZIC2 require SMARCB1 for enhancer binding. SMARCB1-dependent SWI/SNF integrates transcriptional functions of lineage-specific TFs into a core regulatory circuit that depends on the AP-1 subunit c-JUN, whose expression is determined by a SMARCB1-dependent super-enhancer that is lost in ATRT-MYC. In the absence of SMARCB1, lineage-specific TFs are sequestered to promoters, where they maintain core transcriptional programs necessary for cell survival. Targeting residual, promoter-proximal TF activity by a protein degrader of the SWI/SNF ATPase SMARCA4 or small-molecule inhibitors that indirectly inhibit AP-1 and TEAD activity abrogates expression of these networks, reducing cell viability in vitro and prolonging survival in an orthotopic patient-derived xenograft model.

Conclusions: These results demonstrate SWI/SNF complexes are critical for lineage-specific TF binding and activity at both promoters and enhancers. In the context of ATRT, these findings reveal a previously underappreciated therapeutic vulnerability in targeting residual promoter-proximal TF function in ATRT.

Keywords: ATRT; chromatin; epigenetics; rhabdoid tumor; transcription factors.

Graphical Abstract

Figure 1.

Enhancer and transcriptional landscapes in SMARCB1-absent and -restored ATRT cells. (A) SMARCB1 restoration remodels enhancers. Volcano plots demonstrate differentially bound H3K27ac loci in SMARCB1-restored vs. -absent ATRT cells (FDR < 0.05, left). Accompanying histograms detail the genomic distribution of gained H3K27ac peaks relative to the nearest transcription start site (TSS) (right) in BT12 and MAF737a cells. (B) Enhancer dynamics. Relative overlap in active enhancers called by Rank Ordering of Super-Enhancers (ROSE) highlights altered enhancer loci in SMARCB1-absent and -restored cells. (C) Transcription factor motif enrichment. De novo motif enrichment analyses using HOMER of constituent H3K27ac broad peaks contributing to gained enhancers in SMARCB1-restored cells (broad peak FDR < 10⁻⁵), identifying potential transcriptional regulators mediating enhancer activation. (D) Transcriptional consequences of SMARCB1 restoration. Volcano plots illustrating differentially expressed genes (DEGs) between SMARCB1-absent and restored BT12 and MAF737a cells (FDR < 0.05), highlighting transcriptional reprogramming. (E) Functional enrichment of SMARCB1-dependent transcriptional circuitry. Gene ontology (GO) enrichment of predicted AP-1 associated transcriptional regulatory circuitry in BT12 cells, linked to de novo enhancers using the single-nearest gene method. The most highly enriched terms and key ontological families are highlighted.

Figure 2.

SMARCB1 restoration redirects lineage-specific TFs to activated enhancers. (A) SMARCB1-dependent redistribution of FRA2. FRA2 ChIP-Seq peak overlap analysis in BT12 in SMARCB1-absent and -restored BT12 cells. De novo motif enrichment analysis in the union peak set highlights additional TF families that may co-bind with FRA2. (B) TF engagement at SMARCB1-dependent enhancers. Read pileups of SMARCB1-dependent active enhancers associated with FRA2, TEAD1, and ZIC2, demonstrating their recruitment to enhancer loci only after SMARCB1 restoration. (C) TF occupancy at gained enhancers. Relative overlap of TF peaks at gained active enhancers following SMARCB1 restoration, illustrating significant but limited coordination in TF recruitment. (D) Transcriptional impact of TF-associated enhancers. Gene set enrichment analysis (GSEA) of TF-associated enhancer target gene expression assessed by RNA-Seq. Target genes were assigned to enhancers using a curated library of regulatory elements (GeneHancer). (E) Convergence of TF-regulated transcriptional programs. Venn diagram illustrating overlap in TF-associated enhancer target genes upregulated following SMARCB1 restoration, highlighting cooperation in regulatory circuits. (F) Local TF cooperation at developmentally relevant enhancers. Representative genomic loci demonstrating TF co-binding at developmentally pertinent enhancers.

Figure 3.

Transcription factor-mediated interactions define a conserved c-JUN-oriented core regulatory circuitry. (A) Network cooperativity of TF-associated enhancers. Protein-protein interaction (PPI) network analysis using upregulated target genes of enhancers associated with FRA2, TEAD1, and ZIC2-associated enhancers as input. The right panel highlights the union of networks, with the most highly integrated nodes further emphasized. (B) c-JUN core regulatory circuitry. Visualization of single-nearest neighbors of c-JUN, identifying the key upregulated partners of c-JUN that make up the c-JUN core regulatory circuitry (CRC). (C) c-JUN expression following SMARCB1 restoration. JUN mRNA and c-JUN protein expression assessed by RNA-Seq and Western blot, respectively. Significance was assessed using the unpaired 2-tailed Student’s t-test; ***P < .001. (D) Super-enhancer regulation of c-JUN. Ranked enhancer analysis (ROSE) of SMARCB1-restored cells highlighting the JUN super-enhancer. The y-axis in panel F represents the average difference in area (in arbitrary units) within the H3K27ac ChIP-Seq peaks and inputs at each locus. The JUN genomic neighborhood is shown for illustration of the associated super-enhancer. (E) JUN knockdown in SMARCB1-restored cells. Western blot of ATRT cells following SMARCB1 restoration and transduction with either nontargeting (NT) or JUN shRNA, assessing the impact of JUN depletion. (F) c-JUN-dependent cell proliferation. Incucyte proliferation analysis of cells analyzed in E, quantifying the phenotypic impact of JUN knockdown. (G) Transcriptional impact of JUN loss. Gene set enrichment analysis of SMARCB1-restored cells following JUN knockdown, with gene set identifiers shown. BT12 cells were used for all analyses in this figure.

Figure 4.

Loss of AP-1-associated enhancers and the c-JUN core regulatory circuitry are specific to ATRT-MYC. (A) c-JUN expression across cell lines of multiple ATRT subtypes. Western blot of c-JUN expression in ATRT-MYC and ATRT-SHH cell lines following SMARCB1 restoration, illustrating subtype-specific differences in c-JUN regulation. (B) Differential JUN expression across ATRT subtypes. Principal component analysis (PCA) of RNA-Seq data from 25 untreated ATRT specimens demonstrating 3 molecular subgroups, with expression of key subgroup markers and JUN expression shown. (C) Subtype-specific enrichment of c-JUN-associated transcriptional programs. Gene set enrichment analysis of subgroup-relevant genes, c-JUN CRC component genes, and the AP-1-associated TRN in ATRT patient specimens. (D) Patient-specific SMARCB1 mutants. Schematic for experimental design utilizing patient-specific SMARCB1 mutants (right). Expression of HA-tagged SMARCB1 variants was confirmed by Western blot. (E) JUN transcriptional response to SMARCB1 variants. JUN mRNA expression by RNA-Seq in SMARCB1 variant subgroups, highlighting variant-specific impacts on JUN expression. (F) Transcriptional impact of SMARCB1 variant expression. Gene set enrichment for the c-JUN CRC and AP-1 TRN across SMARCB1 variant-expressing cells, illustrating their impacts on network expression. (G) Proliferative effects of SMARCB1 variants. Incucyte proliferation assay of BT12 cells transduced with selected SMARCB1 variants, demonstrating their impact on proliferative dynamics.

Figure 5.

Lineage-specific TFs are sequestered to ncBAF-dependent promoters in the absence of SMARCB1. (A) Genomic distribution of TF binding sites in SMARCB1-absent cells. Relative distribution of TF peaks relative to TSSs (top) and proportion of TF peaks overlapping with H3K4me3 and SMARCA4 peaks (bottom) in SMARCB1-absent cells, highlighting promoter-associated TF occupancy. (B) TF binding at SMARCA4-associated promoters. ChIP-Seq profiles of H3K4me3+ and SMARCA4+ loci co-bound by each TF in SMARCB1-absent cells, demonstrating promoter binding. (C) ncBAF-associated super-enhancers in SMARCB1-absent cells. Ranked enhancers (ROSE) of SMARCB1-absent. Shaded call-outs highlight ncBAF-associated SEs overlapping with TF-bound active promoters (ie, ±2 kb to nearest annotated TSS, H3K4me3+, SMARCA4+). (D) Functional consequences of ncBAF-TF interactions. Overlap of expressed genes associated with ncBAF- and TF-bound peaks identified in C, followed by GO enrichment analysis to identify biological pathways regulated by these interactions. (E) Regulatory network of ncBAF-associated TF-bound genes. Schematic of regulatory circuitry showing protein-coding genes identified in (D) with the most highly integrated nodes shown. (F) Disrupting ncBAF alters TF-dependent transcription. BRD9 expression by Western blot (left) and volcano plot of DEGs by RNA-Seq (FDR < 0.05, right) of ATRT cells treated with either DMSO or dBRD9, illustrating the transcriptional impact of ncBAF depletion. (G) Gene set enrichment analysis following ncBAF inhibition. GSEA of dBRD9-treated cells, showing depletion of TF targets defined in (D) following dBRD9 treatment. BT12 data are shown.

Figure 6.

Targeting of residual AP-1 and TEAD activity in ATRT. (A) Transcriptional response to pathway inhibition. Gene set enrichment analysis of known AP-1 and YAP/TAZ/TEAD target genes in BT12 cells treated with trametinib (a MEK inhibitor that reduces downstream AP-1 activation) and verteporfin (a disruptor of YAP/TAZ-TEAD association), demonstrating pathway-specific transcriptional repression. (B) Cellular proliferation upon pathway inhibition. Incucyte proliferation assays of ATRT cells treated with trametinib or verteporfin at various doses, showing dose-dependent proliferation suppression. (C) Cellular responses to treatment. IC₅₀ values of cells treated with 25 nM trametinib or 500 nM verteporfin for 48 hours, as determined by Cell Titer Blue assay. (D) Survival impact of trametinib treatment in ATRT xenografts. Kaplan-Meier analysis of patient-derived orthotopic xenograft ATRT models treated with trametinib. The Mantel-Cox log-rank test was used for statistical comparisons.