Down, but Not Out: A Role for SMARCB1 in Synovial Sarcoma

The BRG1-associated factor (BAF) family of chromatin-remodeling complexes is among the most frequently mutated chromatin regulators in human cancer (1). Loss-of-function mutations in the core canonical BAF (cBAF) and Polybromo-BAF (PBAF) complex subunit SMARCB1 are the most frequently observed genomic alterations in malignant rhabdoid tumors (MRT) and epithelioid sarcoma (EpS), suggesting a driving role for SMARCB1 loss in tumorigenesis. In addition to SMARCB1 genomic alterations in certain rare cancers, reduced SMARCB1 protein expression is frequently observed in synovial sarcoma, in which nearly all tumors show a t(X;18)(p11;q11) chromosomal translocation that gives rise to an oncogenic SS18–SSX fusion protein. SS18–SSX is incorporated into cBAF complexes where SS18–SSX has been posited to evict SMARCB1, leading to its proteasomal degradation (2). Although the reduction in SMARCB1 expression in synovial sarcoma has long been observed, the respective contributions of the SS18–SSX fusion and SMARCB1 loss to synovial sarcomagenesis are not known.

In this issue of Cancer Discovery, Li and colleagues (3) examine whether homozygous Smarcb1 genetic deletion alone gives rise to synovial sarcoma–like tumors via a localized injection of TATCre into the paws of Smarcb1^fl/fl mice or cross to the Myf5Cre line, traditionally used in mouse models of synovial sarcoma (4). However, in both mouse models of Smarcb1 deletion, tumors strongly resembling EpS rather than synovial sarcoma were observed as determined by the presence of EpS-like histologic features and transcriptional profiles. Thus, SS18–SSX fusion–mediated reduction in SMARCB1 protein expression is not sufficient to drive synovial sarcomagenesis. These results support prior literature showing that the SS18–SSX fusion results in gained BAF complex activity via retargeting of SS18–SSX-containing BAF complexes to activate oncogenic gene transcription (2, 5–7). The authors therefore investigated the role of SMARCB1 in synovial sarcomagenesis in mouse models with TATCre- or Myf5Cre-induced expression of the oncogenic SS18–SSX2 fusion. Here, the authors reveal a dose-dependent effect of Smarcb1 deletion in accelerating tumorigenesis compared with SS18–SSX2 expression alone (Fig. 1, left). However, tumors generated by concurrent deletion of Smarcb1 and SS18–SSX2 expression show histologic features and transcriptional profiles that are distinct from those of synovial sarcoma and more closely resemble MRT. Interestingly, tumors driven by combination SS18–SSX2 expression and Smarcb1 deletion showed decreased expression of genes comprising a known synovial sarcoma transcriptional signature relative to tumors driven by SS18–SSX alone, including synovial sarcoma–specific pathways such as neuronal activity and WNT–β-catenin signaling. Moreover, known transcriptional targets of SS18–SSX (6) were found to be downregulated in the combination tumors compared with tumors driven by the fusion alone. These results suggest that SMARCB1 contributes to the transcriptional and phenotypic characteristics of synovial sarcoma, a finding that challenges prior models that SS18–SSX oncogenic activity is SMARCB1-independent (5, 7).

The prevailing model of SMARCB1 reduction in synovial sarcoma holds that the SS18–SSX fusion mediates eviction of SMARCB1 from BAF complexes, resulting in SMARCB1 proteasomal degradation. This is based on observations that SS18–SSX expression results in a concomitant reduction of SMARCB1 over time and that SMARCB1 protein levels can be rescued by knockdown of the SS18–SSX fusion, overexpression of wild-type SS18, or via proteasomal inhibition (2). In the current work, the authors tested this model by coimmunoprecipitating the tagged SS18–SSX fusion and endogenous SMARCB1 in multiple cell lines, revealing a reciprocal interaction between SMARCB1 and SS18–SSX. To further validate the interaction of SS18–SSX and SMARCB1, the authors examined the levels of SMARCB1 in affinity-purified recombinant BAF complexes containing wild-type SS18 or SS18–SSX and found that SMARCB1 was incorporated at similar levels in wild-type SS18 and SS18–SSX-containing BAF complexes. These biochemical studies provide compelling evidence that SMARCB1 incorporates into SS18–SSX-containing BAF complexes, and argues that it is unlikely that SS18–SSX causes eviction of SMARCB1 or a general defect in cBAF complex assembly.

If SS18–SSX does not evict SMARCB1 from purified BAF complexes, then how are SMARCB1 levels reduced in synovial sarcoma? The authors investigate via transfection of wild-type SS18 or SS18–SSX followed by semiquantitative Western blot analysis of glycerol gradients tracking all three BAF complex variants: cBAF, PBAF, and the recently discovered GBAF complex (8, 9). cBAF and GBAF, but not PBAF, incorporate SS18–SSX, while cBAF and PBAF incorporate SMARCB1, which is naturally absent from GBAF. The authors observed a significant reduction in the cBAF-specific subunits ARID1A, ARID1B, and DPF2 along with SMARCB1 in cells transfected with SS18–SSX relative to cells transfected with wild-type SS18, with no significant change in PBAF and GBAF subunits. Conversely, in human synovial sarcoma cell lines, shRNA knockdown of the SS18–SSX fusion or proteasomal inhibition resulted in elevated cBAF subunit expression and SMARCB1 expression. Together, these findings suggest that SS18–SSX expression mediates proteasomal degradation of cBAF complexes, resulting in reduced SMARCB1 protein expression (Fig. 1, right), agreeing with several aspects of the prior model of SMARCB1 reduction by SS18–SSX (2). An important validation of the wholesale cBAF degradation proposed by this work will be determining whether cBAF subunits such as ARID1A, ARID1B, and DPF2 are reduced at the protein level in human synovial sarcoma tumors, which can be addressed in future studies via IHC for cBAF subunits in synovial sarcoma tumor samples.

A consequence of cBAF degradation is the relative increase in the fractional abundance of GBAF and PBAF complexes in human synovial sarcoma lines, which the authors showed could be rescued by knockdown of the SS18–SSX fusion (Fig. 1, right). These results have important implications for understanding the mechanism of oncogenic gene transcription in synovial sarcoma. Previous work demonstrated that SS18–SSX-containing GBAF complexes are targeted to chromatin, but reports differ on whether GBAF complexes are the main driver of SS18–SSX-dependent oncogenic transcription (7), or an SS18–SSX-containing SMARCB1-less cBAF complex (5). Neither of these models excludes a potential synergy between SS18–SSX and SMARCB1 loss in tumorigenesis as shown by the authors. Li and colleagues demonstrate that much of the residual SMARCB1 coimmunoprecipitates with PBAF complexes in addition to cBAF complexes in human synovial sarcoma cell lines and mouse SS18–SSX2-induced tumors. Furthermore, knockdown of SMARCB1 in human synovial sarcoma lines or deletion of Smarcb1 in mouse tumors results in destabilization of PBAF complexes. Assessment of the core component SMARCA4 in mouse tumors driven by SS18–SSX2 alone or the combination revealed that Smarcb1 deletion resulted in a reduction of the total number of SMARCA4 binding sites genome-wide, including reduced SMARCA4 binding at synovial sarcoma–specific binding sites. PBRM1 binding, specific to the PBAF complex, was found to be prominent at promoters in SS18–SSX2 tumors and similarly reduced by Smarcb1 deletion in combination tumors. It is thus possible that the impairment of PBAF complexes or residual SMARCB1-containing cBAF complexes in combination tumors results in additional tumor-promoting activities. Future functional genomic experiments are necessary to understand the relative contribution of cBAF and PBAF to SS18–SSX- and Smarcb1-dependent gene expression in synovial sarcoma. A previous report also observed PBAF complex localization to the promoters of SS18–SSX target genes (5), arguing for a potential cooperative role for PBAF complexes in SS18–SSX-mediated oncogenic transcription. The authors, on the other hand, speculate that PBAF complexes may act as a tumor suppressor by promoting the expression of synovial sarcoma–specific differentiation genes, giving rise to features of both mesenchymal and epithelial differentiation that characterize synovial sarcoma and restrain the more aggressive tumorigenesis observed in combination tumors.

Overall, the work of Li and colleagues provides compelling evidence that SMARCB1 plays a functional role in synovial sarcomagenesis, novel insights on the reduction of cBAF subunits in synovial sarcoma, and demonstrates the relative redistribution of BAF complexes toward greater fractional abundance of PBAF and GBAF complexes. Several questions remain. First, given that SS18–SSX fusion incorporates into both cBAF and GBAF, why is cBAF targeted for degradation but GBAF is not? The answer likely lies within distinct assembly pathways or structural differences between SS18–SSX-containing cBAF and GBAF complexes. An additional explanation could be the specific sensitivity of cBAF subunits to ubiquitylation, such as ARID1A, a known target of E3 ligases (10). Quantitative mass spectrometry experiments will be invaluable for addressing these questions and assessing the absolute abundance of individual proteins and variants in synovial sarcoma. Second, this work further adds to the existing controversy surrounding the mechanism of SS18–SSX-mediated oncogenic transcription, with various data supporting the contribution of SS18–SSX-containing GBAF complexes (7, 9), as well as cBAF (5) and potentially PBAF complexes (3, 5, 6). A detailed analysis of the abundance, localization, and activity of each variant will aid in the identification of BAF subunit genetic dependencies in synovial sarcoma and potential therapeutic strategies. The relative increase in the percentage of BRD9-containing GBAF complexes in SS18–SSX fusion–expressing cells is consistent with the genetic dependency on BRD9 observed in synovial sarcoma and the sensitivity of human synovial sarcoma lines to BRD9 degraders (7, 9). Similarly, the relative increase in BRD7-containing PBAF complexes may suggest a synthetic lethal role for BRD7, consistent with an shRNA screen identifying BRD7 knockdown as a specific vulnerability in human synovial sarcoma cancer cell lines (6). The emergence of BRD7/9 degraders and SMARCA4/A2 inhibitors and degraders represent alternative methods for inhibiting BAF complex function via targeting GBAF/PBAF complexes or all functional forms, respectively. As BAF-targeted therapeutics advance toward clinical trials, the findings in this study may provide relevant insights into understanding the relative utility of disrupting one or more BAF variants in synovial sarcoma with dual inhibitors or combination strategies.

No disclosures were reported.