The Cell Cycle: a Key to Unlock EZH2-targeted Therapy Resistance

MRT and ES are aggressive tumors characterized by poor prognosis and limited response to conventional therapies, including chemotherapy and radiation, emphasizing the need for new therapeutic strategies. Cancer-sequencing studies identified frequent alterations in epigenetic regulators, while preclinical studies verified their crucial role in these cancers. These discoveries lead to clinical trials testing small molecules that attempt to exploit therapeutic vulnerabilities driven by epigenetic dependencies.

One of the most frequently mutated genes in MRTs and ES is SMARCB1 (1). SMARCB1 is one of the core components of the chromatin remodeling SWI/SNF (BAF) complex, which regulates chromatin accessibility through remodeling of nucleosome–DNA interactions. The epigenome is a highly interconnected system with both cooperative and antagonistic mechanisms that fine tune regulation of gene expression. For example, the function of the SWI/SNF complex is antagonized by the Polycomb Repressive Complex 2 (PRC2), which mediates methylation of lysine 27 on histone 3 (H3K27), a repressive chromatin mark. Normally, the balance between chromatin remodeling and histone-modifying complexes ensures precise gene expression across various cell types at the appropriate time during development. Disruption of this equilibrium leads to dysregulation of gene expression and may promote malignant transformation or bypass tumor suppressive mechanisms. Specifically, loss of SWI/SNF function, such as through SMARCB1 mutations, creates a dependency on PRC2's enzymatic activities for cell survival (2, 3). This dependency suggests that inhibiting PRC2 could selectively kill cancer cells with SWI/SNF mutations.

The development of small-molecule inhibitors of EZH2, such as tazemetostat, marked a significant opportunity to exploit the synthetic lethality between SWI/SNF mutations and PRC2 dependency therapeutically. Preclinical studies showed that EZH2 inhibitors could attenuate SMARCB1-deficient tumor growth, culminating to FDA approval for treating MRTs and ES (4). Despite this promising data, clinical outcomes were muted, with an objective response hovering around 15%, and most treated sarcomas developing resistance to tazemetostat (5). This discrepancy between preclinical data and clinical reality underscores the complexity of patient response to epigenetic therapies and highlights the need for deeper understanding and strategies to overcome drug resistance.

To uncover the genetic basis of resistance to tazemetostat within the context of SMARCB1 deficiency, Kazansky and colleagues (6) conducted targeted sequencing of 500 genes in samples from MRT and ES patients both pre- and posttreatment. This comparative analysis revealed distinct patterns of somatic mutations between responders and nonresponders. Notably, one patient who developed resistance after initially benefitting from tazemetostat, had acquired a heterozygous missense mutation in EZH2 at tyrosine 666 (Y666N). A previous genetic screen in lymphoma cells identified the importance of this residue in conferring resistance to EZH2 inhibition (7); however, this is the first noted spontaneous occurrence of this mutation in patients after treatment. The authors found that the Y666N mutation does not impair the enzymatic activity of EZH2 in rhabdoid tumor cells but appears to hinder tazemetostat from binding to the enzymatic SET domain. Engineered expression of Y666N into rhabdoid tumor cell lines rendered them resistant to tazemetostat, underscoring its significance. Interestingly, cell lines harboring this mutation remained susceptible to inhibition by other components of the PRC2 complex, such as EED, allowing alternative therapeutic avenues in the future.

A second patient that acquired resistance exhibited biallelic loss-of-function mutations in RB1. RB1 functions as a negative regulator of the cell cycle in a complex with E2A by repressing expression of genes required for the transition from G₁ to S-phase. Phosphorylation of RB1 by CDK4 and CDK6 dissociates it from E2A, allowing cell-cycle progression. CDK4 and CDK6 themselves are blocked by CDKN2A and CDKN1A, two other known tumor suppressors, while RB1 and EZH2 regulate each other's expression and may also interact at the protein level (8), highlighting the multiple layers of regulation of this pivotal cell-cycle regulator and entry point. Kazansky and colleagues discovered that cells lacking RB1, circumvented the growth-inhibitory effects of tazemetostat, not through EZH2's direct action but likely by bypassing the drug-induced cell-cycle arrest at the G₁ phase. These results were evidenced by an observed shift in cell-cycle distribution, as RB1-null cells favor progression into the S and G₂–M phases of the cell cycle and showed increased expression of S and G₂–M genes like CDK2 and AURKB despite tazemetostat treatment.

Resistance to tazemetostat was also observed after experimental targeted deletion of CDKN2A and CDKN1A, which function upstream of RB1 (Fig. 1), pointing to the critical role of the RB1–E2F axis, and not just RB1 itself, in the efficacy of EZH2 inhibition. Mutations in CDKN2A, CDKN1A, and other RB1 regulating factors (e.g., ANKRD11) were prevalent among tazemetostat nonresponders, further corroborating that the RB1–E2F axis is a key determinant in the cellular response to EZH2 inhibition. Interestingly, despite effective inhibition of EZH2 activity after RB1 inhibition in rhabdoid tumor cells, the authors also observed morphologic changes and sustained transcriptional responses (e.g., Epithelial-Mesenchymal Transition gene sets), suggesting that resistance mechanisms may decouple drug-induced differentiation from cell-cycle control. Overall, these results suggest that effective epigenetic cancer therapy may require a comprehensive strategy that accounts for cell-cycle regulation and differentiation mechanisms.

To identify functional markers of response and resistance to tazemetostat, the authors performed gene expression analysis in tazemetostat responding and nonresponding patient tumors, before and after treatment. Pretreated tazemetostat resistant tumors exhibited enrichment in S and G₂–M cell-cycle signatures, consistent with the important role of the RB1–E2F axis in patients that acquired resistance to tazemetostat. In addition, they identified a small set of consistently up- and downregulated genes associated with RB1 loss, including PRICKLE1, which has previously been implicated as a biomarker of poor prognosis in other cancers (9), and which may be used as a marker of response in these tumors.

To overcome tazemetostat resistance, the authors targeted cell-cycle kinases CDK2 and AURKB, which function downstream of the G₁–S checkpoint, and which are upregulated in tazemetostat-resistant tumors (Fig. 1). The authors found that in vitro, nearly all sarcoma cell lines they tested were susceptible to the AURKB inhibitor, barasertib, regardless of resistance to tazemetostat, while the combination of tazemetostat with barasertib caused a greater cell-cycle arrest than either drug alone. Notably, the combination of tazemetostat with barasertib effectively mitigated resistance in vivo and markedly reduced tumor growth. These results provide new opportunities to improve the efficacy of epigenetic therapies through rational drug combinations, offering a promising direction for MRT and ES patients who have developed resistance to current treatments.

The effort by the authors to study rare cancers like MRTs and ES are commendable, and the combination of tazemetostat with AURKB inhibitors in overcoming resistance in vivo are encouraging and invite cautious optimism. The rarity of the tumors and the nature of the studies, however, impose certain limitations, such as the relatively small sample size and targeted sequencing, which constrains the ability to uncover the full spectrum of resistance or differentiation mechanisms. Furthermore, while the authors identified higher levels of certain immune-related gene sets in posttreatment patients that responded to tazemetostat, the xenograft modeling that was used to test drug combinations was devoid of a functional immune system. This is particularly important here given the role of EZH2 in regulating antitumor immune mechanisms in many solid tumors (10). In most studies, EZH2 inhibition tends to result in a stronger antitumor immune response, such as through increased expression of antigen processing and presentation factors. The effects of EZH2 inhibition on antitumor immunity, however, can be nuanced and cell type dependent. Future studies, therefore, need to provide better understanding of the role of EZH2 on the tumor immune response against MRT and ES tumors to better tailor possible future combinations with immunotherapy.

Multiple preclinical studies support a model of synthetic lethality between SWI/SNF and PRC2 in multiple tissues and cancer types (4). Despite strong preclinical studies, most patients on the tazemetostat trial did not respond to treatment, underscoring a broader problem of translating preclinical findings into clinical success. It is possible that the positive results reported in preclinical studies are not entirely representative of the relationship between SWI/SNF and PRC2 as negative results tend not to be reported. The reality is that these epigenetic regulators are complex and have many functions: from context-dependent protein complex compositions to noncanonical functions, and numerous downstream effectors. All of these can affect different aspects of cancer biology, such as cancer progression, angiogenesis, metastasis, antitumor immunity, and response to therapy. Exploiting the synthetically lethal relationship between the SWI/SNF and PRC2 complexes in cancer, therefore, is not trivial and requires deep mechanistic understanding. While this study opens new avenues for targeted therapies and gives hope to MRT and ES patients with SMARCB1 mutations, there is still a lot of work to be done to fully understand the relationships and mechanisms of these epigenetic regulators so that we can more effectively target them or exploit the dependencies they may create.