Atypical teratoid rhabdoid tumor (ATRT): disease mechanisms and potential drug targets

1. Introduction

Atypical teratoid rhabdoid tumor (ATRT) is a highly malignant central nervous system (CNS) neoplasm of early childhood [1]. Approximately 70% of ATRTs are diagnosed in the first year of life, and over 90% of cases occur in children younger than three years of age. Although ATRT accounts for only 1–2% of all pediatric CNS tumors, it comprises 40–50% of all CNS malignancies in children younger than one year of age. ATRT is slightly more common in males, with a male-to-female ratio of approximately 1.5, and is most often diagnosed in Caucasians. It may arise anywhere in the CNS and should be considered as a diagnosis when evaluating any aggressive-appearing tumor of the brain or spinal cord in a very young pediatric patient. Metastatic dissemination at diagnosis is common and may be found in 20–40% of cases [2–6].

ATRT is characterized by a primitive histologic appearance with multi-lineage differentiation, including varying components of glial, mesenchymal, and epithelial differentiation [1]. Histologically, ATRT is an embryonal tumor that displays loss of nuclear SMARCB1 or SMARCA4 expression or, for unresolved lesions, a DNA methylation profile that aligns with ATRT [1,7]. Seminal molecular studies over the past 30 years have revolutionized our understanding of the molecular basis of ATRT, facilitating a transition from a descriptive, morphology-based assessment toward a molecular (epi) genetic understanding that has revealed potential targeted treatments for this lethal childhood cancer [8–10]. Here, we focus our discussions on disease mechanisms of ATRT and how its unique molecular changes may be exploited using novel therapeutics, ultimately offering hope for patients with ATRT and their families.

2. Basic biology

Pioneering research of rhabdoid tumor predisposition syndrome (RTPS) in the late 1990s and early 2000s identified the loss of part or all of the long arm of chromosome 22 as a recurrent molecular event in rhabdoid tumors including ATRT [11–13]. Subsequent molecular analyses demonstrated biallelic SMARCB1 loss-of-function alterations as the key genetic event in ATRT pathogenesis [14–16]. In its role as a core subunit of the canonical SWItch/sucrose non-fermentable (SWI/SNF) chromatin-remodeling complex (also known as the canonical BAF [cBAF] complex), SMARCB1 is required for cell lineage determination and normal cell differentiation [17–19]. Functional deficiency of SMARCA4, a shared BAF ATPase, in the few SMARCB1 wild-type cases further underscores the importance of BAF in ATRT tumorigenesis [8,20]. Despite their highly malignant and heterogeneous appearance, the genomes of ATRTs are remarkably stable and contain limited exomic alterations [15,21,22].

BAF complexes destabilize histone-DNA interactions in an ATP-dependent manner, opening neighboring stretches of DNA and allowing transcription factors and chromatin-remodeling enzymes to either activate or repress target genes [23]. Loss of BAF has profound effects on cell viability, proliferation, and differentiation depending on cell context and the exact BAF perturbation. Pertinent to ATRT, SMARCB1 loss results in histone subunit 3 lysine 27 acetylation (H3K27Ac) loss at promoter-distal enhancers necessary for cell-lineage specification and differentiation [18,24]. Conversely, promoter-proximal, cell lineage-specific super-enhancers associated with non-canonical BAF (ncBAF) are retained in ATRT and are essential for cell survival [19,25,26]. Deranged polycomb-related signaling as a result of SMARCB1 deficiency has been identified as a key tumorigenic driver in ATRT by multiple studies [21,27,28]. Abundant research further implicates SMARCB1 in the regulation of numerous other pathways of cancer and development, including RhoA-dependent regulation of cell migration [29,30], cell cycle regulation via Aurora Kinase A [31], P16INK4a [32], TP53 [33,34], and RB [35,36], and downstream activation of other oncogenic pathways such as SHH [37], WNT/β-catenin [38,39], and c-MYC [40–42] (Figure 1).

Figure 1.

Basic BAF complex composition and epigenetic function. The canonical SWItch/sucrose non-fermentable (SWI/SNF) chromatin remodeling complex (canonical BAF [cBAF] complex) is a multi-subunit modulator of gene transcription implicated in vital cellular functions such as cell differentiation, lineage determination, and self-renewal. In the absence of a functional SMARCB1 subunit, less BAF complex is recruited to typical enhancers, which is accompanied by loss of histone acetylase activity and deranged polycomb-signaling at these sites of transcriptional regulation. Super enhancer activity is selectively retained at genes that contribute to immortality and cell proliferation in the absence a functional SMARCB1 subunit (created with BioRender.com).

In contrast to the chromosomal stability of ATRT, recent comprehensive molecular studies have demonstrated that SMARCB1-deficient ATRTs are comprised of at least three molecularly distinct, clinically relevant subtypes with shared and subtype-specific epigenomic features, while SMARCA4-deficient ATRT constitutes a fourth subgroup that does not harbor SMARCB1 alterations [20,21,43,44]. While multiple naming paradigms have been proposed, recent consensus defines the three SMARCB1-deficient subgroups as ATRT-SHH, ATRT-TYR, and ATRT-MYC [3]. ATRT-SHH tumors tend to occur in toddlers predominantly in the supratentorial compartment. They contain focal SMARCB1 alterations, have a hypermethylated phenotype, and are defined by a primitive neuronal gene expression pattern and NOTCH pathway activation [21,43,44]. ATRT-TYR tumors usually occur in the posterior fossa of infants and tend to have broad deletions affecting chr22q11.2, are hypermethylated, and display a mesenchymal gene expression profile characterized by deranged bone morphogenic protein (BMP) and melanogenesis signaling pathways, including micropthalmia-associated transcription factor (MITF) and tyrosinase (TYR) [21,43,44]. ATRT-MYC tumors are more common in older children and may occur in either the supra- or infratentorial compartment or the spinal cord. These tumors usually contain focal mutations or deletions in SMARCB1, but they display a hypomethylated phenotype, and are marked by dysregulated MYC signaling and derepression of HOX family genes [21,43,44]. SMARCA4-mutant ATRT occurs in young infants and is associated with a higher frequency of germline mutations and a distinct DNA methylation signature in comparison to SMARCB1 mutated tumors [20,45].

Clinically, prognosis appears to differ by molecular subgroup, but this is heavily influenced by other clinical factors that are not evenly distributed among molecular subgroups, namely patient age and the presence of metastatic disease at diagnosis. Two separate studies have demonstrated TYR-subtype tumors have an improved overall prognosis, either when analyzed independently or compared to all non-TYR-subtype ATRT. Frühwald et al. identified age less than one year as an independent risk factor for overall survival in addition to non-TYR subgrouping [46]. Upadhyaya et al. showed that ATRT-SHH is most commonly associated with metastatic disease at diagnosis, leading to inferior outcomes compared to ATRT-TYR and ATRT-MYC. In this study, given that patients with ATRT-TYR are often the youngest, infants without metastatic disease were analyzed as a separate subgroup, and in this analysis, outcomes did not differ by molecular subtype [47]. These observations demonstrate the complex interplay between clinicopathologic factors and molecular subtype in ATRT and highlight the need for further investigation into the unique role, if any, of molecular subgrouping in prognosis.

3. State of the art treatment

The mainstay of treatment in ATRT is an aggressive multi-modal approach consisting of maximal safe surgical resection with subsequent adjuvant chemotherapy and radiotherapy. Advances in each of these modalities have improved four-year overall survival to just under 50%, on average, leaving substantial room for improvement in future therapeutic avenues [5,48–50]. Most referral centers in the USA would use the Children’s Oncology Group (COG) ACNS0333 protocol for newly diagnosed ATRT, which utilizes two courses of multiagent chemotherapy followed by three courses of high-dose chemotherapy (HDCT) followed by peripheral blood stem cell rescue and craniospinal irradiation (CSI) after maximal safe surgical resection [6]. ACNS0333 resulted in a significant survival benefit over a historical cohort, but both event-free and overall survival remained below 50% at four years, and survivors suffered typical chemo- and radiotherapy-associated adverse effects. Similar outcomes have been achieved in some smaller studies using conventional and intrathecal chemotherapy in combination with radiotherapy, calling into question the need for HDCT in selected patients [48]. Given the associations of older age and ATRT-TYR subgrouping with improved prognosis, the risk-benefit ratio of HDCT in these patients must be carefully weighed, given the potential for similar therapeutic benefit using regimens that spare ablative chemotherapy. While radiotherapy remains a mainstay of contemporary treatment for ATRT and has generally been associated with improved tumor control [4,46], the known association with poor neurodevelopmental outcomes in very young patients demands a careful consideration of treatment risk in this modality compared to HDCT [5]. To assess the non-inferiority of HDCT compared to conventional chemotherapy and focal radiotherapy, the European Society for Pediatric Oncology recently initiated a randomized phase-III clinical trial for children with ATRT (EudraCT 2018–003335-29), which also includes the comparison of neurocognitive outcomes among treatment arms as a secondary study objective.

Many patients with ATRT who survive the early stage of the disease will remain refractory to conventional chemo- and radiotherapy, and for these patients, novel targeted treatments are on the horizon that may offer promise as a therapeutic option. Current work is aimed at defining both highly efficacious as well as molecular subtype-specific therapies that target underlying molecular drivers of tumorigenesis (Figure 2). Histone deacetylase inhibitors (HDACis) have demonstrated efficacy across molecular subgroups of ATRT. HDACis such as suberoylanilide hydroxamic acid (SAHA, also known as vorinostat) and LAQ824 were shown to be effective against multiple ATRT cell lines in vitro [44,51], and vorinostat demonstrated a radiosensitizing effect in a rhabdoid tumor xenograft model [52]. Recruitment for two early phase clinical trials investigating vorinostat in a variety of cancers including ATRT has been completed with myelosuppression being the dose-limiting toxicity (NCT00217412, NCT01076530 [53]). As described above, SMARCB1 loss leads to cell cycle progression, and both the Aurora Kinase A inhibitor alisertib and the CDK4/6 inhibitor ribociclib have been and are currently being used in clinical studies of children with ATRT (NCT02114229, NCT03387020, NCT03434262) [54,55].

Figure 2.

Summary of potential therapeutic strategies. Groundbreaking findings in the fields of BAF and enhancer biology have revealed a central role for epigenetic dysregulation in the pathogenesis of ATRT. A variety of different therapies targeting secondary epigenetic vulnerabilities are currently being investigated in preclinical studies and clinical trials. Among others, these pathophysiologically relevant and potentially subgroup-specific therapeutic approaches include inhibition of AURK, CDK4/6, MYC, and polycomb, which may be therapeutically targeted by alisertib, ribociclib, dasatinib, tazemetostat, respectively. An alternative strategy is to target residual BAF, and a few select compounds (e.g. FHD-609 and FHD-286) are likewise in clinical trial for cancers that contain a mutation in a BAF complex subunit (created with BioRender.com).

To date, only a few, limited-scale drug screens have been performed for ATRT. Torchia et al. screened a variety of small-molecule inhibitors targeting multiple molecular subtype-specific pathways on cell lines that transcriptionally correspond to ATRT-SHH and ATRT-MYC subtypes [44]. In this study, ATRT-SHH cell lines were more critically dependent on epigenetic regulators involved in chromatin remodeling with EZH2 and bromodomain inhibition being selectively toxic to ATRT-SHH cell lines. Tumors of ATRT-TYR and ATRT-MYC subgroups (formerly ‘Group 2’ subtypes) were more critically dependent on PDGFR and other receptor tyrosine kinase pathways. The multi-tyrosine kinase inhibitors dasatinib and nilotinib were selectively toxic to group-2 cells in vitro, and in an orthotopic xenograft model of a group-2 cell line, dasatinib prolonged survival [44]. Further cementing the importance of RTK signaling in ATRT-MYC, a CRISPR screen of extracranial rhabdoid tumors, which are most similar to ATRT-MYC by DNA methylation, similarly demonstrated RTK dependence [56]. A phase I/II study of dasatinib in combination with cytotoxic chemotherapy for CNS tumors is ongoing (NCT00788125). Taken together, results from laboratory research and early clinical studies underscore the therapeutic potential of ATRT-targeted therapies and suggest selective vulnerabilities in each molecular subtype.

4. Expert opinion

The benefit of aggressive multi-modal cytoreductive and cytotoxic therapy for ATRT has cemented the role of surgical resection, craniospinal radiotherapy, and ablative chemotherapy in the upfront treatment of ATRT. Nevertheless, the numerous developmental adverse effects of this treatment and high relapse rate underscore the importance of identifying highly efficacious, targeted therapies for this disease.

ATRT is fundamentally a disease of selective enhancer loss, so it stands to reason that targeting secondary epigenetic vulnerabilities is a pathophysiologically relevant and potentially tumor-specific therapeutic approach. However, the road to targeted epigenetic therapy in ATRT is not straightforward and is limited not only by our understanding of rhabdoid tumor biology but also by available therapeutic options.

Polycomb inhibition using the clinically available EZH2 inhibitor tazemetostat is one approach that has gained traction in the pediatric oncology community. EZH2 inhibition is attractive not only due to its targetability but its biological relevance: in developmental systems, BAF has been shown to antagonize polycomb-mediated gene repression [57], and EZH2 inhibition using RNA interference or pharmacologic inhibition has been shown to variably inhibit ATRT cell growth in vitro and in vivo [28,58]. In a phase I clinical trial or relapsed or refractory SMARCB1-deficient tumors, tazemetostat treatment resulted in a clinically measurable response in 19% of patients with ATRT [59]. Final results of clinical trials of tazemetostat in ATRT and other tumors (NCT02601937, NCT03213665) are necessary in determining the long-term efficacy of this treatment.

An alternative approach to targeting epigenetic vulnerabilities in ATRT is to inhibit residual, ncBAF-dependent enhancer activity that is thought to contribute to tumorigenesis. The oncogenic role of the ncBAF-specific subunit BRD9 in a range of malignancies and its demonstrated role in maintaining potentially proto-oncogenic, promoter-proximal enhancers in ATRT makes this protein an appealing therapeutic target [25,26]. A small molecule inhibitor of BRD9 (FHD-609) is now in clinical trial for SS18-SSX fusion-containing synovial sarcoma (NCT NCT04965753), in which SS18-SSX fusions preclude cBAF assembly. An inhibitor of the BAF ATPase subunits SMARCA2 and SMARCA4 (FHD-286) is likewise in clinical trial for uveal melanoma (NCT04879017) and multiple hematologic malignancies (NCT04891757). The results of these trials will be imperative in determining whether targeting BAF is a viable therapeutic strategy in the wide range of cancers that contain a mutation in a BAF complex subunit.

In sum, a wealth of research over the past decade has revealed a central role for epigenetic dysregulation in the pathogenesis of ATRT. Clinically, ablative chemotherapy, craniospinal radiotherapy, and aggressive surgical resection are the mainstay of upfront treatment of ATRT, which has substantially improved outcomes compared to historical cohorts. Moving forward, the next generation of therapeutic approaches should implement recent groundbreaking findings in the fields of BAF and enhancer biology into clinically efficacious treatments for early refractory and recurrent disease.