BET inhibitors in metastatic prostate cancer: therapeutic implications and rational drug combinations

Abstract

Introduction:

The bromodomain and extra-terminal (BET) family of proteins are epigenetic readers of acetylated histones regulating a vast network of protein expression across many different cancers. Therapeutic targeting of BET is an attractive area of clinical development for metastatic castration-resistant prostate cancer (mCRPC), particularly due to its putative effect on c-MYC expression and its interaction with the androgen receptor (AR).

Areas covered:

We speculate that a combination approach using inhibitors of BET proteins (BETi) with other targeted therapies may be required to improve the therapeutic index of BET inhibition in the management of prostate cancer. Preclinical data has identified several molecular targets that may enhance the effect of BET inhibition in the clinic. This review will summarize the known preclinical data implicating BET as an important therapeutic target in advanced prostate cancer, highlight the ongoing clinical trials targeting this protein family, and speculate on rationale combination strategies using BETi together with other agents in prostate cancer. A literature search using Pubmed was performed for this review.

Expert opinion:

Use of BETi in the treatment of mCRPC patients may require the addition of a second novel agent.

1. Introduction

Epigenetic regulation of gene expression in prostate cancer has been well studied over the past several years. The epigenome is controlled by ‘writer,’ ‘eraser,’ and ‘reader’ proteins, which results in a dynamic molecular interplay affecting gene transcription. Methylation and acetylation are common ‘marks’ found on histones bound with DNA, which can be ‘written’ (via methyltransferases or acetyltransferases) or ‘erased’ (via demethylases or deacetylases). Pharmacologic targeting of these enzymes has been a focus of clinical development across several malignancies, but their use in solid tumors is of unclear benefit to date. ‘Reader’ proteins, such as the bromodomain and extra-terminal (BET) family of proteins, translate the epigenetic markings via interaction with various transcription factors into a pattern of mRNA and hence protein expression. BET proteins have been implicated in regulating prostate cancer cell growth and are an attractive target in patients with metastatic castration-resistant prostate cancer (mCRPC). Inhibitors of BET proteins (BETi) are being studied in early-phase clinical trials examining the safety of these agents in the mCRPC patient population. However, concurrent development of combination strategies for BETi together with other targeted agents is also ongoing. Here, we review the preclinical data supporting the use of BETi in prostate cancer and identify rational therapeutic partners which may enhance the clinical benefit of this drug class.

2. BET inhibition in prostate cancer

Members of the BET family of proteins include BRD2, BRD3, BRD4, and BRDT. Within the N-terminal domain of each protein are two conserved bromodomain motifs, BD1 and BD2, which recognize and bind acetylated histones [1]. The extra-terminal (ET) domain, located in the C-terminus, likely has a role in protein–protein interactions. BET proteins can bind positive transcription elongation factor b (P-TEFb), which facilitates RNA polymerase II-dependent transcription of a number of downstream targets [2–4]. The ubiquitous expression of BET proteins in different tissue types makes it a clinically relevant target across many disease states (reviewed in [5]). With respect to oncology, inhibition of BET proteins, particularly BRD4, was shown to epigenetically downregulate the expression of c-MYC, which prompted further study in prostate cancer and other MYC-driven malignancies [6].

Current treatment strategies for mCRPC focus on targeting the androgen receptor (AR) signaling pathway. The development of new, non-AR-targeted therapies remains an unmet clinical need and may overcome conventional mechanisms of therapy resistance. The interdependent relationship between AR and c-MYC has been well documented in preclinical prostate cancer models. Studies have demonstrated that c-MYC is a downstream mediator of AR signaling, but can also drive cell growth in the absence of androgen suggesting a role for c-MYC in the development of castration resistance [7,8]. Molecular modeling suggests that AR signaling may decrease whereas c-MYC molecular signatures are more prominent with disease progression [9]. In addition, c-MYC may regulate the expression of AR splice variants (including AR-V7) leading to ligand-independent AR signaling [10].

Based on the presumed mechanism of c-MYC downregulation, BETi was tested in several in vitro models of prostate cancer. Wyce et al. demonstrated a growth-inhibitory effect of I-BET762 on the AR-positive human prostate cancer cell lines: LNCaP, VCaP, and 22RV1 [11]. Minimal effects were seen in AR-negative cell lines (i.e. PC3, DU145, NCI-H660) in this study. Following BETi treatment, apoptosis induction, validated by PARP cleavage, was observed in the VCaP model, whereas G1 cell-cycle arrest occurred in LNCaP cells, identifying at least two different and unique mechanisms of action of this class of drug. Interestingly, while AR protein levels were not decreased in any of the AR-positive cell lines, c-MYC expression was decreased across LNCaP and VCaP cells as well as the BETi-insensitive PC3 cell line, suggesting that reduced c-MYC expression may not equally affect cell growth/apoptosis in different in vitro models of prostate cancer. Moreover, restoration of c-MYC expression in LNCaP cells largely reverse BETi-induced growth inhibition, although it does not do so in VCaP cells. The work of Wyce et al. laid the groundwork for further investigation of BETi in prostate cancer. In addition, these findings supported c-MYC as a key target, but it also suggested diverse mechanisms of action.

Members of the BET family of proteins, BRD 2/3/4, may interact with AR and regulate AR-targeted gene expression [12]. More specifically, BRD4 interacts with the N-terminus of AR, which was abrogated in the presence of BETi. Asangani et al. performed a microarray analysis showing that BETi treatment resulted in decreased AR and MYC expression signatures in AR-positive cell lines. However, stable c-MYC expression did not rescue BETi-induced growth inhibition, suggesting that the main mechanism of action of BETi in prostate cancer is directed through AR signaling. Additional in vivo studies by Asangani et al. suggested that BETi could inhibit the growth of castration-resistant xenografts suggesting a potential role of BETi in the treatment of mCRPC.

More recently, multiple preclinical studies demonstrated that mutations in SPOP, an E3 ligase substrate binding protein commonly mutated in prostate cancer, resulted in resistance to BETi treatment in prostate cancer models [13,14]. In the work by Janouskova et al., missense SPOP mutations in endometrial cancer promoted sensitivity to BETi suggesting that the use of SPOP mutations as a potential biomarker of response to BETi may be tumor-type-specific [14]. The mechanism of SPOP-mediated BETi resistance may be related to changes in levels of BET proteins. However, SPOP is known to have several functions including a role in mediating PI3K and AR signaling which requires further exploration [15].

Multiple clinical trials studying BETi with open cohorts for patients with mCRPC are ongoing (Table 1). A Phase I trial of OTX015 in patients with lymphoma and multiple myeloma showed adverse events of thrombocytopenia and anemia in 96% and 91% of subjects, respectively, prompting concern over hematological toxicities of this class of drug [16]. Based on these reported adverse events, therapeutic levels of BETi as a single agent may not be achieved in solid tumor patients, including mCRPC patients. BETi treatment may require a combination strategy for optimal clinical activity. We propose several rational drug partners, which may enhance the clinical benefit of BETi in advanced prostate cancer (Figure 1, Table 2).

Table 1.

Ongoing BET inhibitor clinical trials enrolling patients with metastatic, castration-resistant prostate cancer.

Single agent BET inhibitor
A dose-finding study of OTX105/MK-8628, a small molecule inhibitor of the bromodomain and extra-terminal (BET) proteins, in adults with selected advanced solid tumors (MK-8628-003) – NCT02259114
Study of BMS-986158 in subjects with select advanced solid tumors – NCT02419417
A study evaluating the safety and pharmacokinetics of ABBV-075 in subjects with cancer – NCT02391480
Dose escalation study of GSK2820151 in subjects with advanced or recurrent solid tumors – NCT02630251
A two part study of RO6870810. Dose-escalation study in participants with advanced solid tumors and expansion study in participants with selected malignancies – NCT01987362
A study of ZEN003694 in patients with metastatic castration-resistant prostate cancer- NCT02705469
A study to investigate the safety, pharmacokinetics, pharmacodynamics, and clinical activity of GSK525762 in subjects with NUT midline carcinoma (NMC) and other cancers – NCT01587703
A Phase 1/2, open-label safety and tolerability study of INCB057643 in subjects with advanced malignancies – NCT02711137
Combination BET inhibitor
A study of ZEN003694 in combination with enzalutamide in patients with metastatic castration-resistant prostate cancer – NCT02711956
Safety, tolerability, pharmacokinetics, and pharmacodynamics of GS-5829 as a single agent and in combination with enzalutamide in participants with metastatic castrate-resistant prostate cancer – NCT02607228

Figure 1.

Novel drug combinations with BET inhibitors in prostate cancer. The BRD family of proteins (Center) interacts with P-TEFb and mediates RNA Pol II dependent transcription of targeted genes, including c-MYC. We propose several therapeutic partners for use in combination with BETi, including inhibitors of Wnt signaling (Bottom Left), Enzalutamide/AR-targeted therapy (Upper Right), PI3K inhibitors (Bottom Right), and PARP inhibition (Upper Left).

Abbreviations: AR – Androgen Receptor, BETi – Bromodomain and Extra- Terminal inhibitor, P-TEFb – positive transcription elongation factor b, RNA Pol II – RNA Polymerase II, V7 – Androgen Receptor Splice Variant −7, RTK – Receptor Tyrosine Kinase, PI3K – Phosphoinositide 3-kinase, PI3Ki – PI3K inhibitor, PARP – Poly ADP Ribose Polymerase, PARPi – PARP inhibitor, β-cat – beta-catenin.

Table 2.

Proposed BET inhibitor combinations in prostate cancer.

Drug combinations	References
BETi+ AR-targeted therapies	[35,36]
BETi +AKT/PI3K inhibitors	[45]
BETi+Wnt signaling inhibitors	[50,51]
BETi+ PARP inhibitors	[60]

2.1. Novel AR-targeted therapies

Metastatic prostate cancer is initially treated with androgen deprivation therapy (ADT), which inevitably leads to castration-resistance over time. Novel AR-targeted therapy (ATT), such as abiraterone acetate and enzalutamide, increases the survival of mCRPC patients irrespective of prior chemotherapy [23–26]. However, sequential use of ATT results in decreased clinical activity with the second-line agent compared to the first-line hormonal agent [27–29]. A common mechanism of resistance to ATT involves the expression of a key splice variant of AR, termed AR-V7. AR-V7 contains an intact N-terminal and DNA-binding domain, but lacks the expression of the ligand-binding domain, which results in constitutively active AR signaling in the absence of engagement with circulating androgen [30–32]. The detection of AR-V7 in circulating tumor cells confers resistance to ATT, but does not predict resistance to chemotherapy [17,18,33,34]. Non-chemotherapy-based treatments for AR-V7-positive patients are understudied and represent an unmet medical need [32].

AR-V7 retains expression of the N-terminal domain suggesting that BETi attenuation of BRD4-AR interaction may be a potential therapeutic option for AR-V7-positive prostate cancer. To this end, treatment with BETi blocked AR splice variants from binding chromatin and resulted in downregulation of their expression in one study [35]. Consistent with this finding, enzalutamide-resistant prostate cancer cell lines retained sensitivity to BETi [36]. In this study, Asangani et al. confirmed BETi could induce loss of AR-V7 potentially secondary to decreased expression of known splicing factors, SRSF1 and U2AF65. Thus, clinical use of BETi may overcome enzalutamide resistance when used in combination with the latter, but these data also provide a rationale for single-agent use of a BETi upon progression on novel ATT (Figure 1, Upper Right Panel). Clinical activity of BETi in AR-V7 positive prostate cancer has not been tested to date. Several clinical trials are now investigating a combination strategy of ATT + BETi in mCRPC patients (and many of these are examining CTC-based AR-V7 expression), but preliminary data is not yet available (Table 1).

2.2. PI3K pathway

Loss of PTEN occurs in approximately 50% of prostate cancers and has been linked to disease aggressiveness and prostate-cancer-specific death [37]. It is a dual-specific protein and lipid phosphatase best known for antagonizing PI3K signaling through the conversion of phosphatidylinositol-(3,4,5]-trisphosphate (PIP3) to phosphatidylinositol-(4,5)-bisphosphonate (PIP2) [38]. Upon PTEN loss or inactivation in prostate cancer, the resulting activation of the PI3K-AKT-mTOR signaling pathway has been shown to negatively regulate AR signaling [39]. Moreover, inhibition of AR has been shown to activate the PI3K pathway demonstrating a reciprocal feedback mechanism between these two pathways [39]. Clinical studies using combinations of ATT and PI3K inhibition are currently ongoing (summarized in [40]).

Concomitant gain/activation of the c-MYC locus and loss of PTEN correlates with lymph node metastasis and relapse after radiotherapy, suggesting an important relationship between the MYC and PI3K signaling in defining indolent versus lethal disease [41,42]. Prior studies have shown that c-MYC expression can be regulated downstream of PI3K/AKT signaling [19,43]. A murine model of forced MYC expression in the setting of PTEN deletion resulted in genomic instability and lethal metastatic prostate cancer, which recapitulated key histologic and phenotypic features of human disease [44]. In a murine breast cancer model involving a PIK3CA activating mutation and PTEN deletion, cells were resistant to treatment with single-agent PI3K or BET inhibition [45]. Exposure to a pan-PI3K inhibitor (PI3Ki) resulted in a transient decrease in Akt phosphorylation, which was restored by the activation of other receptor tyrosine kinase pathways as a mechanism of resistance. Using a combination approach of PI3Ki plus BETi, impairment of AKT phosphorylation was maintained, resulting in sustained inhibition of MYC expression and enhanced cell kill. Stratikopoulos et al. also demonstrated a similar effect in the PTEN-negative prostate cancer cell lines, LNCaP and PC3. Given the relationship between the PTEN/PI3K/AKT pathway with AR and MYC signaling, a novel therapeutic approach in metastatic prostate cancer may be to combine BETi with either PI3Ki or AKT inhibitors concurrently with ADT, particularly in a PTEN-deleted prostate cancer population (Figure 1, Bottom Right Panel). A dose-escalation Phase I trial would be necessary to determine a relevant biologic dose of both inhibitors and to assess adverse toxicities.

2.3. Wnt signaling

Wnt signaling has been well studied as a potential target for prostate cancer therapeutics. Much of the earlier molecular work on Wnt investigated the canonical signaling pathway with a focus on the role of beta-catenin as an oncogenic factor. AR and betacatenin have been shown to interact in vitro facilitating AR transcriptional activity [46–48]. Mutations in beta-catenin were found in a small subset of prostate cancer specimens and may contribute to enhanced protein stability resulting in tumor-specific protein overexpression [20]. More recently, data have emerged suggesting that the non-canonical Wnt signaling pathway may also play a role in prostate cancer. In a pilot study of 13 mCRPC patients, RNA-seq was performed on single prostate circulating tumor cells [21]. Following progression on an AR-targeted agent, activation of the non-canonical Wnt signaling pathway was observed, specifically via upregulation of Wnt5a protein expression [21]. Wnt5a has been implicated in multiple cancers and may signal through both the canonical and non-canonical Wnt pathways [49]. Given the vast signaling network involving downstream mediators of Wnt, effective targeting of this pathway remains an unmet challenge and a potentially interesting therapeutic approach in appropriate patients.

In a leukemia model, BETi decreased c-MYC expression resulting in early clinical responses, but c-MYC protein levels were restored upon BETi resistance [50,51]. Both studies identified the canonical Wnt signaling pathway as a mediator of resistance to BETi. Activation of Wnt signaling led to reestablishment of BRD4 target gene expression, which abrogated the effect of BETi (Figure 1, Bottom Left Panel). It is not known if concurrent therapy with an inhibitor of Wnt signaling and BETi would delay resistance or improve therapeutic benefit. Another study in mesenchymal stem cells showed that BET-induced downregulation of Wnt2 and FZD2, both implicated in canonical Wnt signaling, resulted in cell cycle inhibition [52]. While these data provide compelling evidence for combination therapy, finding the appropriate inhibitor of Wnt signaling in human cancer is less clear. Further molecular analysis from mCRPC patients treated on a clinical trial with a BETi will be needed to confirm these in vitro observations and to clarify the putative role of Wnt in mediating BETi response or resistance.

2.4. Poly ADP ribose polymerase (PARP)

Pharmacological inhibitors of PARP (PARPi) have been shown to induce synthetic lethality in cells harboring bi-allelic defects in genes involved in homologous recombination (HR) (i.e. BRCA1, BRCA2, ATM, and others) (reviewed in [53,54]). Somatic mutations in DNA-repair genes have been identified in both primary [55] and metastatic castration-resistant [56] prostrate cancer, with prevalence rates of approximately 8–10% and 20–25%, respectively. With respect to germline mutations, a recent study found that the incidence of inherited DNA-repair gene alterations in metastatic prostate cancer was significantly higher (about 12%) than in both men with localized prostate cancer (5%) and in the general population at large germline (2%) [57]. The majority of prostate cancer patients with germ line DNA repair mutations have somatic loss-of-heterozygosity of the corresponding allele, leading to bi-allelic inactivation and loss-of-function of the corresponding protein [22,58].

These findings served as the rationale to test PARPi (olaparib) in heavily pretreated mCRPC (TOPARP-A trial) [59]. Although the presence of a DNA-repair defect in the tumor was not an eligibility requirement, the response rate to olaparib was 33% in the unselected population (16 of 49; 95% CI 20–48%), including a prostate-specific antigen (PSA) decline of over 50% (PSA₅₀) germline in 10 out of 49 patients. The trial did identify somatic and germline mutations in DNA-repair genes, and the presence of inactivating DNA-repair defects correlated with improved response rates (88% vs. 6%) as well as improved progressionfree (9.8 vs. 2.7 months) and overall (13.8 vs. 7.5 months) survival compared to DNA-repair proficient cancers.

In the TOPARP-A trial, objective responses were observed in a small subset of DNA-repair proficient tumors potentially suggesting a role in epigenetic regulation of DNA repair genes. Using HR-proficient ovarian and breast cancer models, concurrent use of BETi and PARPi resulted in downregulation of BRCA1 promoting cell death [60]. These data provide proof-of-principle for a combination strategy of BETi with PARPi in an unselected patient population in an effort to increase the objective response rates induced by PARPi alone (Figure 1, Upper Left Panel). Interestingly, BRD4 may directly regulate the cellular response to genotoxic stressors (i.e. ionizing radiation) via an effect on chromatin structure. In one study, use of a BETi resulted in open chromatin and facilitated DNA repair, which protected against radiation-induced apoptosis in several cancer cell lines suggesting a protective effect of BETi against DNA damage-induced cell death [61]. However, in different models, BETi restored radiosensitivity suggesting that the role of BRD4 in directly regulating DNA repair may be cell-type specific [62]. The use of BETi as a radiosensitizer remains unclear. Nonetheless, further exploration of BETi and its effect on epigenetic regulation of HR DNA repair gene expression in prostate cancer is warranted.

3. Conclusion

Multiple strategies incorporating BET inhibition are currently in clinical development for the treatment of metastatic prostate cancer. This review has highlighted four potential avenues of combination therapy using BET inhibition together with other approaches: novel AR-targeting therapy, PI3K/AKT pathway inhibition, Wnt pathway inhibition, as well as PARP inhibition. While BETi monotherapy trials are awaiting safety and efficacy data, consideration should be given now to rational combination strategies in an effort to maximize clinical benefit for patients with mCRPC. Clearly, optimal clinical activity is likely to be observed in only a subset of patients and not the entire mCRPC population. Therefore, early Phase 2 combination trials should incorporate biomarker-selection strategies to maximize the chance of observing an efficacy signal. Such biomarker selection will depend on the particular combination strategy being employed (e.g. Wnt pathway activation when combining BETi plus Wnt inhibitors), and would ideally be achieved using a liquid biopsy-based approach. The future of BET-targeting therapeutic strategies in advanced prostate cancer is upon us.

4. Expert opinion

The clinical development of non-AR-targeted therapies for metastatic prostate cancer is an unmet clinical need. Patients with metastatic castration-resistant disease are commonly cycled through oral AR-targeted therapies followed by chemotherapy. In years past, several small molecule inhibitors of various pathways have been tested and failed to improve clinical outcomes. Recently, PARP inhibitors have been shown to improve survival in patients harboring a mutation in a HR DNA repair gene, but this represents a minority of the prostate cancer population. The recent publication of preclinical studies showing significant benefit of BETi in multiple prostate cancer models spurred the opening of prostate cancer cohorts in existing BETi clinical trials. Prostate-cancer-specific clinical trials quickly followed suit. The preclinical data supporting the use of BETi in prostate cancer patients suggest these agents may target both AR-mediated transcription as well as other potential drivers of cancer growth such as c-MYC. However, the exact mechanism of action in the clinical treatment of prostate cancer remains to be determined. It will be important to see if BETi achieves most of its clinical benefit through AR in which case these agents may be affected by similar mechanism of resistance as other AR-targeted therapies. Given the importance of c-MYC in prostate cancer, any clinical effect of BETi mediated through c-MYC downregulation would be revolutionary, as c-MYC has been untargetable in the past. Since multiple clinic studies studying BETi as a single agent are ongoing in prostate cancer, we are hopeful that pre- and on-treatment biopsies will be able to answer this question.

Use of epigenetic therapy as a single agent in solid tumor malignancies has been unsuccessful to date. Based on this history, we expect that the addition of other small molecule inhibitors may be necessary to enhance the effect of BETi. This review identifies and describes potential combination strategies for BETi in prostate cancer. Next-generation, AR-targeted therapies (i.e. enzalutamide) are already being tested in the clinic in combination with BETi, suggesting multi-agent approaches are being considered. A key question that will be answered in these early studies involving ATT is whether the generation of AR-V7 can be inhibited or delayed by the addition of BETi. Aside from chemotherapy, no effective treatment for AR-V7-positive prostate cancer is currently available making BETi an attractive agent for such cancers. Such an approach may extend the clinical utility of abiraterone or enzalutamide, but BETi would still be considered under the umbrella of other AR-targeted agents.

We believe the greatest benefit of BETi may be found when partnering with agents beyond those that affect AR, such as those described in this review (i.e. PI3K inhibitors). PTEN is lost in nearly 50% of prostate cancers allowing for clinical benefit in a large population of men with prostate cancer. Similar to epigenetic drugs, use of single-agent PI3K inhibitors is not a main focus of clinical development programs. However, preclinical data suggest one potential resistance mechanism to BETi is via signaling through the PI3K-AKT pathway. Concurrent inhibition using both BETi and PI3k inhibitors may result in a synergistic effect on cell death in PTEN deleted or PI3K mutated cancers. Moreover, lower doses of each agent may be required to induce such an effect, which can potentially reduce off-target effects. Similar sentiments can be expressed about the use of PARP inhibitors or targets of Wnt signaling with BETi. These combination approaches will need additional preclinical study, particularly using in vivo models so that dose-limiting toxicities may be identified before clinical testing.

BETi are being developed by multiple pharmaceutical companies for use in many malignancies. In the prostate cancer field, we need to keep an open mind about the notion that existing small molecule inhibitors, once thought to be obsolete with respect to clinical development, may still have value. If these agents can be tested with BETi and shown to be safe, there is little risk and potentially a large reward for using such combinations in patients with advanced metastatic, castrationresistant prostate cancer.

Article highlights.

• Bromodomain and Extra-Terminal (BET) family of proteins are key readers of acetylated histones that regulate gene expression.

• Clinical development of BET protein inhibitors is ongoing for use in multiple malignancies, including prostrate cancer.

• BRD4 associates with the androgen receptor, potentially serving as an important target of BETi in prostate cancer.

• Novel combination strategies with BETi may be required to maximize clinical benefit in the treatment of metastatic prostate cancer.

This box summarizes key points contained in the article