Inhibitors of the transactivation domain of androgen receptor as a therapy for prostate cancer

Abstract

The androgen receptor (AR) is a modular transcription factor which functions as a master regulator of gene expression. AR protein is composed of three functional domains; the ligand-binding domain (LBD); DNA-binding domain (DBD); and the intrinsically disordered N-terminal transactivation domain (TAD). AR is transactivated upon binding to the male sex hormone testosterone and other androgens. While the AR may tolerate loss of its LBD, the TAD contains activation function-1 (AF-1) that is essential for all AR transcriptional activity. AR is frequently over-expressed in most prostate cancer. Currently, androgen deprivation therapy (ADT) in the form of surgical or chemical castration remains the standard of care for patients with high risk localized disease, advanced and metastatic disease, and those patients that experience biochemical relapse following definitive primary treatment. Patients with recurrent disease that receive ADT will ultimately progress to lethal metastatic castration-resistant prostate cancer. In addition to ADT not providing a cure, it is associated with numerous adverse effects including cardiovascular disease, osteoporosis and sexual dysfunction. Recently there has been a renewed interest in investigating the possibility of using antiandrogens which competitively bind the AR-LBD without ADT for patients with hormone sensitive, non-metastatic prostate cancer. Here we describe a class of compounds termed AR transactivation domain inhibitors (ARTADI) and their mechanism of action. These compounds bind to the AR-TAD to inhibit AR transcriptional activity in the absence and presence of androgens. Thus these inhibitors may have utility in preventing prostate cancer growth in the non-castrate setting.

1. Introduction

Full-length androgen receptor (AR) is a modular, ligand-activated transcription factor belonging to the super family of nuclear steroid receptors. These receptors serve as master regulators that function as signaling hubs to tightly control gene expression and mediate a multitude of cellular functions in response to their cognate ligand¹. Androgens are the main ligands for AR transactivation² and are responsible for regulating a wide variety of diverse biological functions including cardiovascular, musculoskeletal and hematopoietic systems³. However androgens are best known for their role in sexual differentiation and development of male reproductive organs including the prostate. Blocking transactivation of AR has been a therapeutic approach for over five decades for the clinical management of pathologies driven by this receptor such as prostate cancer.

2. Structure of androgen receptor

The AR is comprised of three main functional domains (Figure 1): the C-terminal ligand-binding domain (LBD); the DNA-binding domain (DBD); and the N-terminal transactivation domain (TAD). A flexible hinge region connects the DBD to the LBD. The LBD is a folded protein with its crystal structure resolved to reveal 11 α-helices and 4 β-sheets and contains activation function-2 (AF-2) which provides an interface for protein-protein interactions⁴. Unlike related steroid hormone receptors, the AR AF-2 contributes little to no transcriptional activity, and the AR relies almost entirely upon the N-terminal TAD⁵. In the absence of androgen (ligand), the AR is primarily localized in the cytoplasm and bound to heat-shock chaperone proteins. Upon androgen binding, the AR is released from chaperone proteins and undergoes a conformational change with dimerization mediated by interactions between its N-terminal TAD and C-terminal LBD (N/C interaction)⁴. AR dimer translocates to the nucleus where it recognizes cognate DNA regions termed androgen-response elements (AREs) within enhancer and promoter regions of target genes through two zinc fingers contained within its DBD^4,6.

Figure 1 – AR functional domains.

Schematic representation of the 919 amino acid residues androgen receptor (AR) including the ligand-binding domain (LBD), the DNA-binding domain (DBD) and the N-terminal transactivation domain (TAD). Important regions in the TAD are highlighted including Activation Function-1 (AF-1) and transcriptional activation units (Tau) 1 and 5 with their corresponding residue numbers. FxxLF and WxxLF motifs are also indicated. Notably, the TAD features highlighted cysteines and tryptophans, pivitol for the interaction with ralaniten and masofaniten, respectively. Created with BioRender.

Transcriptional activity of the AR is dependent upon its N-terminal TAD which mediates interactions with basal transcriptional machinery and coregulators. Unlike the LBD or DBD, the TAD is characterized by large regions of intrinsic disorder^4,7,8. Protein-protein interactions are facilitated by the existence of molecular recognition features (MoRF) contained within the TAD. These are short binding regions contained within intrinsically disordered regions that promote disorder-to-order transitions upon interactions with binding partners^8,9. ²³FxxLF²⁷ and ⁴³³WHTLF⁴³⁷ are two such MoRFs which are found in the TAD that play important roles in AR transcriptional activity by mediating interactions with coregulatory binding partners as well as interdomain N/C interaction. Nearly 300 binding partners have been identified for the AR^10,11, and the ability of the AR to coordinate a highly tailored response to a variety of cellular cues is in large part due to the intrinsic disorder of its N-terminal TAD. The plasticity of an unfolded structure permits the TAD to adopt numerous unique conformations enabling rapid and transient protein-protein interactions with many structurally diverse binding partners; such as CBP/p300, TFIIF^5,10 and STAT3¹² as well as other transcription factors and corepressors^10,13.

Unsurprisingly while the AR can tolerate the loss of its LBD, its TAD is critical for a functional receptor^5,14. TAD transcriptional activity can be further mapped to the AF-1 region, and more specifically to two transactivation units: tau-1 (residues 101-370) and tau-5 (residues 360-485). While tau-1 is thought to mediate androgen-dependent transcriptional activity, tau-5 allows transcriptional activity in the absence of androgen^5,14. Variants of AR that lack the LBD such as AR-V7 are clinically relevant in prostate cancer. This splice variant is constitutively active and its expression and activity are considered to contribute to resistance that evolves in response to therapeutic approaches that target AR-LBD^14-16.

3. The AR as a master regulator of gene expression

The full-length AR is directly responsible for the positive and negative regulation of the expression of hundreds of different genes in response to androgen^17-20. The expression of androgen-regulated genes plays roles in maintaining male sexual tissues and especially the prostate ²¹. Many of these androgen-regulated genes are involved in promoting growth and survival, and the AR signaling pathway is often exploited both in the initiation and in the progression of prostate cancer. It is estimated that AR-driven disease accounts for approximately 80-90% of prostate cancers at initial diagnosis²², and often remains a key oncogenic driver throughout disease progression. This is evidenced by the fact that the classical androgen-induced gene KLK3 that encodes prostate-specific antigen (PSA) has long been used as a surrogate biomarker in the clinic to track disease progression and clinical response to AR-targeted therapies²³.

While the AR primarily acts as an oncogene in prostate cancer, androgen-repressed genes are also thought to play major roles in proliferation and invasion (ie, CCND1, hTERT and MET)^24-26. However the mechanism of androgen-repression of gene expression is poorly understood and may involve multiple different mechanisms involving protein-protein interactions with different AR domains. For example, the mechanism of androgen repression of CCND1 gene expression involves androgen-bound AR recruited to a negative ARE and an SP1-binding site with recruitment of a repression complex that includes DAX-1 and HDAC-1²⁶. Nonetheless, it is well established that AR is a driver of prostate cancer from the seminal work of Dr. Charles Huggins who received the Nobel prize for his work demonstrating that castration results in the clinical improvement of prostate cancer symptoms and reduced tumor burden²⁷.

4. Androgen Deprivation Therapy and AR-LBD inhibitors

Androgen deprivation therapy (ADT) remains a crucial component of current treatment regimens for patients with high-risk localized or metastatic prostate cancer, and for those patients who experience biochemical relapse following failure of primary treatment^28-30. ADT was traditionally achieved using surgical castration or treatment with exogenous estrogen in an effort to eliminate testicular androgens. Both fell out of favor in the 1980s due to adverse psychological and physiological side effects and the advent of gonadotropin-releasing hormone (GnRH) agonists (eg. leuprolide, goserelin, triptorelin, histrelin)³⁰. GnRH agonists work by exploiting a negative feedback mechanism of the hypothalamic-pituitary-gonadal axis. While these drugs initially induce the production and release of lutenizing hormone and testosterone, castrate levels of testosterone are eventually achieved as expression of cognate GnRH receptors are lowered in response^3,30-32. The initial testosterone surge could lead to enhanced tumor growth if not managed by co-treatment with AR-LBD inhibitors, or use of an lutenizing hormone releasing-hormone (LHRH) antagonist (degarelix, relugolix) instead of an agonist^33-35. Unfortunately, ADT is associated with a number of co-morbidities which must also be considered when offering treatment^3,36-38. The most common adverse effects associated with ADT are increased risk of cardiovascular disease³⁹ and associated mortality; hot flashes; decrease in bone mineral density leading to increased risk of osteoporosis and fractures; sexual dysfunction caused by erectile dysfunction and reduced libido; decreased muscle mass and increased fat deposition; and decreased cognition and mood alterations with increased risk of depression^{37, 38} (Figure 2). Some of these side effects such as a loss of bone mineral density, are due to the fact that following castration estrogen levels which would normally be synthesized via testosterone metabolism are also lowered. Estrogens are highly important in the regulation of male bone metabolism, and estradiol levels are inversely correlated with fracture risk⁴⁰. This is evidenced by the fact that chemical castration using estrogens is not associated with treatment-related osteoporosis⁴¹.

Figure 2 – ARTADIs effectively inhibit AR in the non-castrate setting.

Androgens are produced via the hypothalamic-pituitary-gonadal axis and adrenal glands. In the presence of androgen (left), treatment with androgen receptor transactivation domain inhibitors (ARTADIs) block AR activity potentially without the adverse effects associated with androgen deprivation therapy (ADT). The castrate setting (right) can be achieved through ADT in the form of LHRH agonists or antagonists. In the absence of gonadal androgens, numerous adverse effects exist. Despite ADT and antiandrogen treatment, AR activity persists due to AR-V7 splice variants and AR with ligand-binding domain (LBD) mutations. Created with BioRender.

Antiandrogens that bind to the AR-LBD to block transactivation of the AR provide substantial survival benefits when used in combination with ADT. Several commonly used antiandrogens include enzalutamide, daralutamide, and apalutamide. Steroid synthesis inhibitors such as abiraterone acetate also improve outcomes for advanced prostate cancer patients when used in combination with ADT by the reduction of levels of non-testicular androgens. While ADT provides a substantial survival benefit for patients in managing symptoms and delaying disease progression, it is not curative^28,36. Patients will inevitably progress despite castrate-levels of androgens and develop castration-resistant prostate cancer (CRPC), generally within 2-3 years after initiation of ADT. Most CRPC still relies on transcriptionally active AR that may include the following mechanisms: increased expression of AR; gain-of-function mutations in the AR-LBD; and expression of constitutively active truncated splice variants of AR that lack LBD such as AR-V7. Expression of AR-V7 is predictive of a lack of clinical responses to ADT, abiraterone-acetate and antiandrogens^15,16. ADT, abiraterone-acetate and antiandrogens are not effective in tumours that express AR-V7 and in fact induce the expression of AR splice variants¹⁶.

To address these resistance mechanisms, we developed inhibitors to the AR N-terminal TAD (ARTADIs). The first ARTADI reported was in 2008 with the sintokamide family that binds directly to AR-TAD potentially in the tau-1 region^42,43. Report of the discovery of EPI-001 (a racemic mixture of ralaniten/EPI-002 and stereoisomers) quickly followed in 2010 and its prodrug ralaniten-acetate (EPI-506) entered clinical trials in 2015 (NCT02606123). While ultimately unsuccessful due to pharmacokinetic liabilities, the next-generation ARTADI masofaniten is currently being tested in several clinical trials (NCT05075577 and CT04421222). Additional compounds targeting the AR-TADs have recently been reported such SC428 and SC912^44,45. These inhibitors block the transcriptional activities of full-length AR as well as AR-splice variants such as AR-V7.

5. AR transactivation domain inhibitors (ARTADIs) interact with key tryptophan and cysteine residues within the AR-TAD

EPI-001 was proven to bind to AR-TAD first in 2013 using Click-chemistry approaches⁴⁶. Three years later, NMR studies confirmed that EPI-001 binds to amino acids 341-446 within AR-TAD⁴⁷ . The binding site for EPI-001 contains a MoRF at ⁴³³WHTLF⁴³⁷ within tau-5 which is potentially important in mediating protein-protein interactions through disorder-to-order transitions^7,8. The second-generation drug masofaniten (EPI-7386) which is currently in numerous clinical trials for prostate cancer has been reported to interact with W397 and W433 within tau-5⁴⁸. Therefore, these tryptophan residues may play a key role in the binding of some ARTADIs such as ralaniten and masofaniten to tau-5. Of note is that a gain-of-function mutation, W433L was discovered in clinical samples from prostate cancer patients failing antiandrogen therapy. This mutation stabilizes AR N/C interactions, which is a process necessary for AR transcriptional activity, in a gene-specific and cell-specific manner⁴⁹. Since these tryptophans reside in the binding site of ralaniten and masofaniten with potential importance in the binding mechanism, mutations to these residues may impact their efficacies.

In addition to tryptophan, cysteines in AR-TAD have also been suggested to be important in the covalent binding mechanism of ralaniten. In a closed system, ralaniten and its analogues can bind covalently to AF-1 region as proven using click chemistry, pull down assays, and a radiolabelled compound in cellular assays using prostate cancer cells^46,50. Nucleophilic amino acids such as cysteine, lysine, serine, threonine and tyrosine can form covalent bonds. Notably, the thiol group in cysteine is reactive because it is electron-rich and able to become polar⁵¹. Mass spectrometric analyses demonstrated that EPI-001 bound covalently to cysteine residues in the AF-1 region of AR and the data suggested this irreversible binding occurred at C404⁵². In recent years, a covalent binder called EN4 was reported to bind to a cysteine residue in the intrinsically disordered region of the Myc protein and this binding inhibits the transcriptional activity of Myc⁵³. Similarly, UT-143 was reported to bind covalently to C327 and C406 (equivalent to C404 in ref 48) in AR and its splice variants to inhibit their activity in prostate cancer cells⁵⁴. Therefore, cysteines in AR-TAD are potential candidates to be targeted in drug development.

The AR contains 27 cysteine residues, 13 of which are in AR’s intrinsically disordered TAD. In addition to the role of cysteines in covalent binding of small molecules, these residues also regulate protein structure, functions, and signaling with other proteins. These functions involve redox catalysis, nucleophilic interaction, allosteric regulation, and metal binding⁵⁵. Intermolecular disulfide bonds between cysteine residues within the AR protein itself are crucial for maintaining its tertiary structure and regulating conformational changes necessary for ligand-binding and transactivation⁵⁶. Additionally, intermolecular interactions between cysteine residues in the AR and cysteine residues in other proteins or cofactors, contribute to complex formation and functional modulation of AR activity⁵⁷. These interactions may occur through redox-sensitive mechanisms. Cysteine residues are highly susceptible to redox reactions through the thiol (-SH) group present in their side chain, making cysteine a key player in maintaining the redox homeostasis and signal regulation within cells⁵⁸. A recent study showed that mutation of cysteine to alanine at C267, C327 or C406 residues in the AR-V7 reduced its transcriptional activity and triple mutation at these three cysteine residues showed impaired function to form molecular condensate which is essential for formation of transcriptional complexes⁵⁶. These data suggest these three cysteine residues are important for AR-V7 function and its protein stability.

6. Summary

ARTADIs have a unique mechanism of action compared to antiandrogens and do not prevent ligand from binding to the AR-LBD. These inhibitors are structurally distinct from antiandrogens (Figure 3)⁵⁹. Importantly, because these inhibitors bind to N-terminal AR-TAD, their efficacy is not impeded by the presence of androgen or antiandrogens that bind AR-LBD, and they are capable of inhibiting AR-V7 unlike antiandrogens^{44,45,49,59-61}. This difference in mechanism of action allows for ARTADIs to be used in conjunction with existing therapeutic strategies which target the AR-LBD⁶¹. An ongoing Phase I/II clinical trial is currently investigating a combination strategy using the second generation ARTADI masofaniten (EPI-7386) in combination with enzalutamide, versus enzalutamide alone (NCT05075577). Further, sustained androgen ablation or treatment with AR antagonists has been shown to induce the expression of AR-V7, a well characterized mechanism of resistance to current AR-targeted therapies^14-16. As ARTADIs block AR transcriptional activity through action on the AR-TAD, treatment-induced expression of constitutively active AR-splice variants is not expected. Thus, ARTADIs may delay progression to CRPC mediated by expression of AR-V7, but may also allow concurrent treatment with existing antiandrogen therapies. Finally, the AR-LBD is susceptible to mutations which generate a promiscuous receptor capable of being activated by a variety of ligands, including antiandrogens^62,63. ARTADIs not only retain efficacy in this context⁶⁴ but are again not expected to promote the generation of such mutants in the first place – given their unique mechanism of action.

Figure 3 – Chemical structures of anti-androgens and ARTADI inhibitors.

(A) Structures of antiandrogen enzalutamide and steroidogenesis inhibitor abiraterone-acetate. (B) Structures of first-generation ARTADIs (EPI-001, ralaniten/EPI-002 and ralaniten-acetate/EPI-506) as well as second-generation masofaniten/EPI-7386. (C) Structures of SC428 and SC912 reported to target AR-TAD (see refs 44 and 45).

Antiandrogens were initially approved for patients with metastatic castration-resistant disease following biochemical recurrence following initiation of ADT⁶⁵. However due to the success of recent clinical trials, the standard of care has continued to evolve to allow these treatments to be used in earlier settings, generally in combination with ADT^66-69. Currently, antiandrogen based therapies are not approved for non-castrated patients. The lack of benefit of antiandrogens in non-castrated patients was presumed to be due to the inability of the antiandrogen to effectively compete for binding the LBD with dihydrotestosterone⁷⁰. However, given the well-documented adverse effects associated with ADT there is renewed interest in antiandrogen monotherapy without ADT for androgen-sensitive nonmetastatic prostate cancer (EMBARK, NCT02319837; LACOG 0415, NCT02867020)^68,71. As these compounds do not reduce testosterone levels, some side effects such as ADT associated osteoporsosis should be mitigated⁴¹. Antiandrogens, including enzalutamide, significantly increase testosterone and estrogen levels in non-castrated patients and have different adverse events compared to ADT^68,72-75. One-half (50.0%) of patients treated with the enzalutamide monotherapy arm of the EMBARK trial had grade ≥ 3 adverse effects versus 42.7% in the leuprolide (ADT) arm⁶⁸ thereby highlighting the need to carefully consider alternatives to ADT.

While the first-generation ARTADI ralaniten-acetate failed clinical trials due to a poor pharmacokinetic profile (NCT02606123) it was generally well-tolerated⁷⁶. Given that this class of AR antagonist remains effective in the presence of androgens and is not expected to induce AR-V7 expression, there exists the potential therapeutic value of ARTADIs to be used as an alternative option to ADT. For example patients receiving active surveillance that seek a treatment option, patients that have rising PSA after localized therapy, or in combination with any compound that binds to the AR-LBD such as selective AR modulators or bipolar androgen therapy. If effective despite the presence of androgens, ARTADIs should reduce some of the adverse effects of ADT that impact the quality of life of prostate cancer patients. The ability to treat non-castrated patients with an ARTADI has potential near-term clinical impact to reduce tumor burden, maintain sensitivity to inhibitors of the AR-LBD, and diminish adverse side effects associated with castration such as osteoporosis and sexual dysfunction.

• The androgen receptor plays a key role in prostate cancer

• The androgen receptor is transactivated by androgens

• Castration is a critical component in the clinical management of prostate cancer patients

• Inhibitors of the androgen receptor transactivation domain are a new class of drug with a unique mechanism of action

• Inhibitors of the androgen receptor transactivation domain may be beneficial as an alternative to castration