Breaking androgen receptor addiction of prostate cancer by targeting different functional domains in the treatment of advanced disease

Abstract

In the last decade, treatment for castration-resistant prostate cancer has changed markedly, impacting symptom control and longevity for patients. However, a large proportion of cases progress despite androgen deprivation therapy and chemotherapy, while still being fit enough for several more lines of treatment. Overstimulation of the androgen receptor (AR) activity is the main driver of this cancer. Targeting biological functions of the AR or its co-regulators has proven very effective in this disease and led to the development of several highly effective drugs targeting the AR signalling axis. Drugs such as enzalutamide demonstrated that the improvement in anti-tumour efficacy is closely correlated with an affinity for the AR and its activity and have established the paradigm that AR remains activity in aggressive disease. However, as importantly, key insights into mechanisms of resistance are guiding the development of the next generation of AR-targeted drugs. This review outlines the historical development of these highly specific agents, their mechanism of action in the context of defective AR activity, and explores the potential for the upcoming next-generation AR inhibitors (ARI) for prostate cancer by targeting the alternative domains of AR, rather than by the conventional ligand-binding domain approach. There is huge potential in these approaches to develop new drugs with high clinical activity and further improve the outlook for patients.

Fig. 1

The androgen receptor (AR) signalling pathway. The AR protein is structurally made up of three main functional domains: The N-terminal domain (NTD), the DNA-binding domain (DBD) and the ligand-binding domain (LBD). A small hinge region lying between DBD and LBD is not shown. In its inactive form, the AR is bound to chaperone proteins of the Heat-shock protein (HSP) class. In addition to HSPs, Choline kinase alpha (CHKA) has recently been identified as an AR chaperone, bound both in the cytosol and the nucleus. Dihydrotestosterone (DHT) is synthesised from testosterone locally and binds the LBD of the AR, specifically the ligand-binding pocket. Upon binding, a conformation change in the AR occurs, where the HSP chaperones dissociate and the NTD and LBD form interactions (N/C interaction). Nuclear translocation occurs due to conformation change-induced exposure of a nuclear localisation signal (NLS) within the hinge region. Subsequent interactions of the NLS with the nuclear import proteins drive the AR into the nucleus. Inside the nucleus, the AR can exist in both dimer and monomer forms, however when binding to androgen response elements (AREs) of the DNA, the AR almost always exists as a dimer. Androgen-bound transcriptionally active AR dimers assume parallel “head-to-head” and “tail-to-tail” conformations, wherein the NTD, DBD and LBD all form the dimerisation interface. The two NTDs within the AR dimer adopt different conformations and surround the LBDs, with one NTD mainly interacting with its own LBD through intra-molecular N/C interaction, while the second NTD appears to form both intra- and intermolecular N/C interactions by interacting with both LBDs within the AR dimer. Transcriptional initiation can then occur by interactions with coactivators and subsequently transcription of AR target genes by other cofactor enzymes. These genes increase proliferation and survival signals to stimulate cell growth .

Fig. 2

The AR gene/protein structure. The Androgen receptor gene resides on the X chromosome with 8 exons coding for the protein. Exon 1 codes for the N-terminal domain (NTD), exons 2 & 3 code for the DNA-binding domain (DBD) and exons 4-8 coding for the hinge region (HR) and the ligand-binding domain (LBD). The HR harbours the canonical Nuclear Localisation Signal (NLS) which direct the AR into the nucleus upon ligand binding. The DBD contains the P-Box recognition helix (577-581) and the D-Box site (596-600) that regulate specificity to DNA and receptor dimerisation respectively. Activation Function-1 (AF-1) resides within the NTD, with the activation domains TAU-1 and 5, while Activation Function-2 (AF-2) resides mainly within the 12^th and final helix of the LBD .

Fig. 3

Ribbon diagram of the AR LBD with dihydrotestosterone (green) and a small peptide (FxxLF - not shown here- to mimic the N/C interaction). Helix 12 is coloured in cyan. (PDB:1T7M).

Fig. 4

Chemical structures of ARIs for prostate cancer treatment. These inhibitors bind the ligand-binding domain of the AR to inhibit the AR-dependent transcription. The upper panel shows first-generation ARIs, steroidal (cyproterone acetate) and non-steroidal (Bicalutamide, 2-Hydroxyflutamide, Nilutamide). The lower panel shows second-generation non-steroidal ARIs. The bulky benzene ring is common in these structures to effectively fill the ligand-binding domain of the AR [32,49,50,51,43,35,52].

Fig. 5

Cis mechanisms underlying ar gene activation and resistance to treatment. AR point mutations in the ligand-binding domain (Mut-LBD) cause a conformational change in the protein which can then bind to alternative ligands such as glucocorticoids, adrenal androgens and metabolites (receptor promiscuity). Contrastingly, the mutation can cause ARIs to fail by causing an antagonist-to-agonist switch, where the change in conformation allows the LBD -targeting ARIs to bind but to activate growth and survival pathways instead of suppressing. AR gene or enhancer amplifications can occur, where increased numbers of AR protein molecules are synthesised due to a greater copy number of the AR gene or hyperactivation of enhancer-induced gene expression. Consequently, more DHT can bind the receptor thus causing the growth and survival of cells to increase. AR variants (AR-Vs) can be synthesised by alternative splicing and/or AR gene rearrangements where variants can lose some or all the LBD. Without the LBD, receptors cannot be activated by ligands but can become constitutively active, allowing the signalling of the AR to occur continuously [59]. All variants shown contain the NTD and DBD (coded for by exons 1,2 & 3), however, exons 4-8 that code for the LBD are partially omitted or replaced by cryptic exons (CE). AR-V7 is coded by exons 1/2/3/CE3, Arv567es (also known as AR-V12) is coded by exons 1/2/3/4/8/9c and AR-V9 is coded by exons 1/2/3/CE5 [[58], [59], [60]].

Fig. 6

Androgen receptor (AR) ligand-binding domain (LBD) point mutations identified in CRPC patients. The gene encoding the AR has 8 exons, with most of the point mutations occurring in the 8^th exon (shown in red). A few LBD point mutations occur in other exons and are shown in blue. The point mutations cause a change in amino acids when the gene is transcribed and subsequently translated for example L702H indicates that leucine (L) at position 702 changes to histidine (H). These mutations allow for the LBD to change in conformation and/or activity, thus allowing for changes in binding ligands or inhibitors [65].

Fig. 7

Activation and amplification of an AR enhancer drives prostate cancer progression. In primary prostate cancer, enhancers lay dormant due to histone and DNA hypomethylation or hypoacetylation. The enhancer, upstream of the AR gene locus, can be epigenetically activated by histone (H3K27) acetylation which will increase AR expression. In metastatic CRPC, this enhancer is often seen to be activated and amplified where it drives disease progression by further increasing AR expression which thus increases AR protein abundance [76].

Fig. 8

The generation of Androgen Receptor-Variants (AR-Vs) can be caused either by a combination of intragenic rearrangements and alternative splicing or alternative splicing alone with examples shown. a) Intragenic rearrangements are seen in the prostate cancer cell line CWR22Rv1 cause tandem duplication of AR cryptic exons resulting in a greater abundance of cryptic exons leading to the increased likelihood of their preservation in alternative splicing. The fusion occurs at the 66,924,525^th base pair of the first AR gene and the 66,889,976^th base pair of the second AR gene, with the fusion also containing a de novo 27 base pair sequence. b) Alternative splicing of AR pre-mRNA results in the formation of a truncated AR variant, AR-V7. Exons 1, 2, 3 and cryptic exon 3 are translated to form AR-V7, a constitutively active form of the AR [[60],[86]]. Conventional exons shown in pink and purple while cryptic exons shown in green. Exon 1b expressed only when alternative promoter used.

Fig. 9

Chemical structures of AR inhibitors that do not target the ligand-binding domain. AR degraders: Niclosamide , ASC-J9 , UT-34 and ARD-69 . NTD-targeting agents: EPI-001 , SINT1 and QW-07 . DBD-targeting agents: Pyrivinium and VPC-14449 . Not all structures were available for all inhibitors discussed in the review.

Fig. 10

Summary of current and experimental treatments that target the AR signalling axis via different mechanisms. Current therapies such as Enzalutamide competitively inhibit the ligand-binding domain (LBD) while therapies such as Abiraterone acetate inhibit the production of dihydrotestosterone (DHT). Emerging therapies have been able to bind the N-terminal domain (NTD) of the androgen receptor and thus suppress transactivation of AR target genes, bind the DNA-binding domain (DBD) thus preventing interaction with DNA, inhibit AR dimerisation so that it cannot interact with DNA or upregulate AR protein degradation to reduce the abundance of the receptor.