Testosterone in Androgen Receptor Signaling and DNA Repair: Enemy or Frenemy?

In this issue of Clinical Cancer Research, Hedayati and colleagues (1) report a novel, but possibly counterintuitive, strategy of exploiting sequential androgen suppression and stimulation to increasingly sensitize prostate cancer cells to radiotherapy. They attributed this effect to the induction of transient DNA double-strand breaks (DSB) following a combination of androgen deprivation and supraphysiologic levels of dihydrotestosterone. This DNA damage response occurs through an exclusive interaction between the androgen receptor (AR) and topoisomerase II beta (TOP2B). This increased accumulation of DSB following irradiation was significant enough to inhibit tumor growth in vivo.

The combination of androgen suppression and radiotherapy is a time-honored treatment regime with proven efficacy of advancing cure rates in patients with high-risk and locally advanced prostate cancer. Numerous randomized trials, testing a variety of castration approaches, have conclusively confirmed the synergism between androgen suppression and radiation, reporting better tumor control rates (pooled HR 1.67) and reduction of distant metastases (pooled HR 1.63) when compared with radiotherapy alone (2). On the basis of the documented effects on both local and systemic disease, it is widely perceived that combined modality therapy with suppression of the AR axis simultaneously enhances the cytotoxic effects of radiotherapy on the primary prostate cancer and also targets occult metastases.

The mechanistic basis of this clinical observation is starting to mature as increasing evidence indicates that activation of the AR axis exerts an influence on the cellular DNA repair machinery and overall DNA damage response (Fig. 1). In response to genotoxic stress, androgen ligand binding to the AR triggers a cascade of signaling events that promotes the assembly of transcriptional elements leading to the overexpression of DNA repair genes involved in DNA damage sensing, DSB repair, base excision repair (BER), and mismatch repair (MMR; ref. 3). The process of AR-dependent transcriptional regulation is further modulated by the receptor binding with repair proteins, such as Ku, DNA-dependent protein kinase (catalytic subunit; DNA-PKcs), and PARP1; these proteins also function as AR coactivators (4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Interplay between androgen receptor (AR) axis manipulation and DNA repair in prostate cancer. AR is overexpressed in response to radiotherapy (RT), and mediates radioresistance by transcriptional upregulation of DNA repair genes. Ku70, DNA-PK, and PARP1 partake in a positive-feedback circuit by acting as transcriptional coactivators. AR axis suppression counters this process and radiosensitizes by inhibiting DNA double-strand break (DSB) repair. Paradoxically, AR axis stimulation with supraphysiologic levels of dihydrotestosterone (DHT) triggers the assembly of enzymes at genomic fragile sites, contributing to DSB formation. Such androgen manipulation can be a potential radiosensitization strategy through increased generation of DSB. Copy number alterations of NBN (11) is a potential molecular biomarker of radioresistant prostate cancer; Ku70 and DNA-PK are key proteins involved in the synergism between AR axis suppression and RT (5, 7); PARP1 represents a “druggable” target (pink circle). BER, base excision repair; HR, homologous recombination; HRE, hormone response element; LHRH, luteinizing hormone-releasing hormone; MMR, mismatch repair; NHEJ, nonhomologous end-joining; PSA, prostate-specific antigen; SSB, single-strand break.

Given the positive-feedback circuit linking AR axis stimulation and DNA repair, it would suggest, in principle, that targeting the AR axis represents a very sound and attractive strategy for potentiating the DNA-damaging effects of cytotoxic therapies in prostate cancer. Importantly, it was observed in primary prostate cancer specimens undergoing neoadjuvant androgen deprivation therapy (ADT), that inhibition of the AR axis leads to a reduction of Ku protein expression in post-ADT prostate biopsies (5). In this first-in-human proof-of-mechanism study, Tarish and colleagues demonstrated longitudinally that castration primarily affected both Ku and DNA-PKcs expression in response to radiotherapy, leading to significant impairment of the nonhomologous end-joining (NHEJ) pathway of DSB (5). In parallel, Polkinghorn and colleagues and Goodwin and colleagues also observed a radiosensitization effect as a consequence of androgen blockade, and attributed this to the transcriptional downregulation of DNA repair–related genes, with DNA-PKcs being a key target (6, 7).

Stimulation of the AR axis, particularly with supraphysiologic levels of dihydrotestosterone, can also contribute to DSB formation, through an AR-driven recruitment of enzymes to common fragile sites in the genome that are prone to illegitimate rearrangements (Fig. 1). As first observed by Lin and colleagues, ligand-bound AR acts to foster chromosomal rearrangements. This work also demonstrated that AR binding promotes site-specific DSB formation through a novel enzymatic machinery comprising of activation-induced cytidine deaminase (AID) and LINE-1 repeat-encoded ORF2 endonuclease (LINE-1 ORF2; ref. 8). Another mechanism proposed by Haffner and colleagues involves the corecruitment of AR and TOP2B to sites of TMPRSS2-ERG genomic break-points, facilitating formation of transient DSB secondary to TOP2B catalytic cleavage (9). In addition, NKX3-1 may be responsible for accelerating the repair of such breaks or increasing genetic instability with clonal selection of mutator phenotypes, and allelic deletion of this gene has been linked to clonogenic radioresistance and tumor recurrence after radiotherapy (3). The induction of DSB with androgen stimulation offers an additional paradigm to AR manipulation for therapeutic synergism when combined with cytotoxic cancer therapies, and may affect both primary tumor and metastatic phenotypes.

This concept was most recently clinically tested by Schweizer and colleagues where men with low to moderate metastatic burden castrate-resistant prostate cancer were exposed to spikes of supraphysiologic levels of dihydrotestosteone in the background of continuous castration therapy (10). In this study of 16 men, some of whom had progressed on second-generation antiandrogen therapies, clinical responses (both biochemical and radiological) were recorded in 50% of the treated cohort. Consistent with Haffner and colleagues (9), the authors linked the clinical efficacy to incremental accumulation of DSB, as a result of stabilization of AR-induced transient DSB following etoposide (a TOP2 inhibitor). AR “overstabilization” contributing to loss of DNA relicensing, and subsequent mitotic death was also proposed as another mechanistic cause for tumor growth inhibition in vivo. When taken together, both experimental and clinical evidence support the therapeutic synergism between bipolar androgen stimulation and castration in prostate cancer through modulation of DSB induction and repair.

To validate these findings, trials will need to be conducted to address issues pertaining to patient selection, and scheduling of androgen deprivation and stimulation. In clinical practice, many men will receive at least 2 months of LHRH agonist prior to commencing radiotherapy. One combinatorial approach could be to initiate androgen stimulation 1–2 days before radiotherapy, and repeating every 2–4 weeks during treatment, given that stimulation lasts for approximately 2 weeks following an injection of supraphysiologic dihydrotestosterone and track DSB using in situ DSB biomarkers such as γH2AX or TP53 binding protein-1 (53BP-1). Note that the transient nature of DSB induced by androgen stimulation is also an important mechanistic consideration, as majority of these DSB are repaired within 24 hours (9). However, the repression of NHEJ by continuous castration should also impede the repair of these site-specific DSB, resulting in prolonged stabilization of these lethal lesions. These are testable hypotheses with clinical trial specimens.

Of note, the field also needs molecular biomarkers that can identify specific patients with AR-dependent prostate cancers that are aggressive and at risk for recurrence following radiotherapy alone so that they can be offered combined ADT and radiotherapy. Our group recently reported that NBN copy number gain and high percent genome aberration are highly predictive for biochemical relapse following radiotherapy, and such patients may be suitable for intensification with added androgen modulation (11, 12). The safety profile of this more intensive treatment, if incorporating androgen stimulation as well, requires detailed evaluation in prospective clinical studies. Finally, the proposed mechanism(s) of radiosensitization with ADT relies on a functional AR axis; whether these novel approaches will also be as efficacious in hormone-insensitive tumor clones in later stage disease (e.g., metastatic castrate-resistant or neuroendocrine prostate cancers) requires close study.

In conclusion, the synergism between androgen suppression and radiotherapy that has been observed for the past 15 years can now be partially explained by modulation of repair of radiotherapy-induced DSB. New mechanistic insights into the complex interplay between androgen manipulation and DNA repair are now giving rise to novel treatment strategies with radiotherapy or other agents to sensitize aggressive prostate cancers and improve cure.

• Received March 15, 2016.
• Accepted March 24, 2016.

• ©2016 American Association for Cancer Research.