The 5α-androstanedione pathway to dihydrotestosterone in castration-resistant prostate cancer

Abstract

The survival and progression of prostate cancer is generally dependent on expression of the androgen receptor (AR), as well as the availability of endogenous AR agonists. Originating from the gonads, testosterone is released into circulation and is converted by steroid-5α-reductase (SRD5A) in prostate cancer to 5α-dihydrotestosterone (DHT), potently activating AR and driving tumor progression. Advanced prostate cancer is initially treated with gonadal testosterone depletion, which suppresses this cascade of events and typically leads to a treatment response. Eventually, resistance to testosterone deprivation occurs with “castration-resistant” prostate cancer (CRPC) and is driven by the intratumoral synthesis of DHT. The generation of DHT occurs in large part from adrenal 19-carbon precursor steroids, which are dependent on expression of CYP17A1. Although the path from adrenal precursor steroids to DHT was generally thought to require 5α-reduction of testosterone, recent data suggest that it instead involves conversion from Δ⁴-androstenedione by SRD5A isoenzyme-1 to 5α-androstanedione, followed by subsequent conversion to DHT. The 5α-androstanedione pathway to DHT therefore bypasses testosterone entirely. Abiraterone acetate effectively inhibits CYP17A1, blocks the synthesis of androgens and extends the survival of men with CRPC. Further progress in the hormonal treatment of CRPC is dependent on an understanding of the mechanisms that underlie CRPC and resistance to abiraterone acetate.

The androgen axis is central to the progression and treatment of prostate cancer. Essential components of this axis require the expression of the androgen receptor (AR) and the generation of endogenous AR agonists. The androgen signaling pathway is intimately involved from tumor initiation and invasion, to the development of metastatic disease. Translocation of the androgen-controlled TMPRSS2 regulatory region proximal to a member of the ETS-family oncogenes occurs in the transition between high grade prostatic intraepithelial neoplasia and invasive prostate cancer, driving oncogene expression¹. The requirement for expression of these oncogenes, elicited by the androgen axis, continues to very late and resistant states of disease². Therefore, the mechanisms that regulate the androgen axis, from the generation of ligand, to AR expression, to the response of AR-regulated genes, all represent steps that are potential points of intervention for the development of new pharmacologic therapies. A precise understanding of this pathway is required for further advances in the treatment of prostate cancer.

Gonadal testosterone deprivation

Physiologic serum concentrations of total testosterone (T) are generally > 300 ng/dl (10.4 nmol/l)³. In prostatic tissue, T is converted by steroid-5α-reductase (SRD5A) to 5α-dihydrotestosterone (DHT). T is capable of binding AR in the absence of metabolism to DHT but the latter is several fold more potent and is the major androgen bound to AR in the prostate cell nucleus^4,5. Although two isoenzymes exist, in the prostate expression of SRD5A2 is greater than that of SRD5A1⁶. In prostatic tissue, SRD5A enzymatic activity results in DHT concentrations that are several fold higher than T and this ratio is reversed upon treatment with pharmacologic blockade of SRD5A^7,8. The effect of gonadal testosterone deprivation is therefore likely due in large part to the depletion of intratumoral DHT. However, despite 94% reductions in serum T with medical castration, intraprostatic T and DHT are reduced by only 70 and 80%, respectively⁹. The apparent availability of precursors for the synthesis of residual intraprostatic androgens with medical castration provides a clue as to the mechanisms of resistance to depletion of gonadal T¹⁰. Nonetheless, responses to gonadal T depletion therapy occur in the majority of cases, although the response in the metastatic setting is nearly always temporary¹¹.

Castration-resistant prostate cancer

Disease that progresses in the presence of gonadal T depletion is termed “castration-resistant” prostate cancer (CRPC). Multiple lines of evidence suggest that the switch from hormone-responsive to CRPC is regulated by a gain-of-function in AR^12-14. A multitude of mechanisms have been implicated in increasing AR-driven transcription. These vary from alterations in coactivator/corepressor expression, ligand-independent function or ligand-sensitization through growth factors or their receptors, post-translational modification of AR, increased AR expression, AR mutations that broaden the specificity for ligand, and intratumoral steroidogenesis that increases the availability of T and/or DHT ^15-22. All of these factors may contribute to some extent to the development of CRPC in a manner that is probably highly dependent on the molecular pathogenesis of individual tumors. However, the finding of biologically significant concentrations of intratumoral androgens common to the majority of tumors, coupled with clinical responses to depletion of these androgens, implicates intratumoral steroidogenesis as a major and frequent driver of CRPC^21-24.

Essential components of intratumoral steroidogenesis

The synthesis of all steroids originates from cholesterol²⁵. The structural features of the initial substrate and the final product(s) must be considered in the pathway(s) from cholesterol to T and/or DHT. Cholesterol has a 27-carbon, 3β-hydroxyl, Δ⁵-structure (double bond between carbons 5 and 6). Eventual conversion to 19-carbon T and/or DHT necessitates the departure of 8 carbons through 2 enzymes, 3β-hydroxyl oxidation to 3-keto, Δ⁵ isomerization to Δ⁴, and 17-keto reduction to a 17β-hydroxysteroid. In the adrenal, P450scc cleaves cholesterol to 21-carbon pregnenolone, which is then a substrate for CYP17A1 hydroxylase and 17, 20-lyase activity, yielding 19-carbon dehydroepiandrosterone (DHEA). DHEA and its sulfate are the major androgen precursor steroids in serum and the probable major source(s) of intratumoral androgens in CRPC^26,27. In CRPC, DHEA is converted to Δ⁴-androstenedione (AD) by 3β-hydroxysteroid dehydrogenase/isomerase (3βHSD), which is encoded by two isoenzymes²⁸. 3βHSD1 is generally thought to be expressed in peripheral tissues and 3βHSD2 the responsible enzyme in steroidogenic organs^29,30. However, expression of transcripts encoding both isoenzymes has been detected in CRPC tissues^21,31.

It generally had been assumed that the next step in the CRPC pathway is conversion of AD to T^12,32. The presumptive conversion of AD to T was implied in part from the observations that intratumoral T concentrations and expression of AKR1C3, which is capable of converting AD to T, are both increased in CRPC^21,33. Expression of SRD5A1 is increased in CRPC and was generally thought to be required for the conversion from T to DHT^21,32-34.

An alternative possibility to synthesis from adrenal precursor steroids is de novo androgen synthesis from cholesterol, taking place entirely in CRPC tissue. This has been reported in CRPC cell lines³⁵. However, the abundance of adrenal precursors in serum, the requirement for only 2-3 enzymes for the conversion from DHEA to T and DHT, and comparisons of flux through both pathways, together suggest that the adrenals are the main source for intratumoral androgens in CRPC²⁷.

Abiraterone acetate

CYP17A1 enzymatic activity is required for the conversion of 21-carbon steroids to 19-carbon androgens, no matter the relative contribution of the adrenal vs. de novo pathways to intratumoral T and DHT. Abiraterone acetate potently blocks both CYP17A1 hydroxylase and 17, 20-lyase activity³⁶. Initial clinical studies of abiraterone acetate demonstrated declines in serum T and AD concentrations; however, pituitary compensation by luteinizing hormone hypersecretion, resulted in some gonadal testosterone recovery in eugonadal males³⁷. In phase I/II trials in men with CRPC, PSA declines greater than 50% occurred in approximately two-thirds of patients who had not been previously treated with chemotherapy^23,38. Pretreatment concentrations of DHEA, DHEA-S and AD in serum were associated with treatment response²³. In a phase III trial of abiraterone acetate plus prednisone versus placebo plus prednisone in CRPC patients previously treated with docetaxel, overall survival was 3.9 months longer in the abiraterone acetate-prednisone group³⁹. Progression-free survival, PSA response rate and time to PSA progression were all in favor of the abiraterone acetate-prednisone group. On the basis of these data, abiraterone acetate was approved by the United States Food and Drug Administration in April 2011 for the treatment of metastatic CRPC in men previously treated with docetaxel. Notably, progression-free survival in the abiraterone acetate-prednisone arm was 5.6 months, raising the issue of treatment options in resistant tumors. Abiraterone acetate is administered orally, is generally well-tolerated and clinically active, all suggesting that widespread use of this drug will lead to a large population of men with abiraterone acetate-resistant CRPC. Therefore, this is an urgent area of investigation. Although early preclinical data in mouse xenograft models suggest that sustained steroidogenesis is in part responsible, there is very little insight into the mechanisms that may permit androgen synthesis in clinical tumors under abiraterone acetate treatment conditions⁴⁰.

The 5α-androstanedione pathway to DHT

Defining potential points of intervention in abiraterone acetate resistant tumors must be preceded by a firm understanding of the mechanisms underlying abiraterone acetate- and castration-resistance. Increased concentrations of T and overexpression of AKR1C3 in clinical CRPC appear to support the notion that AD is converted to T, which is the immediate precursor to DHT³². However, AD is also a 3-keto, Δ⁴-steroid, similar to T, making it a substrate for SRD5A1 that is possibly even better than T^41,42. An alternative possibility to synthesis through T is that 5α-reduction of AD results in synthesis of 5α-androstanedione (5α-dione), which may be 17-keto reduced to DHT (Figure 1). We have recently shown that AD is preferentially 5α-reduced to 5α-dione, rather than 17-keto reduced to T, in multiple models of CRPC, as well as freshly biopsied tissue from 2 patients with metastatic CRPC⁴³. Any T that is synthesized from AD is actually a poorer substrate for SRD5A compared to AD. Therefore, the preferred route from adrenal precursors to DHT is AD → 5α-dione → DHT (5α-dione pathway), rather than AD → T → DHT (conventional pathway). Furthermore, silencing the expression of SRD5A1 blocks the conversion of AD to 5α-dione and the eventual synthesis of DHT in CRPC⁴³. This suggests that the SRD5A1 up-regulation described in multiple clinical studies of CRPC tissue, serves to increase flux from AD → 5α-dione, rather than T → DHT, as previously assumed³². The increase in expression of AKR1C3, which reduces 17-keto to 17-hydroxysteroids, may serve to convert 5α-dione → DHT, rather than AD → T⁴⁴. These unanticipated findings on the origins of DHT in CRPC suggest that there should be a reevaluation of current strategies of assessing response and resistance to various hormonal therapies, including abiraterone acetate, as well as a reconsideration of the consequences of potential points of pharmacologic intervention.

Figure 1.

The pathway overview of DHT synthesis. The synthesis of intratumoral DHT requires enzymatic modification of the 3-, 5-, and 17-positions of the steroid backbone by 3β-hydroxysteroid dehydrogenase (3βHSD), steroid 5α-reductase (SRD5A) and 17β-hydroxysteroid dehydrogenase (17βHSD) isoenzymes. The widely accepted conventional pathway requires conversion of AD to T (red arrows). An alternative possibility circumvents the requirement for T by 5α-reduction of AD to 5α-dione (green arrows).

Clinical implications of the 5α-androstanedione pathway

An understanding of the relevant and required intratumoral intermediates en route to DHT is necessary in order to accurately characterize androgen depletion downstream of abiraterone acetate. Intratumoral T is probably not the best marker of response or resistance to abiraterone acetate, given that this is not the major DHT precursor⁴⁵. The spectrum of intermediate metabolites as markers of response or resistance should be expanded to include 5α-dione and probably other 5α-reduced androgens, particularly given that several 5α-reduced androgens are reversibly interconvertible to DHT²⁷.

Intratumoral synthesis of DHT through the 5α-dione, rather than the conventional pathway via T, alters the consequences of current and potential pathway inhibitors. Trials of dual SRD5A inhibitors in CRPC only demonstrated very modest clinical activity^46,47. This might be interpreted to indicate that DHT is unimportant in driving CRPC. Alternatively, the effects of blocking the 5α-dione pathway through genetically or pharmacologically inhibiting SRD5A1, results in diverting AD instead to increased synthesis of T⁴³. The diverted pathway resulting in increased intratumoral concentrations of T probably substitutes in part for the inhibition of DHT synthesis, despite the more modest AR agonist activity of the former androgen. One possible solution to this pitfall of SRD5A inhibition is the move one step upstream in the pathway of DHT synthesis. Pharmacologic inhibition of 3βHSD blocks the conversion of DHEA to AD, AR nuclear translocation, expression of AR-responsive genes and cell growth²⁸. Similar to diversion of AD by 17-keto reduction to T with SRD5A inhibition, it is possible that DHEA is also diverted by 17-keto reduction to Δ⁵-androstenediol with 3βHSD inhibition. However, just as with DHEA, Δ⁵-androstenediol must also be metabolized by 3βHSD in order to induce the AR-response, for both wild-type and the LNCaP mutant AR²⁸. Several pharmacologic inhibitors of 3βHSD exist but they all have problems, such as partial AR agonism, that make them untenable for use in the treatment of CRPC¹³.

Conclusion

The presence of intratumoral DHT in CRPC was first noted over 30 years ago. The survival benefit conferred by treatment with abiraterone acetate is the clearest evidence yet that intratumoral androgens are a main driver of the development of resistance to hormonal therapy and progression with CRPC. Although the pathway from adrenal precursor steroids to intratumoral synthesis of DHT was widely believed to require T, the main route instead circumvents T through the 5α-dione pathway and requires expression of SRD5A1. Strategies for the development of better therapeutic approaches should account for the unanticipated dominance of the 5α-dione pathway in CRPC.