It is estimated that one in six men will be diagnosed with prostate cancer (PCa) during their lifetime. Despite the wide application of prostate-specific antigen (PSA) screening, a significant number of patients still present with an advanced form of the disease, and up to 40% of patients treated with primary therapy, with curative intent, will experience disease progression. For a large proportion of these patients, hormonal therapy in the form of medical or surgical castration [i.e. androgen deprivation therapy (ADT)] can improve survival.
Clinicians treating PCa should have a complete understanding of the molecular background and effects of androgens and ARs to optimize treatment of their patients.
The first step in the history of testosterone and PCa was the seminal discovery by Charles Huggins and Clarence Hodges nearly three-quarters of a century ago, in the 1940s, that surgical castration has a beneficial effect on patients with advanced PCa.1 The next step was the discovery concerning peptide hormone production of the brain by Roger Guillemin and Andrew Schally, who jointly received the Nobel Prize in 1977.2 This was a crucial step to enable the development of chemical castration. Since then, we have gained further insight into the intracellular signalling of testosterone and its role as the primary driver for the growth of PCa cells. We have learned that, in addition to peptide hormone production, the androgen receptor (AR) plays a crucial role in the control of the PCa cell. Progress since the 1970s has led to a large variety of ADTs; however, despite being treated with ADT, PCa can develop into a castration-resistant state. There are a number of mechanisms that contribute to the development of castration-resistant prostate cancer (CRPC) including reactivation of ARs despite castrate levels of circulating androgens, neuroendocrine differentiation and transformation of the disease to an AR-negative, and truly hormone-refractory, disease. CRPC usually develops approximately 1–3 years after the start of ADT and patients with CRPC have an average survival of approximately 30 months. The poor quality of life and prognosis of these patients has invigorated research with the goals of understanding the biology behind PCa progression and identifying novel therapeutic targets. Paradoxically, some patients with metastatic PCa experience symptomatic improvement after exogenous testosterone administration3 and some patients with CRPC experience a PSA decline after exogenous testosterone administration.4,5
Effect of androgens in the normal prostate
The majority of testosterone is synthesized under the control of the hypothalamus and the anterior pituitary gland by the Leydig cells of the testis and a minority of testosterone is produced by the adrenal gland. Approximately 3% of the testosterone in blood is bioavailable, while the rest is bound to a sexual hormone that is also bound to globulin and albumin.6 Testosterone is necessary for the sexual differentiation in the fetus and maturation in the adult as well as for the development and growth of the prostate.7
Testosterone is converted intracellularly by the enzyme 5α-reductase in the basal and secretory epithelial cells to dihydrotestosterone (DHT), which is the more potent ligand for the AR. After ligand binding and transactivation, the DHT–AR complex is translocated from the cytoplasm to the nucleus, where the activation of transcription of the relevant genes occurs. A sufficient amount of the activated complex, DHT–AR, is required to keep the homeostatic balance of proliferation and survival signals in the prostate, which consists mainly of two cell types – stromal and epithelial cells. The epithelium consists of basal and luminal cells; however, only the latter express ARs. These luminal cells express the prostate-specific differentiation markers, one of which is PSA. Expression of these marker genes is enhanced by the binding of the DHT–AR complex to the androgen-responsive elements in the nuclei of these luminal cells. The AR in the nuclei of the secretory epithelial cells does not directly regulate the cell's survival or directly affect proliferation and survival of the basal cells; instead, in the healthy prostate, the epithelial cells depend on the paracrine signalling of andromedins, which are growth and survival factors produced by the stromal cells, for regulation of survival. The activated AR within the nucleus of the stromal cells triggers the production of andromedins. These andromedins diffuse back across the basal membrane and enter the epithelial cell to keep them alive. Insufficiency of andromedins results in upregulation of apoptotic signalling via thyroid growth factor receptor in epithelial secretory cells, inducing their degeneration.8
The AR is a protein complex consisting of three domains: the N-terminal domain, the ligand-binding domain and the DNA-binding domain. The ligand-binding domain is required for activation of the AR by androgens and, after translocation to the nucleus, it binds to specific DNA sequences called androgen-responsive elements via the DNA-binding domain, thus promoting protein synthesis. The N-terminal domain is needed for the transcriptional transactivation activity of AR.9
Androgens and androgen receptors in prostate cancer
In PCa cells, certain molecular features of the basal and epithelial cells seem to be highly correlated. In certain circumstances, the previously mentioned paracrine AR signalling mechanism is converted into an autocrine mechanism; therefore, these cells are less dependent on stromal cell-derived factors. The AR acquires a ‘gain of function’ and, after stimulation by androgens, is capable of direct production stimulation of growth and survival factors.10 In other words, the vital cell functions can be autonomously regulated by these cells. Nevertheless, at that stage, the cell depends on the homeostatic balance of androgens in the surroundings.
Initiation of ADT results in a relatively rapid decline in the available serum androgens and, owing to apoptosis of androgen-sensitive cells, temporary tumour remission. However, after 2–3 years, virtually all patients will experience disease progression and ADT will no longer be effective. This was previously considered as a hormone-refractory disease state but findings in the last decade have revealed that, even in this disease state, PCa remains responsive to AR signalling. This resulted in designation of a new stage: CRPC.11 Furthermore, serum testosterone levels do not reflect tissue androgen levels within the prostate, where levels of testosterone and DHT are still sufficient to activate the AR.12 This means that AR signalling and, therefore, AR-related gene and protein expression within the microenvironment of the prostate remain functional.13 This is further supported by the fact that recurrent tumours frequently re-express AR target genes such as PSA and nearly 30% of patients with progressive disease after first-line ADT will respond to secondary hormonal manipulations.14
After initiation of ADT, many, but not all, PCa cells will die. This is evidenced by the eventual disease progression after 2–3 years, showing that PCa cells can develop the capacity to adapt to the altered environment. The resumption of AR-dependent transcriptional activity in CRPC is achieved by AR aberrations, rendering them more sensitive to androgens,15 or by producing the necessary ligands autonomously.
If there is a shortage of testosterone after ADT, the PCa cell is capable of producing testosterone itself via the ‘backdoor pathway’. To achieve this, the cell needs certain enzymes, the most important being cytochrome P450 17 (CYP17).16 Another possibility is to increase the conversion rate of testosterone to DHT by increasing 5α-reductase activity.6
Apart from the ligand, the PCa cell can also influence the AR itself and, if the AR is more sensitive, it can be activated despite low available levels of androgens. One of the first steps that PCa cells take is to simply increase the number of ARs by amplification. In addition, mechanisms of AR degradation can be impaired, leading to an increased stability of the ARs. Another form of modulation is to enhance nuclear translocation of the AR. In addition to rendering the AR more sensitive to its natural ligand, the protein structure of the AR can be changed in such a way that other ligands can activate it. This promiscuous AR can then be activated by coactivators such as adrenal androgens, metabolic products of DHT and even by antagonists such as flutamide or bicalutamide. This explains the clinical observations of the benefit of the ‘withdrawal syndromes’. Overexpression of coactivators can further create or enhance the promiscuity of AR. The third possibility is that the AR is activated in the absence of androgens by growth factors such as insulin growth factor or epidermal growth factor; an AR with these capabilities is termed outlaw AR.17,18
In addition to the three mechanisms mentioned above, there is also the possibility to bypass the AR completely by alternative or parallel survival pathways, thus rendering the PCa cell completely androgen independent.
Another method of increasing the availability of AR is to enhance activation of the receptor. A certain proportion of the AR is bound to a nuclear corepressor (NCOR1) and is therefore rendered inactive. PCa cells can upregulate a protein called SIAH2, which removes NCOR1 from the AR, thereby making it sensitive to ligands.19
Finally, the epigenetic state of PCa cells can also be altered so that the AR can be activated without changing its DNA structure. Enhancer of zeste homologue 2 (EZH2) is also known to be overexpressed in CRPC. Usually, EZH2 silences genes, but in CRPC, a switch in the function of EZH2 leads to gene activation instead of transcriptional repression. This is achieved by alterations in the methylation of the AR.20
Taken together, there are a multitude of mechanisms leading to AR-dependent and AR-independent molecular growth promotion in the PCa cell. Understanding these processes is essential for the development of novel therapeutic strategies for early-stage PCa, and especially late-stage PCa, such as CRPC.
Paradox effects of androgens in prostate cancer
It is well accepted that testosterone fuels PCa cell growth; however, as previously mentioned, there are clinical reports showing an inhibitory effect of testosterone on PCa cells. For a long time this was considered to be a paradox since there was a lack of understanding of the molecular background explaining this phenomenon. Through gain of AR function and loss of AR function in epithelial–stromal cell coculture and coimplantation experiments, it was demonstrated in animals that the AR could function in epithelial basal cells as a tumour suppressor, in epithelial luminal cells as a survival factor and in stromal cells as a proliferation promoter.21 This is explained not only by a change in function of AR but also by the role of the different microenvironment. In addition, it could, hypothetically, be that AR acts as a sort of changeover switch whereby certain coactivators influence the AR in different ways. Since we know that PCa can develop at different sites in the prostate, it may be that the stage of differentiation is different and the action of the AR varies accordingly. On the way from a paracrine to a cell-autonomous autocrine mechanism, these changes may be not uniform but overlapping so that testosterone and the activated AR–ligand complex act differently at different sites.
Another finding that may explain paradoxical effects of testosterone is its role in DNA replication licensing. In order to maintain cellular viability and genome integrity, chromosomal DNA must be precisely replicated once and only once before cell division occurs.22 Since the genome is too long to start replication from one end continuing to the other end, DNA synthesis has to occur at one of thousands of origins of replication simultaneously.10 The authorization for replication takes place in the early G1 phase and must be cancelled at the end of this phase to avoid rereplication, which could lead to an S-phase arrest of the cell. In PCa cells, the AR acquires a role as a licensing factor, which means that it must be degraded in order to allow relicensing in the subsequent cell cycle. In the early G1 phase, anti-androgens such as bicalutamide can block proliferation of androgen-sensitive PCa cells by deactivating the AR licensing capabilities. If AR licensing took place, degradation would be mandatory at a later stage in G1 phase. Upregulation of the AR protein and/or ligand-induced AR stabilization during this critical time frame results in growth inhibition. This time dependence for AR stabilization provides a theoretical rationale for intermittent androgen blockade.23
New agents targeting the androgen axis in castration-resistant prostate cancer
Traditionally, advanced PCa is treated by surgical or chemical castration, which reduces testosterone to castrate levels (≤ 0.5 ng/ml). This effect lasts for a sustained period of time until CRPC emerges, which emphasises that traditional therapies causing surgical or chemical castration do not abrogate all sources of testosterone as approximately 10% of the body's testosterone is produced by the adrenal glands and by de novo synthesis of testosterone by PCa cells. Therefore, inhibiting testosterone synthesis in addition to blocking testicular function presents a compelling rationale for treating PCa.24 Examples of novel androgen-modulatory therapeutic modalities that are already available to the clinician, as outlined above, are the following.
Abiraterone acetate (Zytige®, Jannsen-Cilag, Neuss, Germany) is an inhibitor of CYP17, which is a key enzyme in androgen synthesis. It is also required in the cell-autonomous androgen production (backdoor pathway). Two studies, each enrolling more than 1000 patients who were given abiraterone acetate and who either did25 or did not undergo chemotherapy26 for metastatic CRPC, reported improved overall survival. This is probably due to a reduction in available androgen levels27 and clearly indicates that CRPC clinically remains androgen responsive.
A large phase 3 study (known as AFFIRM; NCT00974311) was carried out in men with metastatic CRPC who experienced disease progression after previous treatment with docetaxel-based chemotherapy and were subsequently given enzalutamide (Xtandi®, Astellas, Tokyo, Japan). The trial was stopped in November 2011 after a planned interim analysis showed a 35% reduction in the risk of death in the enzalutamide group compared with the placebo group (median survival 18.4 months vs. 13.6 months respectively; hazard ratio 0.631; P < 0.0001). The data of this study also confirm the central role of AR-signalling throughout the disease and emphasize that CRPC is neither androgen independent nor hormone refractory.28
In the near future, studies should be conducted combining abiraterone acetate and enzalutamide to target testosterone and the AR with the aim of increasing survival in CRPC patients. To our knowledge, it is not currently possible to synchronize cells so that they are in the same cell cycle and testosterone administration is not effective in a clinical setting. To date, all therapeutic modalities target the ligand-binding domain of the AR; however, if the N-terminal is blocked, interaction of the AR and androgen-responsive elements is reduced, resulting in impaired protein expression. A small molecule called EPI-001 has the capacity to block the N-terminal and may be a promising way to further impair the androgen axis in PCa in the future.29
We are in a highly exciting era regarding the management of PCa, especially CRPC. A plethora of novel therapies based on our increasing understanding of molecular networks underlying cellular PCa activity are now available. Optimal use of such therapies for patients requires familiarity with the biological processes of both normal and malignant prostate cells.