Department of Medicine, Division of Oncology (M.E.T.), and Department of Surgery, Division of Urology (S.-M.H.), University of Massachusetts School of Medicine, Worcester, Massachusetts 01655
Address all correspondence and requests for reprints to: Dr. Shuk-Mei Ho, Room 4-746A, Division of Urology, Department of Surgery, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, Massachusetts 01655. E-mail: Shuk-mei.Ho{at}umassmed.edu
Prostate cancer (PC) is the leading cancer diagnosis in men in the United States and is second only to lung cancer in male cancer-related death (1). In 2000 there were approximately 190,000 new cases of PC and 32,000 PC deaths. Not long ago, PC was considered an incidental problem of elderly men and as such was given relatively little research attention. In the last 15 yr there has been an amplification of interest in all aspects of PC biology and treatment. This shift stems from an aging population with an improved overall life expectancy. Basic research has been spurred by the development of sophisticated molecular biology techniques coincident with the development of prostate-specific antigen (PSA) serum testing that facilitates earlier detection and assessment of response to therapeutics.
This review focuses on recent advances in our knowledge of the endocrinology of PC and the application of hormone-based therapies. PC is an endocrine-responsive tumor. Huggins and Hodges (2) demonstrated the therapeutic effect of castration in the 1940s, and this approach remains the mainstay of systemic treatment today. Normal and aberrant prostate cellular function is governed chiefly by androgens. However, other nonandrogenic hormones, whose influence on the prostate can be synergistic or antagonistic, may also contribute to the etiology and progression of PC. A major challenge for PC management is the progression of the disease to androgen independence (AI)1 and hormone refractory state. This review addresses the role of steroid hormones in the development and progression of PC.
Synthesis, transport, and metabolism of sex hormones (Fig. 1)
Androgens are required for normal growth and functional
activities of the human prostate. In men, the major circulating
androgen is testosterone (T), with the testes producing over more than
95% and the adrenal gland less than 5% of the sex steroid
(3). In both tissues, the -5 synthetic pathway, which
results in androstenedione and T production, is the predominant
pathway, whereas the
-4 pathway leading to the synthesis of
dehydroepiandrosterone (DHEA) and androstenediol is the
minor pathway (3). Once synthesized, most of
DHEA is inactivated via sulfation while a small fraction
of this weak androgen is converted to androstenedione, then to T, in
peripheral tissues and in the prostate (4). Testicular
steroidogenesis in the Leydig cells is regulated primarily by the
gonadotropin LH, whereas adrenal androgen production is under the
control of ACTH. In healthy men, adrenal androgens contribute little to
prostatic function, yet after orchiectomy or treatment with GnRH
analogs this source of androgen may become significant in promoting PC
growth (5).
T in circulation is bound to serum proteins, primarily sex
hormone-binding globulin (SHBG), albumin and corticosteroid-binding
globulin (6). It has been estimated that only 23% of T
exists as the bioavailable free form (6). Among the three
proteins, SHBG has the highest affinity for T and, therefore, plays an
important role in regulating the amount of free T available to target
tissues. In end organs such as the prostate, T is converted to the more
potent intracellular androgen, 5-dihydrotestosterone (DHT) by
5
-reductase activities. Two 5
-reductases have been reported. Type
I 5
-reductase is present in most tissues of the body whereas Type II
5
-reductase, encoded by the SRD5A2 gene located on chromosome 2p23
(7), is the dominant isoenzyme in genital tissues,
including the prostate (8). Recent studies have, however,
reported expression of Type I 5
-reductase in prostatic epithelial
cells (9). When finasteride, a specific Type
II 5
-reductase inhibitor, was administered to men for treatment of
benign prostatic hyperplasia (BPH) serum DHT was suppressed by 70% and
prostate DHT by as much as 8590% (8). The remaining DHT
in the prostate likely results from the activity of Type I isoenzyme.
Two large international multicenter, Phase III trials have documented
the safety and efficacy of finasteride in the treatment
BPH (8). However, the efficacy of finasteride
in preventing PC development will have to await final reporting of data
from the Prostate Cancer Prevention Trial, which has been scheduled to
complete its primary end point in October 2004 (10).
Intracellular DHT is rapidly metabolized to 3,
17ß-androstenediol (3
-diol), which can be back-converted to
DHT, or to 3ß, 17ß-androstenediol (3ß-diol), which is instantly
and irreversibly converted to the water soluble, inactive triol
steroids. The key enzyme responsible for this inactivation pathway is
the Type II 3ß-hydroxysteroid dehydrogenase encoded by the HSD3B2
located on chromosome 1p13 (7).
Lastly, it is well established that circulating androgens are also converted to estrogens extragonadally (7590%) at various peripheral tissues mediated by the enzyme aromatase (coded by the CYP19 gene) (11). Interestingly, both aromatase protein and message have been identified in the human prostate, suggesting local aromatization could provide a source of estrogen to the prostate (12, 13). In fact, intraprostatic synthesis of estrogens has been implicated in the etiology of BPH in men (14, 15), but its significance in prostate carcinogenesis is currently uncertain.
Epidemilogic evidence implicating involvement of endogenous sex
hormones in prostate carcinogenesis (Fig. 1)
Because the human prostate is an androgen-dependent organ it is logical to presume that prostate malignancy develops under abnormal androgenic stimulation. This notion is, to some extent, supported by observations that eunuchs do not develop PC (16) and that a higher incidence of PC is found in men who used androgens as anabolic agents or therapeutics (17, 18, 19). Additionally, exposure to maternal estrogen during fetal life leads to the development of squamous metaplasia in the fetal prostate (20), and animal model systems have implicated estrogens as cancer-causing agents in the rat prostate (21). Collectively, these findings raise the possibility that sex hormones contribute to prostatic carcinogenesis. In addition, because androgens undergo extensive metabolic activation and inactivation, both extra- and intraprostatically polymorphisms in genes encoding metabolic enzymes involved in these processes may influence PC risk.
|
Another valuable epidemiologic data set that has shed light on the contribution of sex steroids to prostate carcinogenesis was derived from studies on susceptible populations (22). Overall, a weak but somewhat consistent association between higher circulating T and/or estrogens and high-risk ethnic/racial groups has been observed. For example, Ross et al. (30) found higher T, free T, and estrone in 50 healthy young African-American men (high risk) when compared with 50 young European-American males (low risk). In agreement, de Jong et al. (31) found 71% higher circulating total T levels in Caucasian-Dutch men (high risk) than in Japanese men (low risk). These data would suggest higher circulating levels of sex steroid as PC risk factors. However, in a follow-up study in 1992, Ross et al. (32) found serum T levels in Japanese men (low risk) were not lower than those of the United States whites and blacks (high-risk groups) (32). Similarly, Wu et al. (33) demonstrated higher serum T levels in Asian-Americans (low risk) than in European-American men (high risk). Although circulating sex hormone levels in adult males have not been established as definitive etiological factors for PC, fetal or adolescent exposure to high levels of sex hormones may have long-lasting impacts. Serum T levels were 47% higher in black women than in white women during pregnancy, and E2 levels were 37% higher (34). E2 levels were lower in young black boys than in white boys but higher in older black boys and young black men than in their age-matched white counterparts (35). The notion that sex hormones may have "imprinting" effects on PC susceptibility during adulthood is an intriguing question that warrants further investigation.
Lastly, several polymorphisms in genes associated with sex hormone
activation/inactivation have been implicated in the etiology of PC. To
date, studies have been focused on the following genes: 1) the SRD5A2
gene that encodes Type II 5-reductase responsible for androgen
activation (36, 37, 38); 2) the CYP17 gene that encodes
17
-hydroxylase and 17,20-lyase activity accountable for regulation
of androgen (39) and estrogen (40)
biosynthesis; and 3) the HSD3B2 gene that encodes 3ß-hydroxysteroid
dehydrogenase responsible for DHT catabolism (41, 42). It
has been shown that prevalence of high/low activity alleles in specific
ethnic/racial groups could partially explain some of the observed
difference in PC susceptibilities among these populations
(22).
Androgen receptor (AR) and prostate carcinogenesis (Table 1)
The AR, a member of the superfamily of ligand-dependent nuclear transcription factors (43, 44), was first described in 1969 (45) and cloned in 1988 (46). The gene is located on the X-chromosome at Xq1112, contains 8 exons, and spans a length of approximately 90 kb of DNA (47). Transcription occurs from one of two initiation sites (48, 49), producing two AR transcripts (10 kb and 7 kb long) with different 3' untranslated sequence lengths that are expressed in most target tissues (50). The primary amino acid (aa) sequence is approximately 910919 in length, with a calculated molecular weight of 98 kDa (51). Similar to other steroid receptor proteins, the full-length AR protein contains four functional domains: the amino-terminus regulatory domain, a DNA-binding domain, a hinge region, and the ligand-binding domain (52).
|
Results from immunohistochemistry studies reveal that AR is present in primary site and metastatic PCs regardless of stage and grade, as well as in hormone refractory cancers (58, 59, 60, 61, 62). AR immunoreactivity is heterogeneous within and among cancer foci and bears no apparent correlation with disease prognosis or with duration of responsiveness to hormone therapy (59, 60, 62, 63, 64). However, some studies have noted that an increase in AR heterogeneity or decrease in AR positivity in PC is associated with higher grade and poorer prognosis (59, 60, 63, 64, 65). In contrast to epithelial (adenocarcinoma) cancers, prostate tumors composed primarily of small cells and neuroendocrine cells are devoid of AR (66, 67). Thus, it is possible that clonal propagation of AR-negative malignant cells may give rise to some androgen-independent prostatic carcinoma (66).
Similar to the roles it plays in the normal prostate, AR activation in androgen-responsive PC leads to a complexity of proliferative, apoptotic, and angiogenic events, which collectively control tumor growth (68). It has been shown that androgens enhance expression of cyclin-dependent kinase 2 and 4 and down-regulation of the cell cycle inhibitor p16 (69). Furthermore, AR activation in PC cells leads to up-regulation of p21(WAF1/CIP1), a presumed antiapoptotic factor (70). In this regard, expression of p21 is significantly increased in high Gleason score, fast proliferating, advanced PCs (71). AR seems to exert both positive and negative regulation on the expression of the antiapoptotic factor Bcl-2 (72), and on the synthesis of the growth inhibitor transforming growth factor-ß (73, 74, 75). Androgens also stimulate IGF-binding protein (IGFBP)-5 synthesis and indirectly affect IGF-I action. Likewise, by inhibiting the enzyme neutral endopeptidase, AR may indirectly control neuropeptide-induced PC cell growth (76). Recently, the role of androgen in angiogenesis modulation has also been explored and data relevant to PC treatment have emerged (77, 78, 79, 80).
AR amplification is rarely found in primary cancer but is common in recurrent therapy-resistant cancer [20% (81), 30%, (82)]. Thus, the androgen-deprived environment may favor clonal expansion of cancer cells expressing a high level of AR due to gene amplification. AR amplification is frequent in cancer that had prolonged (>12 month) response to the initial androgen deprivation therapy (81). It is also associated with favorable treatment response to second-line combined androgen blockade (CAB), longer survival times, and a higher level of PSA expression (81, 83, 84). Taken together, these findings suggest that PCs with AR amplification and overexpression of wild-type AR have survival advantage in androgen-deficient environment. Yet, they are also more differentiated, as evident from their high PSA expression, and, therefore, exhibit better and longer responses to antiandrogen therapies.
A parallel mechanism for progression to AI is afforded by mutation of the AR gene. Over 300 (374, last counted in 1999) somatic and germ line mutations have been described in AR-related diseases (Ref. 85 ; http://www.mcgill.ca/androgendb/, ftp.ebi.ac.uk/pub/databases/androgen, http://www.MC33@musica.mcgill.ca). They affect both the coding and 5' and 3' untranslated regions of the AR gene. In PC, a mutation hot spot in exon 5 of the AR gene has been identified (85). Unlike loss-of-function mutations found in the androgen insensitivity syndrome or spinal and bulbar muscular atrophy, PC-associated AR mutations often bestow the receptor with hypersensitivity, promiscuous usage of ligands, and gain-of-function (86, 87, 88, 89, 90, 91, 92, 93). These mutational changes, in general, either permit the receptor to be trans-activated by lower concentrations of androgen, by antiandrogens, and by nonandrogenic ligands. PCs with mutant receptors are, therefore, better apt to survive an androgen-deprived environment (52). In comparison, mutations that cause inactivation of AR are rare. One study, however, revealed that 22% of cases of latent carcinoma in Japanese men contained inactivating mutations of the AR gene, whereas these mutations were not found in latent cancer in American men. Based on these findings, it has been postulated that inactivating mutations may prevent progression of early cancer in Japanese men (94).
Overall, the frequency of AR mutation in localized PC is low (88, 95, 96) but increased with increasing stage even before androgen ablation therapy (96, 97, 98, 99). These findings suggest that AR mutation is perhaps not a causative factor of tumor initiation but occurs before androgen deprivation therapy rather than as a result of the therapy. Androgen deprivation therapy may provide a selective environment for clones with mutations. In support of the latter theory, AR mutations are commonly found in metastatic cancers following androgen ablation therapy, particularly if an antiandrogen has been used (88, 100, 101, 102). Intriguingly, the types of mutation found in the hormone refractory cancers seem to be dependent on the kind of hormonal therapies given to the patients. For example, Taplin et al. (102) have recently reported a high incidence of AR mutations that can be stimulated by flutamide in bone metastasis obtained from patients who received CAB with flutamide, but not in specimens from patients treated with androgen ablation monotherapy. An important lesson to learn from this data are that any antiandrogen therapy could contribute to tumor progression by exerting selective pressure for clones with AR mutations that bequeath growth advantages in the environment induced by treatment. The promiscuous nature of some mutated ARs in PC raises concerns regarding the use of megestrol acetate (MA), a progestational antiandrogen, and minidose diethylstilbestrol (DES) as alternative programs for androgen blockade (5). It does not seem, however, that patients with AR mutations have a shorter survival than those without AR mutations. Lastly, AR mutation may be a mechanism for acquiring novel phenotypes for clonal establishment in distant metastasis sites.
The polyglutamine region, encoded by CAG repeats, in exon 1 of the AR is polymorphic, and its length (831 repeats) is inversely correlated with trans-activation of the receptor, with 40 or more repeats being associated with androgen insensitivity (54). Because the majority of PCs are AR positive, polymorphic differences or mutational changes in polyglutamine length may affect cancer risk and clinical progression. Short CAG repeat length (<22) polymorphism is associated with increased PC risk (103, 104). These alleles are most commonly found in African-American males (75% with short alleles; median length, 18), less frequent in European-Americans (62% with short alleles; median length, 21), and least common in Asian-Americans (49% with repeat short alleles; median length, 22). Thus, AR with shorter CAG repeat lengths are more prevalent in racial groups with higher PC risk. Short CAG repeat lengths also correlate with early onset of PC (105, 106), suggesting that early tumorigenesis is dependent on a more active AR. Other exon 1 AR polymorphisms, the GGN- and GGC-repeat lengths, may also influence PC risk (103, 107, 108, 109), but the functional significance of these polymorphisms is currently unknown. Somatic reduction in glutamine repeats from 24 in nontumor tissues to 18 residues cancer was found in one patient (110). Interestingly, this patient also had an agonistic response to flutamide, suggesting alteration in the length of the polyglutamine stretch may influence trans-activational potential of antiandrogens. A possible mechanism for this to occur may be due to an influence of the CAG repeat size on coregulator interaction with the N-terminal domain of AR (111). An epidemiologic study showed that germ line shortening of the CAG repeat length correlated with advanced clinical stage and poor prognosis (112).
Another important mechanism of PC progression involves ligand-independent activation of AR by growth factors such as IGF-I, keratinocyte growth factor (KGF), epidermal growth factor (EGF), and by cellular signaling regulators such as cAMP, butyrate, interleukin 6, bombesin, and activators of the protein kinase A signaling pathways (113, 114, 115, 116). One or more of these pathways may prove to be an important mechanism sustaining androgen-dependent PC growth after castration or LH-releasing hormone (LHRH) analog treatment. The fact that AR continues to be expressed or even up-regulated in advanced PC provides evidence that AR remains functionally active either in a ligand-dependent or ligand-independent manor (81, 117).
It has now become apparent that the trans-activational activity of a steroid receptor such as the AR is dependent on interaction with receptor-specific or general coregulators (118). Recently identified AR-specific coregulators include ARA54, ARA55, ARA24, ARA 160, and FLH2 (119, 120). Other more general steroid receptor coregulators such as CBP, SRC-1, ARA70, and TRAM-1 as well as oncogenic molecules such as BRCA-1, RB, and Her2/neu have been demonstrated to influence AR trans-activation (118). Altered regulation or mutation of coactivators is a potential mechanism for altered PC growth. The presence of coactivator abnormalities and their functional significance to PC progression are under intense investigation.
In summary, the genesis and progression of PC likely requires a wild-type AR. ARs with higher trans-activating potentials are promotional for these events. Thus, polymorphic differences or mutation-induced contraction of the polyglutamine repeat in exon 1, mutations that leads to hypersensitivity or gain in function of the receptor, and gene amplification are all probable mechanisms for cancer progression. Androgen deprivation therapy introduces new selection criteria for clones expressing different AR phenotypes, and modifies the AR status of the cancers. Ligand-independent AR activation and altered modulations of AR trans-activation by coregulators are likely involved in promoting progression of androgen-dependent PC to an AI state.
Involvement of other hormonal factors
Hormone receptors other than AR may be involved in PC biology. An
analysis of estrogen receptor (ER)- and -ß in normal prostate,
dysplasia, cancer, and hormone refractory PC (HRPC) has demonstrated
expression of ER-ß in HRPC (121). In the PC cell lines
PC-3 and DU0145, ICI 182,780, a selective ER antagonist was able to
inhibit growth. This inhibition was abrogated by an ER-ß antisense
oligonucleotide (122). Future trials will explore the use
of ICI 182,780 in HRPC.
Specific polymorphisms in the vitamin D receptor gene have been identified to offer protection against PC and influence PC susceptibility (123, 124). 1,25-Dihydroxyvitamin D3 deficiency has been implicated in the genesis of prostatic cancer (125). This active metabolite of vitamin D maintains differentiation of prostate epithelial cells and retards the development of PC. In vitro 1,25-dihydroxyvitamin D3 was able to inhibit PC growth in both an androgen-dependent and androgen-independent manner (126). In addition, despite initial concerns regarding potential hypercalcemia, vitamin D analogs are being tested in clinical trials (127).
PRL, GH, and LH, acting alone or in concert with androgens, are known
to regulate normal physiological functions of the prostate
(128). Recent experimental and clinical data now show that
they may play important roles in PC development and progression
(129, 130, 131, 132). Future investigations evaluating hormones as
etiological factors in PC development need to address the roles of
these hormones. Mechanisms responsible for prostate carcinogenesis and
the development of HRPC are likely diverse, however, an
understanding of hormone-controlled growth factors and cell cycle
regulators will hopefully advance therapeutics beyond castration (Table 1).
PC diagnosis and local treatment
With the use of PSA for early diagnosis and the transrectal
ultrasound for prostate biopsy, the majority of PCs are now diagnosed
before overt metastases. The PSA gene, located on chromosome 19 was
cloned in 1987 (133). PSA is a member of the kallikrein
gene family (14 members) and functions as a serine protease. PSA
expression is primarily limited to prostate epithelial cells, although
some expression has been observed in benign and malignant breast cells
(134). In the serum, PSA is bound to
-1-antichymotrypsin. For unclear reasons, more
PSA is bound to
-1-antichymotrypsin in PC
compared with benign prostate enlargement (BPH). The free PSA to total
PSA ratio (<18%) can improve diagnostic specificity by 20% when the
serum PSA is elevated in the range of 410 (135). The
total PSA, percent-free PSA, and digital rectal examination will guide
the recommendation for prostate biopsy, however, the diagnosis of PC
can be difficult and may require more than one biopsy series.
PSA expression is regulated by AR. PSA is an outstanding tumor marker; a possible role for PSA in carcinogensis is under active investigation (136). There have been reports suggesting that PSA may play a beneficial role to inhibit PC growth as well as reports demonstrating that PSA may actually stimulate growth (137, 138). These conflicting results likely stem from the limitations of in vitro cell line work, nonphysiolgic enzyme concentrations and impurities (136). Further investigation may prove the usefulness of PSA as a therapeutic target. Androgen-regulated genes are diverse and together with PSA include many important growth factors and cell cycle regulators (58). Androgens stimulate the expression of EGF receptor, KGF, NEP, ARA 70, CDK 2 and 4, p21, hGK 2, fatty acid synthase, PSA, vascular endothelial growth factor, and IGFBP-5. Androgens repress the expression of transforming growth factor-ß, p16, and bcl-2. There can be either stimulation or inhibition of p27 and IGFBP-5. The balance between these androgen-regulated factors continues to be investigated in the context of PC initiation and progression.
At the time of diagnosis patients are usually stratified into low risk vs. high risk of recurrence based on descriptive features. Risk factors for recurrence include a Gleason score of 7 higher, PSA of 10 or more, and stage T2 (palpable nodule) or greater (139). Therapies are based on these prognostic factors, together with the patients age and general health. In the future, gene chip technology may provide a more accurate biologic profile to guide management.
When cancer is localized to the prostate, standard treatment options include prostatectomy and radiation delivered either by conformal external beam radiation therapy (EBRT) or brachytherapy I125 or palladium (103). The addition of medical castration (see below, primary hormone therapy) to prostatectomy and prostate radiation has had a divergent impact on outcomes. Three months of medical castration before prostatectomy improved pathologic tumor stage and margin status, but there was no improvement in disease-free or overall survival and, therefore, is not routinely recommended (140). In contrast, medical castration in conjunction with EBRT has become standard care for locally advanced or high-grade tumors (141). In a European cooperative group trial, Bolla et al. (141) compared 3 yr of medical castration plus EBRT to ERBT alone. They demonstrated a 37% improvement in disease-free survival and 12% improvement in overall survival at 5 yr for combined hormone/EBRT (141). The lack of benefit for hormone therapy with surgery compared with radiation may result from the divergent duration of hormones used (3 months vs. 3 yr). In fact, the optimal duration of hormone therapy with EBRT is presently unknown and being investigated in the context of clinical trials.
The biology of the interaction of castration combined with EBRT is also under investigation. A retrospective analysis of tumor from patients randomized to EBRT with or without 6 months of hormone therapy in RTOG 8610 suggested that prostate tumors required a wild-type p53 to benefit from combined therapy (142). If p53 status correlates with response in larger prospective analyses, p53 could be a future target for gene therapy in PC.
Adjuvant therapy
Unlike breast cancer, there is no proven benefit for adjuvant hormone therapy in localized (lymph node-negative) PC. A large trial, completed by AstraZeneca, randomized postprostatectomy patients to 2 yr of the antiandrogen bicalutamide (casodex, 150 mg) vs. placebo. This trial was composed of three separate studies with somewhat different entry criteria in the United States, Europe, and Scandinavia. The published results of these trials are not yet available, but early evaluation in the European and Scandinavian trials seems to show reduction in time to progression. Survival cannot yet be assessed. With further follow-up if benefit is confirmed, the prescription of adjuvant antiandrogen therapy may become standard for PC. The national clinical trials cooperative groups have taken a similar but more aggressive approach to adjuvant therapy. Men at high risk for failure after prostatectomy are randomized to 2 yr of CAB therapy with a GnRH agonist and antiandrogen vs. CAB plus six cycles of mitoxantrone chemotherapy. If proven beneficial, this approach will be analogous to the use of adjuvant ER antagonists and chemotherapy in breast cancer.
Primary hormone therapy (Table 2)
Castration, commonly referred to as "hormone therapy," is standard first-line therapy for advanced PC. Castration was initially accomplished by orchiectomy or DES. DES, although very affordable, fell out of favor because of cardiac toxicity and is no longer routinely available in the United States. GnRH analogs are available as agonists and antagonists. The GnRH agonists have been in clinical practice since 1985. The agonists provide a castrate level of T by desensitizing the pituitary to native GnRH stimulation. The availability of 3- to 4-month depot preparations of GnRH analogs has allowed this method of castration to become the preferred treatment for many patients. A 1-yr depot preparation of lupron is also now available. Recently, a GnRH antagonist has completed Phase III testing and is awaiting approval for routine clinical use. Castrate T levels are obtained immediately without potential risk of T flare with GnRH antagonist compared with approximately 3 weeks to obtain castration with GnRH agonists (143).
|
Because of the availability of PSA for following PC, castration therapy, which traditionally was reserved for symptomatic metastatic disease, is being used earlier in the natural history of the disease. A recent study demonstrated benefit for prostatectomy plus castration compared with prostatectomy alone for patients with involved lymph nodes (146). Castration therapy, however, is not curative, and within a variable time period (1848 months) the majority of patients will demonstrate escape from therapy with a rising PSA. The molecular mechanisms responsible for hormone independence, which have been discussed above, are not fully known and are likely heterogeneous.
Before discussing second-line hormone options, castration by monotherapy and CAB will be defined. Monotherapy is the use of orchiectomy or GnRH analog alone, and CAB includes one of these plus an antiandrogen. Antiandrogens function at the cellular level by binding to HSP and preventing activation of AR (147). Theoretically, antiandrogens, when combined with testicular suppression, can block stimulation from adrenal androgens that contribute approximately 510% of total T. A prospective trial comparing GnRH analog alone or with flutamide suggested a survival benefit for CAB, especially for patients with limited metastatic disease (148). A recent well-powered prospective study, however, showed no survival benefit for the addition of an antiandrogen to orchiectomy (149). Survival was no better despite the fact that 74% of the CAB group had a nadir PSA of less than 4 ng/dl compared with 61% of the orchiectomy group (149). Most practitioners, therefore, recommend orchiectomy alone, but some debate continues for the initial use of an antiandrogen with a GnRH analog.
Second-line hormone therapy (Table 3)
Previous dogma held that, unlike breast cancer, once PC escaped from initial castration there were no further effective therapeutic hormonal options. There has been renewed interest in and demonstrated benefit for second-line hormone therapies for PC (150). This interest has been driven by multiple factors, including patient preference for change in therapy when the PSA starts to rise despite no clinical symptoms and patient and physician reluctance to use chemotherapy early. It should be noted, however, that there have been no clinical trials to assess survival benefit for secondary hormonal therapies or to assess the proper sequence to prescribe these treatments.
|
Patients treated with CAB should have the antiandrogen discontinued when the PSA starts to rise. Kelly and Scher (151) were the first to report a series of patients who had a withdrawal response when flutamide was stopped. In their series of 35 patients, 29% showed a greater than 80% decline in PSA, pain improved, and there was documented improvement in measurable disease in two patients. The average duration of response was 5+ months, and the average duration of prior flutamide exposure was 14 months. Additional investigators have documented flutamide withdrawal response rates ranging from 1580% for an overall response rate of 35% (96 of 276) (152). The withdrawal response has also been reported in smaller numbers of patients on bicalutamide, megestrol, DES, cypertone acetate, estramustine, and 13-cis-retinoic acid (152).
It has been theorized that the antiandrogen withdrawal syndrome is the result of point mutations in the hormone-binding domain of the AR (153). These types of AR mutations have been found from clinical PC specimens (91, 96, 102). The mutations allow activation of the receptor by a range of nonphysiologic ligands including flutamide, progesterone, estrogen, and cortisol (90, 93). To date, however, there has not been direct correlation of mutations with antiandrogen withdrawal. In fact, abstracted results from a Cancer and Leukemia group B (CALGB) trial of simultaneous vs. sequential antiandrogen withdrawal and adrenal blockade with AR analysis from bone marrow metastases did not show a correlation of AR mutation with antiandrogen withdrawal (154). It has become standard care when the PSA starts to rise on CAB to discontinue the antiandrogen and observe for withdrawal response. It is recommended based on drug half-life to observe the patient for 1 month after flutamide discontinuation and for 2 months after stopping bicalutamide. This observation period will prevent confusion regarding responses to subsequent therapy.
The addition of a first or second (after discontinuation of first) antiandrogen when the PSA begins to rise is common practice. There have been two Phase II trials evaluating the use of second-line bicalutamide (150 mg daily) (155, 156). In both of these trials response was correlated with prior exposure to flutamide. There was a 3843% response (PSA decline, >50%) for patients previously treated with flutamide compared with 615% response with no prior flutamide. The average duration of response was 5 months, although one patient who started second-line bicalutamide with a PSA of 18 remains in clinical and serologic remission at 60 months (155).
There are anecdotal reports of patients responding to a third antiandrogen although the response rate is likely low (<10%) and of brief duration. The three antiandrogens approved for use in the United Statesflutamide, bicalutamide, and nilutamideall function by binding to AR-associated HSPs and preventing AR from binding to T or DHT (157). Mechanistically, therefore, it is unclear why some tumors will be sequentially inhibited.
At this point, no clear recommendation can be made regarding the proper sequence of and number of antiandrogens to use as second-line therapy. Practice is frequently governed by convenience and reimbursement issues. Future clinical trials will hopefully provide guidance.
Inhibition of adrenal androgen synthesis has reported response rates of between 1944% as second-line therapy in PC (158). Adrenalectomy and hypophysectomy are no longer used to reduce adrenal androgens because of the associated morbidity. There are three options for inducing a medical adrenalectomy: aminoglutethimide, ketoconazole, and hydrocortisone. Aminoglutethimide was initially developed as an anticonvulsant and subsequently found to cause adrenal insufficiency. Ketokonazole is an antifungal agent that was found to cause gynecomastia. Both drugs inhibit adrenal androgen synthesis by blocking adrenal hydroxylases and aromatases (159). Aminoglutethimide (250 mg four times a day) and ketoconazole (400 mg three times a day) are often combined with hydocortisone to prevent addisonian symptoms. Studies have suggested that the combination of aminoglutethimide and hydrocortisone was no more effective than hydrocortisone alone, which is a potent inhibitor of adrenal androgen production (160).
Two Phase II trials have reported a response rate of 4455% for the combination of antiandrogen withdrawal and simultaneous adrenal blockade (161, 162). As mentioned above, the CALGB cooperative clinical trials group has recently completed a randomized trial evaluating combination vs. sequential maneuvers. Results to be presented at the 2001 American Society of Oncology meeting reveal no significant improvement or survival for combining antiandrogen withdrawal with simultaneous adrenal blockade. Insight into the effectiveness of sequential therapy can be obtained from another trial recently competed at the University of Massachusetts Medical Center (Worcester, MA) and Beth Israel Deaconess hospitals (Boston, MA). This Phase II trial demonstrated no response to adrenal blockade for patients progressing on second-line bicalutamide (150 mg) (Bubley, G. J., S. P. Balk, and M. E. Taplin, submitted for publication). This small trial suggests that when patients had failed primary hormone therapy and the AR remained blocked with bicalutamide, further reduction in adrenal androgens was of no benefit.
DES has been evaluated as second-line hormone therapy in PC. One reported mechanism of action for DES in patients treated with orchiectomy or GnRH agonist is suppression of adrenal androgens (163), and another is induction of tumor cell apoptosis (164). In a Phase II trial of DES (1 mg) in 21 patients there was a 43% PSA (50% decline) overall response rate (165). Patients with only one previous hormone therapy responded 62% compared with 13% of patients who had two or more hormone treatments. No anticoagulation was used, and there was one (5%) deep venous thrombosis. In a Canadian trial of DES (1, 2, or 3 mg) combined with coumadin (1 mg) there was an unacceptable toxicity with 31% deep venous thrombosis, 7% myocardial infarction, 7% transient ischemic attacks, and one patient with angina or congestive heart failure (166). In clinical practice the use of DES is limited by its potential toxicity and limited availability outside clinical trials.
Another form of estrogen is the herbal preparation PC-SPEC. PC-SPEC contains eight herbs and has estrogenic activity (167). PC-SPEC has demonstrated activity in hormone-sensitive PC, reducing T and PSA (167). PC-SPEC-induced castration has the same morbidity as traditional castrating agents but, because of its herbal nature, has gained popularity among patients. The effectiveness of PC-SPEC in the second-line setting is presently under investigation. Early results of a Phase II trial of PC-SPEC in 23 patients demonstrated a PSA response (>50% decline) of 52% with a median duration of response of 2.2 months (range, 17) (168). Practical issues pertaining to this compound include lack of standardization of active compounds in PC-SPEC and the cost because it is not covered by insurance plans.
Hormonal agents that have shown minimal effectiveness as second-line therapy include MA and tamoxifen. A CALGB trial randomized 149 patients to either 160 mg or 640 mg MA (169). There was no survival benefit for the high-dose arm, and only 14% of all men had a greater than 50% drop in PSA. Similarly, in a Phase II study of tamoxifen in 30 patients there was only one partial response (170). These agents, therefore, cannot be recommended enthusiastically for consideration as second-line therapy.
Many patients with PC will have disease responsive to second-line hormone therapies. The clinician should first document a castrate level of T and direct therapy to reduce T if it is not in the castrate range. A sequential approach to the second-line options discussed above can then be initiated. Responding patients may benefit for months to years from second-line hormone therapies. The duration of response will likely shorten with each subsequent hormone maneuver. This approach is rational because at this time, although chemotherapy has high response rates, there has been no demonstrated benefit for early chemotherapy. More effective hormonal and targeted biologic approaches will likely develop as our understanding of the molecular and cellular mechanisms of PC evolves.
Summary
Mechanisms responsible for prostate carcinogenesis and the development of HRPC are diverse. As our understanding of hormone-controlled growth factors and cell cycle regulators expands, targeted therapeutics will move beyond standard castration and improve the length and quality of life for PC patients.
Footnotes
This work was supported by NIH Grants CA15776 and CA62269 and Department of Defense Grant DAMD17-98-1-8606 (to S.-M.H.).
Abbreviations: aa, Amino acid; AI, androgen independence; AR,
androgen receptor; BPH, benign prostatic hyperplasia; CAB, combined
androgen blockade; CALGB, Cancer and Leukemia group B; DES,
diethylstilbestrol; DHEA, dehydroepiandrosterone; DHT,
5-dihydrotesterone; E2, estradiol; EBRT, external beam radiation
therapy; EGF, epidermal growth factor; ER, estrogen receptor; HRPC,
hormone-refractory prostate cancer; HSP, heat shock protein; IGFBP,
IGF-binding protein; KGF, keratinocyte growth factor; LHRH,
LH-releasing hormone; MA, megestrol acetate; PC, prostate cancer; PSA,
prostate-specific antigen; SHBG, sex hormone-binding globulin; T,
testosterone; TAF, transcriptional activation function.
Received January 4, 2001.
Accepted April 29, 2001.
References