Intermittent androgen suppression for prostate cancer - Nature

Prostate Cancer and Prostatic Diseases (1998) 1, 289±296 ß 1998 Stockton Press All rights reserved 1365±7852/98 $12.00

http://www.stockton-press.co.uk/pcan

Review

Intermittent androgen suppression for prostate cancer: rationale and clinical experience M Gleave, SL Goldenberg, N Bruchovsky and P Rennie Division of Urology, Vancouver General Hospital, University of British Columbia, Vancouver, Canada

Research on hormonal treatments of prostate cancer over the past 20 y have focused on maximizing androgen ablation through combination therapy. Maximal androgen ablation increases treatment-related side-effects and expense and fails to prolong time to androgen-independent (AI) progression. The rationale behind intermittent androgen suppression (IAS) is based on: (1) observations that androgen ablation is palliative, not curative in most patients where quality of life must be considered; (2) assumption that immediate androgen ablation is superior to delayed therapy in improving survival of patients with prostate cancer; and (3) the hypothesis that if tumor cells which survive androgen withdrawal are forced into a normal pathway of differentiation by androgen replacement, then apoptotic potential might be restored and progression to androgen independence may be delayed. Observations from animal model studies suggest that progression to androgen-independence involves adaptive responses to androgen deprivation which are, in turn, modulated by intermittent androgen replacement. Supported by these animal model observations, several centers have now tested the feasibility of IAS therapy in non-randomized groups of patients with prostate cancer using serum PSA as trigger points. Experimental and clinical data suggest that prostate cancer is amenable to control by IAS. IAS may offer clinicians an opportunity to improve quality of life in patients with prostate cancer by balancing the bene®ts of immediate androgen ablation (delayed progression and prolonged survival) while reducing treatment-related side effects and expense. Whether time to progression and survival is affected in a bene®cial or adverse way is being studied in randomized, prospective protocols. The purpose of this article is to review the rationale behind IAS, compare observations from published series, and discuss potential indications and treatment strategies using IAS.

Keywords: prostate cancer; intermittent therapy; hormone therapy; prostate speci®c antigen; androgen independent progression; intermittent androgen suppression

Rationale for intermittent androgen suppression Approximately 80% of prostate cancer patients achieve symptomatic and objective responses following androgen suppression and serum prostate speci®c antigen (PSA) Correspondence: Dr M Gleave, D-9, 2733 Heather Street, Vancouver General Hospital, Vancouver, B.C. V5Z 3J5, Canada. Received 2 February 1998; revised 23 July 1998; accepted 23 July 1998

levels decrease in almost all patients. Surgical or medical castration results in a median progression-free survival of 12 ± 33 months and a median overall survival of 23 ± 37 months in patients with stage D2 disease.1,2 However, for reasons that remain unde®ned, the apoptotic process induced by androgen ablation fails to eliminate the entire malignant cell population.3 Another limitation of conventional androgen ablation is that it accelerates the rate of progression of prostate cancer to an androgenindependent state,3 and after a variable period of time averaging 24 months, progression inevitably occurs with

Intermittent androgen suppression for prostate cancer M Gleave et al

290

increasing serum PSA levels and androgen-independent growth. Over the past two decades, most efforts have focused on maximizing the degree of androgen suppression therapy by combining agents that inhibit or block both testicular and adrenal androgens. Unfortunately, large randomized studies and meta-analyses have concluded that maximal androgen ablation increases treatment-related side-effects and expense and has not been shown to prolong time to androgen-independent progression in most patients.2,4,5 The idea of inducing more than one regression of androgen-dependent malignancy arose from the observation that the involution of prostate brought on by castration is an active process involving the rapid elimination of a large number of epithelial cells.6,7 It was ®rst postulated that the replacement of androgens even in small amounts would have a conditioning effect on surviving cells allowing them to conserve or regain desirable traits of differentiation.7 In the case of tumors, several lines of evidence obtained by Foulds8 and Noble9 implied that long periods of hormone deprivation simply accelerated progression towards autonomous growth and thus were better avoided. Although no practical method of countering progression by hormonal means emerged from this work, the subsequent demonstration that consecutive episodes of testosterone-induced regeneration of involuted prostate completely restored the susceptibility of the epithelium to further androgen withdrawal,10 was an incentive to try similar experiments on tumors. The potential bene®ts of intermittent, rather than continuous therapy, include improved quality of life through reduced therapy-induced side effects, decreased costs, and possibly improved control over progression to androgen-independence. However, intermittent androgen suppression (IAS) is a feasible treatment option only if immediate hormone therapy is superior to delayed, and if adaptive mechanisms help mediate androgen-independent progression.

Balancing the bene®ts of immediate androgen ablation If time to disease progression and overall survival is similar with immediate (that is at the time of diagnosis of metastatic or recurrent disease) or delayed (namely waiting for the development of symptoms) androgen ablation, then the role and rationale for IAS would be limited. Measuring serum PSA after radical prostatectomy or radiotherapy would be a waste of time and resources, and only patients with clinical symptomatic recurrences would require hormonal therapy. Accumulating evidence, however, suggests that treatment should commence at the time of diagnosis of locally advanced or metastatic disease (Table 1). Although the ®rst VACURG study11 found delayed endocrine therapy equivalent to immediate endocrine therapy, re-analysis by Byar12 and evidence from animal models of prostate cancer indicate that early androgen ablation is more effective.3,13 The Goldie ± Coldman14 hypothesis of increasing somatic genomic alterations and tumor heterogeneity over time provides additional theoretical evidence supporting initiation of therapy as early as possible. Most notably, the MRC study from the UK randomised over 800 men with advanced M0 or M1 disease to immediate vs deferred

Table 1 Rationale for immediate androgen ablation therapy in patients with advanced prostate cancer 1. Biological progression linked to cumulative number of cell divisions and accumulation of mutational events14 2. Dunning R3327 studies suggested improved survival if animals castrated when tumours were low volume13 3. Review of VACURG I: decreased Ca-speci®c death rates12 4. EORTC study: improved survival with adjuvant therapy postradiotherapy16 5. MRC study: Delayed time to progression and improved survival15 6. Signi®cant psychological impact on quality of life: delays development of mets and lowers serum PSA15,16

therapy and showed that palliative surgery for bladder outlet or ureteral obstruction was reduced by close to 50% and disease-speci®c survival improved with immediate therapy.15 Another, more recent study from the EORTC also demonstrated signi®cant improvement in time to disease progression, and 5 y survival (58 vs 78%) in patients with locally advanced disease treated with radiotherapy and either immediate androgen ablation for 3 y or delayed therapy until symptomatic progression.16 Coincident with accumulating evidence supporting immediate therapy is a shift in the target population receiving androgen ablation from metastatic disease to PSA failures after radical prostatectomy or radiotherapy. Use of PSA for early detection has produced a stage migration and 50% decrease in the incidence of stage D2 disease over the last two years.17 Because the median survival in D2 disease is only 2 ± 3 y, the long term sideeffects of androgen ablation do not become apparent and are therefore not clinically relevant. However, use of PSA and stage migration have resulted in earlier diagnosis at younger ages and earlier stages of disease, with the prospect of much longer therapy and risk of chronic complications. Furthermore, use of PSA to detect biochemical recurrences following radical prostatectomy or radiotherapy identi®es men destined to recur who may bene®t from early adjuvant therapy13 ± 16 but who have life expectancies exceeding 10 y. Hence, combinations of stage migration, earlier diagnosis of PSA recurrences, longer life expectancies, and trend towards immediate therapy will force clinicians to balance the potential bene®ts of early adjuvant therapy with risks of development of metabolic complications, as well as the increased expense, associated with long-term continuous androgen withdrawal therapy. These metabolic complications include loss of bone mass (osteoporosis) and fractures, loss of muscle mass and anemia with easy fatigue and decreased energy levels, changes in lipid pro®le with increased risk of cardiovascular complications, glucose intolerance, and depressive and/or irritable personality changes.18 IAS offers clinicians an opportunity to improve quality of life in patients with prostate cancer by balancing the bene®ts of immediate androgen ablation (delayed progression and prolonged survival) while reducing treatment-related side effects and expense.

Modulating adaptive responses mediating androgen independence Prostatic carcinoma is an androgen-dependent tumor, manifested by apoptosis, cell cycle arrest, and changes


in androgen-regulated gene expression after androgen ablation.19,20 Unfortunately, progression to androgenindependence nearly always occurs in prostate cancer cells after androgen ablation. Cellular and molecular mechanisms that help mediate progression to androgenindependence include expansion of pre-existing clones of androgen-independent cells (clonal selection),21,22 up-regulation of androgen-repressed adaptive mechanisms capable of aborting the apoptotic process (adaptation),3,23 androgen-receptor mutations leading to a constitutively active transitional factor,24 or production of factors capable of interacting with the androgen-receptor (protein± protein interactions).25 A complete review of these mechanisms is beyond the scope of this review, but likely all variably operate in heterogenous tumors like prostate cancer. Clonal expansion of AI cells after castration cannot be prevented by current therapeutic strategies because of resistance to traditional cytotoxic chemotherapies. However, it may be possible to inhibit or modulate the adaptive changes in gene expression precipitated by androgen ablation through the differentiating in¯uence of androgen. The ability of benign or malignant prostate cells to undergo apoptosis and express PSA are acquired as a feature of differentiation of prostatic cells under the in¯uence of androgens. Even in malignancy, the ability to undergo apoptosis is acquired as a feature of differentiation under the in¯uence of androgens. Therefore in the absence of androgens, it is impossible for dividing cells to differentiate and become pre-apoptotic again, which explains why recurrent tumor growth after castration is characterized by androgen independence.3,19 The rationale behind IAS is based on the hypothesis that if tumor cells surviving androgen withdrawal are forced along a normal pathway of differentiation by androgen replacement, then apoptotic potential might be restored and progression to androgen independence delayed. It follows that if androgens are replaced soon after regression of tumor, it should be possible to bring about repeated cycles of androgen-stimulated growth, differentiation and androgen-withdrawal regression of tumor.

tumors treated with continuous or intermittent DES had tumor volumes at death smaller than control or castrate rats, but a survival advantage was achieved only in the castrate or continuous DES groups. Although the intermittent regimen was successful in checking tumor volume, it did not yield a survival advantage and in this respect was inferior to castration. The foregoing experiments on the Dunning R3327 prostatic adenocarcinoma draw attention to importance of using cyclic androgen deprivation only for the treatment of androgendependent cancer. Dunning tumors are androgen-sensitive, not androgen-dependent, and apoptosis with tumor regression does not occur following castration. The failure to improve outcome with IAS in Dunning tumors is not an unexpected result. Indeed, IAS may be detrimental because of stimulation of androgen-sensitive cells when testosterone levels increase. Androgen ablation with the addition of cytotoxic agents appears to be the most effective therapy in this model.22

Shionogi mouse carcinoma IAS of the androgen-dependent Shionogi carcinoma has been carried out experimentally by transplanting the tumor into a succession of male mice. The cycle of transplantation and castration-induced apoptosis was successfully repeated four times before growth became androgen-independent during the ®fth cycle.28 The mean time to androgen independence increased 3-fold compared to one-time castration, consistent with a retarding effect of cyclic therapy on tumor progressions. Some caution is required in the interpretation of these results since the procedure of transplanting a regressing tumor from a castrated to a non-castrated animal temporarily may reduce the rate of cell division and give rise to an apparent delay in progression. However, in addition to tumor growth, other markers of androgen dependence in this model, castration-induced apoptosis and TRPM-2 gene expression, are maintained three times longer with IAS compared to continuous therapy, paralleling and supporting the changes in tumor volume.

LNCaP prostate tumor model

Preclinical studies of intermittent androgen suppression Dunning R3327 rat prostatic adenocarcinoma Using the Dunning rat carcinoma as a model of progression to androgen-independence, Isaacs et al 21,22 emphasized the importance of genetic instability, which increases tumor heterogeneity by increasing the number of distinctly different clones of cells, analogous to the Goldie ± Coldman hypothesis of tumor progression.14 Androgen-independent progression results from selective outgrowth of one or more androgen resistant clones after androgen ablation. Treatment of the Dunning R3327 carcinoma by cyclic therapy was studied by Trachtenberg et al 26 but no signi®cant growth reduction with the IAS group was observed and they concluded that IAS was inferior to early castration in inhibiting tumor growth. Similarly, Russo et al 27 reported that Dunning R3327

As in human prostate cancer, serum PSA levels in this model are androgen-regulated and proportional to tumor volume.29 ± 32 After castration, serum and tumor-cell PSA levels decrease by 80% and remain suppressed for 3 ± 4 weeks before increasing again. With intermittent testosterone withdrawal and replacement in castrated mice bearing LNCaP tumors, androgen-independent regulation of the PSA gene was delayed 3-fold.33 Thalmann et al 34 reported evidence which suggests that continuous androgen suppression may facilitate development of androgen-independent osseous metastasis. Characterization of androgen-independent-LNCaP cell lines produced a subline (C4 ± 2) that metastasized to bone, and interestingly, the incidence of osseous metastases appeared higher in castrated than in intact male hosts. Finally, two separate groups of investigators35,36 independently reported isolating androgen-repressed LNCaP cell lines that grew faster in castrated than in intact hosts, and with growth rates that were actually inhibited by androgens.

291


292

Taken together, these data emphasize the potentially diverse biologic effects that androgens have on a heterogenous and adapting tumor population and supports the concept that progression to androgen-independence involves loss of androgen-induced differentiation and/ or upregulation of androgen-repressed growth regulatory pathways.

Clinical studies of intermittent androgen suppression Reports of multiple regressions of hormone dependent malignancy are rare in the clinical and scienti®c literature. Stoll37 drew attention to the balance between inhibition and stimulation of the growth of breast cancer in an elderly woman who was treated intermittently with estrogen. Regression of tumor occurred three times over a period of ®ve years, after which an estrogen withdrawal response resulted in a forth remission. Bruchovsky described sequential responses of a pulmonary metastases of breast cancer in a female patient whose endocrine therapy consisted of radiation ovariectomy, administration of androgen, discontinuation of androgen and adrenalectomy respectively.19 Regressions followed each procedure and the tumor size was successfully held in check. The intermittent regulation of serum testosterone levels for therapeutic purposes in prostate cancer was ®rst attempted with cyclic administration of estrogenic hormone.38 Nineteen patients with advanced prostate cancer received diethylstilbestrol until a clinical response was clearly demonstrated and then it was withheld until symptoms recurred. One additional patient was treated with ¯utamide using a similar schedule. The mean duration of initial therapy was 30 months (range 2±70 months). Subjective improvement was noted in all patients during the ®rst three months of treatment. When therapy was stopped, 12 of the 20 patients relapsed after a mean interval of eight months (range 1±24 months) and all subsequently responded to re-administration of drug. Therapy-induced impotency was reversed in 9 of the 10 men within three months of the break in treatment. An improved quality of life was achieved owing to the reduced intake of diethylstilbestrol, and no adverse effects on survival were apparent. The ®rst nonsteroidal antiandrogens became available around the same time as this study, and most clinical research focused on the combined use of LHRH agonists and antiandrogens.2,4 Little attention has been given to the reversibility of action of LHRH agonists, the signi®cance of which is far-reaching. The potential for a full recovery from therapy makes it possible to alternate a patient between periods of treatment and no-treatment. Furthermore, serial serum PSA measurements, which were not available at the time of the study by Klotz et al,38 permit accurate monitoring of disease activity and serve as trigger points for stopping and restarting therapy. In response to this androgenic stimulus, atrophic cells are recruited into a normal pathway of differentiation where the risk of progression to androgen- independence is reduced. With the associated movement through

the division cycle, the cells become pre-apoptotic again making it possible to repeat therapy. Preliminary results of IAS in 47 patients using reversible medical castration and serum PSA as trigger points were reported by Goldenberg et al 39 in 1995, and updated to include 60 patients by Gleave et al 40 in 1997. Seventy patients with a minimum follow-up of 12 months (mean 46 mo.) have now been accrued, 41 with clinically localized and 29 with metastatic disease. Mean initial serum PSA was 110 mg/L. Treatment was initiated with combined androgen blockade and continued for an average of nine months. Because prognosis is poor in patients who do not achieve normal PSA levels after androgen ablation, only patients with PSA nadir levels below 4 mg/L are eligible for the IAS protocol. Medication was then withheld until serum PSA increased to mean values between 10±20 mg/L. This cycle of treatment and no-treatment was repeated until the regulation of PSA became androgen independent (Figure 1). Thirty-two men are currently in their second cycle and 24 are in their third or higher cycle. The mean time to PSA nadir during the ®rst three cycles was ®ve months. The ®rst two cycles averaged 18 months in length with 45% of the time off therapy, while the third cycle averaged 15.5 months. Serum testosterone returned to the normal range within a mean of eight weeks of stopping treatment. The off-treatment period in all cycles was associated with an improvement in sense of well-being, and the recovery of libido and potency in the men who reported normal or near-normal sexual function before the start of therapy. Androgen-independent progression occurred in 10 out of 29 stage D patients and 7 out of 41 with localized disease after a mean followup of 43 months. Of 21 stage D2 patients, 12 remain on study with stable disease. Nine patients have died, three from non-cancer related illness, with median time to progression and overall survival of 26 and 42 months, respectively. Since these initial reports, other investigators have con®rmed the feasibility of IAS in patients with advanced or recurrent prostate cancer (Table 2). Higano et al 41 reported their experience in Seattle using IAS with a cohort of 22 patients (7 stage D2, 3 stage D1, 2 stage T23, and 10 biochemical failures after de®nitive local therapy). The median follow-up was 26 months, and all patients remain alive with two patients (1 biochemical relapse and 1 D1) progressing to androgen-independence at 33 and 22 months, respectively. Time off therapy for stage D2 patients was 35%, slightly less than the percentage of time off for biochemical relapsers (44%). All patients who were cycled off therapy for three or more months noted some improvement in overall quality of life. Oliver et al 42 in London reviewed 20 patients with localized or metastatic prostate cancer who elected to stop androgen ablation therapy after median of 12 months (range 3±48). After a median of nine months off therapy, 13 patients progressed and were restarted on androgen ablation, and 10 of these remain progression-free after two years. Of particular interest were seven patients who were on therapy for 3, 3, 3, 16, 12, 5, and 12 months and subsequently off hormone therapy for 33, 43, 36, 36, 36, 40, and 42 months, respectively. These authors commented on trends in their small series for better duration off therapy in older patients and in those with nonmetastatic disease.


293

Figure 1 Androgen ablation results in apoptosis and upregulation of previously androgen-repressed genes (for example Bcl-2, TRPM-2). However, not all cells are eliminated and in the absence of androgens, proliferating cells do not differentiate to become preapoptotic again, which results in development of the androgen-independent phenotype. With intermittent re-exposure to testosterone (T ), proliferating cells differentiate and become preapoptotic again, permitting another round of androgen ablation-induced apoptosis and tumor regression (T 7 ). Inserts are northern analyses illustrate constitutive overexpression of PSA gene expression, a marker of androgen-independence in LNCaP tumors, is delayed by intermittent compared to continuous androgen suppression in LNCaP tumours.33

Carrol et al 43 in San Francisco reported on their initial experience with 47 patients with clinically localized prostate cancer treated with an IAS protocol. Patients were followed for a mean of 24 months and completed on average two treatment cycles. Only one patient failed to respond to reinstitution of treatment, and no other patients demonstrated progression during the followup period, illustrating the ability of reversible medical castration to induce repeated cycles of therapeutic response. Average cycle length was 16 months for the ®rst cycle and 11 months for the second. Patients averaged 50% of their time off therapy. A prospective phase II study in post-radiotherapy recurrences has been initiated at 4 centers across Canada and over 100 patients have now been accrued.44 Patients are treated for nine months and cycled off therapy until serum PSA rises back to levels between 8 and 15 mg/L (Figure 2). Endpoints of the study include monitoring length of time off therapy, quality of life measurements, and time to androgen-independent progression. First cycle analysis demonstrated improve quality of life measurements when men were cycled off therapy. Observations from these series add to and are consistent with the initial reports from Vancouver39,40 in terms of response rates and time off therapy. Each of these small

series illustrate that relatively short on-therapy periods may result in off-therapy intervals of several years, which suggests that continuous androgen ablation may represent over-treatment in a signi®cant proportion of men. These IAS series use serum PSA as the marker of response to therapy and tumor progression, as well as a guide for when to restart therapy. Serum PSA increases before evidence of tumor progression becomes apparent in more than 90% of patients with progressive disease treated with continuous therapy.1,47 As a result, bone scans for monitoring disease progression are used only in patients with new symptoms or biochemical progression. Dissociation between biochemical and tumor progression occasionally occurs, however, and stopping therapy with IAS in this scenario would likely accelerate tumor progression. Although unusual, dissociation between biochemical and clinical progression emphasizes the need for Phase III controlled data to better characterize the risks and bene®ts of IAS.

Candidates for intermittent androgen suppression Failure of serum PSA to nadir below 4 mg/L after six months of therapy is associated with a short duration of


294

Figure 2 Composite results on 70 patients with prostate cancer treated with intermittent androgen suppression, of whom 14 are in their ®rst cycle of treatment, 32 in their second cycle, and 13 in their third, 5 in their fourth cycle, and 6 in their ®fth cycle. Mean serum PSA at the start of the ®rst, second, third and fourth cycles was 110 mg/L, 16 mg/L, 12, and 14 mg/L, respectively. The ®rst three cycles averaged 18, 18, and 15.5 months in length, respectively.

response to androgen ablation and poor prognosis,47 and IAS is not offered to these patients. Men with TxN1±3M0 or TxNxM1 who are sexually active, compliant to close followup, or who do not tolerate the side effects of androgen ablation can be considered for IAS as long as they realize it is investigational. Biochemical failures after radiotherapy or surgery are groups most likely to bene®t from IAS, which permits ¯exibility in applying and achieving survival bene®ts of immediate therapy,12±16 while balancing potential longterm side effects and costs of continuous therapy.18

Length of treatment cycle Conceptually, androgen ablation should be continued until maximal castration-induced apoptosis and tumor regression have been induced, but stopped before constitutive development of the androgen-independent phenotype. Premature termination of therapy would permit some cancer cells destined to undergo apoptosis to survive upon reexposure to testosterone. Some investigators43 treated for 1±2 months past PSA nadir, which results in

Figure 3 Schema of intermittent androgen suppression therapy for patients with biochemical recurrences following radiation therapy for clinically localized prostate cancer.

Table 2 Clinical studies of intermittent androgen suppression

Investigator Goldenberg39 Oliver41 Carrol43 Higano42 Gleave40 a

Sample size

Clinical stage

Mean follow up

Cycle 1 start PSA (mg/L)

Cycle 2 start PSA (# of pts)

Cycle 3 start PSA

Mean cycle lengtha (mo.) cycle 1, 2, 3

Mean time off therapy

Mean time to PSA nadir

47 20 47 22 70

local met local met local local met local met

30 mo. n/a 24 mo. 26 mo. 46 mo.

128 n/a 21 25 110

15 (n 30) n/a 7.0 (n 17) 7.7 (n 12) 16 (n 56)

18 (n 15) n/a 20.5 (n 3) n/a (n 0) 12 (n 24)

18, 18, 16 n/a 16, 11, n/ab 15, 18 18, 18, 16

47% n/a 50% 38% 47%

5 mo. n/a 4 mo. 3.5 mo. 5 mo.

Length of time spent on drug therapy and off therapy during each cycle of IAS. Treatment was continued for 1±2 months after nadir PSA reached.

b


shorter òn' cycles and shorter òff' cycles, but similar percentages of time spent òff' therapy. The de®nition of nadir must be consistent, however, and should be the lowest, plateau serum PSA level attained following institution of androgen ablation. For comparison of future series, standardization of treatment regimens are necessary. We have used a nine month treatment òn' cycle for the following reasons. Firstly, PSA nadir and maximal soft tissue regression is not reached until 8 or 9 months in many patients.39,40 Similarly, the duration of time necessary for PSA to reach its nadir in a group of patients with clinically con®ned disease after institution of neoadjuvant hormone therapy is often longer than six months.45,46 Using a lower limit of sensitivity of 0.2 mg/L, PSA nadir was reached in only 34% after three months, 60% after ®ve months, 70% after six months, and 84% at eight months. The initial rapid decrease in PSA results from cessation of androgen-regulated PSA synthesis and apoptosis, while the ongoing slower decline re¯ects decreasing tumor volume. Secondly, immunostaining for proliferation markers (Ki-67 and PCNA) showed minimal or reduced activity in radical prostatectomy specimens after eight months of neoadjuvant therapy compared to pretreatment biopsy specimens.45,48 Thirdly, levels of the antiapoptotic oncoprotein, Bcl-2, increase beginning during the ®rst three months of NHT,49 and remain elevated after eight months of therapy,48 consistent with its role in prevention of castration-induced apoptosis.50 These observations suggest that Bcl-2 is upregulated early after castration and that concerns regarding constitutively increased levels developing between six and nine months of therapy must be balanced by bene®ts of ongoing tumor involution during this period of time. Interestingly, preliminary experiments using RT-PCR revealed that androgen downregulates Bcl-2 expression in LNCaP cells in vitro,51 an observation which may be signi®cant when considering mechanisms by which IAS may delay time to AI progression. Based on time to PSA nadir, reduced prolifation marker staining, and early upregulation of Bcl-2, we treat patients for nine months prior to stopping therapy.

When to restart Trigger points for reinstitution of androgen withdrawal therapy are similar among investigators reporting on their IAS experience.39±44 Although the optimal time remains unde®ned and empirical, time off therapy should be long enough to permit normalization of improved quality of life and testosterone-induced tumor cell differentiation. Re-exposure of tumor cells to testosterone during the off-treatment cycles of IAS is critical to the underlying hypothesis and rationale behind IAS. Furthermore, recovery of testosterone between treatment cycles is necessary for recovery of sexual function, reduced side effects, and normal sense of well-being. Until more information is obtained, PSA trigger points are regarded as tentative settings only. Trigger points are individualized and factors that are considered include pretreatment PSA levels, stage, PSA velocity, presence of symptoms, and tolerance of androgen ablation therapy. In general, in patients with metastatic disease and high pretreatment PSA levels, therapy is restarted when PSA

increases to 20 mg/L; in patients with locally recurrent disease and moderately elevated pretreatment PSA levels, therapy is restarted when PSA reaches 6 ±15 mg/L, and earlier for post-radical prostatectomy recurrences.

Future directions Observations from accumulating experience in Phase II studies suggest that IAS does not have a negative impact on time to progression or survival, both of which appear similar to continuous combined therapy. However, the available information about IAS remains limited and leaves several questions unanswered. The issue of survival is the most important and phase III randomized studies are required to accurately assess the effects of IAS on this critical parameter. IAS appears to improve quality of life by permitting recovery of libido and potency, increasing energy levels and enhancing sense of well-being during off-treatment periods. The animal and preliminary clinical studies have helped identify groups of patients who are most likely to bene®t from IAS, to determine optimal duration of treatment, and to suggest trigger points on when to re-start therapy again. The therapeutic strategy and trigger points in each situation are guided by serum PSA. It is important to look for other signs of progressive disease when serum PSA is used to monitor disease activity, in case of the unusual scenario of PSA-negative disease progression. The possibility of expanding the use of IAS to earlier stage disease should be examined in greater detail. Conceivably, IAS might become an alternative to radical prostatectomy or irradiation for the primary treatment of localized prostate cancer in some men with life expectancies of less than 15 y. The prospect of improving the results of IAS by increasing the length and number of cycles is attractive. Prolongation of the length of each òff' treatment cycle may be possible with 5a-reductase inhibitors like ®nasteride. Augmentation of intermittent therapy might be accomplished by the administration of apoptotic-enhancing agents (for example antisense Bcl-2 strategies), differentiating agents, or gene therapies at speci®c times during a cycle of treatment when the modality-of-choice would have its maximum effect.

References 1 Bruchovsky N. Androgens and antiandrogens. In: Holland JF et al Cancer Medicine 3rd Ed Lea & Febiger: Philadelphia, 1993, pp 884 ± 896. 2 Denis L, Murphy GP. Overview of phase III trials on combined androgen treatment in patients with metastatic prostate cancer. Cancer 1993; 72 (Suppl): 3888 ± 3895. 3 Bruchovsky N et al. Effects of androgen withdrawal on the stem cell composition of the Shionogi Carcinoma. Cancer Res 1990; 50: 2275 ± 2282. 4 Crawford ED et al. Comparison of bilateral orchiectomy with or without ¯utamide for the treatment of patients with stage D2 adenocarcinoma of the prostate: Results of NCI intergroup study 0105 (SWOG and ECOG). Br J Urol 1997; 80: 278 (Abstract). 5 Prostate Cancer Trialists' Collaborative Group. Maximum androgen blockade in advanced prostate cancer: an overview of 22 randomised trials with 3283 deaths in 5710 patients. Lancet 1995; 346: 265 ± 269.

295


296

6 Bruchovsky N et al. Hormonal effects on cell proliferation in rat prostate. Vitamin Horm 1975; 33: 61 ± 102. 7 Bruchovsky N, Rennie PS, Van Doorn E. Pathological growth of androgen-sensitive tissues resulting from latent actions of steroid hormones. J Toxicol Environ Health 1978; 4: 391 ± 408. 8 Foulds L. Neoplastic Development 1, Academic Press: New York, 1969. 9 Noble RL. Hormonal control of growth and progression in tumors of Nb rats and a theory of action. Cancer Res 1977; 37: 82 ± 94. 10 Sanford NL, Searle JW, Kerr JFR. Successive waves of apoptosis in the rat prostate after repeated withdrawal of testosterone stimulation. Pathology 1984; 16: 406 ± 410. 11 Mellinger GT, Veterans Administration Cooperative Urological Research Group. Carcinoma of the prostate: a continuing cooperative study. J Urol 1964; 91: 590 ± 594. 12 Byar DP. Proceedings: The Veterans Administration Cooperative Urological Research Group's studies of cancer of the prostate. Cancer 1973; 32: 1126 ± 1130. 13 Isaacs JT. The timing of androgen ablation therapy and/or chemotherapy in the treatment of prostatic cancer. Prostate 1984; 5: 1 ± 17. 14 Goldie JH, Coldman AJ. A mathematical model for relating the drug sensivitity of tumors to their spontaneous mutation rate. Cancer Treat Rep 1979; 63: 1727. 15 Immediate vs deferred treatment for advanced prostate cancer: Initial results of the MRC trial. The Medical Research Council Prostate Cancer Working Party Investigators Group. Br J Urol 1997; 79: 235 ± 246. 16 Bolla M et al. Improved survival in patients with locally advanced prostate cancer treated with radiotherapy and goserilin. NEJM 1997; 337: 295 ± 300. 17 Stanford JL, Blumenstein BA, Brawer MK. Temporal trends in prostate cancer rates in the Paci®c Northwest. J Urol 1996; 155: 604A. 18 Townsend MF, Sanders WH, Northway RO, Graham SD Jr. Bone fractures associated with luteinizing hormone-releasing hormone agonists used in the treatment of prostatic carcinoma. Cancer 1997; 79: 545 ± 550. 19 Bruchovsky N et al. The endocrinology and treatment of prostate tumor progression. In: Coffey DS et al. (eds). Current Concepts and Approaches to the Study of Prostate Cancer. Progress in Clinical and Biological Research Alan R. Liss Inc.: New York, 1987, pp 348 ± 387. 20 Tenniswood M. Apoptosis, tumor invasion and prostate cancer. Br J Urol 1997; 79: 27 ± 34. 21 Isaacs JT, Wake N, Coffey DS, Sandberg AA. Genetic instability coupled to clonal selection as a mechanism for progression in prostatic cancer. Cancer Res 1982; 42: 2353 ± 2361. 22 Issacs JT. The timing of androgen ablation therapy and/or chemotherapy in the treatment of prostatic cancer. Prostate 1984; 51: 1 ± 17. 23 van Weerden WM. Animal models in the study of progression of prostate and breast cancers to endocrine independence. In: Berns JC et al. Mechanisms of Progression to Hormone-Independent Growth of Breast and Prostate Cancer Parthenon Publishers: New Jersey, 1991 pp 55 ± 70. 24 Taplin ME et al. Mutation of the androgen receptor in metastatic androgen-independent prostate cancer. N Engl J Med 1995; 332: 1393 ± 1398. 25 Sato N et al. Androgenic Induction of Prostate Speci®c Gene is repressed by protein-protein interaction between the androgen receptor and AP-1/c-jun in the human prostate cancer cell line LNCaP. J Biol Chem 1997; 272: 17485 ± 17494. 26 Trachtenberg J. Experimental treatment of prostatic cancer by intermittent hormonal therapy. J Urol 1987; 137: 785 ± 788. 27 Russo P et al. Effects of intermittent diethylstilbestrol diphosphate administration on the R3327 rat prostatic carcinoma. Cancer Res 1987; 47: 5967 ± 5970. 28 Akakura K et al. Effects of intermittent androgen suppression on androgen-dependent tumors: apoptosis and serum prostate speci®c antigen. Cancer 1993; 71: 2782 ± 2790.

29 Gleave ME et al. Serum PSA levels in mice bearing human prostate LNCaP tumors are determined by tumor volume and endocrine and growth factors. Cancer Res 1992; 52: 1598. 30 Gleave ME et al. Predictors of time to androgen-independent progression in the LNCaP prostate tumor model. J Urol 1994; 151: 241 (Abstract). 31 Gleave ME et al. Acceleration of human prostate cancer growth in vivo by factors produced by prostate and bone ®broblasts. Cancer Res 1991; 51: 3753. 32 Sato N et al. A metastatic and androgen-sensitive human prostate cancer model using intraprostatic inoculation of LNCaP cells in SCID mice. Cancer Res 1997; 57: 1584 ± 1589. 33 Sato N et al. Intermittent androgen suppression delays time to non-androgen regulated prostate speci®c antigen gene expression in the human prostate LNCaP tumor model. J Steroid Biochem Molec Biol 1996; 58: 139 ± 146. 34 Thalmann GN et al. Androgen-independent cancer progression and bone metastasis in the LNCaP Model of human prostate cancer. Cancer Res 1994; 54: 2577. 35 Umekita Y, Hiipakka RA, Kokontis JM, Liao S. Human prostate tumor growth in athymic mice: inhibition by androgens and stimulation by ®nasteride. Proc Natl Acad Sci 1996; 93: 11802 ± 11807. 36 Zhou HYE et al. Androgen-repressed phenotype in human prostate cancer. Proc Natl Acad Sci USA 1996; 93: 15152 ± 15157. 37 Stoll BA. Palliation by castration or by hormone administration. In: Stoll BA (ed). Breast Cancer Management Early and Late. Year Book Medical Publishers, Inc.: Chicago 1977, pp 133 ± 146. 38 Klotz LH, Herr HW, Morse MJ, Whitmore WF Jr. Intermittent endocrine therapy for advanced prostate cancer. Cancer 1986; 58: 2546 ± 2550. 39 Goldenberg SL et al. Intermittent androgen suppression in the treatment of prostate cancer: a preliminary report. Urology 1995; 45: 839 ± 845. 40 Gleave ME, Bruchovsky N, Goldenberg SL, Rennie P. Intermittent androgen suppression: Rationale and clinical experience. In: Schroeder F (Ed). Recent Advances in Prostate Cancer and BPH. Parthenon Publishing: London, 1997, pp 109±121. 41 Oliver RT, Williams G, Paris AM, Blandy JP. Intermittent androgen deprivation after PSA-complete response as a strategy to reduce induction of hormone-resistant prostate cancer. Urology 1997; 49: 79±82. 42 Higano CS, Ellis W, Russell K, Lange PH. Intermittent androgen suppression with leuprolide and ¯utamide for prostate cancer. A pilot study. Urology 1996; 48: 800±804. 43 Grossfeld GD, Small EJ, Carroll PR. Intermittent androgen deprivation for clinically localized prostate cancer: Initial experience. Urology 1998; 51: 137 ± 144. 44 Bruchovsky N et al. A phase II study of intermittent androgen suppression in men with rising serum PSA after radiation therapy for localised prostate cancer. J Urol 1998; 159: 335A 45 Gleave ME et al. Maximal biochemical and pathological downstaging requires 8 months of neoadjuvant hormonal therapy prior to radical prostatectomy. J Urol 1996; 155: 213±219. 46 Gleave ME et al. Longer duration of neoadjuvant androgen withdrawal therapy prior to radical prostatectomy in clinically localized prostate cancer: Biochemical and pathological effects. Mol Urol 1997; 1: 199±204. 47 Miller JI et al. The clinical usefulness of serum prostate speci®c antigen after hormonal therapy of metastatic prostate cancer. J Urol 1992; 147: 956±961. 48 Zubovitis JT, Paterson RF, Gleave ME, Jones EC. The effect of neoadjuvant hormone therapy on prostatic adenocarcinoma in radical prostatectomy specimens: an analysis of tumor cell proliferation. Am J Pathology 1998; (Abstract). (in press). 49 Murakami Y et al. Immunohistochemical analysis of Ki-67 and Bcl-2 in prostate cancer: Effect of neoadjuvant hormonal therapy. Br J Urol 1997; 80: 238 (abstract). 50 Reid JC. Bcl-2 and the regulation of programmed cell death. J Cell Biol 1994; 124: 1± 6. 51 Gleave M et al. Antisense Bcl2 oligonucleotides inhibit progression to androgen-independence after castration in the LNCaP tumor model. J Urol 1998; 159: 6A