Purpose: To investigate the safety and feasibility of rapid androgen cycling for men with progressive prostate cancer.

Experimental Design: Schedule 1 included a 4-week induction of androgen depletion, followed by 4-week treatment cycles of a monthly gonadotropin-releasing hormone agonist, testosterone on days 1 to 7, and an estrogen patch on days 8 to 21. Schedule 2 included a 12-week induction of androgen depletion followed by 4-week cycles of gonadotropin-releasing hormone agonist and testosterone, but no estrogens for patients with a prostate-specific antigen (PSA) nadir <1 ng/mL after induction. The primary end point was serially declining PSA trough values over six treatment cycles.

Results: Thirty-six patients were treated; 27 were evaluable after cycling, of whom 8 of 12 (67%) and 9 of 15 (60%) on schedules 1 and 2, respectively, reached the end point. Five patients with PSA >1 ng/mL following induction did not cycle. No patient progressed radiographically or clinically during cycling. Three posttherapy PSA patterns were observed: a decline followed by a rapid increase in trough levels, a sustained decline with a plateau at a detectable nadir, and a decline to an undetectable nadir. Mean testosterone levels were castrate at the time of trough and in the normal physiologic range following androgen repletion. Major toxicities included grades 1 and 2 fatigue, hepatitis, gynecomastia, and hot flashes.

Conclusions: Rapid hormonal cycling is feasible and well tolerated, and successive declines in PSA troughs are achievable. Although the sample size was small, the proportion of patients achieving declining PSA at the end of six cycles was comparable with that reached with continuous androgen depletion therapy.

Androgen depletion therapy (ADT) has been the cornerstone of advanced prostate cancer management during the past six decades, producing anticancer effects that include declines in prostate-specific antigen (PSA), regression of visible tumor masses, improvements on bone scan, and palliation of symptoms (1). Unfortunately, ADT alone rarely results in cure. Based on traditional response criteria, tumor regressions are documented in only 30% to 40% of cases, and the tumor often progresses initially at the same sites of disease that were manifest before ADT. Even when tumor spread is limited to regional nodes or to the prostate itself, complete eradication with ADT alone is rare. These observations show that tumor cells that are resistant and able to survive the effects of androgen depletion are present upon treatment.

In human prostate cancer xenograft models, ADT dramatically increases apoptotic rates, which return to baseline within days (2), whereas surviving cancer cells remain in a viable nonproliferative state (3) until castration-resistant phenotypes become manifest (4). The same rapid increase and decrease in apoptotic rates has been confirmed in studies of normal and malignant human prostate epithelial cells. In a study where daily biopsies of the prostate were done after the start of ADT, maximal apoptosis occurred by day 3 and was virtually absent by day 7 (5). These observations suggest that ADT can be considered as analogous to a single dose of chemotherapy; that is, an intervention that results in fractional cell kill until the surviving quiescent cancer cells undergo adaptive changes that allow regrowth despite castrate levels of testosterone. This raised the question whether overall tumor cell kill might be increased if multiple apoptotic cycles could be induced.

In an attempt to delay the development of castration resistance, investigators have evaluated intermittent androgen depletion. This is a cyclic process in which androgens are depleted, a response is documented, therapy is stopped, and the tumor is allowed to regrow under the influence of androgens. Androgen depletion is resumed when tumor regrowth reaches a predetermined level. In preclinical models, successive cycles of androgen depletion and repletion have been shown to prolong the time of sensitivity to changes in androgen levels and reduced the number of cells showing a castration-resistant phenotype (6, 7). Similar results have been observed in the clinic, and toxicities are reduced during the off periods. Intermittent androgen depletion has been studied by many groups (811), and ongoing large-scale phase 3 trials are addressing whether the approach is superior, equivalent, or inferior to more traditional continuous treatment (reviewed in refs. 12, 13).

Rapid androgen cycling is based on the hypothesis that shorter and more frequent cycles of androgen depletion and repletion might increase overall cell kill. For this study, 7 days of androgen repletion was chosen, in part because preclinical studies showed that proliferation rates in human prostate cancer xenografts assessed by Ki-67 staining and intratumoral fluorodeoxyglucose uptake by positron emission tomography increased within days following testosterone administration (14), and that the maximum increase in proliferation in a transgenic model occurred by day 4 or 5 (15). The time to maximal proliferation observed in these models mirrors the time to clinical flare of symptoms following testosterone administration to patients with castration-resistant disease (16). In addition, shorter cycles might diminish the frequency of the gain-of-function changes in the androgen receptor observed in tumors progressing during long-term continuous ADT (17, 18). In this first report exploring feasibility and safety, the primary aim was to determine the proportion of patients with PSA levels that decline over successive castration phases for a minimum of six cycles.

This was an Institutional Review Board–approved, single-institution study. All patients signed an Institutional Review Board–approved informed consent.

Patient eligibility

Eligible patients had a diagnosis of prostate cancer that was histologically confirmed; at study entry, patients had an increasing PSA alone or an increasing PSA and clinical metastases following radical surgery or radiation therapy for treatment of a clinically localized disease or had clinical metastases at the time of diagnosis. Increasing PSA was defined as a ≥50% increase in value to a level ≥2 ng/mL, based on at least three PSA determinations obtained >2 weeks apart. Prior ADT for up to 10 months was permitted as neoadjuvant before radiation therapy or as a course of planned intermittent therapy. Additional eligibility requirements included (a) a Karnofsky performance status of ≥70%, (b) adequate organ function, and (c) a serum testosterone level within the reference range. Patients who had received prior chemotherapy, immunotherapy, or therapeutic radiation for metastatic disease were excluded.

Study design

Two sequential treatment schedules were explored (Fig. 1). Schedule 1 enrolled 13 patients, and schedule 2 enrolled 23 patients.

Fig. 1.

Schematic of treatments. Hypothetical testosterone (dashed lines) and PSA (solid lines) levels. A, schedule 1, after rapid androgen depletion was achieved with ketoconazole and GnRH analogue, cycle 1 began with androgen repletion for 1 wk, followed by estrogen for 2 wk. B, schedule 2, 3-mo induction with bicalutamide and GnRH analogue, with cycling limited to patients who achieved a PSA nadir below 1.0 ng/mL.

Fig. 1.

Schematic of treatments. Hypothetical testosterone (dashed lines) and PSA (solid lines) levels. A, schedule 1, after rapid androgen depletion was achieved with ketoconazole and GnRH analogue, cycle 1 began with androgen repletion for 1 wk, followed by estrogen for 2 wk. B, schedule 2, 3-mo induction with bicalutamide and GnRH analogue, with cycling limited to patients who achieved a PSA nadir below 1.0 ng/mL.

Close modal

Schedule 1. To rapidly deplete androgens to castrate levels, patients received ketoconazole (400 mg thrice daily) on days 1 to 14 (19) and a monthly injection of a gonadotropin-releasing hormone (GnRH) analogue (leuprolide 7.5 mg i.m., goserelin 3.6 mg s.c., or their equivalent) on day 1. After a 4-week induction, all patients proceeded to cycling (one cycle = 28 days), during which patients received the monthly GnRH analogue on day 1 and testosterone (AndroGel 1%, 5 g topical daily) for 1 week. During weeks 2 and 3, patients used an estradiol patch (Climara 0.1 mg/d topically every 7 days). No therapy was given during week 4. The rationale for administering estrogen was derived from our institution studies that showed that squamous metaplasia was present in prostatectomy specimens obtained after 3 months of diethylstilbestrol, which was not observed after GnRH analogue therapy alone. This suggested a direct toxic effect of the estrogen on tumor cells (20). Cycles were repeated until the occurrence of progression of the disease as defined below.

Schedule 2. Changes to schedule 1 were adopted because of the toxicities encountered with ketoconazole and estradiol. The induction period was extended to 12 weeks of GnRH analogue (leuprolide 7.5 mg i.m. or goserelin 3.6 mg s.c. every 4 weeks) on day 1 and bicalutamide (50 mg daily) on days 1 to 28 and no treatment for the remainder of the induction period. The extension provided enough time for bicalutamide, which has a long half-life (21), to wash out before cycling. To restrict cycling to tumors that might be more sensitive, cycling was limited to patients with a PSA level of ≤1 ng/mL after the induction period. The cycling phase was simplified to a monthly GnRH analogue injection on day 1, followed by testosterone repletion (AndroGel 1%, 5 g topical daily) on days 1 to 7. No additional therapy was given from days 8 to 28. The cycle was repeated every 4 weeks until progression.

Patient evaluation

Patients were examined and assessed for safety and tolerance a minimum of once a month during induction. While cycling, patients were evaluated weekly for the first 2 weeks of the first cycle and monthly for subsequent cycles. Patients were required to keep and submit a drug diary to record testosterone gel administration and toxicities. PSA levels were measured at the start of the therapy, immediately before testosterone administration (trough level), and immediately before testosterone discontinuation (peak level).

Patients who had metastatic disease on baseline imaging (computed tomography, magnetic resonance, or bone scan) had follow-up studies done at 3-month intervals, whereas those with no radiographic abnormalities at baseline had follow-up studies at 6-month intervals or earlier if indicated. PSA levels were determined by a two-site immunoenzymometric assay using the Tosoh Medics Nexia analytic system (South San Francisco, CA). The low-end analytic sensitivity of the PSA method is 0.05 ng/mL.

Testosterone measurement

Total testosterone concentrations were determined by a solid-phase RIA (Diagnostic Products Corp., Los Angeles, CA) with an analytic sensitivity of 10 ng/dL. During schedule 1, total testosterone levels were measured at baseline and before (trough) and after (peak) androgen repletion therapy. For schedule 2, total testosterone was determined monthly during induction, as well as before and after cycling. Free testosterone was calculated from total testosterone, sex hormone–binding globulin, and albumin. Sex hormone–binding globulin was determined by immunometric analysis on a DPC Immulite analyzer (Diagnostic Products). Blood count and biochemical profile were also monitored.

Statistical analysis

Disease progression was defined as an increase in trough PSA relative to the previous cycle that was confirmed in a subsequent cycle. Treatment failure was defined as progression confirmed by the start of the seventh cycle. Six cycles were chosen because previous research has shown that PSA levels generally nadir by 6 months after ADT (22). The primary end point of the trial was the proportion of patients who did not have treatment failure. The proportion of patients who achieved undetectable PSA levels was also recorded (23). No overall response categories were assigned, although changes in soft tissue disease and/or bone scan were monitored.

The score test from the logistic regression model was used to determine if log-transformed PSA and testosterone measurements, obtained for each patient during the first six cycles, were associated with progression. The maximum value of the testosterone peaks and the minimum value of the testosterone troughs were used in the model.

The demographics, disease status, and prior treatment histories of the 36 patients treated are detailed in Table 1.

Table 1.

Patient characteristics

Schedule 1
Schedule 2
Baseline characteristics (all patients)N = 13N = 23
Mean age, y (range) 63 (52-75) 67 (55-84) 
Median performance status, % (KPS) 90 90 
Median Gleason score (range) 8 (6-9) 8 (6-9) 
Mean baseline PSA level, ng/mL (range) 22 (2.9-136) 63 (2.6-547) 
Mean baseline testosterone level, ng/dL (range) 330 (181-490) 382 (181-654) 
   
Disease state (evaluable patients)
 
N = 12
 
N = 20
 
Increasing PSA, n (%) 5 (42) 5 (25) 
Noncastrate metastatic, n (%) 7 (58) 15 (75) 
   
Primary therapy (evaluable patients)
 
N = 12
 
N = 20
 
Radiation therapy alone, n (%) 2 (17) 8 (40) 
Radical prostatectomy, n (%)* 10 (83) 10 (50) 
None 2 (10) 
   
Prior ADT, n (%) 5 (42) 5 (25) 
Schedule 1
Schedule 2
Baseline characteristics (all patients)N = 13N = 23
Mean age, y (range) 63 (52-75) 67 (55-84) 
Median performance status, % (KPS) 90 90 
Median Gleason score (range) 8 (6-9) 8 (6-9) 
Mean baseline PSA level, ng/mL (range) 22 (2.9-136) 63 (2.6-547) 
Mean baseline testosterone level, ng/dL (range) 330 (181-490) 382 (181-654) 
   
Disease state (evaluable patients)
 
N = 12
 
N = 20
 
Increasing PSA, n (%) 5 (42) 5 (25) 
Noncastrate metastatic, n (%) 7 (58) 15 (75) 
   
Primary therapy (evaluable patients)
 
N = 12
 
N = 20
 
Radiation therapy alone, n (%) 2 (17) 8 (40) 
Radical prostatectomy, n (%)* 10 (83) 10 (50) 
None 2 (10) 
   
Prior ADT, n (%) 5 (42) 5 (25) 

Abbreviation: KPS, Karnofsky performance status.

*

Alone or with local radiation therapy.

ADT in addition to primary therapy.

Schedule 1. Thirteen patients were treated, of whom one developed hepatitis on ketoconazole and was taken off the study before cycling. Of the 12 evaluable patients (five with increasing PSA and seven with clinical metastases), 8 patients (67%; three of the five with an increasing PSA and five of the seven with metastases) had declining PSA trough values through the first six cycles and met the primary end point. Among the patients who reached the end point, all five patients with clinical metastases (three with soft tissue disease and two with bone disease) showed improvement or stable disease on imaging posttreatment. Treatment was considered to have failed in 4 of the 12 (33%) evaluable patients (Table 2A), including two of the seven with metastatic disease (one with soft tissue lesions, one with bone lesions).

Table 2.

Outcomes of evaluable patients

Completed >6 cyclesPSA progression <6 cyclesInduction nadir PSA >1
A. Schedule 1    
    No. patients  
    Mean no. cycles completed (range) 17 (10-32) 5 (3-5)  
    Mean % PSA decline from baseline to completion of induction, % (range) 80 (50-97) 83 (69-94)  
    No. patients with prior ADT  
    Mean duration of prior ADT, mo (range) 5.3 (4-6) 6 (2-10)  
    Mean time from the end of prior ADT to enrollment on study, mo (range) 18 (18-24) 11.5 (4-19)  
    
B. Schedule 2    
    No. patients 
    Mean no. cycles completed (range) >21 (12 to >44) 5 (3-6) NA 
    No. patients with prior ADT 
    Mean duration of prior ADT, mo (range) 2 (1-3) NA 
    Mean time from the end of prior ADT to enrollment on study, mo (range) 18 (12-28) 7.5 NA 
Completed >6 cyclesPSA progression <6 cyclesInduction nadir PSA >1
A. Schedule 1    
    No. patients  
    Mean no. cycles completed (range) 17 (10-32) 5 (3-5)  
    Mean % PSA decline from baseline to completion of induction, % (range) 80 (50-97) 83 (69-94)  
    No. patients with prior ADT  
    Mean duration of prior ADT, mo (range) 5.3 (4-6) 6 (2-10)  
    Mean time from the end of prior ADT to enrollment on study, mo (range) 18 (18-24) 11.5 (4-19)  
    
B. Schedule 2    
    No. patients 
    Mean no. cycles completed (range) >21 (12 to >44) 5 (3-6) NA 
    No. patients with prior ADT 
    Mean duration of prior ADT, mo (range) 2 (1-3) NA 
    Mean time from the end of prior ADT to enrollment on study, mo (range) 18 (12-28) 7.5 NA 

Abbreviation: NA, not applicable.

At progression, 9 of the 12 patients continued on GnRH agonist therapy without testosterone supplementation. Four patients showed a subsequent decline in PSA to a level equal to or lower than the trough nadir PSA documented while on study, two of whom reached undetectable nadir PSA levels both while cycling and while on GnRH monotherapy off the study.

Notably, after 16 cycles, the PSA level was undetectable, with no evidence of radiographic disease in a 64-year-old man with a T3b Gleason 9 tumor, biopsy-proven metastatic disease in lymph nodes, and baseline PSA level of 12 ng/mL. Treatment was discontinued at his request, and the patient was observed. Thirty-four months later, he has no evidence of disease on scan, and his PSA remains undetectable, with a testosterone level in the reference range.

Schedule 2. Twenty-three patients were treated. Three patients were not evaluable, one because bicalutamide was continued beyond the 28-day limit, one because metastases from a second primary tumor were identified, and one because an arterial thrombus developed. Of the 20 evaluable patients, 5 (20%), all of whom had clinical metastasis, did not achieve a PSA of <1 ng/mL after the 12-week induction period and did not cycle.

Of the 15 patients who did cycle (5 with increasing PSA and 10 with metastases), 9 (60%) met the primary end point (3 of 5 with increasing PSA and 6 of 10 with metastases; Table 2B). Of the six patients with clinical metastases who met the end point, four had soft tissue disease, and three had bone disease (one patient had both). All showed improvement or stable disease on imaging posttreatment. Six of the 15 who cycled (40%) did not meet the end point and were treated with GnRH agonist alone at or before completing six cycles; this group included 4 of 10 with metastatic disease (two with soft tissue lesions and two with bone lesions), who were stable and improved posttreatment. Off study, 18 of the 20 evaluable patients continued on GnRH agonist therapy, 14 showed declines in PSA, and 3 reached an undetectable nadir.

Considering the two treatment schedules, a total of 27 patients were evaluable after cycling. No patient progressed radiographically or clinically during cycling. Among the 17 patients who had undergone a radical prostatectomy, 11 (65%) met the end point, and 8 of 17 (47%) achieved an undetectable PSA. Five of the nine patients who had undergone radiation therapy met the primary end point, and two achieved an undetectable PSA (<0.05 ng/mL). Notably, five of the nine patients treated with external beam radiation therapy cycled for at least 10 months with trough PSA levels below 0.3 ng/mL.

Overall, three PSA patterns were observed among patients who cycled. A serial decline in peaks and troughs to an undetectable nadir was observed in nine patients (3 of 10 treated for increasing PSA and 6 of 17 with clinical metastases). Figure 2A shows an example of such a PSA profile. A PSA decline to reach a plateau at a detectable level (Fig. 2B) was observed in eight patients (three with increasing PSA and five with clinical metastases). The third pattern, a PSA increase before the seventh cycle (Fig. 2C), corresponds to the definition of treatment failure. This pattern occurred in 10 patients (four with increasing PSA and six with clinical metastases). Nine of these patients had little or no initial decline in PSA, but one patient on schedule 2 initially achieved an undetectable PSA before the subsequent increase.

Fig. 2.

Patterns of PSA change following androgen depletion and repletion. Serial PSA values in three patients with recurrence after radical prostatectomy. A, serial declines in PSA peaks and troughs to an undetectable nadir. B, serial declines to nadir that remains detectable. C, transient decline followed by serial elevations in peaks and troughs. Because the elevation occurred by the sixth cycle, this patient was classified as having treatment failure.

Fig. 2.

Patterns of PSA change following androgen depletion and repletion. Serial PSA values in three patients with recurrence after radical prostatectomy. A, serial declines in PSA peaks and troughs to an undetectable nadir. B, serial declines to nadir that remains detectable. C, transient decline followed by serial elevations in peaks and troughs. Because the elevation occurred by the sixth cycle, this patient was classified as having treatment failure.

Close modal

Testosterone levels. Among patients who cycled on either treatment schedule, 99% of all trough testosterone measurements were <50 ng/dL (i.e., in the castrate range), 83% were <20 ng/dL, and 58% were ≤10 ng/dL. Patients who cycled consistently achieved peak free and total testosterone levels within the reference range during testosterone repletion. Testosterone values greater than the upper limit of the reference range (>758 ng/dL) were documented in 9.6% of patients on schedule 2 who met the end point and in 3.6% of patients who did not meet the end point. There were no acute toxicities associated with the elevated testosterone levels. No correlation was found between baseline, peak, or trough testosterone levels and the achievement of the primary end point (P = 0.33 and P = 0.32, respectively) or an undetectable PSA (P = 0.77 and P = 0.99; Table 3).

Table 3.

Summary of hormone data for patients on schedule 2

Testosterone measureMet end point (n = 9)Did not meet end point (n = 6)
Baseline, ng/dL 368 ± 91 410 ± 152 
Peak, ng/dL 381 ± 311 295 ± 169 
Trough, ng/dL 6 ± 10 11.2 ± 17 
Baseline-free, pg/mL 70.4 ± 8.4 66.3 ± 27.4 
Peak-free, pg/mL 75.3 ± 49.8 54.8 ± 33.8 
Trough-free, pg/mL 2.1 ± 6.5 1.3 ± 2.4 
Testosterone measureMet end point (n = 9)Did not meet end point (n = 6)
Baseline, ng/dL 368 ± 91 410 ± 152 
Peak, ng/dL 381 ± 311 295 ± 169 
Trough, ng/dL 6 ± 10 11.2 ± 17 
Baseline-free, pg/mL 70.4 ± 8.4 66.3 ± 27.4 
Peak-free, pg/mL 75.3 ± 49.8 54.8 ± 33.8 
Trough-free, pg/mL 2.1 ± 6.5 1.3 ± 2.4 

NOTE: Reference ranges for men over 50 years old are 181 to 758 ng/dL for total testosterone and 47 to 244 pg/mL for free testosterone. Values in table expressed as mean ± SD.

Adverse events.Table 4 describes the adverse events observed for both treatment schedules. The most common effects occurring during schedule 1 were grades 1 and 2 hepatitis, fatigue, and gynecomastia, which are known effects of ketoconazole and estrogens. As noted previously, one patient was taken off the study after 10 days of ketoconazole. Two patients developed pneumonic processes on chest plain film radiographs that were unrelated to the study treatment. Both remained in the study. In schedule 2, grade 1 fatigue, hot flashes, hyperglycemia, and urinary frequency were the most commonly reported adverse events. There were no grade 3 or 4 toxicities. Anecdotally, patients experienced improved energy and sexual performance during the week of androgen repletion.

Table 4.

Adverse events

EventGrade 1Grade 2Grade 3Grade 4Total
A. Schedule 1 (N = 13)      
    Fatigue 10 (77) 0 (0) 0 (0) 0 (0) 10 (77) 
    Hot flashes 2 (15) 3 (23) 0 (0) 0 (0) 5 (38) 
    Transaminitis 6 (46) 3 (23) 0 (0) 0 (0) 9 (69) 
    Pneumonia 1 (8) 0 (0) 0 (0) 1 (8) 2 (15) 
    Decreased libido 1 (8) 1 (8) 0 (0) 0 (0) 2 (15) 
    Gynecomastia 6 (46) 2 (15) 0 (0) 0 (0) 8 (63) 
      
B. Schedule 2 (N = 23)
 
     
    Fatigue 10 (43) 4 (17) 0 (0) 0 (0) 14 (61) 
    Hot flashes 15 (65) 4 (17) 0 (0) 0 (0) 19 (83) 
    Urinary frequency 18 (78) 1 (4) 0 (0) 0 (0) 19 (83) 
    Elevated transaminases 3 (21) 0 (0) 0 (0) 0 (0) 3 (21) 
    Pneumonia 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 
    Decreased libido 1 (4) 0 (0) 0 (0) 0 (0) 1 (4) 
    Gynecomastia 3 (13) 0 (0) 0 (0) 0 (0) 3 (13) 
EventGrade 1Grade 2Grade 3Grade 4Total
A. Schedule 1 (N = 13)      
    Fatigue 10 (77) 0 (0) 0 (0) 0 (0) 10 (77) 
    Hot flashes 2 (15) 3 (23) 0 (0) 0 (0) 5 (38) 
    Transaminitis 6 (46) 3 (23) 0 (0) 0 (0) 9 (69) 
    Pneumonia 1 (8) 0 (0) 0 (0) 1 (8) 2 (15) 
    Decreased libido 1 (8) 1 (8) 0 (0) 0 (0) 2 (15) 
    Gynecomastia 6 (46) 2 (15) 0 (0) 0 (0) 8 (63) 
      
B. Schedule 2 (N = 23)
 
     
    Fatigue 10 (43) 4 (17) 0 (0) 0 (0) 14 (61) 
    Hot flashes 15 (65) 4 (17) 0 (0) 0 (0) 19 (83) 
    Urinary frequency 18 (78) 1 (4) 0 (0) 0 (0) 19 (83) 
    Elevated transaminases 3 (21) 0 (0) 0 (0) 0 (0) 3 (21) 
    Pneumonia 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 
    Decreased libido 1 (4) 0 (0) 0 (0) 0 (0) 1 (4) 
    Gynecomastia 3 (13) 0 (0) 0 (0) 0 (0) 3 (13) 

Note: Values in table expressed as n (%).

This report describes the first in a series of studies exploring the use of rapid androgen cycling in prostate cancer patients with an increasing PSA or noncastrate metastatic disease. Results show that monthly androgen depletion and repletion is feasible and can result in successive declines in PSA troughs and associated radiographic improvement.

Although the overall number of patients treated was small, results from schedule 1 suggested that the initial 4-week castration period may have been too short, and ketoconazole-associated toxicity occurred frequently enough to warrant a change to the induction schedule. As a result, schedule 2 instituted a 12-week induction period, initiated by 4 weeks of bicalutamide. Although five patients (20%) did not reach the specified PSA nadir ≤1 ng/mL after induction therapy and did not go on to cycling, for those patients who cycled, posttreatment PSA patterns on this schedule were essentially unchanged.

The rapid depletion and repletion of androgens were designed to increase net tumor cell kill by inducing multiple apoptotic events, while limiting the period of tumor regrowth. The hypothesis is that androgen repletion would cause dormant prostate cancer cells that had survived androgen depletion to enter the cell cycle, so that they would undergo apoptotic cell death when androgens were subsequently depleted (9, 24). Overall cell number would decrease in sequential cycles because of the 7-day exposure to androgens because the number of cells gained through proliferation during repletion would be less than the number killed during the previous period of androgen depletion. Serially declining PSA peaks and troughs in sequential cycles are consistent with this view, with the caveat that PSA levels alone do not provide an adequate measure of overall cell number. However, because regressions of tumor were documented concurrently and measured peak testosterone levels in the same patient were similar between cycles, the responses to rapid cycling are unlikely to be simply a PSA effect.

An additional objective of the study was to increase the proportion of patients who achieved an undetectable PSA. Although an undetectable PSA does not indicate cure, it is a useful benchmark by which to evaluate treatment effects and measure improvements in a treatment regimen (23, 25). In one retrospective study of continuous ADT, 57% of patients with an increasing PSA and 37% of patients with clinical metastases reached a PSA nadir that was undetectable (<0.05 ng/mL). Overall, the current study resulted in an undetectable PSA nadir in 10 of 27 assessable patients (37%; 95% confidence interval, 24-59%). Considering that more than two thirds of the patients enrolled had clinical metastases (Table 1), this proportion is comparable with that obtained with continuous ADT.

In the current study, all patients who did not achieve an undetectable PSA nadir eventually had disease progression. The result is consistent with previously noted associations between an undetectable PSA nadir and outcome. Compared with an undetectable PSA nadir after ADT, a PSA nadir of more than 0.2 ng/mL, combined with a short pre-ADT doubling time, has been linked to a higher prostate cancer–specific mortality, independent of the initial local therapy (25).

For the patients whose PSA nadir was a detectable level, it is unknown at this point whether the continued PSA production signified that the PSA-producing cells were resistant to androgen depletion or whether the degree of androgen depletion was insufficient to reduce signaling through the androgen receptor. In this trial, although 92% of the patients were documented to have testosterone trough levels below 20 ng/dL, and 76% were documented to have testosterone trough levels below 10 ng/dL, no between-measured peak and trough testosterone levels and outcome were observed. These measurements do not, however, indicate the levels of androgen in the tumor. In a study in which intraprostatic androgens were measured, ADT lowered intratumoral ligand levels by only 75%, and the residual androgens documented in the prostate were sufficient for androgen receptor signaling to continue (26). In a second report, persistent PSA and androgen receptor gene expression, which was observed in 7 of 18 tumors isolated from radical prostatectomy specimens after 3 months of neoadjuvant ADT, was associated with early PSA relapse (27). Both reports suggest that combining strategies that deplete ligand levels with strategies that inhibit the ligand binding to the receptor may improve outcomes. These approaches, however, would not address the management of tumors that are independent of androgen receptor signaling for growth and survival.

The practice of cycling between androgen-depleted and androgen-repleted states is not new. Intermittent androgen suppression was introduced initially as a strategy both to delay the outgrowth of castration-resistant cells and to minimize toxicity (8, 28). As currently used in an investigational setting, intermittent androgen suppression seeks to maintain a state of medical castration 2 to 3 months beyond the point of maximal PSA decline, at which time ADT is discontinued (29). As testosterone levels increase, PSA levels also increase; ADT is restarted at a predetermined PSA level, and the cycling continues. Multiple cycles of regression and proliferation have been documented, and a randomized controlled trial assessing the clinical benefits of this strategy is ongoing. Although there are concerns about allowing testosterone levels to increase in a patient with prostate cancer, preliminary analyses of the intermittent androgen suppression approach showed that outcomes are similar to those achieved with continuous ADT, the standard treatment in most contexts (11, 30). This finding contrasts with results found when testosterone is given to patients with castrate metastatic disease who have been on long-term ADT. In this disease state, androgen repletion results in a clinical flare in symptoms, and in a prospective randomized trial, overall survival for patients who received testosterone followed by chemotherapy was shorter than that for patients who received chemotherapy alone (16). The patient population treated in the current study is different because the tumors were not resistant to androgen depletion at the outset. The fact that patients in this study showed declines in PSA level when they were shifted to continuous ADT upon progression further shows that the cells were not resistant to androgens at the end of the cycling period.

During the stress of rapidly declining androgen levels, tumor cells may be more sensitive to targeted signaling inhibitors such as gefitinib (31) or RAD-001(32, 33). Chemotherapy may also be more effective against cycling as opposed to dormant, nonproliferating cells. Investigations are under way for treatment approaches combining rapid androgen cycling with (a) docetaxel given when testosterone levels are in the physiologic range (34) and (b) an anti-CTL antigen-4 monoclonal antibody, an inhibitor of T-cell down-regulation (35) as a means of augmenting any intrinsic anti–prostate cancer immunity. The combination immunotherapy with androgen cycling is attractive because apoptotic cells that are taken up by dendritic cells are capable of priming an immune response independent of knowledge of the target antigen. In a variety of infectious disease models, induction of apoptosis in antigen-bearing cells induced the priming of strong CD8+ T-cell immunity (36); the apoptotic cells functioned, in effect, as an endogenous immune adjuvant. The fact that the overall proportion of patients achieving an undetectable PSA level in this study was similar to the proportion among patients experiencing continuous ADT provides a platform for further exploration. The ultimate objective is to use monthly androgen cycling in combination with other treatment strategies to augment the proportion of patients who achieve undetectable PSA levels, a prerequisite for cure.

Grant support: Prostate Cancer Foundation, Memorial Sloan-Kettering Cancer Center Specialized Programs of Research Excellence in Prostate Cancer, K30, American Society of Clinical Oncology Young Investigator Award (D. Feltquate).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. Peter Nelson and Janet Novak for critical reviews of the manuscript.

1
Scher HI, Leibel SA, Fuks Z, Cordon-Cardo C, Scardino PT. Cancer of the prostate. In: DeVita VT, Hellman S, Rosenberg SA, editors. Cancer principles and practice of oncology. Philadelphia: Lippincott-Raven Publishers; 2005. p. 1192–259.
2
Kim D, Gregory CW, French FS, Smith GJ, Mohler JL. Androgen receptor expression and cellular proliferation during transition from androgen-dependent to recurrent growth after castration in the CWR22 prostate cancer xenograft.
Am J Pathol
2002
;
160
:
219
–26.
3
Smitherman AB, Gregory CW, Mohler JL. Apoptosis levels increase after castration in the CWR22 human prostate cancer xenograft.
Prostate
2003
;
57
:
24
–31.
4
Agus DB, Cordon-Cardo C, Fox W, et al. Alterations of cell cycle regulators in prostate cancer: response to androgen withdrawal and development of androgen independence.
J Natl Cancer Inst
1999
;
91
:
1869
–76.
5
Ohlson N, Wikstrom P, Stattin P, Bergh A. Cell proliferation and apoptosis in prostate tumors and adjacent non-malignant prostate tissue in patients at different time-points after castration treatment.
Prostate
2005
;
62
:
307
–15.
6
Akakura K, Bruchovsky N, Goldenberg SL, Rennie PS, Buckley AR, Sullivan LD. Effects of intermittent androgen suppression on androgen-dependent tumors. Apoptosis and serum prostate-specific antigen.
Cancer
1993
;
71
:
2782
–90.
7
Pantuck AJ, Zismon A, Tso C-L, et al. Intermittent androgen deprivation therapy: schedule modifications based on a novel in vivo human xenograft model.
Prostate Cancer Prostatic Dis
2000
;
3
:
280
–2.
8
Klotz LH, Herr HW, Morse MJ, Whitmore WF. Intermittent endocrine therapy for advanced prostate cancer.
Cancer
1986
;
58
:
2546
–50.
9
Akakura K, Bruchovsky N, Goldenberg SL, Rennie PS, Buckley AR, Sullivan LD. Effects of intermittent androgen suppression on androgen dependent tumors.
Cancer
1993
;
71
:
2782
–90.
10
de Leval J, Boca P, Yousef E, et al. Intermittent versus continuous total androgen blockade in the treatment of patients with advanced hormone-naive prostate cancer: results of a prospective randomized multicenter trial.
Clin Prostate Cancer
2002
;
1
:
163
–71.
11
Calais da Silva FM, Calais da Silva F, Bono A, et al. Phase III intermittent MAB vs continuous MAB [abstract 4513].
Proc Am Soc Clin Oncol
2006
;
24
:
220s
.
12
Rashid MH, Chaudhary UB. Intermittent androgen deprivation therapy for prostate cancer.
Oncologist
2004
;
9
:
295
–301.
13
Wright JL, Higano CS, Lin DW. Intermittent androgen deprivation: clinical experience and practical applications.
Urol Clin North Am
2006
;
33
:
167
–79, vi.
14
Agus DB, Golde DW, Sgouros G, Ballangrud A, Cordon-Cardo C, Scher HI. Positron emission tomography of a human prostate cancer xenograft: the association of changes in deoxyglucose accumulation and response to hormonal therapy.
Cancer Res
1998
;
58
:
3009
–14.
15
Shaffer DR, Viale A, Ishiwata R, et al. Evidence for a p27 tumor suppressive function independent of its role regulating cell proliferation in the prostate.
Proc Natl Acad Sci U S A
2005
;
102
:
210
–5.
16
Manni A, Bartholomew M, Caplan R, et al. Androgen priming and chemotherapy in advanced prostate cancer: evaluation of determinants of clinical outcome.
J Clin Oncol
1988
;
6
:
1456
–66.
17
Koivisto P, Kononen J, Palmberg C, et al. Androgen receptor gene amplification: a possible molecular mechanism for androgen deprivation failure in prostate cancer.
Cancer Res
1997
;
57
:
314
–9.
18
Scher HI, Buchanan G, Gerald W, Butler LM, Tilley WD. Targeting the androgen receptor: improving outcomes for castration-resistant prostate cancer.
Endocr Relat Cancer
2004
;
11
:
459
–76.
19
Trachtenberg J, Pont A. Ketoconazole therapy for advanced prostate cancer.
Lancet
1984
;
2
:
433
–5.
20
Armas OA, Aprikian AG, Melamed J, et al. Clinical and pathobiological effects of neoadjuvant total androgen ablation therapy on clinically localized prostatic adenocarcinoma.
Am J Surg Pathol
1994
;
18
:
979
–91.
21
Furr B, Tucker H. The preclinical development of bicalutamide: pharmacodynamics and mechanism of action.
Urology
1996
;
47
:
14
–25.
22
Bruchovsky N, Goldenberg SL, Akakura K, Rennie PS. Luteinizing hormone-releasing hormone agonists in prostate cancer.
Cancer
1993
;
72
:
1685
–91.
23
Beekman K, Morris M, Slovin S, et al. Androgen deprivation for minimal metastatic disease: threshold for achieving undetectable prostate-specific antigen.
Urology
2005
;
65
:
947
–52.
24
English HF, Kloszewski ED, Valentine EG, Santen RJ. Proliferative response of the Dunning R3327H experimental model of prostatic adenocarcinoma to conditions of androgen depletion and repletion.
Cancer Res
1986
;
46
:
839
–44.
25
Stewart AJ, Scher HI, Chen MH, et al. Prostate-specific antigen nadir and cancer-specific mortality following hormonal therapy for prostate-specific antigen failure.
J Clin Oncol
2005
;
23
:
6556
–60.
26
Nelson PS. Androgen supplementation in the aging male: implications for prostate disease. In: Presentation at the 12th Annual Prostate Cancer Foundation Scientific Retreat. Scottsdale, AZ; 2005.
27
Ryan CJ, Smith A, Lal P, et al. Persistent PSA expression after neoadjuvant androgen deprivation: an early predictor of relapse or incomplete androgen suppression.
Urology
2006
;
68
:
834
–9.
28
Lane TM, Ansell W, Farrugia D, et al. Long-term outcomes in patients with prostate cancer managed with intermittent androgen suppression.
Urol Int
2004
;
73
:
117
–22.
29
Goldenberg SL, Bruchovsky N, Gleave ME, Sullivan LD, Akakura K. Intermittent androgen suppression in the treatment of prostate cancer: a preliminary report.
Urology
1995
;
45
:
839
–44; discussion 44–5.
30
Hussain M. PSA as an outcome marker in de novo metastatic prostate cancer. In: 2005 Multidisciplinary Prostate Cancer Symposium. Orlando, FL; 2005.
31
Sirotnak FM, She Y, Lee F, Chen J, Scher HI. Studies with CWR22 xenografts in nude mice suggest that ZD1839 may have a role in the treatment of both androgen-dependent and androgen-independent human prostate cancer.
Clin Cancer Res
2002
;
8
:
3870
–6.
32
Mellinghoff IK, Tran C, Sawyers CL. Growth inhibitory effects of the dual ErbB1/ErbB2 tyrosine kinase inhibitor PKI-166 on human prostate cancer xenografts.
Cancer Res
2002
;
62
:
5254
–9.
33
Eigl BJ, Eggener SE, Baybik J, et al. Timing is everything: preclinical evidence supporting simultaneous rather than sequential chemohormonal therapy for prostate cancer.
Clin Cancer Res
2005
;
11
:
4905
–11.
34
Slovin SF, Carducci M, Beekman KW, et al. The men's cycle plus docetaxel (Doc) in prostate patients with rising PSAs in the non-castrate state [abstract 4564].
Proc Am Soc Clin Oncol
2005
;
23
:
393s
.
35
Egen JG, Kuhns MS, Allison JP. CTLA-4: new insights into its biological function and use in tumor immunotherapy.
Nat Immunol
2002
;
3
:
611
–8.
36
Chattergoon MA, Shames JP, Weiner DB. Engineering cross-presentation in vivo.
Expert Opin Biol Ther
2003
;
3
:
887
–94.