Targeted therapy in prostate cancer—Are we our own worst enemy?


  • Nancy A. Dawson MD

    Corresponding author
    1. Department of Medicine, Lombardi Comprehensive Cancer Center, Georgetown University, Washington, DC
    • Lombardi Comprehensive Cancer Center, Georgetown University Hospital, Washington, DC 20007===

    Search for more papers by this author
    • Fax: (202) 444-9429.

  • See referenced original article on pages 2478–87, this issue.


Targeted therapies are being assessed in castrate-resistant prostate cancer, but their development is slow, and the correct targets are not always clear. Given the propensity for prostate cancer to metastasize to bone, therapies that target the bone microenvironment are a logical choice.

The current management of castrate-resistant prostate cancer (CRPC) includes second-line hormone manipulations and cytotoxic chemotherapy. Docetaxel, the only chemotherapy drug that has been approved based on a survival benefit, results in a median prolongation of life of only 2.5 months.1 New therapies aimed at improving overall survival clearly constitute an unmet need.

Targeted therapies are emerging as active agents in other malignancies. In just the past few years, the vascular endothelial growth factor receptor and mammalian target of rapamycin inhibitors have revolutionized the treatment of metastatic renal cell carcinoma. Targeted therapies currently are being assessed in CRPC, but their development is slow, and the correct targets are not always clear. Given the propensity for prostate cancer to metastasize to bone, therapies targeting the bone microenvironment are a logical choice.

Enothelins in the bone microenvironment not only contribute to preservation of the vicious cycle of tumor migration to bone and its stimulation through release of various bone-derived growth factors; they also contribute to the sensation of pain. The endothelin receptor antagonists primarily block the binding of endothelin-1 (ET-1) to the endothelin A receptor (ETAR), but they also can block binding to the endothelin B receptor (ETBR). Binding to these receptors can have somewhat opposing effects on cancer progression, with ETBR mediating apoptosis and ETAR mediating the stimulation of mitogenesis, angiogenesis, osteoblast formation, and nocioception.2 Atrasentan is a selective ETAR antagonist that shows supplementary activity at the ETBR. Nelson et al previously reported on atrasentan in CRPC in the metastatic setting.3 In that trial, atrasentan did not reduce the risk of disease progression relative to placebo despite evidence of biologic effects on prostate-specific antigen (PSA) and bone alkaline phosphatase as markers of disease burden.

In their current article, these investigators report on a phase 3, randomized, placebo-controlled trial of atrasentan in patients with nonmetastatic CRPC.4 Given the expectation that targeted therapies are more likely to result in disease stabilization, it is well justified to evaluate this drug in a setting of lower tumor burden. On the surface, the results from their study are negative. The primary endpoint of time to progression (TTP) was 93 days longer in the atrasentan arm, but this did not reach statistical significance, with a P value of .288. Nor did the survival difference of 74 days meet statistical significance (P= .219).

However, on subset analysis, an interesting result emerges based on geographic distribution. The median time to progression was 590 days in patients who were treated in the US (US patients) and 847 days in non-US patients. The probable reason emerges as premature discontinuation of the drug. US patients were twice as likely to discontinue drug early compared with non-US patients (40.8% vs 21.9%; P < .001). Although the primary reason given was adverse events, the median increase in PSA at the time of discontinuation in the atrasentan arm was 12.5 ng/mL in the US patients and 20.3 ng/mL in the non-US patients, suggesting the rising PSA level was the real trigger to stop the drug. Although an increase in PSA was not considered disease progression according to the protocol, the discontinuation rate based on PSA progression was 4 times higher in the US patients (5.3% vs 1.1%). There also is a hint from this study that atrasentan is interrupting the vicious cycle, with a greater percentage of patients having new skeletal disease as the reason for disease progression in the placebo arm (44.3% vs 36.2%) and a 250-day difference in the time to initial skeletal metastases (1008 days in the atrasentan arm vs 757 days in the placebo arm); but, again, the difference was not statistically significant (P = .103).

So, how should we interpret these results? Should we abandon atrasentan and the endothelin receptor antagonists in general as ineffective; or, instead, should we try to unravel why the study was negative? I favor the second choice. To analyze the results successfully, hurdles must be overcome if we are to move these agents forward. Three challenges surface readily: Would a drug that only targeted ETAR without the antiapoptotic effects of ETBR antagonism been more effective? Would a combination of drugs that attack the bone microenvironment through different mechanisms improve outcomes? Last, but not least, is there any way to reverse the current PSA-driven decision making in the treatment of prostate cancer?

ZD4054 is a specific ETAR antagonist that demonstrates no detectable inhibition of the ETBR.5 The mature results of a randomized phase 2 trial of ZD4054 in men with prostate cancer and bone metastases who were pain-free or mildly symptomatic was presented at the American Society of Clinical Oncology 2008 Genitourinary Cancers Symposium.6 In total, 312 patients were randomized to ZD4054 at a dose of 15 mg daily, ZD4054 at a dose of 10 mg daily, or placebo. It is noteworthy that, in this global trial, although there was no difference in TTP in the intent-to-treat population, a prolonged overall survival was observed in the ZD4054 treatment groups at the time of a planned second analysis. The median overall survival was 23.5 months in the group treated with ZD4054 at a 15–mg dose and 24.5 months in the group treated with ZD4054 at a dose of10 mg compared with 17.3 months in the placebo group (P = .05 and P < .01, respectively, favoring the 2 ZD4054 arms). In that study, the reason for discontinuation of drug/placebo appeared to be similar in the 3 arms. The criteria for progression in the trial differed from the currently reported atrasentan study. The events indicative of progressive disease in the atrasentan trial were new skeletal or extraskeletal lesions or an event attributed to metastatic disease whereas in the ZD4054 trial, progression was defined as symptoms that required the initiation of a new therapy, cancer pain that required opiates, progression of soft tissue metastases, or death in the absence of progression. An additional important difference was that the atrasentan trial allowed crossover from placebo to drug and the ZD4054 trial did not, clearly masking any potential survival benefit from atrasentan.

A second important consideration is that the endothelin receptor antagonists may be more appropriate in combination with other targeted therapies. Bisphosphonates also interrupt the vicious cycle in the bone microenvironment by inhibiting osteoclasts and the release of bone factors that stimulate the local growth of migrated tumor cells. In a xenograft model of prostate cancer cells secreting ET-1, zoledronic acid combined with atrasentan had additive effects on the inhibition of tumor growth and lower PSA levels compared to the effect of either drug alone.7 The current Southwest Oncology Group (SWOG) trial in CRPC assesses the importance of this potential synergy by randomizing patients to docetaxel with or without atrasentan with stratification for the concomitant use of zoledronic acid.

The final and most challenging issue raised by the atrasentan trial is how to maintain patients on treatment when PSA (also known as physician/patient-stimulated anxiety) levels rise. If in fact a therapy is effective in terms of prolonging survival but requires longer drug exposure, then how can we achieve this? What can our non-US colleagues share in their communication with patients that will facilitate this goal? Is the reality that, in the US, targeted therapies can be assessed only alone in patients with end-stage disease to look for glimmers of activity and that the only studies we can complete are go-along studies, such as the SWOG trial, in which these agents are added to standard therapy? These questions may require global clinical trial think tanks to answer.

We owe it to our patients to figure out how best to help them, because we are certainly co-responsible for the PSA psychosis that exists. One step certainly should include mass education to inform individuals that there are no data to support that ‘PSA only’-directed changes in therapy improve patient outcome. A delay in the initiation of an alternate effective therapy will not make it less likely to work. In the interim, let us not abandon these targeted drugs; rather, let us focus on how to design and complete studies that allow these drugs a fair trial.