Medivation (MDVN)'s MDV3100 is an orally available small molecule that is currently in phase I-II trials for the treatment of hormone-refractory prostate cancer [HRPC], which is defined as advancing disease in a patient who already has castration-level serum androgens. MDV3100 is an androgen receptor [AR] antagonist. It is the consensus belief of urologists and urologic oncologists that the androgen receptor is integral to the development of malignant prostate disease: the activated AR binds to the DNA of malignant prostate cells, inducing their proliferation and spread. To this point, successful therapy has revolved around chemical castration, using anti-androgens such as finasteride to inhibit androgen formation and decrease its binding-activation of the AR.
This approach has a good initial response rate, but over time prostate cancers tend to recur and progress, despite serum evidence of suppressed androgens. There is convincing molecular evidence and consensus opinion that the AR is reactivated in HRPC and once again stimulates DNA expression and proliferation; whether this is due to binding of a non-androgen binding site on the receptor, or to local (autocrine) production of androgen by the transformed cell, or to another mechanism is uncertain.
To date, treatment for HRPC has been disappointing, with expected survival estimated at 7-16 months, and no significant improvement with current therapies. Bicalutamide (Casodex) is an AR antagonist that has agonist properties when the AR is present in high concentration, as it is in HRPC, and thus has not had sustained success. Chemotherapy with various combinations of docetaxel, mitoxantrone and prednisone has not been curative, and has been demonstrated to prolong life by several weeks only. MDV3100 is distinct from previous therapies: the molecule has a moiety that facilitates binding to the AR, and a separate moiety that induces a molecular conformation of the AR that makes it impossible to bind to DNA, thus making it anti-proliferative.
Elements that Support Positive Phase I-II Data and Subsequent Phase III Trial Initiatives
- MDV3100 is a novel molecule with a markedly downstream effect on the AR-mediated cascade. There are two active sites on the MDV3100 molecule: a binding site that attaches to the AR, and a separate moiety that induces conformational change in the AR protein, making it impossible for the AR to bind (and activate) DNA. MDVN had the benefit of having bicalutamide come before it, and with that experience the company was able to carefully screen its molecular library for a prospective molecule that has no agonist properties. Likewise, as a complete agonist of the AR, MDV3100 might not only have anti-proliferative effects, but it may have apoptotic (induced cell death) effects – it has been observed in pre-clinical studies that complete de-activation of the AR induces programmed cell death, which may explain why studies of MDV3100 to date demonstrate tumor reduction, and not just tumor arrest. In this manner, MDV3100 should be able to inhibit ligand (androgen)-binding sites, non-ligand-binding sites, and androgen response elements that interact with DNA and promote malignant proliferation. Likewise, the MDV3100 molecule should be able to induce apoptosis and promote tumor shrinkage. This complement of actions separates MDV3100 from previously developed therapies for prostate cancer.
- The data currently available from the on-going phase I-II MDV3100 trial has had significant reductions in PSA levels with no reports of significant toxicity. In unpublished data made available by MDVN, MDV3100 has reached the third tier of a seven-tier dose escalation study having demonstrated a dose-dependent reduction in PSA (prostate-specific antigen) levels. The lowest dose cohort achieved a 45-66% reduction in PSA after one month of treatment, and the second-lowest dose cohort achieved a 75-89% PSA reduction at one month. As part of the FDA dose-escalation study requirement, MDVN is required to expand the cohort at any dose in which toxicity is observed before progressing to the next dose cohort: the study has so far proceeded to the third tier without any recorded toxicity or expansion of cohort. Using similar drugs as a basis of comparison, bicalutamide (Casodex) is, like MDV3100, an AR antagonist, and it has a favorable therapeutic window with significant toxicities generally appearing at supra-therapeutic doses. As reported by MDVN, the dose escalation study has already arrived at a dose consistent with a possible clinical application.
- The pre-clinical data from animal model studies demonstrated positive, dose-dependent reductions in tumor burden. In a paper presented at the 2007 American Society of Clinical Oncology/Prostate Cancer Research Institute conference, mice transfected with LNCaP tumor (a widely accepted animal model for HRPC) showed significant, MDV3100 dose-dependent reductions of tumor burden. Dose cohorts ranged from 1 mg/kg/day to 50 mg/kg/day. All cohorts demonstrated tumor shrinkage (from 13% to 68%). Half of the highest dose cohort had complete resolution of tumor burden. Additionally, all of the mice gained weight throughout the study, and there was no drug-associated toxicity reported in any cohort (the highest dose cohort demonstrated toxicity related to the delivery system, however).
- There is general acceptance of PSA levels as a valid clinical biomarker. Published guidelines for clinical trial design for the treatment of prostate cancer is better established for preventive regimens than for treatment regimens. With that being mentioned, authoritative bodies within the field of urology do accept PSA levels as a valid endpoint, particularly in the evaluation of recurrence and/or tertiary prevention. In these scenarios, PSA has been shown to be a valid surrogate endpoint for distant-metastasis-free survival, disease-free survival, and cause-specific survival. This acceptance lends validity to the on-going phase I-II MDV3100 study, which utilizes PSA levels as a primary endpoint. Published guidelines for clinical trial design also suggest that high-risk patients be recruited when there is a goal other than primary prevention – in the MDV3100 study, all patients have hormone-refractory, progressive disease – and that single-arm studies are acceptable for phase II. It should be noted that the phase I-II MDV3100 trial is collecting measurements beyond these minimum endpoint criteria: the study is also observing clinical disease burden, radiographic disease burden, CPK levels, and circulating tumor levels.
- There has been established pre-clinical success with other molecules that also suppress AR-mediated gene transcription. Suppression of prostate cancer biomarkers has been demonstrated by numerous molecules that antagonize the AR: androgen receptor decoy molecules, ablative agents of the AR (such as histone deacetylase inhibitors), interleukin-6, non-steroidal AR antagonists (such as bicalutamide and flutamide) and several other molecules have achieved pre-clinical success in achieving suppression of molecular tumor markers. Though only the non-steroidal AR antagonists have demonstrated clinical success – and limited success, at that – the pre-clinical success of other AR antagonist molecules appears to validate the concept of MDV3100.
Arguments Against Positive Phase I-II Data and Subsequent Phase III Trial Initiatives
- The toxicity profile of MDV3100, specifically as it relates to its dose-dependent therapeutic profile, remains an unknown entity. The pre-clinical and clinical trials to date have demonstrated an excellent dose-dependent response: In mice, for example, tumor burden decreases with increasing doses, to the point of eradication in some subjects. And in humans, the on-going dose escalation study demonstrates a concomitant decrease in PSA levels, with lower doses in the study already approaching PSA eradication – a necessary prerequisite for curative therapy in humans. However, the maximal clinical benefit will be limited by the maximally tolerated dose, and this is not yet known. It is possible, for example, that subjects in the higher dose tiers of the phase I-II study will not tolerate MDV3100 without toxicity, thereby narrowing the therapeutic window of the drug, or that toxicity will develop over time frames not observed in early phase trials. The only relevant precedent for androgen receptor antagonists, however, is bicalutamide (Casodex), which is well tolerated at very high doses and has a relatively mild toxicity profile: diarrhea, impotence and decreased libido, hot flashes, gynecomastia and breast tenderness. In rare cases, though, bicalutamide is associated with elevated liver enzymes and hepatitis.
- Current MDV3100 trials in humans have been of a very short duration with a very small number of study participants. To date, each tier of the phase I-II MDV3100 trial has less than 10 patients – it is possible, of course, that the heretofore positive data may not extrapolate to larger study populations. Also, the data available so far has been gathered after only one or two months of therapy – it is possible that the benefits of MDV3100, like bicalutamide before it, attenuate over time. It is also possible that toxicity accumulates over time, and over longer studies the heretofore favorable toxicity profile of MDV3100 will similarly diminish.
- Previous and current experience with non-steroidal AR antagonists (bicalutamide, nilutamide, flutamide) has been disappointing. Non-steroidal AR antagonists are currently used for HRPC. Most experience has been with bicalutamide, which has modest benefits initially, and then stops working. This observation is consistent with the fact that bicalutamide is a mixed agonist-antagonist, and at higher concentrations of AR (as seen in prostate cancer), bicalutamide actually activates the AR. It is possible that MDV3100 could develop a similar profile as other AR antagonists – it might have only modest clinical benefits, and it might even evolve into an agonist over time. However, the MDV3100 molecule is, structurally, of a distinctly different class than the non-steroidal AR antagonists: as mentioned above, MDV3100 was engineered to minimize agonist properties, and it has a separate, non-binding molecular moiety that induces conformational change of the AR.
- Eventual adoption of MDV3100 by urologists and oncologists will depend upon more than just reduction of PSA levels and tumor burden – subsequent studies will need to demonstrate improvement in survival and quality of life measures. This is an issue that can only be addressed in future phase III studies. To that end, careful consideration must be given to how MDVN designs its phase III trial (assuming MDV3100 reaches that point). Even a well-designed trial, of course, does not guarantee that the drug will pass the challenges a good study will present. With that being said, MDVN maintains that the following preliminary plans would be executed in a phase III trial: 1) The trial will probably have two parallel studies: one of patients who have received previous chemotherapy, which would probably take 10-12 months to complete, and one of patients who did not receive previous chemotherapy, which would probably take 14-16 months to complete; 2) the study will measure not just PSA levels, radiographic and clinical evidence of tumor burden, and adverse events, but will also measure overall survival; and 3) the trial will probably measure quality of life standards, specifically bone pain and analgesic use. If, in fact, MDVN follows through on these comprehensive measures AND the study drug succeeds, this would likely satisfy urologists and oncologists who treat prostate cancer.