Researchers build a 'super' antiandrogen

NEW YORK—Charles Sawyers, MD, head of the new Human Oncology and Pathogenesis Program at Memorial Sloan-Kettering Cancer Center, is perhaps best known for his kinase inhibitor research leading to the development of imatinib (Gleevec) and dasatinib (Sprycel), drugs of unprecedented benefit for patients with chronic myelogenous leukemia.

Yet he began his career investigating hormone-refractory prostate cancer (HRPC), and prior to joining MSKCC was director of the Prostate Cancer Program Area at UCLA's Jonsson Comprehensive Cancer Center. Now, in a preclinical animal model of androgen-resistant HRPC, he has employed molecular approaches similar to those he used to study acquired resistance to targeted therapy of CML—and has made a profound discovery with the potential to change antiandrogen treatment of prostate cancer.

At the Chemotherapy Foundation Symposium XXV, in a special session entitled "The 21st Century Cancer Revolution," Dr. Sawyers shared provocative findings from a mouse xenograft study of RD162, the parent compound of MDV3100 (Medivation, Inc., San Francisco), a novel small molecule targeting the androgen receptor (AR) that is now under investigation in a multicenter phase I-II clinical trial.

The new agent is truly a next-generation antiandrogen, he told ONI: It only functions as an AR antagonist, in contrast to the antagonist-agonist conversion known to develop in men treated with bicalutamide (Casodex), the standard antiandrogen drug. And it shuts down AR activity not via increased AR binding affinity, but by actually changing AR conformation so it cannot bind to DNA.

Current understanding of HRPC, based on work in animal models, Dr. Sawyers said, is that it is still AR dependent: Knockdown of the AR impairs tumor growth, and HRPC progression occurs with AR overexpression.

Recent investigations by his group and others into the molecular basis for restored function of the AR in HRPC xenograft models suggests it results mainly from increased AR messenger RNA and protein levels (80% of cases), via receptor amplification or other mechanisms, and from AR gene mutation (10% of cases).

The findings, he added, are consistent with data from human samples showing higher AR levels in men with hormone-refractory vs hormone-sensitive cancer.

"Another part of the story that is less widely known is that bicalutamide has partial agonist properties," Dr. Sawyers said. "This came out in our study when we showed in model systems that bicalutamide would induce the expression of genes like PSA and several other target genes in systems where the androgen receptor level was higher, as we see in hormone-refractory disease."

Clearly, he said, second-generation antiandrogens must be effective in patients with tumors expressing high AR levels, but also "must not have the liability that bicalutamide has—that is, this antagonist/agonist conversion phenomenon."

When mutagenesis studies he and others conducted in animal models confirmed that the ligand binding domain still needs to be intact for AR function in HRPC, "that was a very useful and 'good news' finding because it meant the most well understood part of the molecule in terms of drug development was still relevant to get a drug that might function in patients at this advanced stage," he said.

Searching for candidates

At this point in the project, Dr. Sawyers enlisted the expertise of colleague Michael Jung, PhD, professor of chemistry at UCLA, who searched the medical and patent literature to identify potential candidates for testing.

They employed chemical structural modeling and cell-based screens to find molecules that not only had high binding affinity for the AR but also had structural groups that conferred stability and enabled them to function as antagonists. More than 200 compounds were screened over a period of several years.

A "chemical scaffold" was identified for a candidate compound with approximately 10-fold greater potency than bicalutamide, Dr. Sawyers said.

This scaffold structure was further optimized to identify a compound with high antagonist and no agonist properties that could be delivered orally in the xenograft model of HRPC, with enhanced absorption and an extended half-life.

Samuel Waxman, MD, who moderated the meeting session, told ONI that this novel approach for androgen deprivation is reminiscent of that of fulvestrant (Faslodex), which targets the estrogen receptor for degradation and loss of function.

"Both have the advantage of not being a receptor agonist or cross-resistant with prior therapies that decrease the production of hormone," said Dr. Waxman, Distinguished Service Professor, Mount Sinai School of Medicine, New York, and founder and scientific director of the Samuel Waxman Cancer Research Foundation.

In HRPC models in which bicalutamide is completely inactive, the identified compounds, RD162 and its clinical analog MDV3100, had half-lives in serum comparable to bicalutamide, and both inhibited PSA release from LNCaP cells engineered to express high AR levels.

In the HRPC mouse model, MDV3100 significantly impaired growth of LNCaP xenografts expressing high levels of AR: At the end of the 28-day treatment period, tumor volumes in mice treated with MDV3100 at 50 mg/kg orally daily were reduced by a mean of 68% from baseline (P < .01) (see Table).

In contrast, bicalutamide was inactive; tumor volumes in mice treated with bicalutamide at 50 mg/kg daily decreased modestly in the first 16 days of treatment, but then climbed steadily through the remainder of the treatment period for a 37% increase from baseline.

In addition, three of seven animals treated with the MDV3100 had nonmeasurable tumors at the end of the study period, compared with zero of seven animals treated with bicalutamide, Dr. Sawyers reported. MDV3100 was well tolerated, and mice treated with the agent gained weight while on study.

RD162, the parent compound to MDV3100, appears to have antiproliferative activity, as indicated by its "dramatic" reduction of Ki67 staining in the mouse xenograft model of HRPC. It has apoptotic activity also, as measured by TUNEL staining (see Figure 1), and significantly reduced tumor size in the mouse model, Dr. Sawyers reported.

He added that RD162, unlike bicalutamide, blocked AR function in an in vivo imaging study in the LNCaP mouse xenograft model of HRPC. In this model, he and his research team used an AR-driven PSA–luciferase reporter gene to "provide a pharmacodynamic readout of androgen receptor action," he said (see Figure 2).

Surprisingly, fluorescence polarization studies have revealed that the 10-fold increased potency of RD162 compared with bicalutamide is not a result of superior binding affinity for the AR; the binding affinities are comparable.

"The answer seems to be something that we're really quite excited about, and that is an alteration in the conformation of the AR that is induced upon binding to RD162," such that the AR is unable to move to the nucleus and bind DNA, Dr. Sawyers said. This conformational change in the AR, reducing the efficiency of its binding to target genes, was inferred in chromatin immunoprecipitation assays his group conducted, he explained.

Dr. Sawyers said he is unaware of any other antiandrogen with this novel mechanism of action.

John Wright, MD, PhD, of CTEP, told ONI that the results suggest MDV3100 "ultimately might benefit patients in earlier 'hormone sensitive' stages of the disease without necessitating androgen ablation." He noted that AR blockade "is one of several avenues being evaluated to disrupt this target that include inhibition of siRNA, small-molecule inhibitors of AR chaperone proteins, and more effective compounds to decrease levels of testosterone, the AR ligand."

Early clinical results

Toxicology studies of MDV3100 have been completed, Dr. Sawyers said, and a Medivation-sponsored multicenter phase I-II clinical trial of the agent in castration-resistant prostate cancer is now under way at MSKCC (Howard I. Scher, MD, PI), University of Washington (Celestia S. Higano, MD), and Oregon Health Sciences University (Tomasz Beer, MD).

In November, Medivation announced early clinical data showing that MDV3100 was well tolerated and reduced PSA levels in a dose-dependent manner in six patients with HRPC who were enrolled in the first two dose groups of the ongoing dose-esclation phase I-II trial.

In the lowest dose group, PSA levels declined by 45% to 66% after 2 months of treatment with the agent. In the second lowest dose group, PSA levels declined by 75% to 89% after the first month of treatment.

Study completion and a final report of the results are expected in 2008, according to Medivation.

Dr. Sawyers stated that he is a consultant to Medivation and holds stock in the company.