Clinical and laboratory investigations of castration-resistant prostate cancer indicate that researchers may be near to identifying subtypes of the disease.
"Unlike many other cancers, we don't yet have a clinically relevant scheme to subclassify our patients," said Charles L. Sawyers, MD, of Memorial Sloan-Kettering Cancer Center (MSKCC) at the 101st Annual Meeting of the American Association for Cancer Research (AACR), held April 17-21 in Washington, DC. "Are there different subtypes for prostate cancer? The precedent is clear in breast cancer, which is now divided into at least three major subgroups." However, he noted, transcriptional studies in prostate cancer have yet to yield the robust findings seen in breast cancer.
Toward Subclassification
Dr. Sawyers is co-inventor of MDV3100, Medivation, Inc's oral androgen receptor antagonist. He focused much of his presentation on advances toward subclassification that have emerged from research on that drug, which could be a model for differentiating subtypes of the disease.
Last year, Dr. Sawyers and colleagues reported that after conventional hormonal treatments fail in prostate cancer, androgen receptor transcriptional activity remains crucial to tumor growth and survival. They also showed that histone deacetylase (HDAC) inhibitors, including vorinostat (Zolinza) and LBH589, now in clinical trials, could block androgen receptor activity in castration-resistant prostate cancer models.1
More recently, his team has mapped DNA binding to androgen receptors across the genome, which revealed a reduced binding by about 1 log for MDV3100 compared to bicalutamide (Casodex), the standard treatment.
Key Clinical Trials
Three days before Dr. Sawyers' AACR presentation, The Lancet published online results of a dose-escalation phase I/II MDV3100 study, which showed positive results in patients with castration-resistant prostate cancer. Among 140 men--75 in whom chemotherapy failed and 65 chemotherapy-naive--researchers reported ≥ 50% reductions in PSA levels in 56% of the men. Median PSA responses lasted 41 weeks in chemonaive patients, 32 weeks for all participants, and 21 weeks for postchemotherapy patients.2
The 1,200-patient AFFIRM trial--a randomized, double-blind, placebo-controlled phase III study of MDV3100--is underway in the United States, the United Kingdom, and 13 other countries.
Preclinical Findings
Clues to distinguishing known cancer subsets have come from disease tissue. Prostate cancer poses an obstacle to this approach because advanced disease largely manifests as bone metastases, which are difficult to access. Dr. Sawyers' laboratory has used preclinical models in efforts to identify subsets. He cited several recent findings from his team that might lead to differentiating types of prostate cancer:
Researchers compared two genetically engineered mouse models. In one, the Myc gene is expressed under control of the probasin promoter. Castration produced a partial response, but combining MDV3100 with castration yielded a dramatically stronger response. "This model is extremely sensitive," Dr. Sawyers said. "In contrast, the Pten deletion model showed minimal response."
After restoring the damaged PI3K Chinese signaling pathway to baseline, mice treated with MDV3100 had dramatic disease regression--at least 80% in the four of them treated so far.
The finding indicates that loss of the pathway could serve as a biomarker for resistance and that restoring PI3K activity might restore androgen-sensitive dependence to such tumors.
Alterations in DNA copy numbers potentially could provide useful predictive factors. In human tumors, the MSKCC team found high copy-number alterations in the NCAO2 gene, whose homolog is amplified in breast cancer. These changes occurred in 21% of primary tumors.
Using this approach in human prostate tumors, Dr. Sawyers and his colleagues identified two patterns-cluster 2 with few alterations, and cluster 5 with many changes. They compared the two subsets by time to disease progression.
"Cluster 2 did well out to almost 100 months, with 90% still in remission, vs cluster 5 doing very poorly," he said. "If this can be consistently reproduced, there may be information in genome copy numbers in prostate cancer that can be prognostic."
References
1. Welsbie DS, Xu J, Chen Y, et al: Histone deacetylases are required for androgen receptor function in hormone-sensitive and castration-resistant prostate cancer. Cancer Res 69:958-966, 2009.
2. Scher HI, Beer TM, Higano CS, et al: (2010). Antitumour activity of MDV3100 in castration-resistant prostate cancer: A phase 1-2 study. Lancet 375:1437-1446, 2010.
