At the 2014 AACR Annual Meeting held in San Diego, CA this spring, Prostate Cancer Foundation-funded investigators discussed recent progress and exciting new studies regarding prostate cancer biology and treatment.

May 29, 2014 — Here, we highlight some of the most important research that was presented, including novel ways to combat drug resistance to hormone therapy, combining drug treatments to improve patient response rates, and the discovery of master gene regulators of prostate cancer.

New therapies mean new challenges in drug-resistance

Dr. Charles Sawyers, President of the American Association of Cancer Research

Dr. Charles Sawyers, the outgoing President of the American Association for Cancer Research (AACR), focused his keynote speech around two critical issues. One: in this era of new prostate cancer drugs with greater potency, researchers are observing increased variety in how cancer cells develop drug-resistance. Two: to accelerate the discovery of therapies that overcome these new drug-resistance mechanisms, researchers need to increase their sharing of genetic and clinical data.

The androgen receptor (AR) is the main molecular factor that drives continued prostate cancer growth and survival, and targeting AR — including the androgen hormones that activate AR — is the primary therapeutic strategy for the treatment of this disease. Primary treatment with androgen-deprivation therapy (ADT) leads to tumor regression. However, recurrence is all too common, as tumors evolve to grow in spite of treatment. In about 90 percent of these cases, primary ADT-resistance is conferred by upping the amount of AR expressed by the tumor, by 300 to 500 percent.

In 2012, two new drugs were FDA-approved for treatment of prostate cancer patients whose tumors have become resistant to primary ADT: Zytiga (abiraterone acetate) and Xtandi (enzalutamide). Both drugs target different points along the AR-axis, and are potent enough that they have clinical activity in this treatment-resistant setting (primary ADT-resistance). Nevertheless, tumor resistance to these drugs too, is inevitable, via “secondary” escape mechanisms that include roundabout ways to regain AR-activities (by either AR-gene mutations or AR-like molecules such as the glucocorticoid receptor stepping in to take over tumorigenic activities of AR), or reprogramming of the tumor cells into alternate cell types that don’t rely on AR to grow and survive. This latter mechanism could involve a complete change in tumor cell phenotype (transdifferentiation) into neuron-like cells, or epithelial-mesenchymal transition (EMT), a process where epithelial prostate cancer cells reprogram to become a mesenchymal cell type with more highly invasive and metastatic abilities.

Many other, more rare, drug-resistance mechanisms occur, and Sawyers emphasized how imperative it is to identify these, so that new therapies can be developed and/or appropriately prescribed. As an example of the power that studies of this type can give to patients and their physicians, Sawyers described the recent identification of a mutation in AR that causes Xtandi to activate instead of suppress AR, thus Xtandi treatment promoted growth of tumors with this mutation. Understanding how Xtandi physically interacts with normal and mutant forms of AR allowed Sawyers and colleagues to design a new set of compounds that inhibit both forms of AR.

It is estimated that 100,000 tumors will need to have their genomes sequenced to identify mutations that occur in as few as 2 percent of tumors. To identify mutations that confer drug resistance, the genomes of tumor specimens from before and after the development of drug resistance, must be sequenced and compared. At least 250-1000 patients would need to be examined to discover mutations that cause drug resistance in 5 percent or less. Hand-in-hand with this effort is the need to identify biomarkers that can predict tumor response to various treatments. To identify biomarkers of response that occur at frequencies of 1-20 percent, potential biomarkers from 250-5000 patients would need to be examined. These efforts will require significant resources and expertise, and as a closing note, Sawyers brought to attention the Global Alliance for Genomics and Health, a growing alliance of 129 institutions from 20 countries, which are working to create an infrastructure and culture for sharing genomic and clinical data so that researchers can join together to reach these goals significantly sooner.

Understanding drug resistance and maximizing anti-androgen targeting to seek cures

Dr. Peter Nelson of the Fred Hutchinson Cancer Research Center

Dr. Peter Nelson, of the Fred Hutchinson Cancer Research Center in Seattle discussed additional studies on mechanisms of tumor resistance to Xtandi and Zytiga. Nelson is part of a team that runs a rapid autopsy program, in which prostate cancer patients who succumb to their cancer donate their bodies immediately after passing away, so that their lethal tumors can be excised and studied. In studying these tumors, Nelson found that drug resistance can sometimes arise from metabolic adaptation: where metastatic tumors evolve to produce high levels of androgens that maintain AR activity in spite of AR-axis inhibiting drugs such as Zytiga or Xtandi. Often, this occurs by enhancement in the levels of enzymes that metabolize androgens from precursor molecules such as cholesterol. Some patients were found to have a hereditary mutation in one of the most critical enzymes in this process (3βHSD) that causes it to become super-active and generate even more androgens, confirming a discovery previously made by PCF-funded Young Investigator, Dr. Nima Sharifi. Other tumors attain drug resistance by expressing short, variant forms of AR that do not need androgens to become activated, but are always “on.” When Nelson examined tumors from rapid autopsy patients, he found that if drug-resistant tumors did not have abnormally high levels of androgens (due to changes in enzymes or other reasons), they instead often expressed these AR-variants, as their mechanisms of drug resistance.

Nelson raised an important question: if researchers are able to develop drugs that can extinguish AR signaling completely, will we begin to see cures? One problem he highlighted is that in current treatment protocols, AR-targeting treatments–none of which cause 100 percent extinction in AR-activity–are used sequentially, which allows time for the tumor to develop resistance at each step. Concurrent treatment with multiple drugs that target different aspects of AR-activity might be more effective, and some promising results have been seen in clinical trials.

In a recent pilot clinical trial published in the Journal of Clinical Oncology, 35 intermediate to high-risk prostate cancer patients were randomly assigned to a combination of two, three, or four different AR-targeting drugs. Men who received all four drugs has substantially lower levels of androgens and lower PSA (a measure of AR activity), and they experienced more complete and near-complete responses than men receiving fewer drugs in combination. In another ongoing clinical trial, combination therapy of Lupron plus three vs six months of Zytiga was tested. In the six-month Zytiga trial-arm, 10 percent of patients had complete responses and 24 percent experienced near-complete pathological responses. These treatment strategies are very promising, but Nelson was only cautiously optimistic as patients can develop prostate cancers that are completely independent of AR, for instance by transdifferentiation into neuroendocrine cancers, as Sawyers had discussed. In these cases, is it important to understand what mechanisms cause prostate cancer cells to no longer need AR.

To discover such mechanisms that confer AR-independence, Nelson and colleagues performed a study that identified two genes, PP2R1A and PP2R2C, that when inhibited, allowed the growth of prostate cancer cells in the absence of androgens. Loss of PP2R2C was associated with more aggressive disease and mortality of prostate cancer patients. Nelson said, if tumors evolve to lose expression of these genes, they may gain ADT-resistance. “In the future, we can anticipate some cures by completely ablating the AR pathway, but that AR “bypass” resistance mechanisms are likely to emerge as well,” he concluded.

Studying mice side-by-side with humans to determine best medicines

Dr. Pier Paolo Pandolfi of the Beth Israel Deaconess Cancer Center

Dr. Pier Paolo Pandolfi of the Beth Israel Deaconess Cancer Center, MA, has created a “Co-Clinical Trial” platform, which takes the approach of studying mice side-by-side with human patients undergoing clinical trials to improve patient treatment plans. This is a three-pronged approach in which patients (arm 1) are enrolled into clinical trials for various therapies. Tumors are removed from patients and surgically implanted into immune-deficient mice (allowing human cells to grow in mice), and mice are treated with the same drugs and clinical trial format (arm 2) as the human patients. Concurrently, mutations in a patient’s tumor are determined, and mice that have been genetically engineered to have the same mutations and develop prostate cancer (arm 3), are also treated with the same clinical trial protocol. Because tumors grow much faster in mice than in humans, the results from these mice will be acquired sooner than patient responses can be measured. The mice therefore can act as surrogates for human responses. This fast-tracked response data can be used to guide the development of clinical trials for humans.

Dr. Pandolfi demonstrated his use of the co-clinical trial concept to study the “immune landscape” of the tumor. He is determining if different cancer cell mutations affect the types, numbers, and functions of immune cells that enter the tumor. Furthermore, because some types of immune cells kill tumors while other types promote tumor growth, this research will determine how the immune landscape affects disease progression and the effectiveness of therapies.

Over 30 different genetically engineered prostate cancer mouse models that lack or overexpress various human prostate cancer-associated oncogenes and tumor suppressor genes, are being studied to address these questions. Pandolfi presented studies on the role of ZBTB7A, a prostate tumor-suppressor gene that is deleted in a subset of advanced prostate cancers. He created mice genetically engineered to lack the ZBTB7A gene, and found that, when combined with a loss of the PTEN tumor suppressor gene, prostate tumors grew faster. The ZBTB7A gene suppressed the production of CXCL5, a cellular “attractant” molecule, so-called because it attracts a type of immune cell called “myeloid-derived suppressor cells,” (MDSC) into prostate tumors. These MDSC cells produce several tumor-promoting molecules, including S100A9, which activates the pro-oncogenic transcription factor NF-κB, which goes on to make more CXCL5, in a sort of vicious tumor-growth cycle. When mice were treated with an agent that kills MDSCs, the levels of these molecules were reduced and tumors shrank.

So, the loss of ZBTB7A in prostate cancer cells causes tumor-promoting immune cells (MDSCs) to be drawn into tumors and produce factors that cause activation of oncogenes that feed this tumor-promoting cycle even further. However, these effects were not seen in mice that instead lacked the p53 tumor suppressor gene, which had a different type of immune cell accumulate in their tumors. These studies demonstrate that different tumor suppressor molecules and oncogenes do indeed alter the types of immune cells that are attracted into tumors and thereby affect clinical outcome. Pandolfi pointed out that many steps in this processes can be targeted by drugs, including the MDSCs, CXCL5, and S100A9. An S100A9-inhibitor, Tasquinimod, is in phase 3 clinical trials for advanced prostate cancer. Further studies in this area will identify not only new drugs, but new methods–targeting immune cell types–for treatment of prostate cancer patients.

Discovering master regulators of prostate cancer in mice and men

Dr. Cory Abate-Shen, Columbia University

Dr. Cory Abate-Shen, of Columbia University, has also been studying prostate cancer patients along with genetically engineered mice to discover genes that are master regulators of prostate cancer in both species. To do this, Abate-Shen and her group assessed the expression of genes in 185 normal human prostate, and primary and metastatic prostate cancer samples, and from 384 samples from mice with different tumor-suppressor and oncogene mutations that develop a spectrum of prostate cancer stages. This included mice that had been treated with various drugs.

The team created high-level computational algorithms to evaluate this gene expression data. This research determined that most prostate cancer genes in mice and humans are not exactly the same, but their activities are highly similar. Still, a number of genes that play a role in prostate cancer in both species were identified. These included oncogenic and tumor-suppressor alterations in genes known to drive prostate-cancer in humans: ERG and AR, and the common tumor-suppressor protein p53. Also, two genes, FOXM1 and CENPF, have been identified as a pair of molecular regulators that synergize, or work together, to promote tumor growth. Inhibiting both FOXM1 and CENPF in mice repressed tumor growth and killed tumor cells. Individually, both of these genes promote division of cells, while synergy occurs because CENPF directs FOXM1 to the place it needs to be on DNA, in order to turn on the expression of genes that promote prostate cancer. Importantly, alterations in these genes were found to co-occur in about 25 percent of prostate cancer patients. Abate-Shen and colleagues analyzed the expression of CENPF and FOXM1 in primary tumors from a cohort of 900 patients who had been followed for 20 years after diagnosis and treatment. She found that levels of these genes predicted patient survival and time to metastasis both individually, and in synergy. The rich amount of data generated from this large study of prostate cancer mouse models with different mutations and with different genetic backgrounds is being used to identify drugs that may inhibit the tumorigenic effects of CENPF and FOXM1. This could lead to clinical trials with new drugs for patients that have aberrations in CENPF and FOXM1.

Overall, this is a highly promising time in prostate cancer research. Scientists are moving forward along many avenues, and with new resources, technologies, and methodologies, this AACR meeting left the prostate cancer research community very hopeful that cures are in sight.

Dr. Andrea Miyahira has a PhD in cancer immunology, and is Manager of Scientific Programs at the Prostate Cancer Foundation.