Year 9: Pilot Projects

Predicting response and resistance to AR-axis inhibition in castration resistant prostate cancer

Elahe Mostaghel, MD, PhD, Fred Hutchinson Cancer Research Center

Castration resistant prostate cancer (CRPC) is variably characterized by elevated tissue androgens, expression of steroidogenic enzymes capable of mediating androgen biosynthesis, persistent to elevated androgen receptor (AR), and the recently described expression of constitutively active AR splice variants. The extent to which therapeutic efficacy of agents targeting the AR axis is unknown.

Hypothesis: We hypothesize that differences in androgen levels, the steroidogenic transcriptome and androgen receptor/splice variant expression underlie tumor-specific differences in response and resistance to agents targeting the AR axis in CRPC. We hypothesize that baseline differences in these parameters can be used to determine optimal treatment regimens.

Aim 1. Establish the ‘castration recurrent profile’, including tumor androgens, steroidogenic enzyme expression and AR/variant levels in a panel of LuCaP xenografts spanning a range of baseline differences in these parameters.

Aim 2. Determine whether response and mechanisms of resistance to the AR pathway inhibitors VN-124, dutasteride and MDV3100 correspond to those predicted by tumor-specific differences in AR axis parameters.

If successful, this proposal will provide a strong impetus to determine the ability (and feasibility) of tissue biopsy or CTC-based methods of tumor sampling to yield real-time pre-treatment data profiles. Importantly, our data indicate that individual CRPC tumors appear to maintain a tumor-specific, functionally relevant and identifiable AR and steroidogenic transcriptome, suggesting that valid data indicative of inherent steroidogenic potential and treatment response can be obtained via in situ sampling of human tumor metastases.

Alternative RNA splicing of the androgen receptor gene in prostate cancer

Xuesen Dong, MD, MSc, PhD, University of British Columbia, Vancouver Prostate Centre

It is widely accepted that AR becomes more sensitive to its ligands and remains ligand-activated in castration-resistant prostate cancers. This concept directs the research to design stronger AR antagonists or androgen synthesis inhibitors, to achieve better outcomes for cancer  patients. However, we identified a novel AR splice variant, ARv567es, which constitutively actives AR targeted genes without androgen binding. It is produced by alternative RNA splicing that skips the 3’ splice site of intron 4 of AR pre-mRNA, resulting exon 5-7 deletion in mature AR mRNA. Castration induces ARv567es levels. ARv567es bearing tumors progress despite of castration. Therefore, developing therapeutic strategies for castration-resistant cancers should consider blocking ARv567es formation, which requires understanding of AR alternative splicing program. We recognize there are several AR variants with similar functions. However, we chose ARv567es because it occurs with a high frequency in clinical tumors and is the only variant that is clearly regulated by tumor androgen levels.

Alternative splicing is primarily processed by SR and hnRNP splicing factors. Once recruited, they determine a splice site to be excised. We hypothesize that castration changes SR/hnRNP recruitment to the 3’ splice site of intron 4 leading to ARv567es formation. We propose to apply RNA binding and CHIP techniques to 1) define SR/hnRNP recruitment to the 3’ splice site in pre- and post-castration conditions, and 2) compare the recruitments of SR/hnRNP proteins to the 3’ splice site between ARv567es positive and negative tumors. Information obtained from these studies would lead to novel treatments for castration-resistant prostate cancers.

Metformin and prostate cancer specific outcomes

Jonathan L. Wright, MD, University of Washington School of Medicine

Metformin is a commonly used medication in the management of type II diabetes mellitus. Metformin has received considerable attention recently as a potential chemopreventive agent for cancer progression. Although there is growing evidence of an anti-neoplastic effect in prostate cancer (PCa), the potential effect on PCa outcomes following primary treatment has not been well described.

We hypothesize that metformin use is associated with a reduction in disease progression after primary treatment for localized PCa. To investigate this hypothesis, we will use the Veterans’ Affairs VISN20 Data Warehouse, which is a comprehensive electronic database of all medical records, including pathologic, pharmacy, vital measures and laboratory results for over 4500 men diagnosed with PCa between 2001 and 2009 in the seven VA medical centers in the Pacific Northwest. We propose the following specific aims:

  1. To determine differences in the rate of biochemical progression for men treated and untreated with metformin among cohorts of newly diagnosed patients undergoing either radical prostatectomy or XRT.
  2. To explore the feasibility of using the national VA electronic medical records for low cost long-term follow-up of PCa outcomes and unexpected adverse effects for conducting a future phase III trial of metformin among newly diagnosed PCa patients.

This study will provide epidemiologic evidence for metformin as a secondary prevention for men with PCa. These results will serve as pilot data for a study using the national VA database to study PCa-specific mortality and metformin and also for future human interventional trials.

Evaluation of neuroligin 1 as a biomarker of prostate cancer aggressiveness

Eva Corey, PhD, University of Washington

Recurrent prostate cancer (PCa) progresses from androgen-sensitive tumors that respond favorably to androgen ablation to castration-resistant tumors that grow even under androgen ablated conditions. The response to androgen ablation varies between patients and at present, there are no biomarkers to determine the magnitude and duration of this response. Identification of biomarkers that distinguish between tumors that will respond well to androgen ablation and tumors that would not, is one of the highest priorities in research and clinical communities, and would have huge impact for PCa patients.

Our proposed studies are based on a serendipitous finding that PCa xenografts that do not respond well to castration express high levels of neuroligin 1, while xenografts that shrink after castration express low to no neuroligin 1 mRNA. Our hypothesis is that neuroligin 1 is a biomarker of responses to castration and might play a role control of prostate tumor cell responses to androgen withdraw.

Objectives of our proposal are:

  1. to determine whether neuroligin 1 is a marker of PCa responses to castration using tissue microarrays containing patients’ samples and PCa xenografts.
  2. to investigate the role of neuroligin 1 in the response of PCa tumors to castration in animal model.

Results of our studies will provide information whether neuroligin 1 distinguishes PCa tumors that respond well to androgen ablation from those that do not, and need more aggressive treatment. Patients with responsive tumors could be spared the negative side effects, while those with androgen ablation nonresponsive tumors would benefit from more aggressive effective treatments. Confirmation of our hypothesis will have highly significant impact on treatment strategies in PCa. In addition, results of our studies might be used as preliminary data in application for further funding to delineate mechanisms of development of castration-resistant PCa and biological roles of neuroligin 1 in PCa progression.

Identification of DNA methylation markers of prognostic value in prostate cancer

Min Fang, MD, PhD, Fred Hutchinson Cancer Research Center

The heterogeneous nature of prostate cancer necessitates new methods of risk stratification for optimal therapeutic decisions. Changes in DNA methylation patterns in prostate cancer presents a possible means for risk stratification. Localized prostate cancer is characterized by DNA hypermethylation of specific genes, whereas metastases display global hypomethylation, suggesting that genome-wide methylation patterns change during prostate cancer development. Currently, no well-defined DNA methylation markers exist to discriminate high-risk prostate cancer characterized by a strong propensity for lethal recurrence and metastasis from low-risk tumors that remain indolent long-term. This pilot study aims to identify DNA methylation markers that stratify patients into specific prognostic risk categories. We adopted a new DNA methylation profiling method, termed comprehensive high-throughput array-based relative methylation analysis (CHARM), which provides accurate and quantitative genome-wide DNA methylation assessment. The advantage of CHARM over existing methylation profiling methods is its greater sensitivity for differential methylation obtained from CpG dinucleotides located within 2kb from CpG islands, compared to examining methylation of CpG islands themselves. We have successfully used this method to discriminate high-risk from low-risk acute myeloid leukemia. Here, we propose to perform CHARM on 10 samples each from prostate cancer patients with Gleason scores above and below 6, respectively, along with 10 age-matched normal controls. We will further validate the top differentially methylated regions identified using quantitative pyrosequencing. This pilot study promises to identify clinically useful epigenetic markers for prognosis. The results will also contribute to ongoing integrated studies with DNA copy number profiling, RNA expression profiling, and exon sequencing studies led by other SPORE investigators on mechanisms of prostate tumorigenesis and metastasis.

Mechanisms of ERG-mediated PC development

Valeri Vasioukhin, PhD, Fred Hutchinson Cancer Research Center

A significant proportion of human prostate cancers carry a chromosomal rearrangement resulting in the overexpression of the ETS transcription factor ERG. Multiple experiments in cell lines and mice demonstrated an important role of ERG in prostate cancer formation and progression; however, the molecular mechanisms of ERG in prostate cancer are not well understood. We present here confidential, preliminary data indicating a functional interaction between ERG and the Hippo signal transduction pathway.

We propose to investigate the role of the Hippo pathway in prostate cancer and analyze the significance of the changes in the Hippo pathway signaling in ERG-mediated prostate cancer formation and progression.