Photo by Theresa Naujack
A simple and reliable blood test that can detect early-stage disease might have saved the lives of the more than 30,000 American men who will die this year of prostate cancer.
But no accurate, highly predictive test is available for the most commonly occurring cancer in men other than skin cancer, leaving scientists eager to develop a method that is sensitive, definitive and easy to administer.
With that goal in mind, investigator Dr. Peter Nelson and colleagues in the Human Biology Division have set out to identify the molecular warning signals that tell doctors when patients are at the earliest stages of disease. In addition to their role in diagnosis, such molecules may prove to be targets for cancer-fighting drugs.
Nelson, who joined the Hutch faculty in 1999, said one of his lab's goals is to find molecular markers that don't have the limitations of the widely-available prostate-specific antigen (PSA) test, which often can give a positive result for benign conditions as well as cancer. PSA is a protein produced by the prostate that typically is present in elevated amounts in patients with prostate cancer.
"An ideal marker is highly specific and sensitive," said Nelson, who is also an assistant professor at the University of Washington. "PSA is not specific for cancer. It can be elevated in men who have prostatitis or benign prostatic hypertrophy."
In addition, Nelson said, many prostate cancers grow slowly and never will impact the life of the host. Other prostate cancers grow rapidly and are quickly lethal.
"The PSA test does not discriminate between these forms, and thus we are simultaneously overtreating and undertreating these patients," he said. "The ideal tumor marker for prostate cancer would indicate which cancers are lethal, and we would treat them even more aggressively than we do, knowing their high risk potential.
"For patients with non-aggressive cancers, we would simply put patients on a program of surveillance or intervene with dietary changes and thus spare them the morbidity associated with surgery or radiation therapy."
Nelson reasoned that to look for new markers in a vast sea of proteins required limiting the scope of the search to a subset of molecules containing the best candidates. He came up with two criteria for his marker search: the proteins should be uniquely present in prostate tissue, and protein levels should be modulated by hormones known to affect prostate development.
PSA, while an imperfect marker, is an androgen-regulated protein, although it is not is not exclusively present in prostate tissue, Nelson said.
"Both normal prostate and tumor tissue is exquisitely sensitive to androgens, a class of hormones that includes testosterone," he said. "We want to understand the network of genes involved in androgen response in the prostate."
Depriving prostate tumors of androgen can be an effective therapy for some forms of cancer, Nelson said.
"If you remove androgen, most prostate tumor cells die," he said. "About 90 percent of cancers initially respond to this treatment. But in most cancers, there are a small number of cells that do not depend on androgens for growth, or the tumor cells adapt to a low androgen state over time and become androgen-independent."
Eventually these androgen-independent cells grow to become the majority of the cancer, and at that point no therapies to date have prolonged a patient's survival.
To identify new prostate cancer markers and potential drug targets, Nelson uses a large-scale genomics approach he initiated as a research fellow at the UW with Dr. Leroy Hood, now at the Institute for Systems Biology in Seattle.
Most of that work seeks sets of genes that are regulated differently in normal, pre-cancerous and cancerous prostate tissue and presents this information in a user-friendly format, Nelson said.
"We called this resource the Prostate Expression Database or PEDB, and it can be used as a virtual tool to analyze gene expression," he said. "The resource is housed at the Hutch with a Web site utilized by other academic and government laboratories as well as pharmaceutical companies."
Nelson has used this resource to design DNA arrays, devices for analyzing expression of thousands of genes simultaneously, which contain prostate-specific genes. This work proceeds in conjunction with the Dr. Jeff Delrow, manager of the Hutch DNA array facility.
To carry out experiments, Nelson's group grows normal and cancerous prostate cell cultures in the presence or absence of androgen. Genes that are active will be copied into intermediate molecules called RNA, which is collected from the cells. Nelson mixes the RNA with the array slides to identify the subset of genes on each slide that were active in each growth condition.
"We found 10 genes that are expressed almost exclusively in the prostate and are regulated by androgen in a manner very similar to PSA," he said. "These may be potential markers for early detection or therapeutic targets."
Two of the genes Nelson's lab has identified contain information for synthesizing two proteins - prostase and TMPRSS2 - that are closely related in structure to PSA.
"Another protein we've identified is PSDR1, which stands for prostate short-chain dehydrogenase reductase 1," he said. "Based on its similarity to other proteins, we think it is involved in steroid metabolism, and androgens are one type of steroid. Our hope is that PSDR1 is involved in androgen metabolism within prostate cells, and this would make PSDR1 a potential drug target."
Further analysis of PSDR1 requires biochemical analysis, and Nelson will be collaborating with UW researchers to synthesize the protein in the lab for further studies.
Until now, Nelson's research has used tissue-cultured prostate cells, which sometimes behave differently than intact tissue. Future work will begin to apply the same genetic technology to analysis of normal and tumor tissue samples, which would be provided by collaborators in the UW department of urology.
As a practicing oncologist at the Seattle Cancer Care Alliance, Nelson hopes that ultimately his research will have a payoff in the clinic.
"About 90 percent of the patients I see have prostate cancer," he said. "What we'd really like is to be able to translate discoveries from the laboratory to patient care."