By David N. Leff

Most American men diagnosed with prostate cancer do not die of the disease. The reason? Simple demographics.

By the time that diagnosis is confirmed, the patient is so advanced in years that other geriatric risk factors are likely to outpace prostatic carcinoma (PC) as the cause of death. Nonetheless, PC is the most frequently diagnosed cancer among males in the U.S. It will take the lives of 40,000 in a typical year, and three or four times as many will be diagnosed with the disease annually.

For the past 15 years or so, that diagnosis has relied heavily on a protein in the blood called PSA ¿ prostate-specific antigen. Prostate cancer mortality has declined substantially since 1990, and many urologists attribute this decline to the ability of PSA screening to detect cancer at a curable stage. Yet the contrary is also likely to be true. PSA often proves to be a false prophet.

¿It can very often be elevated in noncancerous conditions such as BPH ¿ benign prostatic hyperplasia,¿ observed pathologist Arul Chinnaiyan, at the University of Michigan, Ann Arbor. ¿And prostatitis ¿ inflammation of the prostate ¿ could lead to increased PSA in the blood,¿ he added. ¿Different tests, such as free PSA and bound PSA, help distinguish whether it¿s associated with cancer or not.¿

Levels of PSA below 4 nanograms per deciliter are considered normal, whereas readings above 10 ng/dL are strongly indicative of PC. Some 30 percent of patients with PSA levels between these limits will have prostate cancer detectable by biopsy within one year.

¿But in a high percentage of cases,¿ Chinnaiyan pointed out, ¿there is a pretty high false-positive rate in PSA testing. A leading cofounder is BPH, a non-malignant enlargement of prostatic tissue, with many of the same signature symptoms as PC ¿ especially urinary blockage and incontinence. BPH is very frequent,¿ he continued. ¿Almost all older men eventually develop the ailment. It¿s sort of an overgrowth, but there¿s no link between BPH and cancer.¿

Chinnaiyan is senior author of a paper in today¿s issue of Nature, dated Aug. 23, 2001, and titled: ¿Delineation of prognostic biomarkers in prostate cancer.¿

9,984 Elements On One Microarray Chip

¿We were able to molecularly profile a number of prostate cancer tissue specimens,¿ he told BioWorld Today. ¿We looked at about 50 clinical samples, ranging from localized PC to benign conditions all the way up to hormone-refractory metastatic prostate cancer.

¿Using cDNA microarrays displaying 9,984 elements,¿ Chinnaiyan recounted, ¿we were able to classify them based on their gene expression. These DNA chips enabled us to identify 200 or so genes that we feel are significantly associated with some version of prostate or form of prostatic cancer ¿ whether metastatic or localized.

¿We purchased these sequence-verified clones ¿ essentially a random smattering of human cDNA clones ¿ from a library. They represented a fraction of the recently sequenced human genome. So if we assume that the expression of the genome might be 35,000 genes, this would be just a fraction of that number.

¿In this particular study we selected two of the genes to characterize further, namely hepsin and pim-1. Then we looked at the protein level of these genes using high-density, clinically stratified tissue microarrays. These were laid out with hundreds of patient specimens on a glass microscope slide, which could then be profiled in one shot, using an antibody against the two candidate genes that we identified by DNA microarray.

¿One happened to be hepsin, a transmembrane cell protease previously unknown to prostate cancer; the other, pim-1 kinase, was an oncogene never really linked to carcinoma per se. We were then able to show at the proteomic levels that these molecules were upregulated, using arrays in over 100 samples. Even more important, we were able to show that the levels of these proteins could be correlated with a clinical outcome. In this case, the outcome we were measuring was PSA recurrence after prostatectomy. It¿s actually a bad outcome if the PSA level increases after such excisional surgery. So these proteins could be correlated with patient prognosis.¿

The team¿s microanalysis consisted of two parts:

¿The first part was the cDNA microarrays,¿ Chinnaiyan related. ¿We had those thousands of probes on one chip. The second part,¿ he continued, ¿was the tissue microarray, where we had hundreds of elements of tissue from different patients, all on a single microscope slide. Normally, the pathologist might look at one big tissue section from one specific patient. But in this case, we analyzed small, 6-millimeter-sized sections from hundreds of patients, all arrayed on this slide.

¿The University of Michigan tissue bank supplied us with fresh specimens from normal patients and prostate cancer prostatectomy patients. There were 700 to 800 specimens, which when profiled proved to express either hepsin or pim-1. That accounted for about 150 patients replicated approximately six times.¿

Hepsin Protease: Dark-Horse Prostate Biomarker

¿Hepsin in our gene selection criteria came out as being very significantly correlated with prostate cancer. All of our PC specimens ¿ either metastatic or localized ¿ had high hepsin expression. So that was one criterion. And the fact that it is a transmembrane cell surface protease makes it a prostate therapeutic target possibility. Hepsin resides on human chromosome 19 ¿ a very hot genomic neighborhood. It¿s at the top of our list of 200 or so genes. Pim-1,¿ he added, ¿is a serine-threonine kinase identified initially as an oncogene. So our study implicates it, too, in prostate cancer.

¿Our goal,¿ Chinnaiyan observed, ¿is to identify biomarkers that supplement PSA, that will help in confirming a prognosis for an individual patient. Both hepsin and Pim-1 may contribute. PSA will not be able to tell the urologist if somebody¿s prostate cancer is indolent or aggressive. Once a patient has metastatic disease, we don¿t really have a treatment, so the idea is to catch him before the cancer metastasizes. If it¿s aggressive, you need to treat that patient very rapidly. If it¿s indolent, you can watchfully wait. Besides supplementing PSA, hepsin or Pim-1 could theoretically be used on prostate biopsy specimens, to help guide patient prognosis. Hepsin itself has a separate clinical potential. Since it¿s a protease,¿ he concluded, ¿inhibitors or antibodies could be created against it.¿