CAG spells out course of prostate cancers.

From a boy's first whiskers to the first sparse hair follicles on his chest, hormones known as androgens spur the development of sexual characteristics that signal approaching manhood. Decades later, these hormones may fuel the growth of prostate cancer.

Researchers are beginning to understand why. A new study demonstrates that a peculiar pattern of DNA in the genes that encode androgen receptors may govern whether a man's cancer turns deadly or stays relatively benign.

The pattern is a short, repetitive DNA sequence, a so-called CAG repeat, named for the three DNA building blocks that form it. Investigators had previously found CAG repeats in androgen receptor genes on a man's X chromosome.

Researchers have long recognized that the interplay between androgens and their receptors on cell surfaces powerfully influences cell function. When an androgen latches onto a receptor on a prostate cell, for instance, the receptor reacts by prompting the cell to divide. If the cell is a cancer cell, replication may accelerate wildly.

In most men, prostate tumors grow almost imperceptibly, erupting into metastatic monsters in just 10 percent of cases.

Researchers have now found that men with 18 or fewer CAG repeats on their androgen receptor genes are more likely to get prostate cancer than men with stretches of 26 or more. In addition, the tumors of men with fewer repeats are more than twice as likely to turn into deadly malignancies as the tumors of men with more CAG repeats.

"This, to me, is a very fascinating thing biologically," says Philip W. Kantoff of Harvard Medical School in Boston. "As the CAG sequences get shorter and shorter, your risk of developing prostate cancer gets higher and higher. And they are a determinant of how aggressive [the cancer] will be."

Kantoff and his colleagues used blood samples collected from 1,175 men in the Physician's Health Study. About half of the men had developed prostate cancer between 1982 and 1995. The remainder served as a control group. "Men with shorter repeats were at particularly high risk for . . . metastatic and fatal prostate cancer," the scientists conclude in the April 1 Proceedings of the National Academy of Sciences. Kantoff also presented the results at a June 1996 meeting of the American Society for Clinical Oncology in Philadelphia.

Kantoff says that his interest in CAG repeats was sparked when researchers investigating an inherited neuromuscular disorder, Kennedy's syndrome, discovered the repeats in a coding region of the androgen receptor gene. "It's unusual to see DNA repeats in coding regions," he says.

Further study showed that men with Kennedy's syndrome had an unusually large number of CAG repeats, which seemed to render the men slightly less sensitive to the masculinizing effects of androgens. Investigators found they could increase the receptors' sensitivity to androgens in the laboratory by knocking out some repeats.

Two years ago, Kantoff and his colleagues began wondering whether CAG repeats played a role in prostate cancer, which is diagnosed in more than 300,000 men each year. "It's an obvious question, since these receptors modulate androgen function and androgen has something to do with prostate cancer development."

Earlier studies had shown that prostate tumors shrink temporarily if their androgen supply is cut off and that lower concentrations of the hormone reduce the likelihood of prostate cancer.

Last year, in a study of 104 prostate cancer patients, researchers at Memorial Sloan-Kettering Cancer Center in New York found that men with fewer CAG repeats were likely to get cancer at a younger age. "These data suggest that CAG repeat length can affect the risk of developing prostate cancer," the team concluded, cautioning that their findings were tentative.

Howard I. Scher, an author of that study, called Kantoff's report "very nice work" and added, "you can almost look at [CAG repeats] as the throttle or the accelerator. You know the gene to be active. CAG repeats determine how active." Other regulatory factors probably play a role as well, Scher says.

Several major questions remain. For instance, black men-who have fewer CAG repeats on average than white men-are also more prone to prostate cancer, but no one knows yet whether the two factors are related. Kantoff's team could not examine this issue because only 5 percent of the participants in the study were black. He and other colleagues plan to design a study that will permit a closer look.