Anti-inflammatories' cancer effects vary by brand and tissue type.

Two new studies on medications being investigated as cancer treatments indicate that certain of these drugs have secondary effects that could enhance or undermine their antitumor activity. These compounds, which inhibit the inflammatory enzyme cyclooxygenase-2, or COX-2, are currently used to treat pain and arthritis. Researchers have now found that one such drug surprisingly accelerates the growth of some pancreatic tumors. Another COX-2 inhibitor, however, demonstrates unexpectedly strong activity against prostate cancer. These newly recognized effects seem to result from activities beyond the drugs' inhibition of COX-2.

Effects unique to specific drugs in this class could become "a point of distinction," enabling doctors to choose the optimal drug for a given situation, says Andrew Dannenberg of Cornell University's Weill Medical College in New York City.

When cell types that typically produce little or no COX-2 become cancerous, they often churn out excessive amounts of the enzyme, which stimulate the formation of blood vessels that feed a growing tumor. Past studies indicated that aspirin and related anti-inflammatory drugs, which broadly inhibit cyclooxygenase enzymes, and drugs such as celecoxib (Celebrex) and rofecoxib (Vioxx), which selectively block COX-2 production, can prevent or stop certain cancers from growing.

Other research, however, suggested that most prostate cancers and between 10 and 40 percent of pancreatic tumors don't produce COX-2 and so are called COX-2 negative.

To test the COX-2 inhibitor called nimesulide against cancerous pancreatic cells, Guido Eibl of the University of California, Los Angeles School of Medicine and his colleagues applied the drug to cells from COX-2-positive and COX-2-negative pancreatic tumors.

In both lab dishes and mice, the drug impeded production of a compound required for growth of COX-2-positive cancer cells but increased production of that factor in COX-2-negative ones, Eibl reported on March 30 at a meeting of the American Association for Cancer Research in Orlando, Fla. Given these results, he says, doctors should ascertain that a pancreatic tumor is producing COX-2 before they prescribe a COX-2 inhibitor.

"It's a surprising finding ... that really deserves some attention," says Raymond DuBois, a COX-2 researcher at Vanderbilt-Ingram Cancer Center in Nashville. "Certain subsets of the population may respond better to these drugs," he notes.

DuBois adds, however, that the researchers reported no data on celecoxib or rofecoxib, the two prescription COX-2 inhibitors approved in the United States for treating pain and inflammation. He suggests that the new findings could be unique to nimesulide.

Eibl and his colleagues have conducted preliminary tests using nimesulide against COX-2-negative liver, breast, and colorectal cancer cells. They haven't found a cancer-exacerbating or cancer-hindering effect in those experiments.

In the second new study, Dannenberg and his colleagues tested two other COX-2 inhibitors. Celecoxib inhibited growth of cells from two types of prostate tumors that don't express COX-2, says Dannenberg, who acknowledges past financial support from the drug's maker. Furthermore, growth of COX-2-negative human-prostate tumors implanted into mice progressively slowed according to how much celecoxib was administered to the animals, the researchers reported at the Orlando meeting.

In contrast, rofecoxib did not inhibit the growth of the prostate-tumor cells. The findings indicate that biological effects unrelated to COX-2 must account for part of celecoxib's overall impact on cancer, Dannenberg says.