Effect on an enzyme may stop pain, lower fever. (Acetaminophen in Action).

If a medication does its job safely, regulators may approve it even if nobody knows how it works. That's been the case for acetaminophen, commonly sold as Tylenol.

A team of scientists at Brigham Young University in Provo, Utah, is now a step closer to unraveling this medical mystery. The researchers report that acetaminophen targets a previously unidentified enzyme, which they call cyclooxygenase-3, or COX-3.

Two COX enzymes identified in the early 1990s transform arachidonic acid into prostaglandins, which are natural substances that can induce pain and inflammation. Acetaminophen's over-the-counter competitors--aspirin and ibuprofen--stop inflammation by binding to COX-1 and COX-2 and short-circuiting the prostaglandin-making pathway. Newer prescription drugs, celecoxib and rofecoxib, target COX-2 only. These three nonsteroidal anti-inflammatory drugs reduce pain and fever.

Scientists have wondered how acetaminophen can also reduce pain and fever--particularly headaches--while having little effect on inflammation. Moreover, acetaminophen doesn't seem to inhibit COX-1 or COX-2.

While studying whether dogs' brain cells produce COX enzymes in the same way human cells do, Daniel L. Simmons, a biochemist at Brigham Young University, and his colleagues found that the COX-1 enzyme has three variants.

"That was the starting point," Simmons says. The researchers then genetically engineered insect cells to produce the enzyme variants. The team found that in that experimental system, one variant makes prostaglandins from arachidonic acid. This suggested that the variant plays a role in inflammation, says Simmons. Further tests showed that acetaminophen inhibits this same variant, linking the drug to the prostaglandin-making pathway. The researchers dubbed the variant COX-3.

If acetaminophen could block prostaglandin production this way, why doesn't it stop inflammation in, say, a sprained knee? The researchers examined various human-tissue samples to see which harbored COX-3. By looking at the activity of the genes that encode the COX enzymes, Simmons' group found that COX-3 is most abundant in the cerebral cortex of the brain and less prevalent in the heart and other tissues. The researchers report their findings in an upcoming issue of the Proceedings of the National Academy of Sciences.

The new data don't prove that acetaminophen targets COX-3 in people, Simmons says. However, the idea that there is a specific COX variant that makes prostaglandins in the brain and responds to acetaminophen makes sense because the drug works well against fever, which is controlled by brain prostaglandins.

Indeed, scientists have suspected that the beneficial effects of acetaminophen might derive from its better penetration of the barrier that encases the brain and other parts of the central nervous system. As early as 1972, another group of scientists found test-tube evidence that acetaminophen could inhibit COX enzymes--which had yet to be subdivided into COX-1 and COX-2--in brain cells.

The new work is already turning heads. Clark M. Blatteis, a physiologist at the University of Tennessee College of Medicine in Memphis, has been testing the effects of COX-1 and COX-2 inhibitors on prostaglandin production in guinea pigs with fever. He's now adding acetaminophen to his trial to test whether it, too, reduces prostaglandins, in this case by inhibiting COX-3.

Meanwhile, Simmons says there could also be variants of COX-2 susceptible to acetaminophen. This possibility--coupled with the identification of variant forms of COX-1--might explain why there are so many nonsteroidal anti-inflammatory drugs on the market, says pharmacologist Timothy D. Warner of the London School of Medicine and Dentistry. It could be that the structures of the COX enzymes vary from person to person, meaning that a given meditation might work better on some than others, he says.