01 Jun 1999

Hopes have been raised that drugs which inhibit the proinflammatory COX-2 enzyme could be effective in treating rheumatoid arthritis and perhaps also Alzheimer's disease, which many scientists think is exacerbated by inflammatory responses in the brain. A new study from Paul Colville-Nash and his colleagues in London may dampen initial expectations. Writing in this month's Nature Medicine, they raise the possibility that using COX-2 inhibitors as a maintenance medication may actually speed progression of an inflammatory disease.

The researchers investigated the activity of COX-2 by inducing inflammation in the rat pleural cavity, a model used to study the effects of anti-inflammatory agents. In keeping with previous results, they found that after two hours, at the height of inflammation, both COX-2 and its proinflammatory product, prostaglandin E-2 (PGE-2), were at a peak. Surprisingly though, they found that 48 hours later, coincident with the inflammation resolving, the level of COX-2 was at a new peak, this time with a 350 percent increase over the level at the height of inflammation. This was not accompanied by an increase in PGE-2; instead, levels of other prostaglandins-specifically, D-2 and J-2 of the cyclopentanone family-had increased significantly.

Colville-Nash and his colleagues then added a COX-2 inhibitor (NS-398), which reduced inflammation at two hours, but had the opposite effect at 48 hours, exacerbating the inflammation and reducing PGD-2 and PGJ-2 levels. Indomethacin, a traditional nonsteroidal anti-inflammatory drug that inhibits both COX-2 and COX-1, had effects similar to NS-398.

The authors suggest that COX-2 may have a dual role in this model: initially promoting inflammation through the production of PGE-2 and subsequently helping to resolve inflammation via production of PGD-2 and PGJ-2. If this were to be the case, drugs that target COX-2 in inflammatory diseases such as arthritis may be useful for treating the pain and edema of the initial inflammation but may actually contribute to the underlying progression of the disease if used during periods of remission.

Writing a New and Views response in the same issue, Karen Seibert and her colleagues from Searle (makers of the COX-2 inhibitor, Celebrex) caution that this animal model of inflammation may not accurately reflect what occurs in a condition such as human rheumatoid arthritis, and they suggest clinical data are needed to determine how COX-2 inhibitors should be employed.—Hakon Heimer