Why is it that people who regularly use nonsteroidal anti-inflammatory drugs (NSAIDs), such as ibuprofen, are less likely to develop Alzheimer's disease? Some researchers have proposed that NSAIDs might reduce neurotoxic inflammation in the brain. However, Weggen et al. have found another potentially protective mechanism: a decrease in the amyloidogenic Aβ42 peptide in cultured cells or transgenic mice that are treated with NSAIDs.

The Aβ42 peptide is the chief suspect in the pathogenesis of Alzheimer's disease. It is formed when γ-secretase cleaves the amyloid precursor protein (APP), and helps to form the amyloid plaques that are a feature of the disease. γ-Secretase can also produce a shorter peptide, Aβ40, which seems to be less damaging. Weggen et al. treated cultured cells with three types of NSAID — ibuprofen, indomethacin and sulindac sulphide — and found that each of them caused a decrease in the ratio of Aβ42 to Aβ40. However, a number of other NSAIDs, including aspirin, naproxen, and cyclooxygenase 1 (COX1)- and COX2-selective inhibitors, did not have this effect.

Another study recently showed that chronic treatment with ibuprofen can reduce neuropathology in transgenic mice that express APP. Weggen et al. showed that acute ibuprofen treatment could reduce the levels of Aβ42 in these mice, indicating that the reduction in Aβ42 might be the mechanism by which pathology is decreased.

NSAIDs mediate their anti-inflammatory action by inhibiting the activity of COX. However, the ability of NSAIDs to reduce Aβ42 levels seems to be independent of COX activity. Weggen et al. found that when fibroblasts deficient in both COX1 and COX2, which have the same basal levels of Aβ42 and Aβ40 as normal fibroblasts, were treated with NSAIDs, the reduction in Aβ42 still occurred, showing that the ability of some NSAIDs to reduce Aβ42 levels does not rely on COX activity.

When the group looked more closely at Aβ peptides in cell cultures treated with NSAIDs, they found that the decrease in Aβ42 was accompanied by an increase in the levels of a shorter peptide, Aβ38. Treatment with sulindac sulphide and other NSAIDs seems to subtly alter the activity of γ-secretase, producing a shift in the proportions of Aβ peptides produced. This selective action, unlike that of current γ-secretase inhibitors, does not seem to interfere with APP or Notch processing. But NSAIDs have other side effects, particularly gastrointestinal and renal toxicity, so they might not be suitable for long-term use in patients with Alzheimer's disease.

However, there is cause for optimism. As the Aβ42-lowering effect of NSAIDs is independent of their COX-inhibiting activity, it might be possible to develop derivatives that have a strong effect on Aβ42 levels, but do not have the present drawbacks of NSAIDs.