Heart of the matter

Selective cyclooxygenase-2 (COX-2) inhibitors have been the subject of intensive debate since results from the controversial VIGOR clinical trial indicated a potential link between COX-2 anti-inflammatory agents and myocardial infarction. Now, reporting in Science, FitzGerald and colleagues suggest that a disruption in the dynamic balance that exists between the production of the prostacyclin PGI₂ and thromboxane A₂ (TxA₂) in the body might explain these cardiovascular adverse effects.

Both PGI₂ and TxA₂ are synthesized from the prostaglandin precursor PGH₂, which is made from arachidonic acid in a reaction catalysed by the enzyme cyclooxygenase. The synthesis of TxA₂ depends on the COX-1 subtype, which is found in platelets and is involved in protecting the lining of the stomach. COX-2, on the other hand, is involved in the synthesis of PGI₂ in endothelial cells, and mediates pain and inflammation. COX-2 inhibitors, such as rofecoxib (Vioxx) and celecoxib (Celebrex), have been successfully marketed on the basis that the main mechanism by which the non-selective, classical non-steroidal anti-inflammatory drugs (NSAIDs), such as aspirin and naproxen, cause gastrointestinal toxicity is by inhibiting COX-1. However, the publication of the VIGOR trial results in 2001 raised concerns about the safety of COX-2 inhibitors, as a significantly higher relative risk of cardiovascular toxicity was found for rofecoxib than for naproxen. But was this outcome due to a cardioprotective effect of naproxen, owing to its inhibition of the COX-1-dependent production of TxA₂, or to a cardiotoxic effect of rofecoxib, owing to enhanced TxA₂ levels in the absence of PGI₂?

FitzGerald and coworkers used mouse models to investigate the hypothesis that PGI₂ might modulate the cardiovascular effects of TxA₂ in vivo. Deletion of prostacyclin receptors was shown to enhance cellular proliferation and increase platelet activation in response to vascular injury — responses that could cause thrombotic side effects. Prostacyclin-receptor-deficient mice also had raised TxA₂ levels after vascular injury in comparison to wild-type mice, which was indicated by increased urinary secretion of a metabolite of TxA₂.

Conversely, mice in which thromboxane receptor sugnalling was downregulated, either by treatment with a TxA₂-receptor antagonist or by knocking out TxA₂ receptors, were shown to have reduced responses to vascular injury. Finally, in mice that lacked both thromboxane and prostacyclin receptors, the increase in cellular proliferation seen in the prostacyclin-receptor knockouts was eliminated, as was the platelet activation, confirming the contribution of enhanced TxA₂ biosynthesis to the vascular response.

The net implication is that PGI₂ might modulate the cardiovascular response through its influence on TxA₂ levels within the body. These findings could be unwelcome to the developers of Vioxx and Celebrex — Merck and Pharmacia, respectively — as they indicate that the increased risk for developing serious cardiovascular defects observed in the VIGOR trial could be a consequence of selective COX-2 inhibition. Caution should be applied in over-emphasizing the significance of these results, however, as the jury is still out on whether naproxen might have contributed a cardioprotective effect to the outcome of the VIGOR trial.