Reduction of urinary 8-epi-prostaglandin F(2α) during cyclo-oxygenase inhibition in rats but not in man

1. 8-epi-prostaglandin (PG) F(2α) a major F₂ isoprostane, is produced in vivo by free radical-dependent peroxidation of lipid-esterified arachidonic acid. Both cyclo-oxygenase isoforms (COX-1 and COX-2) may also form free 8-epi-PGF(2α) as a minor product. It has been recently seen in human volunteers that the overall basal formation of 8-epi-PGF(2α) in vivo is mostly COX-independent and urinary 8-epi-PGF(2α) is therefore an accurate marker of 'basal' oxidative stress in vivo. 2. To test the validity of this marker in the rat, we evaluated in vivo the effect of COX inhibition on the formation of 8-epi-PGF(2α) vs prostanoids. Two structurally unrelated COX inhibitors (naproxen: 30 mg kg^-1 day^-1; indomethacin: 4 mg kg^-1 day^-1) were given i.p. to rats kept in metabolic cages. In vivo formation of 8-epi-PGF(2α), was assessed by measuring its urinary excretion. Prostanoid biosynthesis was assessed by measuring urinary excretion of major metabolites of thromboxane (TX) and prostacyclin (2,3-dinor-TXB₁ and 2,3-dinor-6-keto-PGF(1α)). All compounds were selectively measured by immunopurification/gas chromatography-mass spectrometry. 3. Naproxen reduced urinary excretion of 2,3-dinor-TXB₁ and 2,3-dinor-6-keto-PGF(1α) but, unexpectedly, also that of 8-epi-PGF(2α) (82, 49 and 52% inhibition, respectively). Indomethacin had a similar effect (77, 69 and 55% inhibition). Esterified 8-epi-PGF(2α) in liver and plasma remained unchanged after indomethacin. 4. These findings prompted us to re-assess the contribution of COX activity to the systemic production of 8-epi-PGF(2α) in man. We gave naproxen (1 g day^-1) to healthy subjects (four nonsmokers and four smokers). Urinary 8-epi-PGF(2α) remained unchanged in the two groups (9.63 ± 0.99 before vs 10.24 ± 1.01 after and 20.14 ± 3.00 vs 19.03 ± 2.45 ng h^-1 1.73 m^-2), whereas there was a marked reduction of major urinary metabolites of thromboxane and prostacyclin (about 90% for both 11-dehydro-TXB₂ and 2,3-dinor-TXB₂; > 50% for 2,3-dinor-6-keto-PGF(1α)). 5. To investigate whether rat COX-1 produces 8-epi-PGF(2α) more efficiently than human COX-1, we measured the ex vivo formation of 8-epi-PGF(2α) and TXB₂ simultaneously in whole clotting blood. Serum levels of 8-epi-PGF(2α) and TXB₂ were similar in rats and man. 6. We conclude that a significant amount of COX-dependent 8-epi-PGF(2α) is present in rat but not in human urine under normal conditions. This implies that urinary 8-epi-PGF(2α) cannot be used as an index of near-basal oxidant stress in rats. On the other hand, our data further confirm the validity of this marker in man.