1. It has been proposed that in inflammatory conditions, in which both the inducible isoforms of nitric oxide synthase (iNOS) and cyclo-oxygenase (COX-2) are induced, inhibition of NOS also results in inhibition of arachidonic acid metabolism. In the present study we have investigated whether mercaptoalkylguanidines, a novel class of selective iNOS inhibitors, may also influence the activity of cyclo-oxygenase (COX). Therefore, the effect of mercaptoethylguanidine (MEG) and related compounds on the activity of the constitutive (COX-1) and the inducible COX (COX-2) was investigated in cells and in purified enzymes. Aminoguanidine, NG-methyl-L-arginine (L-NMA) and NG-nitro-L-arginine methyl ester (L-NAME) were also studied for comparative purposes. 2. Western blot analysis demonstrated a significant COX-1 activity in unstimulated J774 macrophages and in unstimulated human umbilical vein endothelial cells (HUVEC). Immunostimulation of the J774 macrophages by endotoxin (lipopolysaccharide of E. coli, LPS 10 micrograms ml-1) and interferon gamma (IFN gamma, 100 u ml-1) for 6 h resulted in a significant induction of COX-2, and a down-regulation of COX-1. No COX-2 immunoreactivity was detected in unstimulated HUVEC or unstimulated J774 cells. Therefore, in subsequent studies, the effect of mercaptoalkylguanidines on COX-1 activity was studied in HUVEC stimulated with arachidonic acid for 6 h, and in J774 cells stimulated with arachidonic acid for 30 min. The effect of mercaptoalkylguanidines on COX-2 activity was studied in immunostimulated J774 macrophages, both on prostaglandin production by endogenous sources, and on prostaglandin production in response to exogenous arachidonic acid stimulation. In addition, the effect of mercaptoalkylguanidines on purified COX-1 and COX-2 activities was also studied. 3. In experiments designed to measure COX-1 activity in HUVEC, the cells were stimulated by arachidonic acid (15 microM) for 6 h. This treatment induced a significant production of 6-keto-prostaglandin F1 alpha (6-keto-PGF1 alpha, the stable metabolite of prostacyclin), while nitrite production was undetectable by the Griess reaction. MEG (1 microM to 3 mM) caused a dose-dependent inhibition of the accumulation of 6-keto-PGF1 alpha, with an IC50 of 20 microM. However, aminoguanidine, L-NAME or L-NMA (up to 3 mM) did not affect the production of 6-keto-PGF1 alpha in this experimental system. In experiments designed to measure COX-1 activity in J774.2 macrophages, the cells were stimulated by arachidonic acid (15 microM) for 30 min; this also induced a significant production of 6-keto-PGF1 alpha and MEG (1 microM to 3 mM), aminoguanidine (at 1 and 3 mM), but neither L-NAME nor L-NMA inhibited the production of prostaglandins. 4. In experiments designed to measure prostaglandin production by COX-2 with endogenous arachidonic acid, J774.2 cells were immunostimulated for 6 h in the absence or presence of various inhibitors. In experiments designed to measure prostaglandin production by COX-2 with exogenous arachidonic acid, J774.2 cells were immunostimulated for 6 h, followed by a replacement of the culture medium with fresh medium containing arachidonic acid and various inhibitors. Both of these treatments induced a significant production of 6-keto-PGF1 alpha. Nitrite production, an indicator of NOS activity, was moderately increased after immunostimulation. MEG (1 microM to 3 mM) caused a dose-dependent inhibition of the accumulation of COX metabolites. Similar inhibition of LPS-stimulated 6-keto PGF1 alpha production was shown by other mercaptoalkylguanidines (such as N-methyl-mercaptoethylguanidine, N,N'-dimethyl-mercaptoethylguanidine, S-methyl-mercaptoethylguanidine and guanidino-ethyldisulphide), with IC50 values ranging between 34-55 microM. However, aminoguanidine, L-NAME and L-NMA (up to 3 mM) did not affect the production of prostaglandins.5. In comparative experiments indomethacin, a non selective COX inhibitor, and NS-398, a selective COX-2 inhibitor, reduced (LPS) stimulated 6-keto-PGF1alpha production in J774 macrophages in a dose-dependent manner without affecting nitrite release. Indomethacin, but not NS-398, inhibited 6-keto-PGF1alpha production in the HUVECs. 6.The inhibitory effect of MEG was due to direct inhibition of the catalytic activity of COX as indicated in experiments with purified COX-1 and COX-2. MEG dose-dependently inhibited the purified COX-1 and COX-2 activity with IC50 values of 33microM and 36microM, respectively. Aminoguanidine (at the highest concentrations) inhibited the formation of COX-1 metabolites, without affecting COX-2 activity. High doses of L-NAME (3mM) decreased COX-1 activity only, while L-NMA (up to 3mM) had no effect on the activity of either enzyme. 7.These results suggest that MEG and related compounds are direct inhibitors of the constitutive and the inducible cyclo-oxygenases, in addition to their effects on the inducible NOS. The additional effect of mercaptoalkylguanidines on COX activity may contribute to the beneficial effects of these agents in inflammatory conditions where both iNOS and COX-2 are expressed.