Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

N-Hydroxy-L-arginine (L-NOHA) is a stable intermediate in the two step reaction catalyzed by NO synthase, the enzyme family involved in the formation of NO and L-citrulline from L-arginine and O₂ in mammals (Forstermann et al., 1994; Marletta, 1994; Stuehr, 1997). The first step of the reaction, N-oxygenation of the guanidine function of L-arginine, is specifically and exclusively catalyzed by NO synthase isoforms (Stuehr et al., 1991; Klatt et al., 1993). The second step of the reaction consists in oxidative cleavage of the C=NOH bond of the N-hydroxyguanidine function, with formation of the free radical NO and L-citrulline. Recently, it has been reported that non--amino acid N-aryl-N-hydroxyguanidines are also oxidized by NO synthase with formation of NO (Renodon-Cornière et al., 2002). It has been known for several years, however, that oxidative cleavage of the C=N bond of C=NOH functions of various compounds, including L-NOHA, can be catalyzed also by hemoproteins like horse radish peroxidase, rat liver microsomal cytochrome P450 (P450), hemoglobin, and catalase. This results in generation of stable nitrogen oxides, with the possible intermediate formation of NO (Boucher et al., 1992a,b; Jousserandot et al., 1998; Caro et al., 2001). The latter reactions provide alternative pathways to the second step of the reaction catalyzed by NO synthases. The existence of such pathway(s) might be important in blood vessels to restore NO formation in pathological situations in which endothelial NO synthase expression or activity is impaired. This is especially the case of atherosclerosis, diabetes, hypertension, and various situations of increased coronary risk such as cigarette smoking (for reviews see Harrison, 1997; Li and Forstermann, 2000).

Being a substrate for NO synthase (Stuehr et al., 1991; Klatt et al., 1993; Moali et al., 2000), L-NOHA can produce endothelium-dependent vascular relaxation as a result of its metabolization to NO and L-citrulline by the endothelial NO synthase. This mechanism, which is antagonized by NO synthase inhibitors, has been reported to account for the major part of relaxation caused by L-NOHA in bovine pulmonary and porcine coronary arteries (Wallace et al., 1991; Abdul-Hussain et al., 1996). In other vessels such as the rabbit aorta, however, L-NOHA produced no endothelium-dependent relaxation by itself; it did not influence endothelium-dependent relaxation caused by acetylcholine, but it induced modest endothelium-independent relaxation (Zembowicz et al., 1992). Robust endothelium-independent relaxations, which were reversed by NO synthase inhibitors, were also produced by L-NOHA in arginine-depleted isolated pulmonary arteries (Wallace et al., 1991). These relaxations were probably due to metabolization by the inducible isoform of NO synthase because they only occurred after prolonged incubation and comparable relaxations were also produced by addition of L-arginine. Thus, the mechanisms of L-NOHA-induced relaxation appear variable from one vessel to the other.

The existence of an NO synthase-independent pathway capable of oxidizing L-NOHA or another compound with a C=NOH bond, formamidoxime, to nitrogen oxides has been suggested in cultured smooth muscle cells from the rat aorta (Schott et al., 1994) and from the rat trachea (Jia et al., 1998), respectively. In both cases, formation of nitrite was blunted by a P450 inhibitor, supporting the involvement of P450. Exposure of the isolated trachea to formamidoxime caused cGMP accumulation and relaxation (Jia et al., 1998), consistent with the view that the pathway leading to NO formation was functional in the tissue as well as in the cultured tracheal cells. In the present investigation, we hypothesized that this could also be the case in the freshly isolated aorta.

Therefore, effects of L-NOHA and nonamino acid compounds bearing a C=NOH function (Fig. 1) were studied here on vascular relaxation and cGMP accumulation in rat aortic rings with or without endothelium. As L-NOHA competed with L-arginine for transport into aortic smooth muscle by a cationic amino acid carrier (Schott et al., 1994), some experiments were performed in the presence of L-(or D-) arginine. The involvement of NO was explored using an NO scavenger (2-phenyl-4,4,5,5-tetramethyl-imidazoline-1-oxyl-3-oxide; PTIO) and an inhibitor of guanylyl cyclase activation by NO (1H[1,2,4,]oxadiazolo[4,3-a]quinoxalin-1-one; ODQ; 1 µM). The hypothesis that NAD(P)H-dependent enzymes, especially P450s or P450 reductase, were implicated in oxidation of C=NOH bonds was investigated using the flavoprotein-dependent enzymes inhibitor diphenyliodonium, nonselective inhibitors of P450 isoforms (proadifen and miconazole), and 7-ethoxyresorufin, which is not only a suicide substrate of the P450 1A₁ isoform (Tassaneeyakul et al., 1993) but also an inhibitor of NADPH-P450 reductase (Dutton et al., 1989) and other reductases (Jiang and Ichikawa, 1999). In addition, N-nitro-L-arginine methyl ester (L-NAME) was used to inhibit NO synthase, which is also a flavoprotein-dependent enzyme.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Structure of the compounds used in this study. NOHA, N-hydroxy-L-arginine; HG, N-hydroxy guanidine; FA, formamidoxime.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Contraction-Relaxation Experiments in Rat Aortic Rings. Thoracic aorta rings were obtained as described previously (Andriambeloson et al., 1997) from male 12- to 14-week-old Wistar rats bred in our laboratory from Iffa-Credo (France) genitors. In some rings, the endothelium was removed by gentle rubbing with blunt forceps. Rings were then mounted under 2 g of tension in organ baths containing Krebs' solution of the following composition: 119 mM NaCl, 4.7 mM KCl, 1.25 mM CaCl₂, 1.18 mM KH₂PO₄, 11.0 mM glucose, 25.0 mM NaHCO₃ at 37°C bubbled with 95% O₂/5% CO₂ (pH 7.4). After an equilibration period of 60 min during which the Krebs' solution was changed every 15 min, basal tension was readjusted to 2 g. Then contractile capacity was tested with norepinephrine or phenylephrine (1 µM) as indicated. The absence or the presence of a functional endothelium was determined by the respective inability or ability of acetylcholine (1 µM) to relax norepinephrine or phenylephrine precontracted rings. Then a second washing period of 60 min followed during which tension returned to baseline levels. Thereafter, rings were contracted with a single dose of norepinephrine or phenylephrine (0.1 µM, producing approximatively 80% of the maximum contractile response, in both cases) in the presence or absence of the indicated compounds. As there was no difference in relaxing effects of drugs with C=NOH functions when contraction was elicited by norepinephrine or phenylephrine (see Results), either agonist was used. In some cases, however, phenylephrine was preferred because some drugs used to investigate the mechanisms of relaxation altered the stability of norepinephrine-induced contraction. In some rings, L-arginine (1 mM) or D-arginine (1 mM) was added just before norepinephrine. When contractions reached stable levels, L-NOHA or compounds with C=NOH function were added cumulatively at the indicated final concentrations. In some experiments, the following drugs (or corresponding solvent) were added 30 min before the contraction experiments and were also present throughout the experiments: L-NAME (300 µM), ODQ (1 µM), PTIO (100 µM), diphenyliodonium (30 µM), proadifen (10 µM), miconazole (30 µM), or 7-ethoxyresorufin (10 µM).

Determination of Tissue cGMP Content. Aortic rings without endothelium (5 mm) were incubated for 30 min in Krebs' solution supplemented with isobutylmethylxanthine (IBMX; 100 µM) at 37°C and oxygenated with a gas mixture of 95% O₂/5% CO₂. Thereafter, the Krebs' solution was replaced by fresh solution, and rings were incubated in the absence or presence of either L-NOHA (100 µM) or 4-chlorobenzamidoxime (ClBZA) (100 µM) for the next 30 min. The rings were then put into 500 µl of ice-cold hydrochloric acid (0.1M) and homogenized with a Potter glass/glass homogenizer for 30 s followed by sonication (type 75 TS; Ultrason, Annemasse, France) for 15 s and centrifugation at 10,000g for 5 min. Pellet and supernatant were frozen and maintained at 20°C until assayed. The cGMP content was determined in thawed supernatant using a radioimmunoassay described previously (Cailla et al., 1976), modified by separation of free cGMP with activated charcoal (Koch and Lutz-Bucher, 1991). DNA was determined in the pellet according to Setaro and Morley (1976), and the cGMP content of the rings was expressed as fmol · µg¹ DNA.

Expression of Results and Statistical Analysis. Results were expressed as mean ± S.E.M. of n experiments. When possible (i.e., when full relaxation could be reached), the concentration value causing 50% relaxation of precontracted vessel (IC₅₀) was determined by log-logit regression. cGMP content is expressed as fmol · µg¹ DNA. Statistical comparisons of concentration-effect curves were performed using multianalysis of variance. A Student's t test for paired or unpaired data (as appropriate) was used for other statistical comparisons, with p values less than 0.05 considered to be statistically significant.

Drugs and Reagents. L-NOHA was obtained from Alexis Corporation (Läufelfingen, Switzerland) and N-hydroxy-D-arginine from GlaxoSmithKline (Uxbridge, Middlesex, UK). N-Hydroxyguanidines, N-(4-chlorophenyl)-N'-hydroxyguanidine, N,N'-dicyclohexyl-N"-hydroxyguanidine, N-hydroxydebrisoquine, benzamidoxime, ClBZA, 4-nitrobenzamidoxime (NO₂BZA), 4-n-(hexyloxy)benzamidoxime (HXBZA), 4-(methoxy)benzamidoxime (MXBZA), and 4-chloroacetophenone-oxime (ClBK) were synthesized following previously described procedures (Jousserandot et al., 1998; Renodon-Cornière et al., 2002).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Endothelium-Independent Relaxing Effects of L-NOHA and Compounds with a C=NOH Function. Contraction elicited by norepinephrine or phenylephrine remained stable in control rings during the whole experiment, as shown in Fig. 2 in the case of norepinephrine in endothelium-denuded rings. After the addition of L-NOHA, relaxation developed slowly (within approximatively 15 min). The concentration of L-NOHA was therefore cumulatively increased each 15 min. Relaxations caused by nonamino acid compounds with a C=NOH function developed more rapidly (within 3 to 5 min after each drug addition), as illustrated in the case of the most active, ClBZA. The concentration of these drugs was increased when relaxation reached a plateau (each 3 to 5 min).

View larger version (48K): [in this window] [in a new window]   Fig. 2.   Endothelium-independent relaxing effects of L-NOHA and ClBZA in rat aortic rings. The rings were precontracted to comparable levels with norepinephrine (NE; 0.1 or 0.3 µM in the absence or the presence of functional endothelium, respectively). Top panels, original recordings showing the time courses of vascular tone in the absence of functional endothelium; vertical bars indicate addition of solvent (control) or drugs. Bottom panels: concentration-response curves (mean ± S.E.M. of five to eight experiments) of L-NOHA (left) and ClBZA (right).

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Effect of L- or D-arginine (1 mM) on relaxations caused by L-NOHA (left) or ClBZA (right) in endothelium-denuded rat aortic rings. The rings were precontracted with norepinephrine (0.1 µM). Results are mean ± S.E.M. of six to seven experiments. For comparison of data, multianalysis of variance was used. , p < 0.05.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Relaxing effects of compounds bearing C=NOH function in endothelium-denuded rat aortic rings precontracted with phenylephrine (0.1 µM; A) or norepinephrine (0.1 µM; B, C, and D). FA, formamidoxime; HG, N-hydroxyguanidine; BZA, benzamidoxime. Data are means ± S.EM. of four to six experiments.

Involvement of the NO-cGMP Pathway. The data in Fig. 5 show that L-NOHA (100 µM) and ClBZA (100 µM) were both able to cause cGMP accumulation in endothelium-denuded aorta in the presence of the phosphodiesterase inhibitor IBMX. In these experiments, the two compounds were used at the same concentration (100 µM). The elevations in cGMP level (1.9-fold in the case of L-NOHA and 4.2-fold in the case of ClBZA) were consistent with the larger relaxing effect of ClBZA at this concentration (Fig. 2).

View larger version (40K): [in this window] [in a new window]   Fig. 5.   Effects of L-NOHA and ClBZA on cGMP accumulation in endothelium-denuded rat aortic rings. Rings were incubated for 30 min in oxygenated Krebs' solution supplemented with IBMX (100 µM) and either with L-NOHA (100 µM) or ClBZA (100 µM) or solvent. Data are means ± S.EM. of five experiments. Student's t test was used to evaluate the significance of differences. **, p < 0.01; ***, p < 0.001.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Effects of ODQ (1 µM; A and B) and PTIO (300 µM; C and D) on relaxations produced by L-NOHA (left panels) or ClBZA (right panels) in endothelium-denuded rings precontracted with either norepinephrine (0.1 µM; A and B) or phenylephrine (0.1 µM; C and D). Data are mean ± S.EM. of three to seven experiments. For comparison of data, multianalysis of variance was used. , p < 0.01; , p < 0.001.

Involvement of P450s or Other NAD(P)H-Dependent Enzymes. As illustrated in Fig. 7, diphenyliodonium (30 µM) and 7-ethoxyresorufin (10 µM) markedly inhibited relaxations elicited by L-NOHA and ClBZA. By contrast, proadifen (10 µM) and L-NAME (300 µM) failed to alter these relaxations. Miconazole (30 µM) did not either inhibit relaxation caused by L-NOHA. Proadifen, miconazole, and 7-ethoxyresorufin did not affect precontraction level; however, it was reduced by about 40% by diphenyliodonium (not shown).

View larger version (27K): [in this window] [in a new window]   Fig. 7.   Effects of diphenyliodonium (DPI; 30 µM; A and B), 7-ethoxyresorufin (7-ER; 10 µM; C and D), proadifen (10 µM; E and F), miconazole (30 µM; E), and L-NAME (300 µM; G and H) on relaxations elicited by L-NOHA (left panels) and ClBZA (right panels) in aortic rings without endothelium. The rings were precontracted with norepinephrine (0.1 µM), except in the case of experiments shown in F in which phenylephrine (0.1 µM) was used. Data given as mean ± S.EM. of four to six experiments. For comparison of data, multianalysis of variance was used. , p < 0.01; , p < 0.001.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The above data show that compounds bearing a C=NOH function, including L-NOHA, are able to produce endothelium-independent vasorelaxation in the rat aorta. They provide evidence that activation of guanylyl cyclase by NO released from these compounds is involved in this relaxation and that oxidative cleavage of the C=NOH function represents an NO synthase- and P450-independent pathway for NO production in the aorta.

Involvement of the NO-cGMP pathway in aortic relaxation caused by L-NOHA and other compounds with a C=NOH function is supported by inhibition by the NO scavenger PTIO and by the guanylyl cyclase activation inhibitor ODQ, as well as by cGMP accumulation. Whether NO was released in the cytosol (as a free radical or as a related species) or directly transferred to guanylyl cyclase as recently suggested in the case of nitroglycerin (Kleschyov et al., 2002) deserves further investigation. cGMP accumulation caused by L-NOHA was relatively modest compared with the one produced by ClBZA (this study) or other NO donors. Thus, despite inhibition by ODQ, involvement of cGMP-independent relaxing mechanisms, such as S-nitrosylation of K⁺ channels, cannot be excluded, especially in the case of L-NOHA. The stereospecific inhibition by L-arginine, however, is consistent with L-NOHA entering cells via a cationic amino acid transporter (Schott et al., 1994), whereas ClBZA probably enters cells via another pathway. The transporter might limit entry of L-NOHA in cells, accounting for a relatively modest effect on cGMP accumulation and slowly developing and moderate relaxation.

The finding that, in the rat aorta, neither the presence of endothelium nor L-NAME (in endothelium-denuded rings) significantly influenced relaxations elicited by L-NOHA and ClBZA suggest that metabolization by endothelial or nonendothelial NO synthase was of minor importance, if any, in the formation of a relaxing compound from C=NOH bonds in this particular tissue. This is different from what has been reported in other vessels in which endothelial NO synthase plays a major role in relaxation produced by L-NOHA (Wallace et al., 1991; Abdul-Hussain et al., 1996). Tissue or species differences might explain such discrepancy.

Failure of the P450 nonselective inhibitor proadifen to inhibit relaxations caused by L-NOHA and ClBZA do not support the involvement of a proadifen-sensitive P450 in oxidation of C=NOH bond to NO or a related relaxing compound in the freshly isolated rat aorta. Another P450 inhibitor, miconazole, was also unable to blunt L-NOHA-induced relaxation, whereas it inhibited formation of nitrite from L-NOHA in cultured aortic smooth muscle cells (Schott et al., 1994). This suggests that the mechanism of oxidation of the C=NOH bonds was different in freshly dissected aorta from that in cultured cells. Culture conditions might favor expression of P450 isoform(s) or other miconazole-sensitive enzymes which are not expressed (or not expressed at the same level) in the tissue.

Structure activity relationships found in the present investigation show that, in the rat aorta, the order of relaxant activity of amidoxime, ketoxime and N-hydroxyguanidine derivatives is opposite to the order of reactivity of the same compounds for P450-dependent oxidative cleavage of the C=NOH bond, where N-hydroxyguanidines are the most active and the ketoxime the less active one (Jousserandot et al., 1998). In addition, substitutions that increase the ability of benzamidoximes to generate NO in liver microsomes (by introducing an electron donating group in para position of the phenyl group) have no clear effect on the relaxant activity of these compounds (compare NO₂BZA, ClBZA, MXBZA and HXBZA, Fig. 4 and Table 1). The structural determinants required for the relaxant effect are also entirely different from those required for being an NO synthase substrate (Dijols et al., 2001; Renodon-Cornière et al., 2002): the substituted N-hydroxyguanidine, a poorly relaxant compound, is a selective substrate for the inducible NO synthase; by contrast, the identically substituted amidoxime ClBZA and ketoxime ClBK are potent relaxant compounds, whereas they are not substrate for any of the three NO synthase isoforms. Thus, structure activity relationships together with the above-mentioned failure of P450 and NO synthase inhibitors to blunt relaxation lend no support to the hypothesis of the involvement of a P450 or NO synthase in the relaxant effect of compounds with a C=NOH function in the rat aorta.

The present findings that both the nonspecific flavoprotein enzyme inhibitor diphenyliodonium and the NADPH-reductases inhibitor 7-ethoxyresorufin abolished relaxations produced by L-NOHA and ClBZA suggest that the reductase domain of a NAD(P)H-dependent enzyme might be implicated. Even though 7-ethoxyresorufin is regarded as a substrate and selective inhibitor of P450 1A₁ (Tassaneeyakul et al., 1993), it is also an inhibitor of various reductases (Dutton et al., 1989; Jiang and Ichikawa, 1999). Interestingly, it has been previously reported that 7-ethoxyresorufin and diphenyleneiodonium (an analog of diphenyliodonium), but not classical P450 inhibitors such as proadifen, caused inhibition of glyceryl trinitrate-induced vascular relaxation (Bennett et al., 1994; Li and Rand, 1996). The authors suggested that inhibitors like proadifen do not inhibit P450 isoform(s) present in blood vessels or that another enzyme such as NADPH-cytochrome P450 reductase is capable of metabolizing glyceryl trinitrate to NO. The present data lead us to propose a similar hypothesis for the mechanism(s) underlying endothelium-independent relaxation elicited by compounds with a C=NOH function.

It should be stressed that due to the above-mentioned differences in the mechanisms of action of L-NOHA in different blood vessels, the data obtained here in the isolated rat aorta should not be extrapolated to other vessels. Further studies are required to evaluate the involvement of formation of NO from compounds with a C=NOH function in the effects of these compounds in different vascular beds and in in vivo conditions.

In conclusion, the above results indicate that compounds bearing a C=NOH function, including L-NOHA and especially non--amino acid-substituted benzamidoximes can act as efficient NO synthase-independent activators of the cGMP pathway in the rat aorta. They can help identifying a novel pathway for endothelium-independent NO formation in vessels. In addition, they might provide valuable surrogates for impaired endothelial NO activity in vascular diseases with endothelial injury.