Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Responses to endothelium-dependent vasodilator agents such as bradykinin and acetylcholine exhibit several components that involve nitric oxide, prostaglandins, and endothelium-derived hyperpolarizing factor (EDHF) (Vane et al., 1990; Cohen and Vanhoutte, 1995). The contribution of EDHFs to the vasodilator response becomes more apparent when the synthesis of nitric oxide and prostaglandins is inhibited. Indeed, two of the criteria for EDHF-mediated responses are 1) the vasodilator activity remaining after inhibition of nitric oxide synthase and cyclooxygenase and 2) the vasodilator activity being prevented by inhibition of K⁺ channels with specific inhibitors or elevated extracellular K⁺ (Campbell and Gauthier, 2002). The identity of EDHF remains to be confirmed, and it is likely that several EDHFs exist, depending upon the species and vascular bed; the candidates range from a cytochrome P450-derived metabolite of arachidonic acid (Hecker et al., 1994; Campbell et al., 1996; Fisslthaler et al., 1999) to K⁺ itself (Edwards et al., 1998) as well as transfer of hyperpolarization from the endothelium to vascular smooth muscle via gap junctions (Hutcheson et al., 1999; Chaytor et al., 2001). There is considerable evidence for cytochrome P450-derived arachidonic acid metabolites, namely, the epoxides [epoxyeicosatrienoic acids (EETs)], as EDHFs in vascular tissues of several species, including humans (Campbell et al., 1996; Fisslthaler et al., 1999; Halcox et al., 2001). Moreover, a central role for cytosolic phospholipase A₂ in vasodilator responses attributed to EDHF supports the role of an arachidonic acid metabolite (Fulton et al., 1996: Hutcheson et al., 1999).

In addition to the standard endothelium-dependent vasodilator agents, acetylcholine and bradykinin, the acyl-CoA:lysolecithin acetyltransferase inhibitor, thimerosal, has been reported to elicit endothelium-dependent vasodilation that has been attributed to nitric oxide, prostaglandins, and EDHFs (Forstermann et al., 1986a,b; Beny, 1990; Rosenblum et al., 1992). Furthermore, low concentrations of thimerosal have been reported to enhance EDHF-mediated vasodilator responses (Mombouli et al., 1996). As an inhibitor of reacylation, thimerosal would be expected to increase intracellular levels of free arachidonic acid (Burke et al., 1997), the precursor of cytochrome P450-derived EETs that have been proposed as EDHFs. In addition, thimerosal has been reported to increase intracellular Ca²⁺ levels in endothelial cells (Gericke et al., 1993) and act as an inositol triphosphate receptor-sensitizing agent (Montero et al., 2001), effects that should increase the synthesis of nitric oxide, prostaglandins, and EETs. Consequently, we used thimerosal as a tool to further investigate the proposition of EETs as EDHF in the rat isolated perfused kidney. Thus, we reasoned that thimerosal should produce endothelium-dependent vasodilation by stimulating the release of nitric oxide, prostaglandins, and EDHFs and that the residual vasodilator activity following inhibition of nitric oxide synthase and cyclooxygenase should be susceptible to inhibitors of cytochrome P450, specifically epoxygenase, and to inhibitors of K⁺ channels. We tested this possibility by sequentially inhibiting cyclooxygenase, nitric oxide synthase, and cytochrome P450 and determining vasodilator responses to thimerosal and, in some cases, the release of EETs. The results show that both nitric oxide and nitric oxide-independent factors contribute to the vasodilator effect of thimerosal. The nitric oxide-independent component of the response, which was prevented by blockade of K⁺ channels and, therefore, may correspond to an EDHF, was attenuated by inhibitors of epoxygenase, which abolished the thimerosal-stimulated increases in EET release. These studies indicate that EETs contribute to the renal vasodilator effect of thimerosal and provide further evidence for EETs as an integral component of EDHF-mediated responses.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Isolated Perfused Kidney. Male Wistar rats, weight 300 to 500 g, were used for these studies in accordance with National Institutes of Health guidelines. Rats were anesthetized with pentobarbitone, 65 mg/kg i.p., and the right kidney was prepared for perfusion as described (Fulton et al., 1992). Briefly, following a midline laparotomy, the right kidney was cannulated via the mesenteric artery to prevent interruption of blood flow and perfused with warmed (37°C), oxygenated Krebs-Henseleit buffer at constant flow to obtain a baseline perfusion pressure of approximately 50 to 75 mm Hg. The vena cava was ligated above and below the right renal vein and cut to allow exit of the perfusate, and the ureter was transected. In some experiments where the perfusate was collected for the determination of EETs, the kidney was removed from the animal.

6

Release of EETs. In some experiments, EET release from kidneys treated with indomethacin plus L-nitroarginine and those treated with the combination of indomethacin, L-nitroarginine, and either miconazole or MS-PPOH was compared. Thus, 1-min perfusate collections were made immediately before and after the administration of 1 and 10 µg of thimerosal. The EETs were measured as their hydrolysis products, the dihydroxyeicosatrienoic acids (DHETs), by gas chromatography-mass spectroscopy. To 10-ml samples 3 ng of a mix of deuterium-labeled EETs (8,9-, 11,12-, and 14,15-EET-d8, 1 ng each) were added as internal standard. The samples were acidified to pH 4 with acetic acid, and the lipids were extracted with 10 ml of ethyl acetate which was decanted and dried. The extract was incubated with 100 µl KOH (1 M) for 1 h at 60°C to decompose indomethacin, which has an HPLC retention time similar to that of the DHETs. H₂SO₄ (0.5 M; 200 µl) was added to the samples, which were maintained at 60°C for 30 min to convert the EETs to their respective DHETs. After adjusting to pH 4 with KOH, ethyl acetate was used to extract the DHETs, which were dried under nitrogen and dissolved in 50 µl of methanol for separation by reverse-phase HPLC using an HP 1050 instrument (Hewlett Packard, Palo Alto, CA) with a Beckman ODS column (25 cm × 4.6 mm, 5 µm; Beckman Coulter, Inc., Fullerton, CA) and a linear gradient of 60 to 100% acetonitrile containing 0.025% acetic acid over 20 min at a flow rate of 1 ml/min. The peak corresponding to authentic DHETs was collected, the sample was dried, and the DHETs were derivatized to pentafluorobenzyl esters and trimethylsilyl ethers. Pentafluorobenzyl esters were prepared by the addition of 30 µl of 10% -bromo-2,3,4,5,6-pentafluorotoluene (Aldrich Chemical Co., Milwaukee, WI) in acetonitrile and 30 µl of 10% N,N-diisopropylethylamine (Aldrich) in acetonitrile. After 30 min at room temperature, the samples were dried and the trimethylsilyl ethers were prepared by dissolving the samples in 60 µl of N,O-bis(trimethylsilyl) trifluoroacetamide (Sigma-Aldrich, St. Louis, MO) and 20 µl of pyridine and allowing the reaction to proceed for 30 min at room temperature. The samples were dried and dissolved in 50 µl of iso-octane, and 1-µl aliquots were analyzed using an HP5890 gas chromatography-mass spectroscopy. The GC column (DB-1; 10 m, 0.25- mm inner diameter, 0.25-µm film thickness; Agilent Technologies Inc., Wilmington, DE) was temperature programmed from 180°C to 300°C at a rate of 25°C/min. Methane was used as a reagent gas at a flow resulting in a source pressure of 2 torr, and the MS (HP 5989A) was operated in the electron capture chemical ionization mode, monitoring ions at m/z of 481 and 489, which represented the derivatives of the unlabeled and deuterium-labeled DHETs. The 5,6-, 8,9-, and 11,12-DHETs exhibited the same retention time and were quantitated together, whereas 14,15-DHET, which had a different retention time, was quantitated separately. The amounts of DHETs in the samples were calculated by reference to a standard curve.

Removal of Endothelium. In three kidneys perfused with buffer containing L-nitroarginine and indomethacin, the endothelium was removed with 0.1 ml of Triton X-100 (0.5%). After elevation of perfusion pressure with phenylephrine, vasodilator responses to bradykinin, thimerosal, and nitroprusside were determined.

Analysis. All data are expressed as mean ± S.E.M. Because data from different series of experiments were pooled, they were compared by ANOVA of one data set in which individual comparisons were made using a Bonferroni correction. A p value <0.05 was considered statistically significant.

Materials. Bradykinin, nitroprusside, tetraethylammonium, and L-nitroarginine were obtained from Sigma-Aldrich and dissolved in distilled water. Miconazole and indomethacin were also obtained from Sigma-Aldrich but were dissolved in ethanol and 4% sodium bicarbonate, respectively. MS-PPOH was prepared by Dr. J. R. Falck (University of Texas Southwestern Medical Center) and dissolved in ethanol. Triton X-100 was obtained from Calbiochem (San Diego, CA) and diluted in distilled water.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Basal perfusion pressure in the untreated group (n = 3) was 77 ± 3 mm Hg compared with 61 ± 3 mm Hg in the indomethacin group (n = 8); 74 ± 4 mm Hg for the indomethacin and L-NA groups (n = 11); 58 ± 4 mm Hg for the indomethacin, L-nitroarginine, and miconazole groups (n = 7); 62 ± 5 mm Hg for the indomethacin, L-nitroarginine, and MS-PPOH group (n = 5); and 69 ± 3 mm Hg for the indomethacin, L-nitroarginine, and tetraethylammonium groups (n = 4). The elevated perfusion pressures in the respective groups were 188 ± 6, 192 ± 7, 205 ± 6, 173 ± 7, 164 ± 16, and 207 ± 2 mm Hg. In the untreated and indomethacin-treated groups, 7.5 × 10⁷ M phenylephrine was sufficient to raise perfusion pressure; this requirement was reduced to 4 × 10⁷ M when L-nitroarginine was added and reduced to zero with the further addition of tetraethylammonium. In contrast, in kidneys treated with either miconazole or MS-PPOH, the requirement for phenylephrine to elevate perfusion pressure was increased, up to 1.5 × 10⁶ M.

Inhibition of Cyclooxygenase. Indomethacin was without effect on vasodilator responses to bradykinin, thimerosal, or nitroprusside (Fig. 1) when compared with the untreated group. For all of the subsequent interventions, comparisons were made to the indomethacin-treated group.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Vasodilator responses to bradykinin (BK), thimerosal, and nitroprusside (NP) in untreated kidneys (n = 3), those treated with indomethacin (Indo; 5.6 µM; n = 8), and those treated with indomethacin plus 50 µM L-nitroarginine (L-NA; n = 11). , p < 0.05 versus indomethacin-treated group. PP, perfusion pressure.

Inhibition of Nitric-Oxide Synthase. Inclusion of L-nitroarginine in the perfusate in addition to indomethacin reduced the vasodilator responses to thimerosal without significantly affecting those to bradykinin and nitroprusside when compared with the group treated with indomethacin alone (Fig. 1). The response to bradykinin was 92 ± 13 mm Hg and 66 ± 9 mm Hg in the absence and presence of L-nitroarginine, respectively, but the difference did not achieve significance. Similarly, the response to nitroprusside was 78 ± 10 mm Hg in the presence of L-nitroarginine compared with 51 ± 4 mm Hg in the group treated with indomethacin alone (not significant). In contrast, the responses to 1 and 10 µg of thimerosal were reduced from 54 ± 11 and 80 ± 5 mm Hg, respectively, to 14 ± 2 and 52 ± 5 mm Hg, respectively (p < 0.05) in the presence of L-nitroarginine.

Inhibition of Epoxygenase. The addition of miconazole to indomethacin and L-nitroarginine further reduced the vasodilator effect of bradykinin to 25 ± 6 mm Hg (p < 0.05), whereas there was no significant effect on the response to nitroprusside, 59 ± 12 mm Hg (not significant when compared with indomethacin alone or indomethacin and L-nitroarginine). Miconazole also caused a further reduction (p < 0.05 when compared with indomethacin plus L-nitroarginine) in the vasodilator effects of 1 and 10 µg of thimerosal, which reduced perfusion pressure by 4 ± 1 and 10 ± 2 mm Hg, respectively (Fig. 2).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Effects of inhibition of cytochrome P450 with miconazole (Micon; 1 µM; n = 7) and inhibition of epoxygenase with MS-PPOH (28 µM; n = 5) on nitric oxide- and prostaglandin-independent vasodilator responses (n = 11) to thimerosal and bradykinin (BK). , p < 0.05 versus indomethacin plus L-nitroarginine-treated group (Indo + L-NA).

Inhibition of K⁺ Channels. As expected, tetraethylammonium reduced the vasodilator effect of bradykinin to 38 ± 9 mm Hg without affecting the response to nitroprusside, 68 ± 18 mm Hg (Fig. 3). Tetraethylammonium also caused a significant reduction in the vasodilator effect of 1 and 10 µg of thimerosal, 2 ± 2 and 16 ± 9 mm Hg, respectively.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Effects of the K+ channel inhibitor, tetraethylammonium (TEA; 10 mM; n = 4), on vasodilator responses to thimerosal, bradykinin (BK), and nitroprusside (NP) in kidneys treated with indomethacin (Indo) and L-nitroarginine (L-NA). n = 11 for the indomethacin- and nitroarginine-treated group. , p < 0.05 versus indomethacin plus L-nitroarginine treatment.

Release of EETs. Figure 4 shows the increase in total EET release, measured as DHETs, in samples obtained 1 min before and 1min after challenge with 1 and 10 µg of thimerosal. These results are expressed as increases in EET release because of the large variations in release and relatively high basal levels; however, in all kidneys treated with indomethacin and L-nitroarginine (n = 6), thimerosal produced an increase in EET release. Thus, EET release was increased by 4.9 ± 1.8 ng/min and 11.9 ± 6.1 ng/min by 1 and 10 µg of thimerosal, respectively. With the addition of miconazole, the increase in EET release was 2.0 ± 1.9 ng/min (this reflected a high value in one of four experiments) and 0.4 ± 0.4 ng/min in response to 1 and 10 µg of thimerosal, respectively. Similarly, MS-PPOH (n = 5) reduced the efflux of EETs from the kidney stimulated by thimerosal. Thus, in the presence of MS-PPOH, the increase in EET release in response to 1 and 10 µg of thimerosal was reduced to 0.8 ± 0.6 ng/min and 0.7 ± 0.6 ng/min. However, MS-PPOH failed to reduce basal EET release, which was 25.2 ± 3.9 ng/min compared with 19.7 ± 3.7 ng/min in the indomethacin- and L-nitroarginine-treated group. In contrast, basal release of EETs in the miconazole-treated group was reduced to 10.7 ± 1.8 ng/min.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Increases in total EET release, measured as the DHETs, in response to thimerosal in kidneys treated with indomethacin (Indo) plus L-nitroarginine (L-NA; n = 6), indomethacin plus L-nitroarginine, and miconazole (Micon; n = 4) and indomethacin plus L-nitroarginine plus MS-PPOH (n = 5). , p < 0.05 versus indomethacin plus L-nitroarginine treatment.

Removal of Endothelium. Basal perfusion pressure was 70 ± 6 mm Hg and was elevated to 131 ± 8 mm Hg after administration of Triton X-100. After phenylephrine, perfusion pressure was raised to 206 ± 7 mm Hg. Under these conditions, the decrease in perfusion pressure in response to bradykinin (100 ng) was 5 ± 3 mm Hg and that to nitroprusside was 51 ± 13 mm Hg. The vasodilator effect of 1 µg of thimerosal was abolished, whereas 10 µg of thimerosal reduced perfusion pressure by 4 ± 3 mm Hg (in one preparation, thimerosal produced a slight increase in perfusion pressure and, therefore, the vasodilator effect was taken as zero).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results of this study show that thimerosal produces dose-dependent vasodilation of the isolated perfused kidney that is independent of prostaglandin synthesis because indomethacin was without effect. These results are in agreement with those of Forstermann et al. (1986b) and Crack and Cocks (1992) in rabbit aorta and dog coronary artery, respectively, but not those of Rosenblum et al. (1992) using mouse pial arteries, probably reflecting species and tissue differences. The prostaglandin-independent renal vasodilator effect in the presence of indomethacin consists of two components, one mediated via nitric oxide and the other via a cytochrome P450-dependent mechanism that utilizes EETs acting upon K⁺ channels. Inhibition of nitric-oxide synthase with L-nitroarginine reduced the vasodilator effect of thimerosal, showing that part of the response is mediated via nitric oxide. The nitric oxide-dependent component was more apparent with the lower dose of thimerosal, since L-nitroarginine produced a greater inhibitory effect. These results are in accord with other studies that have shown vasodilator/vasorelaxant responses to thimerosal to be dependent on nitric oxide (Forstermann et al., 1986a,b). Similarly, we confirmed that the renal vasodilator effect of bradykinin is also partially dependent on nitric oxide and that when nitric-oxide synthase is inhibited, the vasodilator effect of exogenous nitric oxide in the form of nitroprusside tends to be enhanced. The results with nitroprusside indicate that nitric oxide synthase was inhibited in these studies because removal of background levels of nitric oxide would be expected to increase responsiveness to administered nitric oxide. Also, the concentration of phenylephrine required to increase perfusion pressure was reduced by half in the presence of L-nitroarginine. Finally, an earlier study showed that this concentration of L-nitroarginine was sufficient to abolish the increase in cGMP release from the kidney in response to bradykinin (Cachofeiro and Nasjletti, 1991).

We anticipated that the residual vasodilator effect of thimerosal in the presence of a nitric-oxide synthase inhibitor would involve an EDHF as indicated by Beny (1990). Consequently, we used tetraethylammonium as a nonselective K⁺ channel inhibitor to support this idea. Tetraethylammonium, in the presence of indomethacin and L-nitroarginine, almost abolished the vasodilator effect of the lower dose of thimerosal (in two of four cases the vasodilator effect of thimerosal was converted to a small vasoconstrictor response in the presence of tetraethylammonium), fulfilling one of the criteria for an EDHF-mediated effect. However, because these experiments were conducted in a perfused organ system where it is not possible to distinguish effects at the endothelium versus the vascular smooth muscle, we cannot exclude the possibility that tetraethylammonium affected K⁺ channels on the endothelium to reduce release of a vasorelaxant mediator rather than simply preventing the effect of the mediator on the smooth muscle. Thus, we and others have shown that inhibition of endothelial K⁺ channels reduces EDHF- and nitric oxide-mediated responses (Doughty et al., 1999; Qiu and Quilley, 2001).

The results of these studies also support the concept of EETs as an EDHF in the rat kidney and demonstrate for the first time a cytochrome P450-dependent component to the vasodilator effect of thimerosal that also involves activation of K⁺ channels. The evidence for cytochrome P450 and EETs in particular is considerable. First, miconazole, which is considered a relatively specific inhibitor of epoxygenase (Zou et al., 1994), greatly reduced the vasodilator effect of thimerosal. Because agents such as miconazole have been reported to influence the activity of K⁺ channels, we also used another specific inhibitor of epoxygenase, MS-PPOH. This agent has been shown to inhibit the formation of EETs from arachidonic acid by renal cortical microsomes with little effect on the formation of 20-HETE, a -hydroxylase product (Wang et al., 1998). Like miconazole, MS-PPOH greatly reduced the vasodilator effect of thimerosal that remained following inhibition of cyclooxygenase and nitric-oxide synthase and, thereby, strongly supports a role for EETs. Second, thimerosal increased the release of EETs into the renal perfusate, and this stimulated release of EETs was prevented when kidneys were treated with either miconazole or MS-PPOH. We do not have an explanation for the failure of MS-PPOH to reduce basal release of EETs when miconazole caused a 50% reduction. It is possible that EETs released in response to thimerosal are derived from a source of phospholipids different from those released under basal conditions and that it is the stimulated release that is affected by MS-PPOH. Thus, EETs have been shown to be stored (Capdevila et al., 1987), although our results suggest that the EETs released in response to thimerosal are derived from cytochrome P450-dependent metabolism of arachidonic acid as epoxygenase inhibitors reduced both the vasodilator effect of thimerosal and the associated increase in the release of EETs.

Thimerosal is an inhibitor of acyl transferase and as such would be expected to increase levels of free arachidonic acid that would then be available for metabolic transformation by epoxygenase, which is expressed principally in the endothelium, and by -hydroxylase, which is localized to vascular smooth muscle (Roman, 2002). Thimerosal should, therefore, increase the formation of dilator EETs and constrictor 20-HETE unless it exerts a preferential effect on the endothelium. Removal of the endothelium almost abolished the vasodilator effects of thimerosal in kidneys treated with L-nitroarginine and indomethacin. Under these conditions, the vasoconstrictor effect of 20-HETE would no longer be opposed by endothelial-derived nitric oxide or EETs. Indeed, 20-HETE formation should be increased as a result of removal of an inhibitory influence in the form of nitric oxide (Oyekan et al., 1999). However, it should also be noted that indomethacin has been reported to inhibit the vasoconstrictor effect of 20-HETE in the rat isolated perfused kidney (Askari et al., 1997) and may mask the activity.

The effects of thimerosal in increasing EET release cannot be attributed solely to inhibition of reacylation because it has also been reported to increase the sensitivity of the inositol triphosphate receptor (Montero et al., 2001) and to increase levels of intracellular Ca²⁺ (Gericke et al., 1993). These actions could result in activation of phospholipases to release arachidonic acid as well as stimulation of endothelial nitric-oxide synthase, a Ca²⁺-dependent enzyme. Regardless of the mechanism by which arachidonic acid is released, metabolism via cyclooxygenase as well as cytochrome P450 could be expected unless coupling of substrate to enzyme is distinct.

We measured total EET release after their conversion to DHETs by acid hydrolysis; therefore, we cannot attribute the vasodilator action of thimerosal to any specific regioisomer, although previous studies have shown the 5,6-EET to be the most potent of the EETs. EETs have been shown to be released from the endothelium in response to some endothelium-dependent vasodilator agents and to relax vascular smooth muscle via activation of K⁺ channels, thereby fulfilling the basic requirements for an EDHF (Campbell and Gauthier, 2002). In these studies, we found that removal of the endothelium abolished the vasodilator effect of thimerosal in agreement with the results of Forstermann et al. (1986a). Although the results of the present studies provide convincing evidence for a role of EETs in the nitric oxide-independent vasodilator effect of thimerosal, we cannot exclude the possibility that EETs may function as an intracellular mediator in the endothelium to activate K⁺ channels and increase the influx of Ca²⁺. However, this concept, which was first suggested by Graier et al. (1995) and supported by Rzigalinski et al. (1999), is not supported by studies showing that EETs cause vasodilation that is not dependent on an intact endothelium.

In summary, we have provided further evidence for one or more EETs acting as an EDHF in the rat kidney by showing that thimerosal, an agent that increases free arachidonic acid levels, elicits endothelium-dependent vasodilation that is associated with increased release of EETs and that both effects are attenuated by inhibitors of epoxygenase.