Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Hypoxia and intracellular pH (pH_i) are known to significantly alter vascular responses (Rubanyi and Paul, 1985; Hashimoto et al., 1993; Nagesetty and Paul, 1994). Vascular endothelium plays an important role in modulation of vascular tone by releasing a variety of vasoactive factors, such as endothelium-derived relaxing factor (Furchgott and Zawadzki, 1980; Vanhoutte and Rimele, 1983). The endothelium also probably functions as a sensor of O₂ tension and pH in the arterial wall. Our work and that of others has shown that hypoxia attenuates endothelium-dependent relaxation induced by A23187, Substance P (SP), acetylcholine, and thrombin (deMey and Vanhoutte, 1978; Johns et al., 1989; Hashimoto et al., 1993). The inhibition of endothelium-dependent relaxation by hypoxia probably occurs at the level of the vascular endothelium because relaxation induced by sodium nitroprusside (SNP), a direct activator of guanylate cyclase, is unimpaired by hypoxia (Hashimoto et al., 1993). However, the mechanisms by which hypoxia inhibit endothelium-dependent relaxation are not clear.

Endothelium-derived relaxation factor has been identified as nitric oxide (NO) or related compounds (Ignarro et al., 1987; Palmer et al., 1987). NO activates soluble guanylate cyclase and associated cGMP-mediated relaxation of vascular smooth muscle (Ignarro et al., 1984; Murad, 1986). However, agonist-stimulated endothelium-dependent relaxation also includes component(s) resistant to inhibitors of NO synthase and guanylate cyclase (Nishiye et al., 1989; Rees et al., 1989). The NO pathway-resistant relaxation is abolished by high K⁺ solutions or the nonselective K⁺ channel inhibitor tetrabutylammonium chloride in porcine coronary artery (Nagao and Vanhoutte, 1992), and by a combination of the K⁺ channel inhibitors charybdotoxin and apamin in rat hepatic artery (Zygmunt and Högestätt, 1996) and guinea pig basilar artery (Petersson et al., 1997). This relaxing factor(s) has been termed the endothelium-dependent hyperpolarizing factor (EDHF) in porcine coronary artery and other tissues (for review, see Cohen and Vanhoutte, 1995). Thus, there may be at least two components in SP-induced relaxation, including NO and EDHF in porcine coronary artery.

Using novel methodology involving selective loading of fluorescent dyes, Foy et al. (1997) showed that hypoxia increased pH_i in endothelium in situ on intact arteries. Whether the inhibition of endothelium-dependent relaxation by hypoxia can be attributed to this alkalinization is not known. Moreover, the effects of hypoxia and/or alkalinization on endothelium-derived nitric oxide (EDNO) or EDHF per se are not known. In this study, we thus tested the hypothesis that hypoxia and the concomitant alkalinization differentially affect the responses attributable to EDNO and EDHF. We used novel methodology to measure endothelial cell [Ca²⁺]_i in situ on the intact coronary artery to probe the mechanism of the differential inhibitory effects observed.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Preparation of Arterial Rings. Porcine hearts were obtained from a local slaughterhouse and transported to the laboratory immersed in ice-cold lactated Ringer's solution. The distal portions of left anterior descending coronary artery were dissected from the hearts and were cleaned of fat and adhering connective tissue, with care taken to prevent stretching of the vessel. Blood clots in the lumen were flushed out with ice-cold oxygenated (95% O₂/5% CO₂) Krebs-bicarbonate buffer. The arteries were then cut into rings 4 to 5 mm in length. In some experiments, the arterial rings were everted and the endothelium was removed by gently rubbing with cotton. Arteries were stored in Krebs-bicarbonate buffer at 4°C for up to 2 days; these stored arteries are similar to that of fresh tissue (Hashimoto et al., 1992; Shimizu et al., 1997).

Isometric Force Measurements. Arterial rings were mounted on two stainless steel wires, one stationary and the other attached to a force transducer (Kistler-Morse DSK-6, Bellevue, WA). Output from the force transducer was recorded on dual channel recorder (Linear 1200; Linear Instruments, Irvine, CA) and a computer data collection (BioPac) and analysis (AcqKnowledge III) system (Goletta, CA). The rings were incubated in Krebs-bicarbonate buffer in 15-ml water-jacketed chambers. The solution was maintained at 37°C and bubbled with 95% O₂/5% CO₂ for aerobic conditions. Hypoxic conditions were defined by bubbling with 95% N₂/5% CO₂ (PO₂ ~1-2%). The rings were equilibrated for 90 min and were periodically stretched to obtain a resting tension of 40 mN, optimal for isometric force (Rubanyi and Paul, 1985). After the equilibration period, the arteries were contracted with 80 mM KCl until successive challenges gave stable contractions of equivalent magnitudes. Endothelial integrity and the effectiveness of the denudation technique were assessed by the addition of SP (10 nM) to arteries precontracted with KCl.

2

Measurement of [Ca²⁺]_i in Endothelium on Intact Porcine Coronary. Endothelial [Ca²⁺]_i was measured with the Ca²⁺-sensitive fluorescent dye fura-2 by using a modification technique for endothelial pH_i measurement (Foy et al., 1997). The coronary ring was cut open and sutured endothelial side out onto a U-shaped hook that was connected to a Teflon holder. After equilibration, the tissue was incubated in a 3-(N-morpholino)propanesulfonic acid (MOPS)-buffered physiological saline solution (pH 7.4) containing 10 µM fura-2 acetoxymethylester, 0.02% cremophor, and 2 mg/ml BSA for 20 to 30 min at room temperature in a cuvette (2.4 ml) with stirring. The preparation was then mounted in the sample compartment of a spectrofluorimeter (Photon Technologies, Inc., Santa Clara, CA) configured for front face measurement and continuously perfused with Krebs-bicarbonate buffer (pH 7.4), maintained at 37°C and gassed with 95% O₂ and 5% CO₂. After 15 to 30 min, fluorescence intensity was measured at excitation wavelengths of 340 and 380 nm and an emission wavelength of 510 nm.

[Ca²⁺]_i Calibration. The fluorescence intensity at 340 nm excitation was divided by that measured at 380 nm, and this ratio was used as an index of [Ca²⁺]_i. For statistical analysis, the ratio of the SP response was divided by the maximum ratio obtained with the Ca²⁺ ionophore ionomycin (5 µM) and data are shown as a percentage of ionomycin response.

Measurement of pH_i in Coronary Artery Smooth Muscle. pH_i in coronary artery smooth muscle was measured as described in Nagesetty and Paul (1994). The coronary ring was everted and the endothelium was removed by gently rubbing with cotton. The preparation was isometrically mounted on a U-shaped wire and incubated in Krebs-bicarbonate buffer for at least 1 h. At end of the equilibration period the U-shaped wire with tissue was connected to a Teflon holder and placed in a cuvette (2.4 ml). The tissue was oriented perpendicular to the beam of light for front face measurements in a spectrofluorimeter (Photon Technologies, Inc.) and perfused with Krebs-bicarbonate buffer. After the autofluorescence of the preparation was measured, tissue perfusion was discontinued and 2',7'-biscarboxyethyl-5(6)-carboxyfluorescein tetraacetoxymethylester (BCECF-AM, 5 µM) was added to the cuvette. Fluorescence was measured at an emission wavelength of 523 nm with excitation wavelengths of 439 and 505 nm. Loading was stopped when the 505-nm signal was at least 10 times the baseline at which time the 439-nm signal was at least 3 times baseline. Perfusion of the tissue was reinitiated to wash out the unhydrolyzed BCECF-AM. After a stable baseline was achieved, pH_i was measured.

pH_i Calibration. Absolute values of pH_i were determined with an adaptation of the high K⁺-nigericin technique reported in Thomas et al. (1979). At the end of the experiment, nigericin (5 µM) was added to the cuvette and a calibration curve was constructed with high K⁺-MOPS-buffered solutions of known pH values (6.80-7.81). The ratio of the fluorescence intensities at 505 nm and 439 nm minus the tissue autofluorescence was nearly linearly related to pH_i. Photon Technologies, Inc. software "look up tables" (i.e., segmental lines fitted to the calibration data) were used to convert fluorescent intensity ratios to pH_i values for each individual preparation.

Solutions and Chemicals. The Krebs-bicarbonate solution (pH 7.4) used in these experiments had the following composition: 118 mmol/l NaCl, 4.7 mmol/l KCl, 2.5 mmol/l CaCl₂, 1.18 mmol/l KH₂PO₄, 1.18 mmol/l MgSO₄, 15 mmol/l NaHCO₃, 0.026 mmol/l EDTA, and 11 mmol/l glucose. Decreasing pH from 7.4 to 7.2 was carried out by decreasing HCO₃ in the Krebs' bicarbonate buffer. The MOPS-buffered physiological saline solution used had the following composition: 140 mmol/l NaCl; 4.7 mmol/l KCl; 1.2 mmol/l NaH₂PO₄; 20 mmol/l MOPS; 0.02 mmol/l EDTA; 1.2 mmol/l MgSO₄; 2.5 mmol/l CaCl₂; and 11 mmol/l glucose, pH 7.4. The drugs used in this study were U46619 (in ethanol), SP (in water), L-NNA (dissolved in Krebs-bicarbonate solution), tetrabutylammonium chloride (dissolved in Krebs-bicarbonate solution), indomethacin (in ethanol), quinacrine (in distilled water), SKF-525a (dissolved in Krebs-bicarbonate solution), fura-2 AM (in dimethyl sulfoxide), BCECF-AM (in dimethyl sulfoxide), and nigericin (in ethanol). The final concentration of vehicle, acetone, and ethanol did not exceed 0.1%, below which no vehicle effects were found (Hashimoto et al., 1993).

Statistical Analysis. Data are presented as mean ± S.E. Differences between groups were assessed by one-way ANOVA or two-tailed Student's t test for paired data as appropriate. Differences among means were considered significant when P < .05.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Components of SP-Induced Endothelium-Dependent Relaxation and Effects of Alkalinization by NH₄Cl. We first used a pharmacological approach to identify the components of SP-induced relaxation in KCl and U46619 contractures. Figure 1 presents the force records for a typical experiment and the average values are summarized in Table 1. For contractures elicited by 30 mM KCl, SP (10 nM) relaxed intact but not de-endothelialized coronary arteries. The endothelium-dependent SP-induced relaxations were inhibited by 0.2 mM L-NNA, a potent inhibitor of NO synthase (Fig. 1A). In contrast, in U46619 contractures (Fig. 1B), the magnitude of the decrease in force in response to SP (the peak of relaxation) was not inhibited by 0.2 mM L-NNA. However, the SP response was nearly abolished by further addition of 5 mM TBA, a nonselective inhibitor of K⁺ channels.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   A typical isometric force recording showing the effect of L-NNA (0.2 mM) and TBA (5 mM) on 10 nM SP-induced relaxation in porcine coronary artery contracted with 30 mM KCl or 0.1 µM U46619. A, for KCl contractures, the SP response is nearly abolished by L-NNA, suggesting that it largely reflects EDNO. B, for U46619 contractures, SP still elicits a strong response in the presence of L-NNA, which is blocked by TBA. The component resistant to L-NNA is our operational definition of EDHF.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effect of intracellular alkalinization with NH4Cl (30 mM) on EDNO and EDHF components of SP (10 nM)-induced relaxation. A, effect of 30 mM NH4Cl in 30 mM KCl contractures. B, effect of 30 mM NH4Cl in 0.1 µM U46619 contractures. C, effect of 30 mM NH4Cl in the presence of 0.2 mM L-NNA in 0.1 µM U46619 contractures. L-NNA (0.2 mM), TBA (5 mM), or L-NNA plus TBA were added at least 15 min before stimulation by addition of KCl or U46619. Comparison of A (EDNO) and C (EDHF) suggested that alkalinization inhibits the relaxation to EDNO but not that in response to EDHF. A statistical summary of these types of experiments is presented in Table 1.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Effect of decreasing pHo (from 7.4 to 7.2) on NH4Cl (30 mM)-induced increase in pHi in the smooth muscle of endothelium-denuded porcine coronary artery. A, effects of KCl (30 mM) and NH4Cl (30 mM) on pHi in pH 7.4 Krebs-bicarbonate solution. B, effects of KCl (30 mM) and NH4Cl (30 mM) on pHi in pH 7.2 Krebs-bicarbonate solution. Note that by reducing pHo the steady state pHi after the addition of NH4Cl is close to the original control levels in pHo 7.4 Krebs-bicarbonate solution. Average data for five arteries are presented in Table 2.

Does Alkalinization by NH₄Cl Affect SP-Induced Increase in Endothelial [Ca²⁺]_i? NO synthesis in vascular endothelium has been reported to depend on [Ca²⁺]_i (Mayer et al., 1989; Schmidt et al., 1992). We tested this hypothesis in the intact artery by measuring the effects of SP on endothelial [Ca²⁺]_i, with our protocol (Foy et al., 1997) for loading fura-2 only into the endothelium of an intact porcine coronary artery (Fig. 4A). Addition of 30 mM KCl did not affect the basal fluorescence ratio (340:380), whereas 10 nM SP significantly increased the fluorescence ratio of the endothelium-loaded preparation (Fig. 4A). However, after rubbing the artery to remove the endothelium, addition of SP did not elicit a response. In sharp contrast, when the de-endothelialized artery was loaded with fura-2, the smooth muscle responded to 30 mM KCl with a sustained increase in the 340:380 ratio, whereas SP was without effect (Fig. 4B). Vascular endothelial cells are generally suggested to lack voltage-activated Ca²⁺ channels (Adams et al., 1989). These data attest to the specificity of our dye loading for the endothelial cell layer of the intact artery.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effects of SP (10 nM) on [Ca2+]i (340:380 ratio) in the presence of KCl (30 mM) in porcine coronary artery. A, endothelium. Note that there is little response to KCl, whereas SP elicits a significant effect in the endothelium-loaded preparation. After rubbing to remove the endothelium, SP no longer affects the 340:380 ratio. B, smooth muscle. In contrast to the endothelial cell behavior, KCl elicits a strong increase in the 340:380 ratio in the endothelium-denuded, smooth muscle-loaded preparation, whereas SP has only minor effects. These cell type-specific [Ca2+]i responses support the selective loading of the dye.

View larger version (19K): [in this window] [in a new window]   Fig. 5.   Effect of NH4Cl (30 mM) on the SP (10 nM)-induced increase in [Ca2+]i in the presence of KCl (30 mM) in intact porcine coronary artery. A, SP response in the presence of 30 mM NH4Cl. B, control. For four arteries in each case, endothelial cell [Ca2+]i was 39.9% ± 3.9% of the ionomycin response for the control and 42.5% ± 2.3% in the presence of NH4Cl. Alkalinization did not affect the SP-induced increase in [Ca2+]i.

Effect of Alkalinization by NH₄Cl on SNP-Induced Relaxation. To determine whether alkalinization by NH₄Cl inhibits EDNO-mediated relaxation upstream or downstream of activation of guanylate cyclase, we investigated the effect of 30 mM NH₄Cl on the SNP-induced relaxation in de-endothelialized coronary arteries in 30 mM KCl contractures (Fig. 6). SNP releases NO, which activates guanylate cyclase. SNP elicited a concentration-dependent relaxation that was shifted rightward by 30 mM NH₄Cl. The effects of NH₄Cl were attenuated by decreasing pH_o from 7.4 to 7.2, as was the case for the SP-induced relaxation.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   The effects of decreasing pHo on the attenuation of SNP-induced relaxation by NH4Cl (30 mM). Endothelium-denuded arteries were contracted with 30 mM KCl and SNP was added in a cumulative manner (control, upper curve). Values are expressed as mean ± S.E., n = 5. Alkalinization (+NH4Cl, lower curve) decreases the relaxation to the NO donor, which is partially reversed by decreasing pHo (pH 7.2, middle curve) to compensate for the NH4Cl-induced increase in pHi.

Effects of Hypoxia on Endothelium-Dependent Relaxation. We have shown that hypoxia inhibits SP-induced relaxation when arteries were contracted with KCl (Hashimoto et al., 1993). To determine the effects of hypoxia on EDHF-dependent relaxation, we investigated the effect of hypoxia on SP-induced relaxation in U46619 contractures (Fig. 7 and Table 3). As we have previously shown, hypoxia relaxed steady-state contractions independent of the kind of stimulation. The degree of the hypoxia-induced relaxation in U46619 contractures was larger than that in KCl contractures. In 50-mM KCl contractures, hypoxia inhibited the SP-induced relaxation (Hashimoto et al., 1993). However, the SP-induced relaxation was observed under hypoxic conditions for U46619 contractures and this relaxation was not inhibited by 0.2 mM L-NNA but was inhibited by 5 mM TBA.

View larger version (11K): [in this window] [in a new window]   Fig. 7.   Effects of hypoxia (bubbling with N2 instead of O2) on the EDNO and EDHF components of SP (10 nM)-induced relaxation. A, 30 mM KCl contractures. B, 0.1 µM U46619 contractures. C, Effect of TBA (5 mM) before a 0.1 µM U46619 contracture. D, effect of 0.2 mM L-NNA in 0.1 µM U46619 contractures. L-NNA or TBA were added at least 15 min before stimulation with U46619. This experiment suggests that hypoxia inhibits the relaxation to EDNO but not that in response to EDHF. A statistical summary of these types of experiments is presented in Table 3.

Effects of Hypoxia on SP-Induced Increase in Endothelial [Ca²⁺]_i. Figure 8 shows the effect of hypoxia on the SP-induced increase in endothelial [Ca²⁺]_i in the presence of 30 mM KCl. Addition of 30 mM KCl did not affect a basal fluorescence ratio (340:380). SP (10 nM) significantly increased the fluorescence ratio. Importantly, however, the SP-induced [Ca²⁺]_i response was not affected by hypoxia; 39.9% ± 1.2% versus 42.6% ± 2.6% of the ionomycin response for control and hypoxia, respectively (n = 4).

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Effects of hypoxia on the SP (10 nM)-induced increase in endothelial cell [Ca2+]i in the presence of 30 mM KCl in the dye-loaded endothelium of intact porcine coronary artery. A, SP response during hypoxia. B, control. For four arteries in each case, endothelial cell [Ca2+]i was 39.9% ± 1.2% of the ionomycin response for the control and 42.6% ± 2.6% during hypoxia. Hypoxia does not affect the SP-induced increase in endothelial cell [Ca2+]i.

Is EDHF a P-450-Derived Arachidonic Acid Metabolite? EDHF has not been identified. However, it has been reported that EDHF is synthesized by a P-450-dependent mechanism via arachidonic acid metabolism in various arteries, including bovine coronary artery (Hecker et al., 1994; Campbell et al., 1996) and porcine coronary artery (Hecker et al., 1994). NO synthase also has P-450-like domain. Interestingly, despite the inhibition by alkalinization or hypoxia of the EDNO response, the EDHF response was not similarly affected. Therefore, we investigated whether EDHF also is synthesized via a P-450 pathway in porcine coronary artery (Table 4). Pretreatment with 10 µM indomethacin (a cyclooxygenase inhibitor), 30 µM quinacrine (a phospholipase A₂ inhibitor), or 30 µM SKF-525a (a P-450 inhibitor) did not affect the 0.1 µM U46619-induced contraction. Importantly, the component of the SP relaxation seen in the presence of 0.2 mM L-NNA also was not affected by these interventions.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Hypoxia inhibits the endothelium-dependent relaxation elicited by SP in KCl contractures in porcine coronary artery (Hashimoto et al., 1993). Foy et al. (1997) previously showed that hypoxia increases pH_i in endothelium in situ on intact arteries. This inhibition probably occurs at the level of the vascular endothelium because the relaxation of the smooth muscle by the NO donor SNP, which directly activates guanylate cyclase, was not affected by hypoxia (Hashimoto et al., 1993). In this study, we tested the hypothesis that hypoxia and the concomitant alkalinization inhibit endothelium-dependent relaxation in porcine coronary artery.

We first determined the components of SP responses in KCl and U46619 contractures. In KCl contractures, the SP-induced relaxation was abolished by inhibition of NO synthase. Thus, the major component of the SP-induced relaxation in KCl contractures is probably mediated by NO. However, when arteries were contracted with U46619, the major effect of NO synthase inhibition was to shorten the duration of the SP response (Table 1). Further addition of a nonselective K⁺ channel inhibitor blocked the remaining relaxation. It has been suggested that a hyperpolarizing factor is released from endothelium in various arteries (Brayden, 1990; Chen et al., 1991; Nagao and Vanhoutte, 1992; Shimizu and Paul, 1997). Nagao and Vanhoutte (1992) showed that the use of L-NNA, TBA, or high K⁺ permits differentiation between endothelium-dependent relaxations mediated by NO and hyperpolarization in porcine coronary artery. Thus, in KCl contractures, EDHF would be attenuated and the remaining component of the SP response may be ascribable to NO. However, in U46619 contractures, the SP response is attributable to both NO and EDHF. The contribution of prostacycline is unlikely because indomethacin, an inhibitor of cyclooxygenase, did not affect either the duration or the peak of the SP response.

Hypoxia and NH₄Cl nearly abolished the SP-induced relaxation when arteries were contracted with KCl. We used NH₄Cl to increase pH_i without altering pH_o (Thomas, 1974). This standard technique is based on the rapid diffusion of NH₃ into the cell and its re-equilibration with NH₄⁺. Ando and Fujita (1994) also showed that addition of NH₄Cl inhibited acetylcholine-induced relaxation in rat aorta and the inhibitory effect was blocked by the H⁺ ionophore nigericin, which increases intracellular H⁺ concentration. In this study, decreasing pH_o, such that the increase in pH_i in response to NH₄Cl returned pH_i to the original control levels, attenuated the effects of NH₄Cl on SP-mediated relaxation. This finding implies that intracellular alkalinization is probably the most important factor in the inhibition by NH₄Cl of the SP-induced, NO-dependent relaxation. Thus, hypoxia or intracellular alkalinization by NH₄Cl inhibit SP-induced NO-dependent relaxation.

In contrast, the EDHF response was not affected by either hypoxia or intracellular alkalinization by NH₄Cl. We showed that EDHF-mediated relaxation in porcine coronary artery is inhibited by 4-aminopyridine, suggesting that it probably involves K_v channels (Shimizu and Paul, 1997). Based on our data in this study, these K_v channels are not sensitive to O₂ or alkaline pH, thus limiting the potential members of this K⁺-channel family that may be involved.

The release of NO requires an increase in [Ca²⁺]_i in endothelial cells (Mayer et al., 1989; Förstermann et al., 1991; Schmidt et al., 1992). Acute hypoxia has been shown to reduce Ca²⁺ influx and to decrease [Ca²⁺]_i in bovine pulmonary artery endothelial cells (Stevens et al., 1994) but has been suggested to increase [Ca²⁺]_i in human endothelial cells (Arnould et al., 1992). These reports on endothelial [Ca²⁺]_i used cultured endothelial cells. However, there may be differences between the responses of cultured cells and those of the endothelium in contact with its normal smooth muscle substratum in an intact artery. We measured endothelial [Ca²⁺]_i in situ with fura-2 in intact porcine coronary artery. In arteries in which fura-2 was loaded with our endothelium loading protocol, infusion of high K⁺ solution did not alter the basal fluorescence ratio, in contrast to that observed for fura-loaded smooth muscle of the endothelium-denuded preparation. This result would be anticipated because endothelial cells are generally considered to lack voltage-activated Ca²⁺ channels (Adams et al., 1989). However, SP increased endothelial [Ca²⁺]_i, whereas the smooth muscle was unaffected. These results confirm known smooth muscle and endothelial cell behavior and attest to the selectivity of our loading protocol for the endothelium. Thus, we have developed a novel method of measurement of [Ca²⁺]_i, in addition to pH_i (Foy et al., 1997), in endothelium in situ on intact coronary artery.

Despite the inhibition by NH₄Cl and hypoxia of the SP response in KCl contractures, endothelial [Ca²⁺]_i responses were unchanged. Thus, a mechanism involving attenuation of the endothelial [Ca²⁺]_i response to SP does not underlie the inhibition by NH₄Cl alkalinization or hypoxia of the SP-induced NO-dependent relaxation. The production of EDHF also is reported to be dependent on an increase in [Ca²⁺]_i in endothelial cells (Chen and Suzuki, 1990). Thus, the lack of inhibition by hypoxia or intracellular alkalinization of the SP-induced EDHF response is consistent with their lack of effect on endothelial [Ca²⁺]_i .

Foy et al. (1997) found that in the presence of KCl, hypoxia increased pH_i from 6.8 to 7.0 in the endothelium of intact arteries. Fleming et al. (1994), however, reported that endothelial NO synthase is activated by intracellular alkalinization at basal levels of [Ca²⁺]_i with the pH optimum of 7.5. Therefore, intracellular alkalinization by hypoxia or NH₄Cl may not inhibit NO synthase activity directly. In endothelium-denuded arteries, alkalinization with NH₄Cl shifted the SNP concentration-relaxation relation rightward. In contrast, hypoxia does not affect the SNP-induced relaxation (Hashimoto et al., 1993) or S-nitroso-N-acetyl-DL-penicillamine-induced relaxation (data not shown) in endothelial cell-denuded arteries. This finding is consistent with the lack of effect of hypoxia on pH_i in coronary artery smooth muscle (Foy et al., 1997). These findings suggest that inhibition of the SP relaxation by hypoxia and by NH₄Cl alkalinization involves different mechanisms. Intracellular alkalinization in endothelial cells by hypoxia is not likely to be involved in the attenuation of the SP-induced EDNO-dependent relaxation. Moreover, our results suggest that it is not downstream from guanylate cyclase. Production of NO requires molecular oxygen as one of the substrates for the enzyme. Rengasamy and Johns (1991) demonstrated that hypoxia inhibits NO synthase activity primarily through depletion of oxygen. Therefore, depletion of oxygen may be the most important mechanism of the inhibition by hypoxia of SP-induced EDNO-dependent relaxation.

In contrast to hypoxia, alkalinization with NH₄Cl shifted the SNP concentration-relaxation relation rightward in endothelium-denuded arteries. This shift was partially reversed by decreasing pH_o. This finding suggests that at least part of the inhibition of the EDNO component of the SP relaxation by alkalinization may involve guanylate cyclase activation or other mechanisms of smooth muscle relaxation downstream from this point.

In this study, we showed that hypoxia inhibits SP-induced EDNO but not EDHF relaxation. NO is synthesized by NO synthase, which has a cytochrome P-450 monooxygenase-like domain (Bredt et al., 1991). Interestingly, EDHF also has been suggested to be a cytochrome P-450-derived arachidonic acid metabolite in various arteries, including bovine coronary artery (Hecker et al., 1994; Campbell et al., 1996) and porcine coronary artery (Hecker et al., 1994). However, there are many reports that argue that EDHF is not a cytochrome P-450 metabolite of arachidonic acid (Corriu et al. 1996; Fukao et al., 1997; Yamanaka et al., 1998). Edwards et al. (1998) reported that K⁺ is an EDHF in rat arteries. Thus, whether EDHF is a product of the P-450 pathway is controversial. Cytochrome P-450 requires molecular oxygen as a substrate, and the oxygen sensitivity of the enzyme is similar to that of NO synthase (McGiff, 1991; Moncada and Higgs, 1993). However, EDNO- but not EDHF-mediated relaxation was inhibited by hypoxia. Petersson et al. (1998) also reported that EDHF-mediated relaxation to acetylcholine or the calcium ionophore A23187 were not inhibited by hypoxia in guinea pig basilar artery. These findings suggested that EDHF may not be synthesized by the P-450 pathway in these arteries.

In conclusion, we have shown that hypoxia and intracellular alkalinization by NH₄Cl inhibit EDNO- but not EDHF-mediated relaxation. The mechanism of inhibition of the EDNO component of relaxation by hypoxia and by alkalinization are probably attributable to different mechanisms. Our results suggest that EDHF may not be produced by a P-450 pathway in porcine coronary artery because the SP-induced EDHF relaxation was not inhibited by hypoxia or cytochrome P-450 inhibitors. It is interesting to speculate that EDHF mechanisms may have evolved to compensate for the failure of EDNO relaxation mechanisms under hypoxia or alkaline conditions.