To evaluate the role of a potential interaction between superoxide anion (O2−) and nitric oxide (NO) in regulating kidney function, we examined the renal responses to intra-arterial infusion of a superoxide dismutase mimetic, tempol (0.5 mg·kg−1·min−1), in anesthetized dogs treated with or without NO synthase inhibitor, Nω-nitro-l-arginine (NLA; 50 μg·kg−1·min−1). In one group of dogs (n = 10), tempol infusion alone for 30 min before NLA infusion did not cause any significant changes in renal blood flow (RBF; 5.2 ± 0.4 to 5.0 ± 0.4 ml·min−1·g−1), glomerular filtration rate (GFR; 0.79 ± 0.04 to 0.77 ± 0.04 ml·min−1·g−1), urine flow (V; 13.6 ± 2.1 to 13.9 ± 2.5 μl·min−1·g−1), or sodium excretion (UNaV; 2.4 ± 0.3 to 2.2 ± 0.3 μmol·min−1·g−1). Interestingly, when tempol was infused in another group of dogs (n = 12) pretreated with NLA, it caused increases in V (4.4 ± 0.4 to 9.7 ± 1.4 μl·min−1·g−1) and in UNaV (0.7 ± 0.1 to 1.3 ± 0.2 μmol·min−1·g−1) without affecting RBF or GFR. Although NO inhibition caused usual qualitative responses in both groups of dogs, the antidiuretic (47 ± 5 vs. 26 ± 4%) and antinatriuretic (67 ± 4 vs. 45 ± 11%) responses to NLA were seen much less in dogs pretreated with tempol. NLA infusion alone increased urinary excretion of 8-isoprostane (13.9 ± 2.7 to 22.8 ± 3.6 pg·min−1·g−1; n = 7), which returned to the control levels (11.6 ± 3.4 pg·min−1·g−1) during coadministration of tempol. These data suggest that NO synthase inhibition causes enhancement of endogenous O2− levels and support the hypothesis that NO plays a protective role against the actions of O2− in the kidney.
- renal hemodynamics
- sodium excretion
- superoxide dismutase mimetic
superoxide anion (O2−), a free radical, is a product of cellular metabolism generated due to partial reduction of molecular oxygen (5, 7). The presence of superoxide dismutase (SOD) enzyme in the living tissue usually reduces O2− generated in the body (1, 6, 13). However, recent experimental studies point to the notion that nitric oxide (NO) also acts as an antioxidative agent in the body (12, 32). In the kidney, O2− generation was also shown to cause decreases in blood flow and sodium excretion (36). Recently, we have demonstrated that SOD inhibition with diethyldithiocarbamate in anesthetized dogs reduces both renal cortical and medullary blood flow as well as urinary sodium excretion, which were significantly enhanced during NO synthase inhibition (16). These findings indicate that NO plays a protective role against the actions of O2− in the kidney. It has also been shown that O2−, which is locally produced in the juxtaglomerular apparatus, limits NO signaling from the macula densa and enhances tubuloglomerular feedback responses (32, 34). Because O2− and NO are continually produced and exert opposing effects in the kidney, it is possible that the basal production of O2− and its interaction with NO are regulated physiologically in the kidney. Although O2− has been implicated in the pathogenesis of many renal diseases (1, 13, 29), the possible involvement of O2− in the physiological regulation of renal function is not yet clearly defined.
The present investigation was designed to examine the hypothesis that the O2− level in the kidney is reciprocally regulated by intrarenal NO activity and that the interaction between these free radicals influences renal hemodynamics and excretory function. In these experiments, the renal responses to scavenging of O2− elicited with intra-arterial administration of a SOD mimetic, 4-hydroxy-tetramethyl piperidime-1-oxyl (tempol), were evaluated before and during NO synthase inhibition with Nω-nitro-l-arginine (NLA) in anesthetized dogs (16, 17). Tempol is a low-molecular-weight nitroxide compound that is a membrane-permeable, metal-independent, stable SOD mimetic that reduces endogenous O2− levels as reported in many in vitro and in vivo studies (4, 27, 28, 32, 34).
The experiments were performed in accordance with the guidelines and practices established by the Tulane University Animal Care and Use Committee.
Experiments were conducted in mongrel dogs (11–21 kg body wt) of either gender. These dogs were given supplemental amounts of sodium chloride (1.5 g·kg body wt−1·day−1 for 3 days) added to the normal laboratory diet to achieve a positive sodium balance as reported in earlier studies in our laboratory (16, 17). On the day of experiments, pentobarbital sodium was administered intravenously at 30 mg/kg body wt for induction of anesthesia and was supplemented throughout the experiment as needed. The level of anesthesia in animals was frequently checked by eliciting corneal reflex during the entire experimentation period to determine the need of supplemental doses of pentobarbital sodium. A cuffed endotracheal tube was inserted and connected to an artificial ventilator set at a rate of 18 strokes/min with a stroke volume of 15 ml/kg body wt. Body temperature was maintained within normal range using an electric heating pad. Systemic arterial pressure (SAP) was monitored via a catheter inserted into the right femoral artery and connected to a Statham pressure transducer (P23DC) and recorded on a polygraph (model 7D, Grass Instrument) as well as by the Biopac data-acquisition system. The left femoral artery was cannulated for collection of blood samples. The left jugular and the right femoral veins were cannulated for administration of inulin and isotonic saline, respectively. Isotonic saline was infused into the right femoral vein at a rate of 0.5 ml/min during the entire experimental period.
The left kidney was exposed retroperitoneally and denervated by tying and cutting all the renal nerves projecting to the kidney from the aorticorenal ganglion (15, 16, 17). Renal blood flow (RBF) was measured by placing an electromagnetic flow probe (Carolina Medical Electronics) around the renal artery, which was isolated from surrounding tissue. A curved 23-gauge needle cannula was inserted into the renal artery and connected to a pressure transducer for measurement of renal arterial pressure. Additional catheters were connected to the needle cannula for continuous infusion of heparinized saline (0.4 ml/min) as well as to prevent clotting in the cannula tip and for the intrarenal administration of drugs (tempol, NLA). Urine was collected from a catheter placed in the left ureter. After completion of the surgical procedures, a dose (1.6 ml/kg) of 2.5% solution of inulin in normal saline was administered into the jugular vein at least 45 min before the initiation of the experimental protocol followed by a continuous infusion (0.03 ml·kg−1·min−1) for the entire protocol period to determine the glomerular filtration rate (GFR).
After a stabilization period of 45 min after completion of the surgical procedures, the experimental protocol began with two consecutive 10-min urine collections. An arterial blood sample (2 ml) was collected at the midpoint of each urine collection period to measure plasma inulin, sodium, and potassium concentrations. In one group of dogs (n = 10), an infusion of tempol was initiated intra-arterially (0.5 mg·kg−1·min−1; Sigma Chemical) in the kidney and continued for the duration of the experimental period. This dose of tempol was found effective in reducing urinary 8-isoprostane (a biological marker for endogenous O2− activity) excretion in these anesthetized dog preparations (see Fig. 5). It was also reported that similar doses of tempol caused reduction in 8-isoprostane excretion rate that was associated with an effective reduction in blood pressure in spontaneously hypertensive rats (27). Ten minutes after the initiation of the tempol infusion, two 10-min urine collections were made with corresponding arterial blood sample collections. Then NLA (50 μg·kg−1·min−1; Sigma Chemical) was added to the infusion line with tempol to examine the responses to combine reduction in NO and O2− levels on renal hemodynamics and excretory function (16, 17). After 30-min combined infusion of NLA and tempol, another two consecutive 10-min urine samples were collected. In five of these dogs, tempol was infused alone for 60 min before the addition of NLA to evaluate any long-term effects of tempol on renal parameters. During this prolonged period of tempol infusion, urine samples were collected every 10 min. After this prolonged infusion, combined infusion of NLA and tempol was also made in these dogs as mentioned earlier. To examine the specificity of the O2−-scavenging effects of tempol in these responses, a similar protocol was also carried out in three separate dogs with infusion of another nitroxide compound, 3-carbamyol proxyl (3-CP; 0.5 mg·kg−1·min−1; Sigma Chemical) instead of tempol. 3-CP is structurally similar to tempol, but it has minimal O2−-scavenging activity (9).
In another group of dogs (n = 12), a reverse order was considered for infusions of both tempol and NLA. In these dogs, after control collection periods, an intra-arterial infusion of NLA (50 μg·kg−1·min−1) was initiated. After a 30-min infusion of NLA, two consecutive 10-min urine samples were collected. Tempol (0.5 mg·kg−1·min−1) was then added to the NLA infusion line to examine the responses to O2− scavenging during NO synthase inhibition. After 10 min of the initiation of tempol infusion, two 10-min urine collections were made during tempol plus NLA infusion. In 7 of these 12 dogs, urine samples were also collected and prepared for 8-isoprostane measurement. These urine samples were collected into glass vials containing butylated hydroxytoluene to prevent further oxidative production of 8-isoprostane during collection and preservation of the samples (19, 23, 36). The samples were kept cold during collection by packing the sample vials in insulated boxes filled with ice and then kept frozen at −70°C until analyzed. These samples were also analyzed for measuring concentration of NO metabolites, nitrate/nitrite (17). As 8-isoprostane analysis was not considered in the initial design of the protocol, urine from the other five dogs in this group and also dogs from other groups used in this study was not collected methodically and thus was not considered for this analysis. In five other dogs, 3-CP (0.5 mg·kg−1·min−1) was infused instead of tempol after infusion of NLA to examine the specificity of O2−-scavenging effects on the observed responses to tempol.
At the end of each experiment, the electromagnetic flow probe was calibrated in situ by collection of timed blood samples into a graduated cylinder at different flow rates from a catheter placed in the renal artery. The kidney was then removed, stripped of all surrounding tissue, blotted dry, and weighed so that the calculated parameters could be expressed per gram of net kidney weight. Flame photometry (Instrumentation Laboratory) was used to determine the sodium and potassium concentrations in plasma and urine. Inulin concentrations in plasma and urine samples were determined by the anthrone colorimetric technique using a spectrophotometer (Ciba Corning Diagnostic). GFR was calculated using standard inulin clearance techniques. Urinary concentrations of 8-isoprostane and NO metabolites, nitrate/nitrite, were determined using enzyme immunoassay kit (Assay Design). All the data reported in results and shown in the figures are the average of the two clearance periods during each condition of the experimental protocol. Values are reported as means ± SE. Statistical comparisons of the differences in the responses within the group were conducted using one-way repeated-measures ANOVA followed by Newman-Keuls test. Comparisons of the responses between the groups were made using unpaired t-test. Differences in the mean values were deemed significant when probability values were ≤0.05.
During the control collection periods in the dogs used in the present study (n = 30), the mean values of plasma sodium, potassium, and hematocrit were 146.8 ± 1.2 meq/l, 3.5 ± 0.2 meq/l, and 45 ± 2%, respectively. There were no appreciable changes in these plasma parameters during the course of the experimental protocols.
Renal responses to tempol infusion before NLA administration.
Infusion of tempol into the renal artery before NLA infusion (n = 10) did not cause any significant changes in renal parameters as illustrated in Figs. 1 and 2. There were no differences between the responses to tempol infused for 30 min (n = 5) and for 60 min (n = 5). Thus these results were pulled together and presented as a single response to tempol in the figures. It was observed that tempol infusion alone caused a slight but significant reduction in SAP (137 ± 3 to 133 ± 4 mmHg, P < 0.05). As illustrated in Figs. 1 and 2, there were no significant changes in RBF, renal vascular resistances (RVR), GFR, urine flow (V), sodium excretion (UNaV), and fractional excretion of sodium (FENa). Tempol also did not cause significant changes in urinary excretion of potassium (UKV; 0.5 ± 0.1 to 0.6 ± 0.1 μmol·min−1·g−1). After the addition of NLA to the intra-arterial infusion line with tempol for >30 min, there was an increase of 34 ± 5% (P < 0.05) in RVR and decreases of 24 ± 1% (P < 0.05) in RBF, 26 ± 8% (P < 0.05) in V, 45 ± 11% (P < 0.05) in UNaV, and 36 ± 15% (0.05 > P > 0.1) in FENa (Figs. 1 and 2) without any significant change in GFR or UKV. Addition of NLA to the tempol infusion reversed the decrease in SAP (137 ± 7 mmHg).
Renal responses to tempol infusion in dogs pretreated with NLA.
Intra-arterial infusion of NLA alone before tempol infusion (n = 12) resulted in an increase in SAP (141 ± 4 to 146 ± 4 mmHg; P < 0.05). Figures 3 and 4 illustrate the responses to NLA alone on renal hemodynamics and function. There was a decrease of 29 ± 4% (P < 0.05) in RBF and an increase of 48 ± 9% (P < 0.05) in RVR without changes in GFR. NLA infusion alone in these dogs also caused decreases of 47 ± 5% (P < 0.05) in V, 67 ± 4% (P < 0.05) in UNaV, and 61 ± 6% (P < 0.05) in FENa without a significant change in UKV (0.5 ± 0.03 to 0.5 ± 0.04 μmol·min−1·g−1). These responses to NLA were similar to those reported earlier from our laboratory (16, 17).
Although tempol infusion before NLA administration did not cause appreciable changes in renal parameters (Figs. 1 and 2), administration of tempol to the NLA-pretreated dogs resulted in significant increases in V, UNaV, and FENa as illustrated in Fig. 4. However, there were no significant changes in RBF, RVR, or GFR (Fig. 3). During tempol infusion in NLA-pretreated dogs, SAP as well as UKV also did not change significantly.
Comparison of the responses to NLA before and during tempol.
There were some differences in the responses to NLA between the group in which NLA was given in the presence of tempol and the group in which NLA was first administered alone before tempol infusion. It was observed that the magnitudes of the renal excretory responses to NLA in tempol-pretreated dogs were significantly less (V, 26 ± 8 vs. 47 ± 5%, P < 0.05; UNaV, 45 ± 11 vs. 67 ± 4%, P < 0.05; FENa, 36 ± 15 vs. 61 ± 6%, 0.05 > P > 0.1) than the responses observed with administration of NLA alone. However, the differences in the magnitudes of the renal hemodynamic changes to NLA (RBF, 24 ± 1 vs. 29 ± 4%; RVR, 34 ± 5 vs. 48 ± 9%) were not statistically significant.
Responses to NLA and tempol infusion on urinary 8-isoprostane and nitrate/nitrite excretion.
As mentioned earlier in methods, the urine samples from the seven dogs in which tempol was infused after NLA administration were analyzed for 8-isoprostane and nitrate/nitrite concentration. Figure 5 illustrates these responses in urinary excretion rate of 8-isoprostane (UISOV) and nitrate/nitrite (UNOxV). It was observed that intra-arterial infusion of NLA in these seven dogs resulted in increases in UISOV. After addition of tempol to NLA infusion, UISOV returned to the control levels. As expected, NO synthase inhibition by NLA administration caused decreases in urinary excretion rate of NO metabolites (UNOxV). During coadministration of tempol and NLA, UNOxV did not change significantly from the values obtained during NLA infusion alone.
Effects of 3-CP infusion on renal hemodynamics and function.
In a separate group of dogs, infusion of 3-CP either before (n = 3) or during (n = 5) infusion of NLA did not cause any appreciable changes in renal parameters. The values of RBF, GFR, V, and UNaV (4.0 ± 0.1 ml·min−1·g−1, 0.82 ± 0.10 ml·min−1·g−1, 10.5 ± 3.3 μl·min−1·g−1, and 2.3 ± 6 μmol·min−1·g−1, respectively) obtained during control periods were not significantly different during 3-CP infusion alone (4.1 ± 0.2 ml·min−1·g−1, 1.0 ± 0.2 ml·min−1·g−1, 9.5 ± 2.7 μl·min−1·g−1, and 2.0 ± 0.6 μmol·min−1·g−1, respectively). Similar to what was observed in the earlier group, intra-arterial infusion of NLA alone before 3-CP infusion in five dogs resulted in usually observed decreases in RBF (3.7 ± 0.2 to 2.8 ± 0.3 ml·min−1·g−1), V (7.6 ± 2.3 to 3.9 ± 1.6 μmol·min−1·g−1), and in UNaV (1.6 ± 0.5 to 0.8 ± 0.5 μmol·min−1·g−1) without significant changes in GFR (0.87 ± 0.17 to 0.75 ± 0.12 ml·min−1·g−1). However, in contrast to the findings with tempol infusion, 3-CP infusion in these NLA-pretreated dogs did not cause any notable changes in V (4.1 ± 1.6 μmol·min−1·g−1) and UNaV (0.7 ± 0.4 μmol·min−1·g−1), as well as in RBF (2.5 ± 0.2 ml·min−1·g−1) and GFR (0.71 ± 0.12 ml·min−1·g−1).
In the present study, it has been observed that intra-arterial administration of an SOD mimetic, tempol, causes diuretic and natriuretic responses in dogs pretreated with NLA but not in the intact condition before NO inhibition. Infusion of another nitroxide compound (3-CP) having similar chemical structure as tempol but with minimal SOD mimetic activity (9) did not cause such changes in renal excretory function in NLA-treated dogs, indicating that these responses to tempol were linked to its O2−-scavenging effect. The dose of tempol used in this study was seen as sufficient to normalize the enhanced 8-isoprostane excretion rate (Fig. 5A) during NO blockade, indicating its maximal efficacy as an antioxidant. The dose for NLA was also used to achieve maximal inhibition of renal NO formation as in other studies conducted in our laboratory (16, 17). As 8-isoprostane excretion is considered as a marker for endogenous O2− activity (19, 23, 24), this finding indicates that the reduction in NO level facilitates the increment in O2− activity in the kidney. Urinary excretion of NO metabolites, nitrates/nitrites (UNOxV), decreased during NLA administration as expected and remained unchanged during coadministration of NLA and tempol (Fig. 5B). As scavenging of O2− by tempol in NLA-pretreated dogs caused diuresis and natriuresis without changing UNOxV, it is conceivable that such changes in excretory function were not related to NO but to the reduction in O2− level. It was also noted that pretreatment with tempol generally attenuated the excretory responses to NLA administration. These findings suggest that the renal excretory responses to NO synthase inhibition are at least partially influenced by the concomitant enhancement of endogenous O2− level.
There is a possibility that these effects of tempol may be linked to its action on renal sympathetic nerves (30) or antioxidant-induced changes in the release of norepinephrine from the nerve terminals (14). However, the present experiments were conducted in denervated kidneys to minimize any influence from concomitant alterations in renal sympathetic activity. The procedure for renal denervation employed in the present study was previously shown to completely abolish the changes in renal variables due to elicitation of cardiovascular reflexes that are mediated by alteration in renal sympathetic nerve (15). In the present study, we did not look at the possible changes in any humoral factor(s) in response to tempol administration. However, the involvement of such possible humoral factor(s) in the responses to tempol in the present study seems unlikely as we did not observe any appreciable changes in renal parameters during tempol infusion before NO inhibition. Further experiments may be needed to clarify this issue. It may also be argued that these renal responses to tempol may be attributed to the possible formation of H2O2 due to the use of this SOD mimetic (4). However, such possibility is unlikely as it has been reported that the effect of H2O2 is rather antidiuretic and antinatriuretic (4), an effect opposite to what has been observed in the present study.
The fact that there were no appreciable changes in renal parameters during tempol administration before NO inhibition (Figs. 1 and 2) indicates that the basal tissue concentration of O2− may be kept to a minimal level in the intact condition due to efficient antioxidant effect of endogenous NO (3, 8, 18), and thus there would be no O2−-scavenging effects of tempol. In other studies in rats (27, 28), it was also demonstrated that tempol administration had no significant effect on systemic blood pressure in Wistar-Kyoto rats but caused hypotension in spontaneously hypertensive rats. The results of the present investigation and from others (27, 28) indicate that the effects of scavenging O2− depend on the existing endogenous level of O2− production in the experimental animal. Although no supportive evidence is currently available, it can be speculated that the function of the enzymes responsible for endogenous production of O2− may be limited by intact NO synthase activity, and these oxidative enzymes may become upregulated during NO synthase inhibition. Another recent study from our laboratory has also demonstrated functional evidence for the presence of enhanced NO bioactivity in mice lacking the gene for gp91PHOX subunit of NAD(P)H oxidase (11). Thus it is reasonable to speculate that a disruption of NO synthase activity may be critically linked to an upregulation of NAD(P)H oxidase and/or other oxidative enzymes.
It was observed that tempol-induced increases in urine flow and sodium excretion in NO-inhibited dogs occurred in the absence of changes in RBF or GFR (Figs. 3 and 4). Previous studies have also shown that SOD inhibition (enhancement of O2− level) in dogs could cause reductions in urine flow and sodium excretion without appreciable changes in RBF or GFR (16). Collectively, these data demonstrate that intrarenal O2− can exert a direct influence on renal tubular reabsorptive function, leading to water and sodium retention. The exact mechanism by which O2− can induce antidiuresis and antinatriuresis is not yet clear. However, studies using isolated tubule preparation by Ortiz and Garvin (22) have shown that O2− can stimulate NaCl reabsorption by the thick ascending limb through a mechanism that is linked to protein kinase C activation in the epithelial cells. The findings in the present investigation clearly demonstrated that endogenous O2− exerts such direct action on tubular reabsorptive function primarily under conditions of reduced NO activity. These results are in agreement with our previous findings that the renal excretory responses to SOD inhibition are enhanced during NO synthase inhibition and provide further support to the renoprotective role of NO over the actions of endogenous O2− (16).
8-Isoprostane is generally believed to be formed endogenously due to oxidation of membrane phospholipids by a potent oxidant, peroxynitrite, that is produced during in vivo chemical reaction between NO and O2− (2, 10, 24). However, we observed an increase in urinary excretion of 8-isoprostane when endogenous NO production was inhibited by NLA administration (Fig. 5). This increase in 8-isoprostane excretion was ameliorated completely by the O2− scavenger tempol, indicating that such increased formation of 8-isoprostane may be the result of direct peroxidation of lipids by O2− (19, 23). Although tempol infusion normalized the enhanced excretion rate of 8-isoprostane and caused somewhat of a reversal of renal excretory responses to NO inhibition (Figs. 4 and 5), it did not cause reversal of the vasoconstrictor response to NLA administration (Fig. 3). This finding may not be entirely unexpected as the improvement of endothelial dysfunction due to the scavenging of endogenous O2− has been suggested to be linked with enhanced NO bioavailability (18, 35). In other words, the vasodilator effect of scavenging O2− is primarily due to the action of NO in the vasculature. Thus tempol administration during effective blockade of NO synthase is not expected to cause a renal vasodilator effect. It may be argued that the applied dose of tempol might have been insufficient to cause a vasodilatory effect. However, this possibility is unlikely as this dose of tempol was maximally effective in reducing the increased level of 8-isoprostane excretion to a near control value. The fact that tempol did not cause any vascular effect despite its effective blockade of enhanced O2− activity during NLA administration suggests that NO inhibition exerted a dominant effect over the action of O2− in mediating renal vasoconstriction.
In conclusion, the results of the present investigation indicate that under conditions of reduced NO activity there is an increase in endogenous O2− level that influences renal excretory function. These data further support the hypothesis that endogenous NO provides a renoprotective effect against the actions of O2− by acting as an antioxidative agent and thus suggest an interactive role for NO and O2− in the physiological regulation of kidney function.
The findings in the present study and also from our previous study (16) indicate an important role for O2− and its interaction with NO in the regulation of renal function. It has been shown that NO activity in many biological systems is enhanced by addition of the enzyme SOD, suggesting that interactions of O2− and NO are a frequent occurrence in biological tissues (21, 25). Importance of O2− and NO interaction in the regulation of many biological events has been increasingly appreciated due to the fact that under normal conditions, O2− is a constant product of cellular metabolism (1, 6, 13). In fact, it was proposed in 1989 (31) that vascular endothelium had an intrinsic capacity to generate O2− for regulatory purposes such as inactivation of NO. As a gaseous free radical, NO is highly reactive with other free radicals such as O2− (10); thus it could be argued that the free radical reactivity of NO is in many ways an important physiological regulator of cellular function in vivo. It has been demonstrated that NO has an antiatherosclerotic effect by acting as a powerful inhibitor of membrane lipid peroxidation because of its ability to scavenge O2− (3, 20, 26). Previous studies also provided evidence that NO reacts with oxygen free radicals and thus antagonizes the prooxidant properties of heme proteins in the biological tissues (8). An NO donor has also been shown to provide a cytoprotective effect in the ischemic tissue by reacting with excess oxygen free radicals in ischemia-reperfusion injuries (18). Thus it is conceivable that the interaction between NO and O2− has a role in maintaining a desired state of balance between oxidative and antioxidative processes in the kidney, any imbalance of which would lead to derangements in renal function.
This study was supported by National Heart, Lung, and Blood Institute Grants HL-51306 and HL-66432.
We are grateful to K. A. Taher and Dr. M. Z. Haque for technical assistance and D. Olavarrieta for excellent secretarial support in preparing the manuscript.
Present address for A. Nishiyama: Dept. of Pharmacology, Kogawa Medical University, 1750-1 Ikenobe, Miki-cho, Kita-gun, Kagawa 761-0793, Japan (E-mail: [email protected]).
- Copyright © 2004 the American Physiological Society