INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

INHIBITION OF NITRIC OXIDE synthase (NOS), in particular the neuronal isoform (nNOS), substantially attenuates hypercapnia-induced cerebrovasodilation (18, 29, 39). This indicates that NO, presumably via guanosine 3',5'-cyclic monophosphate (cGMP, see Ref. 31), plays an essential role in that response. It was recently shown that repletion of NO or cGMP, in the presence of a NOS inhibitor, could restore cerebrovascular CO₂ reactivity. That led investigators to conclude that NO (or probably more precisely, cGMP) plays a permissive role in hypercapnic vasodilation in the brain (19, 20, 29). The "permissive approach" assumes that the system has been returned to a state that existed prior to NOS inhibition. However, on the basis of a recent report from our laboratory, that assumption may not be valid (37). In that study, we showed that suffusion of 8-bromoguanosine 3',5'-cyclic monophosphate (8- BrcGMP), at a dose that did not by itself elicit vasodilation, reversed the loss of pial arteriolar CO₂ reactivity accompanying nNOS inhibition. We subsequently found that CO₂ reactivity in the "recovered" system could be greatly attenuated by suffusions of inhibitors of both the Ca²⁺-dependent (K_Ca) and the ATP-sensitive (K_ATP) K⁺ channels. In the absence of NOS inhibition (i.e., normal conditions), the K⁺-channel blockers were ineffective. Thus despite similar levels of CO₂ reactivity, the mechanisms responsible for that response were different when comparing the NOS-inhibited/cGMP-repleted state to "normal" conditions.

In the present study, we addressed the hypothesis that the nascent K⁺-channel dependence of the "recovered CO₂ reactivity" situation (i.e., nNOS inhibition plus 8-BrcGMP) relates at least in part to an increase in the influence of the miconazole-inhibitable cytochrome P-450 epoxygenase pathway. The rationale for that hypothesis derives from a number of published observations. First, a cytochrome P-450 epoxygenase product [i.e., an epoxyeicosatrienoic acid (EET)] has been shown to increase activity of K_Ca (8, 14) and K_ATP channels (13) in vascular smooth muscle (VSM) cells. Second, cGMP may modulate the responsivity of VSM cells to EETs (26). Third, H⁺ can modulate VSM K_ATP and K_Ca channel activities (17, 22). We therefore considered the possibility that the combination of an increased epoxygenase activity and normalized cGMP levels may "permit" greater K⁺-channel stimulation when H⁺ concentration is elevated by hypercapnia. To that end, pial arteriolar responses to hypercapnia were recorded before and after nNOS inhibition [via 7-nitroindazole (7-NI)], following subsequent cGMP repletion and the addition of the epoxygenase inhibitor miconazole. In additional experiments, the effects of miconazole, in the absence of nNOS inhibition, on CO₂ reactivity and on vasodilatory responses to K⁺-channel openers were examined.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The study protocol was approved by the Institutional Animal Care and Use Committee. Sprague-Dawley rats, 250-350 g, were used. Anesthesia was induced with halothane, and the rat was paralyzed (curare), tracheotomized, and mechanically ventilated. Surgical anesthesia for insertion of bilateral femoral arterial and venous catheters consisted of 0.8% halothane-70% N₂O-30% O₂. After catheterization, the animal was placed in a head holder and the skull was exposed to permit placement of a closed cranial window. The cranial window design and surgical implantation were described in detail in previous publications (23, 40). The 11-mm-diameter acrylic windows were placed over a 10-mm craniotomy and fixed to the skull with cyanoacrylate gel. The windows were equipped with three ports [inflow, outflow, and intracranial pressure (ICP) monitoring]. After window placement, the halothane was discontinued and a loading dose of intravenous fentanyl was given (10 µg/kg). Anesthesia during the study was fentanyl (25 µg · kg¹ · h¹ iv) plus ventilation with 70% N₂O-30% O₂. Cannulas were secured into the inflow, outflow, and ICP-monitoring ports of the cranial window, and the space under the window was filled with artificial cerebrospinal fluid (aCSF). The composition of the aCSF is provided elsewhere (23, 40). The aCSF was suffused at 1.0 ml/min and was maintained at a temperature of 37°C, PCO₂ 40-45 mmHg, PO₂ 50-60 mmHg, and pH 7.35. The ICP was controlled at 5-10 mmHg by adjusting the height of the outflow cannula. The reactivity of 25- to 50-µm pial arterioles on the exposed cortical surface was assessed via measurement of diameter changes. A microscope (Nikon) and color video camera (Sony) arrangement was equipped with an epi-illumination, dark-field system (Fryer, Huntley, IL). Magnifications of ×800 were displayed on a video monitor. Measurements of vessel diameters were made using a calibrated video microscaler (Optech).

In all experiments, initial diameter measurements were made after a 30-min period of cortical suffusion with drug-free aCSF (initiated at 1 h after halothane). The rats were divided into groups based on subsequent experimental manipulations. In group 1, hypercapnia was imposed for 5 min, diameters were measured, and normocapnic conditions were restored. After 15-20 min, diameters were again measured and the nNOS inhibitor 7-NI (40 mg/kg ip) was administered. Forty-five minutes later, a 5-min period of hypercapnia, followed by 15-20 min of normocapnia, was again imposed. Suffusion of the cGMP analog 8-BrcGMP (5 µM) was then initiated and after 20 min hypercapnia was reimposed. On return to normocapnia, miconazole (10 µM) was added to the suffusate and the hypercapnic response was measured 20-30 min later. The target PCO₂ in hypercapnia was 65 mmHg. That value was chosen for two reasons. First, it lies at the approximate midpoint of the linear portion of the cerebral blood flow response curve to PCO₂ elevations (i.e., 40-80 mmHg) (41). This permits us to express the pial arteriolar response as CO₂ reactivity (%diameter change per mmHg PCO₂ increase), thereby reducing variability should we miss the hypercapnic target by ±5-10 mmHg (see Ref. 37). Second, raising PCO₂ from 40 to 80 mmHg represents the hypercapnic range over which the percentage reductions in cerebrovascular CO₂ reactivity elicited by NOS inhibition is constant (41). The miconazole was dissolved in ethanol before its further dilution in the aCSF solution (giving a final ethanol concentration of 0.1%). The concentration of miconazole used represents a dose that is effective toward the epoxygenase, without affecting NOS activity (1). In three rats from group 1 the pial arteriolar response to adenosine (10 µM) was tested (see Ref. 37) prior to the initial hypercapnic exposure and again in the presence of 7-NI, 8-BrcGMP, and miconazole before the final hypercapnic exposure. In group 2 rats we evaluated the influence of 0.1% ethanol, followed by miconazole, on CO₂ reactivity. The same protocol used for monitoring pial arteriolar responses to hypercapnia in group 1 experiments was used in these evaluations, although suffusate compositions were different. Thus after baseline determinations of CO₂ reactivity, a 20-min period of suffusion of a 0.1% ethanol (in aCSF) solution was applied. After a second CO₂ reactivity measurement was obtained, miconazole (10 µM) was added to the suffusate, and a third hypercapnia exposure was initiated 20 min later. In a third group of rats we examined whether miconazole had any effects on K_ATP- or K_Ca-channel function. The inclusion of this group was based on the possibility that miconazole may block VSM K_Ca or K_ATP channels directly (3, 36), in addition to the desired effect of blocking epoxygenase activity. However, a direct K⁺-channel effect of miconazole is not universally supported (8, 14). The implication from these findings is that whether or not miconazole affects K⁺-channel function directly may depend on the experimental model employed. Thus we thought it important to test in our present system the possibility that miconazole may repress the vasodilating actions of K_Ca- or K_ATP-channel openers. In these experiments, suffusions of either the K_ATP- channel opener levcromakalim (3 and 30 µM) or the K_Ca-channel opener NS-1619 (5 and 50 µM) were performed [see Wang et al. (37)]. After baseline diameters were restored, a 20-min period of miconazole suffusion was initiated and the K⁺-channel openers were reapplied. A fourth group of animals was used to test the possibility that 8-BrcGMP supplementation, in the absence of NOS inhibition, might affect pial arteriolar CO₂ reactivity or the way miconazole influences CO₂ reactivity. In those experiments, the CO₂ response was tested in the presence of drug-free aCSF, after a 20-min suffusion of 5 µM 8-BrcGMP, and subsequently after a 30-min suffusion of 5 µM 8-BrcGMP plus 10 µM miconazole.

The miconazole and 8-BrcGMP were obtained from Sigma (St. Louis, MO). 7-NI was from ICN Biologics (Aurora, OH). NS-1619 was obtained from Research Biochemicals (Natick, MA), and levcromakalim was a gift from SmithKline Beecham Pharmaceuticals (Brockham Park, UK). For the pial arteriolar diameter data, a repeated-measures ANOVA with a post hoc Tukey's analysis was used. All values are reported as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In group 1 rats the arterial PCO₂ and mean arterial blood pressure values measured during each experimental stage (i. e., initial, at 45 min post-7-NI, at 20 min of 8-BrcGMP suffusion, and at 20 min of combined 8-BrcGMP and miconazole suffusion) are summarized in Table 1. The data were obtained just before and at 5 min of hypercapnia. The same data sets for groups 2 and 4 rats are given in Tables 2 and 3, respectively. No significant variations in normocapnic or hypercapnic CO₂ partial pressures or mean arterial blood pressure levels were observed throughout the experiments. In all cases PO₂ was maintained between 100 and 150 mmHg.

Baseline (normocapnic) pial arteriolar diameters in group 1 showed a slight but not statistically significant tendency to increase over the course of the study. Thus the diameters measured immediately before induction of hypercapnia during each of the four experimental stages in group 1 were 35.5 ± 2.5 (initial), 32.5 ± 3.0 (45 min post-7-NI), 36.3 ± 3.6 (20 min 8-BrcGMP suffusion), and 40.9 ± 3.0 µm (20 min after miconazole plus 8-BrcGMP), respectively. Virtually no variations in normocapnic pial arteriolar diameters were seen over the course of the group 2 studies. The mean values at each stage were 33.4 ± 4.0 (initial), 32.5 ± 3.5 (0.1% EtOH), and 32.8 ± 3.3 µm (miconazole in 0.1% EtOH). In the studies testing the effects of miconazole on K⁺-channel opener vasodilating responses, the initial and 20-min post-miconazole pial arteriolar diameters in the levcromakalim experiments were 33.5 ± 3.6 and 34.1 ± 3.6 µm, respectively. In the NS-1619 experiments, the values were 34.6 ± 4.3 µm (initial) and 31.4 ± 3.8 µm (post-miconazole). In the group of experiments in which we evaluated the effects of 8-BrcGMP and 8-BrcGMP plus miconazole on CO₂ reactivity in the absence of 7-NI, the mean diameters were 32.7 ± 3.4 µm (initial), 33.9 ± 3.8 µm (20 min after 8-BrcGMP), and 35.3 ± 3.9 µm (20 min after 8-BrcGMP plus miconazole). A general conclusion from these findings is that neither 7-NI, 8-BrcGMP, miconazole, nor 0.1% EtOH, by itself, altered pial arteriolar diameters.

Inhibition of nNOS activity with 7-NI produced a ~70% reduction in pial arteriolar CO₂ reactivity (Fig. 1). Subsequent suffusion with 8-BrcGMP was accompanied by an increase in the CO₂ reactivity toward the normal level. Both findings are consistent with the literature (19, 37). When the cytochrome P-450 epoxygenase inhibitor miconazole was added to the suffusate, the hypercapnic response was reduced by ~80%. That reduction in the CO₂ response was not due to a general loss of vasodilating function, because the pial arteriolar response to adenosine was unaffected by the combination of 7-NI, 8-BrcGMP, and miconazole (33.3 ± 2.0% diameter increase initially vs. 32.4 ± 0.8% in the presence of the drug combination). Topical application of miconazole, in the absence of nNOS inhibition and in the presence and absence of added 8-BrcGMP, had no effect on CO₂ reactivity (Fig. 2). Similarly, suffusion of the miconazole vehicle (0.1% EtOH) alone was without influence on the hypercapnic response (Fig. 2A). Thus miconazole attenuates pial arteriolar CO₂ reactivity only in the presence of nNOS inhibition and exogenous cGMP. These findings are remarkably similar to results obtained in a recent study from our laboratory, where we showed that blockade of K_Ca or K_ATP channels also was able to attenuate hypercapnic pial arteriolar dilation only under the unique condition of nNOS inhibition and cGMP repletion (37). Although the present study suggests that activation of the epoxygenase pathway could account for the nascent K⁺-channel participation, we could not eliminate the possibility that miconazole was acting directly on the K_Ca and K_ATP channels. Thus another set of experiments was performed to examine whether miconazole had any effect on the pial arteriolar dilations elicited by the putative K_ATP- and K_Ca-channel agonists levcromakalim (3 and 30 µM) and NS-1619 (5 and 50 µM), respectively. As shown in Fig. 3 (levcromakalim) and Fig. 4 (NS-1619), a 20-min suffusion of 10 µM miconazole did not diminish the vasodilations produced by those agents.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   Effects of 7-nitroindazole (7-NI, 40 mg/kg ip) followed by sequential topical applications of 8-bromoguanosine 3',5'-cyclic monophosphate (8-BrcGMP, 5 µM) and 8-BrcGMP + miconazole (MIC, 10 µM) on hypercapnia-induced pial arteriolar relaxations (expressed as CO2 reactivity in %diameter change/mmHg arterial PCO2 increase). * P < 0.05 vs. baseline.  P < 0.05 vs. 7-NI plus 8-BrcGMP. Values are means ± SE; n = 9 experiments.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effects of miconazole and/or 8-BrcGMP on pial arteriolar CO2 reactivity in absence of neuronal nitric oxide synthase (nNOS) inhibition. A: sequential suffusions of drug-free artificial cerebrospinal fluid (aCSF), miconazole vehicle (0.1% ethanol in aCSF), and miconazole (10 µM). B: sequential suffusions of drug-free aCSF, 8-BrcGMP, and 8-BrcGMP plus miconazole. Values are means ± SE; n = 5 experiments.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effects of miconazole (10 µM) on pial arteriolar dilations elicited by ATP-sensiive K+-channel opener levcromakalim (3 and 30 µM). Values are means ± SE; n = 6 experiments.

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Effects of miconazole (10 µM) on pial arteriolar dilations elicited by Ca2+-dependent K+-channel opener NS-1619 (5 and 50 µM). Values are means ± SE; n = 6 experiments.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major findings of the present study were 1) the normalization of pial arteriolar CO₂ reactivity elicited by cGMP repletion in the presence of nNOS inhibition could be reversed by inhibition of epoxygenase activity with miconazole, and 2) in the absence of nNOS inhibition and cGMP repletion, miconazole had no influence on hypercapnia-induced pial arteriolar relaxation. This seeming dichotomy in the actions of miconazole on hypercapnia-induced pial arteriolar dilation is similar to the effects recently reported by us for K⁺-channel blockers (37).

On the basis of present findings, it is tempting to conclude that the nascent K⁺-channel dependence of hypercapnia-induced cerebrovasodilation, revealed in our earlier paper (37), derived from increased epoxygenase activity. In attempting to explain this apparent association, one must consider the key factors present in both studies, i.e., reduced NO (via inhibition of nNOS) coupled with "normalized" cGMP levels at the time that H⁺ levels are increased. Logically, because of the reported capacity for NO, independently from cGMP, to repress P-450 monooxygenase activity (6), one might assume that removal of NO (via 7-NI) plays some role in the apparently increased epoxygenase activity. Yet, administration of NO donors has been shown to be as effective as cGMP analogs in restoring cerebrovascular CO₂ reactivity in the presence of NOS inhibition (19). Moreover, the NO-restored hypercapnic responsivity remains sensitive to K⁺-channel inhibition (37). The fact that both NO and cGMP supplementation are capable of restoring the hypercapnic response and converting it to a K⁺-channel-dependent process implies that the NO actions occur through cGMP. Indeed, published findings suggest that cGMP may modulate hyperpolarizing factor (presumably an EET) actions toward K⁺ channels (26). Moreover, VSM K⁺-channel function is influenced by cGMP-dependent protein kinase G-mediated phosphorylations (9, 16, 32, 34) as well as increases in H⁺ levels (17, 22). However, these findings do not necessarily permit us to conclude that epoxygenase products and K⁺ channel function are linked when considering the mechanisms responsible for cGMP analog-induced restoration of the hypercapnic response. Thus delineation of the specific roles played by cGMP in promoting normalization of CO₂ reactivity via epoxygenase and K⁺-channel-dependent processes must await further study.

The 5 µM dose of 8-BrcGMP applied in these experiments, when considered in relation to the 0.1-0.2 µM levels of cGMP found in rat periarachnoid CSF before and after administration of vasodilating doses of NO-donating agents (38), would appear to be excessive. However, that dose of 8-BrcGMP did not alter pial arteriolar diameters. In fact, to elicit vasodilation in rat pial arterioles in vivo, the dose of topically applied 8-BrcGMP must exceed 10 µM or perhaps even 100 µM (30, 37). This is consistent with the ED₅₀ value (~50 µM) for 8-BrcGMP relaxation of peripheral arteries (12). The requirement for relatively high 8-BrcGMP levels to evoke a vasodilating response may be due to its modest lipophilicity (butanol/water partition coefficient 1.1 vs. 0.3 for cGMP) and the fact that it is not completely immune to phosphodiesterase action (7, 12). Therefore, "physiological" doses of 8-BrcGMP (i.e., those producing vasodilation) must be in the high micromolar range, and the present 5 µM dose is well within physiological limits.

Concerns have been raised with respect to the nonselectivity of miconazole and other imidazole antimycotic agents. Some investigators have reported a NOS-inhibitory action of miconazole (42). On the other hand, Alkayed et al. (1) recently reported that even at 20 µM concentrations, miconazole had no effect on brain NOS activity. In the present study, 10 µM miconazole was used, which in all probability places us well below the threshold for NOS inhibitory actions. Furthermore, even if miconazole had some inhibitory influence on NOS activity in the present study, it was not of any physiological significance because miconazole alone did not affect hypercapnia-induced pial arteriolar relaxation. If miconazole did affect NOS, a reduced pial arteriolar response would be expected in consideration of the well-established capacity of nonselective and nNOS-selective inhibitors to attenuate cerebrovascular CO₂ reactivity (18, 29, 39). Other reported "nonepoxygenase" actions of miconazole include blockade of K_Ca channels, K_ATP channels (36), and inhibition of Ca²⁺-ATPase activity (25). Direct actions of miconazole toward K⁺ channels could not have influenced present results, as demonstrated by the findings summarized in Figs. 3 and 4. Similarly, it is unlikely that miconazole effects in the present study were in any way related to actions on VSM Ca²⁺-ATPase function. Inhibition of Ca²⁺-ATPase has been shown to be associated with vasoconstriction of cerebral arteries (28). In this study, miconazole by itself did not elicit reductions in pial arteriolar diameters. Thus, as a number of reports have shown (e.g., Refs. 1 and 8), inhibition of epoxygenase activity and subsequent EET synthesis by miconazole, at the dose used in the present study, remains as the major action of this drug.

Although the miconazole effects observed in this study are probably due to blocking EET synthesis, the data do not permit us to determine which of the multiple EETs formed by epoxygenase action are involved or their cellular source. The 5,6-, 8,9-, 11,12-, and 14,15-EET regioisomers are synthesized in vascular endothelial cells and astrocytes (2, 4, 8). All have been variously shown to possess VSM relaxant effects in a number of tissues, including the brain (11, 24), and a capacity to promote VSM hyperpolarization (15). In many cases, the hyperpolarizations and vasodilating actions can be prevented by addition of K_Ca-channel blockers, but not K_ATP-channel blockers (10, 15, 33). One report did present evidence of K_ATP-channel blockade preventing EET-induced VSM hyperpolarization (13), but those data were obtained in mesenteric VSM. Thus there is at present no evidence for EET-related activation of K_ATP channels in cerebral VSM. The possibility remains that the nascent sensitivity to K_ATP- channel blockade we observed in our recent study (37) may be the result of an EET-independent process. Another possibility is that EETs may affect K_ATP (or K_Ca) channels indirectly. Cerebral vasodilations elicited by EETs can be blocked by indomethacin (11, 24), leaving one with the prospect that a cyclooxygenase product may provide a link between the miconazole effect (reduced EET synthesis) and reduced K⁺-channel activity. In support of this, vasodilator prostanoid-induced arterial relaxations in the brain (5) and periphery (21, 27) have been shown to be attenuated by K_ATP- channel blockade, as well as blockade of K_Ca channels. One might also consider that the cyclooxygenase product responsible for K⁺-channel stimulation may be a reactive oxygen metabolite, rather than a prostanoid (35). Whether the recovered CO₂ reactivity, along with the EET- and K⁺-channel dependence of that response, relates to a cyclooxygenase product or for that matter any other endogenous substance must await additional experimentation.

In conclusion, we showed in the present study, as in a recently published report from our laboratory (37), that the substantial nNOS inhibitor-induced attenuation of pial arteriolar CO₂ reactivity could be reversed on cGMP repletion. The restored response was again attenuated after cytochrome P-450 epoxygenase blockade with miconazole. In the absence of nNOS inhibition plus cGMP supplementation, miconazole had no effect on the CO₂ response. The epoxygenase products, EETs, are known to increase K⁺-channel activity. Thus these results may provide some explanation for our previous finding that K⁺-channel blockers were capable of attenuating hypercapnia-induced vasodilations only following cGMP-associated restoration of the response. Even if these two observations are not directly linked, the results of this investigation do at least provide additional evidence that hypercapnia-induced cerebrovasodilation can occur through multiple alternative mechanisms. As the present and our previous study (37) have demonstrated, conditions can be created whereby CO₂ reactivity is normal but where blockers of various signal transduction pathways can have widely differing effects on the response.