No studies have specifically addressed whether cAMP can influence nitric oxide (NO)/cGMP-induced cerebral vasodilation. In this study, we examined whether cAMP can enhance or reduce NO-induced cerebral vasodilation in vivo via interfering with cGMP efflux or through potentiating phosphodiesterase 5 (PDE5)-mediated cGMP breakdown, respectively, in cerebral vascular smooth muscle cells (CVSMCs). To that end, we evaluated, in male rats, the effects of knockdown [via antisense oligodeoxynucleotide (ODN) applications] of the cGMP efflux protein multidrug resistance protein 5 (MRP5) and PDE5 inhibition on pial arteriolar NO donor [S-nitroso-N-acetyl penicillamine (SNAP)]-induced dilations in the absence and presence of cAMP elevations via forskolin. Pial arteriolar diameter changes were measured using well-established protocols in anesthetized rats. In control (missense ODN treated) rats, forskolin elicited a leftward shift in the SNAP dose-response curves (∼50% reduction in SNAP EC50). However, in MRP5 knockdown rats, cAMP increases were associated with a substantial reduction in SNAP-induced vasodilations (reflected as a significant 35–50% lower maximal response). In the presence of the PDE5 inhibitor MY-5445, the repression of the NO donor response accompanying forskolin was prevented. These findings suggest that cAMP has opposing effects on NO-stimulated cGMP increases. On the one hand, cAMP limits CVSMC cGMP loss by restricting cGMP efflux. On the other, cAMP appears to enhance PDE5-mediated cGMP breakdown. However, because increased endogenous cAMP seems to potentiate NO/cGMP-induced arteriolar relaxation when MRP5 expression is normal, the effect of cAMP to reduce cGMP efflux appears to predominate over cAMP stimulation of cGMP hydrolysis.
- nitric oxide
- multidrug resistance protein 5
- phosphodiesterase 5
cyclic nucleotides play key roles in promoting cerebral vasodilation. The stimulus for increased generation of cAMP often is receptor (e.g., prostanoid, β-adrenoceptor, adenosine)-linked activation of adenylyl cyclase (AC). For cGMP, the vasodilating stimulus, whether receptor linked (e.g., acetylcholine) or receptor independent (e.g., hypercapnia), in most cases, involves NO-mediated activation of soluble guanylyl cyclase (sGC). In adult animals, evidence favoring cyclic nucleotide cross-talk in cerebral vascular control can be derived from studies in hypercapnia. First, there is considerable overlap in the ability of nitric oxide (NO) synthase (NOS) and vasodilating prostanoid synthesis inhibitors to attenuate hypercapnia-induced cerebrovasodilation (28). Second, hypercapnia elicits increases in cerebral levels of both cAMP and cGMP. The increases in both cyclic nucleotides can be blocked by administration of either a NOS inhibitor or a prostanoid synthesis inhibitor (27). Thus preventing the increase in one cyclic nucleotide can completely repress the increase in the other. Those findings strongly suggest cross-talk among the pathways involved.
When attempting to understand cross-talk regulation of cyclic nucleotide concentrations, one must consider that such regulation may occur at the level of cyclic nucleotide synthesis, phosphodiesterase (PDE)-mediated degradation, or even cellular efflux. The PDE isoforms showing the highest expression in vascular smooth muscle are the PDE3, PDE4, and PDE5 varieties (24). With respect to cross-talk, the cAMP-preferring enzyme PDE3 has been given the most attention, because cGMP can inhibit its activity. Thus cGMP is capable of limiting cAMP loss and, in doing so, may potentiate vasodilating responses to AC-activating agents. Concerning cAMP-selective PDE4, there is no evidence of any direct influence of cGMP or cGMP-dependent protein kinase.
The relevant literature regarding cAMP modulation of cGMP hydrolysis is sparse. There is limited evidence that enhanced cAMP generation can increase cGMP levels in vascular smooth muscle cells (12, 29). In cerebral vessels, the cAMP effect might involve actions toward PDE5 (20). On the other hand, there is evidence that cAMP, via its kinase (PKA), can increase PDE5 activity, thereby accelerating cGMP hydrolysis (17).
Results suggesting a cAMP-related reduction in cGMP loss could also be explained on the basis of an inhibition of cGMP efflux. The findings of a recent study from our laboratory indicated that cGMP efflux may play an important role in regulating cerebral vascular smooth muscle cell (CVSMC) cGMP levels (32). In the present study, we tested the hypothesis that cAMP enhances NO-induced cerebral vasodilation via interfering with cGMP efflux in CVSMCs. To that end, we examined, in adult male rats, whether knockdown of the cGMP efflux protein multidrug resistance protein 5 (MRP5) in pial arterioles [using antisense vs. missense oligodeoxynucleotide (ODN) applications] affected NO/cGMP-induced dilations in the absence and presence of cAMP elevations via forskolin. To tease out potential cAMP actions on MRP5 from actions toward PDE5, the responses in the presence of forskolin were tested in the absence and presence of PDE5 blockade combined with MRP5 knockdown. Results indicated that cAMP has opposing effects on NO-stimulated cGMP increases. On the one hand, cAMP appears to limit CVSMC cGMP loss by restricting cGMP efflux. On the other hand, cAMP may potentiate PDE-mediated cGMP breakdown. The effect of cAMP to reduce cGMP efflux appears to predominate over cAMP-related stimulation of cGMP hydrolysis.
The experimental protocol was approved by the Institutional Animal Care and Use Committee. Age-matched male Sprague-Dawley rats (300–400 g) were used (supplied by Charles River; Wilmington, MA). Pial arteriolar reactivities were evaluated using a closed cranial window and intravital microscopy system (31). The windows were placed ∼24 h before study. The procedure for “chronic” placement of cranial windows in experiments utilizing topical applications of ODNs was described in a previous study from our laboratory (31). Three hundred microliters of a solution containing either 5 μM MRP5 antisense (5′-AATATCCTTCATCTTCAC-3′) or 5 μM missense (5′-CTCAACTTCAACTATCTT-3′) ODN were injected into the space under the cranial window. Six bases (3 at the 5′-end and 3 at the 3′-end) were phosphorothioated so as to minimize nuclease-mediated ODN breakdown. In a previous study (32), it was established that this approach resulted in a selective >60% reduction in MRP5 mRNA in pial tissue. On the day of study, after anesthesia induction with halothane and paralysis (curare), the rats were tracheotomized and mechanically ventilated. Bilateral femoral arterial and venous catheters, used for arterial blood gas measurement and drug infusion, were inserted under continuous anesthesia with 0.8% halothane-70% N2O-30% O2. After catheterization, the rat was placed in a head holder, and the cranial window inflow, outflow, and intracranial pressure monitoring cannulae were connected. Halothane was discontinued, and a 10 μg/kg fentanyl bolus was given intravenously. Anesthesia was maintained during the study with fentanyl (25 μg·kg−1·h−1 iv) and 70% N2O-30% O2. The space under the window was filled with artificial cerebrospinal fluid (aCSF, pH 7.35) that was equilibrated with a gas mixture consisting of 20% O2-5% CO2-balance N2. The aCSF solution was suffused at 0.5 ml/min and controlled at 37°C. Body temperature was maintained at 37°C with a servo-controlled heating pad, and mean arterial pressure and intracranial pressure were monitored continuously during the experiment.
Vascular reactivity was assessed by measuring the diameters of pial arterioles (30–50 μm). The vessels were viewed by video microscopy, and measurements were made using a calibrated video microscaler (see Ref. 31). In all experiments, the initial diameter measurements were made >1 h posthalothane and after 40 min of drug-free aCSF suffusion. In most experiments, dose-dependent pial arteriolar reactivities to suffusions of a NO donor [S-nitroso-N-acetyl penicillamine (SNAP); at 0.01, 0.1, 0.3, 1.0, and 10 μM] were assessed. After the return to baseline, a forskolin (0.2 μM) suffusion was initiated. That dose elicits only a ∼5% increase in pial arteriolar diameters. After 1 h, and in the continued presence of forskolin, pial arteriolar reactivity to SNAP suffusion was again examined. Diameter measurements from three separate pial arterioles were obtained (and averaged) after 5-min suffusion at each concentration. In all cases, comparisons were made between the antisense and missense ODN-treated rats.
The above rats were separated into five additional subgroups. First, rats chronically treated with Nω-nitro-l-arginine (l-NNA; at 100 mg·kg−1·day−1 ip over 4 days) or vehicle were studied. This chronic NOS inhibition paradigm was applied to prevent desensitization of the vessels to the vasodilating actions of cGMP. That is, recent findings from our laboratory (32) suggested that MRP5 knockdown-induced interference with cGMP efflux, in the presence of continued baseline cGMP generation, is accompanied by a prolonged increase in intracellular cGMP levels and a reduction in the vascular reactivity to subsequent acute elevations in NO/cGMP. That change could be prevented by chronic blockade of basal NO-induced cGMP generation. Second, we evaluated pial arteriolar reactivities to the sGC activator YC-1 in chronically NOS-inhibited antisense and missense ODN-treated rats. YC-1 was suffused at concentrations of 3.0, 10, and 30 μM. These evaluations were done to confirm that the changes in vascular reactivity observed were indeed cGMP related. YC-1 was prepared in DMSO, and subsequent dilutions were made with aCSF. The final DMSO concentrations were, therefore, 0.03%, 0.1%, and 0.3% at the three YC-1 doses. We have established in published (30) and preliminary experiments that within the above concentration range, DMSO does not affect pial arteriolar caliber. Third, we sought to examine whether the changes in vascular reactivity in the presence of forskolin were indeed due to cAMP and not to an unspecified action of forskolin, as suggested by the findings of Schultz et al. (23). Thus pial arteriolar responses to SNAP, in “NOS normal” MRP5 antisense and missense ODN-treated rats, were evaluated in the absence and, subsequently, in the presence of forskolin cosuffused with the AC inhibitor SQ-22536 (50 μM; Calbiochem; La Jolla, CA). That SQ-22536 dose was established in pilot studies (n = 3), where it was found to cause an 84 ± 8% and 84 ± 4% reduction in the pial arteriolar vasodilating response to 1.0 and 10.0 μM forskolin, respectively. Fourth, we endeavored to isolate the hypothesized cAMP actions toward MRP5 from the suspected potentiating effects of cAMP on cGMP hydrolysis via PDE5. To that end, in antisense and missense ODN-treated chronically l-NNA-treated rats, SNAP dose-response curves were generated first in the absence and then in the presence of the selective PDE5 blocker MY-5445 (30 μM), followed by MY-5445 plus forskolin (0.2 μM). MY-5445 was suffused for 30 min before reevaluation of the SNAP response. The dose of MY-5445 used (which is about two orders of magnitude greater than the published IC50 value for MY-5445) was established in preliminary evaluations as the dose producing a slightly greater than threshold (<5%) vasodilating response. The relative insensitivity of pial arterioles to the vasodilating actions of PDE5 inhibition, via a fairly high dose of MY-5445, is not surprising considering that, in chronically NOS-inhibited animals, baseline cGMP production is low. Within each subgroup, antisense and missense ODNs were derived from the same batch, and treatments were randomized. All experiments within one subgroup were completed before experiments were initiated in another subgroup. This allowed for greater subgroup homogeneity with respect to age and weight. Finally, we sought to examine whether MRP5 had any influence on cAMP efflux, as suggested by some recent reports (1, 21). Thus we compared dose-dependent pial arteriolar responses to forskolin (0.1, 1.0, 10, 30, and 100 μM) in antisense versus missense ODN-treated rats.
All drugs and reagents were obtained from Sigma (St. Louis, MO) and dissolved in aCSF unless otherwise stated. ODN were obtained from Sigma Genosys (St. Louis, MO). Values are presented as means ± SE. Comparisons of arteriolar diameter values within groups were made using one-way repeated-measures ANOVA combined with a post hoc Student-Newman-Keuls analysis. Analyses of diameter changes between groups were made using Student's t-test. A P value <0.05 was considered as significant.
With only minor exceptions, arterial Pco2, pH, and mean arterial blood pressure in the study groups were within normal limits and did not show any significant differences when initial and final values were compared over the course of the experiments (Table 1). Arterial Po2 values (not shown) were maintained above 100 mmHg in all rats studied. Also provided in Table 1 are the initial pial arteriolar diameters in the various groups. It should be noted that nearly all groups given chronic l-NNA treatment exhibited elevated mean arterial blood pressure values compared with the NOS normal groups.
NO donor-induced vasodilations: effects of cAMP elevations in the absence and presence MRP5 knockdown.
These experiments examined whether MRP5 is a target for cAMP influence on pial arteriolar dilations elicited by NO donor (SNAP)-induced increases in CVSMC cGMP generation. To that end, responses to SNAP in the absence and presence of the AC-activating drug forskolin (0.2 μM) in antisense versus missense ODN-treated rats were compared. These rats were further divided into NOS-normal (vehicle treated; Fig. 1) and chronically NOS-inhibited (Fig. 2) rats. Chronic NOS inhibition was used to prevent the suspected occurrence of long-term elevations in baseline cGMP levels associated with prolonged reductions in cGMP efflux. Recent findings from our laboratory (32) indicated that cGMP changes of this type resulted in a desensitization of pial arterioles to the vasodilating actions of subsequent cGMP increases. Thus by limiting basal cGMP generation (via chronic NOS inhibition), that desensitizing action could be largely prevented. Additional support for this desensitization phenomenon can be taken from the observation that, in NOS-normal rats, the maximum dilation value was significantly lower in the antisense- versus missense-treated group, whereas that pattern was reversed in the rats chronically treated with l-NNA (Table 2). This is identical to findings reported in our recent study (32). Forskolin, by itself, elicited similar and rather modest diameter increases in the missense- and antisense-treated NOS-normal and NOS-inhibited groups (2.6 ± 2.3%, 4.8 ± 1.7%, 3.7 ± 1.9%, and 4.5 ± 1.9%, respectively). In both control and l-NNA-treated groups, there was a significant leftward shift in the SNAP response in the missense-treated rats in the presence of forskolin. This was reflected in a significant >50% reduction in the EC50 values, with no change in the maximal responses (Table 2). On the other hand, in the MRP5 knockdown animals, in the presence of forskolin, we observed a significant reduction in the maximal responses to SNAP, with no statistically significant variations in the EC50 values (Table 2). Furthermore, the forskolin-associated reduction in the maximal response was greater in the NOS-inhibited (∼50%) versus vehicle-treated (∼35%) group. Thus, in the presence of normal MRP5 function, elevations in cAMP potentiate NO donor/cGMP-induced dilations. On the other hand, when MRP5 expression is reduced, acute increases in cAMP are associated with a reduced pial arteriolar reactivity to NO donor applications. This is suggestive of a MRP5- and efflux-independent effect of cAMP, leading to reduced intracellular cGMP levels or interference with CVSMC sensitivity to the vasodilating actions of cGMP. As is shown below, results seem to support the former, i.e., cAMP enhances the breakdown of cGMP via PDE5.
Vasodilating responses to the sGC-activating agent YC-1: effects of cAMP elevations in the absence and presence of MRP5 knockdown.
These studies were designed to verify that the changes in SNAP reactivity shown above did not include any cGMP-independent actions of NO. Pial arteriolar responses to the direct activator of sGC YC-1 in the absence and presence of forskolin in antisense versus missense ODN-treated rats are summarized in Fig. 3. All rats were chronically treated with l-NNA to prevent losses in cGMP sensitivity associated with MRP5 knockdown. These findings are similar to those seen with the NO donor SNAP (Fig. 2) to the extent that, compared with the initial response, in the presence of forskolin, we observed a significantly greater pial arteriolar response to YC-1 (at 3, 10, and 30 μM) in the missense-treated rats (Fig. 3A) but a significantly lower YC-1 response (at 3, 10, and 30 μM) in the antisense-treated group (Fig. 3B).
cAMP effects on PDE5.
These experiments were designed to examine whether PDE5 is influenced by cAMP. Thus pial arteriolar responses to SNAP were monitored in antisense versus missense ODN-treated rats in the absence and presence of the selective PDE5 blocker MY-5445. To avoid the confounding influence of MRP5 knockdown-related NO/cGMP desensitization and to limit baseline MY-5445-induced vasodilation (see earlier), these animals were subjected to chronic NOS inhibition. After the addition of MY-5445, pial arteriolar diameters increased 2.9 ± 1.8% in the missense ODN-treated rats and 2.1 ± 0.9% in the antisense ODN-treated rats. As one might predict, MY-5445 suffusion was accompanied by a potentiation of the responses to SNAP in both antisense and missense ODN-treated animals (Fig. 4). This was mainly reflected as a leftward shift in the SNAP dose-response curves in both groups. Furthermore, the significant reductions in the EC50 values (Table 3) were virtually identical in the antisense (38%) and missense (34%) ODN-treated groups. The similar MY-5445 effects in those two groups are noteworthy, because they suggest that MY-5445 does not act by interfering with MRP5-linked cGMP efflux, as indicated for other PDE5-selective inhibitors in some (but not all) reports (11, 21). If that were the case, then MY-5445 should have been associated with a lesser decrease in the EC50 for SNAP in the antisense versus missense ODN-treated group. When forskolin was subsequently added to the suffusate, we observed a further potentiation of the SNAP response in the missense ODN-treated rats (increase in the maximal response; Table 3) but no change in the antisense ODN-treated rats from the response seen in the presence of MY-5445 alone (Fig. 4). The latter finding should be contrasted with the substantial diminution of the SNAP response in the presence of forskolin (without PDE5 inhibition) in antisense-treated and l-NNA-inhibited rats (Fig. 2 and Table 2). This provides fairly compelling evidence that the reduced response to SNAP, in the presence of MRP5 knockdown and forskolin (shown in Fig. 2), relates to a potentiating action of cAMP toward PDE5. As an additional note, and in confirmation of results given in Table 2 and Ref. 32 (chronically NOS-inhibited rats), the initial maximal SNAP response (Table 3) was again found to be significantly higher in the antisense versus missense ODN-treated rats.
Forskolin effects in the absence and presence of AC inhibition.
Figure 5 shows the dose-related responses to SNAP in the absence and then in the presence of forskolin + the AC blocker SQ-22536. No significant changes were seen in the EC50 or maximal response values for SNAP-induced pial arteriolar dilation. Thus, in the control (missense treated) group, the EC50 values before and during forskolin + SQ 22536 suffusion were 0.17 ± 0.03 and 0.15 ± 0.03 μM, respectively, whereas the maximal responses were 43.7 ± 1.8% and 39.8 ± 1.7%, respectively. In the antisense ODN-treated rats, the EC50 values were 0.18 ± 0.03 and 0.14 ± 0.03 μM, respectively. The maximal response values were 37.2 ± 2.2% and 33.1 ± 2.5%, respectively. The lack of any significant changes in the dose-response curves for SNAP when cAMP synthesis was blocked strongly suggests that the results summarized in Figs. 1–3 and Table 2 relate specifically to forskolin-induced elevations in cAMP levels.
MRP5 influence on forskolin-induced vasodilation.
These evaluations (in chronically NOS-inhibited rats) were performed in an attempt to detect whether MRP5 knockdown interferes with cAMP efflux, thereby potentiating vasorelaxant responses to forskolin. Unlike MRP5 knockdown effects on NO/cGMP reactivity, no enhancement in cAMP vascular reactivity was observed (Fig. 6). Thus pial arteriolar reactivities to forskolin were not significantly different when dose-response curves were compared in antisense versus missense ODN-treated rats. That is, similar values were obtained in the missense and antisense ODN-treated rats for EC50 (2.56 ± 0.44 and 2.24 ± 0.39 μM, respectively) and the maximal response (66.9 ± 3.0% and 68.2 ± 9.1%, respectively).
The key findings of this study can be summarized as follows. First, in the presence of increased intracellular levels of cAMP and normal expression/activity of the cGMP efflux protein MRP5, NO/cGMP-induced cerebral vasodilation is enhanced. Second, in the face of reduced MRP5 expression/activity, elevations in cAMP are accompanied by an attenuated vasodilating response to NO/cGMP. Third, if one restricts PDE5-mediated cGMP breakdown, the above attenuation does not occur. These findings, therefore, suggest that cAMP has opposing effects on sGC-mediated cGMP increases. One influence relates to cAMP limiting vascular smooth muscle cGMP loss by restricting cGMP efflux. The other cGMP-regulating action of increased cAMP appears to be to facilitate cellular cGMP breakdown/loss via an action toward cGMP-specific PDE5. Because the net effect of elevating cAMP levels in the missense-treated control animals is an increase in NO/cGMP vascular reactivity, the efflux blocking effect would appear to predominate over the PDE-stimulating effect. On the other hand, if cAMP is prevented from acting on cGMP efflux (MRP5 knockdown), a capacity for cAMP to potentiate cGMP hydrolysis is revealed.
In attempting to understand the mechanisms involved in cAMP modulation of NO/cGMP-mediated cerebral vasodilating function, some clues might be derived from the SNAP dose-response information summarized in Tables 2 and 3. In advance of evaluating the cAMP-related data, it is of some importance to restate the rationale [provided in our recent publication (32)] for subjecting rats to chronic NOS inhibition. Thus we found that knockdown of MRP5 was accompanied by a diminished sensitivity of pial arterioles to the vasodilating actions of NO, sGC activators, and cGMP analogs. Relative to the missense group, that reduced sensitivity (exemplified by a lower maximal response) could be converted to an enhanced sensitivity (higher maximal response) if one limited basal CVSMC cGMP generation via chronic exposure to l-NNA. This implied, as discussed in our recent study, that the loss of sensitivity related to a prolonged increase in CVSMC cGMP content, resulting from overnight reductions in cGMP efflux. Consistent with this supposition, chronic l-NNA treatments predictably had no effect on SNAP dose-response data (Table 2; EC50 and maximal response) in rats administered missense ODN. Obviously, reductions in SNAP reactivity in antisense ODN-treated animals would greatly complicate comparisons between antisense versus missense ODN-treated groups. This stresses the need to employ a chronic NOS inhibition paradigm in studies involving MRP5 knockdown.
In analyzing the forskolin effects summarized in Tables 2 and 3, the focus will be on rats chronically treated with l-NNA. With that in mind, there are several findings that merit emphasis. First, in MRP5 knockdown rats (Table 2), the presence of forskolin-induced elevations in cAMP results in a substantial reduction in the maximal response to SNAP (i.e., NO/cGMP efficacy) but no change in the EC50 (i.e., NO/cGMP potency). Because the capacity to influence MRP5, but not PDE5, has been minimized, one might postulate that these data primarily reflect cAMP effects on PDE5. This is in accord with data obtained in visceral smooth muscle (17), where forskolin-induced elevations in cAMP levels enhanced PDE5 activity and limited the cGMP increases elicited by increasing concentrations of NO (reduced efficacy). Additional support for that assumption can be derived from Table 3. That is, if one interferes with the ability of cAMP to influence PDE5 (via high-dose MY-5445 administration), in addition to MRP5, then forskolin-induced elevations in cAMP do not alter SNAP-induced dilations (no changes in the maximal response or EC50). Curiously, when both PDE5 and MRP5 are intact (Table 2; missense), cAMP-associated potentiation of SNAP-induced dilations involves only a reduction in EC50.
Second, PDE5 inhibitor applications in missense ODN-treated animals (Table 3) should minimize PDE5 influence, leaving MRP5 influence intact. In this case, therefore, forskolin-related changes in SNAP reactivity may be viewed as an action toward MRP5. In those experiments, cAMP elevations were associated with an enhanced SNAP efficacy (Table 3). Third, irrespective of MRP5 status, the addition of a PDE5 inhibitor, in the presence of baseline cAMP levels, is accompanied by greater SNAP reactivity, exemplified by a significant reduction in EC50 (increased potency) but no change in efficacy (Table 3). In smooth muscle tissue, such a leftward shift in NO donor dose-relaxation curves, in the presence of PDE5 inhibition, has been reported by others (15, 25). The enhanced SNAP potency seen with PDE5 inhibitors, as opposed to reduced efficacy of SNAP, when cAMP interacts with PDE5, implies separate sites and mechanisms of action for cAMP versus MY-5445. That is, an effect on EC50 (potency) alone is often the hallmark of a competitive action, whereas an altered maximal response (efficacy) indicates a noncompetitive effect.
In summary, the pharmacodynamic data raises some intriguing, yet confounding, issues. For example, why does forskolin (cAMP) appear to affect NO/cGMP reactivity through changes in efficacy, but not potency, under experimental circumstances where cAMP actions can be somewhat isolated to either MRP5 (increased maximal response; Table 3) or PDE5 (decreased maximal response; Table 2) yet alters the potency, but not efficacy, when both proteins are intact? Perhaps this reflects the complexity of the mechanisms regulating intracellular cGMP content as well as cAMP influences on other proteins involved in controlling CVSM tone. Clearly, further work is needed.
Additional factors to bear in mind, when considering the results of the present study, include 1) cAMP modulation of sGC activity, 2) cAMP/PKA-related repression of contractile function, 3) implications of possible MRP5-mediated cAMP extrusion, and 4) alternative pathways for cGMP efflux. With respect to the first factor, findings obtained in nonvascular tissue in vitro suggest a pathway whereby AC activation results in PKA-mediated phosphorylation of sGC, sensitizing sGC to NO-related activation (e.g., Ref. 14). Yet, no evidence exists to indicate the presence of a similar process in vascular cells in vivo. The possibility that cAMP/PKA may potentiate SNAP responses through its documented inhibitory actions on vascular smooth muscle cell contractile mechanisms (e.g., Ref. 18) certainly cannot be ignored. Nevertheless, in the present study, if one limits the possible effects of increased cAMP to 1) increased sGC activity, 2) decreased cGMP efflux, or 3) increased PDE5 activity, and if one accepts the evidence supporting cAMP potentiation of PDE5 activity (see later), then either 1 or 2 (or both) could explain the forskolin-associated increase in the SNAP response seen in control (missense ODN treated) rats. Thus if increased sGC activity were the predominant action, relative to decreased cGMP efflux, removal of MRP5 should have elicited very little change in the forskolin effect on SNAP reactivity. However, in light of the markedly divergent effect of forskolin on the NO donor response seen in rats given MRP5 antisense ODN (i.e., repression) versus rats treated with missense ODN (i.e., enhancement), the second possibility (decreased efflux) would appear to predominate. In addition, despite evidence indicating that cAMP/PKA can enhance endothelial NOS-derived NO generation (2, 4, 9, 16), at least in vitro, such an action of cAMP could not explain the increased responses to SNAP and YC-1 in the present study, because those drugs act downstream from NOS, at sGC. Of some note, in isolated vascular tissue, potent stimulators of AC (forskolin and isoproterenol) were found to potentiate the cGMP increases associated with activation of both the soluble and particulate forms of GC (26, 29). Because only the soluble form is activated by NO, this would not only support an effect of cAMP downstream from NOS but downstream from sGC as well. Such data would not be inconsistent with an action toward cGMP efflux.
Some consideration might be directed toward evidence indicating that MRP5 also participates in cAMP efflux and that plasma membrane proteins, in addition to MRP5, are capable of mediating cyclic nucleotide efflux. Regarding the former, MRP5 has been shown to effect cAMP efflux in a number of in vitro systems (1, 21), although the MRP5 affinity for cGMP exceeded that of cAMP by one to two orders of magnitude. In the present study, MRP5 knockdown was associated with virtually no change in the vasodilating response to forskolin-induced elevations in cAMP. This indicates that MRP5 does not play any meaningful role in cAMP efflux in CVSMCs in vivo. One possible implication of this is that the enhanced arteriolar reactivity to SNAP in control (missense treated) rats (Figs. 1A and 2A), in the presence of forskolin, may not be explainable on the basis of a direct competition between cAMP and cGMP for MRP5-mediated cellular export. This issue is discussed further below.
There are other proteins that have been linked to cGMP cellular export. Among these, MRP4 has received the most attention. Like MRP5, MRP4 is expressed in cerebrovascular tissue (33) but displays a lesser affinity for cGMP than MRP5 (6, 11). Another recently identified cyclic nucleotide efflux protein is MRP8 (10), but little or nothing is known with respect to its cerebrovascular expression. However, because the ODN treatments in the present study only targeted MRP5, the specific role of MRP4 (or MRP8) in cGMP regulation in CVSMCs must await further experimentation.
To date, a limited number of studies have addressed the issue of cAMP influence on PDE5-mediated cGMP hydrolysis. In a majority of these reports, the indication was that cAMP, via PKA, was capable of enhancing PDE5 activity (3, 7, 17). In contrast, the results of other publications have suggested either no effect (22) or an inhibitory effect (12) of cAMP/PKA on PDE5 activity. This disparity in findings could relate to the diversity of the in vitro experimental models used, which range from smooth muscle cells to recombinant PDE5 to crude tissue extracts. In consideration of the fact that the experiments performed in the present study addressed potential cAMP/PKA effects on PDE5 activity in intact cerebral vascular tissue, the results obtained could be viewed as uniquely characteristic of cerebral vessels in vivo. However, one cannot completely ignore possible contributions from cAMP/PKA-mediated effects on cGMP-hydrolyzing PDEs, other than PDE5, that were not examined in the present study (e.g., PDE1, PDE9; see Ref. 8). Nevertheless, the present experiments did focus on PDE5-specific mechanisms. That is, in the presence of forskolin, when the capacity to influence cGMP efflux was impaired (MRP5 knockdown), we observed an attenuation in the vascular responses to the sGC activators SNAP and YC-1. It was postulated that this reflected a cAMP-related increase in PDE5-mediated cGMP loss. That postulate was further supported by the observation that if one interfered with PDE5 function, the cAMP-induced repression of cGMP reactivity was prevented.
With respect to possible cAMP influences on cGMP efflux, information in the literature is sparse. In one study (11) employing MRP5-transfected hamster lung fibroblasts, cAMP was found to interfere with cGMP transport but only at very high doses. In another study (23) using inside-out human erythrocyte vesicles, it was found that cAMP enhanced ATP-dependent cGMP transport (i.e., efflux) at low concentrations but reduced transport at high concentrations. The results of these studies appear to suggest a considerably wide degree of variability and tissue selectivity in cAMP effects on cGMP cellular export. Also, the fact that these studies utilized nonvascular in vitro preparations limits their relevance with respect to cerebrovascular tissue in vivo. Our findings in pial arterioles suggest that MRP5 only mediates the efflux of cGMP (Figs. 2, 3, and 6; see also Ref. 32). Because both cyclic nucleotides do not appear to share the same transport system, it is difficult to envisage a competitive relationship between them. Indeed, if the actions of cAMP were competitive, with respect to the cGMP efflux process, then one might expect the increased SNAP reactivity seen in the presence of forskolin and the PDE5 inhibitor (relative to PDE5 inhibition alone; Table 3 and Fig. 4A) to display the reduced EC50 that characterizes a strictly competitive interaction. The fact that it was SNAP efficacy, and not potency, that was increased by the above combination suggests that the mechanism of cAMP cross-talk regulation of cGMP at the level of cyclic nucleotide efflux is something other than just a simple competition between cyclic nucleotides.
Another factor that merits some consideration relates to evidence suggesting that forskolin-induced inhibition of cGMP transport may be a function of forskolin acting independently from its ability to increase cAMP generation (23). To address this possibility in the present study, the effect of forskolin on SNAP reactivity was tested in NOS-normal antisense and missense ODN-treated rats in the presence of the AC inhibitor SQ-22536. In both groups, the combined addition of forskolin and the AC inhibitor resulted in no further change in the SNAP response from that seen before forskolin exposure. Thus the increased NO donor response in the missense ODN-treated rats and the decreased response in the antisense ODN-treated animals (reflected in Fig. 1) were, indeed, related to forskolin-induced elevations in cAMP. Although this same strategy was not applied to the NOS-inhibited groups, it is reasonable to assume that the forskolin-associated changes observed (Fig. 2) can also be ascribed to elevations in cAMP.
In conclusion, as illustrated in Fig. 7, the results of this study indicate at least two principal sites for cAMP cross-talk regulation of NO/cGMP-mediated relaxation of cerebral arteriolar smooth muscle in vivo. The first involves a stimulatory action on PDE5-mediated cGMP breakdown. The second involves an apparent inhibitory effect on MRP5-mediated cGMP extrusion. The latter seems to provide the stronger influence, because in the presence of normal MRP5 expression and PDE5 activity, increases in endogenous cAMP generation enhance vascular reactivity to NO/cGMP.
This study was supported by National Heart, Lung, and Blood Institute HL-52594.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 by the American Physiological Society