Cyclic nucleotide crosstalk and the regulation of cerebral vasodilation

Abstract

The vasodilating capability of cerebral vessels is vital to brain survival. There are multiple mechanisms participating in the regulation of cerebral vascular smooth muscle relaxation. The cyclic nucleotide-related signal transduction pathways are particularly important to the vasodilating function of cerebral arteries and arterioles. Those pathways are involved in the vasodilations elicited by a wide variety of stimuli, which include hypoxia and hypercapnia and agonist/receptor interactions (e.g. muscarinic, β-adrenergic, and prostacyclin receptors). There is considerable control exerted upon the activities of the cyclic nucleotide signal transduction cascades. One of those control mechanisms relates to the abilities of cyclic 3′-5′ adenosine monophosphate (cAMP) and cyclic 3′-5′ guanosine monophosphate (cGMP) to modulate each other's synthesis, degradation, and actions. That capacity for “crosstalk” between the two pathways provides the focus of the present review. To facilitate our discussion of cGMP/cAMP crosstalk regulation, we have placed some emphasis on hypercapnia-induced cerebral vasodilation. The review considers crosstalk at 4 different levels. First, the capacity for each cyclic nucleotide to repress the degradation of its counterpart via actions on phosphodiesterases (PDEs) is discussed. Second, consideration is given to crosstalk regulation of cGMP/cAMP synthesis. Third, cGMP- or cAMP-dependent protein kinases (PKG and PKA, respectively) are discussed with respect to: (a) the potential of each cyclic nucleotide to activate the counterpart kinase, and (b) the ability of PKG/PKA to elicit common actions. Fourth, some attention is given to the role of compartmentalization of: (a) the enzymes catalyzing cyclic nucleotide synthesis and degradation, (b) the enzymes responsible for cyclic nucleotides-mediated phosphorylations, and (c) the protein targets of those kinases.

Introduction

There are many signal transduction pathways contributing to the regulation of vascular smooth muscle (VSM) tone. The activities of most of those pathways are exquisitely controlled by a complex “fine-tuning” system where components of one signal transduction cascade may interact with components of another. That interplay has been labeled as “crosstalk”. The cyclic nucleotide cascades have received considerable attention in that regard. Most of the experimental evidence supporting crosstalk between the cyclic 3′-5′ guanosine monophosphate (cGMP) and cyclic 3′-5′ adenosine monophosphate (cAMP) pathways in VSM has been obtained in studies on peripheral vascular tissue. Comparatively little literature has been devoted to cyclic nucleotide crosstalk in cerebral VSM. Nevertheless, based upon the evidence from those limited publications, there is no reason to suspect that cyclic nucleotide crosstalk occurs to a lesser extent in cerebral VSM. As an example, there is considerable overlap in the ability of nitric oxide synthase (NOS) and cyclooxygenase (COX) inhibitors to attenuate hypercapnia-induced cerebrovasodilation (Wang et al., 1994b). Because NO- and prostanoid-induced vascular relaxations largely rely on cGMP and cAMP generation, respectively, those findings suggest crosstalk among the cyclic nucleotide pathways. This is not simply a random example. The ability of cerebral vascular tissue to dilate is an extremely important physiologic mechanism for permitting the brain to function adequately in the face of a variety of stresses. Vasodilation restricts the reductions in the delivery of essential nutrients (i.e. oxygen and glucose) when their blood levels are reduced, enhances nutrient delivery when cerebral metabolic demands are increased, and enhances the removal of unwanted metabolic products (i.e. CO₂). Indeed, nitric oxide (NO) and prostanoid-dependent processes have been variously indicated to participate in promoting cerebrovascular relaxation under a variety of conditions, including: hypoxia, hypoglycemia, hypercapnia, and recovery from cerebral ischemia (Clavier et al., 1994; Iadecola et al., 1994a, Iadecola et al., 1994b; Ichord et al., 1994; Iwamoto et al., 1992; Leffler et al., 1994; Pelligrino et al., 1995; Simpson and Phillis, 1991; Wang et al., 1994b). Thus, cyclic nucleotide-related signal transduction pathways and the crosstalk between them are critically important to cerebral vasodilating capacity and the maintenance of normal brain function.

Recent work in our laboratory has been directed toward investigating the overlapping roles of the cyclic nucleotides in CO₂-induced cerebral vascular relaxation. The purpose of this review is to integrate those findings with published information in an attempt not only to provide the reader with an appreciation of the complexity of the CO₂ response (cause for the controversy in the literature?), but in a broader sense, also to use that information as a framework for the discussion of cyclic nucleotide crosstalk in the control of cerebral VSM relaxation in general.

To begin a discussion of crosstalk between the cGMP and cAMP signal transduction pathways in the cerebral circulation, it is appropriate to provide a definition for that term. A more traditional definition of crosstalk would be the ability of signal transduction components from one pathway to influence components of another pathway. We would prefer to employ a less restrictive definition—one that includes actions at common targets (“cross activation”) or actions at separate targets, but ultimately producing an identical action on a single effector. In addition, to facilitate discussion of cyclic nucleotide pathway crosstalk, we think it important to provide a general description of the principal steps involved in the cyclic nucleotide cascades. For cGMP, our focus will be on cGMP generated in VSM via the NO-stimulated soluble guanylyl cyclase (sGC). The actions of cGMP, to a large measure, are mediated through activation of the cGMP-dependent protein kinase (PKG), which, in turn, regulates the function of target proteins via phosphorylation. For cAMP, the receptor and G-protein-associated enzyme, adenylyl cyclase (AC), represents an initial step. Like cGMP, the VSM functional effects of cAMP are primarily mediated by kinase activation. However, not only does cAMP promote activation of the cAMP-dependent protein kinase (PKA), but can activate PKG as well (see below).

Based on the broader definition of crosstalk and the general pathway structure given above, the review will be organized according to the following categories: (1) Reciprocal effects of the cyclic nucleotides on the levels of their counterparts. This would involve the ability of one cyclic nucleotide to modulate the synthesis or degradation of the other. The discussion in that area will include the capacity for cAMP (or cAMP pathway components) to affect the synthesis of NO and/or cGMP and vice-versa (i.e. cGMP effects on AC activity). The major emphasis in this section, however, will be on the capacity of the cyclic nucleotides (or cyclic nucleotide pathway components) to affect the hydrolysis of their counterparts through actions on phosphodiesterases (PDEs). (2) Cross activation of cyclic nucleotide kinases and overlap in kinase effects. This section will deal with the implications of the evidence that cAMP, at levels within the physiologic range, can activate both PKA and PKG in VSM and the possibility that the two kinases may have overlapping influences (common sites of phosphorylation or “cooperative” phosphorylation). (3) Intracellular compartmentalization of enzymes mediating cyclic nucleotide synthesis, hydrolysis, and cyclic nucleotide-related phosphorylation/dephosphorylation reactions. Such compartmentalization represents an additional level of crosstalk control and merits some attention. It is important to emphasize, however, that the issues relevant to cyclic nucleotide crosstalk presume that the signal transduction components reside in the same cell (i.e. VSM). Although that is often the case, there are many exceptions. In particular, cAMP-related systems are documented to be far more widespread than cGMP-related systems (e.g. Francis and Corbin, 1994). Nevertheless, for the sake of expedience, all of the issues discussed below assume the presence of cGMP and cAMP cascades within the same cell.

Hypercapnia is one of the most potent vasodilating stimuli in the cerebral circulation of mammals. However, the specific mechanisms involved in the cerebral VSM relaxation induced by hypercapnia, in particular the contributions from the cyclic nucleotide cascades, are far from settled. For example, most, but not all, relevant publications indicate a role for NO (and, by extension, cGMP) in hypercapnic cerebral vasodilation (reviewed in Iadecola et al., 1994a, Iadecola et al., 1994b). Also, it is not possible to generalize a role for cAMP-related pathways. Thus, although vasodilator prostanoids appear to play a substantial role in cerebral arterial CO₂ reactivity in neonates and many adult species (Eriksson et al., 1983; Hoffman et al., 1982; Leffler et al., 1994; Sakabe and Siesjo, 1979; Wang et al., 1994b), some species are variably affected or unaffected by COX inhibition (Busija and Heistad, 1983; Kontos et al., 1981; Markus et al., 1994). That uncertainty regarding the contributions of cAMP to the hypercapnic response extends into the other major cAMP-linked systems in the cerebral circulation—adenosine and β-adrenergic receptors. That is, studies employing adenosine receptor antagonists have produced results both in support of (Estevez and Phillis, 1997; Simpson and Phillis, 1991) and opposing (Emerson and Raymond, 1981; Hoffman et al., 1984; Pelligrino et al., 1995) adenosine participation in hypercapnic cerebral vasodilation. Moreover, β-adrenergic mechanisms do not appear to participate in the CO₂ response (Dahlgren et al., 1981). Much of the disparity in findings may be related to redundancy in the mechanisms linking increases in H⁺ to vasorelaxation, thus complicating interpretation of experimental findings when using conventional pharmacologic strategies (e.g. enzyme inhibitor, receptor antagonist, or ion channel blocker administration). That redundancy may, to some extent, be a function of crosstalk between the cyclic nucleotide pathways, as the findings with NOS and COX inhibitors would suggest.

Recent reports indicate that the NO/cGMP participation in hypercapnia-induced cerebrovasodilation is largely permissive (Iadecola and Zhang, 1996; Iadecola et al., 1994a, Iadecola et al., 1994b; Okamoto et al., 1997). That is, if one administers, in the presence of a NOS inhibitor, a NO donor or cGMP analogue, at a dose that produces little or no vasodilation by itself, one can restore cerebrovascular CO₂ reactivity to normal levels. Thus, rather than CO₂ increases eliciting vascular relaxation via enhanced generation of cGMP, a certain level of cGMP must be present to provide an “optimal environment” for vasodilation to occur. One possible “permissive” action of cGMP is via crosstalk with the cAMP pathway at the level of cAMP hydrolysis. That action would “permit” greater levels of cAMP to be achieved or maintained in the presence of vasodilator prostanoid-supported adenylyl cyclase activation. Thus, if a portion of the cGMP dependence in hypercapnia relates to promoting increased prostanoid-stimulated cAMP accumulation, this could explain the overlap in indomethacin and NOS inhibitor effects. On the other hand, in adult rats (Pelligrino and Wang, 1997a, Pelligrino and Wang, 1997b) and neonatal pigs (Leffler et al., 1994), cAMP also has been reported to play a permissive role in hypercapnia. Thus, one cannot ignore the possibility that cAMP, generated through vasodilator prostanoid actions, may be “permissive” by virtue of reducing hydrolysis of the cGMP generated through NO actions. The issue of cyclic nucleotide crosstalk at the level of hydrolysis will be discussed below.

Interestingly, hypercapnia, in rats, has been shown to increase brain tissue cGMP levels—a response that was blocked in the presence of a NOS inhibitor (Irikura et al., 1994). Furthermore, in a preliminary study, we found that hypercapnia-induced dilation of pial arterioles in vivo was accompanied by increases in the levels of cGMP and cAMP in cortical periarachnoid cerebrospinal fluid (pCSF) (Wang and Pelligrino, 1997). Both the vasodilation and the cyclic nucleotide responses were blocked by NOS inhibitor and COX inhibitor treatments in a non-additive fashion. Certainly, multiple cell types, apart from VSM, will influence pCSF and brain tissue sample cyclic nucleotide changes. Nevertheless, the presence of significant positive correlations between cerebrovascular CO₂ reactivity and tissue or pCSF cyclic nucleotide levels in the presence and absence of NOS or COX inhibition is suggestive of a functional link. Thus, at least part of the NO/cGMP and prostanoid/cAMP dependence of the hypercapnic response may relate to classical or “obligatory” roles for the cyclic nucleotides. Therefore, irrespective of whether one describes the cyclic nucleotide role in hypercapnia-induced cerebral vasodilation as permissive or obligatory (or a combination of both), the substantial overlap in cGMP and cAMP pathway inhibitor effects on vascular responses and cyclic nucleotide changes strongly suggests the presence of crosstalk.