Cardiovascular Implications in the Use of PDE5 Inhibitor Therapy

Cardiovascular implications in the use of PDE5 inhibitor therapy

DH Maurice*

Department of Pharmacology & Toxicology, Queen’s University at Kingston, Kingston, ON, Canada

Cardiovascular smooth muscle cells (SMCs) exist as resting or activated cells. Resting SMCs produce contractile proteins and are nearly transcriptionally inactive; activated SMCs are transcriptionally active and are involved in pathological processes such as atherosclerosis. Soluble guanylate cyclase, protein kinase G, and protein kinase A are present in SMCs, but their levels can be decreased in activated cells. Phosphodiesterase 3 (PDE3) activity is abundant in cardiovascular tissues; both PDE3A and PDE3B are involved in cyclic adenosine monophosphate (cAMP) hydrolysis in these tissues. Cyclic-AMP-hydrolyzing PDE activities are altered during the phenotypic transition of SMCs from the resting to the activated phenotype. Similar changes have been observed in cyclic guanosine monophosphate cGMP-hydrolyzing PDEs, although the impact of these alterations on PDE5 inhibitor-mediated effects requires further study. This report presents the changes in PDE expression that accompany phenotypic modulation of SMCs and discusses the potential impact of these events on PDE5-mediated cell functions. International Journal of Impotence Research (2004) 16, S20–S23. doi:10.1038/sj.ijir.3901210

Keywords: phosphodiesterase; smooth muscle cells; cyclic AMP; cyclic GMP; protein kinase

Introduction Quiescent/resting SMCs, normally present in healthy blood vessels that perfuse most organs, contract and relax in response to pulsatile differ- In addition to physiologically based differences in ences in the blood flow and in response to the the expression of individual phosphodiesterases pharmacologic and physiologic stimuli. This re- (PDEs) in the cells of cardiovascular tissues, the sponse controls the diameter of blood vessels and, as overall cardiovascular health status of patient such, plays an important role in regulating systemic populations also may impact significantly on both blood pressure. Quiescent/resting SMCs express the the types and levels of individual PDEs in these contractile proteins required for the generation of tissues. At present, it is unclear if these differences contractile forces but are, for the most part, affect pharmacologic agents aimed at PDEs or other transcriptionally inactive and display a very low enzymes involved in regulating cyclic nucleotide rate of proliferation and of directed migration. In levels in these cells. contrast, activated/synthetic SMCs are transcrip- This article reviews the current state of knowledge tionally active and exhibit high levels of prolifera- concerning (a) expression of PDEs in cells of the tion and directed migration. Such cells are typically cardiovascular system, (b) changes in PDE expres- involved in processes related to initial establish- sion under certain pathophysiological situations, ment of cardiovascular structures (vasculogenesis) and (c) the potential implications of altered PDE or in the process of neovascularization often expression on the effects of PDE5 inhibitors. More associated with vascularization of newly formed specifically, the regulated expression of PDEs in two tissues in health or disease (angiogenesis). Indeed, distinct phenotypes of smooth muscle cells (SMCs) activated/synthetic SMCs are present during the (quiescent/resting and activated/synthetic) will be development of the cardiovascular system, accumu- described. late in atherosclerotic plaques, and can proliferate and migrate to repopulate the lumen of diseased arteries after successful balloon angioplasty. The two cyclic nucleotides, cyclic adenosine *Correspondence: DH Maurice, PhD, Department of monophosphate (cAMP) and cyclic guanosine Pharmacology & Toxicology, Botterell Hall, A221, Queen’s monophosphate (cGMP), participate in the regulated University at Kingston, Kingston, ON, Canada K7L 3N6. functioning of SMCs of each phenotype. Interest- E-mail: [email protected] ingly, levels of some of the enzymes involved in PDE5 inhibitor cardiovascular effects DH Maurice S21 the synthesis of the cyclic nucleotides in SMCs calcium-calmodulin-regulated PDE1 family en- (eg, soluble guanylyl cyclase) or the effects of the zymes, and PDE5, a cGMP-specific PDE that repre- cyclic nucleotides in cells (e.g., protein kinase G sents the target of agents used in the treatment of (PKG), protein kinase A (PKA)) vary in SMCs of erectile dysfunction (sildenafil, vardenafil and tada- differing phenotypes. Indeed, activated/synthetic lafil), all hydrolyze SMC cGMP. PDE1A and PDE1B SMCs generated by the culturing of quiescent are differentially expressed in SMCs of differing arterial SMCs, or in situ as a result of vascular phenotypes. The species from which the SMCs are damage, demonstrate markedly lower levels of isolated also impact the relative amounts of each several enzymes involved in mediating cGMP PDE1 isoform. effects than are found in quiescent/resting SMCs. In humans, quiescent/resting SMCs express both Recent work from several laboratories has begun to PDE1A and PDE1B, while activated/synthetic SMCs elucidate differences in PDE expression in SMCs express only PDE1A.6 At present, most data are under various conditions. In this context, variants of consistent with levels of PDE5 expression being the PDE3 and PDE4 families play important roles in unaltered between quiescent and activated SMCs. the hydrolysis of cAMP in SMCs. While both However, given the potential importance of changes quiescent/resting and activated/synthetic SMCs each in PDE5 expression during phenotypic modulations, express both PDE3 gene products (PDE3A and and the impact of such differences on PDE5 PDE3B),1,2 levels of PDE3-based cAMP hydrolysis, inhibitor-mediated regulation of SMC function, a and of PDE3A protein expression, is substantially thorough and detailed analysis of this issue in lower in activated/synthetic SMCs than in quies- human SMCs is needed. cent/resting SMCs.3 In contrast to PDE3, levels of In addition to the phenotype-based difference in PDE4 activity and of PDE4D expression are similar in PDE activity and expression in SMCs, the subcel- quiescent/resting and activated/synthetic SMCs. lular expression pattern of these enzymes and their One of the most interesting modes of regulating regulated expression by various signaling systems PDE3 involves the binding of cGMP to the catalytic also differ. For example, while PDE3A and domain of these enzymes, thereby causing compe- PDE3B are each expressed in SMCs, PDE3A alone titive inhibition of cAMP hydrolysis by these is found in cytosolic fractions, while both the enzymes. This inhibitory effect of cGMP on cAMP PDE3A and PDE3B are expressed in particulate hydrolysis accounts for findings that nitric oxide fractions of these cells. Although phosphorylation (NO)-releasing vasodilators increased platelet and by PKA activates both the PDE3 variants, only SMC cAMP, and that these agents synergistically PDE3B levels increase following prolonged eleva- raised cAMP when used in combination with tions in cellular cAMP.2 activators of adenylyl cyclase, such as prostaglandin Similarly, PDE4D activity and expression increase 4,5 E1 or isoproterenol. subsequent to elevated cAMP levels in both quies- Phosphodiesterase 1C (PDE1C), a calcium–calmo- cent/resting and activated/synthetic SMCs, although dulin regulated enzyme that hydrolyzes both cAMP the different PDE4D variants involved are pheno- and cGMP, is also expressed in human SMCs in type dependent. Thus, although quiescent/resting a phenotype-dependent manner. Indeed, while SMCs increase PDE4 activity through the PKA- PDE1C is not expressed in human quiescent/resting dependent phosphorylation and increase expression SMCs, expression of PDE1C is markedly induced of the endogenously expressed PDE4D variant in activated/synthetic human SMCs. High levels of (PDE4D3), activated/synthetic SMCs respond by PDE1C expression were observed in primary cul- activating PDE4D3 and by inducing the expression tures of SMCs derived from explants of human of two PDE4D variants not expressed in SMCs prior newborn and adult aortas, and in SMCs cultured to such treatments (PDE4D1 and PDE4D2).8 from severe atherosclerotic lesions.6 PDE 1C repre- Chronic in vivo exposure to the NO-releasing sented the major cAMP hydrolytic activity in these vasodilator, nitroglycerin, induces expression of SMCs. Since inhibition of PDE1C leads to the Ca2 þ /calmodulin-stimulated PDE1A1, which pre- suppression of human SMC proliferation, PDE1C ferentially degrades cGMP in the rat aorta. Rats expression may serve as a useful marker of human made tolerant to nitroglycerin by continuous infu- SMCs proliferation,7 and PDE1C inhibitors may sion for 3 days showed an increase in Ca2 þ / target proliferating SMCs in relevant situations such calmodulin-stimulated PDE1A1 and a 2.3-fold in- as atherosclerosis. crease in PDE1A1 in their aortas.9 Selective inhibi- In conclusion, enzymes of the PDE1, PDE3 and tion of PDE1 partially restored the sensitivity of PDE4 families of PDE enzymes participate in the tolerant vessels to subsequent nitroglycerin expo- hydrolysis of cAMP in SMCs, and the species from sure. This suggests that increased PDE1A1 activity which these cells are isolated and their phenotypes can decrease cGMP levels and that this mechanism will alter significantly the contribution of each may contribute to the reduction of nitroglycerin- family to cAMP hydrolysis. mediated vasodilation (Figure 1). Several PDEs participate in the hydrolysis of PDE2, a cGMP-stimulated PDE that hydrolyzes cGMP in SMCs. Thus, PDE1A and PDE1B, two either cAMP or cGMP, is not expressed in SMCs.

International Journal of Impotence Research PDE5 inhibitor cardiovascular effects DH Maurice S22 VEC

EDRF VSMC

GTP cGMP PDE3

PDE4 5'AMP cAMP 5'AMP Figure 2 cGMP inhibition of PDE3 and effect on PDE4 in endothelium-intact arteries.

cells. Indeed, in healthy blood vessels with intact endothelial cell layers, VEC-generated NO regulates vascular tone by regulating SMC cGMP levels. Since, in this context, NO-mediated cGMP might be expected to inhibit constitutively a significant portion of PDE3 activity, one could hypothesize that PDE4 inhibitors might produce a more marked effect Figure 1 PDE1A1, but not PDE5A, is upregulated in tolerant rat on blood pressure than might be expected based on aorta. their relatively poor relaxant effects in in vitro experiments with endothelium denuded blood vessel strips (Figure 2). Indeed, evidence consistent with such a mechanism has been reported pre- However, this enzyme is expressed in some vascular viously.12 Since PDE5 inhibitors would be expected endothelial cells (VECs), human platelets and to increase cGMP in many cell types, including cardiac myocytes (cardiomyocytes), cell types es- SMCs, VECs and platelets, in which levels of both sential to cardiovascular control. As these cell types cAMP- and cGMP-hydrolyzing PDE activities play express PDE2, in addition to other PDEs: (PDE1, crucial roles in regulating cell functions, ‘cross-talk’ PDE3, PDE4 and PDE5 in VECs), (PDE3 and PDE5 in between PDEs in these cells will likely have marked platelets), and (PDE3 and PDE4 in cardiomyocytes), effects on their activity in tissues. A detailed their responses to cGMP-elevating agents, such as analysis of these effects will undoubtedly be NO releasing vasodilators, or the atrial natriuretic possible with the highly selective PDE5 inhibitors peptide (ANP), depend on the relative levels of now available. hydrolysis catalyzed by these enzymes in these cell The overall cardiovascular status of a patient types. This process has been well studied in blood significantly impacts the levels of the individual platelets. Thus, in the presence of an activator of PDEs expressed in the cells of tissues of this system, adenylyl cyclase, low concentrations of NO-releas- and these differences likely will permit ‘cross-talk’ ing vasodilators (e.g., nitroprusside, molsidomine) between cAMP and cGMP signaling systems, un- cause a synergistic increase in cAMP due to anticipated from experiments conducted with inhibition of cAMP hydrolysis by cGMP acting at ‘healthy’ cardiovascular tissues. 4,10 PDE3. However, at higher concentrations of the Important questions include: NO-releasing vasodilators, the PDE3-mediated effect becomes saturated and a cGMP-stimulated hydro- (1) Can the dynamics of short-form spliced var- lysis of cAMP and cGMP by PDE2 predominates. iants significantly alter the specificity of a PDE Indeed, in some circumstances, activation of PDE2 inhibitor? abolished the cAMP-increasing effects of PDE3 inhibitors in cells expressing both PDE2 and Several factors impact this question. First, since PDE3.10 PDE-family specific inhibitors act by interfering Similarly, in cardiac fibroblasts, NO production with active site binding of substrates, it is unlikely decreased cAMP accumulation largely by the cGMP- that these agents will discriminate between long and mediated activation of PDE2.11 Because VECs short forms of PDEs to an extent sufficient to allow express both PDE2 and PDE3 and release NO, and splice-form-based drug selectivity. However, since quiescent/resting SMCs depend in large part on synthetic/activated SMCs respond to prolonged PDE3 and PDE4 for cAMP hydrolysis, dynamic challenges with cAMP-elevating agents by inducing regulation of systemic blood pressure may depend short forms of PDE4D, while quiescent/resting SMCs on tight regulation of VEC and SMC cAMP through do not, and these variants can be anchored at coordinated regulation of such activities in these different intracellular sites, effects of inhibitors

International Journal of Impotence Research PDE5 inhibitor cardiovascular effects DH Maurice S23 may not be predicted solely on the ‘total’ PDE cGMP in most cells might influence this dynamic to catalytic activity of individual cell types. attenuate any anticipated reduced effects of PDE5 (2) In cells with PDE2, PDE3 and PDE5, will a inhibitors. PDE5 inhibitor alter the balance between All these questions form fertile ground for further cGMP and cAMP levels? studies on the dynamics of PDE expression and cyclic nucleotide-mediated regulation of cardiovas- In blood platelets, the cell type in which this issue cular function. has been most carefully studied, the answer is most certainly yes. Indeed, results from these studies would predict that PDE5 inhibitors would influence the activities of PDE2 and PDE3 and, in certain References cases, alter the relative importance of these enzymes in cells. 1 Liu H, Maurice DH. Expression of cyclic GMP-inhibited (3) Given the role of adenylate cyclase in cross- phosphodiesterases 3A and 3B (PDE3A and PDE3B) in rat activation between elevated levels of cAMP tissues: differential subcellular localization and regulated expression by cyclic AMP. Br J Pharmacol 1998; 125: and PKG, what is the potential for adenylate 1501–1510. cyclase activators in future therapies for ED? 2 Palmer D, Maurice DH. Dual expression and differential regulation of phosphodiesterase 3A and phosphodiesterase Unless it can be shown that selective activation of 3B in human vascular smooth muscle: implications for adenylyl cyclase can be triggered in response to phosphodiesterase 3 inhibition in human cardiovascular erectogenic stimuli, it is unlikely that sufficient tissues. Mol Pharmacol 2000; 58: 247–252. selectivity could be achieved to allow activators of 3 Dunkerley HA et al. Reduced phosphodiesterase 3 activity and phosphodiesterase 3A level in synthetic vascular smooth adenylyl cyclase to be used in ED. muscle cells: implications for use of phosphodiesterase 3 (4) In tissues with relatively high endogenous inhibitors in cardiovascular tissues. Mol Pharmacol 2002; 61: 1033–1040. levels of PDE5, what might be the future 4 Maurice DH, Haslam RJ. Molecular basis of the synergistic therapeutic implications of PDE5 inhibition inhibition of platelet function by nitrovasodilators and on altered smooth muscle phenotypes in activators of adenylate cyclase: inhibition of cyclic AMP pulmonary hypertension? breakdown by cyclic GMP. Mol Pharmacol 1990; 37: 671–681. 5 Maurice DH, Haslam RJ. Nitroprusside enhances isoprenaline- Selective vasodilation of the pulmonary vascula- induced increases in cAMP in rat aortic smooth muscle. ture with PDE5 inhibitors has been reported. Given Eur J Pharmacol 1990; 191: 471–475. 6 Rybalkin SD et al. Calmodulin-stimulated cyclic nucleotide the limited number of options available to physi- phosphodiesterase (PDE1C) is induced in human arterial cians for the control of pulmonary hypertension, the smooth muscle cells of the synthetic, proliferative phenotype. relative role played by PDE5-mediated hydrolysis of J Clin Invest 1997; 100: 2611–2621. cGMP inhibition, as opposed to activation of soluble 7 Rybalkin SD et al. Regulation of cGMP-specific phosphodies- guanylyl cyclase, in the control of pulmonary terase (PDE5) phosphorylation in smooth muscle cells. J Biol Chem 2002; 277: 3310–3317. vasodilation is warranted. 8 Tilley DG, Maurice DH. Vascular smooth muscle cell phos- (5) Might reduced levels of soluble guanylate phodiesterase (PDE) 3 and PDE4 activities and levels are regulated by cyclic AMP in vivo. Mol Pharmacol 2002; 62: cyclase in atherosclerotic SMCs alter the effect 497–506. of PDE5 inhibitors in patients with cardiovas- 9 Kim D et al. Upregulation of phosphodiesterase 1A1 expres- cular disease? sion is associated with the development of nitrate tolerance. Circulation 2001; 104: 2338–2343. An extensive literature exists describing reduced 10 Dickinson NT, Jang EK, Haslam RJ. Activation of cGMP- levels of several cGMP-regulated enzymes, includ- stimulated phosphodiesterase by nitroprusside limits cAMP ing soluble guanylyl cyclase, in SMC rendered accumulation in human platelets: effects on platelet aggrega- tion. Biochem J 1997; 323: 371–377. synthetic/activated by culturing. Some of these 11 Gustafsson AB, Brunton LL. Attenuation of cAMP accumula- enzymes are also less abundant in in vivo situations tion in adult rat cardiac fibroblasts by IL-1beta and NO: role of in which synthetic/activated SMCs are abundant. cGMP-stimulated PDE2. Am J Physiol Cell Physiol 2002; 283: Although intuitively one would predict that re- C463–C471. 12 Lugnier C, Keravis T, Eckly-Michel A. Cross talk between NO duced levels of soluble guanylyl cyclase might and cyclic nucleotide phosphodiesterases in the modulation influence the impact of PDE5 inhibition in cells, a of signal transduction in blood vessel. J Physiol Pharmacol tight coupling between synthesis and hydrolysis of 1999; 50: 639–652.

International Journal of Impotence Research