Controlling Myocyte cGMP — PDE1 Joins the Fray

cGMP is a central intracellular second-messenger regulating numerous cellular functions. In the cardiac myocyte, cGMP mediates effects of nitric oxide and atrial natriuretic peptide, whereas its counterpart, cAMP, mediates catecholamine signaling. Each cyclic nucleotide has a corresponding primary targeted protein kinase, PKA for cAMP, and PKG for cGMP. PKA stimulation is associated with enhanced contractility and can stimulate growth, whereas PKG acts as a brake in the heart, capable of countering cAMP-PKA-contractile stimulation and inhibiting hypertrophy¹. Importantly, the duration and magnitude of these signaling cascades are determined not only by generation of cyclic nucleotides, but also by their hydrolysis catalyzed by phosphodiesterases (PDEs). PDE regulation is quite potent – often suppressing an acute rise in a given cyclic nucleotide back to baseline within seconds to minutes². It is also compartmentalized within the cell, so that specific targeted proteins can be regulated by the same “generic” cyclic nucleotide³. For many years, the only PDE in the crosshairs for cardiac biologists was PDE3, a principally cAMP-targeted PDE whose inhibition served as the basis for drugs such as milrinone as a heart failure therapy. However, this list was recently expanded with the recognition of PDE4 as a regulator of beta-receptor signaling and excitation-contraction coupling⁴, and PDE5 for its regulation of cardiac stress responses¹. With the study of Miller et al⁵, in the current issue if Circulation Research, we can now add PDE1 to the list of hypertrophy regulators – via its modulation of cGMP in the myocyte.

The mammalian PDEs comprise a 21 gene superfamily of enzymes grouped into 11 iso-enzymes (PDE1- PDE11) based on sequence homology, enzymatic properties and sensitivity to inhibitors². These iso-enzymes harbor different specificities to cAMP, cGMP or both, and are also differentially expressed in a variety of tissues. PDE1, PDE2, PDE3, PDE4, PDE5 and PDE9 are expressed in the heart and among these, PDE5 and PDE9 are highly specific to cGMP² (Table). Recent studies have demonstrated a role for cGMP modulation by PDE5 in the heart¹, though the role for PDE9 is unknown at present. PDE5 is up-regulated in failing human⁶ and hypertrophied mice ventricles⁷. PDE5 inhibitor ameliorates cardiac hypertrophy in mice⁷, and genetic silencing of PDE5 inhibits cardiac myocyte hypertrophy in vitro⁸. Cardiac PDE5 over-expression leads to exacerbated remodeling after myocardial infarction⁶. Furthermore, the cardio-protective effects from PDE5 inhibitors have been reported in various animal models of cardiac pathology, including myocardial infarction, ischemia-reperfusion injury, doxorubicin cardiomyopathy, and cardiomyopathy associated with dystrophin deficiency¹.

Table.

PDEs in the heart

PDE isoenzyme		Km (mM) cAMP	Km (mM) cGMP	Heart expression
PDE1	Ca2+/CaM regulated dual specificity	PDE1A:73–120 PDE1B: 10–24 PDE1C: 0.3–1.2	PDE1A:2.6–5 PDE1B:1.2–5.9 PDE1C:0.6–2.2	yes
PDE2	cGMP-stimulated dual specificity	30–50	10–30	yes
PDE3	cGMP-inhibited dual specificity	0.02–0.15	0.18	yes
PDE4	cAMP-specific	2.9–10	-	yes
PDE5	cGMP-specific	-	1–6.2	yes
PDE9	cGMP-specific	-	0.17–0.39	?

PDE5 is not the only cGMP hydrolyzing PDE in myocardium, as basal activity represents ~30% of total myocardial cGMP-esterase activity in mice⁷ and dogs⁹. Although this activity can rise with chronic pathologic remodeling (e.g. increasing >50% of the total cGMP esterase activity in the pressure-overloaded mouse heart⁷), a large proportion of cGMP-esterase activity is still attributable to other PDEs. A primary candidate has been the dual substrate PDE1. Unlike PDE5 which is stimulated by cGMP binding and PKG phosphorylation, PDE1 is activated by calcium-calmodulin, thus is an appealing target for stress-stimulated regulation. Recent studies have proposed that PDE1 (particularly PDE1c) is a more prominent regulator of cGMP hydrolysis in vitro in human normal and failing myocardium ^{10, 11} – raising interest in defining its role in myocytes in greater detail.

The study of Miller et al. has now revealed an important role of PDE1A in cardiac myocytes and its response to pro-hypertrophic stimulation. PDE1 has three primary isoforms, PDE1A-C. PDE1A and PDE1B preferentially hydrolyze cGMP with greater affinity than cAMP, while PDE1C hydrolyzes both cAMP and cGMP with equal affinity. Prior studies had revealed expression and in vitro enzyme activity^{10, 11}, but to date, a physiologic role of PDE1 in myocytes and intact hearts has remained unknown, due to the lack of specific inhibitors for the enzyme. Using one such inhibitor (IC86340, developed by ICOS, and unfortunately no longer available), Miller et al. demonstrated an anti-hypertrophic effect in both isolated neonatal and adult rat cardiac myocytes stimulated with phenylephrine or isoproterenol. The IC50 of IC86340 shows high specificity for PDE1C (0.06μM), PDE1B (0.21μM), PDE1A (0.44μM) as compared to other myocyte PDEs (over 100μM for PDE2, 3, 4, 5, 9). Though prior reports had focused on PDE1C as the dominant isoform¹¹, the current work clearly highlighted PDE1A. The anti-hypertrophic effect of IC 86340 was comparable to that using the PDE5 inhibitor sildenafil in neonatal rat cardiac myocytes, and intriguingly both combined resulted in a further silencing of the hypertrophic effect, suggesting different and likely compartmentalized targeted pools of cGMP and distal signaling are involved. The current study also showed up-regulation of PDE1A expression in mice exposed to sustained neurohormones or pressure-overload. Such upregulation, as has been observed with PDE5, could itself depress cGMP, contributing pathologic cardiac remodeling.

The study by Miller et al. leaves open the underlying mechanisms by which PDE1 inhibition resulted in suppression of hypertrophy. However, some of the same pathways already reported from natriuretic peptide stimulation or PDE5 inhibition could play a role. These include inhibition of Gq-coupled signaling and in particular inhibition of the calcineurin-NFAT pathway by PKG. The key regulator appears to be regulator of G protein signaling (RGS) – both RGS2 and RGS4. These proteins are GTPase accelerators, functioning as negative regulators of Gq activation. Mice lacking RGS2 have exacerbated hypertrophic response to pressure-overload that is not inhibited by sildenafil¹². Overexpression of RGS4 in mice lacking natriuretic peptide receptor type A rescues the hypertrophic phenotype in the latter coupled to calcineurin-NFAT inhibition¹³. The present study examined responses to isoproterenol which can stimulate Cn-NFAT pathway, but does not replicate Gq signaling presented by pressure-overload. More studies will be needed.

Another unexplored question is whether and how different pools of cGMP are targeted in localized sub-cellular pools by PDE1 versus PDE5. Such compartmentation would be consistent with what is now recognized to be a general feature of PDE regulation³. Compartmentalized cGMP regulation in myocytes has already been demonstrated for PDE5 and another cGMP targeting (and dual esterase) PDE2. For example, nitric oxide stimulated soluble guanylyl cyclase generates cGMP that is hydrolyzed by PDE5 localized at Z-bands of adult cardiac myocytes, and inhibiting PDE5 blunts acute beta-adrenergic responses coupled to enhanced PKG activation¹⁴. In contrast, ANP stimulation of the receptor guanylyl cyclase generates cGMP that has no impact on beta-adrenergic responses¹⁴, and appears targeted more by PDE2 at the sarcolemmal membrane¹⁵. We do not yet know where PDE1 is localized intra-cellularly, whether it is differentially coupled to the two cGMP synthetic pathways, if it modulates acute adrenergic stimulation, and/or whether it targets different or overlapping pools of cGMP to that by PDE5. All are intriguing questions to pursue.

There are some controversies at present regarding which PDE PDE5 or PDE1 is likely to be more dominant in human hearts for cGMP regulation. The discussions remain largely speculative, based purely on in vitro enzyme assays, with no functional data regarding PDE1¹⁰ and some data regarding PDE5¹⁶. Since signaling via PDEs is likely compartmentalized, activity measured in a test-tube from cell extracts may not reflect the in vivo activity. The only human functional data to date with a cGMP-PDE has been with sildenafil in normal volunteers, wherein Borlaug et al showed that acute PDE5 inhibition suppresses dobutamine stimulated contractility¹⁶, just as observed in dogs⁹ and in mice¹⁴. The role of PDE5 inhibition in ameliorating pathologic heart disease is currently being tested in the RELAX trial (NCT00763867), an NIH-sponsored multicenter trial of sildenafil for treating heart failure with preserved ejection fraction. The major limitation of PDE1 studies to date has been the very limited availability of selective inhibitors (none are commercially available). As noted, the drug used by Miller et al. has been available to some at very limited quantities and there is little left. New agents in amounts that facilitate whole animal and even human studies are needed.

In summary, we can add PDE1 to PDE5 as a likely important regulator of cGMP-hydrolysis and mediator of cardiac hypertrophy. Both appear to contribute as regulatory “brakes” to pathological remodeling, though the relative role and targeted signaling which may be species dependent, remains to be sorted out. The potential to combine both inhibitors as a therapy is intriguing and worthy of further studies. Future in vivo studies employing both genetic manipulation of each cGMP-hydrolyzing PDEs and use of chronic selective suppression will hopefully provide needed insights.