Sildenafil’s protective effect against cardiac hypertrophy

Left ventricular hypertrophy is a leading contributor to cardiovascular morbidity and mortality, increasing the risk of heart failure, myocardial infarction and stroke by three- to ten-fold [1]. Moreover, the prevalence of cardiac hypertrophy is increasing, probably reflecting the advancing age of the world’s population and the global epidemics of diabetes, obesity and hypertension. Hypertrophy occurs in nearly a third of patients with arterial hypertension and is common to most forms of valvular and ischemic heart disease [2,3]. Although hypertrophy has been viewed as providing some initial functional compensation to increased pressure or volume stress, studies increasingly suggest that the response is never really adaptive when the stress imposed is pathologic. Prolonged hypertrophic stimulation can lead to frank heart failure, and even short-term exposure to pathologic versus physiologic stress triggers maladaptive signaling cascades [4]. Recent antihypertension clinical trials have found that the reversal of hypertrophy independently predicts lower cardiovascular mortality risk [5,6], while mice harboring genetic mutations that suppress hypertrophy often respond more favorably to pressure stress [7,8].

The evolution of hypertrophy involves a broad array of signaling modifications, involving kinases, phosphatases, transcription factors and miRNAs, and virtually all of these pathways are presently being pursued as potential therapeutic targets [2]. Hypertrophy is also negatively regulated by a variety of enzymes and, among these, one that has gained particular attention, owing in part to the availability of clinically effective, small-molecule modulators, is that related to cyclic guanosine monophosphate (cGMP) and its primary target kinase, protein kinase-G (PKG) [2,9].

In the cardiovascular system, cGMP/PKG plays a central role for vascular tone and proliferation, and in the heart it provides a brake-like effect to blunt cardiac growth and stress-response signaling [9,10]. It was long recognized to be anti-hypertrophic, first based on responses to nitric oxide and natriuretic peptides (NPs) that stimulate cGMP synthesis [11], and later by genetically engineered models that altered the pathway [12–15]. However, these effects were fairly modest, with long-term efficacy limited by counter-tolerance mechanisms and/or hypotension, since stimulation of cGMP/PKG induces systemic vasodilation. Interest was reignited with the evidence that clinically effective agents that block phosphodiesterase (PDE)5 (e.g., sildenafil, tadalafil and vardenafil), a cGMP-selective PDE, also suppress pressure-stress-induced hypertrophy [16]. Such agents are widely used to treat erectile dysfunction and, increasingly, pulmonary hypertension, both targeting vascular beds with high PDE5 expression levels to induce vasodilation. However, in the normal heart at rest, PDE5 protein expression is very low and its inhibition has minimal functional impact, which had suggested a minor role, if any [17]. This was the dominant view until studies began to show that PDE5 inhibitors could protect the heart against pressure overload, as well as other cardiac stress syndromes, such as ischemia/infarction [18–21]. These and other data have helped to rekindle interest in modulating cGMP via PDE5 inhibition for treating heart muscle disease (i.e., not just as a vasodilator), with at least one major multicenter trial (NIH-sponsored, Phosphodiesterase-5 Inhibition to Improve Quality of Life and Exercise Capacity in Diastolic Heart Failure [RELAX] [101]) now underway to evaluate the efficacy of sildenafil for treating heart failure with a preserved ejection fraction. The evolving story has yielded some intriguing observations and we are still at the fairly early stages in our understanding of the mechanisms involved.

cGMP/PKG signaling & a role for compartments

The myocyte has two mechanisms to generate cGMP, one coupled to soluble guanylyl cyclase (sGC), which is stimulated by nitric oxide, and the other from a receptor-linked GC activated by NPs [9]. Both methods of enhancing cGMP have antihypertrophic effects in vitro and in vivo. The latter has primarily been tested using genetic models, in which one or the other pathway is disturbed by a gain or loss of function model [12–15]. As noted, the magnitude of modulation of cardiac pressure-overload-induced hypertrophy associated with altering cGMP, synthesis while significant, has generally been modest, raising questions concerning the real importance of this pathway to the heart. However, cGMP is also controlled by its hydrolysis to 5´GMP by members of the PDE superfamily [22]. Of the 11 primary isozymes, PDE5 is specific for cGMP and was the first, as well as to date the best, characterized cGMP-PDE isoform. Its expression in cardiac myocytes is 1000-times lower than in the lung, yet it can regulate several aspects of heart function [23]. For example, its acute inhibition suppresses the acute β-adrenergic-stimulated contractility, the only cardiac effect demonstrated in both experimental animals and humans to date [23–25].

There are several features of PDE5 that appear relevant to its cardiac regulation. One is localization. It is normally expressed at higher levels in the cardiac myocyte z-disk region and this localization appears important for its modulation of β-adrenergic-stimulated contractility [23,25,26]. Its hydrolysis of cGMP also appears to be fairly selective for that generated by sGC, whereas NP-stimulated cGMP is little impacted, at least in normal myocytes [27]. This further highlights compartmentalized regulation. Loss of nitric oxide synthase (NOS)3 activity and/or expression results in its relocalization away from z-disks [23,26] and, with this, the capacity of PDE5 inhibitors to modulate adrenergic stimulation and chronic pressure-overload hypertrophy. The diffuse expression in cells with chronic NOS inhibition is reversible if NOS activity is restored or sGC is chronically and directly activated [26]. Precise protein binding partners responsible for this localization and intracellular movement remain unknown, but this topic is being actively studied. Another important feature of PDE5 regulation is its enhanced expression and activation with chronic cardiac stress and failure. This has been demonstrated experimentally in mouse left ventricle (LV) [16] and rat right ventricle (RV) subjected to pressure overload, and in humans with chronic hypertrophy (both RV and LV) [28] and those with dilated heart failure [29]. Upregulation can provide a mechanism whereby the normal role of cGMP/PKG signaling as an ‘anti-stress brake’ is diminished, while responses to PDE5 inhibition are correspondingly enhanced. In keeping with this notion, mice overexpressing PDE5 in myocytes display a worsened response to myocardial infarction [29]. Lastly, PDE5 catalytic activity for cGMP is itself stimulated by cGMP-binding in the N-terminus GAF domain and PKG phosphorylation at Ser92 (also in the N-terminus) [17]. PDE5 inhibitors, such as sildenafil, are cGMP mimetics that function by competitive binding at the catalytic site but, in this case, the rise in cGMP and resulting PKG stimulation now improves binding of sildenafil to the catalytic site. This may be an important mechanism for enhancing and prolonging tissue sildenafil effects beyond those predicted by plasma-level pharmacokinetics.

The inhibition of pressure-overload-induced hypertrophy by PDE5 inhibitors was first reported using oral sildenafil treatment in mice (100 mg/kg/day mixed in rodent food) [16]. While the dose seems high, it in fact yields free plasma concentrations comparable to humans taking 50–100 mg. More recently, we showed that sildenafil can treat more advanced pre-established pressure-overload hypertrophy, stopping progressive LV remodeling, reversing fibrosis and improving myocyte contraction and calcium handling [30]. With the lack of a cell-targeted gene deletion model, the precise cellular target in such pharmacologic studies remains uncertain. Sildenafil can modulate vascular tone, although its coronary effects are modest. Indeed, its first clinical trial was an angina study that was not positive, but the side effect led to its use for erectile function. PDE5 is also expressed in fibroblasts, which could contribute to its impact on remodeling [31,32]. Direct evidence for antihypertrophic effects of PDE5 gene knockdown has been reported in neonatal cardiac myocytes. Importantly, the effect was similar to that observed with sildenafil, with no additive effect from both [33].

Beneficial effects of sildenafil on the heart extend well beyond pressure-overload hypertrophy. The Kukreja laboratory has studied its effects on ischemic models extensively, including infarction and ischemia/reperfusion injury, where a primary anti-apoptotic effect is observed [19,21]. This has been attributed to the activation of extracellular response kinase and glycogen synthase kinase, and corresponding stimulation of mitochondrial K_ATP channels [34,35]. Furthermore, sildenafil improves abnormal fetal gene expression in dystrophin-deficient (mdx) mouse hearts [36]. In humans, sildenafil improves cardiopulmonary exercise capacity in heart failure patients with acute or long-term (12 weeks) treatment, which may involve more than effects on the pulmonary vasculature [37,38].

Antihypertrophy targets of PDE5 inhibition

The physiological action of cGMP is mediated by PKG, cGMP-regulated PDEs and cGMP-gated cation channels, the first two being relevant to cardiac myocytes [9]. Although direct proof in models with inducible and targeted myocyte gene deletion of PKG remains lacking, most current evidence supports PKG activation as playing a central role in the antihypertrophic effects of cGMP and its regulation by PDE5 inhibitors. PKG can phosphorylate subunits of the L-type calcium channel to blunt calcium current [39]; although whether or not this is modified by PDE5 regulation remains uncertain. PKG also phosphorylates: troponin I, resulting in myofilament calcium desensitization [40]; titin [41], which modifies myocardial diastolic stiffness; and regulators of G-protein signaling (RGS2 and RGS4) that modulate G-coupled stimulation [42,43]. The calcineurin–NFAT pathway was the first prohypertrophic signaling cascade shown to be negatively targeted by PKG activation in isolated myocytes [44], and similar suppression was observed from PDE5 inhibitors both in vitro and in vivo[16]. The mechanism remains unclear but data suggested a proximal target. However, mice lacking the catalytic subunit of Cn (CnAβ) still display a hypertrophic response to pressure overload (albeit less than littermate controls) that is potently inhibited by sildenafil treatment [45].

One pathway that appears to be essential for sildenafil antihypertrophic activity early in response to pathologic stimulation is RGS2 [46]. RGS2 enhances restoration of the heterotrimeric Gq-protein complex to terminate its signal activation. It is activated by PKGIα via direct protein binding and phosphorylation [42]. Mice lacking RGS2 develop rapid marked hypertrophy, chamber dilation and early mortality from pressure overload, which is coupled with enhanced Gαq-signaling cascades. Sildenafil activates PKG and thus RGS2, inhibiting Gq-related signaling cascades in wild-type hearts subjected to pressure overload, whereas similar PKG activation by sildenafil does not impact RGS2-deficient hearts [46]. Whether this mechanism applies in later stages of disease remains unknown and awaits conditional gene deletion models. RGS4 also enhances GTPase activity, targeting both Gq and Gi, and has been implicated in cGMP–PKG anti-hypertrophy, coupled to NP receptor activation [43]. However, the role of RGS4 in more pathophysiologic disease models, such as pressure overload, remains unknown.

One theme that has evolved from these studies is that the antihypertrophic efficacy of cGMP/PKG requires the presence of PKG-targeted cascades, and Gq-coupled signaling plays a major role in this regard. Without such stimulation and associated signaling, upregulating cGMP/PKG pathways has little impact. For example, RGS2^-/- mice develop similar levels of adaptive hypertrophy without cardiac dysfunction in response to swimming exercise as their littermate controls [46]. This stress does not trigger Gq-related cascades. Furthermore, treatment of RGS2^-/- mice with an inhibitor of phospholipase Cβ restores their ability to adapt to pressure overload [46]. This same mechanism probably explains why PDE5 inhibition has less impact on the early phases of moderate pressure overload in mice, which lack Gq activation, than on more severe loading applied over a similar time course [47]. Only when the moderate overload is prolonged to the point that Gq activation occurs can antihypertrophic effects from PDE5 inhibition be observed. These observations have potential implications for the clinical use of these agents to treat heart disease. However, genetic hypertrophy from mutations in sarcomeric proteins have not been shown to date to involve the same type of stress-response cascades observed with hypertensive/hypertrophic disease [48]. Thus, whether cGMP/PKG enhancement would be efficacious in such disorders remains unclear. This behavior could also potentially influence responses in individual patients with heart disease.

Does PDE5 regulate the right & left heart in the same way?

After its success for erectile dysfunction, sildenafil was subsequently tested for its capacity to benefit pulmonary hypertension [49,50]. Experimental and human studies demonstrated benefits on right heart that were initially thought to be coupled with its reduction of pulmonary vascular resistance [51,52]. The demonstration of direct cardiac effects in the left heart also raised the potential for an additional mechanism of benefit in pulmonary artery hypertension patients – improving right heart function with reduced hypertrophy. However, several recent studies have questioned whether this occurs. Schäfer et al. found that sildenafil failed to inhibit RV hypertrophy from pressure overload caused by pulmonary artery banding, whereas it did ameliorate RV hypertrophy induced by monocrotaline [53]. In another study, sildenafil treatment improved RV systolic function but induced more hypertrophy in a similar pulmonary artery banding model [54]. It is possible that intracellular signaling regulation coupled to PDE5 in the RV versus LV is different. In an intriguing study, Nagendran et al. found that with RV pressure overload, PDE5 inhibition increased cAMP levels but did not appear to elevate PKG, via coinhibition of cGMP-sensitive cAMP-hydrolyzing PDE3 [28]. In their investigation, sildenafil stimulated cAMP/PKA to the same extent as dobutamine. This contrasts sharply with observations in the LV, where PDE5 inhibition did not impact myocardial cAMP levels in the LV exposed to pressure overload [16]. It remains possible that NOS, GC or PDE5 may be differentially regulated in the RV myocyte, but this is yet to be clearly demonstrated.

Summary

Experimental models have shown that chronic PDE5 inhibition with sildenafil treatment and subsequent PKG activation inhibits cardiac hypertrophy development and stops progressive remodeling in well-established cardiac hypertrophy. The mechanism of the former may be RGS2 activation by PKG, leading to Gαq signal inhibition, including Cn/NFAT. Ongoing studies are testing other Gαq-coupled cascades, including the role of transient receptor potential canonical channels, that have been linked to Cn/NFAT activation. Discovery of additional novel targets within the sarcomere and coupled to other central signaling cascades is anticipated. It is likely that the mitochondrial changes revealed in ischemia studies will also play a role in the antihypertrophic signaling process. The issue of RV versus LV disparities is very intriguing, but this needs to be addressed at the molecular level. The ongoing NIH-funded trial will help to test the relevance of the cardiac PDE5 modulation data to a very relevant human disease (nearly 50% of all heart failure). The role of PDE5 inhibition in systolic dysfunction remains unclear, but in settings where PKG-targetable cascades play a pathophysiologic role, one would anticipate some benefits as well.

Financial & competing interests disclosure

This work was supported by American Heart Association Scientist Development Grant (Eiki Takimoto), and grants from the NIH HL-093432 (Eiki Takimoto) and HL-089297 (David A Kass, Eiki Takimoto). The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.