Cardiac cGMP Regulation and Therapeutic Applications

Abstract

cGMP plays a central role in cardiovascular regulation in health and disease. It is synthesized by NO or natriuretic peptide activated cyclases and hydrolyzed to 5′GMP by select members of the PDEs (phosphodiesterase) superfamily. The primary downstream effector is cGMP-dependent protein kinase, primarily cGK-1a (cyclic GMP-dependent protein kinase 1 alpha) also known as protein kinase G 1a in the heart and vasculature. cGMP signaling is controlled in intracellular nanodomains to regulate myocyte growth, survival, metabolism, protein homeostasis, G-protein–coupled receptor signaling, and other critical functions. The vascular effects of cGMP signaling have been dominated by its lowering of smooth muscle tone, but other cellular processes are also engaged. Localization of cyclases and corresponding PDEs within intracellular domains, along with their varying expression across different cell types, adds multiorgan complexity to cGMP signaling. This diversity can be leveraged therapeutically by targeting selective pathway components to impact some but not other cGMP signaling effects. Here, we review the generation and regulation of cGMP by PDEs and cyclases, focusing mainly on their role in cardiac physiology and pathophysiology. Current therapeutic uses of cGMP modulation and ongoing trials testing new potential applications are discussed.

Cyclic nucleotides are ubiquitous second-messenger molecules that regulate normal cellular function and homeostasis and play critical roles in disease. There are 2 primary species: cAMP and cGMP. In the heart, these molecules maintain a somewhat Yin-Yang relationship, cAMP is often associated with stress response activation of myocyte function, metabolism, calcium cycling, and growth, whereas cGMP acts more like a cellular brake that can counter cAMP-dependent changes while stimulating its own signaling. For example, cGMP activation dampens β-adrenergic and neurohormonal G-protein–coupled receptor agonism,¹ cation channels linked to fibrosis and pathological growth,² stimulates autophagy and protein quality control,³ and influences mitochondrial function, metabolism,⁴ and contractile function.⁵

Given that their influence on cellular functions varies among cell types, it is not surprising that cGMP signaling occurs within compartmentalized nanodomains inside a cell. This enables different and precise signaling despite being triggered by the same cyclic nucleotide. Cyclic nucleotides are hydrolyzed by PDEs (phosphodiesterases) that are also localized in cellular subdomains to control their impact on the primary downstream effectors—PKG (protein kinase G) and PKA (protein kinase A). For cAMP, localization is achieved via protein supercomplexes coordinated by AKAPs (A-kinase anchoring proteins).⁶ While cGMP also exhibits localized signaling, no parallel GKAP (G-kinase anchoring proteins) and organization has been identified.

Here, we review cGMP synthesis and its regulation in healthy and diseased hearts. Augmented signaling related to this pathway is already leveraged therapeutically, and ongoing studies are exploring broader applications. Specific PDEs (PDEs 1–3, 5, and 9–11) and guanylate cyclases (GC-1, activated by NO, and GC-A/B, activated by natriuretic peptides [NPs]) that control cGMP levels and signaling are examined individually. Recent findings on localized signaling effects and mechanisms are also discussed. Finally, we review current and emerging applications of cGMP-enhancing therapies for treating various cardiovascular disorders.

PDEs and cGMP Regulation

The PDE superfamily has 11 primary members, each with its own molecular structure, regulatory mechanisms, and selectivity for one or both cyclic nucleotides (Figure 1). The catalytic site in the C terminus is shared across all PDEs, with specific differences conferring substrate affinity and selectivity among them. Of those considered selective for cGMP hydrolysis, only 2 are found in the heart: PDE5A and PDE9A. Four others (PDE1, PDE2, PDE3, and PDE10) also hydrolyze cAMP, and the extent they modulate cAMP versus cGMP varies with their isoforms, posttranslational modifications, and cyclic nucleotide cross talk.

Figure 1.

Schematic diagrams of PDEs (phosphodiesterases) with Michaelis constant (Km) values, organelle localization, and inhibitors. Akt/PKB indicates protein kinase B; C, cytosol; CaM, calmodulin; CAMKII, CaM-dependent kinase II; G, Golgi body; Mem, membrane; Mic, microsome; Mito, mitochondria; N, nucleus; NE, nuclear envelop; NHR, N-terminal hydrophobic region; Nuc, nucleus; PAS, period, aryl-hydrocarbon receptor (ARNT) and single-minded domain; PAT7, nuclear localization signal; PKA, protein kinase A; PKC, protein kinase C; REC, response regulator receiver; SM, sarcomere; SR, sarcoplasmic reticulum; T-t, T-tubule; and UCR, upstream conserved region.

PDE1

The PDE1 subfamily is comprised of 3 isoenzymes, PDE1A, PDE1B, and PDE1C, with PDE1A and PDE1C being expressed in the heart and PDE1B predominantly in the brain.⁷ PDE1A and PDE1B are more selective for cGMP (PDE1A: Km [Michaelis Constant] 1–5 µmol/L cGMP versus 50–100 µmol/L cAMP; PDE1B: 3 versus 24.3 µmol/L), whereas PDE1C has similar affinity for both (0.6–2.2 µmol/L for cGMP and 0.3–3.5 µmol/L for cAMP).⁸ PDE1C is the most dominant isozyme in the human heart, whereas PDE1A is more prominent in small rodents.⁷ PDE1C localizes to Z lines and M lines within cardiomyocytes, regulating cyclic nucleotide levels that influence contractility and cardiac stress responses (Figure 2).⁹ PDE1C is also prevalent in vascular smooth muscle cells, where it modulates their response to vasoconstrictors like norepinephrine and angiotensin II by regulating the counterbalancing impact of cGMP on smooth muscle tone.¹⁰ PDE1A is expressed in fibroblasts where its inhibition suppresses profibrotic signaling from pathological stimuli and heart disease.¹¹ PDE1 is unique among all PDEs by requiring calcium/CaM (calmodulin) binding to relieve an autoinhibitory domain. It is also regulated by phosphorylation via PKA or CaMKII (CaM-dependent kinase II), which reduces its CaM affinity, creating a negative feedback loop¹² (Figure 1). Pathological increases in PDE1C expression in heart failure (HF) are associated with myocyte apoptosis and cardiac hypertrophy by degrading cAMP, disrupting PKA signaling, and reducing Akt (protein kinase B)-mediated survival pathways. PDE1C also promotes cardiac fibrosis through transforming growth factor β activation in fibroblasts, contributing to myocardial dysfunction.⁷ PDE1A exhibits limited cAMP hydrolysis activity when membrane bound, but this increases when it is localized to the cytosol.¹³

Figure 2.

Compartmentalization of major PDEs (phosphodiesterases) in cardiomyocytes. PDE1 localizes at M and Z lines; PDE2A at the mitochondria, sarcomere, and nuclear membrane. PDE3A is found at the sarcoplasmic reticulum and in the nucleus, while PDE3B associates with mitochondria at Z bands and T-tubules. PDE4D interacts with β-arrestin, desensitizes β-AR, while PDE4B with AKAPs (A-kinase anchoring proteins) regulates PKA (protein kinase A). PDE5A is localized at the Z disk and PDE9A at mitochondria and sarcomere. These PDEs regulate compartmentalized cAMP and cGMP signaling, influencing calcium handling and cardiac contractility via PKA and PKG (protein kinase G) pathways. AC indicates adenylyl cyclase; ANP, A-type NPs; ATP, adenosine triphosphate; B-AR, beta-adrenergic receptor; BNP, B-type NPs; CNP, C-type NP; GC1/2, Guanylate Cyclase 1 and 2; GC-A, guanylyl cyclase-receptor A; GC-B, guanylyl cyclase-receptor B; GTP, guanosine triphosphate; HDAC1, histone deacetylase 1; LTCC, L-type calcium channel; PLB, phospholamban; RyR; ryanodine receptor; and SERCA, sarcoplasmic/endoplasmic reticulum calcium ATPase.

PDE2A

PDE2A hydrolyzes both cGMP and cAMP (Km, 10 versus 30 µmol/L) and is expressed as 1 of 3 splice variants: PDE2A1, PDE2A2, and PDE2A3 (Figure 2). PDE2A1 is found in the cytosol, whereas PDE2A2 is located in mitochondria, where it is thought to regulate energy metabolism and the Golgi apparatus, plasma membrane, sarcomeric Z disk, and nuclear membrane, allowing for spatially specific control of cardiovascular signaling.^14,15 PDE2A3 is primarily expressed in the brain. While constitutive PDE2A hydrolyzes cGMP more than cAMP, this shifts to cAMP>cGMP when cGMP binds to a GAF-B (cGMP-specific PDEs, adenylyl cyclases, and FhlA) domain in the N terminus (Figure 1). In this way, PDE2A regulates signaling cross talk between cGMP and cAMP.¹¹

In HF, elevated PDE2A expression reduces cAMP coupled with β-adrenergic stimulation, impairing calcium uptake into the SERCA2a (sarcoplasmic reticulum by Ca²⁺-ATPase-2a) by reducing phospholamban phosphorylation.¹⁶ Overexpression of PDE2A in myocardium improves chronic hyper-β-adrenergic models of heart disease and myocardial infarction.¹⁷ It also reduces β-adrenergic–stimulated ventricular arrhythmias by inhibiting EPAC (Exchange Protein Directly Activated by cAMP) and CaMKII signaling.¹⁸ Conversely, PDE2A inhibition reduces pathological hypertrophy and cardiac remodeling from pressure overload stress,¹⁹ so its net impact may be disease dependent.

PDE2A also influences cellular energy balance, mitochondrial dynamics, and apoptosis by promoting Drp1 (dynamin-related protein 1) phosphorylation at S616 via PKA, resulting in mitochondrial fragmentation. Overexpression of PDE2A heightens fragmentation, while inhibition supports mitochondrial elongation, membrane potential preservation, and protection against ionomycin-induced cell death.¹⁵ Additionally, PDE2A interacts with MIC60, part of the mitochondrial contact site and cristae organizing system complex, to regulate Parkin-mediated mitophagy.²⁰ It modulates mitochondrial membrane potential, mitochondrial permeability transition, and calcium import. Reduced oxidative phosphorylation is observed in PDE2-overexpressing transgenic mice.¹⁴

PDE3

PDE3 hydrolyzes both cAMP and cGMP and is a key regulator of cardiac and vascular function in humans and larger mammals.¹¹ PDE3 has 2 isoforms, PDE3A and PDE3B, that differ in their N-terminal regions (Figure 1). PDE3A is located at the nucleus and sarcoplasmic reticulum (SR), where it influences cAMP-mediated control over calcium cycling and contractility. PDE3B localizes to T-tubules and mitochondria, regulating excitation-contraction coupling and myocardial energy dynamics by modulating cAMP and, consequently, PKA signaling (Figure 2). In PDE3B^−/− hearts, reduced reactive oxygen species and decreased calcium-induced mPTP (mitochondrial permeability transition pore) opening ameliorate ischemia/reperfusion (I/R) injury by activating mitochondrial calcium-activated potassium channels and caveolin-3–enriched signalosomes, promoting cardioprotective protein expression.²¹

In human cardiomyocytes, PDE3 primarily hydrolyzes cAMP within membrane-bound fractions, influenced by local intracellular Ca2⁺ and cAMP levels. This can alter its signaling in conditions such as HF, where cAMP levels are typically reduced.¹³ Diminished PDE3A expression worsens I/R injury and promotes cardiomyocyte apoptosis via a feedback loop with ICER (inducible cAMP early repressor), intensified by β-adrenergic activation.²² Reciprocally, cardiomyocyte-targeted overexpression of PDE3A1 in mice reduces infarct size and myocyte apoptosis from ischemia/reperfusion injury, with effects similar to those from β-blockers.²³ By contrast PDE3B knockdown is cardioprotective.²¹ Inhibiting both PDE3 isoforms increases PKA-mediated phosphorylation of HDAC-1 (histone deacetylase 1), stimulating hypertrophic gene expression promoting hypertrophy,²⁴ but also reduces post-I/R myocardial damage by inhibiting mitochondrial permeability transition pore opening²⁵ and myocardial apoptosis by activating p38 mitogen-activated protein kinase.²⁶

PDE3 hydrolysis of cGMP has not historically been considered a major contributor to cGMP/PKG signaling, particularly in the resting state. cGMP acts as a competitive inhibitor to cAMP in the catalytic binding site of PDE3; thus, low cGMP levels augment cAMP/PKA signaling.²⁷ Recent studies using a cGMP-responsive Förster resonance energy transfer sensor overexpressed in mouse cardiomyocytes detected significant constitutive cGMP elevation upon PDE3 inhibition.²⁸ This was recapitulated in early-stage (30 days old) cardiomyocytes derived from human induced pluripotent stem cells, whereas longer matured cells (90 days) had little regulation by PDE3 but more from PDE1, PDE2, and PDE5.²⁹ This suggests important differences in cGMP control by PDE3 between small rodent and human.

PDE5A

PDE5A is most highly expressed in vascular smooth muscle in the lungs and in the corpus cavernosum but much less in systemic vascular smooth muscle. It is also expressed at low levels in cardiomyocytes in normal hearts, but this increases with dilated HF.³⁰ PDE5A normally localizes to Z-disk regions in cardiomyocytes (Figure 2), although in dilated HF, pressure overload, and with chronic inhibition of NOS3 (NO synthase type-3, eNOS), its distribution becomes diffuse throughout the cytoplasm,³¹ and this alters its physiological role. PDE5A signaling in cardiomyocytes primarily hydrolyzes cGMP derived from NO-dependent signaling, as myocytes and hearts lacking NOS3 or treated with NOS inhibitors such as NG-Nitro-L-arginine methyl ester (L-NAME) lose this regulation.³² Normal NOS3 activity is required for PDE5A localization to Z disks in a cGMP-dependent manner.

The impact of PDE5A inhibition on myocardial function and various signaling pathways is summarized in Figure 3.^11,30 PDE5A regulates adrenergic-stimulated myocardial contractility in mammals including humans.^1,11 In hypertrophic and HF syndromes, PDE5A inhibition improves cardiac function and suppresses maladaptive remodeling.⁵ PDE5A inhibition also enhances mitochondrial function in diabetic cardiomyopathy through NO-induced Sirt1 (Sirtuin-1)-PGC1α (peroxisome proliferator-activated receptor-gamma coactivator 1α) signaling, indicating roles in energy metabolism and cell survival.³³ Its activity is regulated allosterically by cGMP binding to an N-terminal GAF-A domain that stimulates cGMP hydrolysis in the catalytic region and favors further PDE5A activation via PKG phosphorylation at S92. PDE5A inhibitors, such as sildenafil, bind to the catalytic site with higher affinity than cGMP. This results in increased cGMP and PKG levels, turning a negative feedback loop into a positive one (Figure 1).

Figure 3.

Structural and functional dynamics of NOS (NO synthase) in NO-mediated signaling pathways. NOS is a heterodimer with an N-terminal oxygenase domain and a C-terminal reductase domain, separated by a CaM (calmodulin)-binding domain. Its activity is regulated by phosphorylation from Akt and other kinases and enhanced by CaM binding. The enzyme transfers electron from nicotinamide adenine dinucleotide phosphate (NADH) to flavin adenine dinucleotide (FAD), then to flavin mononucleotide (FMN), and finally to heme iron, reducing Fe3⁺ to Fe2⁺. O₂ binds to Fe2⁺ and reacts with L-arginine to generate NO and citrulline. Tetrahydrobiopterin (BH4) is crucial for enzyme function and likely facilitates electron transfer. NO availability is regulated by CYB5R3 (cytochrome b5 reductase 3), which reduces Hb (hemoglobin) from Fe3⁺ to Fe2⁺. NO activates PKG (protein kinase G), initiating downstream signaling pathways relevant to heart failure (HF) and cardiac function. AKAPs indicates A-kinase anchoring proteins; ANP, A-type NPs; AR, androgen receptor; BNP, B-type NPs; CAV-1, caveolin-1; CNCGs, cyclic nucleotide-gated channels; CNP, C-type NP; C-Src, cellular sarcoma; CYB5R3, cytochrome b5 reductase 3; eNOS, endothelial nitric oxide synthase; GC-1A, guanylate cyclase-1 alpha; GTP, guanosine triphosphate; LTCC, L-type calcium channel; mitoKATP, mitochondrial ATP-sensitive potassium channel; MLC, myosin light chain; MLCK, myosin light chain kinase; MLCP, myosin light chain phosphatase; mTOR, mechanistic target of rapamycin; MyBPC, myosin binding protein C; NFAT, nuclear factor of activated T-cells; PDE5A, phosphodiesterase 5A; PI3K, phosphoinositide 3-kinase; PLB/PLN, phospholamban; RGS2/4, regulator of G-protein signalling 2 and 4; SERCA, sarcoplasmic/endoplasmic reticulum calcium ATPase; Tnl, troponin I; TRPC1/3/6, transient receptor potential canonical channels 1, 3, and 6; and TSC2, tuberous sclerosis complex 2.

The major cardiovascular impact of enhanced cGMP upon PDE5A inhibition is the activation of PKG1a (Figure 3). In vascular smooth muscle, PKG1a modifies inositol trisphosphate receptor–associated cGMP-kinase substrate (IRAG) inhibiting calcium release from the sarcoplasmic and endoplasmic reticulum. It also interacts with myosin light chain phosphatase (MLCP) and the GTP-binding protein RhoA (Ras Homolog Family Member A), all reducing vascular muscle tone.³⁰ PKG1a also phosphorylates transient receptor potential canonical channels, TRPC3 and TRPC6, inhibiting their calcium influx and for TRPC6 modulating cardiac mechano sensing. For both channels, phosphorylation in the N terminus by PKG1a reduces calcium conductance and stress-induced changes in hypertrophy, fibrosis, and arrhythmia.³⁴ In cardiomyocytes and fibroblasts, PDE5A inhibition also stimulates PKG1a phosphorylation of tuberin (TSC2 [tuberous sclerosis complex 2]), enhancing its inhibition of the mTOR (mechanistic target of rapamycin complex-1). This reduces pathological hypertrophy, improves function, and enhances autophagy in mice with cardiac pressure overload.³⁵ PKG1a also phosphorylates and binds to regulators of G-protein–coupled signaling, RGS2, and RGS4, both GTPases that suppress Gαq/11 signaling to inhibit pathological hypertrophy.¹ The CHIP (carboxyl-terminus of HSC70 [Heat Shock Cognate 70 kDa Protein]-interacting protein) is also phosphorylated by PKG, stabilizing its protein levels promoting clearance of misfolded proteins, and reducing postinfarction proteotoxicity.³ In sarcomeres, PKG phosphorylates titin, reducing stiffness and enhancing diastolic compliance,³⁶ and MyBPC (myosin binding protein C) to enhance sarcomere kinetics and adrenergic-stimulated contraction.

Despite these many intriguing myocardial effects, studies of PDE5A inhibitors in various human heart diseases, including HF with preserved (HFpEF) and reduced (HFrEF) ejection fraction, have yet to show benefit.³⁷ PDE5A is expressed in the human heart and myocytes,³⁰ so this lack of impact may be attributed to species-dependent variations in localization and subcellular signaling dynamics.

PDE9A

PDE9A has the highest specificity for cGMP (Km of ≈170 nmol/L) among all PDEs (Figure 1). It differs from PDE5A by specifically hydrolyzing cGMP generated by NP coupled signaling in cardiomyocytes³⁸ and neurons (Figure 4).³⁹ PDE9A has 21 splice variants, underscoring a complex regulatory role across cellular subdomains. In cardiomyocytes, PDE9A localizes to T-tubular regions and mitochondria, supporting a role in regulating signaling within the dyadic cleft^4,11,38 (Figure 2).

Figure 4.

Mechanisms of natriuretic peptide (NP) receptor signaling and regulation in the cardiovascular system. A-type NPs (ANP), B-type NPs (BNP), and urodilatin activate GC-A (guanylyl cyclase-receptor A), while C-type NP (CNP) activates GC-B (guanylyl cyclase-receptor B). GC-A signaling is compartmentalized near T-tubules, inhibiting LTCC and TRPC3/6, to reduce calcium influx. PDE2 regulates this pathway, limiting phospolamban (PLB) phosphorylation with minimal impact on contractility. GC-B signaling has broader effects, phosphorylating PLB and TnI, resulting in a lusitropic response. PDE3 regulates GC-B–mediated cGMP hydrolysis, while PDE2 regulates both GC-A and GC-B pathways. PDE9A modulates NP-mediated cGMP, affecting mitochondrial function. GC-C regulates adenylate cyclase maintaining cardiovascular homeostasis by controlling NP levels. Akt indicates protein kinase B; B-AR, beta-adrenergic receptor; CaM indicates calmodulin; GC, guanylate cyclase; GTP, guanosine triphosphate; LTCC, L-type calcium channel; NPR-A, B and C, natriuretic peptide receptor-A, B and C; PDEs, phosphodiesterases; PDE3/PDE9A/PDE3B/PDE2A, phosphodiesterase 3, 9A, 3B, and 2A; PKG, protein kinase G; RyR, ryanodine receptor; SERCA, sarcoplasmic/endoplasmic reticulum calcium ATPase; and TRPC, transient receptor potential canonical channel.

Inhibition of PDE9A in mice with a genetic knockout or a pharmacological inhibitor that are subjected to chronic pressure overload show less pathological hypertrophy, fibrosis, and enhanced cardiac function.³⁸ Oral administration of a clinical PDE9A inhibitor (PF-04447943, 40 mg/kg per day; Ki [inhibition constant] = 2.8 nmol/L for human PDE9A versus 5.3–99 µmol/L for other PDEs) improves cardiac function and also reduces obesity, liver steatosis, and enhances fat browning in male and ovariectomized female mice fed a high-fat diet combined with pressure overload.⁴ These metabolic changes are driven by the upregulation of PPARα (peroxisome proliferator-activated receptor α), which enhances the transcription of fatty acid metabolic genes (eg, Cpt1a, Acot1, Acadm, and Hadh).⁴ Interestingly, this impact of PDE9A inhibition displays a marked sexual dimorphism being negligible in intact females. This appears related to estrogen’s overlapping control over fatty acid metabolic genes via ER (estrogen receptor) activation, essentially competing with PPARα-mediated gene control and, so reducing the impact of PDE9A inhibition.⁴ Unlike PDE5A that inhibits vasodilation in the lung and corpus cavernosum, PDE9A does not significantly affect vasodilation in humans. Some studies in sheep and mice report an impact on arterial tone, but others show no impact.^4,38,40,41

The selectivity of PDE9A for NP-stimulated cGMP suggests potential added value if both methods are combined, enhancing NP signaling while inhibiting PDE9A. This is supported in an ovine model of acute cardiac failure (short-term tachypacing), whether NP is enhanced by infusion or suppressing its proteolysis by neprilysin.⁴² This synergy may be particularly effective in conditions characterized by elevated cardiac stress that increases NP secretion, such as dilated HF, HFpEF, or morbid obesity. PDE9A expression is also increased in both forms of HF,³⁸ supporting greater impact from its inhibition in these conditions. Many patients with HF currently receive the NEP inhibitor sacubitril combined with the angiotensin receptor blocker valsartan. Adding a PDE9A inhibitor may further enhance their benefits without causing hypotension.

PDE10A

PDE10A regulates both cAMP and cGMP, with higher affinity but a lower Vmax for cAMP. The PDE10A2 isoform is most prevalent in the heart, and expressed at low levels in normal and exercised hearts, but increases in pathologically stressed hearts.⁴³ This isoform features dual GAF domains and a PKA phosphorylation site (Figure 1), which adapt to changes in cyclic nucleotide levels, influencing intracellular signaling dynamics.¹¹ The enzyme’s activity is regulated by cAMP levels: high cAMP concentrations inhibit cGMP hydrolysis by binding to the catalytic domain, whereas low cAMP levels activate PDE10A via the GAF domain, making it a cAMP-inhibited cGMP PDE.¹¹ In this respect, PDE10A functions as the counterpart to PDE2A. Studies using the selective PDE10A inhibitor TP-10 show a diminution of pathological hypertrophy, fibrosis, and improved heart function in mice with chronic neurohormonal stimulation or pressure overload.⁴³ PDE10A has also been shown to contribute to doxorubicin-mediated cardiotoxicity, by increasing topoisomerase-2β expression, mitochondrial dysfunction, and DNA damage by antagonizing cGMP/PKG signaling. It also contributes to cardiomyocyte atrophy by potentiating forkhead box O3 signaling by cAMP/PKA and cGMP/PKG-dependent signaling.⁴⁴ Importantly, in a model of ovarian cancer, PDE10A inhibition with TP-10 treatment ameliorated the adverse effects of doxorubicin on the heart without interfering with its ability to attenuate tumor growth.⁴⁴

cGMP and cAMP Cross Talk via PDE Modulation

An important aspect of cyclic nucleotide regulation is the ability of signaling associated with one nucleotide to influence the other (Figure 2), a balance that can be disrupted in HF syndromes. The strategic localization of PDEs and the ability of several isoforms to modulate both cyclic nucleotides based on their relative levels, binding to regulatory domains, or activation of their primary kinases, provides a key mechanism for this cross talk. For example, PDE2A is allosterically activated by cGMP to enhance cAMP hydrolysis near the L-type calcium channel (LTCC), blunting calcium influx and myocardial contraction in response to stimuli.¹¹ The opposite is achieved by cGMP binding to the PDE3A catalytic site, competing for cAMP binding to increase net cAMP.¹¹ The cGMP concentration is critical for this modulation: low levels (0.1–10 nmol/L) inhibit PDE3A cAMP hydrolysis and enhance contractility, while higher levels (100 nmol/L–1 µmol/L) activate PDE2A to degrade cAMP to dampen β-adrenergic receptor (AR)–stimulated contractility.⁴⁵

Altered cross talk between cyclic nucleotides also results from the displacement of proteins responsible for their generation and hydrolysis. For example, in HF, β3-AR delocalization and reduced PDE2A function disrupt cAMP feedback signaling, worsening pathological remodeling that would otherwise be countered by NO stimulation. Local PDE2A and PDE3A redistribution causes β3-ARs to move from T-tubules (transverse tubule) to the sarcolemma surface, impairing cAMP regulation to augment β-adrenergic stimulation. PDE3A regulates cAMP degradation, contributing more at low cAMP (≈0.1 µmol/L) without Ca²⁺ and less at high cAMP levels with Ca²⁺ present, while cGMP inhibits both PDE3A and PDE1, further complicating the regulatory network.¹³

For cAMP signaling, AKAP proteins also play a role in cyclic nucleotide cross talk by forming local signalosomes (Figure 2).⁶ For instance, AKAP15/18δ anchors PKA to phospholamban at the sarcoplasmic reticulum, enhancing calcium uptake and contraction/relaxation during adrenergic stimulation. Targeting AKAP15/18δ or associated PDEs with peptide disruptors selectively enhances this signaling while minimizing side effects.^46,47 While a parallel GKAP for cGMP/PKG signaling remains undiscovered and likely does not exist, evidence still supports localized signaling and its alteration by NO availability and oxidative stress.⁴⁸ For example, PKG1a oxidation occurs between cysteine 42 in its 2 homo-monomers, altering intramyocyte localization of the kinase and its response to stress. At rest, cardiomyocytes display a diffuse distribution of PKG1a, but upon stimulation with a Gaq agonist such as endothelin-1, or acute pressure overload, the kinase is translocated to the plasma membrane.⁴⁹ If PKG1a oxidation is prevented by expressing it as a C42S mutant, localization to the outer membrane is sustained while chronic oxidized PKG1a again takes on a cytosolic distribution. This impacts consequent responses to PDE5A inhibition, which counters chronic pressure overload if PKG1a is oxidized but not if the C42S redox dead mutant is expressed.⁵⁰ Other targets of PKG1a such as TSC2 are also differently altered if the kinase can be oxidized or not, providing a mechanism coupling myocardial oxidative stress to pathological cardiac remodeling.⁵¹

Guanylate Cyclases and Cardiac Regulation

There are 2 groups of guanylate cyclases: GC-1/2, which are primarily activated by NO, and GC-A (guanylyl cyclase-receptor A) and GC-B (guanylyl cyclase-receptor B), activated by NPs. Each uniquely modulates cGMP pathways in the heart, sharing some downstream effects while selectively influencing others.

GC-1 (Figure 3) is found in the cytosol and plasma membrane domains and is composed of α and β subunits, each with 2 isoforms (α1/α2 and β1/β2), allowing for different heterodimer combinations across tissues. In cardiomyocytes, membrane-localized GC-1 is mostly found in caveolin-enriched microdomains near T-tubules, enabling rapid cGMP production in response to NO, which in turn influences cardiac muscle function.^16,52 The cyclase’s core heme group binds NO generated primarily by NOS3 and some by NOS1. This catalyzes the conversion of GTP to cGMP. In cardiomyocytes, NOS3 is localized to caveolae, whereas NOS1 is found at the SR and mitochondria. NO synthesis by NOS3 is linked to myocyte stretch enhancing cardiac response to mechanical load. NOS2 activation is associated with inflammatory cytokines and is a major source of nitrosative stress—but less so cGMP/PKG signaling. GC-1 contains a central heme that must be in the reduced ferrous state for proper NO binding and cGMP synthesis. Ferrous heme is inserted into the apo-sGCβ subunit, facilitated by GAPDH and heat shock protein (Hsp90). GAPDH transfers heme to the Hsp90–apo-sGCβ complex for ATP-dependent insertion. Once inserted, Hsp90 and GAPDH dissociate, allowing sGCβ to pair with sGCα to form an active sGC heterodimer essential for the NO response and cGMP production.⁵³

In cardiac and endothelial cells, ERs activate NOS3 via PI3K (phosphatidylinositol 3-kinase)-Akt phosphorylation at S1179, stimulating NO synthesis. Reduced estradiol levels postmenopause diminish this signaling and this is associated with a decline in the cardiovascular protection observed in younger women.⁵⁴ Variations in the NO-GC and cGMP-dependent PKGI genes are linked to vascular structural changes and remodeling, demonstrating a direct genetic impact on vascular health.⁵⁵ Reciprocally, genetic predisposition to enhanced NO/cGMP signaling is associated with a lower risk of coronary heart disease, peripheral arterial disease, and stroke.⁵⁶ In hypertension, pressure overload, and HF, oxidation of the heme in GC-1 occurs impairing NO binding and cGMP synthesis. Such stress is linked to the dislocation of GC-1 from caveolin-enriched plasma membrane environments, which is associated with depressed cGMP generation by GC-1.⁵⁷

Receptor Guanylyl Cyclase

Receptor guanylyl cyclases GC-A and GC-B localize to the plasma membrane, where they transduce signaling from NPs into cGMP synthesis. GC-A is activated by A-type and B-type NPs (ANP, BNP) via the NPR-A/NPR1 (natriuretic peptide A/1) receptor, whereas GC-B is activated by C-type NP via NPR-B/NPR2. A third receptor, NPR-C, is primarily expressed on endothelial cells, where it clears NPs and is also coupled to some G-protein signaling functions.⁵⁸ Both GC-A and GC-B form complexes with their respective receptors, residing in the inner plasmalemmal membrane (Figure 4).

There are a total of 7 pGCs (particulate guanylate cyclases) (GC-A to GC-G), but only GC-A and GC-B are thought to play major roles in the cardiovascular system, enhancing vasodilation, myocardial energetics and fuel metabolism, contractility, and remodeling.⁵⁸

GC-A

GC-A plays a central role in natriuresis, helping regulate blood volume and pressure and buffering cardiac, renal, and vascular diseases. It is predominantly expressed in kidneys, heart, vasculature, adipocytes, and the brain.⁵⁹ GC-A supports fluid balance and inhibits the renin-angiotensin-aldosterone system, promoting myocardial relaxation and reducing pathological remodeling. ANP activates GC-A mainly through cGMP-dependent pathways that suppress prohypertrophic and profibrotic pathways. Chronic stimulation, however, can lead to GC-A desensitization, reducing cGMP-dependent modulation of multiple downstream targets that in turn results in maladaptive cardiac remodeling to stress. Scanning ion conductance microscopy and Förster resonance energy transfer–based cGMP biosensors have localized GC-A to the T-tubules.⁶⁰ Its functionality is enhanced through allosteric sites, receptor phosphorylation by PKC (Protein Kinase C), and glycosylation at asparagine residues 2 and 13. GC-A regulates lipid and glucose metabolism, protecting against metabolic syndrome by promoting lipolysis and the browning of white adipocytes. It enhances mitochondrial oxidative metabolism and fat oxidation by upregulating genes for PGC1α and PPAR(α/δ),⁴ while increasing glucose uptake via activating Akt/mTOR signaling.

Genetic variations in GC-A and GC-B genes, such as single nucleotide polymorphism rs5068 and rs198389, have been linked to hypertension and cardiovascular disease risk.⁶¹ Although recombinant ANP and BNP have been clinically tested for acute HF management, their effectiveness has been limited by factors such as blood pressure reduction, short biological half-life, and degradation via NPR-C internalization. Novel synthetic NPs have been developed to address some of these limitations; while none have yet become clinical therapies, others are still in development.

GC-B

GC-B is activated by C-type NP and plays an important role in cardiovascular homeostasis by enhancing endothelial repair and initiating anti-inflammatory responses essential for vascular integrity. GC-A and GC-B exhibit antihypertrophic properties; GC-A stimulation has a minimal effect on myocardial contraction, while GC-B influences it, likely due to differences in localization. Förster resonance energy transfer–based studies show that GC-B is uniformly distributed, enhancing lusitropic effects via PKG-mediated phosphorylation of phospholamban and troponin I (TnI), whereas GC-A remains localized in T-tubules.⁶⁰ Increased GC-B activity observed in HF models highlights its therapeutic potential for managing HF.⁶² The autocrine and paracrine functions of C-type NP via GC-B contribute to reducing fibroblast hyperplasia and cardiomyocyte hypertrophy and promoting endothelial recovery following damage.

cGMP Pathway Modulation for HF Therapeutics

Active clinical trials testing the impact of a myriad of cGMP-enhancing therapeutic strategies are listed in Table S1. Here, we highlight major efforts underway.

Therapies to Enhance cGMP Synthesis

Drugs targeting GC-1 include stimulators, that amplify NO-initiated signaling, and activators that directly enhance GC-1 activity independent of NO stimulation. The latter is theoretically advantageous if NO signaling is impaired for example due to NOS3 oxidation. NO donors such as nitroglycerin are widely used for acute coronary vasodilation, but their value in HF is less clear. Combining the nitrate generator isosorbide dinitrate with a vasodilator hydralazine improved HF survival in a subset of patients with HFrEF who were self-reported African-American although the exact reason this was effective only in this subgroup remains speculative.⁶³ Nonetheless, the combination remains recommended for class III to IV in this subgroup.

The GC-1 activator cinaciguat initially showed promise for treating HFrEF, but concerns emerged due to substantial and unpredictable hypotension.⁶⁴ GC-1 stimulators were a bit less potent and 1 such drug riociguat was beneficial against pulmonary arterial hypertension, while vericiguat reduced a combined end point of cardiovascular death or HF hospitalization in HFrEF.⁶⁵ In the latter study, the effect size was quite modest, particularly compared to benefits being reported around the same time from sodium-glucose transporter 2 inhibition.⁶⁶ In a subsequent study, vericiguat did not benefit HFpEF, and its present use in HF remains limited. Other efforts to enhance GC-1 activity via organic or inorganic nitrates/nitrites have failed to show benefit.

NP Agonism

A major challenge to the therapeutic use of recombinant NP agonists is delivering a peptide with a pharmacokinetics profile that provides sustained efficacy. Various peptides have been generated that interact with NPR-A or NPR-B receptors and are resistant to peptidases such as neprilysin, allowing subcutaneous administration every 1 or ≥2 weeks. Cenderitide is an example and is being developed to target multiple NP receptors.⁶⁷ However, pharmacological concerns and potential side effects, such as hypotension, remain obstacles. Peptides including ANX042, MANP, CRRL269, NPA7, and C53 are being tailored for specific HF subtypes and cardiorenal syndromes, reflecting a shift toward precision medicine.⁶⁴

Another approach is to inhibit the peptidases that degrade NPs. Sacubitril is currently in wide clinical use in combination with valsartan (angiotensin receptor and neprilysin inhibitor). This has been particularly effective in HFrEF but less so in patients with EF above 40%.

Convergence of HFpEF and Women’s Health

HFpEF currently impacts over 50% of all HF, and its prevalence rising particularly with the global pandemic of obesity and cardiometabolic disease. This syndrome presents substantial challenges due to its multifactorial nature. Historically, HFpEF was predominantly found in women, and while this has shifted to younger ages and more males, women still develop HFpEF before menopause. HFpEF has also become increasingly linked to obesity, and there is heightened interest in developing therapies that counteract maladaptive cardiac remodeling and metabolic defects that dominate. Beyond NP mimetics and angiotensin receptor and neprilysin inhibitors, augmenting NP-stimulated cGMP by inhibiting PDE9A presents a promising strategy.⁴ PDE9A inhibition may help counter cardiometabolic syndrome, including fatty liver disease, obesity, and cardiac disease in both males and postmenopausal females, underscoring a potential value in HFpEF. Ongoing phase II clinical trials are expected to provide initial proof of concept for the effectiveness of this approach.

PDE Inhibitors

The PDE5A inhibitors sildenafil and tadalafil are widely used to treat erectile dysfunction and pulmonary hypertension. Their efficacy can be limited due to inadequate NO-dependent cGMP bioavailability.⁶⁴ Several ongoing trials are testing their efficacy in HF and diabetic HF combined with pulmonary hypertension. Dipyridamole, which inhibits PDE4A, PDE5A, and PDE10, is used for thromboembolic syndromes. The PDE3 inhibitor milrinone is currently used for acute decompensated HF, and cilostazol is used to treat microvascular disease in diabetic patients.^64,68 Pentoxifylline, a xanthine derivative that targets multiple PDEs, is presently used for vascular disease to alleviate claudication. The PDE1 inhibitor lenrispodun is now in phase II clinical trials for Parkinson disease, and while prior studies suggested potential utility for vascular dilation and improved cardiac function in patients with HF,⁶⁹ this indication is not currently being pursued. A multicenter multinational phase IIB trial of PDE9A inhibition (CRD-740) in patients with HFrEF and HFpEF is ongoing. The trial was not designed to assess the effects of this drug on obesity, but the inclusion of subjects with HFpEF particularly in the United States where obesity is a common comorbidity may provide some insight into its impact.

Conclusions

Despite many roles of cGMP signaling in the cardiovascular system that would seem beneficial to counter various cardiac diseases, applications of therapies to leverage the pathway have been met with challenges and disappointments. One primary limitation is the potent effect of cGMP-PKG activation on arterial smooth muscle, which often leads to systemic hypotension. This can dominate and ultimately restrict drug exposure levels that could in turn prevent effective stimulation of various intracellular pathways in desired cells such as myocytes. This is especially true for drugs that increase cGMP synthesis, as the enzymes responsible for this (GC-1, GC-A, and GC-B) are highly expressed in arterial smooth muscle. The widespread expression of PDEs across different tissues also raises concerns regarding unintended physiological effects. Nonselective PDE inhibitors can disrupt critical signaling pathways in the heart, vascular smooth muscle, and platelets, potentially causing adverse effects like increased blood pressure and a heightened risk of arrhythmias. However, PDEs also exhibit substantial cell specificity, which enables more targeted modulation of cGMP signaling in specific cell types. For instance, PDE9A is minimally expressed in arterial vessels, as evidenced by human studies showing no associated decrease in blood pressure. Last, a remaining challenge in using PDE inhibitors is that they have historically blocked the catalytic domain that is shared across intrafamily member isoforms. Yet different isoforms can have distinct substrate selectivities, subcellular localizations, and signaling, where inhibiting one but not another is desirable. This is stimulating efforts to generate allosteric modulators that if successful would enable more selective PDE modulation to enable signaling that maximizes therapeutic benefits and minimizes associated risks.

Article Information

Sources of Funding

The study was supported by the American Heart Association CDA:938718 (S. Mishra) and R35HL-166565 (D.A. Kass).

Disclosures

None.

Supplementary Material

hyp-82-185-s001.docx ^{(45.7KB, docx)}