Phosphodiesterase-5 Inhibition in Heart Failure with Preserved Ejection Fraction: Trading Therapy for Prevention

This article refers to ‘Prevention of the development of heart failure with preserved ejection fraction by the phosphodiesterase-5A inhibitor vardenafil in rats with type 2 diabetes,’ by C. Mátyás et al., published in this issue on pages 326–336. Heart failure with preserved ejection fraction (HFpEF) is rapidly becoming a major burden to healthcare systems in Western societies, with a 1% annual rise in prevalence relative to heart failure with reduced ejection fraction (HFrEF) and, despite numerous efforts, no established therapeutic intervention with prognostic benefit are available.1 Although a myriad of mechanisms contribute to low exercise tolerance in HFpEF, diastolic LV dysfunction remains pivotal to the pathophysiology of HFpEF and reflects both cardiomyocyte stiffness and myocardial fibrosis.2 Correction of diastolic LV dysfunction remains a compelling challenge in modern cardiovascular medicine and was addressed in the study by Mátyás et al.3 published in the current issue of the journal. In 2013, a new theoretical framework was proposed for HFpEF, based on a continuum of evidence in experimental preparations and human LV biopsies. This new paradigm for HFpEF postulated a central role for coronary microvascular endothelial dysfunction, and inflammation as a result of multiple mainly metabolic co-morbidities. Microvascular endothelial inflammation is presumed to trigger an unfavourable molecular cascade ultimately leading to diastolic LV dysfunction because of stiff and hypertrophied cardiomyocytes and interstitial collagen deposition.4 The major player in this cascade was identified to be loss of nitric oxide (NO) bioavailability, which resulted in cardiomyocyte deprivation of cyclic guanosine monophosphate (cGMP) and lower activity of protein kinase G (PKG).5 The new paradigm triggered a tremendous interest in the association between HFpEF and the NO–cGMP–PKG pathway as evident from a PubMed search using HFpEF and NO–cGMP–PKG pathway as key words, which revealed >100 published articles since the launch of the new paradigm. Three primary mechanisms may impair cGMP signalling in HFpEF myocardium: (i) enhanced degradation through up-regulation of cardiac phosphodiesterases (PDEs), especially PDE5; (ii) decreased NO stimulation and/or responsiveness of soluble guanylate cyclase; and (iii) reduced synthesis via impaired natriuretic peptide activation of transmembrane particulate guanylate cyclase receptors.6 PDE5 is the isoenzyme that breaks down cGMP into its inactive 5'-GMP form.7 In the last decade, in vitro8 and in vivo9, 10 studies have claimed that PDE5 inhibition by sildenafil effectively protects against unfavourable LV remodelling by targeting inhibition of hypertrophic and profibrotic stimuli and modulation of cardiomyocyte stiffness (Figure 1). However, overexpression of PDE5, the key substrate for efficacy of PDE5 inhibitors, has not been an unequivocal finding.5, 11 In the present issue of the journal, Mátyás et al.3 report on the myocardial effects of almost lifelong (from week 7 to week 32) administration of the PDE5 inhibitor vardenafil in Zucker diabetic fatty (ZDF) rats, an established model of metabolically induced HFpEF. Vardenafil is more cGMP specific than sildenafil, the most widely used PDE5 inhibitor, and is less likely to interfere with the PDE–cAMP axis. The authors confirmed diastolic LV dysfunction to be mediated by myocardial microvascular inflammation and convincingly demonstrated deranged molecular pathways involved in the NO–cGMP–PKG axis such as overexpression of PDE5. Vardenafil improved both active LV relaxation and diastolic LV stiffness. The latter was achieved by lower cardiomyocyte passive tension and by reduced interstitial fibrosis. Vardenafil not only enhanced PKG activity as evident from an increased p-VASP/VASP (vasodilator-stimulated phosphoprotein) ratio, but also affected endothelial inflammation, nitro-oxidative stress, cardiomyocyte hypertrophy, and apoptosis.12 Vardenafil induced these late-life, ‘pleiotropic’ beneficial effects through preventive actions in early life when the ZDF rats were in a pre-diabetic stage. The positive findings of Mátyás et al.3 clearly contrast with the neutral outcome of some previous clinical attempts to treat HFpEF patients with sildenafil, and suggest myocardial responsiveness to PDE5 inhibition to differ between early pre-clinical risk factor exposure and late clinically manifest HFpEF. A time-dependent difference in myocardial responsiveness to sildenafil is also supported by other expertimental and clinical observations. In the ZSF1 rat (a cross between a ZDF and a spontaneously hypertensive rat), treatment with sildenafil from week 16 to 20 improved diastolic LV dysfunction, albeit less than in the study of Mátyás et al.,3 probably because of the later onset of treatment.13 In the ZSF1 rat, myocardial fibrosis was absent at 20 weeks of age and diastolic LV dysfunction could be attributed to hypophosphorylation of titin at PKG-specific sites,14 an ideal target for PDE5 inhibition. Myocardial responsiveness to PDE5 inhibition in the incipient stages of concentric LV remodelling and HFpEF was also reported in clinical studies. In a placebo-controlled study of 59 men with early features of diabetic cardiomyopathy, i.e. concentric LV hypertrophy, preserved LV EF but impaired MRI measures of LV torsion and strain, 3 months of sildenafil treatment reversed LV remodeling, evident from a significant improvement of the LV mass to volume ratio.15 Similarly, in hypertensive patients with recent-onset dyspnoea, 1 year treatment with sildenafil improved diastolic LV distensibility, evident from lower LV filling pressures at larger LV internal dimension.16 The successful outcome of PDE5 inhibition in early concentric LV remodelling or early HFpEF contrasts sharply with the neutral outcome of the RELAX trial,17 which was unfortunately planned in the ‘pre-phenotyping’ era of HFpEF and probably addressed a population of advanced HFpEF patients presenting with a high rate of co-morbidities such as anaemia and COPD, by definition unresponsive to a NO-overexperessing intervention. The advanced stage of HFpEF was probably evident from the mean plasma NT-proBNP level, which equalled 700 pg/mL, a value 10 times higher than in the aforementioned early-stage diabetic cardiomyopathy patients.15 The study by Mátyás et al.3 ‘resuscitates’ use of PDE5 inhibitors for the prevention of HFpEF. In line with the recent emphasis on phenotype-specific treatment of HFpEF,2 their study supports resumption of clinical investigations on the use of PDE5 inhibitors in early pre-clinical metabolic risk factor exposure and in early HFpEF. If these investigations live up to the expectations, the course of PDE5 inhibition in HFpEF would fit the quote of Dante Alighieri: ‘The path to paradise begins in hell’. Conflict of interest: none declared.