Molecular and Pathophysiological Links Between Heart Failure with Preserved Ejection Fraction and Type 2 Diabetes Mellitus

Patients with type 2 diabetes mellitus (T2DM) are at high risk for cardiovascular diseases. Nonetheless, T2DM is often unrecognized in heart failure (HF) patients. Patients with HF are divided into two major groups: those with HF and reduced ejection fraction (HFrEF) and those with HF and preserved ejection fraction (HFpEF). Additionally, recently, an in-between group has also been established (i.e. HF with mid-range ejection fraction). Collectively, the pathomechanisms for the spectrum of HF phenotypes are not only complex but also divergent, thus explaining the limited efficacy of classical HF therapies, in particular in HFpEF. To improve therapeutic efforts, proper phenotyping of HFpEF has been warranted. Nevertheless, no consensus on an optimal approach to this problem has been reached so far. Of note, T2DM frequently co-segregates with left ventricular (LV) diastolic dysfunction and HFpEF. Corresponding data from the CHARM trials revealed the prevalence of 60% for diabetes (40% known diabetes and 20% pre-diabetes) and 22% for undiagnosed diabetes among patients with HFpEF.1 While cardiovascular outcomes of novel T2DM treatments aiming at locations other than the heart (e.g. by using metformin or sodium–glucose co-transporter type 2 inhibitors) are still to be determined,2 here we wish to illuminate the myocardial nitric oxide–cyclic guanosine monophosphate–protein kinase G (NO-cGMP-PKG) pathway, and stress its potential pharmacological significance in T2DM and HFpEF. Mechanisms of diastolic dysfunction are complex and involve components intrinsic (e.g. defective myofilament characteristics and/or Ca2+ handling) and extrinsic to cardiomyocytes. Previous studies have furnished with compelling evidence for the contribution of the giant myofilament protein, titin, to the dynamic modulation of cardiomyocyte diastolic/passive stiffness.3 For example, phosphorylation of titin at the N2B unique sequence by protein kinase A (PKA) or PKG reduces cardiomyocyte passive stiffness, implicating that a chronic phosphorylation deficit in HFpEF can be linked with impaired PKA and/or PKG signalling. PKG-mediated phosphorylation of titin resulted not only in reduced titin-based myocardial stiffness in vitro, but following short-term cGMP-enhancing treatment with sildenafil and B-type natriuretic peptide (BNP) in a large animal model of HFpEF in vivo.4 Among the potential cardiomyocyte extrinsic mechanisms, earlier findings had emphasized that extracellular matrix alterations make also a significant contribution to the passive stiffness of cardiac muscle during HF.3 Recent evidence suggests that co-morbidities common to HFpEF promote a systemic inflammatory state evoking endothelial dysfunction, reactive oxygen species production, nitrosative stress, depressed NO bioavailability, and thereby reduced cGMP levels and PKG activity.5 We have shown that titin phosphorylation by cGMP and PKG lowers cardiomyocyte stiffness and that this process is disturbed in HFpEF.6 Another important concept in the pathogenesis of HFpEF is that oxidative stress and inflammation (resulting from the uncoupling of NO synthase) leads to LV hypertrophy, LV fibrosis, and hence LV diastolic dysfunction.5 In accordance with our previous findings, co-morbidities raise blood levels of proinflammatory cytokines and induce cardiac microvascular inflammation. The latter alters paracrine signalling from microvascular endothelial cells to adjacent cardiomyocytes and fibroblasts: cardiomyocytes become stiff and hypertrophied due to low cGMP, and fibroblasts differentiate into myofibroblasts under the influence of transforming growth factor β secreted by infiltrating macrophages.7 Myofibroblasts augment interstitial collagen deposition, which together with the presence of stiff cardiomyocytes, induce high diastolic LV stiffness, the main determinant of impaired exercise tolerance and symptoms in HFpEF. Levels of key transcription factors and sarcomeric proteins involved in hypertrophy signalling, diastolic relaxation, stiffness and vasorelaxation could be normalized by PKG-dependent phosphorylation, suggesting that PKG-activating agents may be indeed beneficial in HFpEF.8 Collectively, it is instructive to postulate that more than a single cGMP pool affects cardiomyocyte stiffness in the heart, and that the relative importance of different cGMP pools for HFpEF depends on co-morbidities. Therefore, targeting certain cGMP pools may be a viable therapeutic approach in HFpEF, although several details of this approach remain still elusive. Diabetes mellitus complicated by diastolic dysfunction and HFpEF has been shown to be associated with myocardial inflammation.9 In this context, reactive oxygen species are known to oxidize and inactivate many biomolecules. For example, peroxynitrate oxidizes the soluble guanylate cyclase enzyme (sGC), thereby converting it to a dysfunctional heme free form of the enzyme that is not responsive to NO. In diabetic ZSF1 rats (a hybrid model of hypertension, obesity and T2DM), the development of HFpEF was observed in association with titin hypophosphorylation.10 Of note, an increased diastolic stiffness of the diabetic human hearts was also associated with increased cardiomyocyte resting tension.9 Taken together, experimental and clinical data suggest that modulation of titin-based stiffness via cGMP-enhancing therapy can be useful in correcting a pathologically elevated LV diastolic stiffness, one of the primary characteristics of HFpEF patients. Therefore, strategies aiming at the enhancement of intracellular PKG activity may bring benefit in HFpEF, in particular if that is associated with T2DM6, 11 (Figure 1). Long-term preventive application of the phosphodiesterase (PDE) 5A inhibitor vardenafil effectively prevents the development of HFpEF in a model of diabetic cardiomyopathy-associated HFpEF (ZDF rats) by (i) maintaining diastolic function, (ii) restoring cGMP levels and PKG activation, (iii) lowering apoptosis, and (iv) alleviating nitro-oxidative stress, myocardial hypertrophy and fibrotic remodelling.12 The positive expectations adherent to PDE5 inhibition were further supported by the convincingly improved LV contractility of diabetic cardiomyopathy patients following chronic sildenafil administrations. Importantly, the cardiac benefit appeared to be independent of vasodilatory or endothelial effects, and hence it has been associated with a direct intramyocardial action.13 In view of the above findings, it is to be stressed that the neutral results of the subsequent RELAX trial were obtained in a patient population not selected for diabetes.14 Diabetes, in fact, was more common among HFpEF patients in the RELAX trial (43%) than in other comparable studies at that time. Although a RELAX trial ancillary study later reported that diabetic HFpEF patients had a more severe disease phenotype than those without diabetes,15 levels of NO-cGMP-PKG pathway-related biomarkers (e.g. sGC) were not different between diabetic and non-diabetic patients. Moreover, inequalities in the levels of PDE5 expression of left and right ventricular cardiomyocytes and glycosylation in the lung microvasculature in the presence and absence of diabetes mellitus might have also complicated this investigation. Taken together, despite the mixed outcomes of clinical trials, we are still encouraged to explore the clinical use of PDE5 inhibitors in HFpEF and T2DM.8, 14 Similarly to natriuretic peptides and NO, sGC stimulators and sGC activators increase cGMP synthesis and activate cGMP-dependent PKGs. Direct stimulators of sGC (e.g. riociguat)16 (similarly to the inhibitors of the cyclic nucleotide PDEs) reduce cardiomyocyte passive stiffness. In a rat model of type 1 diabetes mellitus, the sGC activator cinaciguat prevented the development of cardiac dysfunction. Interestingly, in this model cinaciguat also improved LV relaxation and decreased diastolic stiffness.17 Clinical data on sGC activators and stimulators so far provided limited conclusions for HFpEF and T2DM. Nevertheless, clinically important improvements were recently reported in the SOCRATES-PRESERVED trial that investigated the novel sGC stimulator vericiguat in HFpEF patients, among them 49% with diabetes.18 Not the two primary endpoints (changes in N-terminal pro-BNP and left atrial volume), but the quality of life scores were improved when vericiguat was compared to placebo in that cohort. The effects of LCZ696, a first-in-class angiotensin receptor–neprilysin inhibitor (valsartan + sacubitril) were tested in a phase 3 double-blind trial that involved 8442 HFrEF patients with class II, III, or IV HF (PARADIGM-HF study). In this study, LCZ696 was found be superior to enalapril in reducing the risks of death and hospitalization for HF.19 The first results with LCZ696 in HFpEF (PARAMOUNT phase 2 trial) were also encouraging.20 Whether the observed promising effects will translate into improved clinical outcomes in larger patient populations in HFpEF is currently being tested in an ongoing phase 3 trial (PARAGON-HF investigation).21 A number of myocardial alterations can lead to diastolic dysfunction during the pathogenesis of HFpEF. Results of preclinical and clinical studies point towards the involvement of a downregulated NO-cGMP-PKG pathway in the pathomechanism of diabetes-induced LV diastolic dysfunction. Accordingly, pharmacological stimulation of the above signalling cascade appears to be an attractive alternative to classic HF therapies. One may argue that clinical data are less convincing than the preclinical ones in the above context. Nevertheless, in contrast to preclinical investigations, HFpEF patients are rarely homogeneous for the presence or absence of diabetes mellitus in clinical studies, and hence the benefit of a mobilized NO-cGMP-PKG pathway can be limited by the complex pathomechanism in HFpEF patients with co-morbidities other than diabetes mellitus. An additional problem of clinical investigations involving modulators of the NO-cGMP-PKG system is related to the limited availability of data on LV diastolic function. Taken together, it is currently unknown which pharmacological approach will bring the most optimal benefit for HFpEF patients suffering from diabetes mellitus, and hence clinical investigations paying specific attention to hypothetical links between diastolic dysfunction and NO-cGMP-PKG signalling are warranted. This work was supported by grants from the European Commission (FP7-Health- 2010; MEDIA-261409), from the National Research, Development and Innovation Fund of Hungary (NVKP_16-1-2016-0017, National Heart Program, by the GINOP-2.3.2-15-2016-00043 project), and by DFG HA 7512/2-1. The project was co-financed by the European Union and the European Regional Development Fund. The research group of Zoltán Papp is supported by the Hungarian Academy of Sciences. Conflict of interest: none declared.