Research Priorities for Heart Failure With Preserved Ejection Fraction: National Heart, Lung, and Blood Institute Working Group Summary

Abstract

Heart failure with preserved ejection fraction (HFpEF), a major public health problem that is rising in prevalence, is associated with high morbidity and mortality and is considered to be the greatest unmet need in cardiovascular medicine today because of a general lack of effective treatments. To address this challenging syndrome, the National Heart, Lung, and Blood Institute convened a working group made up of experts in HFpEF and novel research methodologies to discuss research gaps and to prioritize research directions over the next decade. Here, we summarize the discussion of the working group, followed by key recommendations for future research priorities. There was uniform recognition that HFpEF is a highly integrated, multiorgan, systemic disorder requiring a multipronged investigative approach in both humans and animal models to improve understanding of mechanisms and treatment of HFpEF. It was recognized that advances in the understanding of basic mechanisms and the roles of inflammation, macrovascular and microvascular dysfunction, fibrosis, and tissue remodeling are needed and ideally would be obtained from (1) improved animal models, including large animal models, which incorporate the effects of aging and associated comorbid conditions; (2) repositories of deeply phenotyped physiological data and human tissue, made accessible to researchers to enhance collaboration and research advances; and (3) novel research methods that take advantage of computational advances and multiscale modeling for the analysis of complex, high-density data across multiple domains. The working group emphasized the need for interactions among basic, translational, clinical, and epidemiological scientists and across organ systems and cell types, leveraging different areas or research focus, and between research centers. A network of collaborative centers to accelerate basic, translational, and clinical research of pathobiological mechanisms and treatment strategies in HFpEF was discussed as an example of a strategy to advance research progress. This resource would facilitate comprehensive, deep phenotyping of a multicenter HFpEF patient cohort with standardized protocols and a robust biorepository. The research priorities outlined in this document are meant to stimulate scientific advances in HFpEF by providing a road map for future collaborative investigations among a diverse group of scientists across multiple domains.

Keywords: diagnosis; heart failure; precision medicine; therapeutics.

Figures

Figure 1.:

Proposed Molecular Mechanisms Underlying HFpEF (A) Comorbidites are common in HFpEF and are thought to lead to systemic inflammation which results in microvascular inflammation, widespread endothelial dysfunction (in multiple organs), and coronary microvascular dysfunction, leading to abnormal systolic and diastolic cardiac mechanics, and poor cardiac reserve. Systemic inflammation also leads to the activation of monocytes and macrophages, which release pro-fibrotic cytokines including IL-10 and TGF-beta, thereby promoting interstitial organ fibrosis, which in the heart increases passive myocardial stiffness. (B) Several factors promote a relative natriuretic peptide deficiency state in HFpEF, including obesity, sedentary lifestyle, African ancestry, insulin resistance, increased androgenicity in women, genetic variation in the NPPA and NPPB genes, and a lower amount of wall stress for the severity of heart failure (compared to HFrEF). (C) Natriuretic peptides are active in adipose tissue, where the relative ratio of the NPRA to NPRC receptors are important in dictating whether beneficial natriuretic peptide effects are possible. With increased NPRA, there is increased cGMP and PKG production, leading to lipolysis and the brown-fat thermogenic program. With increased NPRC, these beneficial effects are minimized, as there is increased natriuretic peptide breakdown. (D) Mechanical and metabolic stressors on the cardiomyocyte lead to T-tubule disruption and abnormal calcium handling within the cardiomyocyte, which leads to intracellular calcium overload and inefficient myocardial contraction and relaxation. (E) Natriuretic peptides act through a receptor guanylate cyclase pathway that results in the creation of cGMP and stimulation of PKG, which has a variety of beneficial effects in the heart and multiple other organs. There is also an intracellular, soluble guanylate cyclase, that is stimulated by nitric oxide, which also leads to increased cGMP and activation of PKG. PDE5 results in the breakdown of the NO-based cGMP pool whereas PDE9 results in the breakdown of the natriuretic peptide-based cGMP pool. (F) Multiple mechanisms present in HFpEF can result in stiffening of titin, the major molecular spring within the cardiomyocyte, thereby leading to increased cardiomyocyte (and subsequently cardiac chamber) passive stiffness. Because of insufficient natriuretic peptides and nitric oxide, PKG is reduced in HFpEF, which leads to hypophosphorylation of key sites within titin and increases its stiffness. ERK-2 (stimulated by increased cardiomyocyte stretch), PKA (stimulated by sympathetic stimulation), CaMKII (stimulated by reactive oxygen species), and PKCα (stimulated by endothelin-1 and angiotensin-II) all can have deleterious pro-stiffening effects on titin. (G) While endothelium-derived NO is reduced in HFpEF, inducible NO synthase (iNOS), which is activated by systemic inflammation, is upregulated and could be a pathogenic factor leading to HFpEF. In a recent study that utilized a novel 2-hit mouse model of HFpEF (Nω-nitro-L-arginine methyl ester [L-NAME, which induces hypertension] + high fat diet [obesity]), iNOS was upregulated, which resulted in S-nitrosylation (nitrosative stress) of the endonuclease inositol-requiring protein 1α (IRE1α), leading to defective splicing of an unfolded protein response effector (the spliced form of X-box-binding protein 1 [XBP1s]). XBP1s, in turn, was reduced in both the rodent HFpEF model and also in myocardial samples from patients with HFpEF, leading to increased levels of unfolded proteins within the cardiomyocytes, which are thought to interfere in normal cardiomyocyte function.

Figure 2.:

Targeted Therapeutics in HFpEF: The Example of Transthyretin Cardiac Amyloidosis Top panel: The ultimate goal for a heterogeneous syndrome such as HFpEF is to sub-phenotype patients using methods such as cardiac or other organ biopsy, blood- or urine-based analyses, imaging, exercise physiology, or machine learning techniques in order to determine the underlying biological mechanism(s) of disease to be able to develop effective phenotype-based treatments. Middle panel: A potentially ideal situation would entail first performing endomyocardial biopsies on patients to determine the etiology of myocardial disease via histological and molecular analysis, after which non-invasive imaging and biomarkers could be developed for less invasive diagnosis. Next, based on the biology of the specific HFpEF subtype, specific, targeted therapeutics could be developed, which ultimately could lead to positive clinical trials with disease-modifying therapies. Bottom panel: Although this situation may seem unlikely to achieve, we have seen the successful application of such an approach to transthyretin cardiac amyloidosis. Diagnosis was first discovered and recognized by endomyocardial biopsy with histology and fiber typing on mass spectrometry. Subsequently, unique patterns on echocardiography (speckle-tracking imaging), cardiac magnetic resonance imaging (inability to null the myocardial with late gadolinium enhancement and severely elevated extracellular volume content on T1 mapping), and bone scintigraphy (with increased heart-to-contralateral lung ratio) allowed for the non-invasive diagnosis of the disease. Investigation of the underlying biology of TTR misfolding, particularly the discovery of a protective mutation (T119M) in carriers of the V30M genetic TTR mutation, led to the development of TTR stabilizers, which have now been found to improve outcomes in a phase 3 randomized controlled trial. The history of TTR cardiac amyloidosis (which can be viewed as a subtype of HFpEF that was previously rarely recognized) lends support to the concept of precision medicine for HFpEF.

Figure 3.:

Summary of the Barriers and Solutions for Successful Translation of Therapeutics for the HFpEF Syndrome HFpEF is the culmination of several risk factors (including age, comorbidities [particularly obesity], physical inactivity, frailty, and environmental stressors) that lead to multi-organ reserve dysfunction (i.e., inability to adequately tolerate stress) which involves not only the heart but also the lungs, aorta, microvasculature, kidneys, gastrointestinal tract, adipose tissue, liver, and skeletal muscle. The complexity of the risk factors, etiologies, pathophysiologies, and multi-organ involvement gives rise to heterogeneous HFpEF sub-phenotypes. Thus far, there has been a lack of translation, and the majority of clinical trials have not led to improvements in symptoms our reduction in adverse outcomes due to several potential barriers, each of which has potential solutions as listed in the figure.

Figure 4.:

Framework for a Proposed Translational Science HFpEF Network A key recommendation of the NHLBI HFpEF Working Group is the formation of a translational science HFpEF network among collaborative research centers. Each center would be composed of both a clinical and basic science component working in collaboration with each other on a specific HFpEF phenotype or mechanism. Regional referral centers would be identified which would refer HFpEF patients to the main center. Automated identification of patients, which would supplement routine patient enrollment practices, would be achieved through natural language processing of the electronic health record and deep learning and computer vision techniques of cardiac imaging and electrocardiograms. A “T” type dataset with broad enrollment of HFpEF patients and control patients with comorbidities would be formed such that a subset of participants would undergo deep phenotyping. Subsequently, all participants would undergo serial blood and urine sampling and imaging, with EHR-based data capture to ascertain outcomes and track the patient journeys using a shared data platform across all sites. Importantly, to increase scientific rigor and reproducibility, each center in the network would serve as a replication/validation center for basic and clinical studies at another center, and there would be crosstalk among all centers in the network to engage in scientific discussions and shared biorepositories.