Why don’t we have proven treatments for HFpEF?

Summary

The lack of effective treatments for heart failure with preserved ejection fraction (HFpEF) represents a large and growing unmet need in cardiology today. A critical obstacle to therapeutic innovation in HFpEF has been the absence of animal models that accurately recapitulate the complexities of the human disease. Here we propose that more comprehensive multi-organ system and functional phenotyping of preclinical models is essential if we are to maximize our chances of discovering and validating novel targets for effective therapeutic development in HFpEF.

Viewpoint

At this time in human history, we are witnessing an unprecedented ageing of populations around the world. By 2050, the number of individuals >65 years will nearly triple to ~1.5 billion, accounting for ~16% of the world’s population¹. While cardiology, and medicine more generally, can appropriately take enormous pride in the progress made in extending human lifespan, we must also recognize that as our populations age, so will the clinical landscape.

Perhaps the quintessential example of how the evolution of human longevity is changing the face of medicine is the emerging epidemic of heart failure with preserved ejection fraction (HFpEF). Often described as a disease of the elderly, trends in HFpEF have closely paralleled the change in global ageing demographics. In the US, the prevalence of HFpEF, relative to heart failure with reduced ejection fraction (HFrEF), is increasing at an alarming rate of 1%/year with the overwhelming majority of patients being >65 years². Importantly, the prognosis for HFpEF remains poor with mortality rates comparable to HFrEF².

Unfortunately, while tremendous strides have been made in improving mortality in HFrEF, no pharmacological therapy, including our armament of neurohormonal antagonists, has demonstrated similar benefits in HFpEF. In fact, no large-scale clinical trial of medical therapy has met its primary endpoint in HFpEF.

Why have we failed to develop effective treatments for HFpEF? While various hypotheses have been proposed, central to all of them is the issue of heterogeneity. At a fundamental level this refers to the multifactorial nature of HFpEF pathophysiology. Initially thought to be due primarily to diastolic dysfunction, beautifully conducted human studies over the past 20 years have called into question this simplistic view, demonstrating that HFpEF encompasses a complex interplay of multiple impairments throughout the body. These involve not only diastolic function, but also cardiac reserves, systemic and pulmonary vascular function, renal function, oxygen carrying capacity, and peripheral oxygen extraction³. Complicating matters further, the degree to which each of these systems is altered, and hence, the relative impact of each one on symptoms, can vary markedly from patient to patient, raising the other major issue of heterogeneity in HFpEF- is this one condition or many?

Perhaps it is not surprising that a clinical condition defined largely by the absence of overt systolic dysfunction, together with common symptoms, such as dyspnea, is heterogeneous at multiple levels. While this has made HFpEF particularly difficult to study, recognizing this core concept has enabled human studies to make steady progress dissecting its pathophysiology. Unfortunately, this appreciation of heterogeneity has not carried over to preclinical investigation. For the most part, the diverse spectrum of findings identified in HFpEF patients have not been explored in animal models, which have primarily focused on diastolic function^{4, 5}. We believe this represents such a major shortcoming in the field that impedes our ability not only to identify key molecular mechanisms but also to develop much-needed therapies for HFpEF.

Although the value of animal models has been questioned in recent years, in many settings they have provided an essential bridge to understanding disease pathophysiology and developing effective therapeutics. This is perhaps particularly relevant in cardiovascular disease, given the intricate connection between disease and cardiovascular physiology, as well as limited access to clinical tissue samples. Indeed, one of the cornerstones of HFrEF therapy originated from a rat myocardial infarction (MI) model developed by Marc Pfeffer and colleagues in the 1970s⁶. This seminal work not only refined our understanding of the renin-angiotensin-aldosterone system (RAAS) in post-MI remodeling, but also demonstrated the efficacy of RAAS inhibition and laid the foundation for clinical trials and deployment of these interventions in ischemic heart disease and HFrEF.

To achieve similar success in HFpEF, we believe a vital tool will be animal models that capture a hallmark feature of the disease, namely multiple system pathologies. While systemic impairments are evident in both HFpEF and HFrEF, the etiology of these deficits appears different, which has significant implications for modeling these diseases in animals. In HFrEF, systemic pathologies are largely driven by a primary defect in systolic function, which can be induced in animals with clinically relevant interventions, such as coronary artery ligation⁶. In contrast, it remains unclear whether a single cardiac or systemic defect can reproduce the many systemic phenotypes seen in HFpEF. Thus identifying models that capture the broad spectrum of phenotypes seen in HFpEF patients provides our best hope of investigating and validating causative mechanisms.

Although seemingly intuitive, this approach to modeling has not been routinely pursued in preclinical HFpEF studies. For a detailed review of animal models of HFpEF, we refer the interested reader to the following literature^{4, 5}. For the most part, these models have focused on using comorbidities commonly associated with HFpEF to induce a single cardiac-specific phenotype frequently seen in patients. For example, induction of hypertension through surgical, dietary, and/or genetic interventions, followed by assessing its impact on cardiac hypertrophy or diastolic dysfunction is a common approach to studying HFpEF in animals. However, is this HFpEF? While diastolic dysfunction is certainly part of the compendium of HFpEF phenotypes, it is only one piece of the puzzle. Moreover, many HFpEF patients actually have normal cardiac dimensions, and not all individuals with diastolic dysfunction have HFpEF⁷. We believe animal models that recapitulate two or more HFpEF phenotypes (e.g. diastolic dysfunction and impaired exercise capacity) would be more faithful to the disease and more likely to yield clinically useful insights. Such models could be used to identify common biological mechanisms driving multiple pathologies. Moreover, when testing novel therapies, such animal models would also permit a multidimensional phenotypic readout of therapeutic efficacy, which could help identify patient subgroups most likely to benefit from a specific intervention.

In order to identify animal models with multiple system impairments, it will be critical to comprehensively phenotype existing and new models for the pathologies that have been implicated in patients. These phenotypes include alterations in cardiac morphology (hypertrophy, fibrosis, capillary density), systolic and diastolic function, cardiac reserves, macro- and microvascular function, pulmonary physiology, immune regulation, oxygen utilization, renal function, and skeletal muscle function. Importantly, some impairments in cardiopulmonary, vascular, or peripheral function in HFpEF are not evident at baseline but require physiologic stress, such as exercise, to unmask. This explains why the hallmark symptoms of this condition are typically elicited only with exertion or hemodynamic perturbations. Many clinical studies of HFpEF have recognized the importance of this concept and incorporated exercise testing to quantify impairments in physiologic reserves that could explain symptoms. However, in contrast, most preclinical studies have relied solely on resting cardiac functional assessments^{4, 5}. Moreover, these assessments are typically conducted in anesthetized animals further confounding the interpretation of hemodynamic and functional results. When combined with cardiac imaging or hemodynamic monitoring, exercise testing can provide tremendous insights into an animal’s physiologic reserves^{8, 9}. Indeed, measuring exercise capacity integrates functional assessment of many of the previously mentioned HFpEF phenotypes, and is closely related to dyspnea on exertion, the cardinal HFpEF symptom. In the search for mechanisms of disease, we believe the discoveries most likely to translate to patients will be those that causally link pathophysiology to functional phenotypes. For all these reasons, we believe exercise testing should be routinely incorporated into the evaluation of animal models of HFpEF.

We propose that more comprehensive characterization of preclinical HFpEF models, through multi-organ system phenotyping and physiologic stress-based functional testing, is necessary to identify better models of the human condition and bridge the large translational gap between animal and human HFpEF studies. In particular, as the utilization of high-throughput screening platforms continues to expand, we will likely find numerous promising biomarkers associated with clinical outcomes and HFpEF phenotypes. Determining which if any of these biomarkers play causal roles in HFpEF and thus represent targets for therapeutic intervention will require well-phenotyped animal models in which we can rigorously study molecular mechanisms and the consequences of intervention. In HFpEF, just as in any other complex multi-system disease, understanding the interaction between the various phenotypes becomes just as important as the phenotype itself. For example, assessing an intervention’s impact on cardiac hypertrophy is not sufficient to establish its role in HFpEF pathophysiology or its efficacy as a therapeutic target. The impact of the intervention would also need to be assessed on other phenotypes, such as diastolic function and cardiac reserves, and ultimately on pulmonary congestion and/or exercise intolerance, the final mediators of HFpEF symptoms. Of course, this can only be accomplished if the multiple phenotypes of HFpEF have already been characterized in our animal models.

In summary, while many questions regarding HFpEF pathophysiology remain, we are making headway through clinical investigation that is expanding our once cardio-centric perspective into one that recognizes multi-system contributors to HFpEF. However, continued progress will require parallel advances in basic and translational investigation. At this juncture, the development and comprehensive characterization of animal models is one of the biggest roadblocks to advancing our understanding of HFpEF, as well as to the identification and evaluation of novel therapeutic approaches for this condition. To address this shortcoming, we believe there should be greater emphasis on the comprehensive evaluation of multiple HFpEF phenotypes both at baseline and in response to relevant physiological stress. Until we have well-established animal models, we are unlikely to change the trajectory of the growing HFpEF epidemic in our aging populations.