The immunology of heart failure with preserved ejection fraction

Abstract

Heart failure with preserved ejection fraction (HFpEF) now accounts for the majority of new heart failure diagnoses and continues to increase in prevalence in the United States. Importantly, HFpEF is a highly morbid, heterogeneous syndrome lacking effective therapies. Inflammation has emerged as a potential contributor to the pathogenesis of HFpEF. Many of the risk factors for HFpEF are also associated with chronic inflammation, such as obesity, hypertension, aging, and renal dysfunction. A large amount of preclinical evidence suggests that immune cells and their associated cytokines play important roles in mediating fibrosis, oxidative stress, metabolic derangements, and endothelial dysfunction, all potentially important processes in HFpEF. How inflammation contributes to HFpEF pathogenesis, however, remains poorly understood. Recently, a variety of preclinical models have emerged which may yield much needed insights into the causal relationships between risk factors and the development of HFpEF, including the role of specific immune cell subsets or inflammatory pathways. Here, we review evidence in animal models and humans implicating inflammation as a mediator of HFpEF and identify gaps in knowledge requiring further study. As the understanding between inflammation and HFpEF evolves, it is hoped that a better understanding of the mechanisms underlying immune cell activation in HFpEF can open up new therapeutic avenues.

Summary

This review summarizes the definitions, health burden, and pathophysiology of the key aspects of HFpEF with a focus on literature pertaining to inflammation and the comorbidity-inflammation paradigm of HFpEF. Several knowledge gaps are highlighted here pertaining to multiple scientific domains important to creating a better understanding of human HFpEF. In addition, we have proposed an update to the inflammation-comorbidity paradigm of HFpEF to shift focus away from nitric bioavailability and toward more general pro-inflammatory pathways (Figure 5). This working model proposes that inflammation is a key inciting event in HFpEF that requires further investigation to determine what pathways are causal to HFpEF pathogenesis.

First, HFpEF is a heterogeneous clinical syndrome. The differing characteristics in subgroups of HFpEF raise uncertainty of whether all HFpEF patients have similar pathophysiology. It is possible that certain subgroups may benefit from different therapeutic approaches, but there is no accepted standard or method for subgroups patients with HFpEF. Second, there is a paucity of data regarding cellular mechanisms of HFpEF, as cardiac tissue from patients with HFpEF is not routinely available. Thus, much of what we know about human HFpEF is inferred from studies on plasma or serum proteins. Third, HFpEF is associated with systemic and myocardial inflammation but the mechanisms and cell types involved are poorly understood. Lastly, animal models of HFpEF are still emerging, and it is difficult to evaluate whether they truly reflect human HFpEF. Currently, there are a sizable number of potential HFpEF models in use, which can complicate generalizability of preclinical studies. However, effective preclinical models may provide valuable insights into mechanisms and causal contributors to the development of diastolic dysfunction.

Diagnosis and current therapies for heart failure with preserved ejection fraction (HFpEF)

Heart failure is a clinical syndrome characterized by shortness of breath, exercise intolerance, and symptoms of fluid retention. The diagnosis of heart failure (HF) is a clinical diagnosis supported by evidence from physical exam, history, laboratory values, and imaging. Broadly, heart failure is then further subclassified based on echocardiographic parameters into heart failure with preserved ejection fraction (HFpEF) 5 and heart failure with reduced ejection fraction (HFrEF). There also exists heterogeneity within HFpEF itself, which has provoked the question of whether HFpEF is a single clinical entity or an amalgamation of diseases with a similar presentation. Multiple attempts at grouping HFpEF into ‘phenogroups’ have been made with differing numbers of groups identified [1,2]. An analysis of patients enrolled in TOPCAT (Treatment of Preserved Cardiac Function Heart Failure with an Aldosterone Antagonist Trial) identified three distinct phenogroups of HFpEF that different based on left ventricular (LV) geometry, arterial stiffness, levels of natriuretic peptides, obesity, and chronic kidney disease [1]. These efforts indicate there is clear variability in the cardiac physiology and comorbidities associated with HFpEF, with likely differing relative contributions for each individual. Phenogroups can also provide prognostic information, but it remains to be seen whether they will respond similarly to new therapies.

HFpEF is a common disease with substantial health burden. The lifetime risk of heart failure in the United States is 20–45% after 45 years of age, depending on racial and ethnic groups [3]. Incidence of heart failure in the United States remained relatively stable from 1990 to 2009, but the contribution of HFpEF to overall heart failure is increasing. Meanwhile HFrEF is declining at a similar rate [4]. Another analysis showed that the proportion of deaths in the United States between 1999 and 2018 from ischemic heart disease declined, while those attributable to heart failure increased. Additionally, the proportion of hypertensive heart disease increased [5]. This reflects a change in the burden of heart disease and heart failure in the United States away from atherosclerotic disease toward that of heart failure. This may be attributable to improvements in both acute and chronic treatment of the manifestations of atherosclerosis as well as changes in other health behaviors and disease burden.

The health consequences of being diagnosed with HFpEF are dramatic. Patients with HFpEF demonstrate increased risk of both cardiovascular and non-cardiovascular mortality compared with those without heart failure [6,7]. In addition, heart failure can be accompanied by frequent hospitalizations for diuresis, especially in advanced stages of disease. Heart failure hospitalization is accompanied by a 3.8% in-hospital mortality rate [8].

Unfortunately, many of the therapeutics that improve mortality and hospitalization HFrEF have failed to show benefits in HFpEF [9]. Drugs targeting the renin–angiotensin aldosterone system have mortality benefits in HFrEF. Meanwhile, ACE inhibitors and angiotensin receptor blockers have not shown a beneficial effect on mortality or hospitalization in HFpEF [10,11]. The aldosterone antagonist spironolactone did improve hospitalization rate with some signal for mortality benefit [12], although it did not meet statistical significance for its primary outcome [13]. Recently, sodium-glucose cotransporter 2 inhibitors (SGLT2i) demonstrated improvements in heart failure hospitalization and became the first HFpEF therapy to meet its primary outcome in a randomized, double-blinded, placebo-controlled, event-driven trial [14]. Intriguingly, there is data to suggest these drugs may derive their cardioprotective properties independent of SGLT2, but this is an area of active research [15,16]. Lastly, there is a role for exercise training in HFpEF, as it has been shown to improve cardiorespiratory fitness and quality of life [17,18].

Pathophysiology of HFpEF

One of the hallmark features of HFpEF is diastolic dysfunction; however, diastolic dysfunction can be present in individuals without heart failure. A large study of elderly individuals showed that measures of diastolic function decline as part of healthy aging [19]. HFpEF has clearly been shown to be a syndrome with complex pathophysiology, and the following sections will delve into the specific literature surrounding alterations in cardiac physiology, cardiac cellular composition, and multi-organ communication that contribute to HFpEF development.

Cardiac and vascular physiology

Cardiovascular physiology is dramatically altered in HFpEF with multiple groups documenting impaired endothelial function, arterial and LV stiffness, loss of cardiac reserve, and alterations in both systolic and diastolic function in patients with HFpEF.

Elevated left atrial filling pressures are a hallmark of heart failure, particularly during exercise [20]. Patients with HFpEF may have elevated LV filling pressures at rest in advanced stages of heart failure, but many patients only exhibit elevated filling pressures when stressed with exercise. In one study, patients with HFpEF did not achieve a higher peak pulmonary capillary wedge pressure (PCWP) compared with controls, but they reached peak PCWP at a much lower workload. In addition, patients with HFpEF had much higher systemic vascular resistance during exercise [21]. A separate group found that patients with HFpEF were unable to increase both stroke volume and heart rate to the same degree as controls during exercise, which was associated with a concomitant impairment in pulmonary artery vasodilation, suggesting that limited cardiac reserve and vascular dysfunction are major determinants of exercise intolerance in HFpEF [22]. Similarly, patients with HFpEF have impairments in reactive hyperemia-induced endothelial dependent vasodilation, which was similar to impairments seen in individuals with hypertension. In addition, abnormalities in chronotropic, contractile, vascular, or endothelial reserve all correlated with peak exercise capacity, suggesting that HFpEF is an emergent phenomenon of multiple derangements in cardiac and vascular function [23].

Diastolic dysfunction is a key aspect of HFpEF pathophysiology, and by definition HFpEF must have a preserved ejection fraction. Preserved ejection fraction, however, is not synonymous with completely normal systolic function. Echocardiography at and during exercise in patients with HFpEF demonstrated abnormalities in both systolic and diastolic functional measures, including reduced radial and longitudinal systolic stain, reduced systolic and diastolic longitudinal functional reserve, and delayed ventricular untwisting [24].

Chamber stiffness is thought to be integral to the pathophysiology of HFpEF. Westermann et al. performed hemodynamic measurements in 90 patients and found that patients with HFpEF exhibited diastolic dysfunction and increased LV stiffness [25]. In addition to increased myocardial stiffness, increased vascular stiffness can have compounding effects on diastolic dysfunction, which is described by measurements of ventricular-arterial coupling. Indeed, patients with HFpEF display reduced aortic distensibility [26] and increased arterial stiffness [27]. The functional consequence of increased stiffness in the LV and vasculature is that it becomes more sensitive to volume status, with small changes in volume creating higher filling pressures. In addition to vascular stiffness, there are also numerous studies supporting an association of coronary microvascular dysfunction in HFpEF, even in the absence of macrovascular CAD [28].

While these studies provide valuable insight into the alterations in cardiovascular physiology during HFpEF, we lack conclusive evidence as to which derangements in cardiac physiology are most important for the genesis of heart failure symptoms. In addition, these studies are often small in sample size due to the requirement for invasive hemodynamic monitoring, potentially precluding generalizability to the broader, heterogeneous population of individuals with HFpEF. As new therapies that prove to be effective for ameliorating HFpEF symptoms become available, tracking these parameters in response to treatment may provide further insight. The current data implicate reductions in systolic and diastolic reserve, impaired vasodilation, and increased atrial and LV stiffness as key alterations associated with HFpEF.

Cardiac remodeling

The myocardium is composed of a variety of cell types, including vascular cells, cardiomyocytes, fibroblasts, and local immune cells, all of which are profoundly affected by hypertension and heart failure development. Two hallmarks of cardiac remodeling in HFpEF are interstitial cardiac fibrosis and cardiac hypertrophy. HFpEF is characterized by reactive fibrosis, an accumulation of collagen in the interstitium. This process is distinguished from the replacement fibrosis seen in HFrEF, where collagenous bands of scar tissue are formed to replace swaths of necrotic tissue. Cardiac fibroblasts when activated are the major source of collagen in the heart and therefore the fibrosis seen in hypertensive heart disease. Increased fibrosis can contribute to increased stiffness and cardiac dysfunction. Stiffness can also result from increases in cellular stiffness, primarily due to alterations in sarcomeric proteins and their post-translation modifications. As cardiac myocytes are largely post-mitotic after the developmental period, cardiac hypertrophy results from cellular growth of cardiomyocytes (Figure 1). However, other cell types and alterations in the extracellular matrix can also affect gross heart weight.

Figure 1. Macro and microscopic changes in HFpEF.

HFpEF is associated with alterations in cardiac physiology and structure with concentric LV hypertrophy, diastolic dysfunction and myocardial stiffness being hallmarks of disease pathogenesis. In addition, changes in cellular composition and extracellular matrix occur including myocyte hypertrophy, increased interstitial collagen deposition, increased abundance of immune cells, and alterations in microvascular structure. Schematic made with Biorender.

Cardiac remodeling and its contributors are well-described in the context of hypertension or increased afterload placed on the heart. While hypertension is not the only stimulus for cardiac remodeling, many studies support an important role for the renin–angiotensin aldosterone system (RAAS). Both angiotensin II (Ang II) and aldosterone stimulate collagen synthesis in isolated cardiac fibroblasts in vitro [29-31]. The spontaneously hypertensive rat develops interstitial cardiac fibrosis, increased passive stiffness, and elevations of collagen types I and III, all of which can be improved by treating with the ACE inhibitor lisinopril [32-34]. In addition, rats receiving uninephrectomy, aldosterone and supplemental salt in the drinking water develop hypertension, hypertrophy, and both interstitial and perivascular cardiac fibrosis [35,36]. The critical role of the RAAS in perpetuating cardiac remodeling was confirmed in humans in a small clinical trial in which 35 patients with hypertension underwent treatment with lisinopril or the thiazide diuretic hydrochlorothiazide for 6 months with LV catheterization and endomyocardial biopsy at baseline and at the conclusion of the study. Patients receiving lisinopril but not hydrochlorothiazide had regression of LV fibrosis and improvement in diastolic function without changes in left ventricular hypertrophy [37]. Regression of myocardial fibrosis was also seen in patients randomized to the angiotensin receptor blocker losartan compared with those receiving amlodipine, despite similar changes in blood pressure [38]. A subsequent study of patients with hypertension showed that losartan both decreased cardiac fibrosis and measures of myocardial stiffness [39]. Aldosterone antagonists have also been shown to decrease circulating markers of fibrosis in patients with HFpEF [40] and HFrEF [41].

To determine whether hypertrophy and fibrosis result merely from hypertension or RAAS activation, Brilla et al. undertook a studying utilizing three models of increased afterload with differing profiles of RAAS activation: renovascular hypertension (elevations of Ang II and aldosterone), aldosterone infusion (elevation of aldosterone alone), and infrarenal aortic banding (normal levels of Ang II and aldosterone). In both infusion of aldosterone and renovascular hypertension, fibrosis was detected in both the left and right ventricles, but cardiac hypertrophy was detected only in the left ventricle. In infrarenal aortic banding, cardiac hypertrophy was observed but not fibrosis, suggesting that cardiac hypertrophy is most closely related to the afterload placed on the left ventricle while cardiac fibrosis may be governed by RAAS activation [42]. However, the majority of studies employing transverse aortic constriction (TAC) show that afterload does indeed induce myocardial fibrosis [43]. Fibrosis, hypertrophy, and LV dysfunction are all dependent on both the degree of constriction [44,45] as well as the duration of treatment [46]. Intriguingly, a study assessing the removal of TAC-induced pressure overload showed that despite hemodynamic normalization and regression of cardiomyocyte hypertrophy, regression of LV fibrosis was incomplete [47]. These findings suggest that transient alterations in hemodynamics may have lasting effects on cardiac structure. Despite no direct manipulation of RAAS in aortic banding models, one long-term study in rats showed that low doses of the ACE inhibitor ramipril was able to prevent myocardial fibrosis without lowering systemic blood pressure [48]. While Ang II can stimulate hypertrophy in isolated cardiomyocytes, elegant studies using conditional knockout strategies in mice have demonstrated that cardiac hypertrophy is a result of increased blood pressure, not direct actions of Ang II-AT1R signaling in the heart [49,50]. Thus, hypertension and increased afterload are important drivers of cardiac remodeling relevant to HFpEF.

Cardiac remodeling can also occur without hypertension in preclinical models, as seen in the context of doxorubicin treatment [51], obesity [43], and streptozotocin-induced diabetes [52]. While RAAS inhibition has also been shown to decrease circulating procollagen levels and LV function in patients with obesity [53] and metabolic syndrome [54], there are also other important mediators of cardiac fibrosis outside of RAAS-related signaling. Of note, there is a large body of literature on transforming growth factor β (TGF-β) and its central role in activating fibroblasts. Indeed, the anti-fibrotic medication pirfenidone has been shown to slightly reduce myocardial extracellular volume in patients with HFpEF, which may be partially due to its effects on TGF-β production and signaling [55]. Interestingly, pirfenidone also has anti-inflammatory effects [56]. Other mediators such as connective tissue growth factor, endothelin-1, and catecholamines have also been shown to stimulate collagen production in fibroblasts. These and other cellular mechanisms underlying fibroblast activation have been well reviewed elsewhere [57,58].

Multi-organ communication

Given that HFpEF is generally considered a systemic disease, there is mounting evidence that peripheral changes in the skeletal muscle, kidney, and adipose tissue can contribute to HFpEF pathogenesis.

Heart failure is associated with impaired skeletal muscle function and accompanying structural changes, such as loss of type I fibers and reduced mitochondrial density of type II fibers, but much of the available literature is focused on HFrEF [59]. HFpEF is increasingly recognized to be associated with muscle loss in the context of adipose tissue expansion, termed ‘sarcopenic obesity’, as sarcopenia has been noted in 1 of 5 patients with HFpEF [60]. A study examining thigh volumes skeletal muscle and adipose tissue, found increased amounts of intermuscular fat in patients with HFpEF even when subcutaneous fat was not significantly different [61]. Preliminary evidence suggests that those with HFpEF may have even worse muscular atrophy than those with HFrEF. Those with HFpEF had reduced mitochondrial size, elevated expression of genes associated with atrophy, and alterations in genes relating to fatty acid oxidation and glucose metabolism when compared with both healthy controls and those with HFrEF [62]. While it is difficult to parse out whether skeletal muscle dysfunction results from deconditioning due to heart failure or vice versa, additional studies on exercise as a form of therapy may help reveal the contribution of skeletal muscle to HFpEF symptom development.

Alterations in hematopoiesis may also contribute to HFpEF. Endothelial progenitor cells are bone marrow-derived cells that circulate in peripheral blood and participate in endothelial repair. HFpEF is associated with decreases in circulating angiogenic T cells and endothelial progenitor cells [63]. Further work focused on understanding the inter-organ communication in HFpEF is warranted to better guide therapeutic development.

The inflammation-comorbidity paradigm of HFpEF

The immune system is broadly divided into innate and adaptive immunity. Innate immunity is important for initial immune surveillance and defense. The innate immune system is non-specific in its ability to respond to pathogens or tissue damage. Innate immune cells such as macrophages, monocytes, and dendritic cells rely on pattern recognition receptors that recognize common motifs of bacteria, fungi, viruses, and other pathogens. These motifs are termed pathogen associated molecular patterns when associated with an infectious agent, but there are also endogenous ligands termed damage associated molecular patterns often associated with tissue damage or inflammatory mechanisms of cell death [64].

Meanwhile, the adaptive immune system is slower in its response to infections or damage but is incredibly specific. Lymphocytes contain receptors that recognize specific epitopes and each lymphocyte is selected to only recognize a particular epitope. T lymphocytes are characterized by expression of T-cell receptors and the co-receptor CD3. T cells are often grouped into T helper and cytotoxic subsets, characterized by expression of CD4 and CD8, respectively. T helper cells help coordinate immune responses and are often further divided based on the characteristic cytokines they produce. For example, T regulatory cells are an anti-inflammatory population of CD4 T cells that have high expression of inhibitory receptors and produce anti-inflammatory cytokines such as transforming growth factor β. Cytotoxic T cells produce effector molecules such as perforins and granzymes to lyse infected or damaged cells, but they can also produce cytokines of their own [65]. Together, the innate and adaptive immune systems coordinate complex immune responses to fight infections, mitigate tissue damage, and promote wound healing.

Both innate and adaptive immunity have been implicated in preclinical models of relevant risk factors for HFpEF. In addition, elevations in circulating cytokines reported across multiple studies support innate immune cell activation. This includes the well-validated heart failure biomarker soluble ST2, which binds the pro-inflammatory alarmin IL-33, and inflammation related proteins such as pentraxin-3 and galectin-3 [66-71]. Moreover, increased numbers of both innate and adaptive immune cells are found in the hearts of HFpEF patients [72]. Lastly, an imbalance of pro-inflammatory IL-17A producing CD4+ T cells (Th17) and anti-inflammatory T regulatory cells has been documented in HFpEF [73]. While there is no direct evidence to support a specific antigen leading to B cell or T cell activation in heart failure, there are many studies supporting a role for cytokine signaling alone leading to increased T-cell effector functions [74].

The comorbidity-inflammation paradigm (Figure 2) was proposed in an influential review by Paulus et al. that speculated the following sequential pathogenesis of HFpEF: (1) multiple comorbid diseases induce a systemic inflammatory state, (2) inflammation induces endothelial dysfunction and reduces nitric oxide bioavailability, (3) resulting low protein kinase G (PKG) activity in cardiomyocytes alters titin phosphorylation and predisposes toward hypertrophy, and (4) stiff cardiomyocytes and interstitial fibrosis lead to elevated diastolic filling pressure and symptoms of heart failure [75].

Figure 2. Schematic depiction of the comorbidity-inflammation paradigm.

This paradigm contends that comorbid diseases such as Type 2 diabetes mellitus (T2DM), obesity, age, hypertension (HTN), and chronic kidney disease (CKD) lead to increased circulating cytokines, decreased nitric oxide availability, and alterations in local cardiac cells. These changes lead to increased stiffness, creating diastolic dysfunction and symptomatic heart failure. Figure created with Biorender.

Endothelial dysfunction is an imbalance of normal endothelial-derived vasodilatory and vasoconstricting molecules, which is most often associated with impairments in nitric oxide (NO) bioavailability. Meanwhile, endothelial activation is often used to describe a pro-inflammatory state in which endothelial cells upregulate adhesion molecules to promote immune cell recruitment. These processes are not mutually exclusive, and both endothelial dysfunction and activation have been implicated in HFpEF.

Nitric oxide signals within cardiomyocytes and smooth muscle cells by activating soluble guanylyl cyclase to promote conversion of guanosine triphosphate to cyclic guanosine monophosphate (cGMP) within the cell. Protein kinase G is a serine-threonine kinase that is activated by cGMP and plays an important role in myocyte relaxation and cardioprotection through modulation of downstream signaling events. Low PKG activity and low levels of cGMP have been observed in myocardial biopsies of patients with HFpEF, which was associated with increased passive stiffness of isolated cardiomyocytes and was independent of soluble guanylate cyclase (sGC) and phosphodiesterase (PDE) 5A levels [76]. These results are suggestive of decreased local nitric oxide bioavailability. Boosting cGMP levels via PDE-5 inhibition by sildenafil in the RELAX trial did not improve exercise capacity or clinical status. There were limited improvements, however, in circulating cGMP levels in those receiving sildenafil, which may partially explain the negative results [77]. Direct stimulation of sGC via vericiguat was also met with negative results when examining changes in N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels or left atrial volume [78]. These results suggest that improving downstream signaling of nitric oxide may not be sufficient to improve diastolic dysfunction and quality of life in HFpEF. Instead, future effects may need to focus on signaling events upstream that first lead to endothelial dysfunction.

Given that nitric oxide boosting therapies have not shown clinical benefit in HFpEF, we propose an updated framework of the comorbidity-inflammation paradigm in which inflammatory mediators drive vascular dysfunction, cardiac fibrosis, and diastolic dysfunction. The mechanisms and pathways potentially important for these effects are still being enumerated, but the prior focus on nitric oxide bioavailability should be replaced with a more general pro-inflammatory cascade. Importantly, we propose that inflammation is a causal feature of HFpEF while in HFrEF, it may be a reaction to underlying myocardial structural damage or ischemia (Figure 23).

In the Multi-Ethnic Study of Atherosclerosis, serum levels of vascular cell adhesion molecule-1 (VCAM-1) were associated with incidence of total heart failure, and upon examination of heart failure subtypes, this effect remained significant for HFpEF, suggesting that endothelial activation precedes overt HFpEF symptoms [79]. A similar study found that increased serum levels of IL-6 and TNF-α both associate with incident heart failure, with a stronger association for HFpEF than HFrEF [80]. Moreover, a broad biomarker profiling effort in a large cohort of HFpEF and HFrEF patients found that HFpEF was associated with pathways related to inflammatory response, neutrophil degranulation, cell adhesion, and extracellular matrix organization [81].

An elegant study by Hahn et al. is worth highlighting given it is one of the only studies with extensive data examining alterations of myocardial gene expression in HFpEF [82]. In this study, RNA sequencing was performed on endomyocardial biopsy specimens from the RV septum in patients with HFpEF, RV septal tissue from HFrEF explanted hearts in patients undergoing transplantation, and RV midseptal tissue obtained from brain-dead organ donors. Transcriptomes from control, HFpEF. and HFrEF could be separated in principal components analysis even when adjusting for sex, age, diabetes, and renal function. Inflammatory- and immune-related pathways were up-regulated in both HFpEF and HFrEF conditions, while HFpEF was specifically associated with upregulated of the adenosine triphosphate synthesis and oxidative phosphorylation pathways. However, many immune related genes had to be excluded in the analysis due to correlation with hemoglobin genes, indicative of peripheral blood contamination during biopsy specimen collection. This highlights the challenge in obtaining clinical tissue specimens for comparison with regards to the immune system. Nonetheless, these results suggest that cardiac metabolism may be an important avenue of future research.

Together, these studies provide supporting evidence that HFpEF is associated with chronic inflammation, but whether inflammation is causal to HFpEF remains to be seen. In the next few sections, we focus on the inciting events detailed in step one of this proposed framework: how risk factors for HFpEF lead to an inflammatory state and what potential pro-inflammatory stimuli have the greatest degree of supporting evidence.

Risk factors for HFpEF and their association with inflammation in humans

While HFpEF and HFrEF lead to a similar clinical syndrome, they are distinct in their pathophysiology, risk factors, and prevalence. However, these two subtypes of heart failure do have a few overlapping risk factors such as coronary artery disease, hypertension, aging, and diabetes. A cohort study comparing predictors of HFpEF versus HFrEF found that female sex, atrial fibrillation, increased urinary albumin secretion, and increased cystatin-C levels were preferentially associated with HFpEF. Meanwhile, male sex, smoking, high sensitivity troponin T, and prior myocardial infarction were associated with HFrEF [83]. In a large study of Medicare beneficiaries without heart failure, age, diabetes and chronic kidney disease were associated with incidence of heart failure regardless of ejection fraction. Male gender was associated with increased risk of HFrEF, but not HFpEF. Meanwhile, obesity and pulmonary hypertension were more strongly associated with risk of developing HFpEF [84]. These differences in risk factor development provide support for differences in the underlying pathophysiology of heart failure subtypes. Additionally, treatments that have proven useful in treating patients with HFrEF have largely failed to show such benefits in HFpEF. Finally, mendelian randomization of heart failure risk factors showed that atrial fibrillation, coronary artery disease (CAD), body mass index (BMI), systolic blood pressure (SBP), and pulse pressure were all significantly associated with HFpEF. Meanwhile, LDL, HDL, T2DM, and eGFR were not significant associated with HFpEF [85]. Chronic obstructive pulmonary disease (COPD) is another risk factor for HFpEF, but accurate diagnosis of these comorbid conditions can be challenging given overlap of symptoms and the fact that many patients with HF are on beta-blockers, which can impact pulmonary testing [86]. Lastly, a meta-analysis of HFpEF studies found that 59% of those with HFpEF have concomitant iron deficiency [87]. Anemia is more common in HFpEF compared with HFrEF, and is independently associated with all-cause mortality and HF hospitalization regardless of EF [88,89].

Many of the risk factors for HFpEF are also associated with chronic low-grade or subclinical systemic inflammation (Figure 4). Thus, it has been postulated that inflammation itself may play a role in the pathogenesis of HFpEF. The focus on this section is on the comorbidity-inflammation paradigm of HFpEF, which posits that systemic inflammation associated with increased burden of cardiometabolic disease leads to increased passive stiffness and elevated filling pressures driving symptomatology in HFpEF.

Figure 4. Major Risk Factors for HFpEF.

HFpEF is associated with a wide range of comorbid disease thought to be important for the initiation and progression of heart failure symptoms. These include aging, obesity, hypertension, coronary artery disease, renal dysfunction or chronic kidney disease, diabetes, anemia, chronic obstructive pulmonary disease (COPD), pulmonary hypertension, and atrial fibrillation. Figure made with Biorender.

Hypertension

Hypertension can contribute to HFpEF pathogenesis in myriad ways. Hypertension affects vascular function, arterial stiffening, cardiac remodeling, sympathetic nervous system activation, and renal function. Therefore, it is not difficult to draw connections between HFpEF pathophysiology and hypertension. Hypertension is also associated with low-grade inflammation. Recently, our group performed an unbiased and comprehensive deep immunophenotyping of peripheral blood mononuclear cells from normotensive and hypertensive subjects. Most of the hypertensive subjects were treated with a variety of common antihypertensive agents. We found that human hypertension is associated with decreases in regulatory T cells, suggestive of decreased tolerance mechanisms or impaired ability to suppress T-cell activation [90]. This finding has been validated and extended by an independent study showing that decreased T regs associates with increases in LV mass [91]. Our group also demonstrated that IL-21 production by CD4 T helper cells correlates with systolic blood pressure in humans [92]. A separate study of circulating immune cells from hypertensive patients identified increases in immunosenescent and pro-inflammatory CD8 T cells, including perforin, interferon gamma and tumor necrosis factor α (TNF-α) positive CD8 T cells. In addition, ligands for the C-X-C chemokine receptor type 3 (CXCR3) were increased in hypertensive individuals, supporting a role for T-cell activation and migration in hypertension [93].

There is limited data supporting a role for B lymphocytes in hypertension; however, there are multiple studies suggesting increases in serum immunoglobulins in patients with hypertension [94-96]. Lastly, we examined transcriptional differences between monocytes from normotensive and hypertensive individuals and identified upregulation of genes related to IL-1β and IL-18 signaling in hypertensive monocytes, consistent with a pro-inflammatory phenotype [97]. The current evidence implicates both innate and adaptive immunity and associations with pro-inflammatory cytokine production in human hypertension.

Obesity

The prevalence of obesity is approximately 42% of all adults in the United States [3]. Obesity is characterized by an expansion of adipose tissue across different adipose tissue depots. Adipose tissue can be a reservoir for immune cells, and obesity is associated with increased adipose tissue immune cell abundance. Studies in human adipose tissue have implicated a unique macrophage phenotype termed ‘metabolically’ activated macrophages that have a mixed phenotype not aligned with the classical M1 and M2 paradigm. These macrophages produce pro-inflammatory cytokines TNF-α and IL-1β and express surface markers associated with lipid handling such as CD36, ABCA1, and PLIN2 [98]. More recent studies using single cell sequencing techniques have identified new markers for lipid-associated macrophages that are enriched in obesity. This population is characterized by high expression of TREM2 and other lipid-handling genes and marked by CD9 and CD63 [99].

Those with obesity driven HFpEF are thought to represent a distinct phenotype with greater degrees of plasma volume expansion and LV remodeling as well as worsened exercise capacity compared with HFpEF due to other comorbidities [100]. It is difficult to separate out the effects that anatomically distinct adipose tissue depots may have on the heart during obesity. Changes in adipose tissue can impact distant organs through adipokines, cytokines or metabolites. The Dallas Heart Study found that visceral adipose tissue but not abdominal subcutaneous adipose tissue was significantly associated with LV remodeling in obese individuals [101]. Whether the inflammation associated with obesity or alterations in adipokines or metabolites might be most important for development of HFpEF is poorly understood.

Of note, there are adipose tissue depots in contact with and surrounding the heart—the epicardial and pericardial adipose tissue, respectively. Intriguingly, epicardial adipose tissue is absent in rodents. HFpEF is associated with increased epicardial adipose tissue and alterations in its composition and structure. Accumulation of epicardial adipose tissue is associated with poor prognosis in HFpEF, greater hemodynamic derangements, and worsened exercise capacity [102,103]. Epicardial adipose tissue expansion correlates with cardiac fibrosis [104], impaired microcirculatory function, and diastolic dysfunction [105,106]. Obesity promotes a pro-inflammatory phenotype within epicardial adipose tissue with increased levels of TNF-α, IL-1β, and IL-6 [107]. Thus, changes in epicardial or even distant adipose tissue depots may contribute to cardiac dysfunction through cytokine or adipokine signaling. Whether direct targeting of epicardial adipose tissue will ameliorate HFpEF symptoms or hemodynamics is still to be determined. There is, however, limited evidence to suggest that weight loss via bariatric surgery, diet, or exercise can improve symptoms, reverse cardiac remodeling, and improve diastolic function [108,109], suggesting that weight loss should be a cornerstone of treatment in those with HFpEF driven by obesity.

Diabetes

An estimated 28 million adults in the United States have diagnosed diabetes with an additional 113 million adults with evidence of prediabetes. Development of diabetes and associated hyperglycemia is associated with vascular damage and is a major risk factor for atherosclerosis. In a study of hospitalized patients with HF exarcerbation and underlying HFpEF, diabetes was associated with longer length of stay and increased 30-day readmission but not 30-day mortality [110]. In addition, diabetes is a systemic pro-inflammatory state. Increased numbers of pancreatic islet macrophages and increased in pro-inflammatory gene expression is a hallmark of Type 2 diabetes [111,112]. In addition, hyperglycemia directly induces IL-1β production in human islets ex vivo [113]. Evidence for a causal role of inflammation in human diabetes is under investigation. In a small trial of 70 patients, the IL-1 antagonist anakinra reduced hemoglobin A1c and improved insulin C-peptide secretion in patients with Type 2 diabetes compared with placebo [114]. However, a much larger study of approximately 10,000 patients treated with the IL-1β neutralizing antibody canakinumab or placebo showed no difference in incident diabetes or measures of hyperglycemia in those with pre-existing diabetes [115]. Thus, whether inflammation contributes directly to insulin resistance—and therefore HFpEF—is a subject of ongoing research.

Renal disease

The overall prevalence of chronic kidney disease is approximately 15% of the adult population in the United States [3], and chronic kidney disease (CKD) is associated with significant morbidity and mortality. A majority of patients with HFpEF have some form of renal abnormality, with one study reporting 62% of individuals with HFpEF having an estimated glomerular filtration rate <60 or evidence of albuminuria. In addition, renal dysfunction in HFpEF was associated with abnormal LV geometry [116]. In an analysis of the PREVEND trial, increases in urinary albumin or in cystatin C are associated with increased incidence of HFpEF but not HFrEF [117]. Lastly, unbiased phenomapping revealed a phenogroup of HFpEF associated with CKD, which had the worst prognosis of the phenogroups identified [2]. Together, these studies suggest that renal impairment and HFpEF are tightly linked.

Renal impairment has long been associated with inflammation and alterations in immune cell function. Biomarkers of inflammation are positively associated with albuminuria and inversely associated with renal function, including IL-1β, IL-6, TNF-α, and C-reactive protein [118]. In particular, IL-6 adds significant predictive power to traditional risk factors when assessing mortality and cardiovascular death in patients with CKD [119].

Aging

Aging is a core risk factor for HFpEF. The term ‘inflammaging’ has been coined to describe low-grade immune activation and dysfunction associated with the process of aging. The process of aging is multifactorial and often associated with muscle loss, adipose tissue expansion, alterations in sex hormones, and an increasingly sedentary lifestyle. Elevated levels of CRP, IL-1β, IL-6, and TNF-α are associated with aging [120].

One of the hallmarks of aging is cellular senescence in which cells lose their potential to divide and become dysfunctional. These senescent cells accumulate with age and can be a source of pro-inflammatory cytokines [121].

Mitochondrial dysfunction has long been associated with aging. Sirtuins are NAD+ deacetylases that have been implicated in cardiac aging and impaired mitochondrial function. A study using human LV tissue found a female-specific decrease in sirtuins 1 and 3 with a concomitant increase in cardiac macrophages and NF-kB signaling [122]. In addition, damaged or dysfunctional mitochondria can be a source of mitochondrial DNA released into the cytosol or extracellular environment. Mitochondrial DNA (mtDNA) can act as a damage associated molecular pattern, activating pattern recognition receptors and initiating a cascade of inflammatory signaling, including TLR9 signaling, cGAS-STING signaling, and NLRP3 inflammasome formation [123]. The human heart displays increased levels of peroxidation and oxidative stress with aging [124]. Whether release of mtDNA is a major inciting event in cardiac inflammation or whether therapeutics aimed at preserving mitochondrial function can prevent cardiac aging is yet to be determined.

Animal models of HFpEF

In mice, HFpEF has been modeled by studying how these major risk factors impact cardiac function, particularly aging, hypertension, obesity, and diabetes. Given the clinical heterogeneity of HFpEF patients, it is unlikely that a single animal model can accurately reflect all human HFpEF. Ideally, animal models of HFpEF should reflect key aspects of pathophysiology that can be quantitated in a rigorous fashion. A key advantage of animal models is the ability to isolate the contributions of each risk factor, yet many risk factors in people occur together such as hypertension, obesity, and diabetes mellitus. Therefore, a variety of animal models may be needed to disentangle the effects of particular risk factors either in isolation or together (Table 1).

Table 1.

Animal models of HFpEF

Model category	Model	Species	Strain	Length	Reference
Hypertension	DOCA-salt	Rat or mouse	C57Bl6/J or Sprague-Dawley	3–4 weeks	[141,194]
	SHR	Rat	Spontaneously hypertensive rat	1 year	[152,153]
	Dahl SS	Rat	Dahl salt sensitive rat	20 weeks	[149-154]
	Ang II	Mouse	C57Bl6/J	4 weeks	[144-147]
Pressure Overload	TAC	Mouse	C57Bl6/N	2–4 weeks	[44-47]
Obesity and Diabetes	Loss of leptin signaling	Mouse	db/db or ob/ob mouse	10–16 weeks	[159-163]
Aging	Accelerated aging	Mouse	Senescence-accelerated mouse prone 8	6 months	[133-135]
	Normal aging	Mouse	C57Bl6/J or Fischer 344 rat	1–2 years	[129-132]
Multi-hit	ZSF1	Rat	Zucker fatty and spontaneously hypertensive rat	20 weeks	[176-179]
	L-NAME+HFD	Mouse	C57Bl6/N	5–15 weeks	[167]
	Ang-II+HFD	Mouse	C57Bl6/J	12 weeks	[169]
	Aldo + db/db	Mouse	db/db mouse	4 weeks	[168]

Aging

Aging plays a major role in the onset of diastolic dysfunction in humans as well as increases in passive myocardial stiffness [125-127]. Mice and rats are often used to study age-related disease. Indeed, C57Bl6/J mice develop age-related diastolic dysfunction, which recapitulates features of human cardiac aging [128,129]. However, diastolic dysfunction in mice was not detected until 15-24 months, likely preventing its widespread uptake as a mouse model of HFpEF. Rats also develop age-associated changes in the myocardium, including increased fibrosis and loss of cardiomyocyte cell number [130,131]. Additional studies have utilized rodent models of accelerated aging which develop signs of HFpEF at earlier ages, including diastolic dysfunction and cardiac hypertrophy [132-135]. Particularly when thinking about genetic manipulations in rodents, aging can be a difficult model due to the required study length. However, there are advantageous aspects to rodent models of aging that recapitulate human disease, such as alterations in bioenergetics. Loss of mitochondrial function and NAD⁺ occurs in the heart with aging with downregulation of enzymes responsible for NAD+ biosynthesis. Repletion of NAD⁺ in a mouse model of HFpEF protected from mitochondrial dysfunction and development of HFpEF [136]. Thus, preclinical models of aging represent an important aspect of HFpEF pathogenesis.

Hypertension

Hypertension has well-known effects on the heart, some of which are driven by increased afterload and some of which are driven by neurohormonal changes. A common model to study diastolic dysfunction has been the DOCA-salt model, in which a rodent undergoes unilateral nephrectomy, exogenous deoxycorticosterone acetate (DOCA) administration, and supplemental salt added to the drinking water. DOCA is an aldosterone precursor and is sometimes replaced with aldosterone infusion or injections in some studies. Initial studies using the DOCA-salt model were performed in rats, showing that DOCA-salt leads to hypertension, cardiac hypertrophy, and cardiac fibrosis [36,137-139]. Later studies performed in mice found similar findings [140]. Multiple groups have described diastolic dysfunction in DOCA-salt rodents in the context of preserved systolic function, highlighting its reproducibility. Moreover, DOCA-salt-induced diastolic dysfunction is responsive to SGLT2 inhibition, suggesting its translatability [141]. The protective effect of SGTL2i in DOCA-salt is independent of an effect on the kidney, as others have found that empagliflozin does not improve albuminuria in DOCA-salt treated animals [142]. In addition, work using a similar model to DOCA-salt with the substitution of aldosterone infusion found similar findings, including LV remodeling and diastolic dysfunction in the setting of hypertension [143].

Perhaps the most common hypertension mouse model is that of Ang II infusion. Mice undergoing Ang II infusion develop hypertension, cardiac hypertrophy, cardiac fibrosis, and diastolic dysfunction [144-147]; however, there is a lack of standardization regarding Ang II dosages and duration of infusion throughout the literature. Other elements of a HFpEF model such as pulmonary congestion and exercise intolerance are unknown. Thus, more characterization is needed before the Ang II infusion model can be adopted as a hypertensive HFpEF model.

Lastly, the use of specific strains of rats have yielded a wealth of knowledge regarding hypertension pathophysiology—most notably the spontaneously hypertensive rat and the Dahl salt-sensitive rat. The Dahl salt-sensitive rat was created through selective breeding of Sprague-Dawley rats whose blood pressure rose when challenged with a high-salt diet (8% NaCl) over multiple generations [148]. When fed a high-salt diet, the Dahl salt-sensitive develops hypertension and concomitant cardiac hypertrophy and cardiac fibrosis. Initially, Dahl salt-sensitive rats develop signs of heart failure and maintain their systolic function, but this strain eventually progresses to overt systolic dysfunction and a HFrEF phenotype [149-151]. Meanwhile, the spontaneously hypertensive rat (SHR) is genetically pre-disposed to the development of hypertension without added stressors [152,153]. Similar to the Dahl salt-sensitive rat, the SHR develops LV hypertrophy and diastolic dysfunction that eventually progresses to systolic dysfunction and falling EF values, although the SHR takes much longer to decompensate (at least one year in the SHR compared with 20 weeks in Dahl salt-sensitive rats) [154-156]. Both of these rat strains may be useful for studying diastolic dysfunction in the context of hypertension, but care must be taken to ensure appropriate timepoints are selected.

Transverse aortic constriction (TAC)

A common model of heart failure, TAC, places increased afterload on the heart in the absence systemic hypertension. TAC models initially lead to cardiac hypertrophy, cardiac fibrosis, and diastolic dysfunction. TAC often progresses to a HFrEF phenotype, with some mouse strains being more susceptible than others [46,157]. The degree of cardiac dysfunction and fibrosis is dependent on the severity of constriction. More mild versions of TAC may better model HFpEF, as they do not display overt systolic dysfunction at 4 weeks [44,45]. Thus, care should be taken with regards to the duration and severity of TAC-induced cardiac dysfunction when interpreting the relevance to HFpEF. Nonetheless, these studies suggest that TAC can recapitulate key aspects of HFpEF pathophysiology if these criteria are well-monitored.

Obesity and diabetes

Metabolic disease is a major component of HFpEF for many individuals, and obesity is an important risk factor for HFpEF. Like hypertension, obesity is a systemic disease that may impact HFpEF pathogenesis through multiple mechanisms via cardiac muscle, skeletal muscle, arterial function, and renal function, among others. Rodent models of obesity and diabetes have been employed to study diastolic dysfunction with varying success. All rodent models of obesity described here are accompanied by insulin resistance, making it difficult to separate the two metabolic conditions.

Pioneering studies in obesity led to the development of two mouse models of genetic obesity based on the hormone leptin: the ob/ob mouse and the db/db mouse. The ob/ob mouse was observed due to a spontaneous nonsense mutation at Jackson Labs in 1949. The db/db mouse was also due to a spontaneous mutation observed at the Jackson Labs in 1966. The mutation lies in a donor splice site that leads to loss of function. Both the ob/ob and db/db are characterized by morbid obesity, hyperglycemia, and insulin resistance [158,159]. Both db/db and ob/ob mice develop cardiac hypertrophy with age with preserved systolic function [160-164]. In ob/ob mice, weight loss via leptin infusion improved cardiac hypertrophy while weight loss via caloric restriction had no effect [161]. This suggests that leptin signaling itself, either directly or indirectly, plays a greater role than increased adiposity for LV remodeling in this model. Intriguingly, neither the db/db or ob/ob mice develop elevations of natriuretic peptides [165,166], a phenomenon which can also be seen in humans in HFpEF due to obesity [165]. Unsurprisingly, both db/db and ob/ob mice display exercise intolerance, but it is unclear to whether this is a reflection of cardiac dysfunction or merely increased adiposity. Additionally, leptin deficiency is a rare cause of human obesity, limiting the translatability of this model. However, both the ob/ob and db/db mouse models can be used to study the obese phenotype of HFpEF.

Multi-hit models

Recent studies have emphasized using multiple hits to induce a HFpEF phenotype, given that many patients present with multiple comorbidities. Variations of this strategy include using pharmacologic inhibition of endothelial nitric oxide synthase with N-nitro-L-arginine methylester (L-NAME) in combination with high-fat diet (HFD) [167,168], aldosterone infusion in the db/db mouse [169], and HFD followed by Ang II infusion [170]. These models combining hypertension and obesity generally induce a HFpEF phenotype with cardiac hypertrophy and diastolic dysfunction. Cardiac fibrosis is not well-described and may differ between the above models. The L-NAME+HFD model was extensively studied to demonstrate that there is cardiac hypertrophy, cardiac fibrosis, reduced myocardial capillary density, impaired exercise tolerance, increased pulmonary congestion, and reduced coronary flow reserve in addition to diastolic dysfunction. Thus, L-NAME+HFD is a well-characterized model of HFpEF that has now been utilized by multiple groups. One drawback of the L-NAME and high fat diet model is that chronic L-NAME administration is not routinely used as a mouse model of hypertension. More often, L-NAME is used to induce endothelial dysfunction and sensitize mice to salt challenge and is used only for 2-3 weeks [171-173]. The L-NAME and high-fat diet model lasts up to 20 weeks with L-NAME administered during the entire protocol. Constant administration of L-NAME continually inhibits endothelial nitric oxide production, thus preventing studies focused on the contribution of endothelial dysfunction to HFpEF. In addition, female mice are largely protected from the L-NAME+HFD model, which is not the case in humans with HFpEF [174]. Notably, the HFD + Ang-II in aged mice model of HFpEF utilized female mice and responded to SGLT2 inhibition, making it a good candidate model for translational studies [170]. The main drawback of this model, however, is that it requires 18- to 22-month-old mice prior to study onset.

The Zucker fatty and spontaneously hypertensive rat (ZSF1) was created by crossing a lean female rat with a mutation in the leptin receptor with the lean male rat harboring a separate leptin receptor mutation and predisposition to spontaneous hypertension. The resulting offspring with both leptin receptor mutations develop elevated blood pressure, obesity, insulin resistance, and hyperglycemia [175]. The ZSF1 rat develops HFpEF within the first 20 weeks of life with concentric LV remodeling, diastolic dysfunction, albuminuria, and pulmonary congestion in the setting of preserved EF [176-179]. Thus, the ZSF1 model is similar to other obese HFpEF models, with the additional contributions of hypertension and renal dysfunction.

A recent study utilized a ‘three-hit’ model by utilizing the L-NAME+HFD model in mice transgenic for low cardiomyocyte levels of a subunit of the L-type Ca²⁺ channel, causing increased Ca²⁺ influx [180]. Their results show that L-NAME+HFD treatment in FVB control mice did induce hypertension but only modest changes in cardiac structure and function, suggesting that aspects of this model may depend on the genetic background, as FVB are considered to be more resistant to metabolic derangements induced by HFD [181].

Combined risk factor models or ‘multi-hit’ models are useful for studying complex HFpEF pathophysiology and may be better suited toward translational studies given the difficulty of isolating the effects of a single risk factor. A major benefit of multi-hit models is increased relevance to the cardiometabolic phenogroup of HFpEF patients in which hypertension, obesity and diabetes are all present.

Inflammation in cardiac remodeling

Like many non-lymphoid organs, the major immune cell type in the heart is the macrophage. Macrophages are a heterogeneous group of phagocytic immune cells with myriad roles in maintaining tissue homeostasis and responding to tissue injury. They are located within the interstitial spaces in direct contact with cardiomyocytes, endothelial cells, and fibroblasts. Often, macrophages are identified by a combination of markers such as CD11b, F4/80, CX3CR1, CD64, or CD68. In addition, CD163 is often used as a marker of tissue resident macrophages. In health, most tissue resident cardiac macrophages resemble alternatively activated anti-inflammatory macrophages [182]. Until the advent of single cell sequencing, we were limited in our ability to fully understand the heterogeneity of tissue macrophages, which adapt to their microenvironment through local cues. In addition, parabiosis studies and new genetic tools to trace the origin of macrophage subpopulations in mice have deepened our understanding of what cells give rise to tissue macrophages in both health and disease.

Elegant work by multiple groups have described two origins of tissue resident cardiac macrophages. One such population is seeded early in development and maintained throughout life by self-renewal [183]. This embryonic-derived population maintains expression of Lyve1, Folr2, and Timd4, and has limited input from monocyte-derived macrophages throughout development [184]. In contrast, macrophages derived from monocytes express Ccr2, at least for a window of time after tissue establishment. These two populations of CCR2+ and CCR2− macrophages have been shown to have differing effects on monocyte recruitment and cardiac function in the context myocardial infarction [185]. There are, however, additional macrophage subpopulations even at steady state, so this framework alone does not fully account for macrophage phenotypes. Embryonic-derived macrophages, however, have been demonstrated to lose their self-renewal capacity over time with monocyte contribution to all cardiac macrophage subsets even in the absence of inflammation [186]. Intriguingly, depletion of all tissue resident macrophages using transient antibody treatment to the CSF-1R (CD115) led to increased fibrosis and decreased angiogenesis in a TAC model [187]. Thus, there is evidence that monocyte-derived macrophages may play differing roles in cardiac remodeling compared with tissue resident macrophages. How macrophage ontogeny contributes to cardiac remodeling in a wider array of cardiac stressors in still under investigation. Moving forward, it may be of greater interest to better define the mechanisms underlying macrophage cell state transitions in the context of cardiac stress to identify potential therapeutic targets rather than focus on macrophage origin.

Inflammation in animal models of HFpEF

While there is limited data on inflammation in mouse models of HFpEF, we have attempted to summarize the evidence linking inflammation and cytokine production to HFpEF induction in animal models. For example, the L-NAME+HFD model of HFpEF is associated with myocardial increases in myocardial expression of IL-1β, IL-6, and TNF-α [167]. In addition, L-NAME + HFD treatment increased numbers of myocardial immune cells (CD45+ cells), total macrophages (CD68+ cells), and macrophages expressing the M2 marker CD206 (CD68+CD206+ cells) [168]. The role of immune cells in this model remains untested. In addition, the HFD + Ang-II model of HFpEF is associated with increased inflammatory biomarkers [170]. The senescence accelerated mouse ages at a rapid rate and develops endothelial dysfunction and increased expression of adhesion molecules in both the heart and aorta, consistent with an inflammatory contribution to diastolic dysfunction [188]. In addition, the ZSF1 rat model of HFpEF is associated with increased expression of endothelial adhesion molecules [189].

The DOCA-salt model is perhaps the most well described model of HFpEF. Anti-inflammatory therapies that block monocyte recruitment and macrophage accumulation also prevented myocardial fibrosis in DOCA-salt [190,191]. There is some evidence to suggest that endothelial dysfunction and activation is critical in inflammation and remodeling, as deletion of the mineralocorticoid receptor in endothelial cells prevented cardiac inflammation and remodeling without affecting blood pressure in DOCA-salt treated mice [192]. The IL-1 receptor antagonist anakinra has been shown to reduce blood pressure and renal fibrosis in DOCA-salt hypertension, but unexpectedly anakinra did not reduce leukocyte infiltration [193]. We recently described the cardiac immune response in DOCA-salt treated animals using single-cell sequencing. We observed increased frequencies of neutrophils, monocytes, and CCR2+ macrophages in addition to increased expression of triggering receptor expressed on myeloid cells 2 (Trem2) in macrophages [194]. Genetic deficiency of Trem2 led to exacerbated cardiac hypertrophy and dysfunction, demonstrating the importance of macrophages in governing local tissue repair responses. Lastly, DOCA-salt treated animals have been shown to upregulate CCR2 ligands in the vasculature. Blockade of CCR2 signaling reduced accumulation of aortic macrophages and reduced blood pressure, highlighting a critical role for macrophages in vascular function [195]. Others have confirmed the potential role of recruited macrophages, as macrophage depletion abrogated blood pressure increases in DOCA-salt treated rats [196]. A study focused on macrophages found that both aging or the aldosterone infusion-uninephrectomy-salt model led to expansion of cardiac macrophage population and associated increases in IL-10 production. Moreover, deletion of IL-10 in macrophage improved indices of diastolic dysfunction [197]. Thus, cardiovascular remodeling and diastolic dysfunction in the DOCA-salt model have clear inflammatory components.

In summary, while there is an extensive literature on the dynamics and signaling of immune cells post-myocardial infarction and within atherosclerosis, there is a paucity of literature on the role of immune cells during chronic cardiovascular remodeling. Both innate and adaptive immune cells have been implicated in preclinical models of acute cardiovascular remodeling, hypertension, obesity, and renal dysfunction. Thus, it stands to reason that they may play important roles in HFpEF (Figure 5).

Figure 5. Summary of immune alterations in human HFpEF.

Plasma proteomic studies have demonstrated increases in circulating pro-inflammatory cytokines such as IL-1β, IL-6, and TNF-α as well as other markers of inflammation such as soluble ST2 (sST2), galectin-3, and pentraxin-3 (PTX-3). An imbalance of Th17 and T regulatory (Treg) cells have also been demonstrated in HFpEF. Lastly, biopsy studies have shown that there are increases in macrophages and T cells in the hearts of patients with HFpEF compared with controls. Cardiac inflammation is also associated with interstitial fibrosis in HFpEF. Figure made with Biorender.

Figure 3. Proposed role of inflammation in heart failure subtypes.

Both HFpEF and HFrEF are characterized by cardiac dysfunction, remodeling, and inflammation. We propose that inflammation is an inciting even in HFpEF pathogenesis but a reactive event in HFrEF. Figure made with Biorender.

Abbreviations

Ang-II

angiotensin II

BMI

body mass index

CAD

coronary artery disease

cGMP

cyclic guanosine monophosphate

CKD

chronic Kidney Disease

COPD

chronic obstructive pulmonary disease

CCR2

CC motif chemokine receptor 2

CRP

C reative protein

CXCR3

CXC motif chemokiine receptor 3

DOCA

deoxycorticosterone acetate

HF

heart failure

HFD

high-fat diet

HFpEF

heart failure with preserved ejection fraction

HFrEF

heart failure with reduced ejection fraction

IL

Interleukin

L-NAME

N-nitro-L-arginine methylester

LV

left ventricle

NT-proBNP

N-terminal pro-B type natriuretic peptide

PCWP

pulmonary capillary wedge pressure

PDE

phosphodiesterase

PKG

protein kinase G

PTX-3

pentraxin-3

RAAS

renin–angiotensin aldosterone system

sGC

Soluble guanylate cyclase

SGLT2i

sodium-glucose cotransporter 2 inhibitors

SHR

spontaneously hypertensive rat

T2DM

type 2 diabetes mellitus

TAC

transverse aortic constriction

TGF-β

transforming growth factor β

TNF-α

tumor necrosis factor α

Trem2

triggering receptor expressed on myeloid cells 2

VCAM-1

vascular cellular adhesion molecule-1

ZSF1

Zucker fatty and spontaneously hypertensive rat.