INNATE IMMUNITY AND THE FAILING HEART: THE CYTOKINE HYPOTHESIS REVISITED

Abstract

Elevated levels of inflammatory mediators have been identified in patients with heart failure, including heart failure with reduced and preserved ejection fraction, as well as acute decompensated heart failure. Moreover, experimental studies have shown repeatedly that activation of inflammation in the heart provokes left ventricular (LV) remodeling and LV dysfunction. Nonetheless, phase III clinical trials that have attempted to antagonize inflammatory mediators have been negative with respect to the primary end points of the trials, and in some patients, resulted in worsening heart failure and/or death. The following review will discuss how recent developments in the field of innate immunity have advanced our understanding of the role of inflammation in the pathogenesis of heart failure and will discuss the negative outcomes of the existing clinical trials in light of this new information.

“First get your facts, then you can distort them at your leisure.”

Mark Twain

The link between heart failure and inflammation was first recognized in 1990 by Levine et al., ¹ who reported elevated levels of tumor necrosis factor (TNF) in heart failure patients with a reduced ejection fraction. Since this original report, there has been an exponential rise in the number of cytokines and chemokines that have been identified in the setting of heart failure with a reduced ejection fraction. Elevated levels of inflammatory mediators have also been identified in acute decompensated heart failure, as well as in heart failure patients with a preserved ejection. Thus, there is evidence of an ongoing inflammatory response in all of the manifestations of clinical heart failure.

The early clinical observations with respect to TNF prompted a series of experimental studies, which demonstrated that the sustained expression of TNF at levels that were observed in heart failure patients was sufficient to provoke left ventricular (LV) dysfunction and LV remodeling.² These and other pre-clinical studies formed the basis for several multicenter clinical trials that utilized “targeted” approaches to neutralize TNF in patients with moderate to advanced heart failure. As reported by the author in Circulation Research over a decade ago,³ the targeted anti-TNF approaches were negative with respect to the primary end points of the trial and/or resulted in worsening heart failure and/or death.^4,5 Over the years, the ensuing debate over the negative outcome of these clinical trials has produced more questions than answers with respect to what role, if any, pro-inflammatory cytokines play in the pathogenesis of heart failure. One of the untoward consequences of the negative outcomes of these trials is that that they have a profound chilling effect on further attempts to target inflammation in heart failure. Fortunately, over the past ten years there has been a much clearer appreciation of the importance of inflammation in the heart because of the pioneering efforts in the field of innate immunity by Charles Janeway (1943-2003) and Ruslan Medzhitov, as well as Bruce Beutler, Jules Hoffman, and Ralph Steinman who shared the Nobel Prize in Physiology/Medicine in 2011 for their work in innate immunity. In the following review, we will discuss how recent developments in the field of innate immunity have advanced our understanding of the role of inflammation in the pathogenesis of heart failure, and we will utilize these new insights to reevaluate the clinical trials that have been conducted in this area.

OVERVIEW OF IMMUNE RESPONSES IN THE HEART

Both innate and adaptive immune responses are activated in the heart in response to tissue injury that results from pathogens or environmental injury (e.g. ischemia or hemodynamic overloading). Whereas the innate immune system provides a global, non-specific defense against pathogens and/or tissue injury, the adaptive immune system provides a highly specific response that is mediated by B and T cells. Studies have shown that the ensuing inflammatory response induced by the innate immune system can be physiologic and result in the upregulation of a portfolio of cytoprotective responses that provide the heart with a short term adaptation to the stress (reviewed in ⁶). Alternatively, the inflammatory response can become dysregulated (i.e., pathophysiologic), leading to collateral myocardial damage that eventuates in progressive left ventricular (LV) dysfunction and adverse LV remodeling. Although Ilya Metchnikoff first proposed in 1901 that the immune system had both physiological and pathophysiological roles,⁷ it has been challenging conceptually, to reconcile these two vastly different facets of the immunological response to tissue injury. The relatively recent insight into the role of the innate immune responses in the heart has permitted a clearer understanding of physiologic vs. pathophysiologic inflammation, as well as a nascent understanding of the biology underlying the development of the chronic inflammation that arises in dysfunctional tissue, which has been referred to a para-inflammation.⁸ As will be discussed below, the concept of para-inflammation likely has direct bearing on the inability to successfully target inflammation in the setting of heart failure.

Cardiac innate immune responses, which are essential for homeostatic responses and tissue repair, are initiated by the detection of pathogen associated molecular patterns (PAMPs) or damage associated molecular patterns (DAMPs) by a fixed number of germ-line encoded pattern recognition receptors (PRRs). Classic examples of PAMPs include the lipopolysaccharides (LPS) of Gram-negative organisms, the teichoic acids of Gram positive organisms, the zymosans of yeast, the glycolipids of mycobacterium, or the double-stranded RNAs of viruses. More recently, it has become clear that cardiac PRRs also recognize the molecular patterns of endogenous host material released by dying or injured myocardial cells. Cells that die by accidental necrosis, regulated necrosis (necroptosis), and/or secondary apoptosis release their cytosolic contents into the extracellular space, thereby initiating a brisk inflammatory response through engagement of an ensemble of extracellular or intracellular PRRs. The time course of the inflammatory response that ensues following tissue injury is remarkably consistent, irrespective of the specific cause of cell injury, and is associated with the rapid influx of neutrophils and subsequently monocytes into the area of tissue injury. This inflammatory response has been referred to as “sterile inflammation,” insofar as the inflammation following tissue injury occurs in the absence of a known pathogenic infection. Thus, the innate immune system evolved not only to detect molecules that were non-self (e.g. PAMPs), but also to detect a subset of intracellular molecules (e.g. DAMPs) that were hidden by the plasma membrane (hidden-self) and not ordinarily found in extracellular fluids in the absence of cell death. This latter observation has provided a potentially important link between tissue injury, activation of pro-inflammatory mediators, and the resulting myocardial response to stress.

Many PRRs encountering PAMPs and DAMPs trigger signaling cascades that activate nuclear factor-κB (NF-κB), activator protein 1 (AP1), and interferon regulatory factor (IRF) transcription factors, that in turn regulate target genes that encode pro-inflammatory cytokines and interferons in the heart.⁹ The portfolio of cytokines implicated in the pathogenesis of heart failure has been the subject of numerous reviews and is summarized in the online supplement. Another subset of PRRs in the heart trigger a distinct pro-inflammatory mechanism that requires assembly of cytosolic protein complexes called inflammasomes.¹⁰ Canonical inflammasomes convert procaspase-1 into the catalytically active protease that is responsible for the production of IL-1β and IL-18, which are sufficient to trigger inflammatory responses in the heart.

PRRs can be subdivided into two major classes based on their subcellular localization. Toll-like receptors (TLRs) and C-type lectin receptors (CLRs) are found on plasma membranes or endosomes, where they can detect the presence of PAMPs or DAMPs. Messenger RNAs for TLRs 1 – 10 have been identified in the human heart. ¹¹ The relative expression levels for TLR mRNAs in the human heart is: TLR4 > TLR2 > TLR3 >TLR5 > TLR1 > TLR6 > TLR 7 > TLR8 > TLR9 > TLR10.¹¹ Although expression levels of TLRs have not been identified in human myocytes, TLR2, 3, 4, 6 mRNA has been identified in cardiac myocytes from neonatal rats.¹² Although little is known with regard to the regulation of TLR expression in the heart, TLR4 appears to be upregulated in the failing human heart.^13,14 Moreover, TLR2 and TLR4 have been shown to have a profound effect on cardiac remodeling in the context of ischemia reperfusion injury and myocardial infarction (reviewed in reference ⁹). The signaling pathways used by the TLRs ae summarized in the online supplement. CLRs are calcium-dependent carbohydrate-binding receptors that contain one or more C-type lectin-like domains. While various members of the CLR family have been demonstrated to play an important role in immune responses, very little is known with respect to the role of CLRs in the heart. Relevant to this discussion, CLRs are able to signal independently, as well as modulate the signaling through the TLRs. The relative expression levels of CLR mRNAs in the human heart, which is similar to the murine heart, is: Bcl-10 > Galectin-1 > mannose receptor (MR) 2 > DC-SIGN (CD209) > Src, > MR1, > Dectin-1, triggering receptor expressed on myeloid cells (TREM)-1 and Card-9.¹⁵

A second class of PRRs resides in intracellular compartments and includes RIG (retinoic acid inducible gene)-I-like receptors, also called RLRs, nucleotide binding and oligomerization domain (NOD) like receptors (NLRs) and absent-in-melanoma (AIM) 2 receptors. ^16,17 NLRs act as cytosolic sensors to intracellular DAMPs and PAMPs. In humans, the NLR family is composed of 22 intracellular pattern recognition molecules that share a central NACHT domain (domain present in NAIP, CIITA, HET-E and TP1) and a carboxy-terminal leucine rich repeat (LRR) region.¹⁸ Analysis of human heart tissues has revealed that nucleotide-binding and oligomerization domain (NOD2 [NLRC2]), NOD1 [NLRC1] and NLR family, pyrin domain-containing protein 2 (NLRP2 [NALP2]), NLRP3 [NALP3], also known as cyropyrin, are expressed. Both NOD1 and NLRP3 have been shown to activate canonical inflammasomes in the heart, and play an important role in adverse cardiac remodeling following ischemia reperfusion injury and myocardial infarction.^10,19 The retinoic acid inducible (RIG)-I-like receptor (RLR) family is composed of RIG-I, melanoma differentiation-associated gene 5 (MDA5), and LGP2. RLRs are localized in the cytoplasm and recognize the genomic RNA of dsRNA viruses and dsRNA generated as the replication intermediate of ssRNA viruses. The expression of RLRs is greatly enhanced in response to type I IFN stimulation or virus infection. At the time of this writing, very little is known with respect to the RLRs in heart, although a recent study has shown that MDA5 may play an important role in protecting the heart from direct viral injury during myocarditis.

When myocardial inflammation is induced by microbial or non-microbial sources, the primary purpose of the inflammatory response is to resolve the source of the disturbance, thereby allowing the heart to adapt to the abnormal conditions in the short-term, and ultimately to restore homeostasis and cardiovascular function in the long-term. If the abnormal conditions are sustained, then an ongoing inflammatory state persists in the tissue and leads to a state of chronic low grade inflammation, which can contribute to further disease progression by virtue of the deleterious effects of sustained inflammation. Although speculative, it is very likely that the chronic expression of pro-inflammatory cytokines and ongoing inflammation that have been demonstrated in the failing heart²⁰ represent the inability of the myocardium to restore homeostasis, leading to a state of ongoing chronic inflammation that is intermediate between the baseline state and acute inflammation. This intermediate state has been termed para-inflammation,⁸ and does not require overt tissue injury or infection to be sustained, but instead represents a graded sustained inflammatory response that remains “switched on” in dysfunctional tissue in an attempt to restore homeostasis and tissue functionality (see Figure 2). It should also be noted that activation of neurohormonal systems in heart failure, such as the renin-angiotensin-aldosterone system and the adrenergic nervous system, are capable of triggering inflammation in the heart, thereby leading to a state of low grade inflammation.^21,22 Unfortunately, at the time of this writing, it is not known whether down-modulating the level of para-inflammation in the failing heart to prevent collateral damage can be accomplished without disrupting critical homeostatic responses that are provided by low levels of inflammation. Given that para-inflammatory responses are graded, ranging from physiologic levels of inflammation to a classic inflammatory response, it will be critically important to have a better understanding of when it is appropriate to target inflammation in the failing heart.

Figure 2.

Para-inflammation. The primary purpose of the inflammatory response in the heart is to resolve tissue injury, thereby allowing the heart to adapt to the abnormal conditions in the short-term, and ultimately to restore homeostasis and cardiovascular function in the long-term. If the abnormal condition is sustained, then a chronic inflammatory state persists in a tissue, which is referred to as para-inflammation. Para-inflammation is a graded response that can restore tissue homeostasis, or if sustained can contribute to further disease progression, by virtue of the deleterious effects of sustained inflammation on cardiac myocytes and the extracellular matrix.

ROLE OF INFLAMMATION IN THE PATHOGENESIS OF HEART FAILURE

The primary interest in deciphering the role of inflammation in heart failure arose from the observation that many of the biological effects of pro-inflammatory cytokines were sufficient to provoke a heart failure phenotype in experimental animals and in humans. The “cytokine hypothesis”²³ for heart failure postulates that heart failure progresses, at least in part, as a result of the deleterious effects exerted by endogenous cytokine cascades on the heart and the peripheral circulation. Thus, analogous to sustained neurohormonal activation in heart failure, the chronic inflammation that occurs in heart failure may also contribute to worsening heart failure by virtue of the harmful effects of sustained inflammatory signaling. It bears emphasis that at the time the cytokine hypothesis was proposed, the role of the innate immune system, as well as the concept that chronic inflammation was both beneficial and deleterious, were not at all well understood. Thus, although the biological underpinning for the cytokine hypothesis remains unchanged, the optimal approach to testing this hypothesis in heart failure patients is far less certain because of the inherent complexity of chronic para-inflammation.

Effects of Cytokines on Left Ventricular Function

The pathophysiological effects of pro-inflammatory cytokines have been reviewed extensively, ^3,24 and will only be discussed here briefly. Pro-inflammatory cytokines were first shown to provoke LV dysfunction in the systemic inflammatory response that occurs during sepsis. Direct injections of TNF were shown to produce hypotension and rapid death within minutes, whereas injections of anti-TNF antibodies attenuated the hemodynamic collapse that occurs during endotoxin shock. Subsequent studies in dogs and rats showed that circulating levels of TNF produced negative inotropic effects in vivo and in vitro (reviewed in reference ³). More recent studies in transgenic mice with cardiac restricted overexpression of TNF showed that forced overexpression of TNF resulted in depressed LV ejection performance that was dependent on TNF “gene dosage”.²⁵

With respect to the potential mechanisms for the deleterious effects of TNF on LV function, the literature suggests that TNF modulates myocardial function through an immediate pathway that is manifest within minutes and is mediated by activation of the neutral sphingomyelinase pathway. This is followed by a delayed response that requires hours to days to develop and is mediated by nitric oxide mediated blunting of β-adrenergic signaling (reviewed in reference ³). Whereas the negative inotropic effects of IL-1 appear to be mediated, at least in part, through the production of nitric oxide (i.e., the delayed pathway), the negative inotropic effects of IL-6 are less well understood. Recent studies suggested that TNF and IL-1 may produce negative inotropic effects indirectly through activation and/or release of IL-18. Remarkably, blockade of IL-18 using neutralizing IL-18 binding protein leads to an improvement in myocardial contractility in atrial tissue following ischemia reperfusion injury.²⁶ Although the signaling pathways that are responsible for the IL-18 induced negative inotropic effects have not been delineated thus far, it is likely that they will overlap those for IL-1, given that the IL-18 receptor complex utilizes components of the IL-1 signaling chain.

Effects of Pro-inflammatory Cytokines on Left Ventricular Remodeling

LV ventricular remodeling refers to the multitude of changes that occur in cardiac shape, size, and composition in response to myocardial injury. Inflammatory mediators have a number of important biological effects that may play an important role in the process of LV remodeling, including cardiac myocyte hypertrophy, alterations in fetal gene expression, activation of collagenolytic matrix metalloproteinases (MMPs), myocardial fibrosis, as well as progressive myocyte loss through apoptosis.³ Antagonism of innate immune receptors (TLR2, TLR4), innate immune signaling pathways (MyD88, IRAK-1, IRAK-4, and NLRP3) and the pro-inflammatory cytokines downstream from these pathways (TNF, IL-1β, IL-18) has been shown to attenuate adverse LV remodeling following acute myocardial infarction (reviewed in reference ²⁷). Studies in chimeric mice, wherein it has been possible to separate the role of innate immune signaling in cells derived from the bone marrow from the effects in the myocardium have demonstrated that activation of innate immune signaling pathways in bone marrow derived neutrophils and monocytes contributes to tissue damage, progressive fibrosis and adverse cardiac remodeling, whereas activation of the same pathways in cardiac myocytes is beneficial through short-term mitochondrial stabilization, enhanced sarcolemma membrane integrity,²⁸ and through conservation of energy secondary to the development of reversible left ventricular dysfunction (reviewed in reference ²⁷). Studies in experimental models wherein the inflammatory signaling is sustained have also provided important insights into the mechanisms for inflammation-induced adverse LV remodeling. For example, a study in rats showed that infusion of concentrations of TNF that overlap those observed in heart failure patients led to a time- dependent change in LV dimension that was associated with progressive degradation of the extracellular matrix.² Studies in transgenic mice with targeted overexpression of TNF have shown that these mice develop progressive LV dilation, and that TNF-induced activation of MMPs is responsible for collagen degradation and progressive LV dilation.²⁹ These studies demonstrated that sustained myocardial inflammation leads to temporal changes in the balance between MMP activity tissue inhibitor of matrix metalloproteinases (TIMPs) and mast cell-mediated TGF-β signaling.³⁰ Collectively, these time-dependent changes favor degradation of the extracellular matrix during the onset of inflammation and progressive myocardial fibrosis following sustained inflammation. Thus, the sustained activation of inflammatory signaling contributes to LV remodeling through a variety of different mechanisms that involve both the myocyte and non-myocyte components of the myocardium.

CLINICAL APPLICATIONS

Inflammatory Biomarkers

The extant literature suggests that inflammatory biomarkers provide important diagnostic and prognostic information across the entire spectrum of heart failure syndromes (reviewed in reference ³¹). Table 1 shows that the pro-inflammatory cytokines that are elaborated in heart failure include members of the tumor necrosis factor (TNF) superfamily (TNSF), members of the interleukin -1 (IL-1) family (IL1-F), and IL-6.³¹ Soluble ST2 (sST2), which is the receptor for IL-33 and is thus a member of the IL-1superfamily (IL-1F) of cytokines, is the first inflammatory biomarker to be approved by the Food and Drug Administration (FDA) for prognosis in heart failure. Importantly, soluble ST2 is secreted by cultured myocytes that are subjected to mechanical strain and is thus an integrated marker of mechanical strain and inflammation. In addition to cytokines and cytokine receptors, a number of inflammatory mediators that were originally identified in immune cells, most notably macrophages, have also been observed in patients with heart failure. The inflammatory mediators in this group that have garnered the most attention in heart failure include galectin-3 and pentraxin-3. Galectin-3, a member of the lectin family, is released by macrophages in response to tissue injury, as well as by damaged and/or dying cells. Galectin-3 is also approved by the FDA as a biomarker for determining heart failure prognosis. Pentraxin-3, a novel inflammatory marker and member of pentraxin superfamily of cytokines, has also recently been identified in patients with heart failure.³² In addition to providing information regarding patient prognosis, the measurement of inflammatory biomarkers in heart failure patients may identify subsets of patients who are most likely to benefit from anti-inflammatory strategies.

Table 1.

Inflammatory Biomarkers in Heart Failure

HFrEF	HFpEF	ADHF
Pro-inflammatory Cytokines
TNF (TNSF2), TWEAK (TNSF12), FasL (TNFSF6), LIGHT (TNSF14), IL-1β (IL- 1F2), IL-2,IL-6, IL-18 (IL-1F8), IL-33 (IL- 1F11)	TNF (TNSF2), IL-6 (?),	TNF (TNSF2), IL-6, IL-18
Cytokine Receptors
sTNFR1 (TNFRSF1A), sTNFR2 TNFRSF1B), gp130 (IL6ST); IL-1ra (IL1F3), sST2 (IL-1RL1)	sST2 (IL1RL1)	sST2 (IL1RL1)
Macrophage
Galectin-3, Pentraxin-3	Galectin-3, Pentraxin-3	Galectin-3, Pentraxin-3

The parenthesis denote the nomenclature for the TNF and IL-1 superfamily of cytokines and cytokine receptors.

Key: FasL = Fas ligand; LIGHT = Homologous to lymphotoxins, inducible expression, competes with HSV glycoprotein D for HVEM, a receptor expressed on T-lymphocytes; gp130 = soluble gp130; IL-1β = interleukin-1β, IL-2 = interleukin 2, IL-6 = interleukin 6; IL-18 = interleukin 18; IL-33 = interleukin 33; IL1-F = interleukin-1 family; IL-1RL1 = interleukin1-receptor-like-1; sST2 = soluble ST2 receptor; TNF = tumor necrosis factor, sTNFR1 = soluble TNF type 1 receptor; sTNFR2 = soluble TNF receptor type 2; TNFSF = tumor necrosis factor superfamily; TNFSFR = tumor necrosis factor superfamily receptor; TWEAK = TNF-like weak inducer of apoptosis; ? = conflicting data.

(Reproduced with permission from Hartupee J, Mann DL. Positioning of Inflammatory Biomarkers in the Heart Failure Landscape. J Cardiovasc Transl Res 2013;6:485-92)

INFLAMMATION AS A THERAPEUTIC TARGET IN HEART FAILURE

Given that elevated levels of pro-inflammatory cytokines mimic many aspects of the heart failure phenotype and that the deleterious effects of inflammatory mediators are potentially reversible once inflammation subsides, investigators have used a variety of different approaches to antagonize inflammatory mediators in heart failure (Table 2). These fall into one of three broad categories: anti-inflammatory therapies, immunomodulatory therapies, and autoimmune strategies.

TABLE 2.

Randomized Anti-Cytokine Clinical Trials in Heart Failure
Study	Number Patients	NYHA Class	Agent	Category	Follow-up (months)	Mean Age	Mean LVEF	% ACE- ARB/BB	Primary endpoint	Outcome
ATTACH5	150	II-IV III, IV	Infliximab	DCM, IHD	7	61	24	100/73	Clinical composite score	High dose had adverse effect on clinical outcomes
CORONA56	5011	II-IV	rosuvastatin	IHD	32.8	73	ns	91/75	CV death, non- fatal MI and stroke	No effect on CV death, non-fatal MI and stroke, decreased HF hospitalizations
EXACT84	253	II-IV	allopurinal	DCM,IHD	6	63	25	85/95	Clinical composite score	No effect on clinical composite score
GISSI-HF58	4574	II-IV	rosuvastatin	DCM, IHD	46.8	68	33*	95/62	Death, Death and CV hospitalization	No effect on Death, Death and CV hospitalization
Gullestad et al.40	56	II,III	Thalidomide	DCM, IHD	3	66	25	100/91	LVEF, LV volumes, symptoms	Improved LVEF and LV remodeling
Hare et al62	405	III,IV	Oxypurinol	DCM, IHD		65	26	95/01	Composite of HF mortality + morbidity + QoL	No overall effect; effect in those with elevated uric acid
Parillo et al.41	102/RCT	ns	Prednisone	DCM	3	43	17	na/na	LVEF	Improved LVEF
RECOVER/ RENAISSANCE/ RECOVER4	1500	II-IV	Etanercept	DCM, IHD	5.7/ 12.9	63	23	98/62	Clinical composite score/ Death, or heart failure hospitalization	No effect on clinical status, death, or heart failure hospitalization
Skudicky et al.34	39	II, III	Pentoxifylline	DCM	6	49	24	100/100	NYHA class, exercise tolerance, and LVEF	Improved symptoms and LVEF
Sliwa et al.33	28	II, III	Pentoxifylline	DCM	6	53	24	100/na	NYHA class and LVEF	Improved symptoms and LVEF
Sliwa et al.85	18	IV	Pentoxifylline	DCM	1	46	15	100/na	Symptoms, cytokines, and LVEF	Improved symptoms and LVEF
Sliwa et al.36	38	I-IV	Pentoxifylline	IHD	6	55	25	100/100	NYHA class and LVEF	Improved symptoms and LVEF
UNIVERSE55	87	II-IV	Rosuvastatin	DCM, IHD	6.5	62	29	98/85	LVEF	No effect on LVEF
Randomized Immunomodulation Clinical Trials in Heart Failure
ACCLAIM76	2426	II-IV	Celacade	DCM, IHD	10.2	64	23	94/87	Death, or CV hospitalization	No effect on death, or CV hospitalization
Gullestad et al.66	40	II-IV	IVIG	DCM, IHD	6	61	27	100/75	NYHA class, and LVEF	Improved clinical status and LVEF
IMAC67	62/RCT	I-IV	IVIG	DCM	12	43	25	90/18	LVEF and symptoms	No effect
METIS72	50/RCT	II-IV	Methotrexate	IHD	3	59	35	85/84	6MWT	No effect

Key: ACE: angiotensin converting enzyme; ARB: angiotensin II receptor blocker; BB: beta-adrenergic receptor blocker; IVIG: intravenous immunoglobulin; LVEF: left ventricular ejection fraction; mo: months; na: not available; ns: not specified, NYHA: New York heart Association; QoL; quality of life

^*10% of the patients in GISSI-HF has an EF > 40%

Modified from Aukrust Gullestad L, Ueland T, Vinge LE, Finsen A, Yndestad A, Aukrust P: Inflammatory cytokines in heart failure: mediators and markers. Cardiology 2012;122:23-35)

Anti-Inflammatory Therapies

The biological effects of pro-inflammatory mediators can be antagonized through transcriptional or translational approaches, or by so-called “biological response modifiers” that bind and/or neutralize soluble mediators (e.g. TNF or IL-1β). Many of these strategies have been explored in phase II-III clinical trials, as described below.

Transcriptional Suppression of Pro-inflammatory Cytokines

Pentoxyfilline is a xanthine-derived agent that is known to inhibit TNF transcription and translation, as well as modulate a broad spectrum of inflammatory mediators. Pentoxyfilline has been studied in a number of small randomized trials in patients with ischemic and dilated cardiomyopathy (Table 2).^33-36 Treatment with pentoxifylline resulted in a significant improvement in NYHA functional class and/or LV ejection fraction in each of these studies. Importantly, the beneficial effects were seen in all NYHA classes of heart failure, in patients with ischemic and nonischemic cardiomyopathy, and in patients treated with ACE inhibitors and beta-blockers. Apposite to the present discussion, the beneficial effects on cardiac function in some of studies were accompanied by decreased circulating plasma levels of TNF.³⁷ Given that pentoxifylline is a nonspecific phosphodiesterase inhibitor, it is possible that the salutary effects of this agent might be unrelated to its anti-inflammatory properties.

Thalidomide (α -N-pthalimidoglutarimide) suppresses TNF production, as well as the production of a spectrum of inflammatory mediators that are implicated in the pathogenesis of heart failure. The mechanism of action of thalidomide with respect to reducing TNF levels appears to be through enhancing mRNA degradation;³⁸ however, the precise mechanism of action of thalidomide is unclear, and contradictory results have been reported regarding its effects on cytokine levels in vivo. Thalidomide was safe and potentially effective in a small open-label dose escalation study in patients with heart failure. There was a significant increase in the 6- minute walk distance and a trend (P = 0.16) toward improvement in LV ejection fraction and quality of life after 12 weeks of maintenance therapy with thalidomide.³⁹ However, dose-limiting toxicity was observed in two patients (50 and 200 mg/day). In a larger placebo-controlled trial of 56 patients with NYHA class II-III heart failure secondary to ischemic and nonischemic cardiomyopathy and an LV ejection fraction ≤ 40% , treatment with up to 200 mg/day of thalidomide for 12 weeks resulted in increased LV ejection fraction and decreased LV end-diastolic volume.⁴⁰ These salutary changes were accompanied by a decrease in circulating levels of matrix metalloproteinase 2, but an increase in circulating levels of TNF. The effect of thalidomide on LV ejection fraction was observed to a greater degree in patients with dilated cardiomyopathy who were able to tolerate higher doses of thalidomide.⁴⁰

Translational Suppression of Pro-inflammatory Cytokines

Dexamethasone is thought to suppress TNF biosynthesis at the translational level but may also block TNF biosynthesis at the transcriptional level. In an early study, Parrillo and colleagues⁴¹ randomized 102 patients with dilated cardiomyopathy to treatment with prednisone (60 mg per day) or placebo. After three months of therapy, these investigators observed a ≥ 5% increase in EF in ~ 50 % of the prednisone treated patients, whereas ~ 25% of the controls had a significant improvement in LV ejection fraction (p = 0.005). However, the mean increase in LV ejection fraction was not significantly (p = 0.054) different in the prednisone treated group (4.3 ± 1.5 %) when compared to controls (2.1 ± 0.8 %). When patients were divided into a “reactive” group (pre-specified as a fibroblastic/lymphocytic infiltration or immunoglobulin deposition on endomyocardial biopsy, a positive gallium scan, or an elevated erythrocyte and non-reactive) and a “non-reactive” group, the authors noted that ~ 65% of “reactive” patients had an improved LV ejection fraction at 3 months, whereas ~ 25% of the “reactive” control patients had an improved LV ejection fraction (p = 0.004). The prednisone-treated nonreactive patients did not have significantly improved LV function (p = 0.51). This study was the first to demonstrate that patients with dilated cardiomyopathy benefit clinically from an anti-inflammatory therapy.

Targeted anti-cytokine approaches using biological response modifiers

Two different targeted approaches have been taken to selectively antagonize pro-inflammatory cytokines in the setting of heart failure (Table 2). The first approach employed a genetically engineered TNF receptor (etanercept) that acts as a “decoy” to prevent TNF from binding to its TNF receptors on target cells, whereas the second approach employed a chimeric monoclonal antibody that neutralizes circulating TNF.

Soluble TNF receptors

Etanercept (Enbrel™) is a genetically engineered humanized protein consisting of two human TNF p75 receptors coupled to a human IgG₁:Fc fragment. Two small short-term studies in patients with stable heart failure showed that treatment with 25 mg biw etanercept resulted in improved quality of life, increased 6 minute walk distance and improved LV ejection performance after 3 months of treatment .³ These early trials formed the basis for two moderate size multicenter trials that employed parallel study designs, but differed in the dose of etanercept that was used. The Randomized Etanercept North AmerIcan Strategy to Study AntagoNism of CytokinEs (RENAISSANCE; n = 900) trial was conducted in North America, whereas the Research into Etanercept Cytokine Antagonism in Ventricular Dysfunction (RECOVER; n = 900) trial was held in Europe and Australia. The primary end point for both trials was a clinical composite score in which patients were classified as improved, unchanged or worsened at 24 weeks. In the RENAISSANCE study, patients were treated with placebo or subcutaneous etanercept 25 mg biw or 25 mg tiw, whereas the RECOVER trial employed doses of 25 mg qw or 25 mg biw of subcutaneous etanercept. A third pre-specified trial, the Randomized Etanercept Worldwide Evaluation (RENEWAL; n = 1500) trial, utilized the pooled data from the RENAISSANCE (biw and tiw dosing) and RECOVER (biw dosing only). The primary end point for RENEWAL was all cause mortality and hospitalization for heart failure. On the basis of pre-specified stopping rules, both trials were terminated prematurely because of the lack of benefit of etanercept on the clinical composite in RENAISSANCE (p=0.17) and RECOVER (p=0.34) (Figure 3A). The pre-specified analysis of RENEWAL showed that there was no effect of etanercept on the primary end point (Figure 3B) of death or chronic heart failure hospitalization (hazard ratio =1.1, 95% CI 0.91 to 1.33, P=0.33).⁴ However, in a post-hoc analysis of the RENAISSANCE trial, patients receiving biw and tiw etanercept experienced, respectively, an increased 1.21 (p = 0.17) and 1.23 (p = 0.13) risk of death/heart failure hospitalization when compared with placebo. Further analysis of the components of the clinical composite score in the RENAISSANCE trial indicated that there was a significantly greater proportion of etanercept treated patients (29%, p < 0.04) in the worsened category at 24 weeks when compared to placebo treated patients (20%). Increases in the risk of death/heart failure hospitalization and a worsening clinical composite were not observed in RECOVER, wherein the dose and duration of etanercept dosing was less. Patients in RECOVER received etanercept for a median time of 5.7 months, whereas patients in RENAISSANCE received etancercept for 12.7 months. On the basis of these findings, the prescribing information for etanercept has been updated and now suggests that physicians exercise caution in the use of etanercept in patients with heart failure.

Figure 3.

Results of the RENAISSANCE, RECOVER AND RENEWAL trials. (A) Analysis of the “clinical status” composite score for the RECOVER and RENAISSANCE trials in the placebo and etanercept groups. (B) Kaplan-Meier analysis of the time to death or heart failure hospitalizations in the placebo and etanercept group (biw and tiw) in the RENEWAL analysis. (Reproduced with permission Mann DL. Activation of inflammatory mediators in heart failure. In: Mann DL, ed. Heart failure: A companion to Braunwald’s heart disease. 2nd ed. Philadelphia: Elsevier/Saunders; 2011. p. 163-84)

Although the precise explanation for the worsening heart failure in the RENAISSANCE is not known, it bears emphasis that TNF receptor antagonists have intrinsic biological activity and, in certain settings, can act as agonists (referred to as a stimulating antagonist ⁴²). We and others have reported that in some settings etanercept can stabilize TNF and increase its bioactivity (see reference ³ for further discussion). Although the stabilizing effects of etanercept might not be problematic in rheumatoid arthritis, wherein TNF is encapsulated within a joint space and peripheral circulating TNF levels are relatively low (compared to heart failure) or are nonexistent, it is possible that an increase in the circulating levels of biologically active TNF in heart failure patients might contribute to worsening heart failure.

Monoclonal Antibodies

Infliximab (Remicade™) is a chimeric monoclonal antibody consisting of a genetically engineered anti-TNF murine Fab fragment fused to a human FC portion of human IgG₁. Although infliximab had been shown to be effective in Crohn’s disease and rheumatoid arthritis, infliximab had never been tested in pre-clinical nor early phase I clinical studies in heart failure patients. The Anti-TNFα Therapy Against CHF (ATTACH) trial was a phase II study in 150 patients with moderate to advanced heart failure (NYHA class III,IV). The primary end-point of the ATTACH trial was the clinical composite score that was also employed in RENAISSANCE and RECOVER.⁴³ Patients were randomized to receive three separate intravenous infusions of infliximab (5 mg/kg or 10 mg/kg) at baseline and at 2 and 4 weeks, followed by an assessment of the clinical composite score at 14 and 28 weeks. Because ATTACH was a pilot phase II study, there was no requirement for a formal Data Safety Monitoring Board to monitor ongoing clinical outcomes during the trial. Analysis of the completed data set revealed that there were increased rates of mortality and heart failure hospitalization, particularly in the group who was receiving the highest dose of infliximab (Fig 4A). On the basis of these findings, the prescribing information for infliximab has been changed and it is now recommended that treatment with infliximab be discontinued in patients with worsening heart failure and that infliximab treatment should not be initiated in patients with heart failure.

Figure 4.

Results of the ATTACH trial. A, Kaplan-Meier rates of death and hospitalization for heart failure. B, Kaplan-Meier rates of hospitalization for any reason. (Key: PBO = placebo; HR = hazard ratio). (Reproduced with permission Mann DL. Activation of inflammatory mediators in heart failure. In: Mann DL, ed. Heart failure: A companion to Braunwald’s heart disease. 2nd ed. Philadelphia: Elsevier/Saunders; 2011. p. 163-84)

Analogous to the discussion above for the RENAISSANCE trial, it is not possible to precisely identify the mechanism for the untoward outcomes in ATTACH. However, the publication of the full trial results from ATTACH has allowed for some potential mechanistic insights that were not available previously.³ As shown in Figure 5A, one of the mechanisms of action of infliximab is to bind to cells expressing TNF on their membrane, and to lyse these cells through complement fixation. Although this type of biological activity is beneficial in eliminating clones of activated T cells in Crohn’s disease, it is predictable that infliximab might be deleterious in heart failure if infliximab bound to TNF that was expressed on the sarcolemma of failing cardiac myocytes (TNF is not expressed in the non-failing heart), which would lead to complement fixation, lysis of cardiac myocyte cell membranes, and cell death.⁴⁴ Analysis of the ATTACH trial indirectly supports this point of view. As shown in Figure 5B, plasma levels of immunoreactive TNF increased at 2 and 6 weeks after treatment with infliximab, as well as after the last dose of infliximab at 6 weeks. Although the increase in TNF levels was attributed to TNF that was bound to infliximab (and hence presumably neutralized), this explanation does not explain the striking 25-fold increase in TNF levels at 10 -28 weeks, when the infliximab levels were declining below detectable levels (see Figure 5B). Moreover, there was a progressive and paradoxical rise in CRP and IL-6 levels over the course of the ATTACH trial, consistent with ongoing tissue injury. Accordingly, one biologically plausible explanation for the increase in patient morbidity and mortality in the ATTACH trial is that infliximab was overtly toxic through complement fixation in the heart.

Figure 5.

Biological properties of infliximab. (A) Infliximab (cA2 G1) is cytotoxic for cells that express TNF on their cell membranes (TNF +), whereas it is not cytotoxic for cells that do not express TNF on their membranes (SP2/O). The mechanism for the cytotoxic effects of infliximab was demonstrated using F(ab)2 fragments of infliximab, which lack the Fc domain and therefore cannot fix complement. As shown the F(ab)2 fragment of infliximab was not cytotoxic for TNF + cells. (B) Levels of immunoreactive TNF in patients who received placebo and infliximab (10 mg/kg) are displayed in relation to the circulating levels of infliximab (data are redrawn from figures 2 and 4⁵). The dotted horizontal lines depict the upper and lower limits of the therapeutic window for infliximab. (Reproduced with permission Mann DL. Activation of inflammatory mediators in heart failure. In: Mann DL, ed. Heart failure: A companion to Braunwald’s heart disease. 2nd ed. Philadelphia: Elsevier/Saunders; 2011. p. 163-84)

Since the completion of these two trials there has been intense debate in the rheumatologic literature with regard to the safety of anti-TNF therapies in patients with rheumatoid arthritis, who are known to have higher risk for cardiovascular complications. Whereas some early studies showed an increased incidence of new onset heart failure in patients treated with infliximab and/or etanercept,⁴⁵ other studies have not shown that these agents are associated with heart failure.⁴⁶ Unfortunately any meaningful interpretation of these conflicting clinical reports is fraught with difficulty and uncertainty, insofar as patients with rheumatoid arthritis are more likely to develop heart failure than age matched subjects without rheumatoid arthritis.⁴⁷ Furthermore, the large administrative data sets that are used to detect heart failure in these retrospective studies rely heavily on the use of ICD-9 codes, which are known to lack the requisite predictive accuracy to exclude the diagnosis of heart failure when present.⁴⁸ The 2012 update of the American College of Rheumatology recommendations for the use of disease-modifying anti-rheumatologic drugs and biologic agents for the treatment of rheumatoid arthritis recommends not using anti-TNF biologics in NYHA class III-IV heart failure patients with an ejection fraction < 50%.⁴⁹

IL-1 receptor antagonist

Anakinra (Kineret™) is an interleukin-1 (IL-1) receptor antagonist that blocks the biologic activity of IL-1 by competitively inhibiting the binding of IL-1 to the Interleukin-1 type receptor. Anakinra has been shown to prevent adverse cardiac remodeling following LAD ligation in mice,⁵⁰ but did not have a significant effect on LV remodeling in small randomized study in patients with acute myocardial infarction.⁵¹ Although the experience with anakinra in heart failure has been limited, two small studies have shown significant improvements in exercise performance in patients with heart failure with a depressed ejection fraction (n = 7) and a preserved ejection fraction (n = 12).^52,53

Other anti-inflammatory agents

The biological effects of pro-inflammatory mediators can also be antagonized using pleiotropic drugs that have anti-inflammatory properties. Three of these therapeutics, statins, N-3 polyunsaturated fatty acids (PUFA), and oxypurinol, have been tested in phase III clinical trials.

Statins

Statins have a variety of pleiotropic effects, including inhibition of inflammatory responses, increased nitric oxide bioavailability, improved endothelial function, and antioxidant properties (reviewed in reference ⁵⁴). Based on the results of several promising retrospective analyses of clinical trials and observational databases suggesting that statins decreased the incidence of heart failure and/or reduced mortality in patients with known heart failure, several large heart failure clinical trials were performed . The UNIVERSE trial (Rosuvastatin Impact on Ventricular Remodeling Cytokines and Neurohormones) examined the effects of rosuvastatin (40 mg/day) on LV remodeling in patients with ischemic and nonischemic dilated cardiomyopathy. Compared to placebo, rosuvastatin was associated with significant reduction of low-density lipoprotein cholesterol but had no effects on LV dimension, LV ejection fraction, nor circulating levels of neurohormones.⁵⁵ Similar findings were reported in the CORONA, (Controlled Rosuvastatin Multinational Trial in Heart Failure) trial, in which 5,011 patients (> 60 years of age) with NYHA functional class II to IV heart failure of ischemic etiology were randomized to 10 mg/day of rosuvastatin versus placebo.⁵⁶ In CORONA treatment with rosuvastatin did not confer a significant benefit with respect to the primary end point, which was a composite of death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke (HR 0.92 [95% CI 0.83 to 1.02]; p = 0.12). Further, there were no significant differences in several of the secondary endpoints including all-cause mortality (HR 0.95 [95% CI 0.86 to 1.05]; p =0.31) and coronary events (HR 0.92 [95% CI 0.82 to 1.04]; p = 0.18), despite a significant decrease in circulating levels of low-density lipoprotein, cholesterol and C-reactive protein (CRP). It is worth noting that the rate of atherothrombotic events was relatively low in the CORONA study, and that the majority of deaths were due to sudden death or worsening heart failure, which reflects the fact that the patient population was comprised of patients with symptomatic heart failure rather than symptomatic coronary artery disease. Thus, the primary composite end point of the CORONA study may not have captured the beneficial effects of rosuvastatin in this elderly group of patients with advanced heart failure. Importantly, treatment with rosuvastatin resulted in a significant decrease in heart failure hospitalizations, which was a pre-specified secondary end point in the CORONA study, thus ending speculation that treatment with statins might lead to worsening heart failure. Moreover, a post-hoc analysis of the CORONA trial demonstrated that rosuvastatin had beneficial effects among those heart failure patients with evidence of increased inflammation at baseline, which was defined as a CRP >2.0 mg/dl.⁵⁷

The GISSI-HF (Gruppo Italiano per lo Studio della Sopravvivenza nell’Insuffi cienza cardiaca-Heart Failure) investigated the efficacy and safety of rosuvastatin in patients with NYHA class II-IV heart failure, irrespective of cause and/or LV ejection fraction.⁵⁸ Patients were randomly assigned to rosuvastatin 10 mg daily (n=2285) or placebo (n=2289), and followed up for a median of 3.9 years. The primary end points of the trial were time to death or admission to hospital for cardiovascular reasons. There was no significant difference in the probability of all cause death in patients who were treated with rosuvastatin when compared to the placebo group (adjusted hazard ratio [HR] 1.00 [95.5% CI 0.898-1.122], p=0.943). Further, there was no significant difference in the composite end point of death or admission to hospital for cardiovascular reasons (adjusted HR 1.01 [99% CI 0.908-1.112], p=0.903).

N-3 Polyunsaturated fatty acids (PUFA)

There is a large body of experimental evidence suggesting that n-3 PUFA has favorable effects on inflammation, including a reduction of endothelial activation and production of inflammatory cytokines, platelet aggregation, autonomic tone, blood pressure, heart rate, and LV function. In a parallel arm of the GISSI-HF study, patients with NYHA class II-IV heart failure were randomized to receive n-3 polyunsaturated fatty acids (PUFA) or placebo. The GISSI-HF trial showed that long-term administration of 1 gram per day of omega n-3 PUFA resulted in a significant reduction in both all-cause mortality (adjusted HR 0.91 [95.5% CI 0.83-0.99], p=0.041) and all-cause mortality and cardiovascular admissions (adjusted HR 0.92 [99% CI 0.849-0.999], p=0.009), in all the predefined subgroups, including heart failure patients in non-ischemic cardiomyopathy group.⁵⁹ Although n-3 PUFA are not endorsed by current practice guidelines, the use of n-3 PUFA may be considered in patients who remain symptomatic despite optimal medical therapy.

Oxypurinol/Allopurinol

Elevated levels of uric acid (UA) are known to predict mortality and the need for heart transplantation in patients with heart failure.⁶⁰ Uric acid is a byproduct of the purine metabolism via the xanthine oxidase (XO) pathway. Serum uric acid levels may increase in heart failure because of increased generation, decreased excretion, or a combination of both factors. Recent studies have shown that uric acid can trigger interleukin-1β– mediated inflammation via activation of the NOD-like receptor protein (NLRP)3 inflammasome, which is a large multi-molecular complex that plays a critical role in the processing of immature interleukin-1β to mature “secretable” form of interleukin-1β . Monosodium urate crystals can also activate the innate immune system through engagement of TLR2 and TLR4.⁶¹

The OPTIME-HF trial was a prospective randomized clinical trial that evaluated the effects of the xanthine oxidase inhibitor oxypurinol in patients with New York Heart Association functional class III to IV heart failure with a LV ejection fraction ≤ 40%.⁶² The end point of the trial was a clinical composite comprised of morbidity, mortality, and quality of life evaluated at 24 weeks. The percentage of patients characterized as improved, unchanged, or worsened did not differ between those receiving oxypurinol or placebo. In a subgroup analysis, patients with elevated serum uric acid (SUA) level of >9.5 mg/dl responded favorably to oxypurinol, whereas oxypurinol patients with SUA <9.5 mg/dl exhibited a trend towards worsening. The NIH-sponsored EXACT (Using Allopurinol to Relieve Symptoms in Patients With Heart Failure and High Uric Acid Levelstrial; NCT00987415) trial tested the hypothesis that treatment with allopurinol would lead to improvements in a composite clinical score in heart failure patients with a reduced ejection fraction and a serum uric acid SUA of > 9.5 mg. Overall, 253 patients were randomized in EXACT. Treatment with allopurinol significantly reduced SUA levels vs. placebo (p < 0.001); however, the proportion of patients who worsened, stayed the same, or improved their heart failure classification was similar for allopurinol vs. placebo (p = 0.25). Moreover, the quality of life (p = 0.16) and submaximal exercise (p = 0.64) was also similar between groups.

Immunomodulation

An alternative approach to targeting specific components of the inflammatory cascade is to employ strategies that dampen the various components systemic inflammatory response. Given the recognition that cross-talk between innate and adaptive immune systems leads to progressive LV remodeling following acute myocardial infarction, and that adverse LV remodeling is driven by activation of monocyte-derived macrophages, dendritic cells and CD4+ T-cells that interact with cardiac auto-antigen-loaded dendritic cells,^63-65 there has been interest in developing broad based immunomodulatory strategies for patients with heart failure. Thus far, three different approaches have been employed in heart failure studies: intravenous immunoglobulin, methotrexate and Immune Modulation Therapy (IMT).

Intravenous immuoglobulin

Therapy with intravenous immunoglobulin (IVIG) has been tried in a wide range of immune-mediated disorders, such as Kawasaki’s syndrome, dermatomyositis, and multiple sclerosis, and most recently dilated cardiomyopathy, wherein the initial results have been encouraging. In a double-blind, placebo-controlled study of 20 ischemic and non-ischemic NYHA class II -IV heart failure patients with an LV ejection fraction < 40% monthly IVIG treatment for 6 months resulted in a significant increase in LV ejection fraction from 26% to 31%, independent of heart failure etiology.⁶⁶ These improvements in functional class and LV function were accompanied by an increase in the anti-inflammatory mediators IL-10, IL-1 receptor antagonist (IL-1Ra), and soluble TNF receptors, as well as a slight decrease in plasma TNF suggesting that IVIG evoked a net anti-inflammatory effect. In contrast to these encouraging results, induction therapy with IVIG in the IMAC (Intervention in Myocarditis and Acute Cardiomyopathy) trial in patients with recent-onset cardiomyopathy (<6 months) and an LV ejection fraction <40% demonstrated no significant effect on LV ejection fraction when compared to placebo.⁶⁷ However, it bears emphasis that there was also an increase in LV ejection fraction from 23% to 42% in the placebo arm, which would have made it difficult to show a statistically significant increase in LV ejection fraction in the treatment arm. Moreover, there were important differences in the IVIG dosing strategies in IMAC and the study by Gullestad and colleagues. That is, while both studies used induction therapy (a total of 2 g/kg IVIG), in the study by Gullestad et al, maintenance therapy (monthly infusions [0.4 g/kg] for a total of 5 months) was also given. Thus, one possible reason for the different outcomes in these two studies is that IVIG maintenance therapy is required for an extended period of time, as has been observed in other chronic inflammatory disorders.

Methotrexate

Epidemiological studies have shown that patients with rheumatoid arthritis have an increased incidence of heart failure, ⁶⁸ and that the heart failure that develops in elderly rheumatoid arthritis patients cannot be explained entirely by traditional cardiovascular risk factors.⁶⁹ Notably, the heart failure that develops in these patients is associated with a concomitant increase in circulating levels of TNF.⁶⁸ Methotrexate, which was originally developed as a folate antagonist for the treatment of cancer, has become a mainstay of therapy in rheumatoid arthritis. Several mechanisms have been proposed including inhibition of T cell proliferation via its effects on purine and pyrimidine metabolism, inhibition of transmethylation reactions required for the prevention of T cell cytotoxicity, interference with glutathione metabolism leading to alterations in recruitment of monocytes and other cells to the inflamed joint, and promotion of the release of the endogenous anti-inflammatory mediator adenosine.⁴⁷ Of note, the use of methotrexate in rheumatoid arthritis has also been associated with reduced cardiovascular events, including heart failure hospitalization, especially in patients of 65 years old or older.⁷⁰ Methotrexate was evaluated in a small (n=71) prospective randomized clinical trial of heart failure patients treated with 7.5 mg qw for 12 weeks.⁷¹ Compared to patients on optimal medical therapy, addition of low dose methotrexate resulted in a significant reduction in the circulating levels of pro-inflammatory cytokines (TNF, IL-6, and MCP-1) and upregulation of the anti-inflammatory cytokines (IL-10 and soluble IL-1 receptor antagonist). There were also improvements in NYHA classification, 6-minute walk test distance, and QOL when compared with baseline values. However, methotrexate had no effect of LV remodeling nor LV ejection fraction after 12 weeks of therapy. The main adverse effects reported for low dose methotrexate were related to gastrointestinal symptoms. Importantly, there were no severe drug toxicities recorded, such as bone marrow suppression or alopecia. The METIS (Methotrexate Therapy Effects in the Physical Capacity of Patients With Ischemic Heart Failure) trial evaluated low dose methotrexate in 50 patients with chronic ischemic heart disease. Patients were given methotrexate (7.5mg) or placebo, plus folic acid (5mg), for 12 weeks. The primary end point was the difference in 6-minute walk test (6MWT) distance before and after treatment. There was no significant difference between groups in distance covered in the 6 minute walk test, nor NYHA classification.⁷² The effects of methotrexate on the rate of heart failure hospitalization (secondary outcome measure) are being evaluated in the ongoing CIRT (Cardiovascular Inflammation Reduction Trial [NCT 1594333]) trial, which examines whether low-dose methotrexate reduces heart attacks, strokes, or death in people with type 2 diabetes or metabolic syndrome that have had a heart attack or known coronary artery disease.

Immune modulation therapy (IMT)

Immune Modulation Therapy (IMT; Celacade™;Vasogen,Inc) utilized a medical device that exposes a sample of blood to a combination of physio-chemical stressors ex vivo. The treated blood sample is administered intramuscularly along with local anesthetic into the same patient from whom the sample is obtained. The physio-chemical stresses to which the autologous blood sample is subjected are known to initiate or facilitate apoptotic cell death. The uptake of apoptotic cells by macrophages results in a downregulation of pro-inflammatory cytokines, including TNF, IL-1β, and IL-8, and an increase in production of the anti-inflammatory cytokines, including TGF-β and IL-10.⁷³ Given the imbalance between pro- and anti-inflammatory cytokines in patients with heart failure,⁷⁴ it was hypothesized that IMT would restore this balance towards normal. In a pilot study employing Celacade™ in 73 patients with moderate heart failure, the investigators noted that the group receiving Celacade™ experienced significantly fewer hospitalizations or deaths, when compared to the placebo group. The decrease in event rate in the treatment arm was accompanied by improvements in quality of life and NYHA clinical classification.⁷⁵ Based on the encouraging results of the early studies the ACCLAIM (Advance Chronic Heart Failure Clinical Assessment of Immune Modulation) pivotal study was conducted in 2426 patients with NYHA class II-IV heart failure patients with ischemic and nonischemic dilated cardiomyopathy. Patients were randomly assigned to receive Celacade (n=1213) or placebo (n=1213) by intragluteal injection on days 1, 2, 14, and every 28 days thereafter.⁷⁶ The primary endpoint was an event driven composite of time to death from any cause or first hospitalization for cardiovascular reasons. There was no significant difference between the Celacade™ and placebo treated patients with respect to the primary end point of the trial (HR 0.92; 95% CI 0.80-1.05; p=0.22). However, in a pre-specified subgroup analysis of patients with NYHA II heart failure and patients without a history of previous myocardial infarction, it was noted that treatment with Celacade™ was associated with a 39% (0.61; 95% CI 0.46-0.80; p=0.0003) and 26% (0.74; 0.57-0.95; p=0.02) reduction in the risk death from any cause or first hospitalization for cardiovascular reasons, respectively, suggesting that IMT may have benefited patients with non-ischemic cardiomyopathy and/or patients with milder heart failure (NYHA class II).

Autoimmunity

Autoimmunity triggered by microbial infections and/or tissue injury has been implicated in the pathogenesis of dilated cardiomyopathy. Although a full discussion of autoimmunity in heart failure is beyond the intended scope of this review, this topic it is mentioned briefly herein insofar as direct communication between the innate and adaptive immune systems has been shown to lead to myocarditis and a dilated cardiomyopathy through a mechanism that involves self-recognition of autoantigens.⁷⁷ Penninger and colleagues demonstrated that dendritic cells (DCs) loaded with a heart specific peptide were sufficient to induce a CD4⁺ T-cell-mediated myocarditis in non-transgenic mice. Remarkably, after resolution of acute myocarditis, subsequent TLR4 stimulation of dendritic cell-immunized mice resulted in relapse of inflammatory infiltrates. Moreover, injection of damaged, syngeneic cardiomyocytes also induced myocarditis in mice if TLRs were activated in vivo, thus providing a potential link between tissue damage, TLR activation, synthesis of pro-inflammatory cytokines and myocardial inflammation.

Clinical observations from a number of groups have also shown that patients with dilated cardiomyopathy develop autoantibodies, which can form immune complexes, activate the complement system, bind to cell surface receptors, and influence downstream signaling in target cells. Autoantibodies directed against the outer loop of the β₁-receptor , the M2 muscarinic receptor, the adenine-nucleotide transporter , cardiac myosin, cardiac troponin I, Na+-K+ ATPase have been demonstrated in patients with dilated and ischemic cardiomyopathies.⁷⁸ However, the presence of circulating autoantibodies alone does not necessarily indicate that the patient has autoimmunity, insofar as low titers of auto-antibodies can also be detected in the healthy population as a part of the natural immunologic repertoire.⁷⁹ This statement notwithstanding, it is remarkable that several clinical studies using immunoadsorption against specific autoantibodies (e.g. the β₁-receptor) have shown beneficial effects in small clinical trials in patients with dilated cardiomyopathy (reviewed in references.^77,80). Moreover, recent advances in aptamer technology have allowed for the development of mutated cyclopeptides that bind to and neutralize circulating anti- β₁-receptors. Use of a cyclic peptide (COR-1) that mimicks the tertiary structure of extracellular loop of the β₁-receptor that auto-antoantibodies bind to, prevented anti- β₁-receptor antibody- mediated myocardial damage, and completely reverted cardiac dysfunction over 3-6 months.⁸¹ The use of the COR-1 cyclopeptide was shown to be safe and well-tolerated in a phase I study in normal subjects (NCT 01043146).⁸² The results of the pilot study of the COR-1 cyclopeptide in patients with dilated cardiomyopathy (NCT01391507) have not yet been reported publicly. Future therapeutic developments in the field of autoimmunity and heart failure will likely benefit from the ongoing work on Toll-like receptors and auto-immunity.⁸³

SUMMARY AND FUTURE DIRECTIONS

As summarized in the foregoing review, the experimental evidence linking activation of the innate immune system to the pathogenesis of heart failure has grown exponentially since the original description in 1990. Accordingly, the conceptual underpinnings for the cytokine hypothesis have grown more robust with time because of advances in the field that have led to a clearer understanding and appreciation of the role of innate immunity in health and disease. This statement notwithstanding, the ability to translate this information to heart failure patients has not met with success in phase III clinical trials, and in some cases has led to worsening heart failure and/or death. Although the reason for the disconnect between the biology of inflammation in the heart and the outcomes of clinical heart failure trials targeting inflammation is not known, Figure 6 illustrates several possibilities that warrant discussion.

Figure 6.

Potential reasons for neutral/negative results targeting inflammation in heart failure. (A) Inflammation is not in the causal pathway of heart failure; (B) Of several causal pathways of heart failure, anti-inflammatory therapy targets a pathway that contributes to the syndrome of heart failure, but is not directly related to clinical outcomes (i.e., is a disease modifier) ; (C) Of several causal pathways of heart failure, anti-inflammatory therapy targets a pathway that contributes to the clinical outcomes, but has direct side effects that lead to worsening heart failure. (dotted lines = off-target mechanism(s) of action that might exist) (Adapted and modified from Fleming TR, DeMets DL: Surrogate end points in clinical trials: are we being misled? Ann Intern Med 1996;125:605-613).

The first and most obvious question raised by the neutral and/or negative clinical trials reviewed above, is whether inflammation is a “correlate” of heart failure, or whether instead it is disease causing. The extant literature shows that circulating levels of inflammatory biomarkers correlate with disease progression (i.e. NYHA class), as well as with clinical outcomes. One possibility depicted in Figure 6A is that while inflammatory markers are elevated in heart failure, they may not be directly responsible for the pathophysiological process(es) that lead to untoward clinical outcomes; that is, inflammation correlates with disease severity, but does not contribute to disease progression in heart failure. Although this possibility cannot be formally excluded, the wealth of pre-clinical data demonstrating that clinically relevant levels of pro-inflammatory cytokines faithfully mimic the heart failure phenotype in a variety of different species, including humans, argues against this possibility as the primary and/or sole explanation. Figures 6B illustrates a clinical scenario in which there may be several disease-causing pathways in heart failure, with one pathway predominately driving clinical outcomes, and the other pathway(s) contributing to but not primarily responsible for clinical outcomes (i.e. disease modifying). In this situation, the effects of anti-inflammatory therapies could be substantially offset by the effects of therapies (e.g., neurohormonal antagonists) that are more directly antagonistic to the pathways driving clinical outcomes. Figure 6B might also explain why circulating levels of pro-inflammatory cytokines decrease in heart failure patients who are placed on evidence based medical and/or device therapies. While plausible, these possibilities do not necessarily explain the observation that targeted anti-TNF therapies worsened clinical outcomes in patients who received the highest doses, and/or who had the longest duration of therapy. Figure 6C illustrates a third possibility in which anti-inflammatory strategies might favorably influence clinical outcomes, but the beneficial effects of these therapies are offset by unintended (e.g. complement fixation with anti-TNF antibodies), unanticipated (e.g., agonistic actions of TNF receptor antagonists), or unrecognized mechanisms of action (e.g. disruption of homeostatic para-inflammatory responses [Figure 2]) of the anti-inflammatory strategy. It is also completely possible that some combination of Figures 6A-C may explain the disappointing results of the clinical trials that have been conducted thus far.

Given the inherent difficulties in developing new heart failure therapies in general, as well as the specific difficulties in targeting chronic inflammation in the setting of heart failure, is there a foreseeable future for developing anti-inflammatory strategies in heart failure? Despite the inauspicious beginning with targeted anti-inflammatory approaches, the expanding body of knowledge in the field of innate immunity and the development of new therapeutic targets in this area, coupled with the ability to utilize inflammatory biomarkers to identify subsets of heart failure patients who have ongoing inflammation despite optimal medical and device therapy, raises the exciting possibility that we ultimately will be able to identify subsets of heart failure patients who will benefit from the striking advances that have been made in this field over the past two decades.

Figure 1.

Damage associated molecular patterns (DAMPs) are derived from dying cells that release their cytosolic content following myocardial injury, from degradation of the extracellular matrix, as well as by immune cells that become activated following tissue injury. (Key: ATP = adenosine triphosphate, HSP = heat shock protein; HMGB1 = high mobility box group 1 protein, IL-1α - interleukin-1α; IL-1R – interleukin receptor NALP = NACHT, LRR and PYD domains-containing protein 3 (cryopyrin); RAGE = receptor for advanced glycation end products, TLR = Toll-like receptor) (Reproduced with permission from Mann DL, The role of innate immunity in heart failure, in Heart Failure: A Companion to Braunwald’s Heart Disease, edited by Mann DL and Felker GM, 2015, pp 109-126. Elsevier/Saunders, Philadelphia, Pennsylvania).

Nonstandard Abbreviations and Acronyms

ACCLAIM

Advance chronic heart failure clinical assessment of immune modulation

AIM

absent-in-melanoma receptor

AP-1

activator protein 1

ATTACH

Anti-TNFα therapy against CHF

CLR

C-type lectin receptor

CORONA

Controlled rosuvastatin multinational trial in heart failure

CRP

C-reactive protein

DAMP

damage associated molecular pattern

GISSI-HF

Gruppo Italiano per lo studio della sopravvivenza nell’ insufficienza cardiaca-heart failure

IFN

interferon

IL-1β

interleukin-1β

IL-18

interleukin-18

IL-1R

interleukin-1 receptor

IMAC

Intervention in Myocarditis and Acute Cardiomyopathy

IMT

immune modulation therapy

IVIG

intravenous immunoglobulin

IRAK

interleukin-1 receptor-associated kinase

IRF

interferon regulatory factor

LPS

lipopolysaccharide

LRR

leucine rich region

METIS

Methotrexate therapy effects in the physical capacity of patients with ischemic heart failure

MR

Mannose receptor

MDA5

Melanoma differentiation-associated gene 5

MyD88

myeloid differentiation primary response gene 88

NACHT domain

domain present in NAIP, CIITA, HET-E and TP1

NOD

nucleotide binding and oligomerization domain

NF-κB

Nuclear factor kappa B

NLR – NOD

like receptor

NLRP

NLR family, pyrin domain-containing protein

NYHA

New York Heart Association

PAMP

pathogen associated molecular pattern

PRR

pattern recognition receptor

PUFA

polyunsaturated fatty acid

RECOVER

Research into etanercept cytokine antagonism in ventricular dysfunction

RENAISSANCE

Randomized etanercept north american strategy to study antagonism of cytokines

RENEWAL

Randomized etanercept worldwide evaluation

RIG – I

retinoic acid inducible gene-I

RLR

RIG-I-like receptor

sST2

soluble ST2

TGF

transforming growth factor

TIMP

tissue inhibitor of matrix metalloproteinase

TLR

Toll-like receptor

TNF

tumor necrosis factor

TREM

Triggering receptor expressed on myeloid cells

UNIVERSE

Rosuvastatin impact on ventricular remodeling cytokines and neurohormones