Repair of the Infarcted Heart: Cellular Effectors, Molecular Mechanisms, and Therapeutic Opportunities

Abstract

The adult mammalian heart has limited endogenous regenerative capacity and heals through activation of inflammatory and fibrogenic cascades that ultimately result in formation of a scar. After infarction, massive cardiomyocyte death releases a broad range of DAMPs that initiate both myocardial and systemic inflammatory responses. TLRs and NLRs recognize DAMPs and transduce downstream pro-inflammatory signals, leading to upregulation of cytokines (such as Interleukin-1, TNF-α and Interleukin-6) and chemokines (such as CCL2) and recruitment of neutrophils, monocytes and lymphocytes. Expansion and diversification of cardiac macrophages in the infarcted heart plays a major role in clearance of the infarct from dead cells and in subsequent stimulation of reparative pathways. Efferocytosis triggers induction and release of anti-inflammatory mediators that restrain the inflammatory reaction and set the stage for activation of reparative fibroblasts and vascular cells. Growth factor-mediated pathways, neurohumoral cascades and matricellular proteins deposited in the provisional matrix stimulate fibroblast activation and proliferation and myofibroblast conversion. Deposition of a well-organized collagen-based extracellular matrix network protects the heart from catastrophic rupture and attenuates ventricular dilation. Scar maturation requires stimulation of endogenous signals that inhibit fibroblast activity and prevent excessive fibrosis. Moreover, in the mature scar, infarct neovessels acquire a mural cell coat that contributes to stabilization of the microvascular network. Excessive, prolonged, or dysregulated inflammatory or fibrogenic cascades accentuate adverse remodeling and dysfunction. Moreover, inflammatory leukocytes and fibroblasts can contribute to arrhythmogenesis. Inflammatory and fibrogenic pathways may be promising therapeutic targets to attenuate heart failure progression and to inhibit arrhythmia generation in patients surviving an myocardial infarction.

1. Introduction

Advances in acute care of patients with myocardial infarction (MI) and the introduction of early reperfusion strategies resulted in a marked reduction in short-term mortality (in-hospital and 30-day mortality) from >30% in the 1950s to ~5-8% in the current era¹,²,³. In contrast to the marked improvement in early mortality and event rates, effects on long-term mortality have been much less impressive⁴. In fact, the improved survival of patients suffering acute MI resulted in an expansion of the pool of patients that develop adverse remodeling and heart failure. Progression of heart failure after acute infarction is dependent not only on the size of the acute infarct, but also on the quality of the reparative response. Thus, heart failure after infarction results, at least in part, from faulty healing.

In patients suffering a large MI, sudden loss of up to a billion cardiomyocytes overwhelms the extremely limited regenerative capacity of the adult human heart. Thus, repair of the infarcted heart requires timely activation and suppression of an acute inflammatory reaction that clears the infarct from dead cells and matrix debris and ultimately results in formation of a collagen-based scar⁵. Immune cells, fibroblasts and vascular cells co-operate to promote healing of the infarcted heart, protecting the ventricle from catastrophic rupture. Perturbations in inflammatory, fibrogenic and angiogenic responses after infarction play an important role in the pathogenesis of adverse post-infarction remodeling and heart failure.

Dissection of the cellular and molecular mechanisms involved in repair of the infarcted heart is important to understand the pathogenesis of post-infarction remodeling and heart failure. This review manuscript discusses our current knowledge on the cellular effectors of cardiac repair, and the molecular signals involved in the reparative cardiac response. We also identify promising therapeutic approaches that may protect the infarcted heart from adverse remodeling by targeting inflammatory and fibrogenic signals involved in cardiac repair.

2. The phases of the reparative response after MI

MI is defined as cardiomyocyte death triggered by an ischemic insult. In most cases, rupture or erosion of a coronary plaque triggers thrombosis leading to acute cessation of flow in the myocardium subserved by the occluded coronary (type 1 infarction). In other cases, ischemia is triggered by an imbalance between myocardial supply and demand caused by an extracardiac stressor, such as hypotension, tachycardia, anemia, or hypertension, in the presence of coronary disease (type 2 infarction). Complete cessation of coronary flow results in almost immediate functional depression (within seconds after the ischemic insult), followed by early ultrastructural changes in ischemic cardiomyocytes (within minutes of coronary occlusion). Ischemic insults longer that 15-20 minutes are sufficient to induce irreversible changes in some subendocardial cardiomyocytes that ultimately culminate in cell death⁶. The longer the ischemic interval, the larger the extent of the infarct, which has been suggested to progress from the subendocardium to the less vulnerable subepicardium as a “wavefront of necrosis” as the duration of the ischemia is prolonged⁷.

Massive death of cardiomyocytes in the infarct zone initiates a well-orchestrated reparative response that can be divided into 3 distinct but overlapping phases: the inflammatory phase, the proliferative phase and the maturation/remodeling phase. Dying cardiomyocytes and degraded extracellular matrix proteins release danger signals that activate inflammatory cascades, resulting in recruitment of neutrophils and monocytes in the infarct. Professional phagocytes clear the infarct from death cells and matrix debris, thus activating endogenous pathways that suppress pro-inflammatory signaling and stimulate fibrogenic and angiogenic macrophage phenotypes. Thus, suppression of inflammation is coupled with the transition to the proliferative phase of cardiac repair, during which fibroblasts expand, convert into myofibroblasts and undergo activation⁸, generating an extracellular matrix network that supports the structural integrity of the ventricle. Formation of neovessels ensures supply of the infarct with blood to support the metabolically active wound. As a matrix network comprised of structural collagens is formed, fibroblasts and myofibroblasts undergo phenotypic transitions to matrifibrocytes, a population of fibroblast-like cells that serve a scar-preserving role, neovessels acquire a coat comprised of vascular mural cells (vascular smooth muscle cells and pericytes) and a mature scar containing large amounts of cross-linked collagen is formed. In large infarcts associated with substantial loss of contractile muscle, replacement of dead cardiomyocytes with collagenous tissue in the infarct zone is accompanied by remodeling of the non-infarcted myocardium, which exhibits macrophage activation and interstitial fibrosis in response to pressure and volume loads. The cellular responses to MI are qualitatively similar in reperfused and non-reperfused infarcts. However, reperfusion attenuates cardiomyocyte loss in the area at risk, while accelerating the myocardial inflammatory response (and its resolution) and promoting earlier activation of fibroblast-driven repair.

3. The inflammatory phase of cardiac repair

3.1. MI triggers both local and systemic inflammatory responses.

MI triggers both systemic activation of inflammation and an intense myocardial inflammatory reaction. Long range signals transfer information from the infarct to the bone marrow and the spleen, thus promoting leukocyte production and mobilization⁹,¹⁰,¹¹. The molecular signals that signal to the immune system the need for leukocyte supply to repair the ventricle remain poorly understood. Angiotensin II activation mobilizes monocytes from the splenic reservoir within hours after MI¹⁰, and cytokines (such as IL-1β) released by the infarct stimulate bone marrow hematopoietic stem cell proliferation and monocyte production¹². Moreover, in the bone marrow, sympathetic nerves may activate β3 adrenergic receptors in mesenchymal stromal cells¹³ thus providing local cues that release hematopoietic stem and progenitor cells and ultimately increase the output of neutrophils and monocytes¹⁴. On the other hand, myocardial signals derived from dying cells localize the inflammatory response to the infarcted heart.

3.2. Molecular pathways that initiate the myocardial post-infarction inflammatory response: the Damage-Associated Molecular Patterns (DAMPs),

Cell death and protease-mediated degradation of the extracellular matrix generate a broad range of damage-associated substances that act as “danger signals” and initiate the post-infarction inflammatory response. These mediators, called DAMPs, bind to pattern recognition receptors (PRRs) located on the surface of surviving myocardial parenchymal cells or and infiltrating leukocytes and stimulate downstream inflammatory cascades that culminate in secretion of pro-inflammatory cytokines and chemokines, and in induction of cell adhesion molecules¹⁵,¹⁶,¹⁷. In addition to their passive release from dead cells and damaged extracellular matrix, DAMPs can also be expressed and secreted by activated immune cells, cardiomyocytes, and interstitial cells under conditions of stress. Over the last 20 years several different DAMPs have been suggested to contribute to the initiation of the post-infarction inflammatory response, including the nuclear protein high mobility group box-1 (HMGB1), heat shock proteins (HSPs), fibronectin extra domain A (fibronectin-EDA), low molecular weight hyaluronic acid, S100 proteins, interleukin (IL)-1α, cardiac myosin, ATP, mitochondrial DNA, dsRNA, extracellular RNA (eRNA) and others^18-22 (Figure 1). The relative significance of each DAMP in activating selected aspects of the inflammatory response remains poorly understood. DAMPs act by binding to PRRs, primarily the toll-like receptor/IL-1 receptor (TLR/IL-1R) cell surface proteins, members of the cytosolic NOD-like receptor (NLR) family, or the membrane receptor for advanced glycation end-products (RAGE). Upon activation, PRRs stimulate downstream mitogen-activated protein kinases (MAPKs) and nuclear factor (NF)-κB cascades^{21, 23-25}, driving induction of pro-inflammatory cytokines and chemokines and upregulation of cell adhesion molecules and facilitating leukocyte trafficking in the infarct. Infiltrating leukocytes serve an important reparative role, clearing the infarct from dead cells and matrix debris and stimulating activation of matrix-producing interstitial cells and angiogenesis. However, excessive, prolonged or dysregulated leukocyte activation may accentuate injury, increasing death of cardiomyocytes, promoting fibrosis and stimulating protease-induced matrix degradation that can lead to left ventricular rupture. Considering that DAMPs may stimulate both reparative and maladaptive inflammatory responses, it is not surprising that various studies inhibiting DAMP-PRR interactions in the infarcted heart have reported both protective and detrimental effects.

Figure 1: Damage-Associated Molecular Patterns (DAMPs) Initiate the post-infarction inflammatory response.

In the infarcted heart, dying cells, damaged extracellular matrix and infiltrating immune cells release a broad range of DAMPs, including the nuclear protein high mobility group box-1 (HMGB1), heat shock proteins (HSP), cardiac myosin, extracellular RNA (eRNA), the S100 proteins S100A8/A9, fibronectin extra domain A (EDA-Fn), low molecular weight hyaluronan (LMWH), and interleukin (IL)-1α. DAMPs bind to pattern recognition receptors expressed on the cell surface of immune cells, vascular cells, fibroblasts and cardiomyocytes, transducing downstream pro-inflammatory cascades and stimulating leukocyte recruitment.

3.2.1. HMGB1.

The non-histone chromatin protein HMGB1 is expressed in mammalian cells and is involved in DNA stabilization and transcription regulation. Upon cell stress and injury, HMGB1 is passively released by necrotic cells²⁶, but also actively secreted by stimulated immune cells²⁷, or by ischemic cardiomyocytes²⁸. HMGB1 release has been documented in both experimental models of myocardial ischemia and in human patients with MI. In animal models of reperfused^{29, 30} and non-reperfused³¹ MI, early myocardial induction of HMGB1 expression is accompanied by a marked increase in serum levels. In acute MI patients, increased HMGB1 levels predict adverse outcome^31-33.

The pro-inflammatory effects of HMGB1 involve binding to RAGE or members of the TLR family (predominantly TLR2, TLR4 and TLR9^{22, 29, 34, 35, 36}) and downstream activation of NF-κB signaling. Moreover, HMGB1 has also been reported to stimulate leukocyte recruitment through direct binding to the CXC chemokine CXCL12/stromal cell-derived factor 1 (SDF-1). HMGB1-CXCL12 heterocomplexes can synergistically enhance leukocyte CXCR4 signaling ³⁷ enhancing inflammation through effects independent of PRR binding.

Studies examining the role of HMGB1 in MI have suggested both injurious and protective actions. In a reperfused MI model, HMGB1 was found to augment injury and increase infarct size, through effects mediated partly via RAGE signaling^{29, 30}. In contrast, in non-reperfused MI models, cardiomyocyte-specific HMGB1 overexpression^{38, 39}, or treatment with exogenous HMGB1 ^40-42 were found to attenuate cardiac remodeling. Protective effects of HMGB1 in non-reperfused infarcts were attributed to improved repair, enhanced angiogenesis and reduced fibrosis. Consistent with these reparative actions, HMGB1 inhibition using a neutralizing antibody perturbed repair in non-reperfused MI, inducing scar thinning, infarct expansion and adverse remodeling³¹. These findings underscore the context-dependent role of DAMP-mediated inflammation in injury and repair of the infarcted heart. Reperfusion salvages many cardiomyocytes from death; however, many of these cells may be vulnerable to the injurious effects of inflammation. On the other hand, in non-reperfused infarcts virtually all cardiomyocytes in the area at risk die; thus, HMGB1 signaling may play a central role in activation of inflammatory and reparative cascades that preserve the structure and geometry of the ventricle protecting from rupture and adverse remodeling.

3.2.2. S100 proteins.

The S100 family of calcium-binding cytosolic is comprised of 25 members with a broad range of intracellular and extracellular functions⁴³. Several members of the S100 family were reported to act as DAMPs, which are released following injury and activate pro-inflammatory signaling by binding to TLRs or RAGE. S100A8 (also known as myeloid-related protein 8 (MRP8)] or calgranulin A) and S100A9 (MRP14) are the best characterized S100 family members in MI. Predominantly expressed and secreted by neutrophils and macrophages, S100A8 and A9 associate to form the S100A8/A9 heterodimer, which functions as a DAMP, signaling through activation of RAGE and TLR4^{44, 45}. Release of S100A8/A9 has been documented in both experimental models of myocardial ischemia and in human patients with acute MI^{46, 47}, ⁴⁸. Both pharmacologic and genetic interventions have suggested that S100A8/A9 may accentuate injury during the early stages of the post-MI inflammatory response by promoting granulopoiesis in the bone marrow⁴⁹ and by accentuating leukocyte infiltration⁵⁰, while promoting reparative actions at later timepoints through recruitment of reparative macrophages⁵¹.

The role of other S100 proteins in the infarcted myocardium remains poorly understood. S100A12 is released in the circulation following MI⁵²; however, its potential role as a DAMP had not been documented. S100A1, the most abundant S100 family member in cardiomyocytes, is also released following MI⁵³, but does not act as a pro-inflammatory stimulus, but is taken up through endocytosis by cardiac fibroblasts to transiently activate TLR4-endolysosomal signaling, promoting an anti-fibrotic phenotype that protects the infarcted heart from adverse remodeling⁵³. S100A4 has also been suggested to exert protective actions, mediated through stimulation of an anti-apoptotic program in cardiomyocytes⁵⁴.

3.2.3. Other DAMPs

Several additional mediators were found to be released in the infarcted myocardium and were suggested to serve as danger signals that activate the post-infarction inflammatory response. IL-1α is constitutively expressed by many different cell types and is upregulated in response to a broad range of pro-inflammatory signals. Cell necrosis results in rapid release of large amounts of IL-1α that can directly stimulate pro-inflammatory signaling cascades by binding to the type 1 IL-1 receptor (IL-1R1). In contrast, apoptosis sequesters IL-1α in the nucleus preventing release of the DAMP and protecting from inflammatory activation⁵⁵. In addition to this mechanism, IL-1α can also be localized on the cell surface, where it may serve as an alarmin that activates neighboring cells in response to injury⁵⁶. Following MI, IL-1α derived from necrotic cardiomyocytes was found to serve as a danger signal that simulates pro-inflammatory MAPK and NF-κB signaling in cardiac interstitial cells through activation of the IL-1R1 pathway, thus promoting adverse remodeling⁵⁷,⁵⁸. Moreover, IL-1α was detected on the cell surface of monocytes from patients with MI and was suggested to play a role in mediating the adhesive interactions between leukocytes and endothelial cells that trigger inflammatory cell infiltration⁵⁹.

Components of the extracellular matrix network have also been suggested to play a role as DAMPs that, when released following injury, activate the inflammatory response⁶⁰. Fibronectin domains, low molecular-weight hyaluronan, and tenascin-C have been suggested to act as alarmins in various pathologic conditions. Data in MI are limited. Fibronectin is secreted by fibroblasts, vascular cells and macrophages infiltrating the infarct and includes an alternatively spliced exon coding type III repeat extra domain A (EDA) that can bind to TLR4 and transduce downstream inflammatory signaling^{61, 62}. Expression of fibronectin-EDA by parenchymal myocardial cells was found to mediate inflammation and dysfunction in the infarcted heart⁶³. In addition to its pro-inflammatory actions, fibronectin-EDA may also play an important role in activation of reparative myofibroblasts⁶⁴, in part through recruitment of active TGF-β in the pericellular matrix⁶⁵.

3.3. PRRs and RAGE mediate the effects of DAMPs in the infarcted myocardium:

There are several major subfamilies of PRRs: the transmembrane TLRs and C-type lectin receptors (CLRs), and the cytoplasmic NLRs, retinoid acid-inducible gene (RIG)-1 like receptors (RLRs) and absent-in-melanoma 2 (AIM2) like receptors (ALR)⁶⁶,⁶⁷. Recently, cyclic GMP-AMP synthase (c-GAS)⁶⁸ and a family of c-GAS-like receptors (cGLRs)⁶⁹ have been identified and found to recognize mis-localized cytosolic double stranded DNA (dsDNA), thus triggering NF-κB- and interferon-mediated inflammatory responses. Following MI, DAMPs released by injured cells and damaged matrix act predominantly through binding to members of the TLR and NLR subfamilies. Moreover, RAGE, an immunoglobulin-like receptor that acts as a PRR transducing DAMP-mediated signals, has also been implicated in activation of post-infarction inflammatory cascades.

3.3.1. TLRs:

The TLRs are expressed in both immune cells and in myocardial parenchymal cells (including cardiomyocytes, fibroblasts, and vascular cells), and serve as the main transducers of DAMP-mediated inflammation following cardiac injury. Out of the 10 TLR family members described in humans, TLR1, 2, 4, 5 and 6 serve as cell surface receptors, whereas TLR3, 7, 8, 9 and 10 are expressed in endolysosomes^{21, 23}. TLR2 and TLR4 are the best-studied members of the family in myocardial disease. Myocardial TLR4 expression is increased after acute MI⁷⁰ and in chronic heart failure^{71, 72}. Moreover, in patients with MI, increased expression of TLR4 and downstream induction of inflammatory cytokines was documented in circulating leukocytes^73-75 and correlated with heart failure progression⁷⁴. Both genetic studies and pharmacologic interventions have suggested that TLR4 plays an important role in extending ischemic injury after infarction by accentuating cardiomyocyte loss and by stimulating inflammatory leukocyte activation^{76, 77},⁷⁸. The cellular mechanisms responsible for the injurious TLR4 actions in the infarcted myocardium have not been systematically examined but may involve effects on both cardiomyocytes and immune cells. TLR2 exhibits similar patterns of regulation and effects to TLR4. In patients with MI, high expression of TLR2 was found in circulating monocytes⁷⁹. Moreover, studies in genetic models demonstrated that TLR2 signaling in leukocytes stimulates pro-inflammatory cascades and increases infarct size in reperfused infarcts ^{80, 81}. Pharmacologic inhibition of TLR2 signaling using a neutralizing antibody had protective effects in both rodent and pig models of MI⁸⁰,⁸². In addition to their role in extension of ischemic injury in models of reperfused infarction, TLR2 and TLR4 also contribute to adverse remodeling in non-reperfused infarction⁸³ through actions on inflammatory and fibrotic responses, which are independent of changes in infarct size⁸⁴.

In addition, endolysosomal TLRs (such as TLR3 and TLR7) have also been implicated in extension of injury after MI. Extracellular RNA-mediated activation of TLR3 signaling was found to extend injury in models of reperfused infarction, increasing the size of the infarct and accentuating inflammatory leukocyte infiltration ⁸⁵,⁸⁶. Moreover, TLR7 was upregulated in infarcted human and mouse hearts. Genetic loss-of-function experiments suggested that single stranded RNA (ssRNA) released by dying cardiomyocytes is taken up by leukocytes and activate endosomal TLR7, thus stimulating pro-inflammatory signaling, and promoting cardiac rupture and adverse post-infarction remodeling⁸⁷.

In addition to their involvement in initiation of the post-infarction inflammatory response, several members of the TLR family (including TLR2 and TLR4) have been implicated in protective preconditioning. Early short-term TLR2 and TLR4 activation prior to ischemia and reperfusion was found to protect cardiomyocytes from ischemic death and may be implicated in ischemic preconditioning^{88, 89},^{23, 90, 91}. These cardioprotective effects are not specific to TLR activation, but have been observed upon early stimulation with several other pro-inflammatory mediators ⁹²,^{93, 94}. The mechanistic basis for the contrasting cytoprotective and pro-inflammatory actions of innate immune pathways in the myocardium remains enigmatic but may reflect cell-specific actions of TLR signaling in cardiomyocytes vs leukocytes, or concentration-dependent actions.

3.3.2. NLRs.

The members of the NLR family are structurally related PRRs that recognize pathogens or danger signals and activate pro-inflammatory cascades⁹⁵. Depending on the composition of their N-terminus, NLR proteins can be classified into subfamilies that contain acid transactivation, pyrin, CARD (caspase activation and recruitment) or BID domains. Several of the pyrin domain-containing NLRs (NLRPs) are implicated in assembly of the inflammasomes, multi-protein complexes comprised of an activated NLR protein, the adaptor protein ASC (apoptosis speck-like protein containing a caspase-recruitment domain) and procaspase-1. This complex facilitates proteolytic maturation of caspase-1 and subsequent cleavage and activation of pro-inflammatory mediators, such as IL-1β and IL-18. Caspase-1 also cleaves gasdermin-D, generating an active N-terminal product that translocates to the plasma membrane, oligomerizes and forms a pore, mediating not only secretion of active IL-1 and IL-18, but also osmotic swelling of the cell that leads to a form of death called pyroptosis⁹⁶. NLRP3 is the best studied of the inflammasome sensors in myocardial diseases and is critically involved in stimulation of inflammatory cascades after MI. All the major components of the NLRP3 inflammasome (ASC, NLRP3 and caspase-1) are upregulated and activated within hours after MI predominantly in infiltrating leukocytes, but also in several other cell types, including fibroblasts vascular cells and border zone cardiomyocytes^97-99. Extensive evidence suggests an important role for the NLRP3 inflammasome in activation of the inflammatory response and extension of injury after MI¹⁰⁰,. Mice with global loss of NLRP3, ASC or caspase-1⁹⁷,¹⁰¹,¹⁰² had reduced infarct size and attenuated ventricular dysfunction in a ischemia/reperfusion model. On the other hand, in models of non-reperfused MI, activation of the NLRP3 inflammasome was found to accentuate inflammation and matrix degradation, thus promising adverse remodeling and dysfunction. Caspase-1 activation accentuated chronic apoptosis of cardiomyocytes in the infarct border zone and enhanced MMP activity causing increased ventricular dilation ¹⁰³. Moreover, activation of the NLRP3 inflammasome in leukocytes was found to promote chronic adverse remodeling by increasing levels of pro-inflammatory cytokines¹⁰⁴. It should be emphasized that protective actions of the NLRP3 inflammasome have also been also reported in models of reperfused infarction and may predominantly involve activation of cytoprotective pathways in cardiomyocytes¹⁰⁵. These effects are similar to the protective actions of TLR-mediated innate immune signals discussed in the previous section.

In contrast to the abundance of evidence on the role of the NLRP3 inflammasome, very little is known regarding the role of other inflammasomes in infarcted hearts. Evidence derived from chronic heart failure models (both ischemic and non-ischemic) suggested that the NLRC4 and AIM2 inflammasomes are activated in failing hearts and may be localized in macrophages and in cardiomyocytes¹⁰⁶.¹⁰⁷. However, the relative contribution of these inflammasomes in activation and regulation of post-infarction inflammation remains unknown.

3.3.3. The c-GAS-STING axis:

c-GAS is an enzyme that acts as a sensor for pathogen-derived or host-derived dsDNA that has entered the cytosol. Upon recognition of cytosolic dsDNA, cGAS synthesizes the nucleotide second messenger cGAMP that directly binds and activates the receptor protein STING (stimulator of Interferon Genes). STING then recruits and activates TANK-binding kinase (TBK)1, subsequently phosphorylating the transcription factor IRF3, which enters the nucleus and induces type 1 interferons. A growing body of evidence has implicated the c-GAS-STING-IRF3 axis in mediating the post-infarction inflammatory response. Uptake of cellular debris by a subpopulation of cardiac macrophages was found to contribute to inflammatory cell infiltration, cytokine upregulation and ventricular dysfunction after MI via activation of a c-GAS/STING/IRF3 cascade¹⁰⁸,¹⁰⁹. Moreover, pharmacologic inhibition of STING in a model of non-reperfused infarction attenuated adverse remodeling and dysfunction¹¹⁰. Although macrophages are considered major cellular targets of the c-GAS-STING cascade¹⁰⁸,¹⁰⁹, the potential role of the pathway in other infarct cell types has not been investigated.

3.3.4. RAGE

RAGE is a PRR expressed as a transmembrane receptor in a broad range of cell types, including immune cells, vascular cells, interstitial cells and cardiomyocytes, but also exists in a soluble form (sRAGE) that can be either cleaved through the actions of proteases on the cell surface, or through alternative splicing of the RAGE gene, resulting in generation of endogenous secreted RAGE (esRAGE). Several DAMPs, including HMGB1 and S100A8/9 can activate cell surface RAGE, stimulating downstream NF-κB- and MAPK-dependent pro-inflammatory cascades²⁵. The central role of RAGE in activation of the post-inflammatory response has been demonstrated by experiments in global RAGE knockout mice, which exhibit reduced infarct size, associated with attenuated leukocyte infiltration and inflammatory cytokine expression after reperfused infarction. These protective actions of RAGE loss decrease adverse remodeling and ameliorate dysfunction^{29, 48}. Although some investigations suggested that the pro-inflammatory actions of RAGE may involve actions on endothelial cells¹¹¹, experiments in bone marrow chimeric mice suggested that leukocytes may be the predominant effectors of RAGE-dependent inflammation after MI⁴⁸.

3.3.5. Downstream pathways activated by PRRs: a system of adaptor molecules fine tunes the inflammatory response.

Cell surface TLRs respond to binding with various DAMPs in a similar manner: ligand recognition results in formation of dimers and triggers recruitment of the adaptor proteins and downstream activation of pro-inflammatory NF-κB and MAPK cascades. Moreover, a wide range of accessory molecules provide essential signals in regulation of TLR cascades, acting as co-factors for ligand recognition and delivery, or mediating crosstalk with other pathways. Except for TLR3, all other TLRs associate with and signal through MyD88 (myeloid differentiation factor 88), an adaptor protein that also transduces IL-1R signaling. Upon ligand binding, MyD88 undergoes dimerization and recruits the IL-1R-associated kinases (IRAKs), thus forming a molecular platform that recruits TRAF6 (TNF receptor associated factor 6) and ultimately activates downstream pro-inflammatory cascades¹¹²,¹¹³. Studies using global KO mice demonstrated a central role for MyD88 in activation of the post-infarction inflammatory response¹¹⁴,¹¹⁵. MyD88 loss reduced infarct size in a model of reperfused infarction, through effects attributed to attenuation of inflammatory injury. In non-reperfused infarction, pharmacologic inhibition of MyD88 decreased post-infarction remodeling, despite the absence of effects on the size of the infarct¹¹⁶. The effects of MyD88 in the post-infarction inflammatory response may involve actions on several different cell types, including bone marrow-derived leukocytes¹¹⁷, resident cardiac macrophages¹¹⁸, cardiomyocytes¹¹⁵ and fibroblasts⁵⁷. A recent study in a model of left ventricular pressure overload suggested important regulatory effects of MyD88 on T lymphocyte phenotype¹¹⁹; however, the significance of these actions in MI has not been investigated. Endosomal TLR4 or TLR3 can also transduce pro-inflammatory signaling through MyD88-independent pathways, by recruiting another cytoplasmic adaptor protein called TRIF (TIR domain-containing adaptor inducing interferon [IFN]-β)⁸⁶,¹²⁰,¹²¹. This pathway activates both NF-κB and IRF3-mekdiated pro-inflammatory cascades.

3.3.6. Activation of the NF-κB system

Several PRR-mediated pathways converge on the NF-κB system, the central transcriptional effector of pro-inflammatory signaling after MI⁹². Activation and nuclear translocation of NF-κB is responsible for upregulation of a broad range of pro-inflammatory cytokines, chemokines, and adhesion molecules in the infarcted heart, thus mediating inflammatory leukocyte recruitment and activation. The mammalian NF-κB family is comprised of 5 members (RelA /p65, RelB, c-Rel, p50, and p52) that form homo- or heterodimers to modulate gene transcription. p50/p65 is the predominant NF-κB complex in the heart, where it exists as an inactive cytoplasmic dimer bound to one of the inhibitor of κB family members, IκBα. NF-κB activation requires activation of IκB kinase (IKK), which induces IκBα phosphorylation and subsequent ubiquitination and proteasomal degradation. IκBα loss unmasks the nuclear localization sequence of the NF-kB sequence, resulting in nuclear translocation of the complex and transcriptional activation. In addition to this canonical pathway of NF-κB activation, a non-canonical pathway has been described that involves slow and persistent activation of NF-κB-inducing kinase (NIK), NIK-mediated p100 phosphorylation and subsequent p100 processing and nuclear translocation of non-canonical NF-κB members, such as p52/RelB¹²². Several TNFR superfamily members (such as the lymphotoxin-b receptor (LTβR) and CD40) are the best-studied activators of non-canonical NF-κB signaling. Dysregulated non-canonical NF-κB signaling in various cell types has been implicated in the pathogenesis of inflammatory and autoimmune conditions¹²².

In animal models of MI, NF-κB is rapidly activated in the infarct zone within the first 1-3 hours after ischemia^123-125. In contrast, NF-κB activation in the remote remodeling myocardium is noted several weeks after infarction, likely reflecting the consequences of pressure and volume loads in animals with heart failure due to large non-reperfused infarcts¹²⁶. Studies using genetic and pharmacologic interventions to target components of the NF-κB system have produced conflicting results. In mouse models, cardiomyocyte-restricted overexpression of mutant phosphorylation-resistant IκBα¹²⁷, cardiomyocyte-specific p65 deletion¹²⁸, IκBα overexpression via gene transfer¹²⁹, NF-κB double-stranded decoy DNA transfection¹³⁰, and pharmacological blockade of IκBα¹³¹, or IKKβ¹³² have supported the notion that NF-κB activation extends the size of the infarct, accentuates leukocyte infiltration and increases ventricular dysfunction after reperfused infarction. These effects may predominantly reflect actions on cardiomyocytes that may involve p65. In non-reperfused infarcts, several investigations using mice with cardiomyocyte-restricted overexpression of mutant phosphorylation-resistant IκBα¹²⁶, the NF-κB signaling inhibitor A20¹³³, and IκBα gene transfer¹³⁴ also suggested detrimental effects of NF-κB, associated with accentuated adverse remodeling and increased inflammation. Studies using p50 KO mice also suggested a role for p50 in post-infarction remodeling in non-reperfused infarcts^{135, 136}.

On the other hand, several other studies have suggested protective actions of the NF-κB system in the infarcted myocardium that may involve pro-survival effects on cardiomyocytes, but also reparative actions on leukocytes. In models of ischemic preconditioning, acute NF-κB activation has been suggested to play an essential role in late cardioprotection induced by ischemic preconditioning^{137, 138}. In the ischemic and reperfused heart, RelB was found to transduce cytoprotective signaling in cardiomyocytes¹³⁹. Moreover, endogenous leukocyte p50 was found to protect the remodeling heart after non-reperfused MI, by attenuating inflammation, reducing matrix metalloproteinase activity and improving repair¹⁴⁰. The conflicting observations between various in vivo studies likely reflect the distinct characteristics of each intervention on specific components and functions of the NF-κB system. Subunit- and cell-specific actions of the NF-κB system, as well as the efficiency and context of the intervention (reperfused vs. non-reperfused MI, timing, etc) need to be carefully considered to interpret the findings of investigations targeting the NF-κB system in MI.

3.4. Pro-inflammatory cytokines and chemokines

Stimulation of innate immune pathways and downstream activation of NF-κB and MAPK cascades results in secretion of a broad range of pro-inflammatory cytokines and chemokines in the infarcted heart. TNF-α, IL-1 and IL-6 are the best-studied pro-inflammatory cytokines in MI and were found to exert a broad range of actions on all cell types involved in cardiac injury, repair and remodeling. Several other cytokines have been suggested to play roles in modulation of immune cell phenotype, thus contributing to the post-infarction inflammatory response. Cytokines are notoriously pleiotropic and multifunctional. For this reason, cytokine functions on the post-MI inflammatory response include both cytoprotective-reparative and injurious actions. Thus, the consequences of cytokine targeting in the infarcted heart are dependent on the context, timing and on the magnitude duration of cytokine inhibition.

3.4.1. TNF-α

TNF-α is rapidly induced¹⁴¹ and secreted¹⁴² in the infarcted myocardium. Experimental studies have suggested both protective and detrimental actions of TNF-α signaling in the infarcted myocardium, reflecting the pleiotropic actions of the cytokine. Global loss of TNF-α or treatment with anti-TNF-α antibody attenuated leukocyte infiltration and reduced infarct size in models of reperfused MI¹⁴³. In contrast, other studies have demonstrated cardioprotective actions of TNF-α signaling¹⁴⁴ that may involve activation of NF-κB-dependent anti-apoptotic pathways in cardiomyocytes¹⁴⁵. The conflicting findings may reflect divergent actions of the TNFR receptors, TNFR1 and TNFR2 in the cell types involved in ischemic injury, inflammation and cardiac repair^{126, 146, 147} and may explain, at least in part, the negative clinical trials of TNF inhibition strategies in heart failure patients.

3.4.2. IL-1

Induction and secretion of IL-1 has been consistently documented in MI patients and in experimental models of coronary ischemia. Extensive experimental evidence from both genetic and pharmacologic inhibition studies suggests a central role for IL-1-mediated signaling in the pathogenesis of cardiac dysfunction and adverse remodeling after MI. Mice with genetic disruption of the type 1 IL-1 receptor (IL-1R1), the only signaling receptor for IL-1α and IL-1β attenuated adverse remodeling after MI, markedly suppressing leukocyte infiltration, but without affecting the size of the infarct¹⁴⁸. Pharmacologic targeting of IL-1 cascades also showed protective effects in experimental models of MI, attenuating adverse remodeling¹⁴⁹, reducing cardiomyocyte apoptosis¹⁵⁰ and suppressing arrhythmogenesis¹⁵¹. IL-1-mediated accentuation of post-MI remodeling and dysfunction has been attributed stimulation of pro-apoptotic signaling and functional depression in cardiomyocytes¹⁵²,¹⁵³,¹⁵⁴, to pro-inflammatory leukocyte activation^{12, 155} to induction of endothelial expression of adhesion molecules that promote leukocyte recruiment¹⁵⁶, and to stimulation of a protease-expressing, matrix-degrading fibroblast phenotype¹⁵⁵. Another member of the IL-1 family, IL-18 has been also suggested to extend inflammatory injury after MI, thus contributing to the pathogenesis of adverse remodeling¹²³,¹⁰³.

3.4.3. IL-6

IL-6, the prototypical member of the gp130 family of cytokines, is expressed by several different cell types, including cardiomyocytes¹⁵⁷, leukocytes¹⁴² and fibroblasts¹⁵⁸, and plays a central role in regulation of inflammatory and reparative responses after MI. The bulk of the evidence suggests that the effects of IL-6 signaling in the infarcted heart accentuate post-MI inflammation and promote dysfunction. Persistent gp130/STAT3 signaling enhanced inflammation in remodeling infarcted hearts¹⁵⁹. Moreover, pharmacologic blockade of IL-6 through administration of an anti-IL6R antibody attenuated dilation and improved function in a model of post-infarction heart failure¹⁶⁰. Understanding of the role of IL-6 in healing infarction is hampered by the notoriously pleiotropic and cell-specific actions of the cytokine and by the distinct effects of trans vs cis IL-6 signaling. Classic IL-6 pathways involve binding to the IL-6R on the cell surface, and subsequent association of the IL-6/IL-6R complex with gp130, which dimerizes and initiates signaling. However, it has been suggested that IL-6 can also act on cells that do not express IL-6R on their surface. ADAM proteases can cleave IL-6R from receptor-expressing cells, generating soluble IL-6R (sIL-6R), which can then associate with IL-6, stimulating gp130 signaling on IL-6R-negative cells. This pathway is called “trans-signaling”, and has been suggested to play a major role in inflammatory responses¹⁶¹. In MI models, inhibition of IL-6 trans signaling was found to be more effective in attenuating inflammation and post-MI dysfunction than global IL-6 blockade¹⁶². The cellular targets of IL-6 in the infarcted heart have not been systematically investigated. IL-6 exerts important modulatory effects on several different cell types, suppressing function¹⁶³ and promoting hypertrophy in cardiomyocytes¹⁶⁴,¹⁶⁵,¹⁶⁶ through the gp130-STAT3 pathway, stimulating fibroblast proliferation and activation ¹⁶⁷,^{168, 169} and mediating both pro- and anti-inflammatory effects on macrophages and lymphocytes¹⁷⁰,¹⁷¹,¹⁷², ¹⁷³,¹⁷⁴.

3.4.4. The chemokines:

Chemokines are a family of small chemotactic cytokines that regulate cell migration and positioning in development, homeostasis and inflammatory injury¹⁷⁵. Based on their structure, chemokines can be classified into 2 major subfamilies (CC and CXC chemokines) and 2 much smaller subfamilies (CX3C and C chemokines). This classification has functional implications: most CC chemokines are potent mononuclear cell chemoattractants, whereas the CXC chemokines that contain the glutamic acid-leucine-arginine (ELR) motif in their aminoterminal region (ELR+ CXC chemokines) serve to recruit neutrophils in sites of inflammation. Several members of the chemokine family are upregulated in healing infarcts and are potent regulators of the inflammatory response¹⁷⁶. In comparison to the much more pleiotropic cytokines, most chemokines exhibit higher cellular specificity, acting predominantly on trafficking of specific leukocyte subpopulations that express the corresponding chemokine receptors.

3.4.5. CC chemokines

Several CC chemokines have been implicated in chemotactic recruitment of proinflammatory monocytes in the infarcted region¹⁷⁷. CCL2/Monocyte Chemoattractant Protein (MCP)-1, the best-studied CC chemokine in myocardial pathology, is rapidly upregulated in the infarcted heart and mediates recruitment of CCR2+ monocytes¹⁷⁷. Pharmacologic or genetic disruption of the CCL2/CCR2 axis in reperfused MI was found to reduce infarct size in some ^178-181, but not in all studies¹⁷⁷, and consistently attenuated post-MI remodeling^{177, 182}. In addition to its protective actions against adverse remodeling, complete CCL2 loss also delayed phagocytic removal of dead cardiomyocytes, perturbing the reparative response¹⁷⁷, highlight the dual role of inflammatory pathways in repair and adverse remodeling. Several other CC chemokines have been implicated in activation of the post-MI inflammatory response, including CCL5^{183, 184}, CCL21 and the CCR9 and CCR1 systems^185-187. CCL17 was also found to exert pro-inflammatory actions in the infarcted heart, not by recruiting inflammatory leukocytes, but through inhibition of regulatory T cell (Treg) infiltration. CCL17-mediated inhibition of Treg chemotaxis was attributed to an interaction between CCL17 and the chemokine receptor CCR4 that inhibits CCL22-mediated chemotaxis¹⁸⁸. It is important to emphasize that CC chemokines are not uniformly pro-inflammatory. Some members of the subfamily may suppress inflammation by recruiting anti-inflammatory subpopulations of monocytes. For example, CCR5 signaling was found to play a role in recruitment of regulatory T cells (Tregs) in the infarct, mediating suppression of the post-infarction inflammatory response¹⁸⁹.

3.4.6. CXC chemokines

Several ELR+ CXC chemokines were found to be upregulated in the infarcted myocardium and may play a role in chemotactic recruitment of neutrophils. The absence of a mouse homolog for CXCL8/IL-8, the prototypical human ELR+ CXC chemokine, hampers understanding of the role of specific members of the family in the post-MI inflammatory response. CXCL8/IL-8 is markedly upregulated in large animal models of MI^{190, 191} and has been implicated in neutrophil recruitment in the infarct in some, but not all studies¹⁹⁰,¹⁹². On the other hand, mouse studies suggested that CXCR2, the chemokine receptor for most ELR+ CXC chemokines, not only mediates inflammatory leukocyte recruitment, but may also exert direct protective effects on cardiomyocytes¹⁹³. In addition, IL-8 and other CXCR2 ligands have been suggested to be potent angiogenic mediators¹⁹⁴,¹⁹⁵. The concept of compartmentalized CCR2-mediated actions that may exert protective effects on cardiomyocytes and vascular cells, but accentuate leukocyte-mediated inflammatory injury has also been supported by experiments investigating the role of another (non-canonical) CXCR2 ligand, the cytokine macrophage migration-inhibitory factor (MIF)¹⁹⁶.

In contrast to the effects of the ELR+ subfamily, CXC chemokines lacking ELR cannot attract neutrophils, but have been implicated in lymphocytes chemotaxis^197-199. Moreover, a robust body of literature suggests that several ELR-negative CXC chemokines may also modulate fibroblast and vascular cell phenotype, exerting angiostatic and anti-fibrotic actions^200-202. CXCL10/ Interferon-γ-inducible protein (IP)-10, the best studied member of this subfamily, is markedly upregulated in both canine and mouse myocardial infarcts^{203, 204}. Studies using a global loss of function model suggested that CXCL10 protects the infarcted heart from excessive fibrotic remodeling, limiting infiltration of the infarct with activated myofibroblast by inhibiting growth factor-induced fibroblast migration²⁰⁵. The anti-fibrotic effects of CXCL10 did not involve its canonical receptor CXCR3 but were mediated through interactions with cell surface proteoglycans²⁰⁶.

CXCL12/Stromal cell-derived factor (SDF)-1 stands out as a homeostatic chemokine with a central role in embryonic development²⁰⁷ and in mobilization, survival and differentiation of progenitor cells. CXCL12 is upregulated in the infarcted heart^208-214 and has been suggested to regulate inflammatory, reparative and cell death pathways through effects involving the CXCR4 and ACKR3/CXCR7 receptors. In addition to its role in inflammatory activation²¹⁵, CXCL12 was also found to exert angiogenic actions^216-227 and to inhibit cardiomyocyte apoptosis^{222, 225, 228}. The broad range of cellular effects of CXCL12 may explain the conflicting findings from studies testing the effects of CXCL12/CXCR4 inhibition after MI²²⁹,^{230, 231}.

3.5. The cellular effectors of the inflammatory response

A broad range of cell types, both resident myocardial cells and recruited cells from the bone marrow or the spleen (Figure 2), are implicated in the post-infarction inflammatory response. These cells co-operate to clear the infarct from dead cells and matrix debris and to set the stage for stimulation of a reparative program.

Figure 2: Temporal course of immune cell accumulation in myocardial infarction (MI) is linked to phase-specific cell functions.

ROS, reactive oxygen species. NET, neutrophils extracellular traps. LCN2, lipocalin 2. IL, interleukin. VEGF, vascular endothelial growth factor. SPP1, secreted phosphoprotein 1. TGF, transforming growth factor.

3.5.1. Cardiomyocytes as pro-inflammatory cells

Dying cardiomyocytes contribute the central stimulus that activates post-MI inflammation by releasing DAMPs that transduce innate immune signals¹⁶. Moreover, surviving cardiomyocytes in the infarct border zone are subjected to mechanical stress and are exposed to a cytokine-rich environment. These unique microenvironmental conditions may promote synthesis and release of inflammatory mediators that may contribute to the intense infiltration of the infarct border zone with leukocytes. Studies in large animal models have demonstrated that cardiomyocyte in the viable border zone may be a significant source of cytokines, such as IL-6¹⁵⁷. Considering that several other cell types can produce large amounts of cytokines the relative contribution of border zone cardiomyocytes as pro-inflammatory cells is unknown.

3.5.2. Resident and recruited macrophages

Macrophages are the most prevalent immune cell type residing in the normal mammalian heart. Primitive macrophages infiltrate the developing heart early during embryogenesis and assist in coronary vessel maturation. Their progeny can persist into adulthood through self-renewal in situ. In addition, and over time, monocytes infiltrate the healthy heart and contribute to the resident cardiac macrophage population. Cardiac macrophages in mice and humans have been classified into at least two major subpopulations according to their differential expression of the chemokine receptor CCR2. While monocytes give rise to CCR2+ macrophages, the CCR2− population harbors embryonically derived macrophages with divergent reports regarding the extent of substitution by monocyte-derived cells according to lineage tracing experiments in mice and analysis of sex-mismatched heart transplants or mutation variant allele frequencies in humans²³²,²³³,²³⁴,²³⁵. Differential gene expression profiles of CCR2+ and CCR2− cardiac macrophages share great similarities between mice and humans, supporting the translational value of the mouse model.

Cardiac macrophages preserve tissue homeostasis through efferocytosis of cardiomyocyte exosomes containing damaged cell organelles, electrical coupling to cardiomyocytes, and defense against pathogens²³⁶,²³⁷. With aging and continuous monocyte recruitment, however, release of the anti-inflammatory cytokine IL-10 by cardiac macrophages can induce a profibrotic phenotype which promotes diastolic dysfunction in mice²³⁸. IL-10 is considered a prototypical M2-like macrophage cytokine; however, the extensive use of single cell technologies for characterization of cell profiles has demonstrated that the dichotomous M1/M2 macrophage nomenclature does not adequately reflect primary cardiac macrophage heterogeneity.

Single cell RNA-seq cluster analysis identifies at least three major macrophage subpopulations in the healthy adult murine heart. Macrophages expressing high levels of TIMD4, FOLR2 and/or LYVE1, but low levels of MHC II and CCR2, are commonly referred to as the resident population. This marker profile is relatively preserved across different tissues and organs, the respective population shows little or slow replacement by recruited monocytes, and transcriptional activation of endocytosis, angiogenesis and regeneration related pathways²³⁴,²³⁹,²⁴⁰. Two LYVE1^low MHC II^high populations can be distinguished based on CCR2 expression with more pronounced or complete monocyte dependence. In line with high MHC II expression, their transcriptional profiles are associated with antigen processing and presentation capacity and inflammatory pathways which are particularly enriched within the MHC II^high CCR2+ subset ²⁴¹. Another minor inflammatory cell cluster features many type I interferon response genes and is termed according to its signature gene Isg15 (interferon-stimulated gene 15) or IFNIC (interferon-inducible cells). These heterogeneous cardiac macrophage populations react differentially in response to MI. In murine models of non-reperfused infarction, most tissue resident macrophages in the infarct area die and are replaced by recruited monocytes and their progeny. In the remote myocardium or with timely reperfusion of the ischemic area, many of the original macrophages survive and expand through local proliferation supplemented by newly recruited cells²⁴²,²⁴³. Notably, MHC II^high CCR2+ macrophages residing in the tissue prior to MI are activated via TLR signaling and attract monocytes and neutrophils to the infarcted heart within few hours following MI¹¹⁸,²⁴⁴,²⁴⁵. The pericardial cavity has been proposed as an additional, monocyte-independent source of infarct infiltrating macrophages²⁴⁶; however, the relevance and magnitude of this process has been challenged²⁴⁷.

Depleting all monocytes and cardiac macrophages by clodronate liposomes impairs proper healing of experimental MI²⁴⁸,²⁴⁹, underscoring the critical role of cardiac macrophages in clearing cell debris and orchestrating granulation tissue formation. Depleting CCR2+ resident cardiac macrophages selectively while preserving the CCR2− resident population prior to MI, or limiting the recruitment of monocytes into the infarct in the inflammatory phase improves subsequent cardiac remodeling, but imposes an increased risk of acute post-MI ventricular arrhythmia¹¹⁸,²⁵⁰. Depleting resident cardiac macrophages while allowing for monocyte influx and replacement is detrimental to MI healing, both in terms of structural remodeling and arrhythmogenesis¹¹⁸, ²⁵⁰. Mechanistically, macrophages phagocytosing apoptotic cells (efferocytosis) in the infarct during the inflammatory phase is essential for preventing arrhythmia and secondary necrosis which counteracts the resolution of tissue inflammation and transition into the proliferative phase of MI healing, during which efferocytosis receptor expression peaks²⁵¹, ²⁵⁰. In addition, amphiregulin released by cardiac macrophages controls connexin 43 phosphorylation and translocation in cardiomyocytes and suppresses arrhythmia²⁵². Dying cardiomyocytes overexpress the “don’t eat me signal protein” CD47 on their surface which binds to Signal-regulatory protein alpha (SIRPa) on cardiac macrophages hindering efferocytosis. Blockade of CD47 at the onset of MI improves clearance of apoptotic cardiomyocytes and cardiac remodelling²⁵³. However, uptake of apoptotic cells can lead to intracellular sensing of nucleic acids, activation of interferon regulatory factor 3 (IRF3) pathways and induction of type I interferons (IFN) which amplify the inflammatory response¹⁰⁸. Phagolysosomal destabilization and oxidative stress-related uncoupling of the thioredoxin-interacting protein within macrophages, and P2X7-receptor binding of extracellular ATP released in the infarct, facilitate potassium efflux and subsequent activation of the NLRP3 inflammasome. Its expression peaks within the first three days post MI. NLRP3 activation and oligomerization leads to the release of mature IL-1β, the hallmark pro-inflammatory cytokine, via Gasdermin pores in the cell membrane²⁵⁴. Gasdermin D deficiency reduces IL-1β and IL-18 release in the infarct²⁵⁵. Other gene signatures of infiltrating monocytes and early or transient cardiac macrophage populations that dominate the inflammatory phase are related to hypoxia, leukocyte migration, T cell activation, interferon-response, glycolysis, proliferation and extracellular matrix degradation ²⁵⁶,²⁵⁷,²⁵⁸,²⁵⁹.

3.5.3. Mast cells

In addition to macrophages the adult mammalian heart also contains smaller populations of resident mast cells²⁶⁰,²⁶¹ and dendritic cells²⁶²,²⁶³. Mast cells are strategically located around vessels and degranulate in response to activating signals, such as DAMPs, adenosine, or components of the complement cascade, thus releasing their pre-formed pro-inflammatory mediators and rapidly stimulating the immune response. TNF-α, histamine and tryptase released by degranulating mast cells may activate macrophages and endothelial cells, thus promoting inflammation¹⁴²,²⁶⁴.

3.5.4. Fibroblasts

The normal heart contains large numbers of interstitial and perivascular fibroblasts ²⁶⁵ that regulate homeostasis of the extracellular matrix network. During the inflammatory phase of cardiac repair, stimulation with TLR ligands, or with cytokines delays myofibroblast conversion and induces a pro-inflammatory and matrix-degrading fibroblast phenotype, associated with secretion of cytokines and chemokines, and with release of matrix metalloproteinases (MMPs)¹⁵⁵. Considering that several other cell types (including cardiomyocytes, macrophages and endothelial cells) can also produce the same pro-inflammatory mediators, the relative significance of fibroblast-derived inflammatory cytokines is unclear. Studies investigating the role of fibroblast-specific inflammatory activation are lacking and the evidence supporting the pro-inflammatory roles of fibroblasts in the infarcted heart is for the most part associative. In experimental models of MI, infarct fibroblasts activate the NRLP3 inflammasome⁹⁷,⁹⁹, and could serve as an important source of active IL-1β. Moreover, infarct fibroblasts secrete large amounts of granulocyte/macrophage colony-stimulating factor (GM-CSF) ²⁶⁶, a mediator involved in monocyte recruitment, and are a major source of IL-6¹⁵⁸.

3.5.5. Endothelial cells.

The heart is a highly vascular organ; thus, endothelial cells are the most abundant non-cardiomyocytes in the adult mammalian myocardium²⁶⁵. Endothelial cells play a central role in the post-infarction inflammatory response by secreting cytokines and chemokines and by interacting with leukocytes, thus promoting their transmigration to the infarct. DAMPs released by dying cells inflammatory cytokines (such as IL-1 and TNF) and autacoids (such as histamine) stimulate endothelial cells in the infarct zone, inducing upregulation of endothelial adhesion molecules, thus triggering adhesive interactions with activated leukocytes. Histamine rapidly mobilizes preformed P-selectin from Weibel-Palade bodies to the endothelial cell surface through effects involving H1 receptors²⁶⁷,²⁶⁸. E-selectin is also upregulated in the infarct endothelium in response to the cytokine-rich environment. Upon expression on the endothelial cell surface, selectins engage in weak transient interaction with their leukocyte ligands, capturing leukocytes and mediating rolling along the endothelium ²⁶⁹,²⁷⁰. Subsequently, rolling leukocytes expressing specific chemokine and cytokine receptors interact with the corresponding ligands, which are bound to glycosaminoglycans on the endothelial cell surface. These interactions induce conformational changes of leukocyte integrins (such as LFA-1, Mac-1 and VLA-4) that strengthen the adhesive interaction ²⁷¹,²⁷², resulting in arrest and firm adhesion of the leukocyte to the endothelial surface. The integrin-mediated adhesive interactions involve binding to endothelial ICAM-1 or VCAM-1, which are upregulated on the endothelial cell surface in response to cytokine stimulation⁵⁹. Leukocyte transmigration follows. Leukocytes actively crawl towards “transmigration hotspots”, regions of the endothelial and pericyte layers that serve as exit points for extravasating cells²⁷³,²⁷⁴,²⁷⁵.

3.5.6. Pericytes

The heart also contains a large population of pericytes ²⁷⁶,²⁷⁷. During the inflammatory phase, cytokine-mediated actions may perturb the capillary barrier function of pericytes increasing microvascular permeability²⁷⁸. Moreover, infarct pericytes were found to exhibit a migratory phenotype²⁷⁹ and may dissociate from their close relation with endothelial cells. The cytokine pro-nerve growth factor (pro-NGF), is upregulated in ischemic hearts and may promote microvascular dysfunction through effects on pericyte p75 NGF receptor signaling²⁸⁰. Although pericytes can express cytokines and chemokines in response to DAMPs, their relative role as inflammatory cells in MI has not been studied. Regardless of their significance as cellular sources of inflammatory mediators, pericytes are likely to serve as key regulators of leukocyte trafficking²⁷⁴. Perhaps, the best-documented function of pericytes in the inflammatory phase of MI is their role in mediating the “no-reflow phenomenon”, the failure of subendocardial microvascular reperfusion of the ischemic myocardium despite full restoration of coronary blood flow²⁸¹. In a rat model of myocardial ischemia/reperfusion, sites of microvascular occlusion were spatially associated with pericytes, suggesting that pericyte constriction may contribute to leukocyte entrapment within the microvessels. Adenosine administration to induce pericyte relaxation supported the proposed role of pericytes in microvascular occlusion²⁸². It should be emphasized that other cell types, including leukocytes and endothelial cells, likely also contribute to the pathogenesis of no-reflow.

3.5.7. Neutrophils

Neutrophils dominate the tissue infiltrate within the first 24h after MI^{5, 283}, Neutrophils are recruited in the infarct in response to a broad range of chemoattractants, including cytokines and chemokines, endogenous lipid mediators (such as prostaglandin E₂ and leukotrienes), histamine, and components of the complement cascade^284-286. Extravasation of neutrophils in the infarct requires activation of adhesive interactions with endothelial cells and pericytes. Neutrophils are major contributors to post-infarction inflammation and act by releasing their granular contents (that include cytokines and proteases), by generating reactive oxygen species (ROS), by forming extracellular traps (NET), and by phagocytosis cells and matrix debris. Neutrophils releasing S100A8/9 alarmins prime NLRP3 activation in an autocrine feedback loop leading to IL-1b production, which in turn stimulates granulopoiesis in the bone marrow by reverse-migrating activated neutrophils⁴⁹,²⁸⁷. Neutrophil derived lipocalin-2 (Lcn2) stimulates ROS production in MI tissue and facilitates ventricular arrhythmia²⁵⁰. Inhibition of myeloperoxidase (MPO) released by neutrophils in the infarct improves cardiac remodeling particularly when initiated within the first 24 hours post-MI²⁸⁸,²⁸⁹. In contrast, neutrophil NADPH oxidase-deficiency did not influence the infarct size²⁹⁰. In analogy to the M1/M2 nomenclature in macrophage polarization, a biphasic response of pro-inflammatory N1 dominating the inflammatory phase followed by anti-inflammatory CD206+ IL-10 producing N2 neutrophils in the proliferative phase has been proposed²⁹¹. However, recent scRNAseq analysis has painted a more nuanced and diversified picture distinguishing up to 6 clusters of post MI neutrophils in a time-dependent manner²⁹². During the acute phase post MI infiltrating Siglecf^low neutrophils express high levels of inflammatory Lcn2, Oncostatin-M and Cxcl3. From day 3 post MI onward, a large fraction of neutrophils in the MI tissue expresses high levels of Siglecf, a surface marker shared with eosinophils, which they acquire in the tissue alongside ICAM1 expression²⁹². Neutrophils have a short lifespan and become apoptotic, requiring clearance by macrophages. Neutrophils may undergo a specific form of cell death termed NETosis while releasing extracellular DNA and cell lytic histones and enzymes. Inhibition of NETosis limits MI inflammation and injury²⁹³. Studies depleting neutrophils after MI have produced conflicting results that may reflect, at least in part, difference in timing and effectiveness of depletion. Neutrophil depletion restricted to the time of experimentally induced MI improves infarct healing by limiting early inflammation²⁸⁹,²⁹⁴. However, prolonged neutrophil depletion during the inflammatory and into the proliferative phase post MI has detrimental effects²⁹⁵. Mechanistically, lack of neutrophil-derived LCN2 inhibits the development of reparative macrophages with high efferocytosis capacity in the late infarct. Similarly, neutrophil-derived Annexin A1 induces a pro-angiogenic cardiac macrophage phenotype post MI and limits inflammatory cell recruitment²⁹⁶. Interestingly, MMP-12 expression by cardiac macrophages increases during the late inflammatory phase, with MMP-12 degrading neutrophil chemoattractants. Thus Ly6Clow macrophages arising from infiltrating Ly6Chigh monocytes²⁴² contain neutrophil influx and facilitate the transition from the inflammatory to the proliferative phase of MI healing²⁹⁷.

3.5.8. Monocytes

The peak of monocyte recruitment in the infarcted heart follows the wave of neutrophil infiltration. In mice, two monocyte subsets are distinguished by differential Ly6C and CCR2 expression. Ly6C^high monocytes infiltrate the injured heart and give rise to cardiac macrophages. Ly6C^Low monocytes derive from Ly6C^high monocytes and patrol the vasculature rather than differentiating into MI macrophages²⁴². Murine Ly6C^high/CCR2+ monocytes correspond to the CD14^high/CD16^low classical and CD14^high/CD16^high intermediate monocyte subsets in humans, whereas murine Ly6C^low/CCR2− monocytes correspond to CD14^low/CD16^high non-classical monocytes in humans²⁹⁸, although scRNAseq and mass cytometry studies have identified additional monocyte subtypes in human blood ²⁹⁹,³⁰⁰. In human populations, high numbers of classical or intermediate monocytes associate with future cardiovascular events predict adverse outcomes after MI³⁰¹,³⁰²,³⁰³.

In-vivo time-lapse imaging in mice suggests that monocytes rolling in coronary venules are the first to infiltrate the infarct within 30 minutes post MI, preceding neutrophil invasion which, however, dominates cell infiltration within the first 24 hours post MI quantitatively²⁴⁴. Subsequently, monocytes are mobilized from the splenic reservoir in an angiotensin II dependent manner¹⁰,³⁰⁴. CC chemokines, such as CCL2 and CCL7, released from endothelial cells, CCR2+ cardiac macrophages and stromal cells recruit Ly6C^high monocytes to the activated endothelium and facilitate transmigration into the injured myocardium¹⁷⁶,¹⁷⁷,¹¹⁸,²⁵⁷,²⁴³. Monocytes and macrophages infiltrating the infarct are a major source of pro-inflammatory cytokines, such as IL-1β, and in turn IL-1β stimulates both medullary and extramedullary monocytopoiesis and cell recruitment to the infarct³⁰⁵,¹²,¹⁴⁸. During emergency myelopoiesis, monocytes acquire granulocyte marker transcripts³⁰⁶. In the chronic phase post MI, endogenous beta-adrenergic stimulation of the bone marrow niche suppresses hematopoietic quiescence, inhibits the retention factors CXCL12 and Angiopoietin-1, and stimulates extramedullary hematopoiesis²⁴³. Nevertheless, the precise pathological significance of monocytes in the healing process of MI remains unclear, primarily due to the absence of experimental tools that allow for the selective study of their role without interfering with their differentiation into cardiac macrophages.

3.5.9. Lymphocytes

Lymphocytes are recruited in the infarcted heart along with monocytes²⁸³,³⁰⁷. Most of the studies on the role of lymphocytes in cardiac repair have focused on their anti-inflammatory actions and their modulatory effects on macrophage phenotype during the proliferative phase of cardiac repair. Data on potential pro-inflammatory actions of lymphocytes are limited and predominantly suggest effects of B lymphocytes. B-cells infiltrate the infarcted heart during the first 24h after MI and persist during the proliferative phase²⁸³. ScRNA-seq showed accumulation of diverse polyclonal B-cell subsets in infarcted mouse hearts without evidence of antigen specificity, paralleled by mild clonal expansion of germinal center B cells in the mediastinal lymph nodes³⁰⁸. The role of B cells in regulation of the post-infarction inflammatory response remains understudied. However, Zouggari et al ³⁰⁹ showed that in the infarcted myocardium mature B-cells produce the chemokine CCL7, mobilizing pro-inflammatory monocytes that infiltrate the myocardium. Inhibition of B-cells through CD20 neutralization improved left ventricular remodeling after MI in mice³⁰⁹. In addition to production of monocyte chemoattractant chemokines, B cells have also been suggested to serve as an important source of TGF-β1³⁰⁸.

Moreover, associative data on the kinetics of lymphocytes in human STEMI patients suggested that CX3CL1/fractalkine may mediate recruitment of a pro-inflammatory subset of effector T cells in the infarct, promoting microvascular occlusion and accentuating injury³¹⁰. The role of T cells in MI healing is discussed in detail in section 4.2.4.

4. The proliferative phase

The transition to the proliferative phase of cardiac repair is marked by suppression of the inflammatory response and subsequent activation of fibrogenic and angiogenic cascades. Macrophages, lymphocytes, fibroblasts and vascular cells co-operate to form a highly cellular granulation tissue, ultimately leading to deposition of a matrix network that protects the infarcted heart from catastrophic rupture.

4.1. Mediators involved in suppression and resolution of inflammation.

Timely suppression of the post-infarction inflammatory response is not a passive process that simply reflects the transient nature of pro-inflammatory cytokine and chemokine induction but is dependent on de novo synthesis and release of anti-inflammatory mediators and on activation of intracellular STOP signals that inhibit the innate immune response. Anti-inflammatory cytokines (such as IL-10) and certain members of the TGF-b superfamily are the best studied mediators that have been implicated in suppression of post-infarction inflammatory response ³¹¹. Defective suppression, resolution and containment of inflammation after infarction may underlie adverse remodeling in patients surviving MI.

IL-10 is a potent anti-inflammatory cytokine that inhibits expression of pro-inflammatory cytokines and chemokines in macrophages³¹² through activation of the STAT3 cascade³¹³. IL-10 is markedly upregulated in healing infarcts³⁰⁷,²⁰⁴, and has been localized in T lymphocytes³⁰⁷ (with particularly high levels in regulatory T cells/Tregs¹⁸⁹), in subsets of macrophages³⁰⁷ and in regulatory B cells³¹⁴. Although studies in IL-10 KO mice support the role of endogenous IL-10 in negative regulation of pro-inflammatory cytokines in healing infarcts³¹⁵, ³¹⁶ the evidence on the functional impact of IL-10 actions in the infarcted heart is conflicting. In a mouse model of myocardial ischemia/reperfusion, IL-10 loss was associated with markedly increased early mortality³¹⁵, whereas in another study using a closed-chest model of reperfused infarction and a much higher sample size showed no effects of IL-10 absence on mortality or dysfunction, despite increased myocardial levels of TNF and CCL2³¹⁶. Several investigations suggested that administration of exogenous IL-10 can improve remodeling after MI, at least in part by attenuating the inflammatory response³¹⁷,³¹⁸.

The anti-inflammatory cytokines IL-4 and IL-13 have also been implicated in negative regulation of post-infarction inflammation and repair after MI and can be secreted, not only by macrophages and lymphocytes, but also by eosinophils³¹⁹,³²⁰ and basophils³²¹ that infiltrate the infarct. Myeloid cell-specific IL-4Rα signaling was found to protect the infarcted heart from adverse remodeling, at least in part by suppressing inflammation and by regulating matrix remodeling³²². Exogenous administration of IL-4 and of IL-13 were suggested to exert protective actions after infarction by promoting an anti-inflammatory and reparative macrophage phenotypes³²³,³²⁴

Several members of the TGF-β superfamily (including TGF-βs and Growth Differentiation Factor (GDF)-15) exert potent anti-inflammatory actions and have been implicated in negative regulation of the post-infarction inflammatory reaction. The anti-inflammatory actions of TGF-βs may involve effects on both macrophages and lymphocytes. Although TGF-βs are potent monocyte chemoattractants, their effects on macrophages are anti-inflammatory promoting an M2-like phenotype³²⁵. TGF-β-mediated inhibition of macrophage expression of pro-inflammatory cytokines and chemokines³²⁶,³²⁷,³²⁸,³²⁹ has been attributed to MyD88 degradation, resulting in attenuated NF-κB activity³³⁰. Moreover, TGF-βs have profound effects on T cells, inhibiting Th1 helper and cytotoxic T cell responses³³¹,³³², while enhancing differentiation of anti-inflammatory Tregs³³³. The in vivo significance of the anti-inflammatory actions of TGF-β in MI are supported by neutralization experiments using gene therapy with the extracellular domain of the type II TGF-β receptor (TβRII). Early TGF-β inhibition worsened dysfunction accentuating the inflammatory response, whereas late disruption was protective, reflecting the role of TGF-β in activation of fibroblasts³³⁴, thus supporting the notion that TGF-β may act as a “master switch” in the transition from inflammation to fibrosis. The anti-inflammatory effects of TGF–β on infarct macrophages involve activation of Smad3, but not Smad2³³⁵,³³⁶.

Another member of the TGF-β superfamily, GDF-15 exerts potent anti-inflammatory actions by inhibiting chemokine-mediated leukocyte integrin activation. In mice, GDF-15 loss is associated with fatal cardiac rupture after infarction, caused by accentuated post-infarction inflammation³³⁷. In STEMI patients, elevated plasma GDF-15 levels were found to predict mortality³³⁸, possibly reflecting activation of an anti-inflammatory pathway in patients with an accentuated post-infarction inflammatory reaction.

A broad range of pro-resolving lipid mediators, such as the resolvins, lipoxins, protectins and maresins³³⁹ potently inhibit inflammation and contribute to resolution of the inflammatory infiltrate³⁴⁰. Whether release of endogenous pre-resolving lipids plays a role in suppression of inflammation after MI is not known. However, several studies have suggested that administration of exogenous mediators, such as resolvins E1 and D1 and 15-epi-lipoxin protect the infarcted heart, attenuating post-infarction inflammation^{341, 342,343}.

4.2. The cellular basis of cardiac repair

4.2.1. The role of macrophages in suppression of inflammation and in repair.

As the influx of monocytes subsides, cardiac macrophages become proliferative and acquire reparative properties, secreting growth factors and matricellular proteins and stimulating fibroblast activation and angiogenesis. In fact, efferocytosis can induce non-inflammatory macrophage proliferation³⁴⁴, and contribute to the re-emergence of steady state macrophage progenies in the infarct²⁵⁸,²⁴¹. Efferocytosis also reprograms macrophages to express inflammation-resolving and reparative mediators³⁴⁵,³⁴⁶. The majority of accumulating infarct macrophages during the proliferative phase of repair are monocyte-derived but form heterogenous subpopulations. One-third of them re-acquire resident cell-like profiles. Other macrophages retain an inflammatory signature whereas many exhibit a lipid-associated macrophage gene signature similar to atherosclerotic plaque-associated foam cells and characterized by triggering receptor expressed on myeloid cells 2 (TREM2) expression. TREM2 can be released by these macrophages and suppresses inflammation in an autocrine and paracrine fashion²⁵⁸,³⁴⁷.

Efferocytosing macrophages also secrete reparative matricellular proteins and growth factors, such as Spp1 (encoding osteopontin) and TGF-βs²⁵⁸,³⁴⁷. Osteopontin induction is a prominent characteristic of reparative infarct macrophages and is mediated by cytokines (such as IL-10)³⁴⁸ and chemokines (such as CXCL4)³⁴⁹. Macrophage-derived TGF-β on the other hand plays a central reparative role by activating fibroblasts and by exerting autocrine actions on macrophages that trigger a phagocytic profile (through activation of Smad3 signaling) and stimulate expression of anti-inflammatory mediators³³⁵. In addition to their actions in regulation of inflammation and in fibroblast activation, infarct macrophages also support the formation of new blood and lymphatic vessels (Figure 3). In the infarcted heart, neovascularization arises mainly from preexisting endothelial cells starting from the border zone and extending into the necrotic core³⁵⁰,³⁵¹. Infarct macrophages serve as a major source of the potent angiogenic growth factor vascular endothelial growth factor (VEGF) A³⁵². Several mechanisms have been implicated in VEGF upregulation in infarct macrophages. First, efferocytosis induces VEGFs in cardiac macrophages, stimulating formation of both blood and lymphatic vessels³⁵²,³⁵³. Second, binding of components of the provisional matrix network, such as fibronectin, to macrophage integrin α5 on the macrophage surface was found to induce VEGF synthesis and release, enhancing infarct angiogenesis and improving function³⁵⁴. Third, neutrophils stimulate an angiogenic VEGF-expressing macrophage phenotype by secreting Annexin A1 that binds to formyl peptide receptor 2 on cardiac macrophages²⁹⁶.

Figure 3: The temporal dynamics of angiogenesis in healing myocardial infarction.

During the inflammatory phase, pericytes in the infarct zone disassociate from endothelial cells (EC), increasing microvascular permeability. During the proliferative phase, efferocytosis and components of the provisional extracellular matrix (such as fibronectin) stimulate an angiogenic phenotype in macrophages, promoting expression of Vascular Endothelial Growth Factor (VEGF) and deposition of angiogenic matricellular proteins. Macrophages and fibroblasts contribute to an angiogenic environment, resulting in formation of abundant microvessels that lack a mural cell coat. As the scar matures, pericytes are recruited through pathways involving PDGF-BB-PDGFRβ and TGF-β–TGFBR2 interactions, thus forming coated mature microvessels that stabilize the scar. PDGF, Platelet=Derived Growth Factor; TGF, Transforming Growth Factor.

4.2.2. Dendritic cells:

Using cardiac lymphatics, immune cells traffic to draining mediastinal lymph nodes³⁵⁵, where dendritic cells (DC) act as antigen presenting cells linking innate to adaptive immune responses³⁵⁶,³⁵⁷. Like macrophages, myocardial DCs are a heterogenous cell population, albeit much less abundant: Two conventional DC (cDC) subsets, CD103+ cDC1 cells and CD11b+ cDC2 cells, a PDCA1+ plasmacytoid (p)DC population of non-monocyte origin, and a population of CD64+ monocyte-derived cells (moDC) are found in the heart by differential marker expression³⁵⁸,³⁵⁷. During the proliferative phase of cardiac repair, cDC and moDC numbers peak in the heart and in draining mediastinal lymph nodes (mLN), with moDC dominating the infarct infiltrate and cDC2 dominating the mLN infiltrate. Post-MI matured and activated cDC2 are most efficient but not exclusive in presenting cardiomyocyte self-antigens to auto-reactive T cells in the mLN³⁵⁷. Controversies exist regarding the role of cDC in MI, likely arising from the use and interpretation of different experimental models. Diphtheria toxin-mediated depletion in CD11c-DTR mice resulted in enhanced and prolonged inflammation in the infarct and adverse cardiac remodeling, suggesting a central role for DCs in negative regulation of inflammation³⁵⁹. However, non-selective DC depletion and dependence of the model on lethal irradiation and reconstitution may represent potential confounders. In contrast, selective cDC depletion using Zbtb46-DTR mice suppressed inflammation in the infarct and improved cardiac function³⁵⁸. Clec9a-deficiency impairs cross-presentation of cDC1 to cytotoxic CD8+ T cells and improves cardiac remodeling. Depletion of pDC in mice does not impact MI healing significantly³⁵⁸. On the other hand, subcutaneous administration of tolerogenic DC generated ex vivo by exposure to cardiac lysate from MI mice can induce MI-specific regulatory T cells which suppress inflammation and improve cardiac remodeling³⁶⁰. Whether the divergent effects of DC in the heart depend on T cell intermediaries or direct interactions with stromal cells is unknown.

4.2.3. The potential role of basophils and eosinophils

In addition to the rapid and intense infiltration with neutrophilic granulocytes (discussed in section 3.5.7), infarcts also recruit much smaller numbers of basophilic and eosinophilic granulocytes that have been suggested to play important cells in regulating inflammation and repair. Similar to neutrophilia³⁶¹, elevated eosinophil counts after infarction are associated with adverse outcome³⁶², whereas high basophil counts were found to predict better outcomes in acute coronary syndromes³²¹ and worse outcomes in stable coronary artery disease³⁶³. Recruitment of basophils and eosinophils in the infarct lags that of neutrophils, and experimental evidence suggests that both these cell types may exert protective reparative actions after myocardial infarction. Basophils are a major source of IL-4 and IL-13 in the infarct tissue and instruct a reparative phenotype in macrophages³²¹. A similar protective mechanism of paracrine macrophage polarization was proposed for eosinophils infiltrating the infarcted heart³²⁰. Eosinophil-derived IL-4 and eosinophil-associated RNase 1 were found to mitigate cardiomyocyte apoptosis and fibroblast activation³¹⁹. These eosinophil-mediated protective effects can be exploited therapeutically by injecting IL-5 in mice undergoing myocardial infarction³⁶⁴.

4.2.4. T-lymphocytes in cardiac repair

Several studies have suggested important roles for T lymphocytes in repair of the infarcted heart. CD4+ T cells infiltrate the infarcted myocardium during the proliferative phase and play a central role in infarct healing. The role of CD4+ T-cells was studied using CD4 KOs and mice that exhibit defective T cell antigen recognition (OT-II mice). Both mouse strains exhibited perturbed repair after MI, associated with left ventricular rupture and adverse remodeling³⁶⁵. T-cells can be activated through antigens binding to its specific T-cell receptor. The above-mentioned almost identical phenotype in CD4-KO and OTII mice suggests antigen-dependent mechanisms in myocardial healing after MI. By screening over 150 class-II-restricted epitopes, myosin heavy chain (MHC) was identified as a dominant cardiac antigen. To address the role of T-cell dependent MHC-autoimmunity T cells specific to a MYHCA_614-629 (myosin heavy alpha chain-derived peptide), termed TCR-M cells, were transferred in wildtype mice before MI. Those cells accumulated in the myocardium and in mediastinal lymph nodes, which was associated with improved cardiac healing³⁶⁶. Characterization on a single cell level revealed that TCRM cells rapidly differentiated in myosin-specific regulatory CD4+ T cells in the infarcted myocardium. This suppressed in situ inflammation and improved healing while supporting stromal cell function associated to tissue repair³⁶⁷. This is especially interesting before the background that TCRM mice, i.e. mice that have a lifelong exposure to TCRM, develop autoimmune myocarditis and finally a phenotype of lethal dilated cardiomyopathy³⁶⁸ This indicates that short episodes of autoimmunity may be beneficial for healing, in contrast prolonged autoimmunity may result in a maladaptive phenotype.

In addition to CD4+ T cells, several other lymphocyte subsets were found to contribute to cardiac repair by regulating inflammation and by modulating macrophage and fibroblast phenotype. Regulatory T cells (Tregs) infiltrate the infarcted myocardium through chemokine-dependent mechanisms¹⁸⁹. Studies using both genetic and antibody-based approaches to deplete Tregs after MI demonstrated that CD4⁺/Foxp3⁺ Tregs promote repair, attenuating post-infarction inflammation.³⁶⁹,³⁷⁰. The effects of Tregs were attributed to modulation of macrophage phenotype towards an anti-inflammatory profile³⁶⁹ and to inhibitory effects on fibroblast-mediated protease expression³⁷⁰. Other studies have suggested important reparative roles for invariant Natural Killer T cells (iNKT cells) in reperfused³⁷¹ and in non-reperfused MI³⁷², mediated in part through release of anti-inflammatory cytokines, such as IL-10. The role of CD8+ T cells in the infarcted heart remains poorly defined. Absence of functional CD8+ cells was associated with severe healing defects and increased inflammation after MI, resulting in left ventricular rupture³⁷³.. However, surviving mice with deficient CD8+ cells had improved post-MI remodeling. The basis for these conflicting observations is unclear.

Thus, several T cell populations (including CD4+ T cells, Tregs, iNKT cells and possibly also CD8+ T cells) are involved in negative regulation of post-infarction inflammation. In contrast, CD4⁻γδT-cells were found to promote neutrophil and macrophage infiltration and exert detrimental effects in cardiac remodeling³⁷⁴. Maladaptive phenotypic changes in T cells have also been implicated in progression of chronic post-infarction heart failure³⁷⁵,³⁷⁶

4.2.5. Activated fibroblasts mediate repair of the infarcted heart.

Expansion of the fibroblast population is the hallmark of the proliferative phase of cardiac repair. In the healing infarct, activated fibroblasts are the main source of structural collagens ³⁷⁷,³⁷⁸, and mediate reparative fibrosis, forming an organized scar that protects the ventricle from catastrophic rupture⁸. Infarct fibroblasts exhibit remarkable heterogeneity. Emergence of fibroblast-like cells that express contractile proteins (such as α-smooth muscle actin (α-SMA)), termed myofibroblasts, is a common characteristic of healing wounds in many different organs, including the heart³⁷⁹,³⁸⁰. Although infarct myofibroblasts exhibit increased fibrogenic acitivity³⁸¹,³⁷⁷,³⁸², the basis for their accentuated matrix-synthetic capacity is unclear, and is not caused by increased α-SMA expression³⁸³. Over the last 5 years, single cell transcriptomic studies have contributed additional insights into the diversity of infarct fibroblasts³⁸⁴, and defined clusters with distinct transcriptional profiles²⁵⁷. High expression of Cthrc1 (encoding the glycoprotein collagen triple helix repeat containing 1) identified a subpopulation of reparative infarct fibroblasts that protect from rupture³⁸⁵. Expression of Cilp (the gene encoding the matricellular protein Cartilage Intermediate Layer Protein 1 (CILP1)) may identify another activated fibroblast subpopulation³⁸⁶ with proliferative properties and a role in maladaptive fibrotic remodeling after infarction³⁸⁷.

The origin of the expanding population of infarct fibroblasts has been hotly debated. Although early investigations suggested that endothelial cells³⁸⁸, or myeloid cells (including hematopoietic myeloid progenitors and macrophages)³⁸⁹,³⁹⁰ may undergo fibroblast conversion after infarction, accounting for significant fractions of infarct fibroblasts, more recent investigations using robust lineage tracing strategies showed that the majority of infarct fibroblasts and myofibroblasts originate from resident fibroblast populations that proliferate in the growth factor-rich environment if the infarct³⁹¹,³⁹². Lineage tracing studies coupled with single cell transcriptomics demonstrated that a small fraction (~4-5%) of infarct fibroblasts are derived from pericytes ²⁷⁹. Although myeloid cells and vascular cells exhibit very limited conversion to fibroblasts, their contribution to fibroblast activation is significant, due to their high expression of fibrogenic growth factors and matricellular proteins²⁷⁹,²⁷⁸,³⁴⁸.

4.2.6. Molecular signals that activate infarct fibroblasts.

Fibroblast activation in healing infarcts involves the co-operation of a broad range of mediators, including neurohumoral mediators, cytokines, growth factors, and matricellular proteins (Figure 4). Angiotensin II potently activates infarct fibroblasts, stimulating fibroblast proliferation promoting myofibroblast conversion, and inducing synthesis of extracellular matrix proteins. These fibroblast-activating effects of angiotensin II are mediated through the angiotensin II type 1 receptor (AT1R) ³⁹³,³⁹⁴,³⁹⁵,³⁹⁶; in contrast the AT2 receptor has been suggested to play an anti-fibrotic role, suppressing AT1R-mediated fibroblast activation³⁹⁷,³⁹⁸. Although angiotensin II is a potent activator of fibroblasts, AT1R blockade in both patients with MI and in experimental models of infarction not only does not perturb repair but exerts protective actions³⁹⁹. To what extent the protective effects of AT1R blockade are mediated through inhibition of maladaptive fibrosis is unclear. Angiotensin II has a broad range of effects on many different cell types, including cardiomyocytes, macrophages, vascular cells and fibroblasts that may be involved in adverse post-infarction remodeling.

Figure 4: Fibroblast activation in cardiac repair.

During the proliferative phase of infarct healing, pericytes, macrophages and lymphocytes secrete a broad range of growth factors (such as Platelet-Derived Growth Factors (PDGF), Fibroblast Growth Factors (FGF) and Transforming Growth Factor (TGF)-β) and fibrogenic cytokines (such as Interleukin (IL)-4, IL-6, IL-10, IL-13) that stimulate fibroblasts promoting myofibroblast conversion and activation of a matrix-synthetic, matrix-preserving phenotype. Neurohumoral mediators (such as angiotensin II) and matricellular proteins that enrich the extracellular matrix (ECM) network also contribute to fibroblast activation. Scar maturation is associated with loss of myofibroblast features and acquisition of a matrifibrocyte phenotype by infarct fibroblasts. Fibroblast apoptosis may also contribute to de-activation of the fibrogenic response.

4.2.7. Growth factors: the role of TGF-βs, Fibroblast Growth Factors (FGF)s and PDGFs.

Several fibrogenic growth factors have been implicated in activation of fibroblasts in the infarcted heart. TGF-β is upregulated and activated in healing infarcts and plays a central role in activation of reparative fibroblasts⁴⁰⁰. Matricellular proteins⁴⁰¹, proteases⁴⁰² and αv integrin-mediated actions⁴⁰³,⁴⁰⁴ generate bioactive TGF–β in pericellular areas, thus contributing to a spatially localized fibrogenic response. Active TGF-β transduces signaling through members of the Smad family of transcription factors, or via non-Smad cascades⁴⁰⁵. In infarct fibroblasts, the Smad3 cascade plays a central role in activation of a reparative fibroblast phenotype and in formation of well-aligned arrays of activated myofibroblasts that protect the infarcted heart from rupture and from adverse remodeling⁴⁰⁶,⁴⁰⁷. Smad3-dependent fibroblast activation triggers an integrin-dependent oxidative response that plays a central role in formation of a well-organized matrix network that protects the infarcted heart⁴⁰⁷. In contrast to the central involvement of fibroblast-specific Smad3 signaling in cardiac repair after infarction, the Smad2 pathway does not play an important role in reparative fibrosis of the infarcted heart⁴⁰⁸.

Several additional growth factors are critical regulators of fibroblast phenotype in vitro; however, their in vivo role remains poorly defined. Expansion of the mast cell population during the proliferative phase of infarct healing⁴⁰⁹ increases local expression of fibrogenic tryptase⁴¹⁰,²⁶⁴ and of chymase, a protease involved in local generation of angiotensin II⁴¹¹. Members of the fibroblast growth factor (FGF) family (such as FGF2) potently activate fibroblast proliferation and migration⁴¹²,⁴¹³; however studies investigating the role of their effects on infarct fibroblasts in vivo are lacking. Moreover, FGF signaling also exert angiogenic effects⁴¹⁴ and protective actions on cardiomyocytes⁴¹⁵,⁴¹⁶ that may outweigh its fibrogenic actions. The Platelet-derived growth factor (PDGF) family has also been implicated in cardiac fibroblast activation⁴¹⁷, promoting their survival and proliferation⁴¹⁸,⁴¹⁹,⁴²⁰. It has been suggested that PDGF signaling may stimulate acquisition of a myofibroblast phenotype in mesenchymal stem cells (MSCs), thus contributing to the expansion of fibroblasts after infarction⁴²¹. In vivo experiments using antibody neutralization strategies suggested that PDGFRα signaling may promote fibroblast activation, whereas PDGFRβ may be involved in infarct angiogenesis⁴¹⁷. The isoform-specific effects of PDGFs on the infarcted heart may have important therapeutic implications. Administration of PDGF-AB in a porcine model of MI was found to exert protective actions, improving the mechanical properties of the scar, and promoting angiogenesis⁴²².

Fibroblast activation during the proliferative phase of infarct healing involves not only the effects of secreted growth factors and neurohumoral mediators, but also stimulation of extracellular matrix-mediated pathways. Macrophages and interstitial cells produce large amounts of matricellular proteins, specialized matrix proteins that do not play a major structural role, but transform the matrix into a signaling hub, interacting with cytokines, growth factors and proteases and modulating signaling through a broad range of integrin and non-integrin cell surface receptors⁴²³. Tenascin-C, the thrombospondins, osteopontin, osteoglycin, periostin, SPARC (secreted protein acidic and rich in cysteine), cartilage intermediate layer protein 1 (CILP)1, and ED-A fibronectin enrich the healing infarct and mediate fibroblast activation and myofibroblast conversion⁴²⁴,⁴²⁵,⁶³,⁴²⁶,⁴²⁷,⁴²⁸,⁴²⁹,⁴³⁰,³⁸⁷. The matricellular proteins do not act exclusively on fibroblasts, but also modulate macrophage and vascular cell phenotype and function.

4.2.8. Fibroblasts and arrhythmogenesis after infarction

In addition to their reparative functions and their role in maladaptive fibrosis after MI, infarct fibroblasts may also be implicated in arrhythmia generation⁴³¹. Several studies have suggested the presence of heterocellular electrical coupling between fibroblasts and cardiomyocytes in the infarct border zone that involves formation of gap junctions⁴³²,⁴³³,⁴³⁴,⁴³⁵. A recent study used a model of exogenous optically-stimulated depolarization to demonstrate that fibroblast-cardiomyocyte coupling can generate arrhythmias after MI⁴³¹. Heterocellular coupling of infarct fibroblasts and border zone cardiomyocytes was suggested to result from synergistic actions between gap junctional proteins and a poorly understood non-gap junctional mechanism (termed ephaptic), involving generation of electrical fields in the interstitium between the coupled cells⁴³¹. Given the close spatial association between fibroblasts and cardiomyocytes in the infarct border zone, this is a plausible mechanism; however, whether it plays a significant role in the arrhythmic events that occur in infarcted and remodeling hearts is not known.

5. The maturation phase.

5.1. Endogenous inhibitory pathways suppress fibroblast activation

Although fibroblast-mediated deposition of structural matrix proteins plays a central role in repair of the infarcted heart, excessive, prolonged or expansive fibroblast activation can have catastrophic consequences, resulting in adverse fibrotic remodeling and causing dysfunction and heart failure progression. Thus, the inhibitory signals leading to formation of a mature scar are of central significance in protection from heart failure progression. During the infarct maturation phase, myofibroblasts reduce their matrix-synthetic activity and ultimately disassemble the α-SMA+ stress fibers³⁸². Although a fraction of infarct myofibroblasts may undergo apoptosis⁴³⁶, the majority of the activated infarct fibroblasts convert to matrifibrocytes, cells with low levels of a-SMA and high expression of cartilage and tendon ECM proteins⁴³⁷. The signals responsible for de-activation of infarct myofibroblasts remain enigmatic. However, several negative regulators of fibroblast activation have been identified and found to play an important role in protecting the heart from adverse remodeling.

TGF-β-mediated fibroblast activation stimulates expression of the inhibitory Smads, Smad6 and Smad7, important negative regulators of TGF-β superfamily signaling⁴³⁸. Smad7 is markedly upregulated in infarct myofibroblasts and protects from heart failure, attenuating adverse remodeling and fibrosis, by restraining collagen accumulation and myofibroblast conversion. The anti-fibrotic effects of Smad7 are only partially mediated through inhibition of TGF-β -induced Smad-dependent and non-Smad-dependent cascades, but also involve direct inhibitory actions of Smad7 on the receptor tyrosine kinase Erbb2⁴³⁹. Associative evidence links several other anti-fibrotic signals with suppression of fibroblast activity in the infarcted heart. Induction of the nuclear TGF-β repressor c-Ski during the maturation phase of infarct healing may be involved in fibroblast de-activation⁴⁴⁰,⁴⁴¹. Activation of the intracellular kinase AMP-activated protein kinase (AMPK)α1 may restrain fibroblast activation and proliferation following MI through effects that were attributed to connexin 43 upregulation⁴⁴². Increased HIF-1a expression in infarct fibroblasts was found to restrain their activation and proliferation, thus preventing excessive fibrotic remodeling after MI⁴⁴³. Moreover, changes in the composition of the extracellular matrix may also transduce de-activating signals. Collagen V was found to restrain integrin expression in infarct fibroblasts, thus attenuating their activity and reducing scar expansion⁴⁴⁴. It is plausible to hypothesize that suppression of fibroblast activity may involve the co-operation of several different pathways that may induce fibroblast quiescence, while promoting a matrix-preserving fibroblast phenotype involve in scar maintenance.

5.2. Coating of infarct neovessels with mural cells mediates vascular maturation.

Scar maturation is accompanied by dynamic changes in the infarct neovasculature (Figure 3). The robust VEGF-driven angiogenic response that characterizes the proliferative phase^{417, 445}, is followed by suppression of angiogenesis and recruitment of mural cells (pericytes and vascular smooth muscle cells) that coat and stabilize the infarct neovessels. Investment of the neovessels with a mural coat involves activation of PDGF-PDGF receptor-β (PDGFR-β)⁴¹⁷ and TGF-β²⁷⁹ signaling pathways Defective mural cell recruitment results in persistence of pro-inflammatory neovessels in the infarct, accentuating post-infarction inflammation and promoting dysfunction.

6. Targeting the inflammatory and reparative response in MI.

Considering their diverse roles in injury, repair and fibrosis, inflammatory and fibrogenic mediators are challenging therapeutic targets in the context of MI. Early attempts to inhibit post-MI inflammation for therapeutic gain were based on the concept of cytotoxic inflammatory injury. A large body of evidence using pharmacologic interventions suggested that blockade of mediators involved in post-MI inflammation (such as the leukocyte integrins) markedly reduced the size of the infarct after infarction³¹¹. However, 3 small clinical trials using anti-integrin approaches failed to reduce the size of the infarct in MI patients⁴⁴⁶,⁴⁴⁷,⁴⁴⁸. Administration of pexelizumab, an antibody targeting the complement system, an upstream activator of the innate immune response, produced equally disappointing results in a large clinical trial in STEMI patients⁴⁴⁹. Moreover, treatment of acute coronary syndrome (ACS) patients with the P-selectin inhibitor inclacumab reduced the release of enzymes associated with cardiac injury, but was associated with trends towards worse clinical outcome⁴⁵⁰,⁴⁵¹. Despite these early disappointments, interest in myocardial inflammation continued to increase, fueled by exciting developments in the field of immunology and new insights into the molecular signals underlying the inflammatory response. The recent experience with anti-cytokine approaches offered a glimmer of hope regarding the future of anti-inflammatory therapies in MI patients.

In the Canakinumab Anti-Inflammatory Thrombosis Outcomes Study (CANTOS) trial, treatment of patients with prior MI and elevated hsCRP (suggestive of active inflammation) with the anti-IL1β monoclonal antibody canakinumab reduced the risk of the composite endpoint (death, non-fatal MI, or non-fatal stroke) by 15% in comparison to standard treatment⁴⁵². Although the study was designed to test the hypothesis that IL-1β-mediated inflammation may be involved in atherothrombotic complications, a pre-specified exploratory analysis showed that IL-1β inhibition may reduce heart failure hospitalizations or heart failure-related death⁴⁵³. The CANTOS data along with early phase trials suggesting protection of STEMI patients from adverse remodeling upon IL-1 blockade with anakinra⁴⁵⁴,⁴⁵⁵,⁴⁵⁶ support the role of IL-1 as a therapeutic target in MI. Other pro-inflammatory cytokines, such as IL-6, may also be involved in adverse remodeling after infarction and may be promising therapeutic targets. In the ASSAIL-MI trial, IL-6 receptor inhibition with tocilizumab increased myocardial salvage in STEMI patients⁴⁵⁷. In addition to approaches selectively targeting specific cytokines, repurposing anti-inflammatory agents with broader mechanisms of action has also been attempted. In the COLCOT trial, low-dose colchicine administration within 30 days after MI reduced major adverse cardiovascular events by 22%⁴⁵⁸. Thus, colchicine has become the first guideline-recommended anti-inflammatory medication in cardiovascular secondary prevention. Whether the beneficial effects of colchicine involve improved repair and attenuated remodeling, in addition to a reduction in atherothrombotic events is not known.

Strategies targeting specific cell types may also hold promise for patients with MI. Building on promising results of B-cell inhibition in experimental MI models³⁰⁹, safety and feasibility of B cell depletion using the monoclonal anti-CD20 antibody rituximab has been demonstrated⁴⁵⁹, paving the way for outcome trials. Chimeric antigen receptor (CAR)+ T cells engineered to target activated fibroblasts have been suggested to improve function in models of interstitial fibrosis induced through neurohumoral activation⁴⁶⁰,⁴⁶¹, and may also hold promise in MI patients with overactive fibrogenic responses. Moreover, amplification of Tregs using a superagonistic anti-CD28 antibody may protect the infarcted heart by suppressing excessive inflammation⁴⁶²,⁴⁶³.

In addition to approaches inhibiting pro-inflammatory mediators and cells to reduce their deleterious effects on the remodeling heart, interventions aimed at enhancing the reparative actions of growth factors have also been suggested. For example, PDGF-AB administration has been suggested to improve repair in both rodent and large animal models of infarction. Although the risks of growth factor supplementation approaches may include accentuation of fibrosis, cautious and selective administration in patients with reparative perturbations may hold promise. CXCL12 treatment may also promote recruitment of reparative cells, increasing infarct angiogenesis, and improving left ventricular mechanics^{216, 219, 220, 228}.

Although several different strategies targeting the inflammatory and reparative responses after MI hold promise, therapeutic translation remains challenging. The duality of the effects of inflammatory mediators after infarction and their involvement in both injurious and reparative responses complicate therapeutic implementation. Rational design of therapeutic interventions targeting inflammatory pathways after MI should take into account important temporal and spatial considerations. Our understanding of the temporal sequence of events during the three phases of cardiac repair suggests that there is a narrow window of therapeutic opportunity for safe and effective targeting of specific inflammatory signals.

The remarkable pathophysiologic heterogeneity of MI patients poses another major challenge for therapeutic translation⁴⁶⁴,³¹¹. Heart failure progression and adverse remodeling after MI are only in part dependent on the extent of initial injury and the size of the infarct. The inflammatory, reparative and fibrogenic responses after MI are greatly affected by many other factors, including age, sex, genetic substrate, the presence or absence of concomitant conditions (such as diabetes, obesity, hypertension etc), the pattern of atherosclerotic disease, and the use of other medications that may interfere with inflammatory or fibrogenic signaling. Some patient subpopulations may have defects in negative regulation and resolution of inflammation and perturbations in fibroblast activation, thus exhibiting prolonged or expanded inflammatory responses and perturbed scar formation. Others may have overactive fibrotic reactions, resulting in increased stiffness and diastolic dysfunction. This pathophysiologic heterogeneity cannot be recapitulated by our animal models, which are better suited for dissection of underlying cellular mechanisms and not as predictors of therapeutic efficacy⁴⁶⁴. What is needed is a pathophysiologic classification of patients on the basis of clinical features, biochemical profile, and/or functional responses to identify individuals with accentuated post-infarction inflammatory responses that may benefit from targeted anti-inflammatory strategies, subjects with overactive fibrogenic pathways who may require TGF–β inhibition and patients with impaired reparative fibrosis who may be treated with a fibrogenic growth factor, such as PDGF-AB. For example, patients with diabetes have a high incidence of diastolic dysfunction following MI, despite a smaller infarct size and comparable systolic dysfunction⁴⁶⁵. On the other hand, aging is associated with a prolonged inflammatory phase and defective reparative growth factor signaling, resulting in formation of scars with low levels of collagenous matrix⁴⁶⁶. In addition to clinical criteria, biomarkers reflecting inflammatory and fibrogenic activation⁴⁶⁷,⁴⁶⁸ and imaging approaches informing on the extent of fibrotic remodeling⁴⁶⁹,⁴⁷⁰,⁴⁷¹ may be used to design personalized therapeutic approaches.

7. Conclusions:

Over the last 30 years, both experimental and clinical studies contributed to our understanding of the role of inflammatory and reparative pathways in infarct healing and post-infarction remodeling. The field survived early skepticism and was strengthened by the failures of initial approaches to target inflammation in MI patients. As several promising candidates for intervention have emerged, there is enough evidence to support cautious optimism on the future of therapeutic translation,