Tumor necrosis factor superfamily molecules in acute coronary syndromes

Abstract

Accumulating evidence suggests that inflammatory pathways play an essential role in all stages of atherogenesis. Inflammatory processes are not only involved in plaque progression, but seem also to play a critical role in plaque rupture. Members of the tumor necrosis factor (TNF) superfamiliy are potent regulators of inflammation and cell survival and consist of 20 ligands that signal through 29 different receptors. Several lines of evidence suggest that TNF-related molecules are involved in the development of acute coronary syndromes (ACS). Most, convincing evidence exists for CD40 ligand-CD40 interaction, but several other members of the TNF superfamily seem also to be involved in this immune-mediated promotion of plaque instability, including LIGHT, receptor activator of nuclear factor κB ligand, and TNF-α. These plaque destabilization pathways involve the bidirectional interaction between platelets and endothelial cells/monocytes, activation of vascular smooth muscle cells, and co-stimulatory effects on T cells, promoting inflammation, thrombus formation, matrix degradation, and apoptosis. TNF-related pathways could contribute to the non-resolving inflammation that characterizes atherosclerosis, representing pathogenic loops that are operating during plaque rupture and the development of ACS. These TNF-related molecules could also represent attractive new targets for therapy in this disorder.

Key words::

• Acute coronary syndrome
• atherosclerosis
• inflammation
• tumor necrosis factor

Abbreviations
ACS	=	acute coronary syndromes
ApoE	=	apolipoprotein
APRIL	=	a proliferation-inducing ligand
CAD	=	coronary artery disease
DR	=	death receptor
FADD	=	Fas-associated death domain protein
GITR	=	glucocorticoid-induced TNF receptor familyrelated protein
IL	=	interleukin
IFN	=	interferon
L	=	ligand
LDL	=	low-density lipoprotein
LIGHT	=	lymphotoxin-like, exhibits inducible expression, and competes with HSV glycoprotein D for HVEM, a receptor expressed by T lymphocyte
LT	=	lymphotoxin
MCP	=	monocyte chemoattractant protein
MI	=	myocardial infarction
MMP	=	matrix metalloproteinases
NFκB	=	nuclear factor κB
OPG	=	osteoprotegerin
PAR	=	proteinase-activated receptor
PBMC	=	peripheral blood mononuclear cells
ox	=	oxidized
R	=	receptor
RANK	=	receptor activator of NFκB
s	=	soluble
SMC	=	smooth muscle cells
TF	=	tissue factor
TGF	=	transforming growth factor
TNF	=	tumor necrosis factor
TNFRSF	=	tumor necrosis factor receptor superfamily
TNFSF	=	tumor necrosis factor superfamily
TRADD	=	TNF receptor-associated death domain protein
TRAF	=	TNF-associated factor
TRAIL	=	TNF-related apoptosis-inducing ligand
Tregs	=	regulatory T cells
TWEAK	=	TNF-like weak inducer of apoptosis

Key messages

• Members of the TNF superfamiliy are potent regulators of inflammation and cell survival and consist of 20 ligands that signal through 29 different receptors.

• Several lines of evidence suggest that TNF-related molecules are involved in atherogenesis and the development of acute coronary syndromes (ACS).

• In relation to ACS, most, convincing evidence exists for CD40 ligand-CD40 interaction, but several other members of the TNF superfamily seem also to be involved in this immune-mediated promotion of plaque instability, including LIGHT, RANKL, and TNF-α.

• These plaque destabilization pathways involve the bidirectional interaction between platelets and endothelial cells/monocytes, activation of vascular smooth muscle cells, and co-stimulatory effects on T cells, promoting inflammation, thrombus formation, matrix degradation, and apoptosis.

Accumulating evidence suggests that inflammatory pathways play an essential role in all stages of atherogenesis. Thus, immune cells dominate early atherosclerotic lesions, their effector molecules accelerate progression of the plaques, and activation of inflammation can elicit acute coronary syndromes (ACS) (1,2). Although the participation of inflammation in atherogenesis has become widely recognized, the identification and characterization of the different actors and their relative importance are not fulfilled. Also, atherosclerosis is one of several inflammatory disorders characterized by non-resolving inflammation caused by various factors such as persistence of the initiating stimuli (e.g. lipids), excessive inflammatory responses that are forming inflammatory loops (e.g. oxidative stress promoting inflammation which again will induce oxidative stress), impaired function or numbers of regulatory or anti-inflammatory cells (e.g. regulatory T cells (Tregs) and Ly6C^lo macrophages), inadequate production of resolution factors (e.g. interleukin (IL)-10 and transforming growth factor (TGF)-β), or a combination thereof (2–4). However, the pathways and cells that contribute to the maintenance of this persistent inflammation are far from clear.

Inflammation and plaque destabilization

Inflammatory processes are not only involved in plaque progression but seem also to play a critical role in plaque rupture by promoting apoptosis, production of matrix-degrading enzymes, and reactive oxygen species (5). Two major types of physical disruption of the atherosclerotic plaque may occur (6). Superficial erosion of the endothelial monolayer uncovers subendothelial collagen and von Willebrand factor, promoting platelet adhesion and activation, thus making a nidus for platelet thrombus formation with subsequent myocardial infarction (MI), counting for one-quarter of fatal coronary thromboses. The most common mechanism of plaque disruption involves rupture of the fibrous cap. The thrombogenic, lipid-rich core of the plaque is normally sequestered from the blood-flow by the fibrous cap, but upon fissure formation, usually at the shoulders of the plaque, exposure and activation of the coagulation cascade in the blood vessel occur, causing approximately three-quarters of all MI (7). Inflammation is instrumental in both of these mechanisms (7). First, endothelial erosion is linked to the proximity of highly activated subendothelial macrophages that may cause endothelial cell death by apoptosis and also weaken the endothelial intercellular integrity by producing matrix metalloproteinases (MMP). Second, a high content of smooth muscle cells (SMC) is thought to make the plaque less prone to rupture, and cytokine-induced apoptosis of these cells may also weaken the overall integrity of the plaque. Inflammatory mediators may also promote plaque instability by impairing collagen production and activating matrix-degrading enzymes. Third, inflammatory cytokines may promote the development of thrombus formation by inducing platelet activation and by enhancing the level of the pro-thrombotic tissue factor (TF) and decreasing the level of thrombomodulin, an endogenous anti-thrombotic mediator. Typical features of the vulnerable plaque are: a large lipid core occupying at least 50% of the plaque volume, a high density of macrophages, a high TF content, and low content of SMC and collagen in the fibrous cap. Although several cell types may contribute to these features of plaques at future risk, macrophage activation seems to be of particular importance. Interestingly, the recently characterized macrophage subtype (Ly6C^lo), expressing CCR2 (receptor for the pro-atherogenic chemokine monocyte chemoattractant protein (MCP)-1/CCL2) and with an inflammatory and matrix-degrading potential, seems to dominate during plaque rupture (8). Thus, the site of rupture is characterized by enhanced inflammation and often occurs in the macrophage-rich shoulder region. Notably, plaque rupture is also influenced by external factors such as turbulence in passing blood and mechanical stress (9,10). Further, in addition to plaque size, the orientation of remodeling plays a role in plaque vulnerability (11–13). Positive remodeling, that is, the plaque expands into the vessel wall, is associated with unstable angina, and several studies have suggested that in 60%–70% of patients with ACS, the culprit site of the acute event had less than 70% (often under 50%) of vessel diameter narrowing (14). Thus, plaques producing non-flow-limiting and less than severe stenosis account for more cases of ACS events than the severe stenotic plaques producing symptoms of stable angina. Moreover, recurrent coronary events, frequently occurring in ACS, are unrelated to the culprit lesion in almost half of the cases (15), supporting the view of wide-spread diseased coronary vessels with multiple vulnerable plaques, reflecting wide-spread coronary inflammation (16,17). It is also important to have in mind that, despite the body of evidence of a link from plaque vulnerability, rupture, and ACS, many plaque ruptures cause neither occlusion of the vessel nor clinical symptoms (6).

The tumor necrosis factor (TNF) superfamiliy: potent regulators of inflammation and cell survival

Based on their abilities to kill murine fibrosarcoma cells, tumor necrosis factors gained their name in 1975, and about 10 years later the two archetypical TNF superfamily (TNFSF) members were isolated and characterized, namely TNF-α/TNFSF1A and lymphotoxin β (LT-β)/TNFSF3. Research during the last three decades have revealed several additional members, and the superfamily now consists of 20 ligands that signal through 29 different receptors (Figure 1) (18). The TNF receptor superfamily (TNFRSF) is expressed by a wide variety of cells, while the TNFSF ligands are known to be expressed predominately by cells of the immune system, including B cells, T cells, natural killer cells, granulocytes, monocytes, mast cells, and dendritic cells, but importantly and with relevance to atherogenesis also by other cells such as cardiomyocytes, vascular SMC, and endothelial cells (18). The TNFSF have unique structural attributes that couple them directly to pathways for cell proliferation, differentiation, and survival (18).

Figure 1. Ligands in the TNF superfamily (TNFSF) and their corresponding receptors in the TNF receptor superfamily (TNFRSF). Note, the TWEAK receptor, Fnt, and the ligand for the NGF receptor, NGF, are commonly classified as members of the TNFRSF and TNFSF, respectively. (APRIL = a proliferation-inducing ligand; BAFF = B cell-activating factor belonging to the TNF family; BCMA = B cell maturation antigen; DCR = decoy receptor; DR = death receptor; EDA = ectodysplasin; GITR = glucocorticoid-induced TNF receptor family-related protein; L = ligand; LIGHT = lymphotoxin-like, exhibits inducible expression, and competes with HSV glycoprotein D for HVEM, a receptor expressed by T lymphocytes; LT = lymphotoxin; NGF = nerve growth factor; OPG = osteoprotegerin; R = receptor; RANK = receptor activator of NFκB; RELT = receptor expressed in lymphoid tissues; TACI = transmembrane activator and calcium-modulating cyclophilin ligand interactor; TRAIL = TNF-related apoptosis-inducing ligand; TROY = TNFRSF19; TWEAK = TNF-like weak inducer of apoptosis; VEGI = vascular endothelial growth inhibitor; XEDAR = X-linked ectodysplasin receptor).

The TNF superfamily ligands are transmembrane type II proteins, with a C-terminal receptor-binding extracellular domain, that may or may not be shed by proteinases to yield a soluble form (19). Functionally, TNFRSF members can be divided in three groups, according to their intracellular signal domains, containing either a death domain (DD), or a TNF receptor-associated factor (TRAF) domain, or no signal domain. The first group, the ‘death receptors’, recruits intracellular DD containing adaptors, such as Fas-associated DD protein (FADD) and TNFR-associated DD protein (TRADD), which activate the caspase cascade leading to apoptosis (Figure 2) (20). Fas/TNFRSF6, the receptor for Fas ligand (FasL)/TNFSF6, and two of the receptors for TNF-related apoptosis-inducing ligand (TRAIL/TNFSF10) (i.e. TRAIL R1/TNFRSF10A and TRAIL R2/TNFRSF10B) are examples of death receptors in the TNFRSF.

Figure 2. TNFR superfamily ligands can induce apoptosis or cell proliferation and inflammation depending on their intracellular signal domains, containing either a death domain (DD) or a TNF receptor-associated factor (TRAF) domain. The figure shows TRAF2 as a prototypical inflammatory TRAF. (ASK = apoptosis signal-regulating kinase; c-IAP = cellular inhibitor of apoptosis; FADD = Fas-associated DD protein; GCK = germinal center kinase; I = inhibitor; IKK = IκB kinase; JNK = JuAn N-terminal kinase; MAPK = mitogen-activated protein kinase; MEKK = MAPK/extracellular signal-regulated kinase kinase; NF = nuclear factor; NEMO = NFκB essential modulator; NIK = NFκB-inducible kinase; RIP = receptor interacting protein kinase; TAK = TGF-β-activating kinase; TRADD = TNFR-associated DD protein).

The second group of TNFRSF members mediates effects associated with cell differentiation and proliferation; by recruiting TRAF family molecules, these receptors activate the transcriptional factor nuclear factor (NF)κB, a potent inducer of cell survival, proliferation, and inflammation (Figure 2) (21). However, clear lines cannot be drawn between the two groups. Hence, TRAF2, an activator of NFκB which induces anti-apoptotic protein synthesis, binds almost all TNFRSF members, and recent discoveries indicate that any TNF signaling simultaneously activates both apoptotic and cell-survival signals, creating a balance of opposing signals that determines cell fate (18). Thus, while TRAIL and FasL, which are potent activators of death receptors and poor inducers of NFκB, promote apoptosis, receptor activator of NFκB ligand (RANKL/TNFSF11) mainly provides survival signals through NFκB. However, NFκB does not only promote anti-apoptotic signaling, it also regulates pro-apoptotic pathways through regulation of death receptors (DR1-6), and death receptor ligands such as FasL and TRAIL (22–25), illustrating the complexity of TNF-mediated cell survival signaling. In addition, the results of TNFSF ligand signaling is clearly also dependent on co-stimuli, time of exposure, and state of cellular activation at the time of activation (26).

The last group of receptors conducts no intracellular signaling but are soluble decoy receptors, i.e. providing a level of regulation by competing with signal transducing receptors for ligands (18). Moreover, some receptors (e.g. CD27, CD30, CD40, CD95, TNF receptor I (TNFRI), and TNFRII) can be found in a soluble form, with inhibitory but potentially also stabilizing effects on their corresponding ligand at least partly depending on the molar ratio between the ligand and soluble receptor (27).

TNFSF-related pathology: too much or too little of a good thing

The TNFSF regulates immunity at several levels, e.g. organization of lymphoid architecture (28) and controlling the activity and survival of cytotoxic effector cells (29). Whereas they are physiologically crucial for normal responses, any inappropriate presence is harmful, and, despite tight regulation, directly pathological contribution by several TNFSF ligands is described in numerous diseases. While TNF-α is an essential element in host defense, excessive TNF-α activity plays a pathogenic role in several inflammatory disorders. Thus, therapeutic approaches that inhibit TNF-α activity (i.e. soluble TNF receptor fusion protein or anti-TNF-α antibodies) have been successful in the treatment of diseases such as severe rheumatoid arthritis (30), inflammatory bowel disease (31), and psoriasis (32), and TNF-α and related molecules have also been implicated in the pathogenesis of systemic vasculitis (33). Moreover, excessive TNF-α production has been implicated in the pathogenesis of various infectious disorders ranging from HIV infection to severe malaria (34,35). With relevance to atherosclerosis, TNF-α has also been implicated in the pathogenesis of several metabolic disorders such as obesity, type 2 diabetes mellitus, metabolic syndrome, and non-alcoholic fatty liver disease (36–38). Also, increased levels of several ligands in the TNFSF have been suggested to be implicated in the development and progression of heart failure (39). On the other hand, TNF-α is of major importance in the host defense against certain intracellular microbes such as Mycobacterium tuberculosis, as illustrated in the increased frequency of mycobacterial infection in patients receiving anti-TNF therapy (40). Another TNFSF ligand, CD40L, is associated with immunodeficiency and autoimmunity due to reduced or increased levels, respectively (41), illustrating that both ‘too much’ and ‘too little’ of these ligands may be harmful. To further complicate the picture, a subnormal inflammatory response can engender a prolonged inflammation. Thus, Hodge-Dufour et al. reported an exaggerated and ultimately lethal inflammatory response of TNF-α^−/− mice to injection of Propionobacterium acnes (42).

TNF-related molecules in atherogenesis

Several-TNF related molecules have been linked to development of atherosclerosis. Experimental studies using atherosclerosis-prone apolipoprotein E-deficient (ApoE^−/−) mice crossed with TNF-α^−/− mice have shown that atherosclerotic lesion size is significantly smaller in the double knock-out than in that of ApoE^−/− mice, associated with decreased expression of adhesion molecules and chemokines (43). However, anti-atherogenic properties of TNF-α have also been reported. Thus, TNFRI-deficient mice fed with atherogenic diet developed larger lesions than wild-type mice, suggesting a protective role of TNFRI signaling (44). A pro-atherogenic role of TNFR1 has also been supported in humans by showing an association between at least one TNFR1 single nucleotide polymorphism (SNP) and exacerbation of aging-related atherosclerosis (45). On the other hand, ApoE^−/− mice that were deficient in TNFRII showed decreased atherosclerosis, suggesting that the two TNF-α receptors may have different function during atherogenesis (46). Studies in gene-modified mice prone to develop atherosclerosis have suggested that also several other TNF-related molecules are involved in atherogenesis. First, blocking of OX40L/TNFSF4-OX40/TNFRSF4 interaction by anti-OX40L antibody has been shown to reduce atherogenesis by inhibition of IL-4-mediated T helper cell type 2-induced isotype switching and subsequent increased levels of antibodies against oxidized low-density lipoprotein (oxLDL) in LDL receptor (LDLR)-deficient mice (47). Second, deficiency of 4-1BB (CD137)/TNFRSF9 was recently shown to reduce the atherosclerotic plaque lesions in both ApoE^−/− and LDLR^−/− mice, attributed to the down-regulation of inflammatory cytokines such as interferon (IFN)-γ, MCP-1, and TNF-α (48). In line with this, treatment of ApoE^−/− mice with a 4-1BB agonist caused enhanced inflammation within the atherosclerotic lesion consisting of T cells with predominance of the CD8⁺ T cell subset (49). Third, the CD40L/TNFSF5-CD40/TNFRSF5 dyad has been strongly linked to atherogenesis, and disruption of the CD40L gene in ApoE^−/− mice abrogated atherosclerosis and caused a stable plaque phenotype (50). A similar pattern has also been found when ApoE-deficient mice with initial plaques or established atheromata were treated with a blocking anti-CD40L antibody (51). Fourth, inactivation of osteoprotegerin (OPG)/TNFRSF11B, a soluble decoy receptor for RANKL that may prevent interaction with its corresponding membrane-bound receptor RANK/TNFRSF11A, has been found to accelerate atherosclerotic plaque progression and calcification in older ApoE^−/− mice (52). On the other hand, treatment with a Fc-OPG fusion protein was found to reduce significantly the calcified lesion area without affecting atherosclerotic lesion size or number in LDLR^−/− mice (53). Fifth, ApoE^−/− Fas^−/− mice showed lupus nephritis, accelerated atherosclerosis, and osteopenia, suggesting a role for Fas in the enhanced atherogenesis in patients with systemic lupus erythematous and other inflammatory rheumatic disorders (54). Sixth, treatment of ApoE^−/− mice with antibodies against TNF-like weak inducer of apoptosis (TWEAK)/TNFSF12 decreased NFκB activation, inflammatory cytokine expression, macrophage infiltration within the atherosclerotic lesions, as well as attenuated vascular and renal injury severity, indicating a pathological role for endogenous TWEAK in atherogenesis and vascular inflammation (55). In addition, several other TNF-related molecules have been found to be regulated in atherosclerotic disorders, both systemically and within the lesion, such as lymphotoxin-like, exhibits inducible expression, and competes with HSV glycoprotein D (gD) for HVEM, a receptor expressed by T lymphocytes (LIGHT)/TNFSF14, and its two receptors HVEM/TNFRSR14 and LT-βR/TNFRSF3 (56,57), and glucocorticoid-induced TNF receptor family-related protein (GITR)/TNFRSF18 and its ligand (GITRL/TNFSF18) (58). Thus, several lines of evidence suggest the involvement of a number of TNF-related molecules in atherogenesis. Some of them seem also to be related to plaque stability and the development of ACS, and this will be discussed in the following paragraphs.

TNF-related ligands are involved in platelet-mediated inflammation

It is well established that platelets contribute to plaque destabilization and ACS by promoting thrombus formation. However, recent studies suggest that these cells also may promote the development of ACS through inflammatory mechanisms (59–61). First, upon activation platelets provide a wide range of growth factors and inflammatory mediators. Second, platelets do not only contain and express inflammatory mediators but may, upon activation, also induce the expression of such substances (e.g. TNF-α and chemokines) in adjacent cells such as monocytes/macrophages, neutrophils, and endothelial cells. Third, platelets not only promote an inflammatory response in leukocytes and endothelial cells but may also themselves respond to inflammatory mediators produced by these cells. In addition to chemokines, activated platelets have been shown to release, and in some degree express, certain members of the TNFSF (i.e. CD40L, LIGHT, and a proliferation-inducing ligand (APRIL)/TNSF13) (62–64). In contrast to the rapid release of α-granule contents, activated platelets release these TNFSF members in a gradual and long-lasting manner. However, while the gradual release of APRIL seems to be similar to the release pattern of CD40L and LIGHT, APRIL is quite differently regulated. Whereas the release of LIGHT and CD40L involves GP IIb/IIIa-dependent mechanisms and action of metal-dependent proteases as well as actin polymerization, this seems not to be the case for the release of APRIL (64). Whatever the mechanisms, the fact that TNF-related molecules are released from activated platelets clearly contributes to their pathogenic role in ACS.

The CD40L-CD40 dyad: an important mediator of plaque destabilization

Several lines of evidence suggest a role for CD40L-CD40 interaction in plaque destabilization and the development of ACS (Figure 3). First, studies in gene-modified mice have shown that CD40L deficiency and CD40L inhibition result in a stable plaque phenotype with only few inflammatory cells and a high percentage of extracellular matrix (ECM), which is reminiscent of a clinically favorable stable atherosclerotic plaque in humans (50,51). Second, experimental studies in CD40L^−/− mice show that this ligand stabilizes arterial thrombi by an integrin-dependent mechanism and that the absence of CD40L may delay arterial occlusion in vivo (65). Third, in vitro CD40L has been shown to promote an inflammatory and pro-thrombotic phenotype in monocytes and endothelial cells and to increase MMP activity in monocytes and vascular SMC (66–69), all effects with relevance to plaque destabilization. Fourth, since the first report by Aukrust et al. (70), several studies have shown increased circulating levels of soluble (s) CD40L in ACS as compared with stable coronary artery disease (CAD) and healthy controls, as recently reviewed by Antoniades et al. (71). There seems to be a gradual increase in sCD40L levels along with ACS progression (70,72,73), and, interestingly, transcoronary sCD40L levels show an early peak after onset of ACS (74,75). In addition, sCD40L levels have been reported to be higher in the culprit coronary artery than in the peripheral circulation, potentially reflecting the activation of a potent local inflammatory process (75–78). Fifth, there are a number of studies suggesting that high levels of sCD40L may give prognostic information in patients with ACS (reviewed in (71)). Thus, Heeschen et al. showed that in 1,088 patients with ACS, elevation of sCD40L levels indicated an increased risk of cardiovascular events, potentially identifying a subgroup of patients at high risk who are likely to benefit from antiplatelet treatment (79). In the MIRACL study, ACS patients were randomized to atorvastatin and placebo in a double-blind fashion. In the 2,908 patients where plasma levels of sCD40L were measured, high levels (> 90 percentile) were identified as a risk factor for recurrent cardiovascular events in the placebo but not in the atorvastatin group (71,80). Seventh, genetic polymorphisms of platelet glycoprotein Ia has been linked to increased risk for premature MI, and, notably, this seems to be related to enhanced release of sCD40L during the acute phase of premature MI (81).

Figure 3. The figure shows the diverse effects of CD40L in relation plaque destabilization and development of acute coronary syndromes. The figure illustrates that soluble CD40L is released from activated platelets and in some degree also from activated T cells. Similar effects, and even more potent effects, are induced by membrane-bound CD40L expressed by platelets, T cells and macrophages, but also by vascular smooth muscle cells and endothelial cells. (MMP = matrix metalloproteinase).

However, the role of CD40L-CD40 in plaque destabilization is still debated. Thus, two large studies indicated that sCD40L had no predictive values in ACS patients (82,83), and it has been suggested that technical problems with the measurements of sCD40L in circulation (e.g. release from platelets ex vivo) could contribute, at least partly, to the somewhat different results that have been reported (84). Moreover, while it has been shown that membrane-bound CD40L, expressed on the platelet surface and T cells, can be involved in the initiation of an inflammatory response in the vessel wall by inducing the expression of adhesion molecules and the secretion of chemokines in endothelial cells, there has been disagreement whether sCD40L from platelets is biologically active. However, Zhang et al. reported that sCD40L released by platelets is able efficiently to activate fibroblasts in vitro, as indicated by regulation of cyclo-oxygenase 2 expression (85), and we have previously shown that serum-derived sCD40L could induce blood mononuclear cells to produce MCP-1 (70). Also, as for other molecules, the effect of sCD40L seems to be dose-dependent. For example, whereas some results revealed that treatment of cultured endothelial cells with 0.1 μg/mL sCD40L exerts anti-apoptotic, proliferative, and pro-angiogenic effects (86), others have reported negative effects of sCD40L (0.1 to 10 μg/mL) on endothelial cells in vitro, in particular induction of apoptosis and increased oxidative stress (87,88). In relation to serum/plasma levels, even 0.1 μg/mL is high, but whether enhanced concentrations of sCD40L may build up locally at the affected area of the vessel wall is not clear. Moreover, it seems that CD40L may mediate CD40-independent effects through the integrin Mac-1 (89), and Bavendiek et al. suggested that CD40L derived from non-hematopoietic cell types may be of even more pathogenic importance in atherogenesis than T cell or platelet-derived CD40L (90), further underscoring the complex biology of CD40L. Finally, although plasma levels sCD40L seem to reflect platelet activation in atherosclerosis and related disorders, we cannot exclude that in healthy individuals or other inflammatory disorders other cellular sources (e.g. monocytes and T cells) could also contribute to the circulating levels of this molecule, at least partly reflecting the degree of systemic inflammation.

Despite some uncertainties, CD40L, which has the unique property of promoting inflammation, thrombus formation, and matrix degradation, activating both endothelial cells, leukocytes, platelets, and vascular SMC, operating in a self-perpetuating feedback loop, should be further investigated as a therapeutic target in atherosclerotic disorders. However, the complete inhibition of CD40L-CD40 signaling in atherosclerosis is not therapeutically feasible because long-term treatment will compromise systemic immune responses and also entails thromboembolic complications. Very recently, Lutgens et al. showed that TRAF6 was essential for the pro-atherogenic effects of CD40 activation, and these authors suggested that inhibition of the TRAF6 binding site on CD40, using small molecules or an antagonizing CD40 antibody that changes the conformation of the CD40-TRAF binding sites, may be an attractive therapeutic approach (91).

LIGHT: a potential ‘new’ mediator in plaque destablization

In addition to CD40L, it has recently been reported that activated platelets release and express LIGHT, another TNFSF ligand, and, importantly, platelet-derived LIGHT seems to be biological active, potentially contributing to plaque rupture and the development of ACS (Figure 4) (63). Thus, recombinant soluble LIGHT at 1–100 ng/mL promoted the expression of the adhesion molecules E-selectin and vascular cell adhesion molecule 1 as well as the chemokines MCP-1 and IL-8 in endothelial cells, and, importantly, these enhancing effects were also seen with platelet-derived LIGHT (63). In fact, although the amount of soluble platelet-derived LIGHT was relatively small as compared to for example sCD40L (i.e. pg versus ng levels), neutralizing antibody against LIGHT was found to attenuate significantly the platelet-mediated enhancing effect of MCP-1 and IL-8 (63). Moreover, we have recently identified proteinase-activated receptor (PAR)-2 as an inflammatory mediator that was markedly enhanced by LIGHT in endothelial cells, acting synergistically with LIGHT to promote enhanced release of IL-8 and MCP-1 (56). This LIGHT-mediated effect on PAR-2 was mediated through the HVEM receptor, involving Jun N-terminal kinase signaling pathways. In addition, LIGHT has been reported to enhance monocyte and macrophage activation, at least partly involving PAR-2-related mechanisms, and this seems also to be the case for platelet-derived LIGHT (56,63). While platelet activation and pro-thrombotic stimuli (e.g. PAR-1 activation) may cause inflammation, inflammation may also cause thrombus formation. Thus, TNF-α and CD40L promote TF expression in monocytes (67,92), and LIGHT has been shown to induce a pro-thrombotic phenotype in monocytes/macrophages and endothelial cells that involves enhanced expression of TF and plasminogen activator inhibitor 1 as well as down-regulation of the potent anti-thrombotic mediator thrombomodulin (57,93). This bidirectional pro-thrombotic and inflammatory interaction between platelets and monocytes/macrophages and endothelial cells, which include several TNF-related molecules, could contribute to a pathogenic loop in plaque destabilization and development of ACS. While we previously have shown enhancing effects of recombinant CD40L on the release of IL-8 and MCP-1 in endothelial cells only at concentrations > 1 μg/mL (69), similar effects of recombinant LIGHT were seen at a concentration of 1 ng/mL (63). Thus, LIGHT seems to be a very potent mediator of platelet-mediated inflammation in monocytes and particularly in endothelial cells. One might speculate that in an inflamed vascular wall soluble LIGHT derived from activated platelets, able to reach high concentrations in the area, will be of importance with regard to initiating platelet-mediated inflammatory responses. Based on its demonstrated potency, even a modest concentration of LIGHT could be of pathogenic importance in atherosclerosis and plaque destabilization. In fact, we have demonstrated that platelets in patients with acute MI express LIGHT in thrombus material obtained at the site of plaque rupture, suggesting that such LIGHT-mediated inflammation is operating in vivo within an inflamed and thrombotic vessel wall (63). Moreover, patients with ACS show enhanced serum levels of LIGHT accompanied by enhanced expression of PAR-2 in peripheral blood mononuclear cells (PBMC), underscoring the relevance of the inflammatory and pro-thrombotic effects of LIGHT to ACS (56). However, data from large patient populations are lacking, and there are no studies on LIGHT in gene-modified mice prone to develop atherosclerosis. Further studies are needed to establish LIGHT as an important mediator of plaque destabilization.

Figure 4. LIGHT, including platelet-derived LIGHT, may promote plaque rupture and thrombus formation through various mechanisms involving interactions between platelets, monocytes/macrophages, T cells, and endothelial cells. (ECM = extracellular matrix; IFN = interferon; MMP = matrix metalloproteinases; PAI = plasminogen activator inhibitor; SMC = smooth muscle cells; TF = tissue factor; Th1 =T helper cell type 1; VCAM = vascular cell adhesion molecule).

The RANKL/OPG/RANK axis and plaque destabilization

Several findings point at the RANKL/OPG/RANK axis as a contributor to plaque destabilization and development of ACS. First, circulating OPG levels appear to be higher in patients with unstable angina and acute MI compared to controls with stable atherosclerosis (94). Second, T cells from patients with unstable angina show enhanced expression of RANKL as compared with stable CAD (94,95). Additionally, during percutaneous coronary intervention, a mechanically induced plaque rupture, PBMC show enhanced RANKL expression (94), further suggesting that increased RANKL expression is a feature of unstable disease. Third, Golledge et al. reported OPG to be expressed at higher levels in symptomatic than in asymptomatic carotid plaques (96), and we have shown enhanced RANKL immunostaining in plaques from ApoE^−/− mice that are lacking the TGF-β receptor on T cells, which develop a more severe and unstable atherosclerotic phenotype relative to ApoE^−/− mice (94). These findings were further supported by strong immunostaining of the RANKL/OPG/RANK triad in thrombus material obtained from the site of plaque rupture in patients undergoing percutaneous coronary intervention (94). Fourth, serum OPG is strongly pred-ictive of long-term mortality and heart failure development in patients with ACS, independent of conventional risk markers (97,98). Fifth, RANKL exhibits several properties with relevance to plaque destabilization such as promotion of inflammatory responses in T cells, induction of chemotactic properties in monocytes, induction of MMP activity in vascular SMC, and RANKL has also been found to have pro-thrombotic properties (94,99–101). Moreover, one condition strongly associated with plaque rupture is calcification, and the involvement of the RANKL/OPG/RANK system in this process could also contribute to its role in plaque destabilization (102). However, studies in OPG^−/−ApoE^−/− mice as well as in ApoE^−/− mice treated with recombinant OPG suggest a role for OPG in plaque stabilization rather than destabilization (52,53). The reasons for these apparent discrepancies are at present not clear, but several potential explanations may exist. Because OPG circulates at much higher levels than RANKL, it may be a more stable overall measure of RANKL/RANK activity than soluble RANKL. Thus, the ability of OPG to predict adverse events in ACS may not be related to its role as a mediator but may rather reflect its role as a stable marker of activity in the RANKL/OPG/RANK axis, mirroring several interaction pathways with relevance to plaque rupture such as inflammation, matrix degradation, and vascular calcification. Moreover, the biological effect of OPG may depend on the molar ratio between RANKL and OPG. Thus, while OPG under low RANKL/OPG ratios attenuates RANKL-mediated effects, it has during high RANKL/OPG ratios been found to enhance the RANKL-mediated effects on MMP levels in vascular SMC (94). However, although some studies suggest the involvement of RANKL/OPG/RANK axis in ACS, more evidence is needed to evaluate the predictive and diagnostic value of serum OPG levels for clinical use as well as the pathogenic importance of these mediators in the process of plaque rupture.

Anti-TNF therapy in atherosclerosis: caution is still needed

TNF-α plays a key role in initiating and modulating inflammatory responses including the regulation of other cytokines, adhesion molecules, TF and other pro-thrombotic mediators, and MMPs. It is expressed by vascular tissue in response to lipoprotein oxidation and by cultured macrophages exposed to oxLDL (103,104). TNF-α is expressed in atherosclerotic plaques, co-localized with foam cells, vascular SMC, and mast cells (104,105). TNF-α has also been found to be associated with TUNEL-positive cells with potential pro-apoptotic effects in SMC and foam cells, favoring plaque rupture and weakening of the fibrous cap (106). In line with this, plasma concentrations of TNF-α are persistently elevated among post-MI patients at increased risk for recurrent coronary events (107). In addition, experimental studies have demonstrated that TNF-α-deficient hypercholesterolemic mice develop less atherosclerosis (43,108). Patients with rheumatoid arthritis are at increased risk for developing ACS and other complications of atherosclerosis, and, interestingly, anti-TNF therapy in these patients has been shown to improve endothelial function, decrease intima media thickness, and to attenuate arterial stiffness (109). There are also a few epidemiological studies suggesting that anti-TNF may delay cardiovascular events in this patient population (110,111). However, randomized trials with clinical end-points are lacking. Moreover, development of ACS in relation to infliximab (anti-TNF-α) treatment has also been reported, potentially reflecting antibody-mediated apoptosis of TNF-α-expressing SMC and foam cells within the lesion, with plaque destabilization as a consequence (112). Further studies are needed to determine the net effects of these biologic agents on the vasculature.

Several TNF-related molecules may be related to ACS

Several other TNF-related molecules than those that are discussed above could also be related to plaque destabilization and the development of ACS. Thus, vascular SMC are susceptible to TRAIL-mediated apoptosis, and, notably, patients with ACS have been shown to bear increased frequencies of CD4⁺ T cells that express TRAIL upon stimulation and kill SMC in a TRAIL-dependent manner (113). These authors also showed that adoptive transfer of such CD4⁺ T cells causes vascular SMC death in human carotid plaque, demonstrating the in vivo relevance of this mechanism in plaque destabilization. FasL-Fas interaction is also an important pro-apoptotic pathway, and circulating levels of sFasL and sFas are increased in ACS, potentially improving diagnostic accuracy and risk stratification (114,115). The expression of TWEAK and its receptor Fn14 that is not classified as a member of the TNFRSF has been detected in macrophage/foam cell-rich regions of atherosclerotic plaques, with overlapping expression patterns of Fn14 and different MMP within the atherosclerotic lesion (116). Together with the ability of TWEAK to promote MMP activity and apoptosis under certain circumstances (e.g. with IFN-α as a co-stimuli), these findings could relate TWEAK to plaque destabilization (117). Moreover, T cells from patients with ACS have enhanced expression of OX40L and OX40 as compared with stable CAD and healthy controls, and serum level of sOX40L has been suggested as a potential marker for predicting the severity of CAD (118). Finally, 4-1BB is expressed in human atherosclerosis and promotes development of plaque inflammation in hypercholesterolemic mice (49), and ACS patients have been found to have increased serum levels and monocyte expression of 4-1BB, significantly correlated with C-reactive protein (119). However, for all these TNF-related molecules, few patients were included in the clinical studies, and prospective data are lacking. Moreover, although animal models have supported a role for these TNF-related molecules in atherosclerosis, their role in plaque destabilization is far from clear, and the existing data must be interpreted with caution.

In addition to the involvement of several TNF-related molecules, several cell types seem to mediate the effects of these molecules in the development of ACS. Although this review have focused on their relation to platelets, monocytes/macrophages, and endothelial cells, it is important to underscore that several TNF-related ligands and receptors are potent co-stimulatory molecules for T cell activation. Thus, as discussed above, the ability of TRAIL and OX40L/OX40 to cause plaque destabilization seems to be mediated through T cell activation (113,118). Moreover, the inflammatory phenotype in ApoE^−/− mice that are exposed to 4-1BB consists of infiltrating T cells and in particular of the CD8⁺ T cell subsets (49). Moreover, ACS has not only increased serum/plasma levels of sCD40L, but also enhanced T cell expression of membrane-bound CD40L (70). Also, regulatory T cells are known to attenuate atherosclerosis (120), and, as GITR engagement on these cells could influence their function (58), GITRL-GITR interaction could potentially also influence atherogenesis through such mechanisms. Moreover, infiltration of dendritic cells within the atherosclerotic lesion has been linked to plaque instability and the development of ACS (121), and, notably, certain TNF related molecules are potent modulators of this cells subset (e.g. interaction of CD40L-CD40, 4-1BBL (CD137L)/TNFSF9-4-1BB, LIGHT-HVEM, and RANKL-RANK) (122–126). Finally, mast cells have recently been linked to plaque destabilization by their ability to induce MMP activity, inflammation, and apoptosis, and it seems that TNF-α is an important actor in these processes (127,128).

Conclusion

Several lines of evidence suggest that TNF-related molecules are involved in the development of ACS. Most, convincing evidence exists for CD40L-CD40 interaction, but several other members of the TNF superfamily seem also to be involved in this immune-mediated promotion of plaque instability, including LIGHT, RANKL, and TNF-α. These plaque destabilization pathways involve the bidirectional interaction between platelets and endothelial cells/monocytes, activation of vascular SMC, and co-stimulatory effects on T cells, promoting inflammation, thrombosis, matrix degradation, and apoptosis. These TNF-related pathways could contribute to the non-resolving inflammation that characterizes atherosclerotic disorders, representing pathogenic loops that are operating during plaque rupture and the development of ACS. These molecules could represent attractive new targets for therapy in this disorder.

Declaration of interest: The authors state no conflict of interest and have received no payment in preparation of this manuscript.