| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print July 8, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Sections of Cardiovascular Medicine and Immunobiology, Yale University School of Medicine, New Haven, Connecticut
1Correspondence: Sections of Cardiovascular Medicine and Immunobiology, Yale University School of Medicine, 300 Cedar Street, TAC-S469, New Haven, CT 06510. E-mail: jeffrey.bender{at}yale.edu
 |
ABSTRACT |
|---|
Key Words: atherosclerosis • inflammation • congestive heart failure • innate immunity • adaptive immunity
|
INTRODUCTION |
|---|
|
AHD |
|---|
|
IMMUNE MECHANISMS IN ATHEROGENESIS |
|---|
B-mediated inflammatory gene transcription, including interleukin (IL)-8 expression, which has T cell chemotactic properties. Cellular oxidative stress is coupled to the modification of lipids by reactive oxygen species (ROS), which are generated in cellular metabolism. ROS can initiate NF-B-mediated transcriptional activation of inflammation genes [9
], thereby, potentially acting as independent triggers.
An important, emerging risk factor for atherosclerosis is insulin resistance, with its associated obesity and dyslipidemia. Insulin resistance has been associated with elevations of tumor necrosis factor 
(TNF-) as part of the observed adipocytokine dysregulation. TNF- can antagonize insulin-stimulated receptor signaling and is elevated in conjunction with low adiponectin levels in insulin resistance [10
, 11
]. Another potential trigger of atherogenesis is angiotensin II (AT-II), which induces EC IL-6 production [12
, 13
]. AT-II also induces oxidative stress, consequently increasing levels of ROS. It is thus apparent that various triggers of atherogenesis augment proinflammatory cytokine levels with consequent activation of immune cells.
Endothelial injury/dysfunction, as one of the initial steps in the initiation of atherosclerosis, can be induced by cytokines, hemodynamic forces, and a variety of vasoactive substances and influenced by traditional atherosclerosis risk factors. ECs play a pivotal role in atherogenesis. They modify LDL [14
], directing macrophage uptake via the scavenger receptors (SRs). ECs express Toll-like receptors (TLRs) [15
], which comprise a family of pattern recognition receptors, also associated with macrophages. EC TLR ligation induces NF-B activation with consequent gene expression of leukocyte adhesion molecules (LAMs), IL-1, and other inflammatory mediators. ECs also express SRs, such as CD36, which can engulf modified lipoproteins [16
]. Thus, through innate immunoreactive receptor engagement, augmentation of EC antigen-presenting function occurs. Although less efficient antigen presenters than macrophages, ECs can present foreign antigens to specific T cells [17
], thereby providing an early link between innate and adaptive immunity.
|
INNATE IMMUNITY IN ATHEROSCLEROSIS |
|---|
, IL-1ß, TNF-, and IL-18 enhance surface expression of intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), E- or P-selectin on EC, SMC, or macrophages [18
, 19
]. mmLDL induces monocyte adhesion to endothelial-connecting segment-1, mediated by activation of 4ß1 integrins [9
, 20
]. The role of adhesion molecules in the form of selectins, ICAM-1, VCAM-1, and integrins has been studied in the initial stages of leukocyte-endothelial interactions. L-, E-, and P-selectin are expressed on leukocytes, ECs, and platelets and ECs, respectively [21
]. They are involved in leukocyte rolling and tethering after initial contact with endothelium [21
]. ß1 and ß2 integrins, leukocyte receptors for VCAM-1 and ICAM-1, respectively, initiate firm arrest on the endothelium. The VCAM-1/integrin 4ß1 and ICAM-1/integrin Lß2 interactions are particularly important in this step [21
]. Integrins have multiple conformational states, which determine their ligand affinity and can be activated by inside-out signaling. Their specificity for ligands, including collagen, laminin, fibrinogen, and fibronectin, is determined by the ,ß heterodimeric pair. One of the earliest vascular alterations induced by an atherogenic diet is increased EC VCAM-1 expression in rabbit and mouse models [22
]. This effect has also been shown with oxLDL [23
]. This is a pivotal step, inducing monocytes and T lymphocytes to bind arterial endothelium [23
]. VCAM-1 engagement also induces signaling, which modulates a change in endothelial shape and facilitates leukocyte emigration [24
]. Hypomorphic variants of VCAM-1 in apolipoprotein E (apoE)-negative mice have reduced atherosclerosis [25
, 26
]. The cytokines IL-1ß, TNF-, and lymphotoxin transcriptionally induce VCAM-1 through a NF-B-dependent pathway [9
]. Soluble forms of EC adhesion molecules, including selectins, ICAM-1, and VCAM-1, can be detected in the systemic circulation. The biological significance of these soluble forms, likely mediated by shedding or proteolysis or both, is unknown, but their levels correlate with cardiovascular disease activity [21
].
The aforementioned adhesion molecules are shown to be focally expressed at sites of atherosclerotic lesion formation [21
]. This has been the subject of intense investigation and has stimulated the concept of atheroprotection, i.e., mechanisms that protect endothelia from injury and consequent atherosclerosis. Laminar shear stress as a result of uniform laminar blood flow is one such mechanism. Arterial branch points, exposed to turbulent rather than laminal flow, are predisposed to lesion development. A number of genes invoked in atheroprotection have shear-stress response elements in their promoter regions [27
]. The shear-stress-responsive endothelial nitric oxide synthase (eNOS) gene, when active, produces NO, which can inhibit NF-B-mediated VCAM-1 gene transcription.
The chemoattractant cytokines, chemokines, recruit leukocytes from the circulation into arterial walls. This occurs through engagement of their G protein-coupled receptors, leading to the aforementioned integrin activation. MCP-1, overexpressed in atheroma, is induced by hypercholesterolemia [5
], which also induces increased expression of the MCP-1 receptor, CC chemokine receptor 2 (CCR2), in monocytes [28
]. Gene-targeting studies in atherosclerotic mouse models [apoE–/– and LDL receptor (LDLR)–/–], using compound mutant mice lacking MCP-1 [29
] or its receptor CCR2 [30
], have diminished mononuclear phagocyte accumulation. IL-8 can play a similar role as a leukocyte chemoattractant in LDLR–/– mice [31
]. IL-8, a granulocyte chemoattractant, also induces firm adhesion of monocytes to endothelium under physiological flow conditions [32
]. IL-8 and IL-15 can facilitate atherogenesis progression by potentiating plaque angiogenesis. These have been studied in human coronary atherectomy tissue and in nude mice, respectively [33
, 34
]. The CXC chemokines induced by interferon-
(IFN-), IFN-inducible protein 10 (IP-10), monokine induced by IFN- (Mig), and IFN-inducible T cell chemoattractant (iTAC) are also expressed in atherosclerotic lesions [35
] and aid in lymphocyte recruitment, increasing the vascular inflammatory response.
The next step in the inflammatory-atherogenic cascade is the transformation of monocytes to resident macrophages and foam cells. Macrophage-colony stimulating factor (M-CSF) is one activator that stimulates this transformation. M-CSF accentuates SR-A expression and increases monocyte cytokine and growth factor production. It is overexpressed by vascular cells in human and rabbit atherosclerotic plaques [36 ]. Monocyte SR-A and CD36 are involved in scavenging modified lipoprotein, producing cytoplasmic vacuoles of a lipid that are the hallmark of foam cells.
M-CSF–/– and apoE–/–/LDLR–/– mice have reduced atherosclerosis compared with the M-CSF+/+ controls [37
]. Plaque lipid-laden foam cells become a major source of cytokines (IL-1ß, IL-6, IL-12, IL-18, TNF-, TNF-ß, IFN-), chemokines (MCP-1, IL-8), growth factors [platelet-derived growth factor (PDGF) and transforming growth factor-ß (TGF-ß)], CSFs [M-CSF and granulocyte macrophage (GM)-CSF], and proteolytic enzymes [matrix metalloproteinase (MMP) and cathepsins; refs. 38
, 39
]. These secreted factors amplify the atherosclerotic inflammatory cascade, including the conversion to unstable lesions, as will be described below.
|
ADAPTIVE IMMUNITY IN ATHEROSCLEROSIS |
|---|
Early atherosclerosis has been studied in animal models with an obvious paucity of human data, given the practical difficulty in obtaining early atherosclerotic lesions in humans. Most of the T cells in atherosclerotic lesions are CD4+ and bear the conventional T cell antigen receptor ß (TCRß). Such T cells comprise two-thirds of all the CD3+ T cells in advanced human atheromas [41
]. Most of these T cells are of the T helper cell type 1 (Th1) subtype and secrete the immune/proinflammatory cytokines IFN-, IL-2, and TNF- and -ß [44
]. These cytokines in turn induce innate immune responses by stimulating macrophages and vascular cells. Furthermore, Th1-stimulatory cytokines IL-18 and Il-12 are produced and expressed by many cells in the atheroma and mediate various proatherogenic processes in vitro and in vivo [45
46
47
]. IL-18 enhances Th1 polarization, directly, by increasing IFN- secretion and by diminishing levels of the Th2 cytokine IL-10 [44
]. The IFN-, secreted by T cells, primes macrophages, lowering their threshold for TLR-dependent activation. T cells also produce TNF-, with potent NF-B-activating properties. IL-1, TNF-, and IFN- can enhance expression of CD40 ligand (CD40L) and CD40, cell-associated members of the TNF/TNF receptor (TNFR) family [48
]. Activated T cells express CD40L (CD154), which binds to its CD40 receptor on macrophages, B cells, ECs, SMCs, and dendritic cells (DCs) [49
]. This, in turn, induces expression of tissue factor (TF), MMP, and adhesion molecules [48
49
50
51
]. IFN--stimulated IP-10, Mig, and iTac are overexpressed by atheroma cells, attracting and binding CXC chemokine receptor 3-bearing T and B cells [35
]. IFN- influence on atherogenesis continues to be controversial; IFN- prevents formation of foam cells in vitro. Regions of atheromas, abundant in T cells that secrete IFN-, show reduction in SR expression with a concomitant decrease in the number of foam cells [52
, 53
]. However, IFN- accentuates SR expression on cultured rabbit SMCs [54
]. IFN- has recently been shown to promote vascular remodeling, one of the typical features of atherosclerosis [55
]. Atheromas contain only small amounts of Th2 cytokines such as IL-4, IL-5, and IL-10 [44
]. This Th1/Th2 balance is maintained by cross-regulation. It is no surprise that IL-10-deficient mice have increased atherosclerosis or that LDR–/– mice with systemic or endothelial overexpression of IL-10 have minimal atherosclerotic lesions [56
]. IL-4 and IL-10 reduce ICAM-1 and VCAM-1 expression [57
, 58
], also contributing to their antiatherogenic effect.
It is interesting that human atheromas also contain DCs, highly efficient antigen presenters to naïve T cells [59 ]. Antigens from atheromas are presented by DCs in local lymph nodes to activate naïve and memory T cells [60 ]. Atheromas also contain mast cells, which accumulate at sites of plaque rupture and can play an important role in acute coronary syndromes, as a result of production of a large number of proteases that have roles in degradation of plaque fibrin and matrix [61 ]. In addition to CD4+ T cells, atheromas also contain CD8+ T cells and some B cells [41 , 62 ]. Their respective functions of cytotoxicity and antibody production likely represent additional immune response manifestations in atherosclerotic lesions, although pathologic details are unknown.
Early apoE-deficient mice atherosclerotic lesions demonstrate clonal T cell expansion [63 ]. A similar process of clonal expansion might occur in early human lesions, but advanced human atheromas contain a heterogeneous population of TCRs [64 ]. The multifactorial nature of atherosclerotic triggers could explain this multiclonality. Analysis of plaque T cells has revealed that oxLDL is a major autoantigen in atherogenesis [65 ]. Furthermore, the presence of anti-oxLDL antibodies in atherosclerotic patients and atherosclerotic animal models points to the important immunostimulatory role of lipoproteins involving cell-mediated and humoral immunity in atherosclerosis [66 ]. This involvement has also been studied in apoE-deficient mice, in which the presence of oxLDL-specific T and B cells in lymphoid tissue has been demonstrated [43 , 45 , 67 ]. In regards to a major histocompatibility complex (MHC) link to atherosclerosis, mouse models (fat-fed C57BL/6 mice) reveal a permissive role for MHC class II I-Ab-restricted Th1 cells and a constraining effect of I-Ek and I-Ak [68 ]. Another set of autoantigens, which are also thought to modulate the immune response in atherosclerosis and are known to be targets of autoimmunity in other chronic inflammatory diseases, is the heat-shock protein (HSP). HSP60 can cause CD14-dependent activation of TLR4 [69 ], similar to bacterial lipopolysaccharide, thereby stimulating the innate immune responses. HSP60 has also been shown to induce a humoral response (anti-HSP60) in atherosclerotic animals, and immunization with HSP60 has been shown to accentuate atherosclerosis in experimental animals [70 ]. Molecular mimicry between microbial antigens and autoantigens has also been shown and may be another mechanism for initiation or potentiation of atherogenesis. The HSP60 of Chlamydia pneumoniae resembles human HSP60, potentially activating vascular cells and innate immune responders [71 ].
|
CYTOKINES IN THE NATURAL HISTORY OF AN ATHEROMA |
|---|
IL-1 and TNF- enhance production of M-CSF, GM-CSF, and G-CSF by SMCs, ECs, and monocytes. These mediators activate monocytic cells, enhancing their differentiation to macrophages and foam cells [36
, 72
, 73
]. Growth factors such as PDGF are strong stimulators of SMC proliferation and migration, both vital processes in the formation of the fibrous component of atheromas. In vitro and in vivo (rat) studies demonstrate the importance of PDGF in SMC migration and proliferation [74
]. IL-1, TNF-, and CD40L stimulate production of PDGF in ECs and SMCs. Conversely, IFN- inhibits SMC proliferation and collagen production, thereby promoting plaque instability [75
].
As mentioned, cell death is an important factor in plaque dynamics. This is mediated primarily by apoptosis, although necrosis contributes as well [76
]. The role of cytokines in cell death is dependent on which cytokine or Th subfamily is dominant. TNF- and IFN- can accentuate the development of ROS, leading to cell death. EC and SMC can also be sensitized to apoptosis mediated by Fas ligand-positive lesional T cells [77
, 78
]. IL-10 can promote or inhibit apoptosis, depending on the target cell type and interaction with other growth factors [51
, 79
].
Clinical sequelae of atherosclerotic plaques are related to their stability. This is determined by their fibrous content, which in turn, is dependent on the production of extracellular matrix (ECM). The balance between SMC-ECM synthesis and ECM degradation determines matrix turnover, by the proteolytic enzymes, MMPs. ECM synthesis is most dependent on SMCs. Type 1 collagen is the single-most important component of the fibrous cap, influencing the strength of the cap. IL-1ß and IFN- have opposing effects. The former potentiates collagen synthesis by increasing the expression of PDGF and TGF-ß, both of which stimulate collagen production [80
]. The latter inhibits basal, IL-1ß, or TGF-ß-mediated collagen synthesis [80
]. In the atherosclerotic mouse (apoE-deficient) model, IFN- receptor deficiency exaggerates lesional collagen content. Furthermore, in human atheromas, T cell-rich areas are associated with decreased type 1 collagen mRNA localization, suggesting the role of T cell-associated IFN- effect on collagen production [81
]. Thus, the effect of cytokines on collagen homeostasis highlights their dual role, exacerbatory or inhibitory, on atherogenesis and plaque instability.
MMPs largely mediate ECM degradation. The MMPs implicated in atheroma progression are those that have been shown to degrade fibrillar collagen (MMP-1, -8, and -13), nonfibrillar collagen (MMP-2, -3, -7, -9, and -10), fibronectin (MMP-2, -3, -7, and -10), and collagen fragments (MMP-1, -2, -3, -7, and -9). MMP-1, -8, and -13 have been associated with lesional collagenolysis, demonstrated by the presence of MMP-specific collagen-cleavage products [82
83
84
85
]. Cytokines can regulate MMP activity at three levels: the transcriptional stage, maturation of inactive zymogen, and by interactions with MMP inhibitors. Proinflammatory cytokines IL-1ß, TNF-, and CD40L enhance expression of fibrillar and nonfibrillar MMPs in atheroma cells [86
, 87
], whereas Th2 cytokines IL-4 and IL-10 reduce MMP-1, -3, and -9 expression [88
]. IL-1 and TNF- can also impact MMP activity by accentuating levels of plasminogen and thrombin, both of which lead to proteolytic activation of MMP-inactive zymogen, or by attenuating levels of thrombomodulin, which has the opposite effect on the inactive zymogens [89
, 90
]. Step-wise activation of inactive MMP zymogens by other leukocyte-derived proteolytic enzymes has been described [91
]. ROS have also been shown to similarly modulate MMP activation. Finally, cytokines can regulate MMP activity by modulating expression of tissue inhibitors of MMPs (TIMPs) or serpin-like TF pathway inhibitor 2. The MMP/TIMP ratio can determine the proteolytic activity. For example, CD40 ligation results in loss of de novo-synthesized, interstitial collagen in human vascular SMC (VSMCs) in vitro as a result of MMP activity [92
]. Conversely, PDGF and TGF- can increase levels of TIMP-1 and -3 in vitro [93
], thereby providing an indirect mechanism by which cytokines modulate MMP activity.
The rupture of the fibrous cap is not adequate to initiate a significant clinical event. There is a requirement for plaque thrombosis, modulated by procoagulants (in particular, TF) and anticoagulants (such as thrombomodulin). TF levels are higher in subjects with unstable angina versus those with stable angina [94
], demonstrating its importance in acute clinical syndromes. Although many atheroma-associated cell types produce TF, the major source is the foam cell. IL-1, TNF-, and CD40L, as well as cytokine-regulated growth factors such as PDGF, augment the synthesis of TF in atheroma-associated cells [95
, 96
]. CD40L-CD40 interactions in ECs and macrophages play an important role in determining thrombogenicity of atheromas, by impacting the TF levels [97
]. This effect has also been studied in experimental animals [98
]. The Th2 cytokines IL-4, -10, and -13 inhibit IL-1ß and TNF--induced TF expression in atheroma cells in vitro [99
, 100
]. Figure 1
provides a schematic representation of the prominent cytokine effects on innate and adaptive immune cells and the evolution of the atheromatous plaque.
|
|
ATHEROSCLEROSIS AND NOS |
|---|
The role of eNOS is the most studied in atherosclerosis and is complex [108 ]. It is known that eNOS-derived NO has various antiatherogenic effects: Decreased MCP-1, VCAM-1 [109 ], and ICAM-1 [110 ] expression inhibits leukocyte-endothelial interactions in early atherosclerosis. NO also can reduce LDL oxidation [111 ] and impair lipoprotein influx into vessel walls [112 ]. However, eNOS catalytic activity can be "uncoupled," which results in eNOS-mediated superoxide production. In this state, electrons from the NOS reductase domain, normally transferred to the heme domain, resulting in citrulline and NO formation, are instead diverted directly to molecular oxygen, resulting in superoxide formation. Also, cellular BH4 levels correlate with eNOS-mediated NO production. apoE knockout (KO) mice have impaired endothelium-dependent relaxation. This is associated with increased vascular superoxide production, improved by vessel incubation ex vivo with sepiaptern (a BH4 precursor) with concomitant reduction in superoxide levels [113 ]. In fact, BH4 levels are basally low in murine models of atherosclerosis [114 ]. In addition, the BH4:oxidized biopterin (BH2) ratio determines the balance of NOS-derived NO versus superoxide [115 ]. This dichotomy of eNOS effects on atherosclerosis has also been demonstrated in transgenic animal models. apoE-eNOS double-KO mice have an increased burden of aortic atherosclerosis and aneurysm formation compared with single apoE KO [116 ]. Conversely, eNOS KO mice can have a reduction in diet-induced atherosclerosis compared with wild-type controls [117 ]. Conflicting results have also been observed in apoE KO mice overexpressing eNOS transgenically [118 , 119 ]. These discrepancies may be related, in part, to different atherogenic diets, inducing distinct degrees of oxidative stress. At this time, the spectrum of clinical and pathologic settings, which disturb the balance of eNOS-derived NO versus ROS, is not understood completely. What is clear is that the role of the various NOS isoforms in vascular disease is complex, reflecting the dual nature of their products, which can have protective and toxic effects on the vasculature.
|
CHF |
|---|
, IL-1ß, and IL-2, as these are the most well-studied and described in the context of heart failure. These mediators are typically considered to have negative inotropic effects. However, the pattern of this response is kinetics-dependent. Immediate response occurs within minutes and can be stimulatory or depressant, depending on experimental conditions. The delayed response, lasting hours to days, is always cardiodepressant and is modulated by secondary mediators. |
CARDIAC CYTOKINES: AN ADAPTIVE RESPONSE TO STRESS? |
|---|
Evidence favoring the adaptive role of cytokines in heart failure comes from gain-of-function studies. Pretreating rats with TNF- protects their hearts from reperfusion injury ex vivo [122
]. Similar benefits have also been seen with IL-1 [123
]. TNFR1 and TNFR2 double-KO mice have an increase in ischemic infarct size, which is associated with exaggerated apoptosis [124
].
Studies with the IL-6 family [IL-6, ciliary neurotrophic factor, cardiotrophin-1 (CT-1), leukemia inhibitory factor (LIF), and IL-11] of cytokines have shown similar results. These cytokines trigger signaling through the homodimerization or heterodimerization of the gp130 receptor. In neonatal cardiac myocytes, CT-1 blunts serum deprivation-induced apoptosis through a mitogen-activated protein kinase (MAPK) cascade. LIF has also been shown to have similar protective effects [125 , 126 ] using the MAPK pathway as well as the Janus kinase and the signal transducer and activator of transcription-mediated signaling pathways [126 ]. Homozygous gp130–/– KO mice die in utero secondary to cardiac dysgenesis. Mice with a ventricular-restricted KO of gp130 have no cardiac abnormalities [127 ]. The lethal effect of the systemic gp130–/– is thought to be secondary to extracardiac abnormalities, resulting in fetal nonviability. Aortic banding experiments (hemodynamic overloading caused by transaortic constriction) in the ventricular-restricted KO demonstrate biventricular enlargement with a marked increase in myocyte apoptosis and resultant death compared with control animals [127 ].
Although the exact mechanisms of these cytoprotective effects have not been defined, cytokines are known to increase the levels of various protective proteins in the heart, including the free radical scavenger manganese superoxide dismutase [122 , 128 ] and Hsps [129 ]. The former likely plays a role in myoprotection against oxidative stress, and the latter is shown to be protective against ischemic injury [130 ].
|
CYTOKINES: TOO MUCH FOR TOO LONG? |
|---|
Animal studies have demonstrated cytokine effects on cardiac function. Intravenous administration of TNF- or IL-1ß mimics the hemodynamic profile of endotoxic shock, which is blunted by coadministration of anti-TNF- antibodies [134
] or IL-1ß receptor antagonist [135
]. Canine studies have shown that TNF- induces a positive, early inotropic response [136
, 137
], followed by an overwhelmingly delayed, depressive response [136
, 138
139
140
], as measured by a left ventricular (LV) ejection fraction [138
, 140
] and with LV pressure-volume indices [136
, 139
, 141
]. Similar results have been obtained with intracoronary administration of IL-1ß-coated microspheres. In these studies, the cytokine-mediated cardiodepressive effects were reversible over days following cessation of exposure [23
, 136
, 138
139
140
].
These in vivo studies confirm that cytokines have a biphasic cardiac effect. The early (as early as 5–15 min), positive, inotropic effect of short duration suggests a direct myocardial effect that does not involve gene expression. The delayed and prolonged effect suggests the requirement of gene expression, protein synthesis and likely secondary mediators. The reversibility of the delayed response may depend on dose and duration of exposure. In vitro and ex vivo studies have consistently shown the presence of the delayed cardiodepressor effect of cytokines on basal [142 143 144 145 ] or stimulated myocardial function [146 147 148 149 ]. Although the early cardiostimulatory response has been variable, there are in fact numerous reports of early inhibitory effects [150 151 152 153 154 155 156 157 158 ].
|
MECHANISMS OF CYTOKINE EFFECTS ON LV FUNCTION |
|---|
and IL-1 activate acid and neutral sphingomyelinases, which hydrolyze membrane-incorporated sphingomyelin to form ceramide [153
, 161
162
163
], which is then deacylated by ceramidase to give sphingosine (see Fig. 2
). In feline myocytes, TNF- activates neutral and acid sphingomyelinases within 15 min, thereby causing increased production of ceramide and free sphingosine [153
]. This observation correlates with TNF--induced, TNFR1-mediated contractile depression [164
]. These effects are blunted by ceramidase inhibition and subsequently superseded by sphingosine supplementation. Responses to sphingosine may explain the immediate, depressive effects of proinflammatory cytokines and have been observed in experimental animals [157
, 165
, 166
].
|
and IL-1ß activate PLA2 within minutes, causing increased ArA levels [145
, 163
, 168
]. ArA in turn activates neutral sphingomyelinase [162
, 167
]. ArA also has variable direct effects—stimulatory at low concentrations and short exposure times and inhibitory at higher concentration and longer durations [169
, 170
] (see Fig. 2
). It has been suggested that ArA mediates the positive and negative inotropic responses of TNF- [160
]. PLA2 inhibition blunts both, although ceramidase inhibition prevents only the negative inotropic effects, leaving a sustained, positive response. This suggests that cytokine-induced PLA2 activation is upstream of sphinogmyelinase induction and that early ArA production likely induces a direct, myocardial-stimulatory effect, followed by ArA-associated sphingomyelinase activation and release of cardioinhibitory ceramide and sphingosine. This remains an area of active investigation.
Cytokines might modulate cardiac contractility through cNOS activation and NO production. This NOS isoform, also called eNOS or NOS3, is constitutively expressed in cardiac myocytes. As opposed to regulation of iNOS, the cytokine-inducible isoform (NOS2), which is transcriptionally regulated over a minimum of hours, cNOS activation can be induced within minutes, primarily through cytosolic calcium increases or changes in the enzyme phosphorylation state. Moderately high levels of TNF-, IL-2, or IL-6 cause severe, rapid (within 5 min), reversible inhibition of cardiac contractility, inhibitable by the NOS inhibitor N-monomethyl-L-arginine [171
]. The rapidity and reversibility of the response suggest a role for cNOS (Fig. 2)
. It appears that cNOS-derived NO can mediate the myocyte-inhibitory effects of IL-2 [172
], IL-6 [157
], TNF-, and IL-1ß [173
, 174
] in vitro. As noted above, cNOS is a Ca2+-dependent enzyme, and the mechanism by which these cytokines activate cNOS is unclear. Also, although these studies argue in favor of an important role of cNOS-NO in acute cytokine-mediated myocardial depression, there are other studies that argue against this [157
, 175
]. These studies have shown the presence of contractile inhibition ex vivo, despite the presence of NOS inhibitors. Given the contradictory results, there may be cytokine specificity related to the intensity of cNOS activation, thereby determining its acute inhibitory effect [176
].
Modulation of cardiomyocyte excitation-contraction coupling is reflected in contractility changes. TNF- decreases peak calcium-transient amplitude and the rate of calcium decline, suggesting a decrease in sarcoplasmic reticulum Ca2+ adenosinetriphosphatase (ATPase) activity without any change in resting calcium levels or the inward calcium current [157
, 175
]. Similar effects are not consistently seen with high doses of TNF-. These depressant effects of TNF- are mediated selectively by the TNFR1 [164
, 177
]. Conflicting studies reporting an early, beneficial cardiac effect of TNF- have shown accentuation of the calcium transient and calcium decline, suggesting an increased sarcoplasmic reticulum Ca2+ ATPase activity [145
, 178
]. There appears to be cytokine-specific effects on excitation-contraction coupling, and TNF-, IL-2 [156
], and IL-6 [157
] modulate sarcoplasmic reticulum Ca2+ ATPase activity, and IL-1ß [158
, 179
] affects the inward calcium current.
Delayed effects
As noted above, proinflammatory cytokines have negative, delayed, inotropic effects on cardiac myocytes, including inhibited basal contractile function and impaired ß adrenergic receptor sensitivity. Although the mechanisms are different, both appear to be iNOS/NO-mediated [143
, 144
, 146
, 148
, 149
, 151
, 180
181
182
].
The NO-mediated mechanism implicated in basal cardiac dysfunction is dependent on S-nitrosylation of regulatory protein thiol residues [183
184
185
186
] (Fig. 3
). S-nitrosylation is a redox-coupled process [187
]. ROS influence thiol modifications, directing the balance between nitrosative and oxidative stress [188
]. In an ex vivo rat heart model, the combination of IL-1ß, IFN-, and TNF- induces xanthine oxidoreductase and NADPH oxidase and iNOS, resulting in superoxide, NO, peroxynitrate, nitrotyrosine, and dityrosine associated with deterioration in contractility in 20 min [143
]. This myocardial depression is inhibitable by administration of a NOS inhibitor, superoxide scavenger, or peroxynitrate decomposition catalyst. Furthermore, NOS inhibitors, given concurrently with cytokines, can improve delayed cytokine-induced myoinhibition [189
] but are unable to reverse loss of contractile function following sustained cytokine exposure, which has resulted in established myocardial dysfunction [190
]. This suggests that oxidative modifications, associated with peroxynitrate and involving covalent protein modifications, may be irreversible. As noted above, NO-mediated cytokine effects can induce varying degrees of myocardial depression based on the duration and extent of cytokine exposure. More studies are required to clarify the irreversibility of these effects.
|
reduces the expected ß-adrenergic-stimulated cAMP accumulation, blunting the expected increase in contractility [191
, 192
]. This effect is refractory to PDE inhibition and remains reversible by 72 h following cytokine exposure. The role of NO-cGMP-dependent signaling in this uncoupling phenomenon is not clearly established. In adult rat ventricular myocytes, exposure to cytokine-enriched macrophage supernatants or to specific cytokine cocktails for 24 h results in loss of typical ß-adrenergic responses [146
, 148
, 149
, 181
, 193
]. This loss is associated with iNOS induction, increased cGMP, nitrite content [146
, 194
], and measurable NO release [193
], inhibitable by NOS antagonists. Myocardial iNOS activity is up-regulated within 2 h of IL-1ß and TNF- [195
, 196
] or IL-6 [197
] exposure. TGF-ß can prevent this cytokine-induced iNOS induction [148
, 193
, 198
, 199
]. In transgenic mice with an overexpression of cardiac-specific TNF-, short-term administration of a selective iNOS inhibitor improves ß-adrenergic responsiveness but does not abrogate mortality associated with the cytokine-induced cardiomyopathy [109
]. Crosses with iNOS-deficient mice produce progeny with improvement in ß-adrenergic responsiveness but without reversal of baseline cardiac dysfunction [182
]. This again suggests distinct mechanisms involved in basal contractility depressions and ß-adrenergic stimulation, the former apparently a more sustained abnormality. The effect of NO on ß-adrenergic activity is thought to be mediated by cGMP-dependent mechanisms (Fig. 3)
. This involves PKG-mediated PDE II stimulation with a resultant augmentation in cAMP degradation and decrease in cAMP/PKA-dependent L-type Ca2+ currents [146
, 183
184
185
186
]. In summary, it is clear that cytokines have a spectrum of effects on the myocardium. These include an early response that is cardiostimulatory or inhibitory, dependent on the cytokine profile. The early response appears to be mediated by the cytokine effect on sphingomyelinases and cNOS. The delayed response is consistently inhibitory to basal and adrenergic-mediated contractility and appears to be mediated by closely inter-related pathways, leading to NO and superoxide production [176 ]. Given the complexity and influences of hemodynamic and neurohumoral responses, in vivo effects, observed on isolated cardiomyocytes or even ex vivo heart preparations, should be extrapolated cautiously to intact, biologic systems.
Cytokines and LV remodeling
Cytokines induce LV remodeling through several mechanisms. They change the ECM milieu by effects on MMPs. Cardiac-specific TNF transgenic mice have a greater ratio of MMP:TIMP levels at 4 weeks than wild-type controls, favoring collagen degradation. This is followed by progressively lower ratios from 8 to 12 weeks, favoring collagen accumulation [200
]. Myocyte hypertrophy [201
] and alterations in fetal gene expression [202
, 203
] have also been shown to contribute to ventricular remodeling. This dichotomous effect suggests that there is a temporal relationship to matrix degradation. Why such a transition occurs is not understood. The exact mechanisms directing this transition are not defined but may be dependent on preferential increases of certain fibrogenic cytokines, such as TGF-ß, upon sustained activation.
|
CONCLUSION |
|---|
Another example is the use of specific, targeted anticytokine therapy in heart failure. It is important to emphasize that the relationship between cytokines and the cardiovascular diseases in which they are implicated is complex. Despite initial promising, small, short-term studies, several anti-TNF- therapies for CHF have failed [204
]. There are many possibilities for these failures, some related to the biochemical nature of the agent used. Infliximab is a complement-fixing, anti-TNF- monoclonal antibody. Through its complement-fixing activity, infliximab could be cytotoxic to cardiomyocytes expressing membrane-bound TNF-, resulting in worsening cardiac function. Etanercept, a TNF- soluble receptor that binds and antagonizes TNF-, can also have a role in stabilizing TNF- in the peripheral circulation. This TNF- "store" is not strongly bound to etanercept, eventually dissociating and leading to a paradoxical, rapid elevation of TNF- levels. This could result in an accentuation of TNF- action and worsening heart failure. It is possible to speculate, based on the above observations, that antagonizing cytokines in heart failure is a misplaced strategy. However, it is more likely that we have not been able to develop agents that can target cytokines in heart failure safely, given the complex relationship between cytokine levels and cardiac adaptation versus maladaptation.
We are only beginning to understand and immunologically manipulate disease states that were not previously considered to be diseases of the immune system. The ability to favorably impact human cardiovascular disease by modulating immune responses is undoubtedly immense but requires further understanding of immunopathogenesis.
|
ACKNOWLEDGEMENTS |
|---|
Received April 8, 2005; revised June 3, 2005; accepted June 6, 2005.
|
REFERENCES |
|---|
TOP
ABSTRACT
INTRODUCTION
