|Home | About | Journals | Submit | Contact Us | Français|
Atherosclerosis, a chronic inflammatory disorder, involves both the innate and adaptive arms of the immune response that mediate the initiation, progression, and ultimate thrombotic complications of atherosclerosis. Most fatal thromboses, which may manifest as acute myocardial infarction or ischemic stroke, result from frank rupture or superficial erosion of the fibrous cap overlying the atheroma, processes that occur in inflammatorily active, rupture-prone plaques.
Appreciation of the inflammatory character of atherosclerosis has led to the application of C-reactive protein as a biomarker of cardiovascular risk, and the characterization of the anti-inflammatory and immunomodulatory actions of the statin class of drugs. An improved understanding of the pathobiology of atherosclerosis and further studies of its immune mechanisms provide avenues for the development of future strategies directed toward better risk stratification of patients as well as the identification of novel anti-inflammatory therapies. This review retraces leukocyte subsets involved in innate and adaptive immunity and their contributions to atherogenesis.
Atherosclerosis, a chronic inflammatory disease, involves both innate and adaptive arms of immunity which modulate lesion initiation, progression, and potentially devastating thrombotic complications 1 2. Thrombosis often complicates physical disruption of the protective collagen-rich fibrous cap overlying the atheroma, exposing circulating clotting factors to procoagulants expressed within lesions as a result of inflammatory activation and initiation of the coagulation cascade 3. Importantly, inflammation also decisively influences the propensity of a given plaque disruption to lead to a sustained and occlusive thrombus that may manifest clinically as an acute coronary syndrome or ischemic stroke by controlling the balance between fibrinolytic enzymes and their endogenous inhibitors 4 5.
Appreciation of the inflammatory character of atherosclerosis has spawned new avenues in basic, translational, and clinical research. CRP (C-reactive protein), an acute-phase reactant released during inflammatory processes, adds to the predictive power of traditional markers of cardiovascular risk 6. Basic research suggests that treatment with statins (3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors) — initially developed to decrease low density lipoprotein (LDL) cholesterol levels — also reduces leukocyte adhesion, accumulation of macrophages, protease production, procoagulant and pro-inflammatory mediator expression, antigen presentation, and T-cell activation 7. Additional support for the anti-inflammatory and immunomodulatory actions of statins came from clinical research. The magnitude of risk reduction associated with statin therapy may exceed that expected on the basis of LDL-cholesterol lowering alone. The CARE (Cholesterol And Recurrent Events) trial first demonstrated that statin therapy lowers plasma levels of CRP in addition to LDL-cholesterol 8. Retrospective evidence supported the utility of targeting the inflammatory marker CRP with statins in normocholesterolemic patients in both primary 9 and secondary prevention 10 of adverse cardiovascular events. Prospective evidence provided by the JUPITER trial (Justification for the Use of Statins in Primary Prevention: an Intervention Trial Evaluating Rosuvastatin) demonstrated that patients with LDL-cholesterol levels considered near optimal but elevated CRP levels benefit significantly from statin treatment in the prevention of adverse cardiovascular events. This direct result of the clinical application of the science of inflammation in atherosclerosis has potentially far reaching implications for everyday medical practice and public health.
To portray the chronic inflammation in atherothrombosis, we review here the leukocytes involved in innate and adaptive immunity with established or emerging roles in the disease process, and detail their cellular and molecular contributions. Beyond macrophages and CD4+ T-cells, new research highlights the important role of regulatory T-cells, dendritic cells, and mast cells in the disease process. Most of the cited experimental work relies on genetically altered atherosclerosis-prone mice, namely apolipoprotein E (ApoE)-deficient mice, which develop hypercholesterolemia and atherosclerotic disease spontaneously 11, and low-density lipoprotein receptor-deficient mice, which require a high-fat diet to develop hypercholesterolemia and atherogenesis 12.
Accumulation of lipid-laden macrophage-derived foam cells characterizes fatty streaks, the initial asymptomatic lesion of atherosclerosis 1. The precursor lesion of atherosclerotic plaques, fatty streaks, have focal increases in the content of lipoproteins within regions of the intima where they associate with constituents of the extracellular matrix such as proteoglycans, slowing their egress 13. This retention sequesters lipoproteins within the intima, protecting them from plasma antioxidants, thus favoring their oxidative modification. Laboratory and clinical data suggest that oxidized or glycated LDL evokes an inflammatory response in the artery wall, unleashing many of the biological processes thought to participate in atherosclerosis initiation, progression, and complication 14.
Endothelial cells (ECs) normally resist leukocyte adhesion. Pro-inflammatory stimuli that include hypercholesterolemia, hyperglycemia, hypertension, and smoking trigger the endothelial expression of adhesion molecules such as vascular cell adhesion molecule-1 (VCAM-1) and P-selectin that mediate the attachment of circulating monocytes and other leukocytes 15 16 17 18. Chemoattractant factors, including monocyte chemoattractant protein-1 (MCP-1) produced by vascular wall cells in response to modified lipoproteins, direct the migration and diapedesis of adherent monocytes. MCP-1 binds to CCR2 on the surface of migrating monocytes to exert this effect. Genetic absence of MCP-1 in LDLR-deficient mice dramatically decreases atherosclerotic disease with marked inhibition of monocyte recruitment 19. Similar results hold true upon deletion of CCR2 expression in ApoE-deficient mice 20. During diapedesis, monocytes release the matrix metalloproteinase MMP-9 that can degrade type IV collagen in the intimal basement membrane to help them enter the growing intimal atherosclerotic lesion 21. Experimental evidence and human observations support the involvement of several other chemokines in monocyte recruitment into the nascent atherosclerotic lesion, including IL-8, which binds CXCR2 22 23, and fractalkine (CX3CL1), which binds CX3CR1 24 25. Importantly, low shear stress also induces expression of MCP-1, fractalkine, and other chemokines involved in atherogenesis 26. Monocytes infiltrate the lesion and localize particularly in the shoulder region where the atheroma grows. In addition, neovessels arising from the artery's vasa vasorum provide another entry route for monocytes into established atherosclerotic lesions 27.
Within the intima, monocytes mature into macrophages under the influence of macrophage colony-stimulating factor (M-CSF), overexpressed in the inflamed intima 28 29. Importantly, M-CSF stimulation also leads to increased macrophage expression of scavenger receptors, pattern-recognition receptors involved in innate immunity, which engulf modified lipoproteins and apoptotic bodies through receptor-mediated endocytosis, leading to their lysosomal degradation. Scavenger receptors involved in macrophage foam cell formation include CD36, CD68, CXCL16, LOX1 (lectin-type oxidized low-density lipoprotein receptor 1), SR (scavenger receptor)-A, and SR-B 1. Accumulation of cholesteryl esters in the cytoplasm produces the characteristic change of macrophages into foam cells. Another type of pattern-recognition receptor, Toll-like receptors (TLRs), directly elicit inflammatory responses 1. In particular, monocytes in human atherosclerotic plaques have markedly enhanced expression of TLR1, TLR2, and TLR4 30. A majority of these monocytes show nuclear translocation of the transcription factor NF-κB (nuclear factor-κB), consistent with their inflammatory activation in lesions 30. Within atherosclerotic lesions, IκB kinase 2 (IKK-2, or IKK-β) phosphorylates IκBα, leading to its ubiquitination and degradation. IKK-2 thereby terminates the inhibitory action of IκBα on NF-κB, and allows the transcription of pro-inflammatory cytokines and proteases 31. A large number of pathogen-associated molecular patterns can activate TLRs. Heat shock proteins (hsp60) 32 and oxLDL 33 34 mediate at least part of their effects within plaques through TLR4 binding. In support of the notion that TLR4 and downstream effectors such as MyD88 (myeloid differentiation primary-response gene 88) mediate inflammatory activation in atherosclerosis, their genetic abrogation reduces disease burden 35 36. Apolipoprotein CIII, a constituent of certain atherogenic triglyceride-rich lipoproteins, can activate cells involved in atherogenesis through TLR2 37. TLR2, expressed on cells not derived from bone marrow, also appears to promote atherogenesis in mice 38 39.
Macrophages proliferate and amplify the inflammatory response through the secretion of numerous growth factors and cytokines, including tumor necrosis factor-α (TNF-α) and interleukin-1β (IL-1β). These 2 key cytokines are central mediators of inflammatory pathways relevant to atherosclerosis (Figure 1). Among a myriad of actions, they induce expression of adhesion molecules such as VCAM-1, chemokines such as MCP-1, growth factors such as M-CSF, and proteases such as MMPs by cellular effectors present within lesions 1. In human lesions, IL-18 colocalizes with mononuclear phagocytes while ECs, SMCs, and macrophages all express the IL-18 receptor 40. Importantly, IL-18 signaling evokes essential effectors involved in atherogenesis, e.g., adhesion molecules (VCAM-1), chemokines (IL-8), cytokines (IL-6), and matrix metalloproteinases (MMP-1/-9/-13) 40. Recent evidence supports selective mobilization during hypercholesterolemia in mice of a pro-inflammatory subset of monocytes that bear high levels of a surface marker denoted Ly6C. These Ly-6Chi monocytes increase dramatically in the blood of fat-fed apoE-deficient mice, and preferentially adhere to activated endothelium, infiltrate lesions, and become lesional macrophages 41. Importantly, pro-inflammatory Ly-6Chi monocytes are CCR2+ and CX3CR1+ and rely on these chemokine receptors to enter lesions 42.
Advanced atheromata may exhibit a paucity of SMCs and abundant macrophages, key histological characteristics of plaques that have ruptured and caused fatal thrombosis. The fibrous cap covering the atherosclerotic plaque owes its biomechanical strength to interstitial collagens (types I and III). Inflammation interferes with the integrity of the fibrous cap by limiting the creation of new collagen by SMCs 43 and by stimulating the destruction of existing collagen fibers (Figure 2). Indeed, CD40L as well as IL-1 produced by T-cells induce macrophages to release interstitial collagenases, including MMP-1 44, MMP-8 45, and MMP-13 44 46, which mediate the initial attack on interstitial collagen. MMP-14, a membrane-associated MMP, activates MMP-13 and also appears to contribute to collagenolysis 47. Macrophage-derived foam cells contain MMP-9, a member of the gelatinase class of the metalloproteinase family, in human plaques 48. The catalytically active MMP-9 may contribute to the dysregulation of extracellular matrix that leads to plaque rupture. MMP activity overwhelms regulation mediated by TIMPs (tissue inhibitors of metalloproteinases) produced by macrophages and other cells in atheromata, hence influencing plaque stability 49 50. In addition, macrophages in lesions constitutively produce the serine protease neutrophil elastase, and release this enzyme upon CD40 ligation 51. By inactivating TIMP-1, neutrophil elastase favors MMP activity and collagen breakdown 51. Direct in vivo evidence in collagenase-resistant atherosclerosis-prone mice confirms the role of MMPs in plaque collagen turnover 52.
Cathepsins, members of the cysteine protease family, also participate in plaque evolution and destabilization 53. Macrophages in human atheromata express most of the proteolytically active cathepsins S and cathepsins K, which display elevated elastolytic activities 54. Experimental atherosclerotic lesions also express cathepsins L, and B 55. Cathepsin S 56, cathepsin L 57, or cathepsin K 58 deficiency in atherosclerosis-prone mice reduces collagen and elastin degradation as well as CD4+ T-cell, macrophage, and smooth muscle cell accumulation and overall plaque burden. Importantly, arteries under physiologic conditions abundantly express cystatin C, an endogenous inhibitor of cathepsins, whereas human atheromata exhibit very low cystatin C levels 59.
Recently, proteases of the ADAMTS (a disintegrin and metalloprotease with thrombospondin motif) family have also garnered attention for their role in matrix protein turnover and progression of atherosclerosis 60 61. The combined effects of all these classes of proteases, released mostly by macrophage foam cells within lesions, favors matrix and fibrous cap remodeling that may lead to plaque rupture with ensuing thrombosis and clinical manifestations 62.
Mast cells although numerically minor constituents of the leukocytic population in the atherosclerotic intima can also populate the adventita 63 64. In human coronary artery specimens, mast cell numbers rise in parallel to the severity of clinical presentation 65, and these leukocytes accumulate in the shoulder region of plaques, where they degranulate and release proteases and cytokines 66. Recent studies suggest they can contribute to lesion progression in mice (Figure 1). These data require cautious interpretation in regard to the human disease, as rodents rely more on innate immunity and possess a more complex variety of mast cell functions and proteases 67.
The chemokine eotaxin, found in atherosclerotic plaques, can mediate lesional mast cell recruitment by binding to cellular CCR3 68. Utilization of a mast cell-deficient mouse demonstrates decreased lesion size, lipid deposition, T-cell and macrophage numbers, and cell proliferation and apoptosis, but increased collagen content and fibrous cap development in a mast cell-derived IFN-γ and IL-6-dependent manner 69. Periadventitial injury leads to lesion progression and activation of mast cells with ensuing intraplaque hemorrhage, macrophage apoptosis, vascular permeabilization, and recruitment of further leukocytes 70. Importantly, treatment with the mast cell stabilizer chromoglycate prevents all these adverse events elicited by mast cell degranulation 70.
Mast cells produce certain MMPs, including MMP-1 71 and MMP-9 72. In addition, the mast cell serine proteases tryptase and chymase can activate MMPs in human carotid endarterectomy samples, and MMP-1 and -3 colocalize with degranulated mast cells in the shoulder regions of atherosclerotic plaques 73. Mast cell chymase also processes pro-MMP-2 and -9 into their active forms 74. These results suggest that the direct release of certain MMPs and the activation of MMPs by mast cell-derived proteases may promote atherosclerotic plaque rupture.
The proportion of intimal mast cells expressing basic FGF (fibroblast growth factor) — a potent angiogenic mediator — rises with increasing severity of atherosclerosis, suggesting a role for these leukocytes in angiogenesis, plaque neo-vascularization, and disease progression 75. Mast cell chymase also functions as an angiotensin-converting enzyme (particularly in rodents) and may thus contribute to the local regulation of vascular tone 76.
Activated mast cells induce endothelial death by chymase-mediated inactivation of focal adhesion kinase (FAK) and Akt-dependent cell survival signaling, as well as TNF-α-mediated apoptosis, functions that contribute to plaque erosion 77. Mast cell chymase can also inhibit smooth muscle cell proliferation and collagen expression 78 and induce SMC apoptosis 79 80. Histamine — abundantly present in mast cells — reportedly induces tissue factor expression in human aortic endothelial and vascular smooth muscle cells 81, promoting the plaque's thrombotic potential.
Degranulation of mast cells may facilitate SMC and macrophage uptake of LDL and their development into foam cells 82 83, mediated in part by chymase-dependent degradation of the ApoB moiety of LDL 84. In addition, mast cells inhibit cellular cholesterol efflux and reverse cholesterol transport, in part by chymase-mediated degradation of certain apolipoproteins such as ApoE 85 and tryptase-mediated degradation of HDL 86.
Nature killer (NK) cells, cellular effectors of innate immunity, play a critical role in the defense against infectious organisms, particularly viruses 87. Contrary to T-cells and B-cells, these bone marrow-derived lymphocytes do not require antigen receptor gene rearrangement during cellular development and do not express T-cell receptors or surface immunoglobulins. NK cells receive dual signals from inhibition and activation surface receptors 87. NK cell inhibitory receptors such as Ly49A are specific for MHC class I molecules on target cells and prevent NK cell activation, cytotoxicity, and cytokine secretion. Cells that have lost their expression of MHC-I molecules — typically virus-infected cells — are susceptible to NK cell attack. NK cells also express activation receptors such as Ly49D and Ly49H, structurally related to the inhibitory receptors, which recognize target cell ligands and can trigger perforin-dependent natural killing. In addition, NK cells express a receptor that binds the Fc portion of antibodies known as FcγRIII (CD16). Cross-linking of Fc receptors by IgG antibody-coated target cells may constitute a second form of activation that signals the NK cell to kill the target.
Limited direct evidence supports NK cell involvement in atherogenesis. The shoulder region of human plaques contains modest numbers of CD56+NK cells 88. Given their presence within lesions, and the abundance of cytokines known to activate NK cells such as IFN-α/β, IL-12, IL-15, and IL-18, these cells plausibly contribute to lesion progression. Upon activation, NK cells in turn secrete numerous cytokines and growth factors including IFN-γ, TNF-α, GM-CSF, and trigger perforin-mediated natural killing 87. Indeed, atherosclerosis-prone mice with genetically impaired NK inhibitory signaling have decreased plaque burden 89.
Neutrophils comprise a minority of the inflammatory cell composition of atherosclerotic lesions. However, neutrophil numbers rise in ruptured human coronary plaques 90, consistent with their role of endocytosing and clearing damaged tissue, but also raising the possibility of their presence before clinical events. In parallel, high circulating neutrophil counts predict myocardial infarction better than any other leukocyte subset, including total white blood cell, lymphocyte, or monocyte count 91. Intraplaque hemorrhage contributes importantly to the progression of atherothrombosis 92, and analyses of human carotid endarterectomy samples suggest intraplaque hemorrhage as an entry route for neutrophils — which constitute 60% of circulating leukocytes — into lesions 93. Atheromatous lesions demonstrate markers of neutrophil degranulation (α1-antitrypsin/elastase complexes, myeloperoxidase, and α-defensins) and the presence of proteases preferentially released by these leukocytes [NGAL (neutrophil gelatinase-associated lipocalin)/MMP-9 heterodimers, and HLE (human leukocyte elastase)], suggesting the presence of active neutrophils within lesions 93 (Figure 1). Atherosclerotic plaques contain NGAL that inhibits MMP-9 inactivation and thereby promotes its proteolytic and matrix-degrading capabilities 94. Another report confirms the colocalization of neutrophils with myeloperoxidase in lesions, suggesting a source for this enzyme beyond a macrophage subset 95. Myeloperoxidase generates the reactive oxygen species hypochlorous acid, which contributes to endothelial apoptosis and tissue factor expression and lesion advancement 96. In addition to mediating chlorination of tyrosyl residues, myeloperoxidase also leads to LDL protein nitration and lipid peroxidation, facilitating the uptake of these modified LDL particles by macrophages and contribution to foam cell formation 97.
Under physiologic circumstances, neutrophils arise exclusively in the bone marrow, and the chemokine ligand CXCL12 (SDF-1) expressed on bone marrow stromal cells and its receptor CXCR4 expressed on neutrophils allow not only their retention but also the homing of senescent neutrophils back to the bone marrow 98. Experimental antagonism of CXCR4 leads to increased circulating neutrophil levels and their enrichment within atherosclerotic lesions in response to CXCL1 expressed within lesions interacting with CXCR2 on neutrophils 99. Neutrophil recruitment also depends on the expression of neutrophilic Mac-1 and endothelial P-selectin for leukocyte rolling 100. These recruited neutrophils can secrete numerous enzymes including proteases. In addition, recruited neutrophils increase lesion size and enhance intraplaque IFN-γ, tissue factor, and CXCL1 levels, further amplifying their recruitment 99. These results suggest intraplaque hemorrhage and the more classic transendothelial route as sources of neutrophil recruitment, enriching the oxidative and proteolytic content and overall inflammatory activation of lesions.
Dendritic cells (DCs), an innate immune cell type, populate atherosclerotic plaques, particularly in the rupture-prone shoulder region of lesions 101, in part under the control of CX3CR1 102. Granulocyte-macrophage colony-stimulating factor (GM-CSF), produced locally in response to oxidized low-density lipoprotein cholesterol (oxLDL), also regulates DC numbers within lesions 103, probably by controlling their differentiation from monocyte precursors. Importantly, stimuli known to accelerate atherogenesis, such as oxLDL or TNF-α, increase DC adhesion to the endothelium and their subsequent transmigration 104. Nicotine increases DC expression of MHC-II, costimulatory molecules, and adhesion molecules, and promotes the production of IL-12 by DCs, thereby promoting a Th1 response 105. In addition, CD11c+ leukocytes with dendritic processes inhabit regions of the normal arterial intima predisposed to atherosclerosis 106. In human lesions, DC numbers increase in parallel to lesion complexity 107, as does the expression of CD83, a marker of DC activation 108.
Hypercholesterolemia may impede the emigration of a proportion of DCs, retaining them in peripheral tissues 109 110 where they locally promote immune responses, for example by re-stimulating effector CD4+ T-cells 111. Contrary to macrophages 112, DCs retain antigen presenting function under conditions typical of atherosclerotic plaques 111. This property may result from superior DC defenses against oxidative stress, displayed by elevated levels of superoxide dismutase and peroxiredoxin-1 113, and an apparent resistance to cholesterol-induced cytotoxicity, displayed by the absence of expression of the transcription factor CHOP (C/EBP-homologous protein), a marker of unfolded protein response (UPR) induction, after loading of DCs with unesterified cholesterol 111. Indeed, DCs embedded in ‘artificial arteries’ made of ECs, VSMCs, and type I collagen stimulate CD4+ T-cells upon exposure to lipopolysaccharide robustly and superiorly to embedded monocytes and macrophages 114.
DCs constitute a heterogeneous family, with different subsets characterized by varying tissue distributions, surface markers, cytokine profiles, and ensuing functions in the orchestration of immune responses 115. Plasmacytoid dendritic cells (pDCs) specialize in sensing bacterial and viral products and produce IFN-α abundantly 115. Through the release of key cytokines, PDCs contribute to the regulation of VSMC numbers within lesions. Upon microbial stimulation, pDCs release IFN-α, inducing tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) on the surface of CD4+ T-cells 116. TRAIL binds to death receptor 5 (DR-5) and thereby mediates VSMC death in an alternate pathway to the canonical CD8+ T-cell mediated cytotoxicity 117. Plasmacytoid-derived IFN-α also amplifies the inflammatory response by enhancing the production of TNF-α and IL-12 by ‘classic’ CD11c+ DCs 118. PDCs may thus provide a potential link between infections and disease progression, an association often evoked in atherosclerosis 119 120.
DCs prime T-cells in secondary lymphoid organs, enabling T-cell antigen-specific differentiation into effectors and the targeting of select tissues in the periphery such as atherosclerotic vessels. To emigrate toward regional lymph nodes, where they regulate adaptive immune responses, DCs must induce expression of CCR7 121 and CCR8 122 in tissues such as plaques.
Co-stimulatory and co-inhibitory molecule expression patterns by DCs drive antigen-dependent activation of naïve T-cells and the initiation of adaptive immunity (Figure 3) 123. Importantly, overlapping co-stimulatory and co-inhibitory molecules control both effector T-cell and regulatory T-cell responses, a complexity discussed elsewhere 123. Two major families comprise the costimulatory molecules: the B7 124 and TNF 125 families that bind to the CD28 and TNF receptor families, respectively 123.
CD80 and CD86 (B7-1 and B7-2), the prototypical and best described co-stimulatory molecules on DCs/APCs, initially deliver their signals by binding to CD28 on T-cells. CD80/CD86 deficiency in atherosclerosis-prone mice reduces lesion development and decreases IFN-γ production by CD4+ T-cells upon presentation of the atherosclerosis-associated antigen hsp60 126, suggesting inefficient priming by DCs. Following their activation, T-cells express CTLA-4 (Cytotoxic T-Lymphocyte Antigen 4). CTLA-4 then binds CD80 and CD86 on DCs, and turns the co-stimulatory signal initially delivered by CD28 into a co-inhibitory one, dampening the T-cell response 124.
Contrary to resting naïve T-cells, effector and memory T-cells express ICOS (inducible co-stimulatory molecule), another CD28 family member 127. ICOS-ligand activates ICOS — which figures critically in Treg function — on DCs, with decreased Treg suppressive function and increased lesional CD4+ T-cell content in ICOS-deficient mice 128.
Programmed death-ligand 1 (PD-L1) and PD-L2, B7 family members expressed on several cell types including DCs, inhibit T-cell activation by binding to programmed death-1 (PD-1) on T-cells 129. PD-L1/PD-L2-deficient mice have increased atherosclerotic burden in conjunction with exaggerated systemic immune responses with lymphadenopathy and elevated numbers of activated T-cells due to enhanced stimulation by DCs 130.
DCs and APCs express OX40Ligand (OX40L), a member of the TNF receptor family that provides co-stimulatory signals to T-cells through the TNF family member OX40, enabling long-lasting T-cell responses 125. Genetic studies identified OX40L as an atherosclerosis-susceptibility locus in mice 131. OX40L-deficient mice fed an atherogenic diet have decreased lesion size, whereas transgenic over-expression of OX40L demonstrates opposite effects 131. Antagonist anti-OX40L antibody treatment of atherosclerosis-prone mice also decreases plaque burden 132. Moreover, polymorphisms in the OX40L gene increase the risk of myocardial infarction 131, further highlighting the possible contribution of OX40L/OX40-mediated co-stimulation in promoting atherosclerosis.
CD137Ligand (CD137L), a TNF family member, also belongs to the TNF receptor family 125. The wide cellular distribution of these molecules complicates the elucidation of their contribution to atherosclerotic disease, but antigen recognition induces CD137 expression on T-cells that receive co-stimulatory signals from CD137L expressed on DCs 125. CD137L/CD137 participates importantly in CD8+ T-cell responses 133. Indeed, agonistic anti-CD137 antibody treatment increases CD8+ T-cell infiltration, pro-inflammatory cytokine expression, and overall lesion size 134.
DCs maintain their antigen-processing and -presentation functions and co-stimulatory capabilities under hypercholesterolemic conditions present in experimental atherosclerosis 111. DCs thereby activate the adaptive immune response and efficiently generate monoclonal and polyclonal effector CD4+ T-cells that may subsequently leave secondary lymphoid organs and reach atherosclerotic vessels 111.
IFN-γ-producing TH1 CD4+ T-cells with αβ T-cell receptors constitute the majority of T lymphocytes present in human 135 and experimental 136 atherosclerotic lesions 1 2 (Figure 1). T-cells enter lesions in response to the chemokines inducible protein-10 (IP-10), monokine induced by IFN-γ (MIG), and IFN-inducible T-cell α-chemoattractant (I-TAC), which bind CXCR3, highly expressed by T lymphocytes in the plaque 137. CD4+ T-cells undergo oligoclonal expansion within lesions, suggesting the occurrence of antigen-driven T-cell proliferation 138. Indeed, CD4+ T-cell clones in plaques recognize oxLDL 139 and hsp60 140.
Experimental evidence supports an important role for CD4+ T-cells in atherosclerosis. Immunodeficient RAG (recombinase activating gene)-deficient 141 or DNA-PK (DNA-dependent protein kinase)-knockout 142 atherosclerosis-prone mice have no lymphocytes and reduced lesions compared to immunocompetent atherosclerosis-prone mice. Adoptive transfer of CD4+ T-cells into immunodeficient animals, however, greatly increases lesion size in parallel to increased T-cell recruitment and MHC-II expression in plaques 142.
In addition to the histopathologic features of macrophage and T-cell accumulation within lesions, several lines of evidence support TH1 predominance in atherosclerosis. Different strains of mice have varying susceptibilities to atherosclerosis 143. TH1-biased C57BL/6 mice develop significantly more atherosclerosis in association with increased serum levels of IL-6 and the acute-phase protein serum amyloid A (SAA) than TH2-biased BALB/c mice 144. Genetic T-bet deficiency (a transcription factor required for TH1 differentiation) 145 or treatment with pentoxifylline (an inhibitor of TH1 differentiation) 146 protects against atherosclerosis. While administration of IL-12 147, the central cytokine driving TH1-cell differentiation, enhances atherosclerosis, genetic deficiency of IL-12 148 or IL-12 blockade through vaccination 149 attenuates the disease. In addition, the genetic deficiency 150 or inhibition 151 of IL-18, which drives TH1-cell differentiation synergistically with IL-12, also decreases disease progression. Deficiency of IFN-γ 152, the prototypical TH1-cytokine, or of the IFN-γ receptor 153, greatly decreases plaque burden, macrophage content, and MHC-II expression within lesions. Administration of recombinant IFN-γ has opposite effects 154. IFN-γ inhibits the proliferation and differentiation of vascular smooth muscle cells 155 and also decreases collagen production by these cells 43, functions that could contribute to the thinning of the collagen-rich fibrous cap (Figure 2). IFN-γ also inhibits endothelial cell proliferation 156 and potentiates the production of pro-inflammatory cytokines by macrophages as well as MHC-II expression 1. Further supporting these observations, human plaques contain an abundance of cells producing the TH1-type cytokines IFN-γ, IL-12, IL-15, IL-18, and TNF-α, but few cells producing the TH2 cytokine IL-4 157. In vivo, IFN-γ appears to augment arteriosclerosis, based on observations on the effects of this cytokine on small human arteries dwelling in immunodeficient mice 158.
Expression of CD40Ligand (CD40L) by activated CD4+ T-cells induces the expression of the procoagulant tissue factor in endothelial cells 159, VSMCs 160, and macrophages 161. CD40L also stimulates production of the interstitial collagenases MMP-1 44, MMP-8 45, and MMP-13 44 by macrophages, which interfere with the integrity of the protective fibrous cap (Figure 2). Recent results suggest that CD40L mediates certain effects in atherosclerosis through Mac-1 binding 162. CD40L on T-cells also activates APCs/DCs 123. Indeed, CD40 ligation increases co-stimulatory molecule expression by APCs, mainly CD80 and CD86, which then bind CD28 on the T-cell and transmit co-stimulatory signaling 163. In addition to T-cells, endothelial cells, VSMCs, macrophages, and platelets all express CD40L and CD40 164 165. As CD40 ligation promotes the inflammatory activation of all the major cell types participating in atherosclerosis, genetic 166 or antibody-mediated 167 disruption of CD40L signaling significantly decreases lesion size, macrophage and T-cell content, as well as VCAM-1 expression, and promotes the formation of a fibrous collagen-rich plaque.
CD8+ T-cells also populate lesions 168, but their relevance and contribution is less well understood 2. Conditions that enhance CD8+ T-cell function and increase their numbers by modulating co-stimulation 134 or co-inhibition 130 promote atherosclerosis in experimental settings. Though CD8+ T-cells kill VSMCs in aortic aneurysms 169, their role in atherosclerosis remains uncertain.
A subpopulation of T-cells that include Natural Killer T (NKT) cells and γδ T-cells express semi-invariant TCRs. NKT cells constitute a heterogeneous family of cells that has characteristics of both NK cells and conventional T cells 170. NKT cells express TCRs that recognize glycolipid antigens presented on CD1d, an MHC class I–like molecule expressed by antigen-presenting cells 170. The normal development of NKT cells also requires CD1d, found in human lesions 171. CD1d deficiency results in reduction of plaque burden in atherosclerosis-prone mice 172 173 174. Administration of a synthetic glycolipid that activates NKT cells via CD1d induces IFN-γ, MCP-1, TNF-α, IL-2, IL-4, IL-5, and IL-6 production within lesions 172. Lipid antigens present within plaques may activate NKT cells, which appear to participate in the early phases of atherogenesis 175.
Although the majority of T-cells express the classic αβ TCR, a subset expresses γδ TCRs with limited diversity. γδ T-cells represent fewer than 1-5% of circulating T-cells, but are enriched in sites of chronic inflammation 176. Importantly, γδ T-cells may recognize a wide array of antigens in the absence of MHC presentation by APCs 176. Some γδ T-cells recognize lipid antigens presented by CD1. These cells reside in lesions 177 and secrete IFN-γ and other cytokines, but the extent of their contribution to the pathobiology of atherosclerosis remains unclear 178.
Regulatory T-cells (Tregs) control other T-cell types and suppress their activation in secondary lymphoid organs, or their effector functions in peripheral tissues such as plaques, either directly or indirectly via actions on APCs (Figure 1). Tregs contribute to the maintenance of tolerance to self-antigens and the regulation of immunity. Of the different Treg subsets, CD4+CD25+FoxP3+ Tregs are the best characterized. 179 FoxP3 serves as a lineage-specific transcription factor involved in Treg suppressive function. These well-characterized ‘natural’ Tregs, generated during thymic development, comprise 5-10% of peripheral CD4+ T-cells. However, naïve CD4+T-cells induce additional Tregs during antigen-specific immune responses in the presence of IL-2 and TGF-β 180. Other surface molecules expressed by Tregs include CTLA-4, GITR (Glucocorticoid-Induced Tumor necrosis factor Receptor), and CD127. These antigen-specific Tregs inhibit effector T-cell activation either by direct contact inhibition or through the release of the anti-inflammatory cytokines transforming growth factor (TGF)-β and IL-10. These cytokines made by Tregs can mitigate atherogenesis in mice.
IL-10-deficient mice fed an atherogenic diet have increased lesion area, effector T-cell infiltration, and IFN-γ expression 181. IL-10-transfected mice demonstrate opposite results, with IL-10 also impeding the modified LDL-mediated endothelial recruitment of monocytes 182. Bone marrow transplantation of transgenic IL-10 overexpressing T-cells also decreases lesion size and inflammation 183.
Inhibition of TGF-β signaling using neutralizing antibodies 184 or recombinant soluble TGF-β receptors 185 accelerates the development of atherosclerotic lesions in ApoE-deficient mice and favors the development of lesions with increased monocyte and lymphocyte accumulation and decreased collagen content. TGF-β exerts its atheroprotective effect by dampening CD4+ T-cell effector function. Indeed, mice carrying dominant-negative TGF-β receptors on CD4+ T-cells (CD4dnTβRII) exhibit increased inflammation and a paucity of mature interstitial collagen fibers within vascular lesions 186 187. CD4dnTβRII effector T-cells inhibit the production of lysyl oxidase, the extracellular enzyme needed for collagen cross-linking, limiting collagen maturation in the atherosclerotic plaque while having little effect on collagen degradation 188. Although numerous cell types produce TGF-β and IL-10, including endothelial cells, smooth muscle cells, macrophages, and platelets, regulatory T-cells may constitute an antigen-specific source of these anti-inflammatory cytokines 189.
In mice, decreased Treg numbers 189 or Treg function secondary to absence of ICOS 128 leads to increased lesional CD4+ T cells and macrophage numbers as well as amplified expression of the pro-inflammatory cytokines IFN-γ and TNF-α. Moreover, patients with acute coronary syndromes have reduced circulating Treg numbers and suppressive function 190. Conversely, the adoptive transfer of Tregs 191 or Treg-induction secondary to measles virus nucleoprotein vaccination (known to induce immunosuppression) 192 inhibits macrophage and T-cell accumulation within lesions. In addition, leptin deficiency in atherosclerosis-prone mice reduces atherosclerotic lesion formation, in association with diminished Th1 responses in addition to a marked increase in the number and suppressive function of Tregs 193. Given that the adipokine leptin increases in obesity, the possibility that leptin limits Treg function presents an additional link between obesity and inflammation. Interestingly, deficiency of the chemokine CXCL10 (IP-10) decreases atherogenesis not only by diminishing the recruitment of CD4+ T-cells, but also by increasing intra-plaque Treg numbers 194.
Therapeutic modulation of Tregs has undergone scrutiny in experimental settings. Anti-CD3 antibody treatment decreases CD3/T-cell receptor complex expression and has immunosuppressive functions by increasing Treg numbers and TGF-β levels, thus reducing lesion initiation and progression 195. Induction of tolerance through the oral administration of oxLDL 196 or hsp60/65 197 to atherosclerosis-prone mice increases Treg numbers in secondary lymphoid organs as well as TGF-β and IL-10 levels, again attenuating experimental atherogenesis. These reports highlight the central role of Tregs in tolerance induction, which may constitute the mechanism of atheroprotection reported in earlier similar studies to hsp65 198 199 and β2-glycoprotein I 200.
B-cells infiltrate the adventitial layer of human coronary lesions 201 and atherosclerosis-prone mice 202 where they may form lymphoid follicles. The plaques of ApoE-deficient mice contain B-cells at all stages of the disease 203 (Figure 1). Splenectomized mice have increased susceptibility to atherosclerosis, a situation reversed by the transfer of B-cells and the production of anti-oxLDL antibodies 204.
Both humans and atherosclerosis-prone animals have antibodies against oxLDL particles 205 206. Germ-line encoded natural anti-oxLDL IgM antibodies produced by B1-cells bind the oxidized phospholipids on oxLDL and also recognize phosphorylcholine in the cell wall of Streptococcus pneumoniae 207. Taking advantage of this molecular mimicry, pneumococcal vaccination of atherosclerosis-prone mice decreases the extent of experimental atherosclerosis 208.
Anti-hsp60 antibodies cross-react between microbial and eukaryotic hsp60/65, a consequence of high sequence conservation. As such, infections, for example by Chlamydia pneumoniae, might result in breaking tolerance to self-hsp60 and promoting auto-immunity and atherogenesis 209. Indeed, experimental results using hsp65 as an immunogen for vaccination 210 211 mediate endothelial cytotoxicity 212 and promote atherogenesis. Over all, B-cells are considered to mediate protective immunity during the development of atherosclerosis, possibly by preventing antigens from reaching lesions.
The evidence reviewed here highlights the extensive role of innate and adaptive immunity in atherosclerosis, from its initiation to its final thrombotic complications. Our improved pathobiologic understanding of this disease allows the detection of patients at high risk and the design and development of novel treatment modalities targeting cellular and molecular mediators. These have already led to the identification of certain inflammatory indicators, such as CRP, as biomarkers of adverse cardiovascular events, allowing a better risk stratification of patients and targeting of therapy 213. In addition, statins have emerged as powerful anti-inflammatory and immunomodulatory agents, with extensive clinical use in primary and secondary prevention. Despite these strides, death from cardiovascular disease continues to increase worldwide, with many patients experiencing cardiovascular events despite statin, anti-platelet and anti-hypertensive treatment 214.
As a systemic and non-selective modulation of immune responses could lead to adverse effects including acquired immunodeficiency with ensuing infectious and oncologic complications, a more subtle and targeted approach of the immune response would probably prove advantageous. Based on the myriad immune mechanisms involved in atherogenesis, a wide range of potential therapies appear on the horizon, with the goal of diminishing the global cost paid by humanity to this scourge.
This work was supported by the Fondation Leducq, Paris, France (Dr. Libby) and by NIH HL087282 (Dr. Lichtman) & HL36436 (Dr. Libby).
René R. S. Packard, Leducq Center for Cardiovascular Research, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, U.S.A.
Andrew H. Lichtman, Division of Cardiovascular Medicine, Department of Medicine, and Immunology and Vascular Research Divisions, Department of Patholog, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, U.S.A.
Peter Libby, Leducq Center for Cardiovascular Research, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, U.S.A.