OUP user menu

★ editor's choice ★

Chemokines in the vascular inflammatory response of atherosclerosis

Alma Zernecke, Christian Weber
DOI: http://dx.doi.org/10.1093/cvr/cvp391 192-201 First published online: 9 December 2009


Atherosclerosis is considered to be a chronic inflammatory disease of the vessel wall that encompasses the accumulation of lipids, and it is critically shaped by the recruitment of leucocytes during all phases of the disease. In addition, the progression of atherosclerosis is determined by a disturbed equilibrium of immune responses. Chemokines and their receptors are instrumental in orchestrating the influx of leucocytes to the vascular wall, but also seem to regulate immune functions. Recent work has shed light on the apparent redundancy and the robustness of the chemokine system and has also provided evidence for its specialized role in the regulation of specific functions and trafficking of leucocyte subpopulations. This review will give a comprehensive summary to highlight those chemokines addressed in different models of atherosclerosis and vascular injury to date. In addition, we will discuss recent developments scrutinizing heterophilic interactions of chemokines that have advanced our understanding of how chemokines control vascular inflammatory responses.

  • Chemokines
  • Atherosclerosis
  • Inflammation
  • Leucocytes

1. Introduction

The vascular inflammatory response after injury or during atherosclerosis involves the mobilization and recruitment of immune cell subsets and progenitor cells governed by adhesion molecules and chemoattractants. Chemokines constitute a family of structurally related chemotactic cytokines that are classified into subgroups based on the position of the first two cysteine residues (CC, CXC, C, CX3C) and bind to correspondingly classified and often to multiple chemokine receptors. The expression of different chemokines and their cellular sources have been investigated in primary and diet-induced atherosclerosis, and in different models of vascular injury. Notably, a given chemokine can be expressed by different cell types, implying some functional overlap but also specialized functions.1

While the processes regulating the mobilization of cells from the bone marrow or the spleen are not fully understood, leucocyte recruitment to the vessel wall has been shown to involve the sequential and partially overlapping actions of distinct signal molecules. Namely, rolling interactions mediated by selectins and their carbohydrate interactions are followed by integrin-dependent leucocyte arrest and by transendothelial migration along a chemotactic gradient. These processes are regulated by chemokines and their respective receptor pairs in various ways. For instance, chemokines can alter the adhesiveness of integrins and support leucocyte arrest, either directly or involving their presentation by binding to proteoglycans. Chemokines can also form chemotactic gradients supporting diapedesis of cells into the vessel wall. In addition, they might also provide important anti-apoptotic survival cues to leucocytes, as shown for CCL5 or fractalkine.2

The development of atherosclerosis is initially characterized by endothelial-cell dysfunction and activation, which triggers the adhesion of leucocytes but also activated platelets to the endothelium and by an increased permeability for plasma lipid components, such as low-density lipoprotein (LDL).3,4 The recruitment of monocytes, their uptake of lipids, and their transformation into foam cells lead to intimal fatty-streak lesions. The continued cell accumulation and the subsequent apoptosis of plaque cells lead to the formation of a necrotic core and result in the progressive narrowing of the arterial lumen. Secretion of matrix proteases and cytokines by plaque cells can then trigger thinning of the fibrous cap and its disintegration with plaque erosion or rupture leading to thrombus formation and vascular occlusion underlying coronary syndromes, myocardial infarction, or stroke.

As opposed to primary atherosclerosis, the acute injury of the vessel wall comprises the acute endothelial denudation and platelet adhesion to extracellular matrix components, as well as a massive apoptosis of smooth muscle cells (SMCs) in the medial vessel wall. The accumulation of phenotypically unique SMCs within the intimal layer to restore the integrity of the arterial vessel wall subsequently leads to intimal hyperplasia and the progressive stenosis of the vessel.57 In a study of balloon-injury in cholesterol-fed rabbits, it was shown that the early intimal monocyte infiltration precedes the accumulation of SMCs, and suggested that early monocyte recruitment subsequently triggers a more sustained and chronic inflammatory response, possibly by the release of cytokines and growth factors, leading to the ongoing monocyte influx and SMC accumulation during neointimal growth.8

Different inflammatory leucocyte subsets populate different stages of atherosclerotic lesions and also participate in both innate and adaptive immune responses during atherogenesis, as epitomized by macrophage-expressed pattern-recognition and scavenger receptors and the T-cell-directed production of antibodies specific for modified lipids by B cells, respectively.4,9 Moreover, different subsets of vascular progenitor cells that can respond to local damage, for example, of endothelial cells and incipient regenerative homing signals can affect plaque composition, vascularization, and stability.9

While the functions of some chemokines discussed herein have been elucidated more then a decade ago, recent work has extended our understanding for the functional versatility and apparent redundancy but also the robustness of the chemokine system and has provided evidence for its specialized role in regulating the recruitment and functions of specific cell populations. These will be reviewed herein by highlighting chemokines in different models of atherosclerosis and vascular injury (Table 1), as well as recent developments that have yielded insight into how chemokines control the vascular inflammatory response.

View this table:
Table 1

Effect of chemokine or chemokine receptor modulation in atherosclerosis

ChemokineReceptorMouse model/modalityEffect on lesion size
CCL2CCR2Ccl2−/− Ldlr−/−
Ccr2−/− BM > apoE3-Leiden↓ Atherosclerosis9–11,14
Ccr2−/− apoE−/−↓ Neointimal hyperplasia15
CCL5CCR1Ccr1−/−apoE−/−↑ Atherosclerosis33
CCR3↔ Neointimal hyperplasia36
CCR5Ccr1−/− BM > Ldlr−/−↑ Atherosclerosis35
Ccr5−/−apoE−/−↔ Atherosclerosis31
↓ Atherosclerosis33
↓ Neointimal hyperplasia36
Ccr5−/− BM > Ldlr−/−↔ Atherosclerosis34
CXCL1CXCR2Cxcr2−/− BM > Ldlr−/−↓ Atherosclerosis41
CXCL8Anti-CXCL1 Ab in apoE−/−↑ Neointimal hyperplasia49
MIFCXCR2Anti-MIF Ab in apoE−/−↔ Atherosclerosis53
CXCR4Anti-MIF Ab in apoE−/− / Ldlr−/−↓ Neointimal hyperplasia60,61
MIF−/− Ldlr−/−↓ Atherosclerosis54
Anti-MIF Ab in Ldlr−/−↑ Regression55
CXCL12CXCR4Cxcr4−/− BM > apoE−/−
Cxcr4 degrakine BM > Ldlr−/− Cxcr4 antagonist treatment↑ Atherosclerosis46
Cxcr4−/− BM > apoE−/−↓ Neointimal hyperplasia63,70
Anti-CXCR4 Ab in apoE−/−
CX3CL1CX3CR1Cx3cl1−/−apoE−/−↓ Atherosclerosis81
Cx3cl−/−Ldlr−/−↓ Atherosclerosis81
Cx3cr1−/−apoE−/−↓ Atherosclerosis82,83
Cx3cr1−/− BM > apoE−/−↓ Atherosclerosis80
Cx3cl1−/−Ccr2−/−apoE−/−↓ Atherosclerosis84
CCL5 peptide inhibition in Cx3cl1−/−Ccr2−/−apoE−/−↓ Atherosclerosis85
CXCL16CXCR6Cxcl16−/−Ldlr−/−↑ Atherosclerosis101
Cxcr6−/− apoE−/−↓ Atherosclerosis100
CXCL19CXCR7apoE−/−arch > WT↓ Regression104
CXCL20+ Anti-CXCL19/ CXCL20 Ab
CXCL9CXCR3Cxcr3−/−apoE−/−↓ Atherosclerosis105
CXCL10CXCR3 antagonist NBI-74330 or CCR5/CXCR3 antagonist TAK-779 in Ldlr−/−↓ Atherosclerosis106,107
CXCL11Cxcl10−/−apoE−/−↓ Atherosclerosis112
CXCL4CXCR3BPeptide disrupting Cxcl4-Ccl5 heteromer interactions in apoE−/−↓ Atherosclerosis119
  • BM, bone marrow; >, transplantation; ↓, reduction; ↑, increase; ↔, no change.

2. CCL2 and CCR2 in vascular inflammation

Initial evidence for the contribution of chemokines in atherosclerosis was derived from two independent mouse models of atherosclerosis. Genetic deletion of CCL2 (monocyte chemotactic protein/MCP-1) or its receptor CCR2 in atherogenic LDL receptor-deficient (Ldlr−/−) or apolipoprotein E-deficient (apoE−/−) mice, respectively, protected from the development of atherosclerotic lesions concomitant with a reduction in macrophage infiltration.1012 A central role of the CCL2/CCR2 axis in the recruitment of monocytes into the vascular wall was further supported by an up-regulation of CCL2 in SMCs and of CCR2 on monocytes in the context of hyperlipidemia.13,14 Similarly, lesion development was retarded in apoE3-Leiden mice reconstituted with Ccr2−/− bone marrow but not in apoE−/− mice with established lesions,15 corroborating the importance of CCR2 on haematopoietic cells in early atherosclerotic plaque growth, whereas at later stages other recruitment signals may predominate.

An up-regulation of CCL2 was also found in medial SMCs after wire-induced injury of the carotid artery in hyperlipidemic apoE−/− mice, and it was observed that CCL2 could be immobilized on surface-adherent platelets,16 where it could promote monocyte arrest in flow. Thus, locally secreted CCL2 can be retained and presented by platelets adherent to the injured vessel wall, possibly via binding to proteoglycans.17 Interestingly, monocyte arrest on early atherosclerotic endothelium in uninjured carotid arteries of apoE−/− mice was not dependent on CCL2 but on keratinocyte-derived growth factor (KC/CXCL1) and CXCR2,18 suggesting a differential and distinctive contribution of CCL2 to monocyte arrest after injury, which may require binding to platelets adherent to the site of injury. In hyperlipidemic apoE−/− mice, antibody blockade of CCL2 diminished neointimal hyperplasia and macrophage infiltration to carotid arteries after injury.16 Under normolipidemic conditions, however, the role of CCL2/CCR2 is less established. Although monocyte accumulation was reduced after arterial cuff or stent placement, neointimal SMC content was decreased and macrophage content unaffected in Ccr2−/− mice or CCL2 antibody-treated rats after endothelial denudation.1922

3. CCL5 and distinct function of its receptors CCR1 and CCR5

Besides CCL2, other CC chemokines, e.g. CCL1 (I-309), CCL3 (macrophage inflammatory protein/MIP-1α), CCL4 (MIP-1β), and CCL5 (regulated on activation normal T cell expressed and secreted/RANTES) have been found to be present in atherosclerotic lesions.2325 CCL5 can be expressed in a number of different cell types, including monocytes/macrophages, T lymphocytes but also SMCs and can mediate the arrest and transendothelial diapedesis of monoytes/macrophages and T lymphocytes.1,26 In addition, CCL5 can be stored and released from α-granules by platelets and its deposition and immobilization on activated aortic endothelium or neointimal lesions constitutes an important mechanism, by which platelets contribute to exacerbation of lesion formation.2729 Accordingly, the administration of a peptidic CCL5 receptor antagonist (Met-RANTES) modulated the inflammatory process during atherogenesis and reduced atherosclerotic lesion formation.27,30

CCL5 binds to several receptors, namely CCR1, CCR3, and CCR5, and distinct functions of the receptors CCR1 and CCR5 have now been addressed. While genetic deletion of Ccr5 in apoE−/− mice does not reduce spontaneous formation of early atherosclerosis,31 Ccr5−/−apoE−/− mice are protected from diet-induced but also late native and advanced atherosclerosis. Notably, this was associated with a reduction in Th1-type immune responses.32,33 Ldlr−/− mice, on the other hand, reconstituted with Ccr5−/− bone marrow, showed only little change in lesion size but an improved plaque stability.34 In contrast, Ldlr−/− mice reconstituted with bone marrow deficient in Ccr1 but also Ccr1−/−apoE−/− mice displayed rather aggravated lesion formation.33,35 Similar results were obtained in models of vascular injury in apoE−/− mice. Whereas Ccr5 deficiency protected against neointima formation after arterial wire-injury, deficiency in Ccr1 did not affect neointimal hyperplasia, attributable to an increase in interferon (IFN)-γ masking deficiencies in cell recruitment.36 These differences might relate to different expression levels of this receptor on different cell types but also to distinct functional prerequisites. Selective receptor antagonists revealed that CCR1 rather than CCR5 mediated CCL5-induced arrest of monocytes, activated T cells or Th1 cells, displaying different receptor expression levels, whereas CCR5 supported spreading of these cells, and both CCR1 and CCR5 contributed to transendothelial migration towards CCL5.37 Moreover, it has been revealed that oligomerization of CCL5 was crucial for CCR1-mediated arrest under flow but not for CCR5-mediated transmigration.38 Interestingly, a distinct subpopulation (about 13%) of murine endothelial progenitor cells (EPCs) express CCR539 and the deletion of Ccr5 in apoE−/− mice was associated with increased EPC numbers, as a putative atheroprotective factor.32 The role of the third CCL5 receptor CCR3 in atherogenesis remains to be elucidated in detail.

4. CXCL1, CXCL8, and CXCR2 promote vascular inflammation

The CXC chemokines CXCL1 (KC or its human homologue growth-related oncogene/GRO-α) and CXCL8 (interleukin-8/IL-8), which bind and activate the receptor CXCR2, have been detected in monocytes/macrophages of atheromatous lesions.4042 Evidence for the involvement of this chemokine/receptor axis in atherosclerosis was obtained in studies using bone marrow chimeras, as mice deficient in CXCL1 or CXCR2 are not viable or extremely susceptible to infection, respectively. Indeed, Ldlr−/− mice transplanted with Cxcr2−/− bone marrow were protected from atherosclerotic lesion formation43 and it was further shown that CXCR2 is more important for macrophage accumulation in established lesions than its ligands CXCL1 alone.44 Beyond the recruitment of monocytes expressing CXCR2,45 these findings might also relate to an early attraction of neutrophils and the recently identified proatherogenic function of this cell type in the initiation of lesion formation.46,47 In addition, it has been found that CXCR2 can recruit CD14+ angiogenic early-outgrowth cells (previously termed EPCs) to injured vessels and that regenerating endothelial cells express CXCR2. By promoting endothelial recovery,48,49 CXCL1 was thus shown to protect from neointimal hyperplasia after arterial injury.50 Whereas infusion or mobilization of EPCs may improve endothelial regeneration and function, the angiogenic potential of CXC chemokines with an ELR motif as CXCR2 ligands51 may promote neovascularization of plaques and thus contribute to their progression and instability. In advanced lesions, these CXCR2-driven mechanisms clearly favour atherogenesis. In part, this may be ascribed to its recently identified function as a receptor for macrophage migration inhibitory factor (MIF).

5. MIF as a key factor in vascular inflammatory processes

MIF has emerged as a key factor in vascular processes giving rise to atherosclerosis. Different proatherogenic factors, including oxLDL, have been shown to induce the expression of MIF in endothelial cells, SMCs, and mononuclear cells/macrophages during the development of atherosclerotic lesions in humans, rabbits, and mice. Associated with the severity of atherosclerotic disease, the expression of MIF has been shown to correlate with increased intima-media thickening and lipid deposition in the aorta of mice, in advanced human carotid artery plaques and in rabbits fed an atherogenic diet.52 First evidence for the in vivo relevance of MIF in disease progression was obtained from apoE−/− mice fed a normal chow, revealing an impairment in the recruitment of macrophages during atherogenesis but an only slight reduction in the development of atherosclerotic plaque area in the aorta.53 Genetic deletion of Mif in Ldlr−/− mice retarded diet-induced atherogenesis.54 Importantly, treatment with a blocking MIF antibody resulted in a regression of established atherosclerotic lesions with a reduced macrophage and T-cell content.55 With the identification of CXCR2 and CXCR4 as functional receptors for MIF, it could be shown that MIF can mediate arrest and chemotaxis of monocytes and T cells. Moreover, through CXCR2 expressed on neutrophils, a weaker chemotactic activity of MIF also extends to neutrophils.55 In light of a role of this cell type in early lesion formation46 and the capacity of neutrophils to trigger subsequent monocyte recruitment,56 MIF might thus also contribute to atherosclerosis by attracting neutrophils. Besides these chemokine-like functions, part of the effect on cell recruitment may also be due to the induction of CCL257 and other inflammatory mediators, such as adhesion molecules and TNF, by MIF,58 aggravating and sustaining inflammatory cell influx. In addition, genetic deletion of Mif was associated with a reduction in the expression of matrix metalloproteinases (MMPs) and cathepsins in vascular SMCs, and MMP-1 and MMP-9 in vulnerable plaques,54,59 so that MIF may contribute to collagen degradation and plaque destabilization.

The role of MIF has also been addressed in models of arterial injury. Following experimental angioplasty by carotid artery dilation in Ldlr−/− mice, antibody blockade of MIF inhibited neointima formation, reduced inflammation, and cellular proliferation but increased apoptosis.60 After wire-injury of carotid arteries in apoE−/− mice, MIF expression was up-regulated in SMCs early after endothelial denudation but at later stages was predominantly expressed in ECs and macrophage-derived foam cells. Whereas only a slight reduction in neointimal area was observed, neutralization of MIF after injury led to a marked reduction in neointimal macrophage content and inhibited their conversion into foam cells. Conversely, the content of SMCs and collagen was increased in the neointima,61 reflecting a remarkable shift towards a more stable plaque phenotype. This might also be related to an MIF-induced migration and proliferation of vascular SMCs.54,62 In this regard, it is interesting to note that the MIF receptor CXCR4 has been implicated in neointimal hyperplasia and the neointimal recruitment of SMC progenitor cells63 and thus might in part also relate to an MIF-induced proliferation of SMCs in neointima formation and restenosis.

6. Differential roles of CXCL12/ CXCR4 in atherosclerosis vs. arterial injury

The essential role of CXCL12 (stromal cell derived-factor-1/SDF-1α) in haematopoiesis, stem cell mobilization, bone marrow engraftment, organ development, and angiogenesis is well documented. In human atherosclerotic plaques, the expression of CXCL12 has been described in SMCs and ECs. Moreover, plasma levels of CXCL12 were documented to be decreased in patients with stable and (to an even greater degree) unstable angina compared with healthy controls.64 Recent genome-wide-association studies have further implied protective functions of this chemokine by identifying a variation on chromosome 10q11 near the CXCL12 gene as a powerful predictor for the susceptibility coronary atherosclerosis.65 Taken together, these data indicate that CXCL12 can exert beneficial, plaque-stabilizing effects in coronary disease. In parallel, the function of CXCL12 and its receptor CXCR4 in murine atherosclerosis has only recently been addressed. Interfering with the CXCL12/CXCR4 axis by a small molecule antagonist, genetic deficiency or lentiviral sequestration of CXCR4 in bone marrow chimeras aggravated diet-induced atherosclerosis in apoE−/− or Ldlr−/− mice and, strikingly, increased the content of neutrophils in the plaque.46 This appeared to result from a disturbed haematopoiesis leading to increased neutrophil mobilization into peripheral blood. Depletion of circulating neutrophils in mice attenuated diet-induced plaque formation and prevented disease exacerbation by Cxcr4 interference, for the first time revealing a functional contribution of this cell population to atherogenesis.46 This might also concur with the recently identified presence of neutrophils in human unstable plaques at sites of rupture and erosion or in extracted thrombi of patients with acute coronary syndromes,4,66,67 and with earlier observations that neutrophil counts are predictive of coronary syndromes and disease severity.68,69 Transiently increased plasma levels of CXCL12 and an early and sustained up-regulation of CXCL12 have also been described in SMCs following mechanical injury in different models.63,70,71 In medial SMCs, CXCL12 has been found to be induced after vascular injury in the context of apoptosis and to mobilize and recruit progenitor cells from the bone marrow giving rise to neointimal SMCs.63 Similarly, neutralization of CXCL12 inhibited the mobilization of blood-borne stem cells giving rise to lesional SMCs in a model of transplant arteriosclerosis, thereby attenuated neointima formation.63,70 While the accumulation of SMCs might thus augment neointima formation after injury, it can be speculated that the recruitment of SMC progenitor cells to atheromatous lesions might contribute to a more stable and SMC-rich plaque phenotype. Indeed, treatment with endothelial apoptotic bodies delivering microRNA-126 resulted in up-regulation of CXCL12 expression through activation of an autocrine CXCR4 signalling loop and thereby mediated atheroprotective effects with increased SMC content.72

CXCL12 can effectively activate platelets in vitro73 but is also expressed by platelets adhering to luminal sites of arterial injury,63,74 triggering the arrest of SCA-1+ or CD34+ progenitor cells at sites of endothelial denudation.63 However, both neointima formation after arterial injury in apoE−/− mice and in transplant arteriopathy displayed only minor effects on re-endothelialization, indicating that effects on smooth muscle progenitor cells might overweigh and be more functional relevant for effects of CXCL12/CXCR4 in neointimal growth.63,75 Likewise, CXCL12 expression was increased in endothelial nitric oxide synthase-deficient mice and likewise correlated with progenitor cell mobilization and adventitial recruitment in a carotid ligation model of neointima formation.76

7. CX3CL1/CX3CR1 in monocyte recruitment and survival

CX3CL1 (fractalkine) is a structurally distinct chemokine fused to a transmembrane mucin stalk.77 While the transmembrane protein acts as an efficient adhesion molecule for monocytes and T cells on activated endothelium, cleavage of the mucin stalk can produce a soluble form of CX3CL1 with chemoattractant activity. Expressed on the endothelium, CX3CL1 might primarily function to induce leucocyte arrest, whereas expression by lesional SMCs with cleavage of the pro-migratory soluble form by metalloproteinases might function to support macrophage and T cell recruitment to growing lesions.78,79 Furthermore, CX3CL1 has recently been identified to act as an important survival factor for monocytes/macrophages in atherosclerotic plaques.80

A central role of CX3CL1 was confirmed in several murine models of atherosclerosis. Whereas Cx3cl−/−apoE−/− mice displayed decreased atherosclerotic lesion formation in the brachiocephalic artery but not the aortic root, Cx3cl−/−Ldlr−/− mice showed less atherosclerosis at both sites with significantly fewer lesional macrophages in both models.81 The function of the CX3CL1 receptor CX3CR1 has been the focus of two independent studies, both showing that Cx3cl1−/−apoE−/− mice displayed a significant reduction in macrophage recruitment to the vessel wall and decreased atherosclerotic lesion formation.82,83 Recently, triple-knockout mice revealed that combined deletion of CCR2 and CX3CL1 leads to a dramatic reduction of atherosclerotic lesions in Cx3cl1−/−Ccr2−/−apoE−/− mice, providing the first in vivo evidence for independent roles for CX3CL1 and CCR2 in direct monocyte recruitment to atherosclerotic lesions.84 Moreover, inhibition of CCL2, CX3CR1, and CCR5 in apoE−/− mice was associated with a marked and additive reduction in atherosclerosis, suggesting that also CX3CR1-, CCL2-, and CCR5-mediated signals play independent and additive roles in atherogenesis.85 In line with an important function of CX3CL1 for monocyte/macrophage survival in atherosclerosis, the Bcl-2-enforced survival of monocytes and plaque-resident phagocytes, including foam cells, restored atherogenesis in Cx3cr1−/− mice,80 indicating that early monocyte apoptosis can critically delay atherosclerotic lesion formation.

In CAD patients, a higher rate of CX3CR1+ cells could be observed compared with healthy individuals.86 The importance of CX3CL1/CX3CR1 in human atherosclerosis is further supported by two common polymorphisms, V249I and T280M, which were associated with inter-individual differences in susceptibility to atherosclerosis. While one study identified V249I to be associated with a lower risk of CAD, implicating CX3CR1 in cardiovascular disease, another study failed to unravel an association of V249I or T280M with CAD but revealed protective effects of T280M during acute coronary syndromes.8789 In addition, T280M but not V249I has been associated with a decreased common carotid artery intima-media thickness.90 Conversely, a study conducted in 365 patients undergoing coronary stenting revealed a link between CX3CR1 polymorphisms and an elevated risk of restenosis.89

Taking the recently appreciated role of neutrophils and endothelial dysfunction into account,9 CX3CL1 shed from ECs after hypoxia/re-oxygenation has been shown to act through CX3CR1 on ECs to increase ICAM-1 expression and promote neutrophil adhesion by activating the Jak-Stat5 pathway.91 Beyond direct functions in monocyte recruitment or survival, CX3CL1 contributes to vascular dysfunction by stimulating the release of vascular reactive oxygen species resulting in reduced nitric oxide bioavailability in isolated rat aortas.92

8. Chemokines differentially control recruitment of monocyte subsets

Circulating monocytes display a remarkable heterogeneity in surface receptors, which is possibly associated with distinct yet unresolved functional contributions. Monocytes can be divided into at least two subsets. Ly-6Chigh (also termed Gr-1hi) and CCR2+CX3CR1low expressing cells can be discriminated from Ly-6Clow (Gr-1lo) and CCR2CX3CR1high monocytes in mice, corresponding to human CD14++CD16 and CD14+CD16+ monocytes, respectively.93,94 While the exact mechanisms are still unclear, Ly-6Chi monocyte numbers dramatically increase in hypercholesterolemic apoE−/− mice consuming a high-fat diet, associated with an increased survival, continued cell proliferation, and impaired Ly-6Chi to Ly-6Clo conversion.95 The combined inhibition of CCL2, CCR5, and CX3CR1 in apoE−/− mice abrogated bone marrow monocytosis and additively reduced circulating monocyte numbers and atherosclerosis despite persistent hypercholesterolemia. Of note, lesion size correlated with circulating monocyte numbers, particularly the Ly6Clo subset and strikingly, these chemokine signals together accounted for most of the macrophage accumulation in atherosclerotic arteries.85

The signalling routes employed by the two monocyte subsets to enter atherosclerotic plaques have been subject to further analysis. While CCR5 was selectively up-regulated in CCR2Ly-6Clo monocytes and mediated the accumulation of this subset in atherosclerotic lesions, CX3CR1—although highly expressed in Ly-6Clo monocytes—was not required for plaque entry. By contrast, CCR2+Ly-6Chi monocytes employed CX3CR1 together with CCR2 and CCR5.96 Whereas Ly-6Chi cells selectively populate sites of experimentally induced inflammation, their Ly-6Clo counterparts can enter lymphoid and non-lymphoid tissues under homeostatic conditions. The analysis of migration and differentiation properties of these two subsets in apoE−/− mice revealed that CCR2+Ly-6Chi monocytes efficiently accumulated in atherosclerotic plaques, whereas CCR2Ly-6Clo monocytes were rather poorly recruited. This finding may offer valuable therapeutic options based on antagonists selectively targeting CX3CR1-dependent plaque entry without blocking CCR2-dependent inflammatory responses.

9. CXCL16 mediates leucocyte recruitment and scavenger functions

As a rather atypical chemokine present in atherosclerotic lesions, CXCL16 (scavenger receptor for phosphatidylserine and oxidized lipoprotein SR-PSOX) is expressed by macrophages, dendritic cells, T cells, but also cytokine-stimulated SMCs and ECs in murine but also human atherosclerotic lesions.97 In addition, it has been shown that low plasma levels of CXCL16 are associated with CAD.98

When cleaved from the cell membrane by metalloproteinase ADAM10, CXCL16 can function as a chemokine. The CXCL16 receptor CXCR6 is expressed on a subset of interstitial lymphocytes, NK T cells, monocytes, dendritic cells, a subset of CD4+ or CD8+ T cells and CD4+ effector memory T cells, NK T cells, and a subset of Foxp3+ regulatory T cells in humans.99 Genetic deletion of CXCR6 results in a reduction in atherosclerosis and was associated with a lower content of CXCR6+ T cells and macrophages and a diminished expression of lesional IFN-γ.100 This is in sharp contrast to findings addressing the role of CXCL16 in lesion formation. Atherosclerotic lesion formation is enhanced in Cxcl16−/−Ldlr−/− mice, and associated with an increased macrophage recruitment.101 These differences might relate to its additional function as a scavenger receptor for apoptotic cells, phosphatidylserine and oxLDL.99 Thus, CXCL16 appears to overall confer atheroprotection,101 which prevails over its potential pro-atherogenic role as a chemoattractant for CXCR6+ cells.100

10. Immunoregulatory chemokines CCL19/CCL21 in vascular inflammation

Compared with healthy controls, CCL19 and CCL21 are both increased in plasma and within carotid atherosclerotic lesions of patients with stable and even further in those with unstable CAD. Similarly, both chemokines are detectable in atherosclerotic plaques from apoE−/− mice.102 Findings on the effect of these chemokines or their receptor CCR7 in atherosclerosis are ambivalent. On the one hand, lesional T cells display high levels of CCR7 and CCL19/CCL21 promote an inflammatory phenotype in T cells and macrophages, as well as increased MMP and tissue factor levels, which would be consistent with pro-atherogenic, plaque-destabilizing, and pro-thrombotic functions.102 On the other hand, studies applying CCL19/CCL21 have demonstrated an inhibition of T-cell proliferation and IL-2 secretion in murine and human CCR7+ T cells, implying possible atheroprotective functions.103 In a mouse model of atheroregression, yet again, transplantation of the aortic arch of apoE−/− mice with established atherosclerotic lesions into wild-type recipients during the combined blockade of CCL19 and CCL21 prevented plaque regression or a reduction of foam cell content.104 Overall, a complex and regulatory role of CCL19/CCL21/CCR7 in the interplay of immune cells during vascular inflammation with an importance for plaque regression can be deduced. Clearly, future studies are warranted to dissect the precise roles of CCL19/CCL21 and CCR7 in atherosclerotic lesion formation.

11. CXCR3 and its ligands regulate T-cell responses in atherosclerosis

CXCR3, expressed on activated T helper type 1 cells, functions as a common receptor for CXCL9, CXCL10, and CXCL11. Among the CXCR3 ligands, CXCL10 (IFN-γ-induced protein of 10 kDa/IP-10), CXCL9 (monokine-induced by IFN-γ/MIG), and CXCL11 (IFN-γ-inducible T cell α-chemoattractant/ITAC) are highly expressed in human atheromas throughout all stages of plaque development.42 Deficiency in CXCR3 has been shown to reduce lesion formation in apoE−/− mice and was associated with an up-regulation of the anti-inflammatory molecules IL-10, IL-18BP and endothelial nitric oxide synthase, and increased numbers of regulatory T lymphocytes within lesions.105 Similar results were obtained in a recent study employing a specific CXCR3 antagonist NBI-74330.106 Besides interfering with CXCR3+ effector T cell but also macrophage recruitment into plaques and modulating the local inflammatory response with an enhancement of markers for regulatory T cells in Ldlr−/− mice, lymph nodes draining from the aortic arch were smaller with fewer activated but more regulatory T cells.106 Moreover, treatment of Ldlr−/− mice with the HIV entry inhibitor TAK-779 antagonizing both CCR5 and CXCR3 reduced atherosclerotic plaque area, T-cell numbers, and IFN-γ content.107 These data provided evidence that CXCR3 crucially affects the recruitment and responses of T cells in atherosclerosis.

Of the CXCR3 ligands, only CXCL10 has been subject of further analysis in atherosclerosis. Constitutively expressed in lymphoid tissues, CXCL10 expression can be strongly up-regulated by IFN-γ in monocytes, macrophages, endothelial cells, and SMCs.108,109 Evidence for an important function were gathered from clinical studied, showing that higher CXCL10 plasma levels correlated with restenosis after percutaneous coronary intervention. In addition, an increased expression of CXCL10 and IFN-γ have been detected in patients with CAD.110,111 In mice, it could be shown that the genetic deletion of CXCL10 resulted in a reduction in aortic lesion formation and a marked reduction in CD4+ and CXCR3+ T cells but elevated numbers of regulatory T cells and increased IL-10 and transforming growth factor (TGF)-β1 expression.112 While the role of CXCL9 and CXCL11 awaits further elucidation, the CXCL10/CXCR3 axis clearly seems to play an important role in regulating T cell responses during atherogenesis.

12. Functional chemokine heteromerization—CXCL4 and CCL5

A splice variant of CXCR3 named CXCR3b has been identified as a receptor of CXCL4 (platelet factor 4/PF4) on endothelial cells.113 In addition, CXCL4 can bind to glycosaminoglycans and LDLR,113,114 which are present in fatty streaks and atherosclerotic lesions in humans, and its expression correlates with the histological and clinical severity of disease.115 CXCL4 is among the chemokines abundantly expressed in platelets and released in high concentrations upon platelet stimulation. In two mouse models of atherosclerosis, the absence of CXCL4 or CCL5 resulted in a significant reduction of lesion formation, clearly demonstrating a pro-atherogenic functions.115,116 However, CXCL4 by itself does not exert classical chemokine functions below micromolar concentrations,117 so that its role in monocyte recruitment might be auxiliary118 rather than autonomous. CCL5 and CXCL4 can form a heteromeric complex, which can occur in α-granules of human platelets.119 This functional complex of CCL5 and CXCL4 is a prototypic example for an emerging variety of heterophilic interactions between chemokines. To date, 13 chemokines have been identified to engage in heterophilic interactions.120,121 Taking the current body of evidence for chemokine heteromerization into account, it is safe to postulate a chemokine ‘interactome’, i.e. the collective of heterophilic chemokine interactions, as a novel regulatory principle in the regulation of leucocyte responses. Acting in concert with presentation on GAGs, chemokine heteromerization might assist to create a context-dependent and tailor-made repertoire of cues in order to direct the actions of specific leucocyte subsets in a particular microenvironment during health and disease. Beyond the use of modified chemokines, the manipulation of heterophilic chemokine interactions could represent a novel and promising approach for the treatment of inflammatory disease.

In a recent study, the importance of CCL5 and CXCL4 heteromerization for atherosclerosis was revealed. Cyclic peptides that interfered with the complexation of CCL5 and CXCL4, named CKEY2 (human) and MKEY (mouse), inhibited the CXCL4-mediated synergistic enhancement of CCL5-induced monocyte recruitment to inflamed endothelium, while leaving CCL5-related functions per se unaffected. Administration of MKEY in hyperlipidemic mice on a high-fat diet resulted in a marked reduction of atherosclerotic lesion formation. Effective inhibition with MKEY required the presence of both genes encoding CCL5 and CXCL4 in mouse blood cells, confirming the specificity of the approach. Beyond revealing the in vivo relevance of chemokine heteromerization, this study highlights the therapeutic feasibility of breaking functional ties between chemokines.119

13. Concluding remarks: concepts of chemokine function

Contemplating the vast amount of evidence on the involvement of chemokines in the vascular inflammation response and atherosclerosis, several alternative and complementary concepts emerge to explain the apparent redundancy and the remarkable robustness of the chemokine system in vascular disease (Figure 1). First, cell-specific effects of distinct chemokine receptors, as exemplified by the role of CCR5 in recruitment of Gr-1lo patrolling monocytes96 to plaques or by the role of CXCR6 in activated T-cell infiltration,100 can explain a complementary action of different chemokine-receptor axis in the complex mononuclear cell infiltration of plaques. A division of labour reflecting specialized roles exerted by different chemokines and receptors at distinct steps of leucocyte recruitment has been demonstrated for the arrest function of CXCR2 and its ligand CXCL1 and CCR1 vs. the preferential function of CCL2/CCR2 in subsequent transmigration.2,18,37 In contrast, a possible concerted action of different chemokine-receptor axis on one cell type is illustrated by the combined effects of CCR2, CCL5, and CX3CL1 on the mobilization of inflammatory Gr-1hi monocytes, which correlates with atherosclerosis.84,85 The introduction of heterophilic chemokine interactome opens a new regulatory dimension to explain the functional diversity and plasticity of chemokines but also allows for selective targeting without side effects.120 In particular, disrupting interactions of the prototypic heteromer CCL5-CXCL4 by a cyclic peptide resulted in a marked reduction of atherosclerosis and a more stable plaque phenotypes in apoE−/− mice.119 With regards to possible stage-specific effects of chemokines during atherogenesis, CCL5 was found preferentially up-regulated in early plaques,122 whereas CXCL1 and CX3CL1 were up-regulated beyond initial stages in plaques caused by low shear stress and blocking CX3CL1 inhibited plaque growth.123 Likewise, stage-specific effects on atherosclerosis have been observed for preferential effects of CCR2 in the aortic root and for CXCR3 in early-type abdominal lesions.105 Finally, atheroprotective effects have been identified for the CXCL12/CXCR4 axis and CXCL16. Together, these mechanisms harbour ample options for effective therapeutic targeting of chemokines in vascular inflammation.

Figure 1

Concepts of chemokine functions. Chemokines can exert cell-type specific effects: while Gr-1high monocytes displaying only low levels of CX3CR1 are recruited via this receptor, Gr-1low monocytes use CCR5 to enter atherosclerotic plaques; CXCR6 yet again predominantly attracts T cells to lesion (A). Atherogenic mononuclear cell recruitment requires different receptor cues at different sites of lesion formation (B) and at different stages of lesion formation (C). CCR5 and its ligand CCL5 are more important in early lesion formation (C). The chemokine receptors CCR2, CCR5, and CX3CR1 independently and additively mobilize monocytes from the bone marrow (D). On the other hand, different receptors can mediate specialized functions during distinct steps of recruitment (E). Whereas most chemokines and their receptors promote inflammation, CXCR4 controls the homeostasis of neutrophils and CXCL16 functions as a scavenger receptor on macrophages, and both can thereby confer atheroprotective effects (F).

Conflict of interest: C.W. and A.Z. are shareholders of Carolus Therapeutics, Inc.


This work was supported by the Deutsche Forschungsgemeinschaft (FOR809).


View Abstract